r google-plus facebook twitter linkedin2 nujij P P M G W Monitor Nieuwsbrief pdclogo man met tas twitter boek

COVER NOTE

1.

Tekst

Council of the European Union

Brussels, 30 April 2018 (OR. en)

8413/18 ADD 1

Interinstitutional files: 2018/0112 (COD)

MI 295 COMPET 250 DIGIT 77 IND 108 TELECOM 102 PI 42 AUDIO 27 JUSTCIV 97 CODEC 649 IA 111

COVER NOTE from: Secretary-General of the European Commission,

signed by Mr Jordi AYET PUIGARNAU, Director

date of receipt: 26 April 2018

to: Mr Jeppe TRANHOLM-MIKKELSEN, Secretary-General of the Council of the European Union

No. Cion doc.: SWD(2018) 138 final PART 1/2

Subject: COMMISSION STAFF WORKING DOCUMENT IMPACT ASSESSMENT Accompanying the document Proposal for a Regulation of the European Parliament and of the Council on promoting fairness and transparency for business users of online intermediation services

Delegations will find attached document SWD(2018) 138 final PART 1/2.

________________________

Encl.: SWD(2018) 138 final PART 1/2

EUROPEAN COMMISSION

Brussels, 26.4.2018 SWD(2018) 138 final

PART 1/2

COMMISSION STAFF WORKING DOCUMENT

IMPACT ASSESSMENT

Accompanying the document

Proposal for a Regulation of the European Parliament and of the Council

on promoting fairness and transparency for business users of online intermediation services

{COM(2018) 238 final} - {SEC(2018) 209 final} - {SWD(2018) 139 final} Table of Contents

1 INTRODUCTION ................................................................................................................... 2

2 PROBLEM DEFINITION ....................................................................................................... 9

3 WHY SHOULD THE EU ACT? ........................................................................................... 31

4 OBJECTIVES: WHAT SHOULD BE ACHIEVED? ........................................................... 32

5 WHAT ARE THE AVAILABLE POLICY OPTIONS? ....................................................... 36

6 WHAT ARE THE IMPACTS OF THE RETAINED POLICY OPTIONS? ......................... 45

7 COMPARISON OF OPTIONS ............................................................................................. 66

8 PREFERRED OPTION ......................................................................................................... 76

9 MONITORING AND EVALUATION OF IMPACTS ......................................................... 83

1 I NTRODUCTION

1.1 Political Context

In its 2015 Digital Single Market Communication 1 the Commission stressed that 'the market power of some online platforms potentially raises concerns, particularly in relation to the most powerful platforms whose importance for other market participants is becoming increasingly critical'. Following a public consultation as well as in-depth research, the Commission presented an overall assessment of the opportunities and challenges in the online platforms environment in a Communication in May 2016 2 . The Communication recognised the value added of platforms as drivers for innovation and growth in the digital economy. Platforms play an important role in the development of the online world and create new market opportunities, notably for SMEs. Online platforms increase consumer choice in terms of products and services set at a competitive price, thereby enhancing consumer welfare. At the same time, the Communication identified a series of concerns relating to potentially harmful trading practices in relations between platforms and their professional users, and announced a more detailed assessment thereof.

Initiated with an open public consultation closed in Spring 2016, this fact-finding exercise included notably (i) a survey completed by 3,549 businesses users of online platforms, complemented by 50 in-depth interviews and several in-depth case studies 3 ; (ii) a study on the terms and conditions of online platforms (iii) a study on issues related to data access in the platform-to-business relations, (iv) workshops with business users of online platforms as well as with online platforms, (v) a significant number of bilateral discussions with stakeholders including online platforms themselves, but also with civil society, (vi) focus groups with business users and with online platforms for options design, as well as (vi) internal research on the legal and economic aspects of online platforms and their business-to-business (B2B) practices. The Commission also organised a stakeholder workshop bringing together online platforms and business associations representing them, which addressed the practices reported during the B2B fact-finding exercise.

In its mid-term review of the Digital Single Market Strategy in May 2017 (Mid-Term Review) 4 , the Commission identified the promotion of fairness and responsibility of online platforms as an area where further action is necessary to ensure a fair, open and secure digital environment. The Commission therefore, committed to 'prepare actions to address the issues of unfair contractual clauses and trading practices identified in platform-to-business relationships, including by exploring dispute resolution, fair practices criteria and transparency. These actions could, on the basis of an Impact Assessment and informed by structured dialogues with Member States and stakeholders, take the form of a legislative instrument. This work will be finalised by the end of 2017. The Commission will also continue to use its competition enforcement powers wherever relevant.'

The European Parliament welcomed the Commission's fact-finding exercise and initiative on B2B practices. It expressed concerns about a series of practices and called on the Commission 'to propose a pro-growth, pro-consumer, targeted legislative framework for B2B relations based on the principles of preventing abuse of market power and ensuring that platforms that serve as a gateway to a downstream market do not become gatekeepers.' 5 The European Economic and Social Committee noted that online platforms benefit from a strong first-mover advantage, and

1 COM(2015), 192 final .

2 COM(2016), 288 final .

3 ECORYS 2017.

4 COM(2017), 228 final .

5 2016/2276(INI) Report of the European Parliament on Online platforms and the Digital Single Market, 31 May 2017.

that those exploiting network effects can become an unavoidable trading partner for businesses. 6 On the same topic, the Committee of the Regions stressed that 'early action to prevent fragmentation in the first place would still be far less difficult than ex-post harmonisation of 28 national framework'. 7

In his Letter of Intent accompanying the 2017 State-of-the-Union address, President Juncker reiterated the Commission's mid-term review commitment as a key part of the Digital Single Market initiatives, announcing an 'Initiative on Online Platforms to safeguard a fair, predictable, sustainable and trusted business environment in the online economy 8 '. In the Commission's Roadmap for a more United, Stronger, and more Democratic Union for the European Council in Tallinn this initiative was announced for 20 December 2017 9 . In response to these announcements, the European Council of October 2017 called for 'increased transparency in platforms’ practices and uses'. 10

This Impact Assessment follows up on the Commission's commitment in the Mid-Term Review. The objective of this initiative is to maximise the potential of the highly beneficial online platform ecosystems. To this end, this Impact Assessment assesses options to improve predictability and redress possibilities for EU business users that trade on online platforms, whilst maintaining an innovation-friendly environment without unnecessary burden for online platforms. In order to achieve this balance, the retained options all incorporate an important staged approach, focusing on transparency and bilateral conflict resolution in a first step, subject to transparency-enabled monitoring.

1.2 Legal context

There is no specific legislation at EU level addressing platform-to-business relationships. The initiative which would stem from this Impact Assessment would be the first action at EU level specifically targeting commercial contracts between online platforms and their business users.

EU Competition law , on the one hand, focuses on anticompetitive behaviour and mergers. The EU antitrust rules tackling anticompetitive behaviour are enforced on case-by-case basis ex post, prioritising inter alia those cases with a potential impact beyond the case itself. The trading practices described in Section 2.1.1 do not necessarily have an anticompetitive object or effect under Article 101 TFEU. Moreover, to be able to rely on Article 102 TFEU to investigate a potential abuse by online platforms of a dominant position, the respective platforms must be dominant in the relevant market. As a result, competition law at EU or national level does not address the type and breadth of issues outlined in this Impact Assessment. This initiative will therefore, aim at complementing the enforcement of EU competition law.

Consumer law, on the other hand, does address a range of potentially harmful practices, at EU level notably through the Unfair Commercial Practices Directive (UCPD) 11 and the Unfair Contract Terms Directive (UCTD) 12 . While these directives define a number of relevant concepts, such as 'professional diligence' and 'good faith', their scope is explicitly limited to business-to-consumer transactions. Conversely, the Misleading and Comparative Advertising

6 COM(2016), 288 final, TEN/601 EESC-2016 – Opinion of the EESC Online platforms and the Digital Single Market —

Opportunities and challenges for Europe, 15 December 2016, Section 4.4.

7 ECON-VI/016, Opinion of the European Committee of Regions' on the Collaborative economy and online platforms, 7

December 2016.

8 President Juncker, Letter of Intent, 13 September 2017, addressed to President Tajani and Prime Minister Ratas.

9 Roadmap for a more united, stronger and more democratic union Tallinn Digital Summit.

10 European Council Conclusions on Migration, Digital Europe, Security and Defence, 19 October 2017.

11 Directive 2005/29/EC of 11 May 2005 concerning unfair business-to-consumer commercial practices in the internal market, OJ L 149/22. 12 Council Directive 93/13/EEC of 5 April 1993 on unfair terms in consumer contracts, OJ L 95/29.

Directive (MCAD) 13 covers certain B2B relations. It aims at protecting business users against misleading advertising and the unfair consequences thereof and lays down the conditions under which comparative advertising is permitted. 14 However, the provisions set forth in the MCAD are limited to advertising practices and do not generally address the P2B trading practices identified in this Impact Assessment.

While voluntary initiatives exist to tackle harmful trading practices in commercial contracts for example, in the food supply chain, these are sector specific. The types of potentially harmful practices that arise in the food supply chain and the unilateral practices described in Section 2.1.1, together with the different business models that operate, are very different and warrant separate treatment.

1.3 Scope of the Impact Assessment 15

This Impact Assessment analyses the relations between online platforms and their business users (so-called 'platform-to-business' relations, abbreviated as 'P2B' hereafter). It does not focus on the relation between consumers and online platforms, but does have regard for effects on consumers of the P2B dynamic, where relevant. The analysis focuses on online platforms that provide intermediation services for transactions between EU business users and consumers located in the EU.

Annex 1 and Annex 8.3 show the compatibility of the P2B scoping definition with other EU policy initiatives and existing legislation. Where necessary, the Commission services will naturally also ensure full coherence of the technical legal definitions used for the different upcoming EU initiatives that – for distinct purposes – touch in some way on the online platform economy (i.e. the present P2B initiative, the New Deal for Consumers and the digital services tax). The Table below is an extract focusing on current major initiatives in the field of taxation and consumer protection.

13 Directive 2006/114/EC of 12 December 2006 concerning misleading and comparative advertising OJ L 376/21.

14 Article 1 ibid.

15 A detailed argumentation on the scope of the Impact Assessment is attached to the analysis in Annex 1.6.

Legal Definition Platforms Platfor instrume in scope ms nt out- of - /draft scope

DRAFT Multi-sided digital interface - the making available to users of a multi B2B, B2C, All

PROPOS sided digital interface which allows users to find other users and to C2C/P2P online online

AL Art. interact with them, and which facilitate the provision of underlying platforms platform

3(1)(a) of supplies of goods or services directly between those users, irrespective of s below Council where the transactions are ultimately concluded for the this Directive purposes of turnover

establish Although borrowing from the definition of online intermediation levying the threshol services in the P2B initiative, this definition of multi-sided digital digital services ing a d: interfaces has a slightly broader scope (as it includes B2B & C2C/P2P tax (DST)

Digital platforms) in light of its purpose which is to identify taxable revenues, > EUR

Services rather than contractual imbalances in bargaining power. Whereas pure 750

Tax C2C/P2P platforms are frequently provided for-profit, which can be million

(Digitax) subject to the digital service tax, they do not exhibit the potentially global

harmful commercial issues targeted by the P2B initiative. The definition revenues

in the Digitax proposal will therefore include online intermediation ; and services for the purpose of levying the digital service tax (DST), but not > EUR 50 conflict with the definition used in the P2B initiative. The slight million

difference in intended scope between the respective proposals is EU implemented in the Digitax proposal by defining the term user as any taxable individual or business, as opposed to using the separate definitions of revenues business users and consumers in the P2B proposal. Apart from this, the definition of multi-sided digital interface will be aligned with the definition of online intermediation services, both of which target the intermediaries' role in facilitating direct transactions between their

users.

DRAFT 'Online market place' means a service provider, as defined in point (b) of Goes beyond B2B

PROPOS Article 2 of Directive 2000/31/EC of the European Parliament and of the "intermediatio online AL Art. Council of 8 June 2000 on certain legal aspects of information society n" as any platfor

2(19) of services, in particular electronic commerce, in the Internal Market service

Directive (‘Directive on electronic commerce’), which allows consumers to providers'

ms

2011/83 conclude distance contracts on the online marketplace’s online interface website could

/EU be covered This definition identifies one specific type of online intermediation

(Consum services for the purpose of tackling the targeted issue of private All B2C and er Rights providers in the collaborative economy not identifying themselves as C2C/P2P online

Directive such vis-a-vis buyers – resulting in the latter not being aware that the platforms as

-revised CRD protections do not apply. This notwithstanding the conclusion of a well as any

CRD) contract on the platform's interface, which can give the impression that website used

a contract is in fact concluded with a trader (i.e. the platform). The to offer obligation that the revised CRD will impose on online market places by services (i.e. means of this definition is accordingly strictly meant to protect app stores, econsumers,

not businesses. The P2B proposal at the same time explicitly commerce sets out that online market places are one type of online intermediation market places, services, with the latter definition clearly going beyond for a different OTAs,

purpose (to protect businesses). The concurrent application of online webshops, to market places and online intermediation services therefore will not the extent they involve any potential conflict. allow online

contract conclusion)

Annex 1.14 further substantiates the appropriateness of the scoping definition. The Impact Assessment covers services offered by online platforms whose business model is to intermediate 16 or facilitate 17 transactions between consumers and business users (so called "multi-sided" platforms and where a (P2B) contractual relationship exists between the platform operator and the business user 18 . The Impact Assessment analyses services offered by online ecommerce marketplaces, app stores, business pages on social media, ride hailing, online travel, hospitality, food delivery and product comparison platforms. 19

The Impact Assessment thus covers on the one hand intermediation services that enable a direct commercial transaction between a customer and a business user to be concluded online, on the market place. Examples of such services covered by the analysis are e.g. Amazon Marketplace, App Store, Google Play, Zalando, Booking.com, Expedia, Deliveroo, Etsy, etc.

The Impact Assessment covers, on the other hand, services that are designed to increase business users' visibility and ultimately facilitate 20 transactions between them and consumers and where the business user enters into a contractual relationship with the platform. Examples of such services include: Facebook (business pages), Google My Business, Immoweb, Autoscout, la Fourchette (restaurant booking), price comparison websites (to the extent that business users present on those websites have a contract with the platform), etc.

More in detail, the initiative would cover the following online platforms:

  • 1) 
    Marketplaces on which a commercial transaction between a customer and a business user takes place

Characteristics: The transaction and payment takes place on the platform.

Common business model: The platform charges a commission.

Services therefore included:

 E-commerce market places (Amazon market place, eBay, Etsy, Zalando, Fnac

MarketPlace, Opodo, Chrono24 Trusted Checkout, Booking.com, Expedia,

Hostelworld, Tripadvisor Instant Booking, Skyscanner Direct Booking, Uber, Airbnb,

Deliveroo, Uber Eats, Upwork, Idealo.de, Kindle Direct Publishing, Vimeo (can rent

movies), Xbox self-publishing games, Facebook – direct buy function integrated in

profiles & Messenger)

 App stores (Google Play, Opera Mobile Store, Samsung Smart TV, LG Smart World,

Sony Playstation, Oculus Gear VR, Alexa Skills)

16 These are services offered on marketplaces on which a commercial transaction between a customer and a business user takes place (transaction takes place on the platform). 17 Online platforms bring together users with the aim to "facilitate" a commercial transaction (which does not necessarily take place on the platform itself). 18 This includes both platforms on which the entire transaction takes place and those where a transaction is initiated, where the customer makes a choice from among different offers, but where the business user can also be contacted to finalize the transaction outside the platform. 19 More detailed market descriptions for all of these examples are presented in Annex 7. 20 The commercial transaction does not necessarily take place on the platform itself ("facilitator" role). The consumer joins the platform for a variety of reasons, sometimes pro-actively looking to choose between a variety of offers (e.g. houses to buy or rent). The consumer may contact the business user directly (e.g. make a call, schedule a visit, etc.). The actual payment can take place outside of the platform. The business user may be charged by the platform in different ways, such as listing fees (the level of which may depend on level of service provided – e.g. additional promotion, improved content, better visibility), charges per click and commissions.

  • 2) 
    Online platforms bringing together users with the aim to "facilitate" a commercial transaction (which does not necessarily take place on the platform itself)

Characteristics and business models:

• The business user has a contractual relationship with the platform.

• The consumer joins the platform for a variety of reasons, sometimes pro-actively looking for a possibility to choose between a variety of offers (e.g. houses to buy or rent).

• The transaction itself does not usually take place on the platform itself. The consumer may contact the business user directly, often through the platform (e.g. make a call, schedule a visit, etc). The actual payment can take place outside of the platform.

• The business user may be charged by the platform in different ways, such as:

o Listing fees (the level of which may depend on level of service provided – e.g.

additional promotion, improved content, better visibility)

o Charges per click o Commissions

Services therefore included (in addition to 1): Facebook (pages, marketplace), Google My Business, Instagram (profiles used by artists, makers), Olx classifieds, Ebay classified ads, Immoweb, Funda, Autoscout, Instagram ('shop now' button), la Fourchette (restaurant booking), SoundCloud (can purchase tracks), price comparison websites (to the extent that business users present on those websites have a contract with the platform).

Examples of services/platforms that are not covered by the above definition of online platforms:

• Peer-to-peer platforms, i.e. without the presence of "business users" (e.g.

WhatsApp, Skype, Facebook messenger, BlaBlaCar, CouchSurfing)

• Activities where business users don't have a contractual relationship with the online platform (such as Facebook profile, Google Search, Twitter, SnapChat).

o Why is Google Search excluded from the above online platform

definition (cf. Section 4.1.3 below)?

While consumers/users of search arguably have a contractual relationship with the platform (by clicking search they enter into a contract), the business users do not have a contractual relationship with the platform: Their websites are crawled, indexed, tagged without the knowledge or active participation of the business. The business model of search engines is to provide information to users and monetize it by showing them advertising. As such, search engines do not "intermediate transactions".

• Non-platform businesses (i.e. without the element of intermediation): Amazon retail, Zalando retail, Spotify, Netflix, Expedia business of purchasing bulk from hotels and reselling on own platform, cloud services.

• Pure B2B platforms (which cannot be accessed by consumers). Examples: SAP hotel booking, Salesforce AppExchange, Microsoft Azure Market Place, GE Predix, Amadues/Sabre's Global Distribution Systems, Siemens AI platform, advertising exchanges (connecting publishing companies and advertising agencies)

o Why are B2B platforms excluded?

  In the B2B context the clients tend to be big, sophisticated companies which are not easily swayed by the platforms choice of ranking. If products or services are delisted, corporate clients can insist that they be reinstated. Corporate clients have more leverage over platform decisions.

• Advertising activities. Advertisers pay for a service that allows them to reach a specific target audience that they can define, usually by means of a tailor-made advertising campaign rather than a direct digital presence on the platform itself. Consumers cannot choose which ads they will see. Ads do not always lead directly to a transaction. The technical tools used to host and serve the advertisement, which include ad serving tools and ad exchanges, are also not visible to the consumer. Examples are Google Doubleclick, Adjug, AOL, Bing Ads.

• Payment platforms (which cannot be used to initiate transactions, or to find products and services). Payment intermediaries fulfil a supplementary service supporting transactions. Like postal companies, payment intermediaries provide a service helping the parties complete the transaction. Once the consumer has already chosen the product or service and agreed on the price they use online payment services to send the money to the seller. Examples are: Apple Pay, PayPal, Klarna, Amazon Pay, Adyen, etc.

• Search engine optimisation software, which is one-sided service provided to businesses, not being visible to consumers.

• Ad-blocking software, as there is no intermediation of transactions. Although advertisers can in some cases pay for being whitelisted, the ad-blocking software does not itself enable transactions or even the actual serving of advertisements.

• Technology platforms connecting hardware and applications. There is no direct contact with both business user and consumer, and these are not directly connected with the provision of goods or services. Examples are: Android, Windows, Linux, Unix, iOS operating systems.

Online payment intermediaries are not online platforms where consumers choose from a variety of offers to conclude a transaction and where B2C transactions are initiated. Payment intermediaries fulfil a supplementary service supporting transactions. Like postal companies, payment intermediaries provide a service helping the parties complete the transaction. Once the consumer has already chosen the product or service and agreed on the price they use online payment services to send the money to the seller.

Online general search engines pro-actively index websites outside any contractual relationship with website operators for the purpose of returning the most salient results to users' search queries. These services are already defined in Directive (EU) 2016/1148 as digital services that allow users to perform searches of "all websites".

Issues relating to the ranking of business users in search services can be exacerbated by a lack of clarity and predictability around the functioning of ranking in online general search engines, as these services are an important source of Internet traffic to business users' presence on online platforms. For example, business pages of restaurants on social media, the online presence of hotels on OTAs as well as mobile software applications available in mobile app stores are all indexed by online general search engines. Online general search engines moreover are often the source of the significant majority of Internet traffic for smaller standalone websites, including those operated by business users outside their presence on other online platforms (e.g. hotels' own websites, or retailers' own webshops). One policy option identified in this Impact Assessment report is therefore to strengthen the effectiveness of the proposed legal transparency obligation on ranking by extending exclusively this provision to the separate category of digital service that are online general search engines.

A comprehensive explanation of the scoping approach is also available in Annex 1.6. Annex 4.1 provides more explanations as to who is affected by the initiative.

2 P ROBLEM D EFINITION

The core problem addressed in this Impact Assessment is that European businesses – online platforms as well as their business users – cannot fully exploit the potential of the online platform economy. For business users, this is due to a number of potentially harmful trading practices and a lack of effective redress mechanisms in the EU to tackle those unilateral trading practices. For online platforms, the underexploited potential is due to the risks they face in scaling-up and operating across the single market due to a potential loss of business users' trust as well as an emerging fragmentation of the single market. As regards the latter, the national platform-measures that have been adopted so far imply a real longer-term risk for the online platform economy as its single market-potential would be undermined by legal regimes that differ between Member States along a potentially long list of parameters (e.g. type of online platforms covered, type of trading practices covered, the use of exemptions, etc.). A key driver of this risk of fragmentation is the general pressure on national legislators and authorities to regulate the novel online platform-business models, which is fuelled partly by the very problems that this initiative aims to directly address: dependent businesses being subject to a range of potentially harmful trading practices in regard of which existing national legislation does not provide effective redress.

The above problem can cause significant harm as it limits sales for EU businesses through online platforms, which has a negative impact especially on the cross-border sales of non-platform businesses, and thus limits consumer choice and the innovation capacity of EU businesses.

Any quantitative estimates of these problems are likely to be conservative, as evidence shows a significant underreporting of issues by business users due to a fear of retaliation. This problem has a strong EU dimension, as online platforms connect buyers and sellers across national boundaries, and therefore enable e-Commerce inside the Digital Single Market.

Figure 1: Overview of the problem analysis

The drivers of the problem can be traced to the core characteristics of multi-sided online platforms intermediating transactions between business users and consumers. A relatively small number of online platforms increasingly provide the main connection between business users and consumers in each sector, which results in an increased dependency of businesses on these online platforms. Strong, data-driven network effects reinforce this dependency and together, these effects lead to an imbalance in bargaining power.

This dependency situation allows for a number of potentially harmful trading practices on the part of online platforms which limit business users' sales through online platforms and undermine their trust. These practices are not associated with any structural changes in the supplying industries. Practices identified and detailed in this IA are: sudden, unexplained changes in terms and conditions without prior notice; the delisting of products and services and the suspension of accounts without clear statement of reasons; issues related to ranking (including paid-for ranking) of businesses and products; unclear conditions for access to, and use of data collected by platforms; the discrimination of businesses and favouring of platforms' own competing services, and most-favoured nation clauses. The current regulatory framework may not be effective in preventing some of these practices, or in providing redress.

At the same time, the emerging regulatory fragmentation in the EU complicates the regulatory environment for online platforms and constitutes a significant risk for the EU platform economy. Compared to other Single Market areas, the platform economy possesses an intrinsically cross-border nature (and, in many cases indeed, global). The highly targeted but diverging national platform-specific legislations which start appearing therefore, establish a real risk of re-fragmentation of the single market.

The number of enterprises affected varies depending on the sector, but can be estimated to reach today at least 1 million merchants in the EU, combining sectors such as online retail, hotels and restaurant businesses, app stores, etc. The unrealised potential of the platform economy (in terms of reduced turnover due to sales not realised by business users) due to the unfair practices at stake, can be estimated to amount to between € 3.97 to € 15.85 billion per year and indirectly implies a loss of commissions for online platforms of between 0.4 and 1.6 billion. (see Section 2.3.2).

The market dynamics are unlikely to change significantly as the number of businesses who would like to use online platforms to reach markets and consumers is expected to grow much faster than the number of online platforms that provide intermediation services. As a result, the dependency and unequal bargaining power will only deepen. As a consequence, the market itself is unlikely to resolve the potentially harmful trading practices and the absence of effective redress mentioned above due to a misalignment of interests (explained in section 2.2). Absent EU action to address P2B issues, regulatory interventions at national level can be expected to increase in the near future. This would lead to an artificial fragmentation of the single market in the naturally EU cross-border P2B space. The resulting market re-fragmentation would prevent platforms from scaling up, thus undermining the potential of the platform economy. Online platforms are important drivers for innovation and digital transformation. A healthy platform economy, with confident business users and growing online platforms, is hence key for digital growth.

2.1 Problems

This section focuses on the three closely interlinked problems observed: potentially harmful trading practices, a lack of redress available in relation to these practices and an emerging refragmentation of the single market.

2.1.1 Potentially harmful trading practices

A study for the European Commission 21 found that nearly half (46%) of business users experience problems with online platforms in the course of their business relationship, with varying impacts. Such problems include potentially harmful trading practices, the main categories of which are set out in the following sections. Of those users that did experience problems, 21% said that these problems occurred often. Heavy users of online platforms, that is to say those that generate over half their turnover via online platforms, are far more likely to experience problems (75%) and more frequently (33% report experiencing problems often).

Evidence from the same surveys also indicates that potentially harmful unilateral trading practices are not limited to the very largest online platforms, or to specific sectors. Respondents to the survey asking to identify online platforms that generated the most issues identified a broad range of online platforms covering different sizes and sectors. Numerous different actors were indeed identified in this way within each of the categories of online platforms, ranging from app stores to marketplaces to online travel agents and social media. Other study results in addition show that in the vast majority of cases business users cannot negotiate contracts, supporting the finding of relative market strength. 22 It is noted in this regard that relatively small online platforms (including in the bracket between micro- and small enterprises, as identified in the Commission's SME definition) can indeed provide access for business users to very large consumer groups, and the example of Instagram that managed a base of 30 million users with 13 employees is one of many examples that supports the above finding that not only the very largest online platforms can exhibit an imbalance in bargaining power that enables potentially harmful unilateral trading practices to occur. 23 Further clarifications on the evidence base are provided in Annex 1.7.

2.1.1.1 Sudden unexplained changes in terms and conditions unilaterally imposed by platforms without prior notice

Online platforms tend to use standard terms and conditions 24 , which business users generally do not have a chance to negotiate. 25 One out of five business users surveyed 26 consider terms and conditions inherently unfair, and for 72% of this 20% gave the main reason for unfairness as the impossibility to negotiate contractual clauses. 27 Online platforms' terms and conditions can also be characterised by a general lack of clarity, even for legal experts. 28 Online platforms argue that, given the large number of individual business users, it is not feasible to negotiate clauses

21 ECORYS 2017, table 0.1 (see footnote 3, page ix).

22 All 100 platforms analysed as part of the study on platforms' terms and conditions used pre-formulated, standard T&Cs, cf.

Ernst & Young, "Contractual Relationships between Online Platforms and Their Professional Users – SMART 2017/0041"

(forthcoming). The Commission's E-commerce sector inquiry in addition showed that only 13% of marketplaces negotiated more than 10% of agreements with professional sellers. 23 Although not a multi-sided online intermediation service, Whatsapp relied on a team of just 35 engineers to maintain its service for 900 million users. Twitch, Youtube and many other successful online platforms similarly employed less than 100 employees while commanding user bases running in the tens of millions. 24 "The majority of the contractual relationships that marketplaces have in place with sellers are based on standard agreements. Only 13 % of the marketplaces indicate that more than 10 % of the agreements they have in force with professional sellers are negotiated individually," Recital 113 of COM SWD(2017), 154 final, Commission Staff Working Document accompanying the Final report on the E-commerce Sector Inquiry ( COM(2017), 229 final ), 10 May 2017. 25 It was also the general conclusion shared by all participants to the Commission's workshop on platforms' terms and conditions on 14 November 2016 that "changes to terms and conditions are non-negotiable: business users have to accept them in full or terminate the contract completely". 26 "Business users were asked to indicate whether they agree or disagree with the statement “The contractual terms, conditions and related practices of a platform are fair” […]. 20% of all respondents indicated that they disagree or strongly disagree with this statement", ECORYS 2017 (see footnote 3, page x). 27 Similarly, it was reported to the Commission by the Booksellers Association of the UK & Ireland that 51% of their members that participated in an internal 2017 survey on P2B trading practices strongly disagreed that they were able to negotiate or tailor contractual terms of the platform to their needs, September 2017. 28 Ernst & Young study (forthcoming), assessed the overall clarity of a sample of c.100 terms and conditions of online platforms; preliminary results indicate a widespread lack of clarity, as interpreted by legal experts.

with each user. The fact-finding supporting this Impact Assessment also showed that business users are exposed to sudden, unilateral changes of terms and conditions. Almost 50% of the terms and conditions investigated even grant online platform operators an explicit right to unilaterally change the terms and conditions. 29 In addition, the content of changes is not always made clear to business users.

Whilst regular changes to terms and conditions are necessary to adapt to changes in the business environment and to legislation, the problem of sudden, unexplained changes can be substantial for the weaker party. 19% of the businesses reported problems related to sudden changes in terms and conditions 30 . Business users argue that frequently they do not have enough time to make the necessary adaptations to their business operations when substantial features of the service are changed. Examples include changes to return and exchange policies of e-commerce platforms communicated through hyperlinks in routine emails to the business users, or announcements of increases in the price of apps by up to 25% to reflect currency fluctuation. 31 Some companies reported losses in turnover caused by such practices varying between 20% and 95%. 32

2.1.1.2 Delisting of products, services or businesses or suspension of accounts without clear statement of reasons

Businesses using online platforms are often reliant on traffic from these online platforms for sales 33 , and the delisting of certain products or services or the overall suspension of their account has a strong impact on their business. While delisting and/or suspension can be justified by a variety of legitimate reasons, including the take-down of illegal 34 or harmful content, or as a consequence of other non-compliance with the terms of service, 35 few, if at all, safeguards are in place for arbitrary delisting or suspension of accounts on online platforms. In particular, business users pointed 36 to a frequent absence of a clear statement of reasons when delisting suspension occurs. Consequently, they have few levers to remedy the situation leading to the sanction, or to seek redress and challenge the delisting or suspension.

Respondents to the survey who reported having had issues linked these to the suspension of their account (11%), and to other access conditions to the online platform including instances of delisting (15%). 37 Such practices were registered most often in the e-commerce and app store environments.

2.1.1.3 Issues related to ranking of business users or their offers

There is a lack of meaningful accountability and predictability for the business user with regards to ranking systems used by online platforms. In the study 38 , 12% of the respondents having encountered problems in their business relationship with the online platform (and 15% of the

29 Ibid.

30 Figure 3.5 of ECORYS 2017 (p. 32) – 19% of the 37% of business users that had experienced problems with platforms, that is to say 7% of total users. 31 ' Apple increases App Store Prices by 25% following Brexit' , The Guardian, 17 January 2017. 32 Figures reported by some of the business users responding to the inquiries of ECORYS 2017 (see footnote 3). The estimates have not been extrapolated to the entire sample. 33 See Section 2.2.4 above. 34 COM(2017), 555 final - 'Tackling Illegal Content Online', 28 September 2017. 35 Delisting products and services can amount to an abuse under the EU competition rules. However, that can only be the case under stringent circumstances, e.g. in the case where the platform would be indispensable for downstream competition. 36 See summaries of workshops with business users of online platforms in Annex 3.3.5. 37 Table 3.5 of the ECORYS 2017 (see footnote 3, page 32). 38 Table 3.6 of the ECORYS 2017 (see footnote 3, page 32).

'heavy users') claimed these were due to biases in the search related practices. According to the study results, such issues occur most frequently in the e-commerce and hospitality sectors.

Business users are heavily impacted by their position on the online platforms' page. The ranking of a product or service in search results on the online platform has an important impact on consumer choice and, consequently, on the businesses' revenues.

In some of the biggest EU Member States, online platforms (as defined in this Impact Assessment) already account for a share of over 40% of total desktop Internet traffic in the ecommerce and hospitality sectors (see section 7.2.5. of the Annexes to this report). The largest part of this share (70%-80%) is accounted for by direct Internet traffic and therefore does not rely on referrals by online general search engines. These figures underline the crucial market gateway that online platforms represent for business users.

Notwithstanding, online general search engines continue to be important as an indirect source of Internet traffic for business users on platforms. For example, business pages of restaurants on social media, the online presence of hotels on OTAs as well as mobile software applications available in mobile app stores are all equally indexed by online general search engines.

Online general search engines in addition originate the vast majority of Internet traffic for smaller, standalone websites. This applies equally to websites run by business users of online platforms outside those platforms. In the e-commerce and hospitality sectors, Internet traffic in the eight largest EU Member States generated by online general search engines accounted for, respectively, >50% and >70% of total desktop Internet traffic received by these websites (which percentages constitute multiples of the share that search traffic accounts for in respect of the 100 most well-known websites in these Member States). A recent Eurobarometer survey on the use of online platforms also found that nearly nine in ten Internet users in the EU use search engines websites at least once a week. 39

At the same time, 66% of EU SMEs also explain that their position in search results of general search engines and online platforms has a significant impact on their sales. 40 Studies 41 also show a significant and positive relationship between the first position of a product in a ranking and the choice of consumers. Conversely, there is a negative effect of low ranking. 42 The top five search results attract 88% of the clicks 43 , while it is very rare – a chance of 1.11% or less – for a user to click anywhere beyond the 10th site in a search result 44 . It should also be noted that online general search engines continue to be the most common starting point for online research both on desktop as well as on mobile devices, 45 and therefore continue to be important for business users of online platforms as well as for standalone websites, which may be part of one and the same company.

Online platforms and online general search engines are distinct types of online services, with the latter indexing websites without necessarily entering into contractual relations with the website

39 Special Eurobarometer 447, Online Platforms of June 2016.

40 Flash Eurobarometer 439 "The use of online marketplaces and search engines by SMEs" of June 2016, p. 21.

41 ECME Consortium in partnership with Deloitte , 'Study on the Coverage, Functioning and Consumer use of Comparison

Tools and third-party verification schemes for such tools EAHC/FWC/2013 85 07', 2015 page 150. According to Flash

Eurobarometer 439 , two thirds of the companies that sell online agree that their position in search results has a significant impact on their sales. 42 One example was reported in discussions with the European Commission by a company developing apps, pointing to an 80% drop in revenue and a 70% drop in traffic rates following a sudden lower ranking on the app store search results. 43 Fairsearch.org based on Online Marketing Research, iProspect, iProspect Search Engine Behavior Study, 3 April 2006. 44 Daniel Ruby, The Value of Google Result Positioning, CHITIKA INSIGHTS (May 25, 2010), http://insights.chitika.com/2010/the-value-of-google-result-positioning/ 45 https://www.thinkwithgoogle.com/advertising-channels/mobile/mobile-path-to-purchase-5-key-findings/

users. Online search engines also generate revenue from advertising rather than from intermediating B2C transactions and they necessarily have to index the entire Internet in order to provide a quality service (i.e. there is no room for specialisation in general search). At the same time, online visibility for small businesses is dependent both on online platforms as well as on online general search engines. There are, at the same time, no indications that the concurrent importance of online platforms and online general search engines will change significantly in light of foreseen technological developments. 46

Online platforms as well as online general search engines explain that ranking algorithms are increasingly complex and are oftentimes at the core of the service innovation proposed to consumers: the better the user experience, the more successful the online platform or online general search engine and, consequently, the benefits also for the business user.

Some online platforms offer instructions and support to their business users for optimising their ranking (See Annex 7.3), including information on how to encode meta-data for the services and products listed, or parameters on e.g. sales and user reviews which would help rank higher. Other online platforms, however, are opaque and vague in their terms and conditions and businesses reported 47 unclear criteria, including special programmes offered to some business users, fastchanging parameters in the ranking of offers, and fear of arbitrary dimming of ranking by the online platform. Uncertainty about the main search parameters, including the risk of demotions, can add to the lack of predictability that both business users of online platforms as well as websites face when trading online. This behaviour has indeed entirely undermined the effect of voluntary efforts to reassure business users. As regards the other ("searcher") side of the online general search market, a recent Eurobarometer survey moreover found that 19% of Internet users in the EU do not trust that the search results provided to them are the most relevant to their query. 48

In addition, rankings of business users and their offers can be influenced by (additional) payments by the business users whose products and services are made more visible in the rankings (paid-for ranking). The increase in visibility afforded by paid-for ranking can be achieved via direct payment for advertising (business pages can for example "boost" their visibility on social media), or sometimes through an increase in the commission paid per transaction (as commonly seen in the hospitality sector). Business users have argued 49 that it is often unclear to what extent the increased commission leads to a higher ranking or frequency of display of higher ranking in consumers' searches. As such, they pay without being certain to what extent the service is delivered to them.

As far as transparency of paid-for results towards consumers is concerned, under the Unfair Commercial Practices Directive (UCPD) 50 , online platforms as well as online general search engines are required to distinguish paid-for results from "organic" search results. 51 However, while informative to the consumer, the distinction between paid-for and organic search results is

46 Google Search is for example integrated with Apple's Siri voice assistant, meaning that possible growth of this new user interface is unlikely to displace the use of online general search engines. 47 See the workshop report "Business-to-business relationships in the online platforms environment - algorithms, ranking and transparency" .

48 Special Eurobarometer 447, Online Platforms of June 2016.

49 "Another big problem is that higher ranking due to the participation in preferred partnership programs is not transparent",

ECORYS 2017 (see footnote 3, pages 38-40), and in bilateral discussions with businesses.

50 SWD(2016), 163 final , Revised Guidance on the Unfair Commercial Practices Directive, 25 May 2016, Articles 6(1) c and

7(2).

51 The Key Principles for Comparison Tools (May 2016), which have been developed and endorsed by stakeholders under the steer of the Commission, also clearly state that advertising and sponsored results must be prominently differentiated from organic comparison results.

not sufficient to reassure business users of fair delivery of the improved paid-for ranking service across the different consumer segments and in comparison to other competing businesses.

On the other hand, wide ranging disclosure of ranking algorithms is generally accompanied by attempts to manipulate rankings ('gaming'), as business users are incentivised to gain a higher ranking without necessarily improving the quality of the product or service.

2.1.1.4 Issues related to data access and use

Further data sharing and use across the value chain is a first condition for maximising the value of data 52 . At the same time, economic theory is not conclusive as to the role of data in network effects around online platforms or the impacts of data flows across the value chain. Online platforms aggregate large amounts of personal and non-personal data 53 , both at the very core of the online platforms' business model 54 , and the online platforms' ability to build and maintain a user base on both sides of the market depends to a large extent on the collection and retention of data.

Preliminary results of the study on data in P2B relations 55 show that business users do not have consistent views as to their level of satisfaction with the data access policies of the online platforms they use. Some argue that they lack access to specific types of information regarding their customers, while others acknowledge that they can access a large variety of data, but that they lack the resources or skills to exploit it. The variety of data types businesses can access is not consistent across online platforms. In some cases, this is a matter of competing offers between online platforms providing similar services: e.g. market analysis either as part of the service to business users or against a fee. In addition, third parties also frequently aim at providing their data-related services to business users active on online platforms.

Limited access to data and limited skills to procure, analyse and exploit data-driven market insights have a negative effect on businesses' ability to grow. 56 At the same time, developing data sharing policies, legal provisions and facilitating technically access to data is costly on the online platforms' side, in addition to potentially affecting, in some cases, the relevance of the intermediation business model of the online platform. 57

Preliminary results of a study on data access commissioned by the European Commission 58 also identified a specific issue for business users mainly active in the hospitality and e-commerce sectors. The vast majority of online platforms do not give business users the opportunity to ask for customers' consent 59 to obtain and process his or her certain personal data, in particular e-mail addresses, even after the completion of a transaction and the payment of the commission

52 The Commission's policy on data sharing is synthesised in the Communication on 'Building a European Data Economy,'

COM(2017), 9 final, 10 January 2017.

53 The range of data collected by platforms include for example data provided by business users to platforms; data provided by end consumers to platforms; data generated by transactions between end consumers and business users via the platform; or data generated by the consumers' general use of the platform beyond specific transactions. 54 See Section 2.2.3. 55 VVA, 'Study on data in platform-to-business relations', ENTR/172/PP/2102/FC, forthcoming, and testified in the datarelated workshop organised by the Commission (see Annex 2 and 3), as well as in bilateral discussions with platforms and business users.

56 Restrictive policies of data access and use also have a negative impact on the market of third party data analytics and brokering services.

57 Costs for technical provisions include expensive processes for data curation, storage and network provisions, security and, potentially, differentiated access provisions, development and maintenance of application programming interfaces and accompanying documentation. When personal data is concerned, additional costs for anonymisation need to be factored in. 58 See footnote 58, VVA 2017 59 In doing so, the EU General Data Protection Regulation (GDPR) principles on purpose limitation (Article 5) and lawful processing (Article 6) would need to be taken into account.

to the respective online platform. In the Flash Eurobarometer 439 60 42% of the respondents said that they usually do not get the data they need about their customers from online marketplaces. The business users claim that not having this possibility increases their dependency on the online platform as a gateway to consumers and prevents them from scaling-up by, inter alia, building an independent customer base 61 , improving their direct marketing or independent online presence as well as offering customer-tailored services 62 . Online platforms responded that they offer ways of contacting customers, in some cases including marketing against a fee, for business users via their platform systems. Their business model is typically built on commission per transaction intermediated by the online platform and both consumers and business users are naturally incentivised to avoid concluding transactions on the platform and preventing direct contact between customer and business users helps to avoid 'free-riding' behaviour. Some online platforms also indicated that they are required to shield customers from direct contact with business users to comply with EU data protection rules.

Beyond claims to access specific types of data, there is a lack of clarity as to the conditions for access and use of data, both regarding online platforms' collection and use of businesses' and transaction data, and the conditions for business users to use data collected from the online platform. 25% of non-heavy users and 33% of the heavy users of online platforms responding to a Study 63 said the problems they have encountered were caused by the lack of transparency of online platforms' policies and practices on data and content. Further research into the online platforms terms and conditions 64 shows that online platforms frequently include general, and often unclear clauses restricting to a certain extent the use of particular types of data by the business user outside of the environment of the platform. The clauses are generally rooted in the protection of the online platforms' trade secrets, databases 65 or to impede the use of data collected from the platform's environment to compete against the online platform's intermediation business model.

2.1.1.5 Discrimination of businesses and favouring of online

platform's own competing services

Online platforms sometimes play a dual role, for example by both providing the online market place and selling their products and services on their own market place. When such online platforms apply differentiated treatment to their own products or services 66 such treatment is generally not made transparent to their business users. The favouring of own products or services by online platforms was identified as one of three most commonly experienced problematic trading practices by business respondents to the public consultation on platforms. 67

Favouring of own products or some business users takes place e.g. through more favourable ranking, use of transaction data to learn from downstream competitors and improve online

60 Flash Eurobarometer 439 'The use of online marketplaces and search engines by SMEs' of April/June 2016 .

61 As described in more detail in Section 2.2.2 on network effects, businesses benefit from platforms' existing customer traffic and platforms take advantage of the attractiveness of the businesses' products and services to attract even more customers. However, albeit both platforms and businesses have their share in attracting customers, in most cases platforms do not allow businesses to establish a direct customer relationship even with those customers who have already transacted with them through the platform. 62 HOTREC Hospitality Europe, Position Paper on the Mid-term Review of the Digital Single Market Strategy . 63 Table 3.6 of the ECORYS 2017 (see footnote 3, page 32). 64 Ernst &Young study, forthcoming (see footnote 51). 65 As confirmed in Directive 96/9/EC of 11 March 1996 on the legal protection of databases, OJ L77/20, and national legislation implementing its provisions. 66 Differentiated treatment can breach competition rules if certain conditions are fulfilled. For example, the Commission has recently imposed a fine on Google for abusing its market dominance as a search engine by giving an illegal advantage to Google's comparison shopping service 'Antitrust: Commission fines Google €2.42 billion for abusing dominance as search engine by giving illegal advantage to own comparison shopping service', 27 June 2017. 67 Synopsis Report on the Public Consultation on the Regulatory Environment for Platforms, Online Intermediaries and the Collaborative Economy.

platforms' own competing service, or by charging additional fees to third party business users but not to online platforms' own services. 68

The "bundling" of auxiliary services like advertising or payment to the online platforms' intermediation service was also reported in this context, as the choice for a business user to use such auxiliary services would effectively be limited to the relevant online platforms' own solutions. 11% of business users surveyed in a study for example linked the problems they experienced to limitations placed on payment possibilities 69 . Some app developers reported adapting to the online platforms' commission on auxiliary payment services by applying a net price increase of 30%. A third of a sample of online platforms used terms and conditions that were not transparent as to the pricing of the main online platform (intermediation) service and auxiliary services. 70

2.1.1.6 Most-favoured nation (MFN) clauses

Issues have also arisen in the context of so-called 'most-favoured nation' ('MFN') clauses 71 , also known as 'parity' or 'price-parity' clauses. These are common in Online Travel Agents ('OTAs'), but also exist to a more limited extent on e-commerce platforms, app stores or price comparison tools. MFN clauses require the supplier to offer a product or service on an online platform at the lowest price and/or on the best terms offered either through its own distribution channel(s) ('narrow' MFN clauses 72 ) or on all sales channels ('wide' MFN clauses).

The ongoing monitoring by competition authorities regarding MFN clauses in the hospitality sector constitutes an important element of the baseline scenario in this Impact Assessment. While the economic literature suggests that MFN clauses can create efficiencies in particular market contexts, certain MFN clauses used specifically by OTAs have been investigated by several national competition authorities. The German competition authority prohibited HRS' and Booking.com's MFN clauses in 2013 and 2015 respectively 73 . In close coordination with the Commission, the French, Swedish and Italian authorities accepted Booking.com's commitment to reduce its wide MFN clauses to narrow clauses EU wide, thereby accepting the use of such clauses in the future. 74 Following the decisions, a group of 10 national competition authorities and the Commission decided to carry out a monitoring exercise to assess their effects. 75 The enforcement measures resulted in increased room price and room availability differentiation on OTAs, but there is no clear evidence that they have led to lower commission rates charged by OTAs. The heads of the European Competition Network (ECN) therefore agreed to keep the online hotel booking sector under review, to re-assess the competitive situation in due time and to coordinate new enforcement actions or market investigations within the ECN. 76 Competition law can therefore provide, in certain instances, the possibility to redress and correct some of the identified problems regarding MFNs on a case-by-case basis.

68 See the workshop report "Business-to-business relationships in the online platforms environment - algorithms, ranking and transparency" .

69 Table 3.5 of the ECORYS 2017 (see footnote 3, page 32).

70 Ernst &Young study, forthcoming (see footnote 51).

71 MFN clauses are included in the "Unfair terms and conditions" category of ECORYS 2017(see footnote 3).

72 Narrow MFN clauses are understood to refer to those clauses that are limited in application to the online platforms' website and the suppliers' own website, thus leaving suppliers free to offer better conditions through offline channels, through emails, through closed user groups, or through other online platforms. 73 The decisions are available here and here. 74 The decisions are available here . 75 ' Report on the Monitoring Exercise Carried out in the Online Hotel Booking Sector by EU Competition Authorities in 2016,' 6 April 2017. 76 Outcome of the meeting of the ECN and DGs , 17 February 2017.

MFN clauses in the online hotel booking sector are also regulated by several national laws. 77 The French 'Loi pour la croissance, l'activité et l'égalité des chances économiques' (Loi Macron) adopted on 5 August 2015 foresees that hotels remain free to grant any rebate or pricing advantage of any kind to customers through their direct sales channels. A new Austrian law amending the Austrian Federal Act against Unfair Competition and the Austrian Price Marking Act and an Italian law, which entered into force respectively on 1 January 2017 and on 29 August 2017 78 , also prohibit any MFN clauses in agreements between OTAs and hotel operators (i.e. wide and narrow MFN clauses, and regardless of the size of the OTA). A draft law containing a similar per se ban of MFN clauses imposed by OTAs has recently been proposed by the Belgian government and has been notified on 4 December 2017 to the Commission under Directive 2015/1535/EU, which establishes a transparency procedure for rules applying to information society services. 79

2.1.2 Lack of effective redress

When business users attempt to solve the potentially harmful trading practices described above, they are often unable to find a solution. According to a study carried out for the European Commission, almost a third (32%) of all problems in P2B relations remains unsolved and a further 29% can only be resolved with difficulties. As regards online general search engines, a recent survey found that 32% of EU businesses selling online disagreed that a reliable dispute resolution system is available to solve disputes with the operator of online general search engines. 80 Reasons for business users not to take any steps at all notably include the perceived ineffective nature of existing redress mechanisms, a fear of damaging the business relationship with the online platform and the difficulty of available procedures. 81

The dependency-induced fear of retaliation of business users (Section 2.2.5) indeed limits the effectiveness of any existing type of redress, whether judicial or out-of-court. In addition, online platforms generally use exclusive choice of law and forum clauses. 82 In the inherently crossborder digital economy, the widespread use of such clauses significantly raises the existing barrier to access justice, as any national court seised by a business user will first need to settle the complex question of whether it is competent to deal with the case at hand regardless of the applicable law and competent forum determined by contract. Existing national B2B fairness legislation, which can in theory be relied upon by businesses in certain Member States to seek relief against alleged potentially harmful behaviour, is therefore significantly impaired in terms of its use in the online platform economy.

Other important factors that limit the effectiveness of judicial redress are linked to (1) lack of knowledge of judicial redress possibilities due to the small size of the companies, (2) disproportionate costs of seeking international judicial redress, especially for the microenterprises and/or where jurisdictional redress would involve the jurisdiction of a third country, and (3) judicial redress being too lengthy. 83

77 Many of the national laws already apply and are actively enforced, notwithstanding that in some cases complaints have been brought alleging that these laws breach EU law. 78 Legge annuale per il mercato e la concorrenza , adopted on 2 August 2017. 79 See notification 2017/570/B under Directive 2015/1535/EU . 80 Flash Eurobarometer 439 "The use of online marketplaces and search engines by SMEs" of June 2016, p. 20. 81 ECORYS 2017 (see footnote 3, page 63). 82 The terms and conditions of the platforms analysed in a recent study for the Commission included, without exception, such exclusive choice of law and forum clauses, see: Ernst & Young study (forthcoming). 83 Also more generally, EU SMEs find the cost of proceedings as the main reason for not using a court to settle a dispute, and 19% of EU SMEs do not use conventional alternative redress mechanisms out of the fear that nothing would come of it and out of the fear to ruin the business relationship, see: Flash Eurobarometer 347 , 'Business-to-business alternative dispute resolution in the EU', November 2012, page 7.

Study results in this regard show that only 4% of business users of online platforms that took action when faced with a problem went to court in the EU, and this formed the only exception of an external redress mechanism used by these business users where even the majority of issues (55%) could not be resolved (45%) or only with difficulties (10%).

Participants in a Commission workshop also indicated that online platforms currently either do not offer internal redress mechanisms, or that such mechanisms are ineffective, in particular for claims where the business user's interest opposes that of the platform. External procedures are found to be ineffective for different reasons, including a fear of retaliation on the side of business users 84 , high costs and the length of procedures. Similarly, some retailers that participated in the Commission's e-commerce sector inquiry stress the importance of the transparency of the notice and take down process on e-commerce market places, and consider that the possibilities of retailers to defend their interest and request review of the decision taken by the marketplace are not sufficient. 85

2.1.3 Existing and emerging regulatory fragmentation of the Digital Single Market

The existing legal framework at both EU and Member State-level 86 does not effectively address the problems identified in this Impact Assessment.

General B2B fairness rules exist in some Member States, but they are not geared towards the platform-specific problems identified above.

Similarly, existing initiatives targeting harmful trading practices in the offline world are designed to tackle practices relevant to the sector or context in which they arise. For example, the Supply Chain Initiative aims at increasing fairness in commercial relations along the food supply chain. A set of principles on good practices in vertical relationships in the food supply chain were devised by industry voluntarily in November 2011. 87 The Commission is considering further action to improve the position of farmers in the food supply chain, 88 in light of the outcome of the work of the Agricultural markets Task Force 89 and the High Level Forum for a Better Functioning Food Supply Chain 90 . This is framed around addressing problems such as fairer payment periods for suppliers, prohibitions on the last minute cancellation of perishable goods, requirements for contributions to promotional or marketing costs, claims for wasted or unsold products and requests for upfront payments to secure or retain contracts. 91 None of these overlap with the platform-specific problems identified above.

Creating a single rule to address potentially harmful practices in the online and offline world would not address the specificities and the problems businesses face in the sector they operate in. It also does not recognise the differences in the business models. 92 This may be because of the very different business models and the fact behaviour may not be seen both offline and online

84 See Section 2.2.5.

85 Recital 498, COM SWD(2017), 154 final , Commission Staff Working Document accompanying the Final report on the E- commerce Sector Inquiry ( COM(2017), 229 final ), 10 May 2017. 86 The current regulatory situation in the EU is described in more detail in Annex 8. 87 The Supply Chain Initiative - Principles of Good Practice in vertical relationships in the Food Supply Chain 88 Commission initiative to improve the governance of the food supply chain with regard to unfair trading practices, one rule regarding producer cooperation and market transparency 89 An independent high-level group reporting to the Commission, composed of 12 independent experts and chaired by former Dutch Minister for Agriculture and University professor, Cees Veerman. 90 This has been set up by the European Commission to help develop policy in the food and drink sector and contribute to a better functioning food supply chain. The forum today comprises of EU country national authorities responsible for the food sector at ministerial level and representatives of the private sector. 91 Improving Market Outcomes, Enhancing the Position of Farmers in the Supply Chain, Report of the Agricultural Markets Task Force, Brussels, November 2016 92 The characteristics of platforms have been described in the drivers described in Section 2.2.

(for example, algorithms play no part in the placement of a product on a supermarket shelf). In fact, a single rule could extend a solution beyond what is necessary to address the problems identified.

At the same time, a number of Member States (Austria, France, Italy, Germany and Belgium) have already adopted, or are considering adopting, online platform-specific legislation.

• As explained in Section 2.1.1.6 above, France, Austria and Italy have adopted laws prohibiting MFN clauses imposed by OTAs on their business users. Belgium in addition notified a draft law to the Commission on 4 December 2017, which contains a similar MFN-

ban. 93 These laws constitute per se prohibitions (i.e. regardless of the market size of these firms) of the use of MFN clauses, without distinguishing between sales channels (OTAs, hotels' websites or offline), by one specific category of online platform (i.e. OTAs) in their relations with business users. This approach contrasts with the competition-law based commitment-approach focusing only on wide parity clauses that was taken by a group of Member States 94 . These national laws banning the use of MFNs by OTAs also differ in their respective design. The French law in this regard contains an additional requirement, as compared to the Austrian and Italian laws, requiring the contract between hotels and OTAs to determine a fixed room price. Online platforms have indicated that such a requirement forces their cross-border operations to be segmented along, in this case, the French borders, which they moreover consider virtually impossible to comply with.

• In 2016, France adopted Law N. 1321 95 , which defines online platforms and requires these firms to provide further transparency towards consumers on e.g. terms and conditions, certain mandatory pre-contractual information or on the way in which goods, content or services are ranked and whether there are any contractual or financial relationships influencing this.

• The Italian Parliament has considered two proposals that aim to regulate some platformrelevant aspects. Proposal N.2520 96 aims to abolish certain restrictions imposed by platforms, specifically app stores, impeding mobile device users' freedom and ability to access or remove apps as well as to switch services 97 . This obligation could imply significant cost for both app stores as well as their business users (independent app developers), as the same content may have to be made available across different platforms and corresponding operating systems (source code). Moreover, it could affect competition between online platforms also to the detriment of their business users, as it would no longer be possible to distinguish on the basis of richness or quality of content. Proposal N. 3564 98 has as its main objective to ensure fairness and transparency regarding security, health, taxation in the collaborative economy. Platforms intermediating connections between consumers and business users are expressly excluded from the Proposal, because the focus is solely on relationships between consumers, but it is not clear is self-employed business users would be unequivocally out of scope. 99

93 See notification 2017/570/B under Directive 2015/1535/EU .

94 Note a full prohibition of MFN clauses is also possible under competition law, but only following an effects-based analysis, as included for example in the relevant prohibition decisions of the German competition authority. 95 Loi pour une République numérique of 7 October 2016, Article 49. For the fairness standard see Article 117-7 II. 96 Proposta di legge " Disposizioni in materia di fornitura dei servizi della rete internet per la tutela della concorrenza e della libertà di accesso degli utenti " (2520), Approved on 7 July 2016. Transmitted to the Senate 97 Various amendments have been tabled regarding this proposal, including one to delete this specific provision. The result of the legislative process is difficult to foresee at this stage. Notwithstanding, the Italian government has notified the draft law to the Commission on 24 October 2017 which contains the relevant provision. 98 Proposta di legge "Disciplina delle piattaforme digitali per la condivisione di beni e servizi e disposizioni per la promozione dell'economia della condivisione " (3564), under discussion in the Italian Parliament as of 05/12/207. 99 The Proposal also requires platforms to publish a policy document comprising its general terms and conditions, which is subject to the opinion and approval of the Italian Competition Authority and which will be included in a "National Electronic

• The German government issued a White Paper on Digital Platforms 100 which envisages the creation of a 'Digital Agency' to safeguard effective and systematic market control of digital platforms and proposes a comprehensive framework for the use of data and to introduce basic transparency and information duties for digital platforms.

None of the above already enacted or envisaged national platform-measures comprehensively cover the set of potentially harmful trading practices identified above, and they in any event suffer from the difficulty to be enforced (especially through private litigation) in the inherently cross-border platform economy. They do not therefore provide effective redress for business users of online platforms against potentially harmful trading practices. Certain types of specific business users (e.g. hotels) may nonetheless, depending on their Member State of establishment, benefit from a higher perceived level of legal protection on targeted issues, which can lead to an uneven playing field in online intermediated trade – even within one and the same Member State. Moreover, some of these Member State measures may raise issues of compatibility with EU law. At the same time, these national platform-measures imply a real longer-term risk for the online platform economy as its single market-potential would be undermined by legal regimes that differ between Member States along a potentially long list of parameters (e.g. type of online platforms covered, type of trading practices covered, the use of exemptions, etc.). A key driver of this risk of fragmentation is the general pressure on national legislators and authorities to regulate the novel online platform-business models, which is fuelled partly by the very problems that this initiative aims to directly address: dependent businesses being subject to a range of potentially harmful trading practices in regard of which existing national legislation does not provide effective redress.

2.2 Drivers

This Section outlines the market dynamics and the drivers of the problem. A detailed analysis of the market structure is presented in Annex 7.2. Annex 1.8 gives an overview of drivers.

2.2.1 Online platforms intermediate an increasing number of transactions and are increasingly the main vehicle for market access

The European e-commerce market has been growing at a dynamic pace. In 2016, two thirds of internet users made online purchases, while the value of the market was estimated at over € 500 billion 101 , which is a 13% increase in comparison to 2015. 102

This growing digital trade is increasingly intermediated by online platforms. The retail value generated by EU third party sellers on platforms represented 22% of total online retail sales in 2016, and in countries such as Germany over 37% of total internet sales were generated by third party sellers 103 . Online platforms that host third party sellers are now leaders of internet retailing. The biggest marketplaces, such as Amazon, Alibaba's Tmall and eBay account for $365 billion in sales worldwide in 2016 104 . Sales of online-only retailers in the EU more than doubled between 2011 and 2016, reaching €111 billion in 2016. Sales over platforms now account for over half of all online sales in retail. According to Euromonitor, the online retail value generated by third party sellers in the EU in 2016 was €54,566.5 million, representing 22% of total online

Register of Digital Platforms of the Sharing Economy". It also includes a blacklist of contract terms, e.g. exclusion of access to platforms without legitimate reason, as well as a definition of traders to ensure fiscal transparency. 100 Bundesministerium für Wirtschaft und Energie, Weißbuch Digitale Plattformen , March 2017. 101 European B2C E-commerce Report 2016 , E-commerce Europe. Ecommerce Europe is a European association representing more than 35,000 companies selling goods and/or services online to consumers in Europe. 102 Ibid.

103 Euromonitor International, Passport Database 2016 Edition.

104 Euromonitor International, New Retailing Research 2017 Edition. Key Trends for the Industry to 2021.

retail. Other estimates suggest that around 60% of private consumption and 30% of public consumption of goods and services related to the total digital economy go via online intermediaries. 105 Online marketplaces enable businesses to take advantage of the growing markets through existing infrastructure that is already trusted by consumers. For this reason, 71% of the consumers who participated in a survey on online platform transparency found online market places to be the preferred source to buy goods or services for private use 106 . Online sales accounted for 9% of global retail sales in 2016 and that figure is expected to rise to 13% in 2021. 107 However, the impact of online retailing is much stronger due to web-influenced crosschannel sales: more than half of total retail sales in Europe in 2020 (€957 billion out of €1 793 billion) is estimated to be influenced by e-commerce, up from €603 billion in 2015. 108

In 2016, online booking channels captured 49% of all travel bookings in Europe. The two biggest online travel agents have now over 60% of European "market share" of OTAs in Europe, although through a large number of important online platforms 109 . Online travel agents are particularly important for small, independent hotels - one study shows that independent hotels make up 67% of total room supply in the EU and that 71% of their online bookings are made through online platforms 110 .

App developers generally distribute their apps through app stores and some studies estimate the EU app industry to amount to €63 billion by 2018 111 , while the global mobile apps revenue is estimated to increase from $69.7 billion in 2015 to $188.9 billion in 2020. 112 The use of social media promote and drive traffic to the services and products offered by business users: 89% of business user respondents to the surveys in a study carried out for the European Commission use social media for business purposes 113 . At a global level and in Europe, Facebook has a clear market lead in this category, claiming over 2 billion of active users through its various owned online platforms (in June 2017, out of 2.5 billion) 114 . Finally, the importance of online platforms can be further illustrated by the example of private motor insurance in the UK, where more than 50% of overall sales volumes during the period 2013-2015 is generated via online intermediaries. 115

The relevance of organic search as a source of traffic grows. In the case of (i) accommodation and hotels, one quarter of all traffic is generated by organic search results, (ii) online retail, 28.6% is generated by organic search results, while (iii) for government sites, 43% of all traffic comes from organic search. 116 The retail sector shows a high degree of dependency. For example in Germany, 43% of total Internet traffic related to eCommerce goes to the top 10 online platforms in this space. Notwithstanding, organic search still does constitute a major source of traffic, including for online platforms – in France, over 33% of total traffic of the top 10 online retail platforms is referred by organic search.

As of April 2017, Google is the leading search engine – it has 88.56% of worldwide desktop market share, with other search engines (such as Yahoo!, bing and Baidu) sharing the remaining

105 Copenhagen Economics , Online Intermediaries: Impact on the EU economy 2015, page 9.

106 LSE & Partners, Behavioural Study on the Transparency of Online Platforms , forthcoming.

107 Euromonitor International, New Retailing Research 2017 Edition, Key Trends for the Industry to 2021.

108 Forrester Research, " European Cross-Channel Retail Sales Forecast, 2015-2020 ":

109 Dealroom, " Online travel: A deep dive ", June 2016.

110 Hotel Management, ' The Digital Marketplace in Europe: Hotels and Third Party Intermediaries In the New Age of Travel ',

2016.

111 GigaOm Research ' Sizing the EU app economy ', February 2014.

112 Statista, Worldwide App Revenues ,

113 ECORYS 2017, figure 3.1 (see footnote 3, page 30).

114 Statista – Number of social media users worldwide from 2010 to 2021

115 Digital comparison tools market study, final report of the UK Competition and Markets Authority, page 22.

116 Figure based on Similarweb's index of the top 100 websites for December 2017.

part of the market 117 . In Europe, Google's market share was 92,5% and bing's 3.3% in May 2017. 118

At present, more than a million EU enterprises trade through online platforms in order to reach their customers. 119 Online platforms have become central to the businesses using them: almost half (42%) of SME respondents to a recent Eurobarometer survey on online platforms use online marketplaces 120 to sell their products and services. Online platforms have also become enablers for cross-border trade. Through the intermediation of online platforms, businesses can de facto reach consumers across the entire European Single Market, as well as in global markets. A recent survey among 49 081 SMEs active on Facebook showed that nearly half of exporting SMEs (45%) report that more than 75% of their international sales depend on online tools or platforms. 121 Research, such as the comparative study on one e-commerce platform and overall international trade flows, shows a significant smaller effect of geographic distance on trade when online platforms intermediate transactions - up to 65% for the study quoted 122 . These findings are supported by the conclusions of a public consultation, where all categories of stakeholders agreed in the public consultation that one of the most important benefits offered by platforms was the access they offer to new market and business opportunities. 123

2.2.2 Successful platforms enjoy unprecedented, strong network effects

Indirect network effects can be at the heart of the business model of online platforms: the increase in the number of users on one side of the platform (e.g. sellers, content creators, service providers) makes it more attractive to users on the other side (e.g. consumers, viewers) and the other way around. In the online world, these network effects are of an unprecedented magnitude, scale and speed. While the increase of cost to provide services to additional users on either side of a networked market grows increasingly slowly, the value of the network increases very rapidly with the number of additional users on either side. Platforms thus create their economic value by attracting and retaining users on both sides of the market, while the investment e.g. in infrastructure for supporting additional users is marginal: when a platform scales to millions of consumers, functions such as customer or business support are frequently automated in order to maintain low scaling costs.

Consequently, there is a tendency towards market concentration around a few big platforms ('market tipping'), where the biggest entry barrier for new competitors is attracting a sufficient number of users on each side of the market. This translates into having a small number of large platforms intermediating transactions (and access to consumers) for a large number of smaller business users, for each type of platform and sector.

Direct and indirect network effects also exist in the offline world. 124 However, research shows significant differences of scale and greater asymmetries induced by network effects in the online

117 Statista – worldwide desktop market share of leading search engines .:

118 StatCounter GlobalStats

119 Ibid.

120 A recent Flash Eurobarometer among European SMEs showed that around 37% sell their products or services online, with

42% of these online sellers using third-party online market places to do so, Flash Eurobarometer 439 'The use of online marketplaces and search engines by SMEs' of April/June 2016 . Also, latest Eurostat figures show that 39% of all European businesses used online social media in 2015, with social networks being the dominant outlet. 121 OECD/World Bank/Facebook, Future of Business Survey, Trade Report , July 2017, page 8.

122 Lendle, Andreas and Olarreaga, Marcelo and Schropp, Simon and Vézina, Pierre‐Louis, There Goes Gravity: eBay and

the Death of Distance (March 2016), The Economic Journal, Vol. 126, Issue 591, pp. 406-441, 2016.

123 Synopsis Report on the Public Consultation on the Regulatory Environment for Platforms, Online Intermediaries and the

Collaborative Economy.

124 Arguably, the more suppliers a supermarket (chain) works with, the more products it will have on sale and the more interesting it becomes for customers, which in turn makes the supermarket (chain) more attractive for suppliers. However, the

world 125 . In certain specific areas such as online publishing, detailed models have demonstrated in quantitative terms that online network effects far outstrip their offline counterparts. 126 Datadriven advantages additionally reinforce the concentration and dependency, as explained in the next Section. 127

2.2.3 Platforms benefit from a data-driven competitive advantage

The virtuous circle of online platforms' growth can also be fuelled by data-driven indirect network effects 128 . Successful platforms can have access to large quantities of fine-grained consumer and business user data, and develop state-of-the-art data and analytics infrastructures to draw intelligence and market strategies out of the insights they obtain. The more users a platform has on each side of the market, the larger the scale of the collected data. The more varied the services offered to a single customer (e.g. buying products, intermediating communication, social networks) the richer the data collected. The combination of scale and variety improves insights e.g. about user profiles and preferences, and may reinforce the 'winnertakes-most' dynamic 129 .

Moreover, one of the key factors that allow platforms to attract users and encourage consumer loyalty is the convenience of use and quality of service: they improve recommendation engines, adjust the matching mechanisms to reflect individual consumer preferences and make it easier to find the right product. Platforms can also enable users to build their online reputation through rating and review systems. All of these features are built and improved through the use of high quality, variety and volumes of data. Consequently, the largest players on each market are also best placed to deliver the best user experience. This can create positive data-driven feedback loops leading to increased returns to scale, scope and network effects, thus accelerating platforms' development and creating a virtuous circle of growth 130 .

2.2.4 Imbalanced bargaining power and dependency of business users on online platforms

The market dynamics described here-above, i.e. a growing intermediation of transaction through online platforms, strong indirect network effects fuelled by data-driven advantages by the online platforms, can lead to an increased dependency of businesses on online platforms as quasi 'gatekeepers' to markets and consumers. While not an issue in itself, this exposes business users to potentially harmful trading practices described earlier in Section 2.1.1.

This tendency can be exacerbated by the imbalance of power in a business user – platform relationship. Indeed, a small number of medium-large platforms intermediate the biggest shares

physical limitations in the number of both suppliers and customers in a supermarket are not comparable to e-commerce market places, where there are virtually no limitations to either customers or traders. 125 Detailed analysis of network effects in the online world in the JRC Scientific and Technical Reports, Institute for Prospective Technological Studies Digital Economy Working Paper 2016/05, ' An Economic Policy Perspective on Online Platforms ', 2016.

126 Michigan Ross School of Business Working Paper No. 1248, ' Quantifying Cross-Network Effects in Online C2C

Platforms ', September 2014, page 19, , for a comparison of the platforms Taobao and the Yellow Pages.

127 In 'Platform Revolution' , Parker, Van Alstyne and Choudary (W. W. Norton & Company, 28 March 2016) contrast digital platforms with conventional “pipeline” businesses that have dominated industry for decades. Pipeline businesses create value by controlling a linear series of activities—the classic value-chain model. Inputs at one end of the chain (e.g. materials from suppliers) undergo a series of steps that transform them into an output that is worth more: the finished product. The engine of the industrial economy remains supply-side economies of scale. The driving force behind the internet economy, conversely, is demand-side economies of scale (network effects). The larger the network, the better the matches between supply and demand and the richer the data that can be used to find matches. 128 See, for example: Prufer, Jens and Schottmüller, Christoph, ' Competing with Big Data ', Tilburg Law School Research Paper No. 06/2017, 16 February 2017. 129 McKinsey Global Institute, ' The Age of Analytics: Competing in a Data-Driven World ', December 2016, chapter 1. 130 See, for example, OECD, ' Data-driven innovation: Big data for growth and well-being ', 6 October 2015.

of transactions in several categories of B2C platforms. 131 This asymmetry between the relative market strength of a small number of leading platforms – not necessarily dominant in the sense of competition law – is combined with a highly fragmented supply-side of many small business users, with the exception of those areas where the natural number of suppliers is limited (e.g. airline ticketing). The final report on the E-commerce Sector Inquiry of the Commission 132 revealed that the overwhelming majority (89% for the EU as a whole) of businesses selling via online marketplaces generated each an annual turnover lower or equal to € 50 000 in 2014, irrespective of the Member State in which they were established. In December 2016, there were 724 000 active developers developing for Google Play, 494 000 for iOS App Store and 69 000 for the Amazon Appstore. 133 Similar trends are seen in the hotel industry, where some 200 000 hotels and 1.8 million cafés in Europe are selling their services on platforms 134 . 92% of these establishments employ fewer than 10 people 135 , while around 60% of hotels have fewer than 25 rooms 136 .

Typically, smaller business users have no ability to organise themselves and negotiate better terms either individually or collectively with the online platform. 137 They generally need to adhere to the terms and conditions pre-set by the platform.

In addition, a study by the JRC 138 shows how the majority of business users multi-home within each platform-segment. The study explains that platforms have little incentive to focus their business strategy in attracting business users rather than customers. 139 The economic literature characterises such dynamics as 'competitive bottlenecks' 140 , where platforms compete aggressively for the buyers, often subsidizing that side, and recoup the costs through higher prices, or lower quality of service on the seller side. 141

Given that business users appear to be commonly multi-homing, switching among platforms could theoretically appear as a solution for a business user experiencing problems with the platform on which he is present. However, this is not necessarily the case since business users engage in multi-homing to reach the maximum number of consumers who single-home (at least for a specific purpose) on different platforms. Switching between platforms could not allow business users to sell to an optimal number of consumers, thus negatively impacting their turnover and their ability to optimise network effects and scale-up possibilities. Multi-homing does not necessarily, therefore, diminish the importance of the platforms' gateway function and does not allow business users to be more independent vis-à-vis platforms – it rather appears a symptom of the bottleneck theory, as explained above.

131 The JRC Scientific and Technical Research Reports, Digital Economy Working Paper 2017-04, 'The competitive landscape of online platforms' . Annex 7 presents a more in-depth market description for a number of sectors. 132 COM SWD(2017), 154 final , Commission Staff Working Document accompanying the Final report on the E-commerce Sector Inquiry ( COM(2017), 229 final ), paragraph 449, 10 May 2017. 133 Number of new developers by year in 2016 - App figures Blog , 'App stores start to mature- 2016 year in review', 24 January 2017.

134 HOTREC Hospitality Europe, Annual Report 2016/2017, page 5. HOTREC is an association representing hotels, restaurants and cafes at European level currently having 40 member associations from 29 European countries. 135 HOTREC Hospitality Europe, Facts and Figures. 136 Eurostat , data set tour_cap_nat. 137 Repeatedly reported in all workshops and interviews organised with business users throughout the fact-finding supporting this Impact Assessment. 138 The JRC Scientific and Technical Research Reports, 'Quality discrimination in online multi-sided markets', forthcoming. 139 Multi-homing, i.e. the parallel use of competing online trading platforms, is common, ECORYS 2017 (see footnote 3). Similarly, nearly two-thirds (64%) of recent comparison site users said they used multiple digital comparison tools, i.e. they multi-homed, the last time they searched for a particular product, UK CMA, Digital Comparison Tools Market Study , update paper, 28 March 2017, page 43. 140 Armstrong, Mark, ' Competition in two-sided markets' (2006) RAND Journal of Economics Vol. 37, No. 3, Autumn 2006. 141 Similar market characteristics were described in relation to digital comparison tools and described in detail in UK CMA, Digital Comparison Tools Market Study , update paper, 28 March 2017.

2.2.5 Business users fear retaliation

Businesses fear commercial retaliation in the case of complaints against the platforms 142 , amplified by the relative dependency and asymmetries. This leads to the impossibility of estimating with precision the scale of the harm, very likely significantly underreported. This underreported friction in platform-to-business relations is set to increase with growing online intermediated trade and may affect business users' in online platforms going forward.

For example, in the online hotel booking segment, a recent market investigation by 10 EU competition authorities and the Commission also found that hotels' fear of retaliation, which retaliation could for example take the form of a less favourable display or the loss of preferred partner status, was quoted by 33% of responding hotels as one of the reasons they maintained the same price between the two most important online travel agents. 143 Other data sources indicate that in the context of online market places, 60% of sellers are fearful of being banned from online platforms. 144

2.3 Consequences

If extrapolated to the 1 million EU businesses selling goods and services via online platforms, the findings of the study carried out for the European Commission 145 and explained above would show that P2B issues tend to impact a large number of business users, e.g. 460,000 enterprises would encounter problems, 200,000 enterprises would consider terms and conditions unfair, and more than 50,000 would encounter issues related to search and ranking. In addition, almost one third of the issues encountered would remain unresolved.

2.3.1 Direct loss in sales through platforms

Nearly half (46%) of all businesses experienced problems with varying gravity and/or disagreements with online platforms 146 , according to a study for the Commission. Amongst the causes listed by business users, several are regarded and analysed in this Impact Assessment as potentially harmful commercial practices. The study also surfaced evidence from business users illustrating the impact of these trading practices on sales through platforms. For example, sudden changes in terms and conditions without sufficient time to adapt led to significant reduction in sales ranging from 20% to 95%. 147 Multiple business users flagged the danger of delisting and suspension for their business, indicating that the viability of the business would be at risk and reported loss of turnover of up to 10% for several weeks or months 148 and, as reported

142 The fear of commercial retaliation in the case of complaints against stronger parties is a common phenomenon in commercial B2B relations: in the retail sector, 87% of suppliers were found not to take action against potentially harmful trading practices beyond a discussion with their customer. Almost two thirds (65%) of these did not take action due to a fear of retaliation (see 'Green Paper on Unfair Trading Practices in the business-to-business food and non-food supply chain in Europe,' COM(2013), 37 final , January 2013). Similarly, in the area of late payment, 39% of respondents to an ex-poste evaluation study of the Late Payment Directive ( ENTR/172/PP/2012/FC, November 2015 ) even mentioned that maintaining good commercial relationships was the main reason for not exercising their rights. 143 ' Report on the Monitoring Exercise Carried out in the Online Hotel Booking Sector by EU Competition Authorities in 2016, published 6 April 2017, paragraph 9. 144 Based on a 2016 survey conducted by Webretailer, a website for businesses who sell through online marketplaces claiming to have circa 20k affiliates worldwide. 145 ECORYS 2017 (see footnote 3). 146 ECORYS 2017, Executive summary (see footnote 3, page ix). 147 It is assumed for the purposes of this estimate that business users would be able to adapt to the changes within one to two months and that the loss in turnover is only temporary. These are conservative assumptions since they disregard the risk of permanent loss of market share, implying that the turnover would not increase back to the previous level once the business has adapted its business model to the change in terms and conditions. 148 For the purposes of the estimate, the business users concerned are assumed to experience a turnover loss of 10% during two to three months.

in anecdotal evidence, having to lay-off up to 20 employees due to a suspension 149 . It is assumed that the negative impacts of ranking 150 lead to a loss in yearly turnover of between 1% and 2%, most of which is permanent due to the difficulty in redirecting sales to other channels. These assumptions have been applied to the total turnover in the different sectors considered, 151 but exclude the issue of ranking in online general search engines for which insufficient evidence is available at present to allow a robust quantification of any systematic negative impacts.

On that basis, the reduction of sales through platforms for EU business users caused by the practices at stake can be estimated to amount to between € 1.27 and € 2.35 billion per year.

2.3.2 Further dampening effect through lack of trust

The dependency-induced fear of retaliation leads to an underreporting of actual problems, while individual potentially harmful trading practices have an important knock-on effect on the wider trust in online platforms. The subsequent uncertainty experienced by business users leads to an economic under-utilisation of the potential of the online platform economy as business users are reluctant to enter into or expand their business relationships with online platforms.

A recent industry survey showed that more than 60% of sellers on the biggest e-commerce marketplace fear being banned. 152 Another industry survey also found that 25% of app developers view the app stores themselves as their greatest threat. 153 Finally, a recent Eurobarometer survey found that 19% of Internet users in the EU do not trust that the search results provided to them are the most relevant to their query. 154

If the additional dampening effect of this uncertainty and fear can be assumed to lead to a further reduction of total sales by business users on marketplaces by a conservative 1-5% 155 , an

149 This statement, while showing the important negative impact that delisting/account suspension may have on business users, does not put into question the need for platforms to proceed to such delisting/account suspension for legitimate reasons. Cf. Section 8, one but last paragraph. 150 No assumptions were made for data and MFN clauses to ensure that the direct loss has not been inflated. The direct loss figure is thus a conservative estimate since the negative impact of these two practices has not been accounted for. 151 The assumed loss was directly applied to the EU turnover generated by business users on e-commerce marketplaces (€55 billion in 2016 according to Euromonitor International, Passport Database 2016 Edition), the retail value of online travel intermediaries including air, attractions, hotels, other lodging and short term rentals and car rentals in 22 Member States (€73.4 billion in 2015 according to Euromonitor International, Passport Database 2016 Edition), the aggregate revenue generated by European app developers on app stores (€16.5 billion in 2014 according to Vision Mobile, The European App Economy 2014 .) Conversely, social networks rather seem to have an indirect effect on the other categories in the sense that they are used by businesses to increase brand awareness, to expand their potential customer bases and to promote sales, for instance by stimulating app usage. The practices listed above taking place on social networks are therefore, assumed to magnify the impacts on the other categories, in the proportion of internet traffic from social networks to the other categories. 152 Based on a 2016 survey conducted by Webretailer, a website for businesses who sell through online marketplaces claiming to have circa 20k affiliates worldwide. 153 Application Developers Alliance, ' Competition in the Mobile App Ecosystem ' survey of 673 mobile app publishers and developers of September 2016. Noteworthy is that platforms such as Google and Facebook are themselves members of the association behind this survey. 154 Special Eurobarometer 447, Online Platforms of June 2016. 155 This range is an assumption made by the Commission services. Data demonstrating the impact of lower trust on business users of online platforms is difficult to obtain due to the scale and diversity of the online platform ecosystem. It can be assumed however that business users respond to trust issues in ways similar to individual consumers. A recent UNCTAD study analysing the response of consumers to lower trust in e-commerce demonstrates that lower trust leads to 15% of consumers engaging in less online transactions, 13% making less purchases and 10% using online platforms less often (see: http://unctad.org/en/pages/newsdetails.aspx?OriginalVersionID=1465). On average, business users are likely to be more sophisticated and knowledgeable compared to individual consumers. It can therefore be assumed that the negative effect of the lack of trust on the activities of business users on online platforms is significantly lower than in the case of consumers. Therefore the 'chilling effect' of lack of trust due to the threat of being subject to the problematic practices is estimated to lead potential sellers to opt for other sales channels and lower the sales on online platforms by 1-5%. Further economic research is on-going to cross-check this assumption with the documented effects of increased protection of businesses and consumers in EU Member States.

estimate of € 2.7 to € 13.5 billion 156 of turnover not realized on online platforms can be arrived at.

Combined with the estimates of direct losses in section 2.3.1, these figures lead to a total estimated reduction in platform turnover by business users of € 3.97 to € 15.85 billion per year. Assuming that online platforms charge, on average, a 10% commission, online platforms would forego commissions of between € 0.4 and 1.6 billion.

These figures are consistent with independent estimates by the JRC. The aggregated impact in the EU economy due to the uncertainty linked to opaque practices by online platforms is estimated by the JRC to be in the range of € 2 to € 19.5 billion per year 157 .

2.3.3 Fewer EU cross-border sales

Online platforms are of great importance for businesses' cross-border sales. On average only 9% of retailers in the EU sell online cross-border today 158 , while more than 50% of SMEs selling through online marketplaces sell cross-border. Online marketplaces thus facilitate cross-border sales in the Digital Single Market especially by the smallest retailers. They reduce trade costs for SMEs, in particular those related to differences between languages and regulatory frameworks, and at the same time provide them with a global presence and reach previously reserved to large (multinational) retailers 159 . It is estimated that the 'distance effect' on trade flows (a measure of trade frictions) is 65% smaller on an e-commerce market place than for total trade due to the effect of the online marketplace in reducing information frictions associated with geographical distance 160 . The importance of online intermediated trade for SMEs is well illustrated by the following example: where eBay created an online webpage integrated into the eBay platform for the small retailers in the small German towns of Diepholz and Mönchengladbach, the 79 retailers participating in these towns’ eBay platforms sold more than 87 500 items with a total value of more than € 3.2 million, and delivered to 84 countries in a year.

Since selling on online platforms reduces the costs of exporting and makes the seller's goods or services easier accessible to customers in other countries, when sellers are forced to divert sales away from online platforms to other channels the share of cross-border sales is likely to fall. Individual sellers are likely to find it difficult and expensive to replicate in-house the services (like product offer translation, multi-lingual customer support, international shipping, regulatory compliance) which the online platform can supply at a significantly lower price due to its scale. Similarly, if business users are reluctant to enter into or expand their business relationships with online platforms, they will most likely sell less cross-border. As a consequence, factors limiting the take-up of online platforms by third-party sellers also limit the growth of cross-border sales.

2.3.4 EU consumers have more limited choice

It is widely acknowledged that online platforms have recently dramatically contributed to increases in consumer access to goods and services, especially cross-border. Around 60% of private consumption and 30 per cent of public consumption of goods and services related to the

156 This figure is a conservative estimate based on the figure of EUR 270 billion of private and public consumption realised via online intermediaries in 2014. Due to the dynamic growth of the market, it is likely to be significantly higher. The conservative nature of the estimate has further been guaranteed by excluding the area of online general search engines, where a significant lack of user trust nonetheless has been observed. 157 Duch-Brown, Nestor, 'Platforms to business relations in online platform ecosystems', Joint Research Centre, Sevilla (publication forthcoming). 158 Eurostat, data set isoc_ec_eseln2. 159 Copenhagen Economics - Economic effects of marketplace bans , a study prepared for eBay, November 2016. 160 An Anatomy of Online Trade: Evidence from eBay Exporters , by Andreas Lendle et al. (2013).

total 'Internet economy' go via online platforms 161 . The value of goods and services purchased by private households and the public sector via online intermediaries was valued at € 270 billion in 2014, corresponding to 2.5 per cent of the total final consumption in the EU-28 countries 162 .

That means that if business users suffer a loss of sales through platforms or if they choose to limit their presence on platforms for reasons of fear or lack of trust, consumers would be more likely to be faced with reduced choice of competitive products/services as compared to a situation where business users would be able and prepared to reap the full potential of the platform economy. As described in the previous Sub-section, this applies especially to crossborder sales of smaller companies.

2.3.5 Innovation capacity for businesses may be undermined

Online platforms are major investors in innovative technologies such as artificial intelligence, internet of things and data analytics 163 . At the same time, enterprises that depend on online platforms to reach customers are de-incentivised and sometimes stopped from innovating in areas that would compete directly with the intermediary's role – e.g. developing online market analysis and strategies based on consumer behaviour and preferences on which many of the platform innovations are based. While this may in certain cases spur innovation from the side of the businesses that use these platforms 164 , they are unlikely to be able to innovate enough to become independent, particularly as platforms invest heavily in innovation in order to cement their market power in their relevant markets 165 .

The consequence is that the incentives to innovate for smaller companies shift to complementary areas to platforms, which does not – in itself – relieve the dependency and the exposure to some of the potentially harmful practices. At present no robust quantitative estimates for these innovation dynamics exist, but the acquisitions and partnerships of online platforms with major deep tech businesses point to an increasing differential in innovation capacity of the intermediaries 166 . Where conditions for accessing and using data are unclear, this can have a chilling effect on business users' investment in developing their capability or in contracting thirdparty services for data-driven innovations 167 . With the rapid developments in data analytics and data-driven business intelligence, access to data is an evolving problem.

2.4 How would the problem evolve absent intervention?

B2C e-commerce revenue in Europe is forecast to amount to ~ € 250 billion in 2017, up from € 108.7 billion in 2012 168 and expected to grow steadily in the future, at rates much higher than the average growth rate of the economy. This in itself makes this sector particularly important to the overall EU economy. This growth is primarily driven by fast evolving consumer demands for ecommerce in the EU. Given, explained above, the platforms' incentives to grow, big platforms can be expected to continue expanding. The overall growth rate for online intermediaries is of

161 Copenhagen Economics , Online Intermediaries: Impact on the EU economy 2015.

162 ibid.

163 Confirmed in FABERNOVEL, Gafanomics: New economy, new rules , 2014. See also ‘Towards a thriving data-driven economy’, COM(2014), 442 final , 2 July 2014 164 UK House of Lords, Online platforms and the Digital Single Market , 10 th Report of Session 2015-16 , page 30f. 165 Batura/van Gorp/Larouche, Online platforms and the Digital Single Market – a response to the call for evidence by the House of Lord's internal market sub-committee , page 6. 166 E.g. in the gaming industry, cf. Atomico report; AI acquisitions. 167 The findings of the Commission workshop on data confirmed the importance of online platforms for innovation and pointed to data skills asymmetry and related unexploited potential. Commission workshop of 16 October 2016: " Business-to href="https://ec.europa.eu/digital-single-market/en/news/data-related-aspects-business-platform-trading-practices-workshop-report">business relationships in the online platform economy-data access, (re)use and portability ". 168 Statista, B2C e-commerce revenue in Europe from 2012 to 2017 (in billion euros).

around 10% per year since 2013 (based on 2012 estimates) 169 and exceeds by far the growth rates in other sectors. The growth of online intermediaries is expected to continue over the coming years facilitated by an increasing use of cloud computing and a rapid growth in ecommerce. It is estimated that 40% of retail online sales will be conducted through online marketplaces by 2020. The use of search engines is also expected to grow, as a result of growth in the number of websites, which have currently reached over 1.3bn, up from 207 million in 2010. 170 As a result, platforms will increasingly develop the potential to become 'gatekeepers'.

This would increase platforms' bargaining power and business users' dependency. The likely aggravated P2B issues could be expected to lead to further regulatory intervention across the EU Member States and undermine business user trust. In a similar way, the number of cases being considered by EU Member States’ courts leads to further divergent or unpredictable outcomes. For example, in Germany and in France there are cases pending in courts, which relate to P2B fairness standards. The outcome in those cases differed from first to second instance and it is unknown what highest court shall render as its final judgment. 171 Furthermore, the fragmentation which results from different legislative approaches between Member States concerning terms and conditions, is exacerbated due to the varying levels of enforcement between them. For instance, in France, there is a political commitment to ensure the efficient enforcement of the P2B legislation in place. In December 2017 the French authority for competition, consumers and repression of fraud, DGCCRF 172 opened an investigation into P2B clauses and their compliance with the legal standards. It has already obliged some companies to remove MFN clauses from their terms and conditions 173 . Whereas in Germany – where supervision of MFN clauses takes place only from a competition law angle - a recent judgment concluded the validity of both wide and narrow MFN clauses under competition rules. 174

At the same time, the market position of the existent larger platforms would strengthen (due to data-driven network effects), which coupled with the single market fragmentation and more limited growth (because of reduced trust in the platform economy) would make market entry difficult for new platforms. If less business users decide to be present on a limited number of platforms, this could lead to reduced quality and choice for consumers in the longer term.

2.5 Conclusion of Problem Definition

EU businesses cannot exploit the full potential of the platform economy because of issues in the platform-to-business relations and emerging re-fragmentation of the single market.

Business users active on platforms face a number of potentially harmful trading practices for which there is a lack of effective redress. According to a study carried out for the European Commission, these trading practices would concern a large number of business users. The results of the study show that 46% of business users encounter problems in their relation with platforms while this percentage is higher (75%) for business users realising more than half of their turnover on platforms. Almost one third of issues remain unresolved while 29% are solved only with difficulties. The potentially unfair trading practices listed in this section risk gradually undermining business trust in the platform economy. Trust is primordial to the platform economy since it allows increasing the number of users on both sides thus optimising data-driven

169 Euromonitor International, Passport Database 2016 Edition.

170 Internet Live Stats, http://www.internetlivestats.com/total-number-of-websites/

171 See overview of case in Annex 8.5: Emerging national legislation for the platforms' environment: relevance for P2B; and in Ernst&Young study, Chapter 2.1.7., Pending cases. 172 Direction générale de la concurrence, de la consommation et de la répression des fraudes 173 See here .

174 OLG Düsseldorf, judgment of 4 December 2017 the OLG Düsseldorf (1e Kartellsenat | VI-U 5/17 (Kart), U (Kart) 5/17)), future revision by the level highest court (BGH- Bundesgerichtshof) is possible.

network effects which fuel online platforms' growth. It can be reasonably assumed that potentially unfair and non-transparent P2B practices are therefore not only detrimental for business users (since they lead to direct loss in sales) but could also negatively impact the growth of the online intermediation sector and reduce platform operators' revenues (through unrealised commissions). The long-term sustainability of the platform economy is therefore closely linked to issues encountered by business users in their relations with platforms.

Platform operators are increasingly faced with emerging national legislations which start fragmenting the naturally cross-border market for online intermediation in the EU. The uncoordinated adoption of national legislations - whether platform-specific or covering B2B issues in general but applicable to platform businesses – may result in divergent regulatory measures across the EU and carry the risk of hampering online platforms' ability to scale up. The EU platform economy is of intrinsic global nature and is by definition cross-border. Scaling-up is core to platforms' business strategies as it allows for stronger network effects. Start-up and small online platform operators would be the most heavily impacted by a fragmented market for online intermediation because of their more limited capacity to comply with different national rules. If emerging re-fragmentation expands to other Member States (which could be expected given the growing online intermediated trade and the increasing importance of platforms as a gateway for SMEs to access new markets), it would negatively impact the emergence of new platforms in the EU. Without facing competition from new market players, existing platforms would reinforce their market strength. This would further increase their bargaining power and could be expected to increase business users' dependency and the size of the problem.

Such dynamics would be detrimental to the Digital Single Market in terms of innovation, growth and consumer benefits. Online intermediated trade has important impact on the digital economy. The less businesses use platforms, the less they seize innovation opportunities and the less they are able to embrace digital transformation. P2B issues are hampering the potential of the platform economy thus preventing it to fully contribute to a well-functioning Digital Single Market.

3 W HY SHOULD THE EU A CT ?

3.1 Legal Basis

Given (i) that the initiative constitutes a core part of the Digital Single Market strategy, (ii) the intrinsic cross-border nature of online platforms, and (iii) the risk of further regulatory fragmentation regarding online platforms, Article 114 TFEU (Title VII Common rules on competition, taxation and approximation of laws) is identified here as the relevant legal basis for this initiative. Further explanation is provided in Annex 1.10.

3.2 Subsidiarity: Necessity of EU action

The intrinsic cross-border nature of online platforms implies that the objectives cannot be reached effectively by Member States alone. Leading online platforms such as Booking.com, Facebook and eBay are legally established in one Member State, but provide access to almost the entire EU population, both from their place of normal residency as well as while travelling across the EU. Importantly, a platform such as Facebook is at the same time used for commercial communications by 90% of the respondents to the Commission's fact-finding on platform-tobusiness relations. EU action, therefore, constitutes the only way to ensuring that the same rules apply to online platforms and the business users active on them, also regardless of the law and forum identified in contractual terms. On the specific set of issues described here, the European Council 'underlined the necessity of increased transparency in platforms' practices and uses' 175 as part of a future oriented regulatory framework for the EU.

3.3 Subsidiarity: Added value of EU action

EU action would ensure that business users trading on platforms can fully leverage the potential of the Digital Single Market, as the same P2B protection will apply to them regardless of which Member States they sell into. It will also facilitate the scaling-up of platform start-ups, as compliance costs are lowered and legal certainty enhanced.

Furthermore, EU action would avoid further fragmentation of the Single Market into different, potentially contradictory frameworks – including the resulting jurisdictional issues. This is expected to increase the incentives for new platforms to develop.

The initiative would, therefore, contribute to releasing the full potential the platform economy could offer in terms of increased competitiveness, innovation, growth and jobs.

Further explanation on subsidiarity is provided in Annex 1.11.

4 O BJECTIVES : W HAT SHOULD BE A CHIEVED ?

Figure 2: Objectives tree

4.1 What are the general policy objectives?

The general policy objective is to ensure the functioning of the Digital Single Market in line with Article 114 TFEU, and considering the inherent cross-border nature of the online platform economy and the dramatically increasing role that online platforms play in intermediating access to the Digital Single Market.

175 European Council Conclusions, 19 October 2017, ST 14 2017 INIT .

Releasing the full potential of the online platform economy therefore constitutes a broad general objective – implying that more businesses operate via online platforms in general, and that more consumers use online platforms to access goods and services.

This requires a fair, trusting and innovation-driven ecosystem around online platforms across the EU, in which business traders have the necessary safeguards to prevent harm from unfair trading practices, general lack of redress, and regulatory fragmentation across the EU.

4.2 What are the specific objectives?

The initiative pursues the overarching objective of establishing a fair, trusting and innovationdriven ecosystem around online platforms in the EU. In particular, it aims at the following specific objectives:

4.2.1 Ensuring a fair, transparent and predictable treatment of business users by online platforms (specific objective 1)

The first specific objective of the intervention would be to provide a clear set of minimum standards that platforms need to provide, notably in terms of transparency on those aspects of their relationship where asymmetries of bargaining power are particularly pronounced. To this end, the initiative aims at defining basic rules for online platforms and their business users. The objective is to facilitate the business users' relations with online platforms thus allowing businesses to concentrate on their core activities and to fully grasp the opportunities offered by the various forms of online intermediation on which businesses rely to access markets. This in turn should lead to a predictable business environment for those enterprises which use online platforms to reach consumers.

4.2.2 Setting effective and agile redress for businesses, adaptable to the evolving market

(specific objective 2)

The second objective is to ensure the enforcement of the above rules by appropriate redress mechanisms, all internal, external and judiciary. These mechanisms should ensure the necessary speed, independence, affordability and anonymity, to overcome the observed regulatory gap in terms of ineffective internal redress offered by platforms, the lack of an external redress mechanism and the limited use of judicial remedies. The aim is to closely monitor the functioning of the mechanisms to ensure their effectiveness and adapt them to the changes observed in the platform-to-business relations.

4.2.3 Preserving a predictable and innovation-friendly legal environment for online platforms within the EU, without placing undue administrative burden on platforms (specific objective 3)

Reaching a critical mass is essential for platforms' business model. The third objective is therefore to define clear requirements at EU level for online platforms, thereby allowing online platforms to operate at a larger EU scale without creating unnecessary and disproportionate burdens. This would set the basis for more consistency in national legislations by providing a common framework of high level rules within which Member States can set national legislations if needed. This would entail helping start-up platforms to scale up by providing a clear overview of the legal requirements they have to comply with, thus ensuring greater regulatory certainty.

Operational objectives are defined in Section 0.

4.2.4 How do the objectives link to the problem?

The overarching objective of a fair, trusting and innovation-driven platform ecosystem is directly linked to the two main problems identified in the problem tree, i.e. emerging difficulties for platforms to operate in the Single Market and impossibility for EU business users to fully exploit the potential of the platform economy.

A healthy platform ecosystem would help address the issues business users face in their relationship with platforms. A predictable regulatory environment for both online platforms and their business users (specific objectives 1 and 3) coupled with effective redress mechanism for business users (specific objective 2) would contribute to releasing the full potential of the platform economy. The overarching objective of establishing a fair, trusting and innovationdriven online platforms ecosystem in the EU would thus contribute to the better functioning of the internal market. A healthier platform ecosystem would also help prevent the fragmentation of the internal market which could otherwise occur as a result of uncoordinated efforts by Member States to solve platform-specific issues at the national level.

The more specific objectives 1, 2 and 3 allow addressing the different problems identified in the problem tree. All three specific objectives aim at creating the appropriate regulatory tools to safeguard the single market dimension of the platform economy and address the emerging fragmentation of the single market.

A more fair, transparent and predictable treatment of business users (specific objective 1) coupled with effective and agile redress possibilities adaptable to the evolving market conditions (specific objective 2) would act on the imbalance of bargaining power and help addressing potentially harmful trading practices and address the lack of effective regulatory tools against potentially harmful trading practices. This could be expected to lead to less P2B issues, which would in turn prevent any need for intervention at national level. This would help preventing further legal re-fragmentation of the single market across Member States, which could create future obstacles to cross-border trade and jeopardise the functioning of the Digital Single Market.

In addition, such improved business environment may also be expected to increase business users' trust in the platform economy and lead to an increased use of online platforms. Given the intrinsic global and cross-border nature of online platforms as well as the importance of online intermediated trade for SME's exports 176 , such increased use of platforms may be expected to lead to more cross-border sales, thus reinforcing the single market dimension.

The objective of a predictable and innovation-friendly legal environment for online platforms within the EU without placing undue administrative burden on platforms (specific objective 3) aims at ensuring that any set of rules to the benefit of business users will be proportionate and non-intrusive for platforms. Possible new rules at EU level - such as the ones presented and assessed in this Impact Assessment - will provide more regulatory predictability for platforms at EU level. It would thus allow preserving the existing cross-border dynamics of the platform economy by setting a common framework for Member States' possible regulatory approaches.

4.3 Consistency of the objectives

This initiative aims at ensuring a fair, predictable and ultimately trusted legal environment for business users and B2C online platforms alike that will limit the occurrence and/or the impact of

176 A recent survey among 49 081 SMEs active on Facebook showed that nearly half of exporting SMEs (45%) report that more than 75% of their international sales depend on online tools or platforms. The cross-border effect of online intermediated trade is further demonstrated in Section 2.3.3.

the problematic P2B practices identified in Section 2.1 thereby safeguarding trust in the platform economy and preventing further legal fragmentation of the Digital Single Market.

The initiative thus contributes to the goals of Digital Single Market Strategy by creating a clear and stable legal environment for B2C online platforms and their business users to tackle market fragmentation and allow all players to tap into the new market dynamics under fair and balanced conditions. 177

The initiative is also consistent with a number of other EU policies and rules. It complements EU policies and rules in the area of consumer protection 178 by aiming at providing a targeted fairness framework also for certain B2B relations, namely for the relations between B2C online platforms and their business users. To this end, it also builds on relevant findings made as part of the Fitness Check of the EU's consumer protection acquis, which exercise explicitly excludes any follow-up in the area of B2B or P2B relations. 179

It also complements Regulation 80/2009 on a Code of Conduct for computerised reservation systems 180 , which contains a set of obligations for a specific type of B2B platforms (computerised reservation systems, also called Global Distribution Systems, GDS) that allow travel agencies to compare information and book tickets from a large number of travel service providers worldwide. The initiative, although building on a different design for conflict resolution, is not in friction with ODR-Regulation 181 and the ADR-Directive 182 . This initiative shares the objectives of those instruments, i.e. to offer a low cost and accessible out of court conflict resolution. However, achieving the same objectives within the P2B-relationship as those of the ODR- Regulation and the ADR-Directive requires a more targeted design for conflict solution. More particularly, this initiative builds on the presumed incentives of platforms to settle disputes with their business clients. The design for conflict solution in this initiative is also more specific if compared to the one of the ODR-Regulation and the ADR-Directive because of the specificity of the problems identified.

Finally, it complements the EU competition rules, which allow tackling anticompetitive behaviour and mergers the potentially harmful trading practices identified in Section 2. 1, as explained further in Annex 8.3.

To the extent that the fair and trustworthy legal environment that this initiative aims at ensuring would involve increased access to and use of personal data by business users of B2C online platform, such access and use would have to be compliant with the requirements of the General Data Protection Regulation ('GDPR'), in particular the principles of purpose limitation and lawful processing, and with Article 8 of the Charter of Fundamental Rights of the European Union ('CFR'). Where platforms act as processors of personal data 183 an obligation of increased transparency in changes to terms and conditions will also support the implementation by

177 See footnotes 1 and 2.

178 In particular, the UCDP, UCTD and the Consumer Rights Directive ( Directive 2001/83/EU on consumer rights), but also the Commission's proposals for a (i) Digital Content Directive (Proposal for a Directive on certain aspects concerning contracts for the supply of digital content, COM (2015), 634 final ), and (ii) a Directive on contracts for online and other distances sales of goods ( COM(2015) 635 final ), which aim at removing contract law related barriers to the Digital Single Market, adjusting the consumer protection legislation to the online environment and increasing consumer trust. 179 Inception Impact Assessment for the initiative "Targeted revision of EU consumer law directives" of June 2017. 180 Regulation (EC) No 80/2009 of 14 January 2009 on a Code of Conduct for computerised reservation systems OJ L 35. 181 Regulation 524/2013/EC of 21 May 2013 on online dispute resolution for consumer disputes provides for an online platform via which disputes can be assigned to the specific competent bodies, OJ L 165. 182 Directive 2013/11/EU of 21 May 2013 on alternative dispute resolution for consumer disputes, OJ L 165. 183 This is often the case: usually, the business platform user will be the data controller while the platform acts as the data processor, see definitions in Article 4 (7) and 4 (8) GDPR.

platforms of the new obligations under the GDPR 184 . This increased transparency in the contractual platform –business relationship will, in turn, positively impact on the data subjects' rights related information on changes in the data processing policy of platforms because it will better enable the data controller to keep the data subject informed about data processing issues.

The EU is committed to high standards of fundamental rights. The specific objective of timely, effective and trustworthy redress for business users contributes to enhancing business users' right to an effective remedy and to a fair trial (Article 47 CFR) as far as it would involve improved access to the judiciary. It would be neutral to this fundamental right, if it provided business users with additional out-of-court redress mechanisms, while simultaneously not impeding the platforms' right to take legal steps including going to court. Moreover, the fair and trustworthy legal environment that this initiative aims to create shall balance the business users' and the B2C online platforms' respective freedoms to conduct a business (Article 16 CFR).

5 W HAT ARE THE AVAILABLE P OLICY O PTIONS ?

5.1 What is the baseline from which options are assessed?

Under the baseline option, EU action would continue to be limited to possible ex-post enforcement of the existing competition and consumer protection frameworks in targeted cases with no new rules at EU level. This scenario is described in Section 2.4. In the baseline scenario, the drivers of the problem description will only gather in strength, and inevitably increase the dependency and the relative market strength of online intermediaries over their business partners. In the baseline option, different online platforms will implement different – potentially contradictory – policies for each of the identified problems. Smaller businesses, who (as is demonstrated above) generally need to multi-home to optimise their revenues, will be confronted with a confusing mix of different practices and problems, depending on country of operation or type and brand of online platform. Specifically, there is no incentive at present for market players to provide for effective dispute resolution across the board - and certainly not outside their own platforms.

Concerning the individual problems, it is possible that the strength of evolution of the underlying drivers even increases the range of potentially harmful trading practices. The baseline scenario also implies that no effective, continuous monitoring of the evolution of potentially harmful trading practices, would take place.

Fragmentation across the Digital Single Market is likely to increase as national legislators seek to address largely cross-border issues with national rules, which are likely to target only specific regulatory interests in individual Member States.

5.2 Description of the policy options

The policy options considered are as follows:

− Policy option 0: Baseline – no EU action taken − Policy option 1: Non-legislative approach / pure self-regulation

184 Article 28 (3) GDPR contains mandatory contractual obligations of data processors to enable the data controller to fulfil its obligations of transparency on data processing in relation to the data subject. Those mandatory obligations would then be supplemented by a general transparency obligation in relation to changes of terms and conditions. It would thereby be clear that changes in data processing implying changes on terms and conditions must be communicated in any event in a timely manner- thereby contributing to clarify contractual obligations between the data controller and the data processor under the GDPR.

− Policy Option 2: Co-regulation implying (i) on transparency: legal principles with significant scope for industry implementation, (ii) on redress: co-regulatory cascade of redress mechanisms; focus on industry action, playing on light-touch reputational levers, and (iii) on monitoring: EU Observatory to monitor emerging issues – partly informed by the new legal transparency obligations – and to inform potential future review of initial light-touch Regulation. Policy Option 2 could take the form of one of the four options 2a, 2b, 2c or 2d such as explained in Table 1 below.

− Policy option 3: Mandatory, binding rules for all aspects − Policy Option 4: Extension of existing rules to P2B

The above list of policy options has been identified on the basis of the following approach. First, a range of substantive policy elements have been considered for each specific issue identified in the problem statement, referred to in Annex 10 as 'content' options as they seek to address the subject matter of the problem at hand. Second, these option elements have been assessed on the basis of their effectiveness, cost efficiency and coherence (Tables 1 and 2 of Annex 10). Third, the specific retained measures for each issue identified have been combined in policy options. Finally, a variety of legislative or non-legislative instruments have been considered for the so identified policy options, ranging from self-regulation to co-regulation, or to full mandatory binding measures (see also Sections 5.3 and 5.4). Option 4 departs from this approach since it is an extension of existing rules to the P2B issues identified.

Table 1 below presents for each option the measures aiming at addressing (i) potentially harmful trading practices, and (ii) inefficient redress alongside (iii) the elements that are part of the envisaged monitoring exercise at EU level.

Table 1: Presentation of the policy options considered

Policy option 0: Baseline – no EU action taken

Policy option 1: Non-legislative approach / pure self-regulation

Transparency measures: -Invitation to industry to develop measures of its choice to address the problematic potentially harmful trading practices identified, particularly focussing on developing principles and best practices for changes to terms and conditions, for delisting/suspension. -Encouragement of industry to improve transparency on data policies, differential treatment and auxiliary services.

-Structured dialogues with industry aiming at addressing emerging issues in paid-for ranking, encouraging voluntary standards and private audits.

Redress: Call on the industry (i) to improve their internal complaint-handling mechanisms accessible for business users, and (ii) to set up an external independent redress mechanism at EU level to provide business users with an additional venue for redress.

Monitoring: An EU Observatory of the Digital Platform Economy ('EU Observatory') would be set up, having as part of its mandate to monitor the evolution and emergence of issues related to data access and use by both platforms and their business users. This would include sharing of both non-personal and personal data, e.g. e-mail addresses, with business users, and to what extent business users request access to such data in full compliance with the GDPR.

Policy option 2a: limited scope of legal transparency principles, maximum focus on voluntary industry action

Transparency: Builds on Option 1, but includes legal transparency obligations for platforms on limited issues, i.e. terms and conditions and delisting. Foresees the following measures:

-improve clarity & availability of terms and conditions. -give reasonable notice period before introducing changes to terms and conditions -list the objective grounds for suspension or termination of use of platform -provide a statement of reasons for any decision to suspend or terminate use of a platform, referring to predetermined objective grounds

-invitation to industry to voluntarily explore practical solutions that improve predictability around: the functioning of ranking mechanisms, including the use of any mechanism that allows business users to influence their prominence against remuneration; any preferential treatment of platforms' own products or services; access to personal and other data; the use of MFN clauses, the verifiability paid-for prominence in ranking (relevant to specific e-commerce areas) potentially developing industry standards and proactively running audits and monitoring the functioning of the wider digital advertising space.

Redress:

-Legal obligation for platforms (i) to provide internal complaint-handling mechanisms, with

detailed mechanism to be specified in industry codes, and (ii) to either list a mediator or make reference to organisations providing mediation services set up by platforms, together with a legal obligation to act in good faith in relation to any mediation attempts. -Call on industry to set up an external organisation that can provide industry-specific mediators at EU level to provide business users with an additional means out of court for redress with legal obligations as to their effectiveness. -Invitation to industry to voluntarily explore developing further recommendations on the

internal complaints-handling mechanism in the form of codes of conduct.

-Right for business associations or representative bodies to seek action in court to obtain injunctive relief to ensure high-level legal obligations on redress (internal complaint-handling & good faith mediation) are complied with.

Monitoring: In addition to the tasks in Option 1, the EU Observatory shall monitor the evolution and emergence of issues related to: preferential treatment of a platforms' own products or services; use of MFN clauses by online platforms and test the reasons provided by platforms to justify their use. The EU Observatory shall act as a repository for public reports on the effectiveness of internal complaint-handling mechanism and refusals by a platform to engage in any

mediation attempts. Platforms are required to report in a non-detailed manner on the use of the internal complaint-handling system. A medium-term review clause would be considered.

Policy Option 2b: Co-regulation with horizontal application of legal transparency principles to all trading practices.

The legal transparency principles would extend to all potentially harmful trading practices: clarity

  • availability of terms and conditions, delisting, ranking, discrimination, data and MFNs. Industry would continue to play important role in developing codes of conduct or standards to provide practical solutions to implement these legal principles. Redress and monitoring measures are identical to those under option 2a.

Policy option 2c: horizontal application of legal transparency principles to all trading practices, scope extension to online general search

Transparency: Building on option 2b, the scope of the legal transparency obligation on the core issue of ranking would be expanded to encompass both online platforms as well as online search engines. The role for industry in providing practical solutions for meaningful ranking transparency

in general search would be ensured through developing codes of conduct or standards to provide practical solutions to implement these legal principles.

Redress builds on option 2a, by additionally granting associations or representative bodies of businesses whose websites are indexed by online general search engines the right to seek action in court to obtain injunctive relief to ensure high-level legal obligation on transparency on ranking in online general search is complied with (c.f. Sections 7.2.1 and 7.2.3).

Monitoring is identical to Option 2a (and hence also to 2b).

Policy option 2d: horizontal application of legal transparency principles to all trading practices, scope extension to online general search and targeted legal obligation on email addresses

Transparency: Building on option 2c, this option adds a legal obligation for platforms to give business users the opportunity to ask, in line with the GDPR, for customers' consent to obtain and process their e-mail addresses after the completion of a transaction and the payment of the commission to the respective platform.

Redress is identical to 2c and monitoring to 2a (and hence also to 2b and 2c).

Policy option 3: Mandatory, binding rules for all aspects

Transparency: Building on both Options 2b and 2d, to cover all the legal obligations in Options 2b and 2d, but elaborated to include all legal or technical details, including those left to codes of conduct in Options 2b and 2d above. Option 4 adds further prohibitions and legal obligations in relation to data and MFNs as follows: (i) an obligation for platforms to extend data access rights

to business users for specific categories of data, (ii) a ban on contractual clauses that prevent business users from retrieving and/or using specific types of data outside the platform, and (iii) prohibition of most-favoured-nation clauses (whether on price, availability, quality) for platforms.

Redress: (1) Legal obligation on industry to set up internal complaint-handling mechanisms to include all legal or technical details, including those left to codes of conduct in Options 2b and 2d above. (2) An obligation for Member States to ensure effective enforcement and efficient dispute resolution of the P2B rules by designating competent authorities, who would be capable of imposing sanctions.

Monitoring is identical to Option 2.

Policy Option 4: Extension of existing rules

Transparency measures: (i) include platform-specific practices in the blacklist in the annex of the UCPD; (ii) include platform-specific practices in the grey list in the annex of the UCTD, (iii) include platform-specific practices in the blacklist in the annex of the MCAD.

Redress measures: (i) extend the scope of the UCPD to cover B2B relationships P2B relations; (ii) extend the scope of the UCTD to cover B2B relationships including platform-to-business relations, (iii) extend the content of the MCAD to also cover existing contractual relationships and to broaden

the scope to particularly address platform-to-business relations. Monitoring: extension of the available monitoring mechanisms under consumer law to businesses.

As explained in the beginning of this section, a wide range of other measures (than those listed in the table) have also been considered for the policy options described above to address the identified problems. However, these option elements have been discarded at an earlier stage and not retained for further examination for various reasons that are listed in Table 3 of Annex 10. The most prominent grounds for discarding these option elements were a lack of effectiveness or disproportionality.

While options 2b and 2c mainly differ on the scope of the ranking and the legal standing obligation, this differential element between the two options is important since it allows expanding the benefits of the initiative to online general search. The difference between the two options is thus important since it allows under option 2d (i) covering business users' dependency on online general search ranking which directly influence businesses websites' visibility and Internet traffic, and (ii) granting legal standing to business associations to act on behalf of professional website owners to enforce this transparency requirement only.

As to the single differential element between options 2c and 2d, it is also a significant one. Option 2d adds a legal obligation for platforms to give business users the opportunity to ask, in line with the GDPR, for customers' consent to obtain and process their e-mail addresses after the completion of a transaction and the payment of the commission to the respective platform. While this additional element may from the outset appear as a detail, it is a strong distinguishing feature between 2c and 2d. Due to the data-related aspect of email addresses, the impacts of options 2c and 2d are very different as explained in Section 6 (Impacts) and Section 7 (Comparison of Options).

5.3 Legislative or non-legislative character of options

Option 1 is a non-legislative instrument, e.g. a Commission Communication, calling on the industry to address the identified problems by means of self-regulation. Option 2 – with its four sub-options - embodies a co-regulatory approach including a set of obligations enshrined in a new legal instrument alongside non-legislative measures requesting further actions by the industry. Option 3 comprises full regulation by a new legislative instrument imposing legal obligations to addressing the problems identified in this Initiative. Option 4 implies the revision of existing EU law to extend the scope of the Unfair Commercial Practices Directive and the Directive on unfair terms in consumer contracts to B2B relationships and to amend the content of those directives and the MCAD to allow them to address platform-specific issues effectively. All options are supported by the establishment of an EU Observatory of the Online Platform Economy flanked by a wider non-legislative Commission initiative as explained in the Table 1 above.

5.4 Discarded Options

Options 1, 3 and 4 are discarded because they are either manifestly ineffective or evidently disproportionate. Particularly, proportionality is of crucial concern for this Initiative, because it embraces a light-touch and principles-based approach as the first step of the underlying two-step approach. Based on the principle of proportionality in Article 5(4) TEU any EU action shall not exceed what is necessary to achieve the objectives of the Treaties.

5.4.1 Non-legislative approach (Option 1)

In choosing this option the Commission would adopt a non-binding document such as a

Communication, inviting industry to develop measures to address the most problematic potentially harmful trading practices. This document would set out the principles and behaviour to be followed by platforms when dealing with their business users. This would include requirements for platforms to ensure transparency to business users in key areas such as the rolling out of changes in terms and conditions, reasons for delisting or suspension, and their data policy. However, the Communication would not specify the measures that online platforms should take to achieve these objectives: these would be left up to online platforms to determine. To solve the problem of ineffective redress, the soft law document would call upon online platforms to adapt their internal escalation procedures, and to set up an EU-level external redress mechanism.

Structured dialogues with industry would aim at addressing emerging issues on paid-for ranking, encouraging voluntary standards and private audits, as well as addressing emerging issues on data access.

In addition, a monitoring strategy would be established to regularly assess the evolution of all the problems above and emerging in the digital platforms economy. The strategy would be implemented through the establishment of the EU Observatory with a mandate to collect evidence, analyse, give guidance on and make policy recommendations for the evolution and potentially need for further regulatory intervention. Issues would include, but not be limited to, discrimination, business disputes and functioning of mediation mechanisms, emerging data issues like access to and sharing of personal and non-personal data, and MFN clauses.

However, pure self-regulation is unlikely to be effective. Limiting EU action to only selfregulation and certain accompanying measures would essentially rely on the platform industry's own incentives and willingness to change the status quo. While both platforms and their business users have an interest to maximise interactions and transactions with end consumers on platforms, their short-term interests in tackling issues arising in their business relationship are only imperfectly aligned. For example, vertically integrated platforms might have a natural vested interest in keeping the conditions for delisting and differential treatment opaque so as to preserve their margin of manoeuvre for favouring own downstream entities without much publicity. Improving the situation for business users, for instance by setting up an external redress mechanism, would also entail certain efforts and sometimes costs for platforms. In addition, there is no guarantee that the majority of platforms would adopt the non-binding best practice, or sign up to the voluntary measures, limiting the effectiveness of this non-legislative option

Previous experience with the Supply Chain Initiative (SCI) in the food sector suggests that purely voluntary initiatives are not suited for creating a functioning independent redress mechanism and enacted fairness rules that are attractive and credible for both sides of the market. The SCI is a joint initiative launched by seven EU-level associations in November 2011 with the aim to increase fairness in commercial relations along the food supply chain 185 . Despite some progress (elaboration of principles of good practice and setting up of a governance group), agricultural providers – the main supposed beneficiaries of the scheme – did not sign up to the scheme because of confidentiality and enforcement concerns. Five years after its creation, the SCI still had no independent chair 186 and the Agricultural Markets Task Force concluded that it needed improvement 'so as to render it more effective and attractive, including for farmers' 187 . In the meantime, 20 Member States have already adopted national legislation and initiatives to combat potentially harmful trading practices in the food supply chain, and that more are planning to do so in the near future 188 .

In the meantime, Member States risk adopting further regulation for B2B relations of online platforms, thereby increasing legal fragmentation. Furthermore, even if self-regulatory rules would eventually be adopted, those would be non-binding and non-enforceable. This would hamper the rules' effectiveness. Similarly, enforcement concerns were one of the reasons for producers, the weaker side of the market, not to join the SCI. A pure self-regulatory option is therefore discarded.

A wide range of other option elements were also considered for the policy options but not retained due to lack of effectiveness or disproportionality (see Table 3 of Annex 10). A full overview of options elements (both retained and discarded) is in Annex 9 and their assessment in Annex 10, tables 1 and 2.

5.4.2 Mandatory, binding rules for all aspects (Option 3)

As explained in Section 5.2, the identified problems could also be addressed in a fullyfledged, exhaustive and binding act that includes all necessary details, legal or technical, in the basic act itself. These detailed rules would leave the industry without any margin to set its own rules. In light of current evidence, however, the more far-reaching data obligations for example are disproportionate. To extend data access rights would entail significant legal and technical costs for platforms. Importantly, the impact of mandated access to specific sets of data for business users on the platforms' business model cannot be precisely quantified. A ban on contractual clauses that prevent the retrieval or use of specific types of data is too broad for a variety of reasons, including the protection of business secrets and more generally, platforms' business models as an intermediary. A wide block ban of such clauses would favour business users (beyond general fairness principles) but create significant risk of harm to platforms and consumers. Similarly, a prohibition on MFN clauses would be

185 See the website of the Supply Chain Initiative .

186 An independent chair was appointed on 8 November 2017.

187 Report of the Agricultural Markets Task Force, Enhancing the position of farmers in the supply chain , page 33.

188 AGRI Council conclusions of 12 December 2016 . Poland in the meantime has also adopted national legislation.

disproportionate in view of the current evidence. Also, a full per se prohibition of MFN clauses would not be compatible with EU competition rules.

This approach would provide significant clarity and redress possibilities to business users of platforms. It does not seem appropriate however for a fast-moving technological, economic and legal environment where industry involvement and more flexibility to necessary adaptations of rules are prone to lead to a more future-proof framework. Detailed legislation indeed risks both being rendered obsolete as a result of technological developments, as well as intervening in a disproportionate manner in the highly beneficial platform-model.

Therefore, the fast-moving nature of the platforms space renders it ill-suited to detailed regulatory frameworks. This finding also underpins the Commission's stated problem-driven and principles-based approach to regulating online platforms. As a result, exhaustive EU legislation is discarded as an option.

5.4.3 Extension of existing EU rules (Option 4)

As already stated in Section 1.1 and described in Annex 8.1 EU Competition law focuses on anticompetitive behaviour and mergers. The harmful practices described in Section 2.1.1 do not necessarily infringe EU competition law rules. Moreover, to be able to rely on Article 102 TFEU to investigate a potential abuse by online platforms of a dominant position, the respective platforms must be dominant in the relevant market. EU Competition law tackling anti-competitive behaviour is in addition enforced on a case-by-case basis ex post, prioritising, inter alia, those cases with a potential impact beyond the case itself. This Initiative therefore aims at complementing the enforcement of EU Competition law in a horizontal and light-touch manner, rather than relying on case-specific interventions.

Notwithstanding the previous paragraph, one of the policy options that have been considered as potentially feasible is the extension of existing EU rules to tackle the potentially harmful trading practices and the inefficient redress identified. Particularly, the UCPD and UCTD aim at protecting consumers against certain unfair commercial practices and contract terms respectively. Given the broad scope and the general nature of the instruments they already cover digital platforms and their business users in relation to their actions geared towards consumers.

It does not, however, seem to be proportionate and efficient to simply broaden the scope of these instruments for various reasons to address both the lack of redress and the potentially harmful trading practices.

First, these instruments address all practices that consumers face vis-à-vis businesses throughout the various transaction phases and regardless of the distribution channel or product. Given that the concerned practices are specific to the platform economy and the business models involved it would be disproportionate to extend the UCPD and UCTD to the general B2B sphere implying application to any business relationship in the EU and not just covering the relationship between platforms and their business users. Furthermore, to be effective it would not be sufficient enough to broaden the scope of the current instruments, because the practices listed in the annexes of the instruments have been designed for the offline world and the B2C relationships in mind focusing on fundamentally different aspects and not covering the specific problems. However, given the current evidence base adding the identified potentially harmful practices to the annexes of UCPD/UCTD would be too invasive at this stage and would go far beyond addressing the problems identified.

In addition, splitting up the identified problems based on the rationales behind both instruments would imply an incoherent approach, because the Annex of the UCPD is a blacklist prohibiting all practices per se, while the Annex of the UCTD is a grey list including only indicatively potentially problematic practices. This incoherence would be fortified, because the UCPD is a so-called 'full harmonisation Directive' and the UCTD is minimum harmonising giving the Member States a margin of discretion in regulation further.

Second, the problem identified as part of the present Initiative can be summarised as the cumulative effect of potentially harmful trading practices occurring in light of the absence of effective redress. The impact of the problem is direct economic harm to businesses and emerging regulatory fragmentation. Additional transparency and monitoring would be required in a first step to determine the precise impact of any individual trading practices. Combined transparency and redress measures would in addition already help to prevent direct economic harm resulting from such trading practices. It would not, however, be justified at this stage to categorise any individual trading practices as unfair per se, which would result from an extension of the consumer protection acquis to B2B relations.

As already set forth in Section 2.1.3., several Member States have adopted general fairness rules for B2B relationships, and 10 Member States have also extended UCPD and UCTD to varying degrees to B2B relationships (a more detailed description is available in Annex 8). It is evident that neither the existing national B2B fairness rules nor the extension of UCPD or UCTD provide an adequate and effective way for business users to resolve any disputes with platforms. If the platform-specific problems identified are not included in or covered by the blacklist of the UCPD or the grey list of the UCTD, business users would still need to bring an action before the courts to enforce their rights based on the general fairness test of UCPD or UCTD. Therefore, the crucial problem of lack of effective redress would not be tackled sufficiently. The practice of imposing exclusive choice of law and forum clauses often used by platforms in their terms and conditions further undermines the effectiveness of the extension, because in these cases business users cannot rely on European rules. Furthermore, given the expected ineffectiveness of the extension of the UCTD and the UCPD adopting a broad scope covering all B2B relationships in the EU is blatantly disproportionate.

Alongside the inefficiency and disproportionality of the scope broadening itself one has also to bear in mind that both the UCPD and the UCTD are cornerstones of EU consumer protection legislation. Any possible extension or review must inevitably take into account the instruments' rank as well as the other existing EU consumer law instruments and cannot be analysed in isolation. An extension of consumer protection instruments to cover certain aspects of B2B relations would also raise the concern whether such extension could have a – negative – impact on the high level of EU consumer protection.

Another option would be to take the MCAD as a basis as it applies to B2B relations, and extend it to tackle potentially harmful trading practices identified in this initiative. The MCAD prohibits traders from engaging in misleading advertising and lays down rules for comparative advertising. The report of the Fitness Check of Consumer and Marketing law concluded on the need to consider changes to the rules on misleading and comparative advertising in business-to-business (B2B) relations, largely confirming the findings of the 2012 Commission Communication announcing the intention to revise the MCAD 189 . The Inception Impact Assessment on a targeted revision of EU consumer law directives, however, underlined that the findings of the Fitness Check would inform future action in the area of

189 Communication from the Commission on Protecting businesses against misleading marketing practices and ensuring effective enforcement, 27 November 2012, COM(2012), 702 final.

B2B relations (notably on platform-to-business relations) within the Digital Single Market context 190 . Simply adding P2B specific rules to the MCAD would indeed not be effective as the existing horizontally applicable provisions of the MCAD require a revision that goes further than the scope of the present initiative limited to P2B relations. Furthermore, the MCAD addresses advertising and marketing in the B2B sphere, which are predominantly precontractual activities. The majority of the practices identified in this Initiative arise in already existing contractual relationships between business users and platforms. To be able to use the MCAD to address the identified practices one would need to revise the rationale and substantive content of the MCAD completely. Equally to the argumentation above on the UCPD and UCTD the business users would need to rely on the general fairness test of the MCAD by taking legal actions, if the platform-specific practices are not included in the blacklist in the Annex. Experiences show that the latter possibility has already proven to be inefficient to address the business users' lack of redress and based on the existing evidence including platform practices in a blacklist would be disproportionate.

As a result of the foregoing arguments the option to create an add-on or opt for a revision of existing EU legislation, namely UCPD, UCTD and MCAD, is discarded and will not be considered further. To be effective and proportionate this Initiative aims at creating a new self-standing instrument to address the identified harmful practices and the inefficient redress.

5.4.4 Conclusion

The non-legislative approach of Option 1 is unlikely to be effective, in the absence of sufficient incentives in the industry to address the problems identified in a across all Member States, and across all platform types. Previous experience with the Supply Chain Initiative (SCI) in the food sector suggests that purely voluntary initiatives are not suited for creating a functioning independent redress mechanism and enacted fairness rules that are attractive and credible for both sides of the market.

Mandatory, binding rules for all aspects (Option 3) are inappropriate for a fast-moving technological, economic and legal environment where industry involvement and more flexibility to necessary adaptations of rules are prone to lead to a more future-proof framework. Overly detailed, binding legislation risks being obsolete due to technological developments and intervenes disproportionately in the highly beneficial platform-model.

The extension of existing EU rules is analysed under Option 4, in the context of possible extensions of the Directives on Unfair Commercial Practices, Trading Practices, or on Misleading and Comparative Advertising. None of the possible extensions however, can be consider effective or proportionate for the specific problems at hand – the UCPD or UCTD instruments focus on a wide array of consumer related practices, which go far beyond the issues identified here. Modifications to the MCAD would not be effective as the existing horizontally applicable provisions of it require revision that goes further than the scope of this initiative limited to P2B relations, and focus furthermore on pre-contractual activities.

5.5 Retained options - co-regulatory Options 2a, 2b, 2c and 2d

As shown in Table 1 in Section 5.2, option 2b builds on option 2a, and options 2c and 2d build on option 2b. Both option 2c and option 2d are composed of the legal transparency requirements in six areas of concern (changes to terms and conditions, delisting, discrimination/preferential

190 Inception Impact Assessment on a Targeted revision of EU consumer law directives .

treatment, ranking, general data policy and MFNs), but options 2c and 2d also include additional provisions. Option 2c provides for an extended scope of the transparency requirement for ranking to also cover online general search engines. Under option 2c, the redress-related measures applicable to online general search engines concern only the ranking transparency obligation and are limited to granting legal standing to business associations to act on behalf of professional website owners to enforce this transparency requirement only. In addition to this scope extension in 2c, option 2d foresees an obligation for platforms to give any business user the opportunity to ask, in line with the GDPR, for customers' consent to obtain and process their e-mail addresses after the completion of a transaction and the payment of the commission to the respective platform.

Co-regulation has the advantage of guaranteeing a predictable legal framework while simultaneously giving sufficient flexibility to industry to shape and decide on their voluntary commitments, respecting the speed of innovation. Experience has also shown that, in line with the argumentation in Section 5.4.1. on a non-legislative approach, pure self-regulation is insufficient, particularly where it is built on purely private and voluntary commitments, and more so in this sector where the short term interests of platforms and business users are not aligned. A clear legal framework is necessary and can be provided by adopting a new legal instrument. This leaves room for industry to develop codes of conduct to further elaborate on their substance by platforms, who have expressed in consultations that they are willing to openly engage in selfregulation. The options could be based on the 'Principles for better co- and self-regulation' developed by a Community of Practice established by the European Commission 191 in its conception as well as its implementation.

The rules contained in the co-regulatory instrument would be self-standing and be immediately relied upon by business users and platforms. They can form the baseline for any self-regulatory addition. Depending on the co-regulatory model chosen, any self-regulatory addition would be screened and added to the body of rules if it strikes a satisfactory balance between the interests of platforms and their business users. This would usefully specify and enlarge the body of applicable and enforceable rules and promote buy-in from industry, both platforms and business users. The co-regulatory technique would enable a good balance to be struck between a limited number of general rules in the legislative instrument and the more detailed or technical rules that could be filled in by industry. This would make the instrument more future-proof since outdated technical rules could be modified more easily than by a revision of the basic act and would benefit from first-hand experience of the industry itself. The additional benefits of co-regulation include increased transparency, the simplification of rules that can automatically have sector support ensuring they can be swiftly adopted and properly implemented and help contribute toward a sense of co-responsibility of the industry and businesses involved, which will have a positive knock-on effect on elements of co-regulation that involve reputational levers 192 . Option 2 with its sub-options is retained for further analysis. The impacts of the co-regulatory options are assessed in Section 6.

6 W HAT ARE THE IMPACTS OF THE RETAINED POLICY OPTIONS ?

This Section summarizes the main impacts of the retained policy options as compared to the baseline, namely co-regulatory options 2a, 2b, 2c and 2d.

6.1 Impact on internal market

191 The Principles for Better Self- and co- Regulation .

192 The Current State of Co-Regulation and Self-Regulation in the Single Market , EESC Pamphlet Series, March 2005.

The impact on the Single Market of the co-regulatory options is two-fold. The substantive rules proposed will provide business users with greater legal certainty when using online platforms to trade in the internal market and provide concrete tools to seek redress in case of problems, thus supporting the growth of the Digital Single Market. Also, the monitoring and review clauses will allow regulators to adapt rules to the observed market reality to help business users find the appropriate support, gradually creating a common understanding of the issues identified and solutions to address them. This would possibly allow some alignment of platform-related rules across the EU. Further fragmentation will also be limited by encouraging Member States to cooperate through the EU Observatory.

6.1.1 Growth

The initiative's aim is to increase legal certainty in the platforms environment. This is why all retained options can be expected to have a positive impact on user trust and on growth of the platform economy. As a result of the proposed measures, more business users can be expected to sell over platforms or to expand their share of online sales through platforms. All retained options would therefore contribute to optimising businesses' turn-over realised on platforms, thus limiting the chilling effect that the currently observed potentially harmful trading practices have on sales (see Section 2.1.1.) A comparison of the total 2016 value of e-commerce with the total offline retail value or 2.56 trillion EUR shows it is likely the current e-commerce growth trend will continue, and the growing importance of online platforms along with it. In light of these statistics, the growth of the platform economy can legitimately be expected to have a positive effect on overall growth in the Digital Single Market.

Increased trust in the online environment will attract more business users, having thereby a positive impact also on online platforms' turnover. As the number of sales carried out over online platforms increase, so too will the commissions received by online platforms. Based on the calculations above, assuming that an average commission charged by platforms is 10%, platforms can be expected to receive additional commissions ranging from EUR 38 million to EUR 70.5 million if exclusively the effect on direct sales through platforms is taken into account. Adding the reversal in the dampening effect increases this estimate to a range from EUR 119 to EUR 476 million. Through their role as enablers of cross-border trade, the growth of online platforms resulting from all retained options will therefore benefit the internal market growth.

As compared to the baseline, user trust (as a trigger for growth) can be expected to increasingly gain in strength from options 2a to 2d, with option 2d bringing potentially a higher trust level since comprising the most comprehensive set of business user–friendly measures. At the same time, option 2d - by allowing business users access to their customers' email addresses carries the risk of free-riding by businesses. Due to these opposite effects that Option 2d would have on business users and platform operators, its impact on growth of the Digital Single Market would be more difficult to predict.

The value of e-commerce in the European Union was estimated to more than 500 billion EUR in 2016, a 13.5% increase from 2015 193 ; in 2017 the growth is estimated at another 14% year on year 194 . 22% of the 2016 e-commerce value is estimated to have been generated by EU third party sellers on online platforms 195 . Trade intermediated through online platforms is expected to

193 European Ecommerce Report 2017 – Ecommerce Foundation http://www.ecommercefoundation.org

194 European Ecommerce Report 2017 – Ecommerce continues to prosper in Europe, but markets grow at different speeds,

Press Release. https://www.ecommerce-europe.eu/press-item/european-ecommerce-report-2017-released-ecommercecontinues-prosper-europe-markets-grow-different-speeds/ 195 European Ecommerce Report 2016

follow an upward trend as most consumers opt for platforms when purchasing goods and services online: in a recent study, 71% of them preferred platforms for their purchases 196 All retained options will reduce potentially harmful trading practices online. The impact of the potentially harmful trading practices identified in the problem statement can be expected to drop by a minimum of 30%. 197 The drop in the impact of potentially harmful trading practices can be estimated to be reflected in the unrealised potential in terms of turnover and of the dampening effect that was previously identified. If one assumes on that basis that a similar share (30%) of unrealised potential in terms of turn-over businesses realise on platforms could be addressed and that the same share (30%) of the dampening effect as estimated in Section 2.3.2 could be reversed, this would lead to a positive impact on the platform economy of respectively between € 381 million and € 705 million per year in terms of increased turn-over, and of between € 810 million to € 4.05 billion per year of reversed dampening effect. These figures are likely to be higher in the future because trade on platforms is growing every year.

The assessment of impact on growth of the options must take into account both the possible positive effects of the increase in trust in the platform environment, as well as downside risks resulting from the costs of the proposed measures and possible repercussions on online platforms' practices, fees to business users and consumer prices. Overall however, the costs created by the initiative are expected to be limited under all scenarios, as described in the section on compliance costs (cf. sections 6.2). Online platforms will therefore have little incentive to pass on costs to consumers or to limit access to (small) business users. These dynamics are assessed in more details in the relevant sections 6.2.3 (impact on businesses) and 6.6 (impact on consumers).

6.1.2 Preserving the cross-border nature of the platform economy – preventing fragmentation of the internal market

While none of the retained options aim at harmonising national B2B fairness legislation, each of the retained options will complement national B2B legislation with high-level transparency rules. Given the more limited scope of option 2a, this option carries a greater risk of fragmentation as opposed to options 2b, 2c and 2d, if MS consider that not the entire set of problems faced by businesses is adequately addressed.

The proposed rules will provide more clarity and regulatory predictability for platforms at EU level as to the requirements they need to comply with. It would allow preserving the existing cross-border dynamics of the platform economy by setting a common framework for Member States.

Given the absence of the regulatory backstop on key issues such as ranking, data access policies, discrimination and MFNs, the effect of EU monitoring pushing on reputational levers to effectuate a fairer and more predictable business environment for business users of online platforms would be limited under Option 2a as compared to the baseline. Indeed, the legal transparency obligations on these issues foreseen under options 2b, 2c and 2d will be accompanied by enhanced external scrutiny of online platforms' trading practices which should incentivise these firms to pro-actively improve the situation for business users, for example by solving issues out-of-court (bilaterally or through mediation). .Option 2a - by leaving four out of

196 See footnote 106, LSE & Partners – forthcoming.

197 This is a conservative estimate that assumes that a fair share of currently unsolved problems would be resolved, along with part of the problems that are currently only resolved with difficulties. Indeed the study on business users of online platforms showed that 30% of all problems in P2B relations remain unsolved and a further 29% can only be resolved with difficulties. A drop of 30% is therefore a safe assumption to cover the resolution of most (if not all) of the unsolved problems, along with a reduction of the cases resolved with difficulties, which also cause damages to businesses.

six of the most frequently observed high-impact trading practices to be addressed through selfregulation - carries a more pronounced risk of fragmentation as opposed to options 2b, 2c and 2d. Given the absence of the regulatory backstop on key issues such as ranking, data access policies, discrimination and MFNs, the effect of EU monitoring pushing on reputational levers to effectuate a more fair and predictable business environment for business users of online platforms would be limited under Option 2a as compared to the baseline. Indeed, the legal transparency obligations on these issues foreseen under options 2b, 2c and 2d will be accompanied by enhanced external scrutiny of online platforms' trading practices which should incentivise these firms to pro-actively improve the situation for business users, for example by solving issues out-of-court (bilaterally or through mediation). Options 2b, 2c and 2d would thus help to ensure a more harmonised approach to platform-to-business relationships within the EU. These options would thus have a positive impact as compared to the baseline scenario in which Member States are increasingly adopting or considering legislation addressing specific platformrelated aspects. The scope, addressees and level of intervention of these national measures vary significantly, which leads inevitably to a fragmentation of the inherently cross-border platform environment. In each option, a key role of the EU Observatory would be to monitor further the internal market dimension of the platform economy. In addition, the EU observatory will help allow building a common EU understanding of what issues are, thus allowing for more consistent regulatory approaches when such are deemed necessary at national level.

6.2 Impact on enterprises

6.2.1 Impact on online platforms

6.2.1.1 Compliance costs

The costs are expected to result from three main factors: (a) the implementation of the different transparency requirements; (b) the setting up of internal redress mechanisms and external organisations that can provide industry-specific mediators; and (c) the reporting obligation on the functioning of the internal redress mechanism.

The implementation of the transparency requirements will result in one-off costs to adapt the implementation and communication of platforms' terms and conditions, and updating these standard contracts where needed (costs related to the legal expertise, revision and publication of their terms and conditions). Once these procedures are carried out, platforms will face running costs when modifying and communicating changes to their terms and conditions. However as changes are not expected to occur more frequently, these costs are likely to be equivalent to those that online platforms currently face. Clearly, it is good business practice, even for very small platforms who want to build a customer base, to have clear and transparent terms and conditions on their different policies on matters such as delisting, ranking or access to data. Options 2b, 2c and 2d foresee legal transparency obligations on more complex issues such as ranking, data access, discrimination and MFNs.

An independent study on contractual arrangements between platforms and their business users examined options which closely match the retained options assessed here. Their independent impact assessment identifies major benefits for business users that would correspond to little or no impact for larger online platforms, especially as the legal framework would leave these firms free to change the rules applicable to their ecosystem, or even to delist large numbers of offerings, as long as this is done in line with pre-defined objective aims and in a nondiscriminatory fashion. Exclusively the rule requiring data on transactions to be provided (which is similar to the data-relevant aspect of policy option 2d), is estimated to imply some initial costs for platforms and possibly impact their willingness to invest in innovation. 198

Option 2c would extend the legal transparency obligation on ranking to the area of online general search, where complex algorithms determine the saliency of search results on the basis of a search index that can cover, in principle, most of the openly accessible Internet.

Providing meaningful transparency on ranking in this fast-moving area implies, in theory, more significant compliance costs. Transparency measures would have to capture the high frequency with which changes to the functioning of online general search engines' ranking mechanisms are implemented, and "translate" their functioning into useable guidance for the appropriate audiences – which may cover a broad spectrum of businesses ranging from technologically illiterate firms to digital natives. However, the major providers of online general search engines already today offer some transparency to inform webmasters how to achieve high quality search results, although the level of detail provided differs significantly. The main transparency tools developed by the three major online general search engines active in the EU are described in Annex 7.2.5. In addition, Search Engine Optimisation (SEO) strategies have developed around online general search engines, which could be leveraged for the purpose of "communicating" effectively with the target audiences.

A legal requirement to be transparent about ranking policies would give more prominence to these existing practices and give them legal character. The majority of tools allowing for search optimisation put in place by online general search engines are publicly available but the level of detail and exhaustiveness differs, and these are not, in general, tailored to the business audience and their availability to the general public is not guaranteed. Option 2d would require online platforms to ask consumers for their consent to share some data with business users. This measure would require a technical adjustment on the side of platforms to allow consumers to express their consent (the data itself can be shared through existing communication channels). The request would be conditional on the completion of a transaction on the respective platform and to the payment of the platform's commission. Nonetheless, if the sellers have their own sales channels, it may also allow them to circumvent the platform for future transactions and disrupt a core aspect of the platform business model. 199

Setting up internal and external organisations that can provide industry-specific mediators mechanisms will also be linked to compliance costs. These are also set out in Annex 4. Regarding internal redress, platforms that already have a dispute settlement mechanism may be required to upgrade their systems to comply with the quality standards set out in the legal act, notably speed and effectiveness (e.g. identifying a clear contact point for submitting complaints). Those that do not will face both set-up and running costs, which may be offset over time as a result of increased or more efficient platform-use. The cost of such mechanisms varies considerably according to the size of the platform. These costs can be estimated to lie in the range of a 0.4 to 1% increase in the cost base for smaller platform companies 200 , and a one-off

198 Ernst & Young study (forthcoming).

199 In its assessment of the impact on platforms of a data sharing obligation, an independent contractor also found that "the collection and analysis of data constitutes an important competitive advantage for platform operators. A limitation thereof could reduce the ability to achieve differentiation from other platforms and, as a result, a platforms willingness to undertake investment. On the other hand, there are no reasons to solely attribute the ownership of information with regard to the transaction between business users and customers to the intermediary, i.e., platform owners. Defining and implementing clear data policies may initially be associated with some costs",see: Ernst & Young, study (forthcoming). 200 Assuming a cost of one additional FTE for small companies having between 50 and 250 employees.

cost of 0.03% of total turnover for larger ones 201 . While the administrative burden resulting from this particular measure may not always be completely offset, platforms may in many cases be able to develop intelligent solutions to lower costs, such as using the same or similar technologies and operational structures for customer support to also provide for internal redress for businesses. They are also likely to have a commercial incentive to follow the example of larger players. Therefore, the actual additional cost is likely to be lower and likely to be on top of sunk costs for investments already made 202 . The cost is therefore expected to be limited in all cases but can be estimated to be largest for the smallest companies, in relative terms 203 . This supports considering an exemption of certain categories of companies. The majority of platform business models generate income from commissions on transactions concluded between business users and consumers.

The obligation to allow for P2B disputes to be escalated internally is not likely to fundamentally change the economics of running an online platform: most platforms already possess such systems, meaning that they are fully compatible with the intermediary business model. The options also preserve the platform operators' flexibility and ability to curate the content. In addition, there is no clear evidence to suggest that the additional costs would necessarily be passed on to either party in the intermediated transactions, because most online platforms in their early stage of development run losses in order to capture a wide user base on both sides of their networked market. The costs of the external organisations that can provide industry-specific mediators will be determined by the set-up chosen by industry. However, they are also likely to be limited, not least because online platforms will be incentive to spread the costs widely across the entire industry. Online general search engines will not be under an obligation to engage in good faith in mediation, and they are therefore equally excluded from the obligation on the Commission to encourage the setting up of industry-specific mediation organisations. This reflect the issue-specific approach taken in respect of online general search engines, which exclusively covers ranking transparency. Contacts with industry show that industry is willing to take part in voluntary initiatives of this kind, mitigating the risk of low-industry buy-in. In addition, the use of mediation has been shown to limit costs (e.g. for litigation) when it is successful, so the shared investment in a trusted mediation body amounts to savings elsewhere.

The compliance cost associated with giving legal standing to business associations to act on behalf of business users to enforce the foreseen legal obligations can be estimated qualitatively as follows. On the one hand, additional legal costs may arise for online platforms if they have to defend against cases brought under the enforcement provision of the rules. At the same time, the regulatory assumption is that compliance with the mostly one-off transparency obligations will be high, especially in light of the proposed monitoring efforts, and the technical legal grey zones would be small, and therefore limited costs arising from litigation would be incurred. Safeguards against frivolous litigation include limiting the nature of cases that can be brought to injunctive relief (and not compensatory), and requiring that associations are non-profit in character.

Finally, the reporting obligation relating to the internal complaint-handling mechanism that is included in all co-regulatory policy options will be designed to limit costs for the platforms concerned. The reporting obligation would cover a limited number of elements such as the total number of complaints received, the subject matter of the complaints, the time period needed to

201 Based on the actual example of a EUR 1.75 million one-off cost for a platform company achieving a EUR 6 billion annual turnover.

202 See Section 7.2.3 dealing with proportionality for more detail.

203 Using the definitions of the European Union of Small and Medium Enterprises (Commission Recommendation of 6 May

2003 concerning the definition of micro, small and medium-sized enterprises C(2003) 1422) small platform businesses would be those with < 50 employees, and a turnover or balance-sheet total of < EUR10 million, while medium platform businesses are those with < 250 staff headcount, and < EUR 50 million annual turnover or < EUR 43 million balance sheet total.

process the complaints and the decision taken. Online platforms can largely automate data collection and reporting. Many collect this type of information already for quality management purposes and would, therefore, only incur very limited costs for transmitting the data to the EU Observatory on a regular basis.

6.2.1.2 Impact on smaller online platforms

SME online platforms will also benefit from the growth of the platforms environment. At the same time, all retained co-regulatory options will entail limited, mostly one-off costs associated directly with changes to contractual terms and conditions to accommodate the legal obligations in relation to transparency (cf. Annex 12.2 for a more detailed description of impacts). The enhanced transparency that would result from the implementation of the foreseen legal transparency obligations can benefit smaller online platforms as there will be more scope for competitive differentiation. At the same time, any costs resulting from possibly increased litigation are expected to be limited, given the principles-based nature of these obligations, and the important scope for implementation by industry. Small online platform providers will accordingly be able to exclude any significant litigation costs by providing transparency at a general level, while staying abreast of, or getting involved - should they voluntarily opt to do so - in industry discussions on codes of conduct on each of the various potentially harmful trading practices.

Option 2a implies in this regard, relative to the other co-regulatory options, the lowest number of legally binding obligations and therefore the most limited cost increase compared to the baseline. Option 2b allows addressing the unfair practices identified while preserving platforms' ability to set freely their business strategies. The compliance cost entailed by Option 2c could possibly be higher for smaller online general search engines which have not developed guidance for businesses on how to optimise their appearance in search results. These firms could however try to limit their compliance costs by using as an example the existing transparency provisions developed by bigger online general search engines. A transparency rule for search engines would indeed help spread best practices and possibly creates incentives for small platforms to use quality saliency of information/data, or to guarantee absence of any conflict of interests (e.g. concerning advertising-based business models) as criteria of competitive differentiators. The main argument against ranking-related transparency obligations is that spammers could game online general search engines, which would be detrimental to society. The proposed transparency obligation would not, however, require disclosing any trade secrets and foresees an important scope for industry efforts to provide practical tools for meaningful transparency – including ways to prevent the gaming of search results. While the marginal cost of adding the transparency requirements should therefore not be overestimated, policy option 2d may create higher costs for small platforms. The need to share email addresses with business users who have obtained consumers' consent may put constraints on small platforms in their development phase, preventing them from scaling up. Option 2d may therefore have a negative impact on small platforms. Option 2a implies in this regard, relative to the other co-regulatory options, the lowest number of legally binding obligations and therefore the most limited cost increase compared to the baseline.

Imposing new legislation could in theory lead to increased difficulties for new market entrants. In this case it can rather be expected that the provisions will provide start-ups with greater clarity on the requirements they need to comply with across the EU when entering the market thus benefitting from a more predictable regulatory environment. Such legal transparency requirements could also provide additional competitive parameters that can be leveraged to distinguish the start-up platforms from established players and thereby enhance market entry. At the same time, while an efficient and speedy internal dispute resolution process could be possibly ensured through automated systems - chatbots can be used for a first screening of complaints, with justified claims eventually being dealt with in person - the creation of an internal escalation mechanism, may require that small platforms put in place processes that go beyond their current capacities.

This type of considerations raises the question whether specific thresholds are needed to exempt some types of enterprises from the proposed regulation. Irrespective of which of the four retained options would appear as the most appropriate on the basis of the comparative analysis performed in Section 7, the evaluation of the regulatory burden of this Initiative requires an assessment of whether micro-enterprises or other categories of companies 204 should be exempted from its scope in line with the Think Small First 205 and the Better Regulation principles 206 . The following section analyses this question.

6.2.1.3 Impacts for Options for thresholds for exemption

Several questions arise when considering options for thresholds: For which measures are thresholds needed? If they are needed, how should they be set? Which measurements or proxies can be used to determine their level and to verify compliance easily?

The main considerations in relation to thresholds concern the degree to which the underlying problem analysis applies to smaller online platforms and the regulatory burden which would stem from the proposed intervention. Different "threshold" possibilities are considered below while Section 8 dealing with the preferred policy option presents the conclusions of the analysis.

(A): A threshold exempting some categories of online platforms from those measures for which a significant impact cannot be excluded, based on impacts assessed above

One option would be to exempt smaller platforms from those measures for which an administrative burden cannot be fully excluded, i.e. from the most burdensome measures in relative terms. This approach would mean that the relevant legal transparency obligations foreseen under the various co-regulatory options apply to all online platforms, while the internal redress mechanism which appears more costly (cf. Section 6.2.1.2) is not applicable to a certain category of smaller platforms.

As to the precise setting of the threshold, the SME definition based on the double criterion of staff headcount and turnover/balance sheet has been considered. The question is whether the threshold should be set at the level of a small or micro-enterprise? Available data does not allow drawing a clear distinction between the impacts of the internal redress on these two types of platforms. In order to avoid any disproportionate burden it seems therefore more appropriate to exempt all online platforms constituting a small enterprise (< 50 employees) from the obligation to provide for an internal complaint-handling mechanism. This would support the scaling up of both start-up and emerging small platforms.

204 Using the definitions of the European Union of Small and Medium Enterprises (Commission Recommendation of 6 May

2003 concerning the definition of micro, small and medium-sized enterprises C(2003) 1422) small platform businesses would be those with < 50 employees, and a turnover or balance-sheet total of < EUR10 million, while medium platform businesses are those with < 250 staff headcount, and < EUR 50 million annual turnover or < EUR 43 million balance sheet total. 205 Principle embodied in the Communication from the Commission to the Council, the European Parliament, the European Economic and Social Committee and the Committee of the Regions - “Think Small First” - A “Small Business Act” for Europe, COM(2008), 394 final. 206 Better Regulation Toolbox complementing the Better Regulation Guidelines presented in in SWD(2017), 350 final.

By excluding platforms with < 50 employees from the internal redress measure, 47% 207 of all EU platforms would be exempted from this specific obligation, while a large proportion of all transactions would nonetheless be covered 208 . This option would reflect the light-touch approach behind the transparency requirements which would apply to all companies. The extra burden associated with the initiative would be limited, and the measure may provide a competitive edge to very small platforms who want to build a customer base; they would be able to attract customers by offering them clear and transparent terms and conditions, similarly to bigger platforms.

Instead of using the staff headcount and/or turnover, an option would be to use different proxies when defining the thresholds, such as number of website visits/month (as in the Loi Lemaire), or number of registered users (as the German NetzDG definition). This would potentially allow targeting the exemption more narrowly to start-up and scale-up platforms that can be guaranteed not (yet) to have significant user bases. These metrics are however disconnected from need to limit the impact of the cost of internal complaint-handling, as even online platforms having significantly less than 50 employees may command relatively large user bases; these firms have frequently been seen not yet to generate any turnover and may precisely be unable to absorb such costs. Also, large variations in user numbers or website visits are possible depending on the date, either seasonal or due to fast growth periods, leading to uncertainty for the business on whether it is concerned or not by the rules. Basing thresholds on number of registered users would also decrease platforms' incentives to increase their user base thus artificially limiting network effects which are at the core of platforms' business models. Although the number of website visits and active users are routinely measured by online platforms themselves, they are not publicly reported. The criteria used in the SME definition seem therefore more appropriate for the purpose of setting a meaningful threshold.

(B): A horizontal threshold exempting some categories of platforms (micro- or small-) from the entire measure

This option may be designed to exempt from the entire measure those platforms which qualify either for micro- or for small-enterprises. Excluding exclusively micro-enterprises from the entire measure would risk putting the threshold too low, thus imposing the more burdensome internal redress obligations to platforms which may not be ready yet to absorb such measure as many such companies are still in the seed phase. As explained above, the proportionality principle rather suggests for not only micro- but also small enterprises to be exempted from those measures for which the regulatory burden may be more significant. On the other hand, if a single horizontal threshold is set for all small platforms, this would imply that almost half of all existing online platforms are exempted. This may appear unjustified given the light touch approach of the legal transparency requirements proposed, the exemption of these platforms from the most burdensome obligation and the fact that platforms having between 10 and 50 employees can already fulfil an important gateway function. Thus, a horizontal threshold applicable to the entire initiative, while clear and simple for implementation, does not appear appropriate to the measures proposed. The latter rather call for a targeted threshold taking into consideration in particular the more burdensome nature of the internal redress rule.

(C) A dual threshold combining (A) and (B)

207 3298 platforms under options 2a and 2b. This number would be 3380 platforms under options 2c and 2d extending the scope to online general search engines. 208 This can be indirectly inferred from a recent DG JUST study on the collaborative economy study showed that only 20 out

of 485 platforms were very large, with over 100 000 daily unique visitors, and the companies that will be in scope of the initiative will therefore account for a very large share of total intermediation.

Another option would be to combine the internal-redress related threshold with a general microenterprise exemption applicable to the entire measure (C).

(D): No threshold – the proposed measure applies horizontally to all platforms

This option would imply that not only the relevant legal transparency requirements would be applicable to all types of platforms but all proposed obligations. In light of the above explained need to account for the more burdensome nature of the internal redress mechanism, option D does not appear to constitute an effective alternative.

Table 3 shows the distribution of employees in EU and Global Platform businesses, including search engines (red colour in the table) 209 .

Table 3: Distribution of employees in EU and Global Platform businesses

Staff headcoun Number of EU Platform Businesses Number of Global Platform Businesses

<10 1772 + 34 = 1806 = 25% 3333 + 82 = 3415

<50 3298 + 82 = 3380 = 47% 6472 + 160 = 6632

<200 3904 + 113 = 4017 = 56% 7871 + 194 = 8099

Total EU Platform Business Total number of Global Platform Businesses = 7012 + 113 = 7125 19526 + 194 = 19720

6.2.2 Impact on non-EU platforms

According to data available to the Commission there would currently be approximately 12 500 non-EU platforms active in the market world-wide 210 . The elements presented for the coregulatory options are applicable to EU platforms and would apply equally to these non-EU platforms, if they intermediate between EU business users and consumers located in the EU. While the measures would not restrict the platforms' freedom to determine the law and forum applicable to contractual issues, these clauses should not prevent the enforcement in (EU) courts of the envisaged P2B rules by business users - obviously in full compliance with private international law, including the Hague Convention or the Rome I, Rome II and Brussels Ibis Regulations. 211 The mediation possibilities which would be offered constitute a considerable change to the current situation in which EU business users do not have any other choice but to revert to the extra-EU court set by a non-EU platform in its terms and conditions. However, it should be noted that submitting to the outcome of mediation proceedings will remain voluntary, both for EU and non-EU platforms.

6.2.3 Impact on business users

Compliance costs linked to the co-regulatory options will not have a short term effect on the fees and commissions applied by online platforms. Compliance costs linked to establishing transparent terms and conditions are, as explained in the previous sections, limited and apply to platforms, not their business users. Certain platforms might decide to use the opportunity created

209 From the Dealroom.co database of November 2017, defining platforms as marketplaces and adjacent categories including classified listings, booking, lead generation, and performance-based business models, as well as search engines.

210 According to the Dealroom database.

211 See section 0

by the implementation of a legislative initiative to increase the fees applied to business users, who may in turn pass these on to consumers. If these increases were to be substantial, the effects described above would be limited by a dampening effect on sales. However, it is unlikely that platforms will significantly increase their levels of fees or commissions, as the compliance costs of the co-regulatory options are limited and because business models in the platform economy frequently accept growth at high internal costs in order to acquire users on both sides of the platform, in order to leverage the network effects.

In the event that platforms consider transferring the additional costs incurred onto the sellers, the additional burden on the business user is foreseen to be highly limited. Using the estimates under 6.2.1, should platforms transfer the entirety of the one-off cost of setting up an internal redress mechanism towards its existing business users, the additional cost increase for each seller is foreseen to be minute 212 . Beyond the increased sales through platforms, the creation of an effective redress mechanism will be particularly positive for business users. 213 In combination with the greater clarity provided by the transparency measures on the reasons for differences with platforms, business users will have a greater chance of quickly and effectively solving disputes with online platforms. This may mean that more business users find grounds under which they can take problematic cases to court, using for instance collective interest litigation. Nonetheless, it will be in the interest of both business users and platforms to make greater use of the possibilities offered by mediation, which both co-regulatory options make more readily available. Mediation has been shown to be a more flexible and cost-effective solution. A European Parliament study on the cost of (non-)mediation has shown that an average cost to litigate in the EU is more than € 10,000, while the average cost to mediate is approx. € 2,500. Therefore when mediation is successful, it could save € 7,500 per dispute 214 .

As regards potential retaliatory actions from platforms towards the businesses active on them, neither of the options retained is estimated likely to trigger them. The fear of retaliation expressed in parts of the stakeholder consultation process concerned measures taken by certain platforms against individual businesses, or groups of businesses, as payback for specific actions. The co-regulatory options retained do not, however, single out particular businesses or specific actions that might single out a particular business to a platform. They rather increase clarity on the grounds for suspension or termination on the use of a platform, how ranking operates including mechanisms that allow business users to influence their prominence etc. and give more time to react to changes in terms and conditions or understand why a decision has been taken to delist or suspend them. Furthermore, the specific measures put forward in the proposal (redress mechanisms, observatory, etc.) contain safeguards against businesses being endangered for making use of the new functions: the use of external mediators must fit the criteria of anonymity and independence, the internal redress mechanisms would be open for use to all the businesses on a platform, therefore making it hard for the platform to reasonably single out and punish against one single business, etc. In addition, a platform choosing not to adopt the new features designed to enhance trust and increase the quality of the experience for a business would find itself at a disadvantage compared to other platforms that choose to attract businesses by improving the quality of the seller's experience, and therefore, risk losing its market share among sellers.

212 Based on the actual example of a EUR 1.75 million one-off cost for a platform company achieving a EUR 6 billion annual turnover, with 45.500 active sellers. 213 Ernst & Young study (forthcoming), chapter 4.3.2.2 also concludes on a strongly positive effect of the transparency benchmark on business users due to reduction of their direct costs resulting from non-transparent T&C; this beneficial effect being more pronounced for small firms. 214 Note of the European Parliament, Quantifying the cost of not using mediation – a data analysis.

None of the options will create significant costs for SME business users. On the contrary, SME users of platforms stand to gain from measures that will provide them with greater ease of doing business and enhanced legal certainty. This includes in particular the measures on transparency and minimum notice periods for changes to terms and conditions, transparency on the reasons for suspension or delisting, and transparency on the criteria for paid ranking results. These measures would lead to savings for smaller business users, as they would be spared the costs linked to reinstating accounts or products that have been blocked. In the case of paid-for ranking results, the increased transparency around the auctions oftentimes used to award increased visibility would be beneficial to smaller business users, as they could either choose not to participate where the resulting ranking is unlikely to be satisfactory, thus saving them the cost of participation, or choose to participate, and gain increased exposure. This was confirmed in the replies to the Commission's consultation of SMEs through the Small Business Act Follow-up Group. 215

Businesses' turnover is in addition directly impacted by the visibility they get on online general search engines. A transparency obligation on ranking in general search, such as foreseen under policy option 2c, would allow more predictability for business users. As explained in section 2.1.1.3, the legal transparency obligation on ranking would thereby additionally cover 20%-30% of total Internet traffic received by online platforms and respectively >50% and >70% of Internet traffic received by retailers' and hotels' own websites. Given the regular and high rate of algorithms changes done by online general search engines, such obligation would help businesses to develop better informed search optimisation strategies. This would be particularly beneficial for enterprises with little or no online presence.

6.3 Impact on public authorities

National authorities will not be directly impacted by any of the co-regulatory options. Over time, the obligation on platforms to list national mediators in their terms and conditions and to engage with them in good faith might lead to more P2B cases being brought before such mediators. Mediation is a private activity which will not impact public authorities. Member States will moreover not be required to adapt their existing certification schemes for mediators; those that already have such schemes in place will simply provide this existing service also for any new mediators that may enter the specialised area of P2B relations. Any possible burden on national court systems is also expected to remain limited as a result of the layered design of the legal redress provisions. All co-regulatory options rely on out-of-court, alternative dispute settlement mechanisms to solve substantive issues arising between business users and online platforms. The legal provision granting standing to business associations is a tool to encourage online platforms to actually engage in these out-of-court mechanisms to effectively resolve disputes without having to resort to national courts. In Member States, the mere threat of possible collective interest litigation is a sufficient deterrent to encourage industry compliance with obligations. 216 It will not be possible for these associations, as representatives of the business users, to rely on the foreseen legal instrument to instigate court cases concerning substantive issues relevant to individual business users. Rather such cases shall be limited to the

215 Based on a response to the questionnaire circulated through the Small Business Act Group on 11 August 2017 and discussions at the Small Business Act stakeholders meeting on 27 September 2017. 216 See comment from the Finnish Competition and Consumer Authority in the Analysis of the state of collective redress in the European Union in the context of the implementation of the Recommendation of the Commission on common principles for injunctive and compensatory collective redress mechanisms in the Member States concerning violations of rights granted under Union Law (request for services JUST/2016/JCOO/FW/CIVI/0099, Lot1/2016/06) prepared by the British Institute of International and Comparative Law (BIICL), in a research consortium with Civic Consulting and Risk & Policy Analysts (RPA), and supported by the Office for Economic Policy and Regional Development (EPRD), (forthcoming).

prevention or termination of non-compliance with the obligations in the foreseen legal instrument, which will be limited to those that do not make the necessary adaptations.

6.4 Impact on innovation, competitiveness, competition

Online platforms are important drivers and enablers of innovation (Section 2.3.5) 217 , thus contributing to digital transformation of the economy and enhancing businesses' competitiveness. Business users, in turn, are important innovators, using the innovation-enabling software "building blocks" and market access provided to them by online platforms. The millions of sellers active on e-commerce market places provide constant feedback to online platform operators on logistical, software and commercial problems encountered and on possible innovative ways to address them. App developers provide a constantly improving richness of content that no single platform could have imagined or engineered. 218 Research confirms this key role played by communities, networks and user involvement in platform innovation. 219 Regulatory action on P2B relations could hamper the innovation capacity of online platform ecosystems only if it were too interventionist, as it would divert resources from innovation activities to regulatory compliance activities. All retained options are designed to be proportionate and do not interfere with platforms' business models. The coherence with innovation is indeed also a specific criterion against which each content option has been checked (see table 2 in Annex 10). Conversely, the impact in terms of innovation on the side of business users can be similarly expected to be positive, as business users trading via platforms can benefit from a more predictable and contestable business environment and focus even more on product and service improvements and innovation.

All co-regulatory options that aim at releasing the full potential of the online platform economy can thus positively impact innovation both on the side of the platforms and of the business users. It is legitimate to expect an increase in platforms' innovation expenses under options 2a, 2b and 2c. Different from them, policy option 2d may have a negative impact on innovation, as data lies at the centre of platforms' business models. The innovation-related considerations for each policy options 2a, 2b, 2c and 2d are detailed in Annex 12.4.1.

In terms of competitiveness, EU-based online platforms will not be undermined by the proposed measures, irrespective of the policy option chosen. On the contrary, the platform ecosystem may be expected to become more competitive as a result of this initiative aiming at greater predictability for the platform ecosystem. Policy options 2a, 2b and 2c will, although to a different extent, have a positive or neutral impact on the three components of competitiveness as defined in the Better Regulation toolbox, i.e. price- , innovation- and international competitiveness. Notwithstanding their broader scope, policy options 2b and 2c are light-touch and do not imply significant costs for online platforms. Also, since all platforms intermediating between EU-based business users and consumers will have to comply with the measure, any risk of European platforms being undercut by platforms not complying with the proposed measure is minimal. Rather than driving operators of online platforms out of Europe, the proposed measure is estimated to increase trust and lead to an increase in the number of businesses present on

217 Platforms are a magnet for innovation. For example, new apps are constantly being developed and made available in app stores. In March 2017, there were 2.8 million apps available in Google Play, 2.2 million in the Apple App Store, 669,000 in the Windows Store and 600,000 in the Amazon Appstore. Statista: Number of apps available in leading app stores as of March 2017.

218 Based on 1.5 million apps listed in the App Store, it is estimated that in order to re-create the same creative richness, it would have required Apple itself 519 000 years' worth of work, see: https://www.slideshare.net/faberNovel/gafanomics href="https://www.slideshare.net/faberNovel/gafanomics-season-2-4-superpowers-to-outperform-in-the-network-economy/42-42The_more_apps_available_the">season-2-4-superpowers-to-outperform-in-the-network-economy/42-42The_more_apps_available_the. 219 Jeremy Rose and Brent Furneaux, "Innovation Drivers and Outputs for Software Firms: Literature Review and Concept Development," Advances in Software Engineering, vol. 2016, Article ID 5126069.

online platforms. The impact of option 2d on competitiveness is more difficult to predict: its immediate impact on business users would be increased revenues due to improved customer base but it may affect negatively platforms. If it leads to their reduced ability to innovate and compete, this effect could be passed on business users in the longer term (cf. annex 12.4.2).

In terms of competition, the co-regulatory options retained would set a standard for higher quality of service thus creating the opportunity for start-up platforms to compete on the basis of the better business environment they would offer to professional users. Options 2a, 2b and 2c would thus likely create the right regulatory environment for increased competition as compared to the baseline scenario on both sides of the market - among platforms and among businesses present on these platforms. Increased competition could be expected to lead to higher quality products/services provided to business- and end-users.

Under option 2a, the mandatory transparency provides additional competitive parameters for start-up platform companies. Even small increases in user trust will equally support the growth of existing platforms as well as the emergence of start-up platforms. Increased transparency though non-binding for some of the issues would possibly give the right signal to more businesses to use online intermediation; this could in return lead to more competition.

In option 2b, the increased trust resulting from the resolution of disputes with online platforms will expand the business user base of existing platforms. This will feed into the existing network effects laid out in the problem statement. The resulting renewed dynamism of the platform economy would a priori allow the emergence of new, small platforms. Legally binding transparency rules on all six issues identified and appropriate redress tools could also be expected to contribute to more competition among business users.

Option 2c would extend these above effects to online general search engines and the business users who use them as a gateway to customers. This option may be expected to have an indirect positive effect through enhanced transparency as a complementary tool to competition law.

The positive effect on online platforms will be more limited in option 2d, where the risk of freeriding by business users is greater; the number of business users active on platforms may not increase as much, if platforms are faced with the need to redesign their business strategies. The dynamics behind each of the four options are explained more in detail in Annex 12.4.3.

6.5 Employment and social impact

The overall impact of the co-regulatory options on social welfare including in particular employment and related social impact are likely to be positive, building further on the 4.7 million jobs which can be roughly attributed 220 as being generated by business users in the platform economy. As described above the measures aiming to full potential of the online platform economy will lead to an increase in turnover for both business users and platforms. The increase of business users' turnover may be expected to lead to increased employment opportunities. This expectation would probably be less valid for online platforms despite the fact that they will see their revenues from commissions increasing (as a result of the increase in sales

220 We consider an average number of employees of 4.7 irrespective of the economic sector concerned. This is a conservative assumption since this figure corresponds to the lowest one of the three sectors for which data is available: i.e. computer programming (i.e. app developers). The corresponding figures are 5.2 for retail and 8.4 for accommodation (source: Eurostat, datasets sbs_na_dt_r2 for retail and sbs_na_1a_se_r2 for accommodation). This number does not account for possible loss of jobs due to the closing of physical stores/activities since such closing is not foreseen as a result from the assessed initiative.

over platforms). Given that platforms are innovation-driven, it would be difficult to predict whether they would use increased revenues to create additional jobs instead of investing in research and development.

In addition, while the primary impacts of the initiative are of an economic nature, some beneficial social impact in particular for self-employed individuals is conceivable. A large part of sellers on e-commerce market places are self-employed individuals, with 89% of all sellers achieving a turnover of less than € 50,000. 221 In addition, both options will cover collaborative economy platforms to the extent these host professional users. Whether a user of a platform is considered professional or not depends on national rules in the EU. However, the proposed obligations would lead to greater transparency, predictability, and certainty for all users. 222 Overall, although difficult to quantify, it is therefore reasonable to assume that the proposed options will provide more predictability also to the benefit of a large number of particularly vulnerable economic actors. The net benefit of the initiative – improved businesses' access to innovation and business opportunities created by inline platforms, increased competition leading to lower prices, higher quality and broader choice for consumers of goods and services offered on online platforms - should outweigh the costs.

6.6 Impact on consumers

The impacts on consumers in terms of costs, choice and trust can be estimated as follows.

In terms of costs for consumer, the nature of the platform economy business-model (focused on consumer acquisition even at a loss, attention economy, generally low switching costs in many markets), combined with the proportionate nature of the obligations in all the retained options indicates that costs for consumers would not increase. On the contrary, the expected increase in the number of platforms and businesses active on these platforms would likely lead to increased choice for consumers on online platforms and increased competition among business users for these consumers, thus better prices and quality.

Option 2d, however, might increase in some consumer costs as platforms may be partially deprived of one of their key assets, i.e. consumer data. As the exact share of consumers who would agree to this measure is difficult to estimate, it is equally difficult to determine whether platforms would pass on possible losses from this measure to consumers.

When analysing impacts on consumers under option 2c, it is important to underline the value added that online general search engines have brought for them – in one-click consumers have access to a huge variety of information, businesses, goods and services. While the ranking transparency obligation does not target "consumer users" (i.e. searchers), a transparency obligation would indirectly contribute to safeguarding and possibly strengthening this positive effect. A transparency obligation would incentivise search engine operators to be more cautious and transparent in those instances where there a conflict of interest could exist between their own services and competitors' services. This would possibly contribute to a more impartial and procompetitive outcome for consumers.

Business-oriented fair practices would also complement consumer protection rules, thus enhancing end-users trust in the platform economy, giving them confidence in buying online

221 COM SWD(2017), 154 final , Commission Staff Working Document accompanying the Final report on the E-commerce

Sector Inquiry ( COM(2017), 229 final ), paragraph 449, 10 May 2017.

222 A large number of platforms active in the collaborative economy are however "hybrids", in that they enable peer-to-peer as well as business-to-consumer transactions.

thus benefitting from larger cross-border offer. This could support the existing trend of growing consumer trust in e-commerce. More than one in two Europeans now buy online (55% of consumers in 2016 223 ). Since 2014, consumers' levels of trust have increased by 12 percentage points for purchases from retailers located in the same country (72.4% of consumers are confident buying online in their own country) and by 21 % for purchases from other EU Member States (57.8%).

6.7 Environmental impact

No direct environmental impact of the measure is expected.

6.8 Impact on fundamental rights

All co-regulatory options fully comply with the Charter of Fundamental Rights ("CFR"), in particular with Articles 8, 16 and 47 CFR. The rationale behind this initiative is to establish a more balanced commercial relationship, while ensuring that platforms maintain full discretion over their business concepts. The underlying analysis showing this compliance is available in annex 12.8.

6.9 Summary of impacts of the retained policy options

Summary of impacts of retained options

Type of impact Option 2a Option 2b Option 2c Option 2d

Internal market Neutral to Positive: reduced As in option 2b: As in previous options

fragmentation positive: legal need to intervene at transparency and 2b and 2c: transparency and national level to redress measures transparency and redress measures resolve them. The will increase legal redress measures will

will increase legal new rules together certainty, and increase legal certainty. with the EU monitoring tools certainty, and

However, with observatory will help will help limit monitoring tools will

various highallow building a further help limit further impact trading common EU fragmentation. fragmentation. practices being understanding of

left to selfwhat issues are, thus regulation also allowing for (ranking, data, more consistent

discrimination, regulatory MFNs), this approaches when options may not such are deemed prevent direct necessary at national

harm to level. businesses. The effect of EU

monitoring / pushing reputational

levers will also be less pronounced in the absence of

legal transparency

223 Consumer Conditions Scoreboard – Consumers at Home in the Single Market. 2017 Edition.

obligations on all

high-impact trading practices. The risk of

fragmentation thus remains large since national

authorities would continue intervening to

solve existing imbalances.

Growth Neutral to Positive: increase in Positive impacts Uncertain: risk of positive: increase trust in the platform as in Option 2b, free-riding by

in trust in the environment will extended also to business users limits platform support growth in the online positive impact on environment will sales realised on general search sales realised on

support growth in online platforms. This environment. online platforms. sales, while impact, positive both While growth in the limited legal for platforms and online intermediation

obligations business users would could be suboptimal, represent contribute to it could possibly be minimal mitigating possible compensated to some compliance costs. impact of compliance extent by growth in

Leaving highcosts. The expected online sales on profile issues growth of the business users' own such as ranking, platform economy in websites.

data, conjunction with discrimination foreseen growth of and MFNs to selfonline activities could

regulation can be expected to however contribute to digital undermine some growth within the

of the trustinternal market. building effect of the legal measures on

terms & conditions, delisting and

redress, following increasing awareness and

concern among businesses.

Platforms Minimal costs: Limited costs: the Limited costs: the Negative to neutral:

The legal impacts of 2a are extension of the While thresholds may

transparency also valid here. The ranking help small platforms obligations on transparency extension of the to develop new terms and legally binding nature requirement and

conditions and of the requirements parts of redress

business models, this

option may impact

delisting imply to all areas of issues would create platforms' business only limited costs would create some some limited

for platforms additional one-off costs for online

strategies. It does

imply relatively high

while allowing and running costs as general search costs for platforms them to benefit engines. The compared to 2a. without creating

from increased These costs remain additional trust benefits to

trust. limited to ensure the for search compensate for costs. effectiveness of the engines which

Smaller platforms would result from obligations and

are exempted the initiative is exempt small from the most expected to platforms from burdensome disproportionate counterbalance internal redress the limited costs. burden (possibility

mechanism for thresholds to be obligation. The reviewed in a second two-step step). The limited approach will regulatory costs are ensure that expected to be thresholds are outweighed by the adapted if increased growth needed to opportunities for respond to the platforms (more evolution of the sales would also fast-changing translate in increased platform commissions for economy. platforms) created by Threshold for positive indirect small platforms network effects. will support Small platforms can emergence of benefit from new players. appropriate thresholds to grow. A

clear and predictable regulatory environment would

also support emergence of new players.

Business users Positive: Positive: Business Positive: In Positive to neutral: In

proposed users would benefit addition to addition to benefits measure, while from greater benefits under option 2c, not binding for all predictability. Clarity identified under option 2d would allow issues, would of platforms' policies option 2a, business users to

offer a higher and improved access businesses would expand their quality to redress would be able to customer base. experience for allow businesses to develop better Potentially negative

business users. better grasp the informed search effects are however innovation and optimisation not to be excluded in business strategies. This the longer term if

opportunities offered would be platforms' by online particularly environment intermediaries. beneficial for dynamics change as a

Business users' sales SMEs and result of option 2d's would grow as a enterprises with impact on platforms' result from their no or emerging business strategies.

increased trust in the online presence. platform economy and stemming strengthened

network effects.

Public Limited costs: As Option 2a. As Option 2a. As option 2a. legal obligations

administrations require mostly

one-off implementation and are easy to

monitor. Moreover, instead of relying on public

enforcement that would have to be financed by

national or EU administrations, the co-regulatory

design of this policy option foresees private

litigation by representative associations and therefore implies

no additional costs. Finally, the Commission will

bear the costs of EU Observatory and portal.

Employment and Neutral to Positive: as in option Positive: As Uncertain: impact

Social impact positive: growth 2b but the option 2b, option would depend on of the platform maintaining effect on 2c could be whether the economy will employment could expected to lead immediate effect of

contribute to be expected to be to increased possible increase in maintaining the stronger due to the employment number of jobs by 4.7 million jobs additional incentives opportunities for business users would

created by provided by the legal businesses outweigh possible job business users of character of dependent on cuts. The latter would platforms. transparency online platforms be due to potential measures for all six and on general decrease in business

issues. Expected search engines. opportunities in the positive social Should the platform economy as impact. positive effect be a result of option 2d –

more limited related obligations. than expected, option 2c would

contribute to maintaining existing jobs.

Innovation Neutral to Positive: growth of Positive: Better Negative: may

positive: any sales will create insight in ranking interfere with action injecting revenues for policies could platforms' ability to trust would innovation both for help business develop new dataincrease

sales, platforms and users grasp driven services and thus creating business users. innovation products. This could revenues for opportunities in turn reduce

innovation. offered online. innovation However, as with opportunities offered the foreseen impact on to business users.

growth, the limited scope of the legal

transparency obligations can undermine some of the trustbuilding

effect of the intervention and thereby limit

the positive impact on innovation.

Competitiveness Neutral to Positive: Positive: As in Negative to neutral:

positive: costs competitiveness option 2b. In limited impact on will be would be reinforced addition, if price-competitiveness particularly through the greater insight in but negative impact

limited and priceenhanced incentives SEO leads on innovation and competitiveness for platforms to businesses to international therefore compete on the basis access new competitiveness.

unaffected; of greater markets and innovation could transparency and embrace increase and increased quality of innovation (cf. support service. Positive above) this could

international impact on trust and translate in competitiveness. presumably growth strengthened of the sector would positive impact

support international on competitiveness competitiveness. through increased

innovation opportunities for business users and

platforms. Start-up and small platforms could use increased transparency and

redress standards (incentivised through the measure) to build

a competitive advantage.

Competition Positive: the Positive: expansion Positive: Negative to neutral: mandatory of user base will Increased proas for option 2c but

transparency support emergence competitive the business model of provides of new, small effect among platforms may be additional platforms. platforms impacted by sharing

competitive through of data, which makes parameters for enhanced the effect uncertain start-up platform transparency. on inter-platform

companies. Even Potential indirect competition. small increases in effect: Uncertainty is also user trust will competition true for business- and

equally support among business end- users – the the growth of users could potential positive existing platforms potentially effect expected could as well as the increase due to be counteracted by

emergence of greater insight in consumers concluding startup ranking policies. fewer transactions platforms. because of the

increased burden to prior respond to email sharing requests.

Consumers Neutral to Positive: improved Positive indirect Uncertain: sharing of positive: linked to P2B relations will effect: could data is a direct burden

growth, the allow maintaining contribute to a for consumers. If limited scope of and possibly more impartial, considered too heavy the legal increasing pro-competitive by them, the emailtransparency

consumers' choice in outcome in the related obligation obligations can terms of quality form of higher may result in reduced undermine some goods and services relevance results purchase on online

of the trustoffered at a being more easily platforms. This could building effect of competitive price. identifiable. negatively affect the intervention Legal transparency indirect network

and thereby limit obligation on ranking effects and lead in the the positive combined with EU long term to reduced impact on monitoring may in consumer choice.

consumer choice. addition have particularly

important indirect positive effect for

consumers that will be able to make more informed

purchasing decisions.

Environment Neutral Neutral Neutral Neutral

Fundamental rights The legal Legal measures on The targeted Article 8 CFR is transparency and technically complex legal safeguarded, as redress measures issues such as transparency compliance with

will help ranking imply some obligation on GDPR would be safeguard the risk of legal ranking in online ensured. The data freedom to uncertainty for online general search sharing obligation

contract (to platforms, which strengthens the however risks

conduct a should however be positive impact negatively affecting

business), by alleviated by the coof policy options online platforms'

improving the regulatory design 2a and 2b on freedom to conduct a situation for that uses business users' business. business users technologically right to conduct

without affecting neutral legal their business, online platforms' principles to be while leaving the freedom to implemented by corresponding determine their industry. These right of providers

business measures combined of online general concepts with the legal redress search engines implemented in measures at the unaffected.

general terms same time ensure the and conditions. appropriate respect These measures for business users'

will also improve freedom to contract the right of as well as for their access to justice right of access to

for business users, by justice.

addressing importing barriers to

accessing crossborder justice.

7 C OMPARISON OF OPTIONS

7.1 Comparative analysis of retained co-regulatory options

This Section discusses the effectiveness, efficiency, coherence and proportionality of the four retained policy options 2a, 2b, 2c and 2d (such as defined in Table 1 and summarised in the introduction to Section 6) in comparison with the baseline scenario. The baseline scenario (option 0) as basis for comparison is described in Section 2.4 and 5.1.

All four retained options consist of measures on transparency and redress, and are intended to leverage the foreseen enhanced monitoring of the platform economy 224 . The options are largely similar with respect to the redress (see Section 7.1.1 in this respect) and monitoring tools proposed. The comparison of the options on the basis of their effects against the baseline would hence mainly be influenced by the nature of the transparency obligations in each of the options: option 2a imposes legal transparency obligations on two out of the six most frequently observed, potentially harmful trading practices, option 2b proposes legal transparency obligations also in respect of the four practices left to self-regulation in 2a; option 2c extends the legally binding transparency obligation on ranking also in a targeted manner to the issue of ranking in online general search, while option 2d adds the obligation for platforms to provide any business user with the opportunity to ask, in line with the GDPR, for customers' consent for the business to obtain and process their e-mail addresses after the completion of a transaction and the payment of the commission to the respective platform. The fact that the vast majority of platforms do not allow this has been identified as a crucial problem for business users related to data as described in Section 2.1.1.

7.1.1 Effectiveness

Objective 1: Ensuring a fair, transparent and predictable treatment of business users by online platforms

Options 2b, 2c and 2d address the potentially harmful trading practices identified in the problem statement thus contributing to ensuring a fair and predictable business environment for platforms’ business users. In particular, (i) the transparency requirements on ranking, discrimination, data and MFN clauses, (ii) the obligation on platforms to provide information about substantial changes in terms and conditions, and to grant firms a reasonable notice period to adapt to the changes announced, and (iii) enhanced redress possibilities for platforms' professional users, together with the continuous monitoring of the platform ecosystem contribute to a fairer, more predictable and trusted business environment based on a set of enforceable rules.

224 The entire set of measures will be further explained in Section 8 for the preferred option(s) .

Options 2b, 2c and 2d are all more effective in reaching the first specific objective as compared to the baseline, as they all provide for greater predictability and transparency on elements of the relationship between platforms and businesses. Given that the scope of the legal measures foreseen increases from policy option 2b to policy option 2d, the effectiveness of each of these three options as compared to the baseline also increases from 2b to 2d.

Option 2d is thus particularly effective in achieving specific objective 1 since it provides for a significant increase in transparency offered to business users who would in addition have access to email contacts of their customers. Option 2a differs in that the action it foresees to address four out of the six harmful trading practices is limited to self-regulation. By leaving significant issues in relation to transparency on data access policies, ranking, MFNs and discrimination to industry, option 2a is, however, likely to be less effective as the participation in self-regulatory schemes is less likely to be uniform across the Digital Single Market. Option 2a could thus be expected to lead only to a modest improvement over the baseline.

Options 2c and 2d foresee the inclusion of online general search engines in the initiative. These two options thus allow the initiative to simultaneously cover the two most important ranking-based originators of Internet traffic (online platforms and online general search engines). Options 2c and 2d appear therefore highly effective in achieving the objective of more transparent and predictable environment for businesses dependent on both online platforms and online general search engines. The effectiveness of Option 2b as regards Objective 1 is between that of Option 2a and 2c and 2d, and therefore intermediate.

Objective 2: Setting effective and agile redress for businesses, adaptable to the evolving market

All retained policy options offer solutions to the lack of redress issue identified (thus contributing to achieving specific objective 2) through the proposed set of enforcement measures. The latter consists of (i) the obligation to provide for effective internal complainthandling, (ii) the obligation to identify mediators and to engage with them in good faith, and (iii) the provision on standing for business associations (collective injunctive redress). In addition, the redress-related measures include the dual call on industry to voluntarily (a) explore developing codes of conducts for the internal complaint-handling, and (b) set up an external independent redress mechanism at EU level responding to an effectiveness legal requirement. Whereas standing for business associations is meant exclusively to enforce the light-touch transparency & redress obligations in the P2B Regulation, 225 the obligations on internal complaint-handling and mediation are intended to allow business users to bring complaints on any substantial issues arising out of their contractual relationship with online platforms (including any alleged act in breach of any legal obligation).

Option 2a, can be expected to contribute to a more limited extent to offering effective redress for the four harmful trading practices of ranking, MFNs, discrimination and data left to selfregulation. In the absence of a regulatory backstop in the form of legal transparency principles

225 Nothwithstanding this, the results of surveys with professionals with relevant expertise in Croatia, the Czech Republic, Denmark, Estonia, France, Greece, Hungary, Italy, Lithuania, Malta, the Netherlands and Poland, consider that access to justice is enhanced by collective redress. See the analysis of the state of collective redress in the European Union in the context of the implementation of the Recommendation of the Commission on common principles for injunctive and compensatory collective redress mechanisms in the Member States concerning violations of rights granted under Union Law (request for services JUST/2016/JCOO/FW/CIVI/0099, Lot1/2016/06) prepared by the British Institute of International and Comparative Law (BIICL), in a research consortium with Civic Consulting and Risk & Policy Analysts (RPA), and supported by the Office for Economic Policy and Regional Development (EPRD), (forthcoming).

and standing to enforce these transparency principles by business associations, business users could reasonably be expected to feel less empowered for these four specific practices to confront online platform providers in light of the very significant fear of retaliation.

As explained in Table 1, options 2c and 2d involve an issue-specific extension of the intervention to ranking in online general search. To ensure proportionality in light of the fundamental differences between online platforms and online general search engines (cf. section 2.1.1.3), the improved redress for issues with online general search engine providers will be limited to legal standing for business users' representative organisations.

The issue of ranking is nonetheless the most business-critical issue that has arisen in this context, hence its inclusion in the retained policy options policy (namely 2c and 2d) in the first place. The apparent reduced scope of issues that may be brought before the out-of-court redress mechanisms in relation to issues encountered with online general search engines (i.e. limited to ranking) as compared to the three categories of online platforms therefore does not negatively affect the effectiveness of policy options 2c and 2d. The latter options actually imply a higher degree of effectiveness in achieving specific objective 2, relative to options 2a and 2b, by virtue of the range of issues which can be challenged in a redress process.

Objective 3: Preserving a predictable and innovation-friendly legal environment for online platforms within the EU, without placing undue administrative burden on platforms

In addition, all three options 2b, 2c and 2d would set a harmonised legal framework for P2B relations on all core potentially harmful trading practices that have been observed, contributing to creating a predictable and innovation–friendly legal environment for platforms (specific objective 3). The platform economy would thus benefit from more consistent regulatory approaches across the EU.

None of the options aims to harmonise national B2B fairness legislations. Nonetheless, the legal framework the co-regulatory options propose, combined with (i) the increased industry involvement through voluntary commitments, and (ii) the collaborative monitoring with Member States would limit the legal fragmentation of the internal market and ensure a more harmonised EU approach to P2B practices across the EU, thereby supporting a more predictable regulatory environment at EU level and increasing businesses' trust thus benefitting the growth of the entire platform sector through amplified network effects.

Option 2a cannot meet the third specific objective of ensuring regulatory predictability for online platforms and general search engines since this option would not sufficiently contribute to addressing the emerging issue of legal fragmentation. This relatively weak effect would result, on the one hand, from less effective and patchier industry-led action on complex issues such as ranking in the absence of a principles-based harmonised legal backstop. On the other hand, leaving the four most controversial commercial issues observed (ranking, discrimination, data and MFNs) fully to self-regulation, while nonetheless introducing a legal framework for the overall P2B relationship within which these occur, risks increasing political pressure on national regulators to take more far-reaching action on these topics.

Options 2b and 2c, through a comprehensive set of measures could be expected to contribute to a general legal framework and EU level and increase platforms' innovation capacity. The option is therefore more effective than the baseline in ensuring an innovation–friendly legal environment for platforms as foreseen in specific objective 3.

This effect is not as clear under option 2d which risks running counter to platforms' innovation incentives. The obligation for platforms to allow business users to ask for customers' consent for further processing of their e-mail addresses impacts data which is a core element of platforms' business models. More information would be needed to assess the effects on competition and innovation as well as on any potential free-riding phenomenon. Option 2d thus risks being relatively ineffective in reaching specific objective 3.

Conclusion on Effectiveness

Option 2a incorporates - relative to the other co-regulatory policy options - the greatest degree of uncertainty in terms of its effectiveness to achieve specific objectives 1 and 2. It will nonetheless provide a legal framework for platform-to-business relations incorporating important redress provisions to help minimize frictions in platform-to-business relations. This framework will also leverage the foreseen enhanced monitoring of the online platform economy, including on those potentially harmful trading practices left to self-regulation. In absolute terms, policy option 2a can therefore be expected to contribute to achieving specific objectives 1 and 2 as compared to the baseline with moderate effectiveness. As regards specific objective 3 however, policy option 2a would have more limited effectiveness since it leaves a relatively large scope for fragmentation to increase.

Option 2b is effective as regards all three objectives: it provides a holistic set of transparency measures for a predictable business environment in areas that matter for businesses, while allowing redress for these issues, and at the same time providing an innovation-friendly business environment for online platforms.

Option 2c is more effective still since its expanded scope also covers online general search the expected positive effects of the transparency obligation and of the legal standing provision as appropriate and proportionate. Increased trust and legal certainty would extend also to the online general search environment. The extension of the scope to online general search engines would help optimising businesses' online visibility across the board. Greater transparency would also increase trust in online general search engines.

Option 2d appears effective in achieving the first and second specific objectives, but significantly less effective on objective 3, by potentially circumventing the business model of online platforms.

7.1.2 Efficiency

Option 2a is expected to lead to rather limited implementation costs. The broader range of issues left to self-regulation by industry suggest that option 2a would be less burdensome for platforms since they would be essentially free to respond to the call for transparency on a number of more complex and important commercial issues. At the same time, the benefits brought by this policy option could also be more limited by leaving four out of the six harmful trading practices out of the scope of the regulation. The limited costs should thus be compared with the more likely limited benefits.

Option 2b is a cost-efficient solution to ensure increased transparency, effective redress and, as under the other co-regulatory policy options retained, an appropriate monitoring of the platform economy. As described in Chapter 6 and Annex 4.1.2, the measures on transparency will require relatively small adjustments to be made by online platforms (one-off costs to adjust their terms and conditions, including legal and communication costs, and limited running costs when their terms and conditions change, which would be similar to the baseline scenario), and would provide greater legal certainty to business users.

The possible risk of any negative impact on legal certainty for online platforms is also limited, given the important scope for industry action in implementing the details, and the generally aligned incentives (as described in more detail in chapter 6). Exempting certain categories of online platforms (e.g. smaller ones) from the potentially more burdensome legal obligation to provide effective internal complaint-handling mechanisms, while enabling these firms to optin, would at the same time ensure that the cascade of legal obligations to provide redress is cost-efficient, in particular also in light of the co-regulatory design of the measure (building on existing best practices) and the largely aligned incentives of platforms and business users to minimize frictions.

The legal redress obligations also build on the out-of-court dispute settlement with a strong track record in solving P2B issues (online mediation) and open these for all business users, regardless of their size. The provision giving legal standing to representative organisations to bring injunctive actions to enforce the effectiveness of internal complaint-handling mechanisms and platforms' mediation efforts will in this regard help nudge even very small business users past their fear of retaliation, as they could anonymously report structural deficiencies. This effect will be further reinforced by the foreseen enhanced monitoring of the online platform economy. Research has also shown that collective redress measures are beneficial to SMEs 226 . Finally, the foreseen monitoring measures would create limited costs for public authorities, which would mostly be covered by running administrative budgets. And although the monitoring measures are light-touch, they have an immediate reputational effect and are important in informing the review of the intervention. Overall, option 2b is an efficient measure that effectively reaches the objectives of the action while limiting costs. While online platforms may face increased public scrutiny in areas of commercial relevance (like data access policies or the use of MFN clauses), increased transparency together with continuous EU monitoring of the platform economy and public scrutiny should yield increased trust and improved competition to the benefit of all market participants.

Option 2c would extend the compliance requirements to online general search engines. As explained in Section 6, these costs would be related to (i) the implementation of the ranking transparency obligations and (ii) the legal standing obligation from which some limited litigation costs could stem. The additional compliance costs are expected to be rather limited both for bigger search engines (which have provided SEO guidelines that could be usefully re-purposed for business users, or serve in some cases as inspiration for ways to provide meaningful transparency), and for smaller ones (since they would be able to equally draw from existing best practices). Assuming that meaningful transparency requirements are implemented, additional litigation costs would be limited to non-compliance cases. It is important to stress that this option does not generate costs due to loss of business or trade secrets related to disclosure of algorithms, as the requirements would be limited to providing necessary and sufficient information to provide businesses with an understanding of the link between ranking and features of their products and services, as well as the necessary predictability.

At the same time, by covering also ranking issues in online general search, Option 2c would create additional benefits. Businesses would be able to develop better informed search

226 See, e.g., M Pakamanis, Journal of International Comparative Jurisprudence, Dec 2016; and Centre for Justice and Democracy: "How Small Businesses Benefit from Class Action", 2013. These conclusions are predicated on an assumption that representative organisations would have standing in the first place to bring such an action.

optimisation strategies, greater insight in which could lead businesses to access new markets and embrace innovation opportunities offered online, thus also enhancing their competitiveness. Option 2c could also be expected to have a pro-competitive effect between search engines and comparison sites through enhanced transparency, as public ranking policies can provide a greater scope for differentiation by start-ups, new entrants as well as existing players. Competition on quality of products and services among business users dependent on search engines for their marketing strategies could potentially also increase due to greater insight in ranking policies – to the extent such business users' website design are currently sub-optimal for achieving visibility. It is not excluded that Option 2c therefore equally contributes to a more impartial outcome for consumers in the form of higher relevance results being more easily identifiable. A transparency obligation set in EU law would strengthen businesses' ability to use such a provision in court proceedings. Finally, it would also be a helpful complement to enforcement tools under competition law since it would allow greater insight in possible discriminatory behaviours. The additional trust for search engines which would result from the initiative could be expected to counterbalance the limited costs.

The implementation costs of 2d stem from the additional measure (as compared to 2c) to allow consumers to express their consent to share some data with business users requires only a technical adjustment on the side of platforms, as the data itself can be shared through existing communication channels. Importantly, such a possibility for business users will be linked to the completion of a transaction on the respective platform and to the payment of the platform's commission. Nonetheless, if the sellers have their own sales channels, it may also allow them to circumvent the platform for future transactions and disrupt a core aspect of the platform business model. The increase in free-riding behaviour by business users could endanger online platforms' business models, resulting in decreases in sales through online platforms, limited innovation, and an increase in prices, including for consumers. As a result, despite its limited implementation costs, the negative impacts of option 2d are not negligible and it is not considered efficient in comparison with the baseline scenario.

Conclusion on Efficiency

All three Options 2a, 2b and 2c are efficient. Option 2a implies limited costs but its benefits would also be limited. Compared to Option 2b, Option 2c would create larger benefits even though these would also be achieved at a somewhat higher marginal cost. Option 2d appears inefficient; it would entail important costs while there is uncertainty as to the benefits it could bring.

7.1.3 Coherence

In their objectives, all retained co-regulatory options 2a, 2b, 2c and 2d are coherent with the Digital Single Market Strategy, because they aim at ensuring a fair, open and secure digital environment such as announced in the Commission's DSM mid-term review. It needs to be noted however that the implementation of option 2d may not de facto lead to a strengthened Digital Single Market, as certain consequence of circumventing the platform's intermediation role are not fully understood. Given the importance of platform innovation and the importance of the platform economy for digital transformation and growth, option 2d may, by intervening in online platforms' business model, have a detrimental impact on platforms' incentives and opportunities to innovate and thus negatively impacting the growth potential of the Digital Single Market.

All retained co-regulatory policy options are also consistent with the Communication on Illegal Content, as measures providing greater clarity reasons for delisting will also support intermediary liability actions. Finally, the measures on redress included in both options are consistent with the New York Convention on Arbitration given that this Convention does not preclude the creation of platform-internal or external dispute settlement mechanisms.

All four retained options are in addition consistent with the Trade Secrets Directive 227 and any information or data lawfully acquired in a P2B2C relationship will continue to enjoy trade secret protection, if it falls under the definition in the Directive. Consistency with the GDPR is also ensured in both options, because the access to personal data remains subject to the full compliance with the GDPR, in particular on the further processing of e-mail addresses in option 3 requiring a valid and informed consent by the data subject as stipulated in Art. 6 of the GDPR.

All options are have been verified to be coherent with current proposals in the area of copyright (including the aspects related to press publishers rights), consumer law, and upcoming proposals in the field of taxation.

In addition, all four options are coherent with competition law. The Commission would continue using its competition law enforcement tools under all four co-regulatory options retained.

Finally, the definitions covered by all four options are coherent with definitions used in EU existing legislation and currently ongoing Commission initiatives. For the purpose of exhaustiveness, Section 8.3 of the Annex contains an exhaustive overview of all current and proposed definitions for comparison purposes. The Table in Annex 1.3 is an extract focusing on comparison/compatibility with major initiatives in the field of taxation and consumer protection.

Conclusion on Coherence

Options 2a, 2b and 2c are all three coherent with other EU policies and with fundamental rights. Option 2d appears not fully aligned with the objectives of the DSM strategy, as the straight circumvention might impact the platform ecosystem and slow down innovation.

7.1.4 Proportionality

EU level action is appropriate due to the intrinsically cross-border nature of online platforms. EU level action is the only way to provide the necessary high-level and harmonised legal framework for platforms to scale up, which is an indispensable condition for their business strategies and overall economic growth. From that perspective, all four retained policy options are appropriate since they provide for an action at EU level.

Option 2a is in principle proportionate since it provides, at limited costs, a comprehensive set of measures that could allow all three specific objectives to be achieved. However, by leaving issues around internal discrimination, data, ranking and MFNs to self-regulation, policy options 2a incorporates the risk of increased fragmentation. The proportionality of option 2a should thus be assessed in light of the extent to which the option contributes to achieving the objectives of the initiative.

227 Directive (EU) 2016/943 of the European Parliament and of the Council of 8 June 2016 on the protection of undisclosed know-how and business information (trade secrets) against their unlawful acquisition, use and disclosure, OJ L 157/1.

Options 2b and 2c offer EU business users trading on platforms an improved and more predictable business environment. At the same time, platforms remain free to set the general policies for their market places and are offered the opportunity to shape their voluntary commitments by way of industry codes of conducts within a principles-based legal framework. These options thus avoid intrusive full-fledged legislation, providing instead an adequate response to these specific problems, whilst safeguarding the innovation capacity of platform firms. In addition, the important harmonising effect of the options will facilitate the scaling-up of online platforms to the benefit of all actors in online intermediated trade as harmonised EU rules will inherently lower compliance costs and enhance legal certainty in particular for cross-border operations.

Option 2c implies obligations also for online general search engines. Due to their somewhat different nature 228 , online general search engines (regardless of their size) have to index the largest possible number of websites, while also competing on the quality of their algorithms, which are therefore subject to, in some cases, thousands of changes every year. The scope for complaints against online general search engines could thus be higher and potentially concern a broader range of issues; the related litigation costs could thus also be important. In order to ensure the proportionality of the measure and avoid burdensome costs for online general search engines, option 2c foresees an issue-specific extension of the intervention to ranking in online general search. In addition to a targeted legal transparency obligation, option 2c exclusively provides that online general search engine providers will be called upon to voluntarily explore developing codes of conducts. Option 2c ensures the proportionality of costs by the light-touch nature of the legal transparency obligation, which does not provide for any algorithms' disclosure and which builds on existing practices in terms of transparency (Search Engine Optimisation). In addition, the redress measures that could be invoked against providers of online general search engines would be limited to the light-touch enforcement mechanism foreseen for the initiative 229 , and concern a more limited legal transparency obligation (limited to ranking) than that envisaged for business users of online platforms.

Option 2d does not appear to be a proportionate solution since it goes beyond what is necessary to achieve the objectives defined. A sufficient level of transparency and predictability for business users could be achieved without risking interfering in the platforms' business strategy. While the technical costs associated with sharing consumers' email addresses are limited, it is difficult to qualify and quantify the further effects this option may have on the platform economy.

The proportionality of the co-regulatory options also lies in their two-step approach which is tailored to platforms' fast changing technological and economic environment. The EU Observatory will follow both the general evolution of the platform economy and the specific issues described in this Impact Assessment, informed amongst other by the legal transparency obligations. Requiring platforms to report in a non-detailed manner 230 , capable of being performed using automated data collection and reporting techniques, is a proportionate way to encourage the use of the internal complaint-handling system to show how effectively it functions. The value of such increased transparency, particularly as the statistics shall be available to the general public, reflects the objective of transparency rules in other sectors, for example, in investor-state arbitration with the aim of, inter alia, increasing accountability and

228 The core business of online general search engines is to index the entire Internet, also outside any contractual relationship with websites, whereas online platforms can grow somewhat more organically with the number of their business users. 229 Redress with regard to online general search engines would be limited to granting representative associations legal standing to act on behalf of the businesses they are representing. 230 For instance a high level report on the total number of complaints received, the subject matter of the complaints, the time period needed to process the complaints and the decision taken.

promoting good governance. 231 Options 2b and 2c will thus ensure that the approach put in place remains proportionate to the issues encountered in P2B relations. Policy option 2c will likely increase the effectiveness of the intervention as opposed to policy option 2b. While this would imply additional costs as compared to option 2b, the measures of option 2c are proportionate to what is strictly needed to contribute to achieving the objectives; i.e. the measures foreseen with regard to online general search engines are limited to the only issue of ranking transparency and to the legal standing for representative organisations to act on behalf of their business members.

If platforms fail to put in place the voluntary approach recommended by the Commission, stricter and more intrusive rules could be put in place if justified in light of observed developments. In addition to constituting a pre-condition to effective monitoring, the legal transparency obligations also have an important potential to increase peer competition as they provide new competitive parameters.

Finally, all retained options foresee the inclusion of a targeted exemption of online platforms constituting a "small enterprise" from the obligation to provide effective internal complaint-handling mechanisms, in line with the "Think Small First" principle. It should be noted that the obligation to provide effective internal complaint-handling mechanisms sets only high-level effectiveness and accessibility criteria, which will leave online platforms free to implement cost-effective technical solutions resulting in lower than average costs. For example, platforms would be free to reuse consumer-facing support mechanisms also to provide complaint handling for the business side of their operations.

The targeted exemption will guarantee that where an administrative burden resulting from the Initiative cannot be fully excluded, only companies that are sufficiently mature to absorb this burden are covered. The internal redress – related obligation will at the same time extend to all platforms that have a very large number of business users and which are therefore most likely to face capacity constraints in complaint-handling in the absence of proper procedures.

In addition to a targeted exemption, the proportionality of the retained policy options could be further ensured by horizontally exempting very small online platform firms from the overall P2B initiative. Depending on the threshold that may be used, a significant number of smaller online platform companies in the EU would not, in that case, face any additional burden, while being able to benefit from the enhanced user trust that the Initiative should yield 232 . At the same time, the cost-benefit ratio of such an exemption can be estimated to be relatively neutral, as the additional burden on very small platforms to provide for transparency obligations in their terms of reference can be expected to be minimal.

Conclusion on Proportionality

Option 2a is a proportionate choice since it implies limited burden on platforms and no obligations on online general search engines.

Options 2b and 2c appear both as proportionate since they allow meeting the objectives of the initiative while imposing relatively reasonable level of burden.

231 The UNCITRAL rules on Transparency in treaty-based Investor-State Arbitration provide for extensive disclosure in case of investor-state disputes, with details such as the notice of arbitration being sent to a repository.

232 According to the Dealroom database there are 7,012 EU platform businesses, 1,919 of which would have less than 10 employees.

Option 2d appears disproportionate since it implies relatively high burden for fulfilling the objectives of the initiative.

7.2 Conclusion

The following table summarises the comparison of the retained policy options 2a, 2b, 2c and 2d in terms of effectiveness, efficiency, coherence, and proportionality.

It is important to stress that the definitions covered by all four options are coherent with definitions used in EU existing legislation and currently ongoing Commission initiatives. For the purpose of exhaustiveness, Section 8.3 of the Annex contains a detailed overview of all current and proposed definitions for comparison purposes.

Comparison of

policy options Effectiveness Efficiency Coherence Proportionality

Option 0 0 0 0 0

Option 2a + ++ +++ +

Option 2b ++ +++ +++ +++

Option 2c +++ +++ +++ +++

Option 2d ++ - ++ -

Option 2a - Foresees self-regulation on ranking, discrimination, data & MFNs and legal principles on the issues related to terms and conditions and delisting; excludes issues encountered in online general search as well as the data-related obligation for platforms to share email addresses with their business users Option 2a is of limited effectiveness. While it would increase legal certainty through legal transparency and redress measures, the risk of direct harm to businesses and of fragmentation remains important since some high-impact trading practices (ranking, data, discrimination, MFNs) may not be sufficiently tackled through self-regulation. Reputational levers of EU monitoring would also be underexploited in the absence of legal transparency obligations on all high-impact trading practices. In terms of efficiency, this option while achieved at a lower cost may underperform with regard to the achievement of the objectives set by the initiative thus leading to limited benefits. The limited effectiveness of Option 2a has been taken into consideration in the scoring of proportionality. Option 2a is in principle proportionate since it provides, at limited costs, a comprehensive set of measures that could allow all three specific objectives to be achieved. However, by leaving issues around internal discrimination, data, ranking and MFNs to self-regulation, policy options 2a incorporates the risk of increased fragmentation. The proportionality of option 2a should thus be assessed in light of the extent to which the option contributes to achieving the objectives of the initiative.

Option 2b - Foresees legal principles on all issues; excludes issues encountered in online general search as well as the data-related obligation for platforms to share email addresses with their business users Option 2b is effective since it would lead to a more predictable, transparent and innovationfriendly environment for business users. The new rules together with the EU observatory will help allow building a common EU understanding of what issues are, thus also allowing for more consistent regulatory approaches when such are deemed necessary at national level. Increased trust in the platform environment will impact in a positive way both platforms and business users, contributing to mitigating possible compliance costs. In terms of efficiency, the regulatory costs are expected to be outweighed by the increased growth opportunities for platforms created by positive indirect network effects which would support in growth in sales through online platforms. Option 2b appears as an effective, efficient and proportionate option coherent with other EU policies.

Option 2c - Foresees legal principles on all issues; covers issues in online general search, excludes the data-related obligation for platforms to share email addresses with their business users Option 2c has a higher overall effectiveness compared to Option 2b, but also achieved at a higher cost. It allows expanding the expected positive effects of the transparency obligation and of the legal standing provision to online general search as appropriate and proportionate. Online search engines also exhibit dependency-enabled unilateral behaviour targeted by the initiative, as ranking practices directly influence websites' visibility and internet traffic received. The limited extension of the scope to the transparency obligation for ranking and to the enforcement provision on legal standing for representative bodies is a proportionate and effective way to ensure clarity as regards the complementarity of the initiative with competition law while option 2c implies higher costs as compared to option 2b - since it creates costs for online general search engines - it also creates additional benefits. The inclusion of online general search engines in the initiative allows simultaneously covering the two most important ranking-based originators of Internet traffic (online platforms and online general search engines). Option 2c would help introducing more transparency and predictability for business users across the board thus addressing the issues identified around ranking transparency more comprehensively. Option 2c would ultimately create positive impacts in terms of online visibility for business users, increased pro-competitive effect (between comparison sites and search engines), and preserved quality of search results for end-users. In addition, associations or representative bodies that have a legitimate interest in representing businesses whose websites are indexed by online general search engines would have the right to seek action in court to enforce – exclusively – the legal transparency obligation on ranking in online general search. Option 2c is proportionate since the measures it proposes are targeted to the sole ranking-related issue identified in online general search. Option 2c foresees therefore a scoped transparency obligation (limited to the single issue of ranking), and a minimal redress measure limited to granting representative associations with legal standing to act on behalf of their business members in relation to the ranking transparency issue. Finally, option 2c is coherent with other EU policies and fundamental rights.

Option 2d - Option 2d is particularly effective in fulfilling specific objectives 1 and 2 since it allows for an exhaustive set of measures to address all issues identified. Option 2d appears however less effective in reaching specific objective 3 of creating a predictable and innovation– friendly legal environment for platforms. The extended data-related obligation for platforms to share customers' email addresses may have an impact on platforms' business models thus possibly reducing innovation incentives. This option is thus disproportionate and not fully coherent with the objectives pursued by the DSM strategy.

8 P REFERRED O PTION

Option 2c is the most effective of the retained options in reaching the specific objectives of the intervention, most notably in terms of ensuring fair, transparent and predictable treatment of business users. Option 2c allows expanding the expected positive effects of the transparency obligation and of the legal standing provision to online general search. It allows the inclusion of online general search engines in the initiative thus simultaneously covering the two most important ranking-based originators of Internet traffic. Businesses would be able to develop better informed search optimisation strategies, which would be particularly beneficial for SMEs and enterprises with no or emerging online presence. This option is of comparable efficiency as option 2b, which represents the closest regulatory alternative to option 2c. In addition, option 2c provides for equal treatment between online platforms and general online search engines, as regards dependency-induced potentially harmful ranking practices. Consequently, to maximise the effectiveness of the policy intervention and to ensure a level playing field, Option 2c is selected as the preferred option.

8.1 Overview of the measures

The Preferred Option is a co-regulatory set of measures combining obligations imposed in a legal instrument, self-regulatory measures by platforms to set up an independent mediator, and the establishment of an EU Observatory of the Online Platform Economy to monitor the problems identified. It is informed by and builds on the on best practices such as the Audiovisual Media Services Directive (AVMSD) and the experience in the Supply Chain Initiative (SCI).

Targeted EU-wide transparency measures are both the most proportionate way to effectively tackling established potentially harmful trading practices, as well as a precondition to effective monitoring (cf. Section 8.1.1 below on transparency and Section 8.1.3 on monitoring). Such transparency measures also have the potential to improve competition – and indirectly the quality and redress available to business users – by pushing reputational levers. This is also critical to developing a more trusted business environment. EU-wide measures to ensure effective out-ofcourt redress possibilities for business users are in addition required to minimise inadvertent economic damage arising out of any P2B issues. Such redress measures underpin the effectiveness of the proposed transparency measures. Transparency measures such as an obligation to state objective reasons for delisting in turn constitute an absolute precondition to enabling effective redress. This increased clarity develops trust and helps businesses overcome the fear of retaliation. The high-level nature of the transparency and redress rules would allow industry, if it chose to do so, to additionally develop codes of conducts, which could spell out legal and technical details of their practical implementation sensitive to the fast changing technological and economic environment in which they operate 233 . The transparency and redress actions will also be flanked by the separate establishment of an EU Observatory of the Online Platform Economy to ensure effective monitoring in close collaboration with Member States. The harmonised transparency and redress measures combined with collaborative monitoring will help prevent further fragmentation by avoiding direct economic harm to EU business users and by allowing the preparation of well-informed responses at EU-level – which may be justified to regulate emerging potentially harmful trading practices in a second step.

8.1.1 Enhanced transparency to tackle potentially harmful trading practices

  • 1. 
    Legal obligation to inform business users of significant changes to contractual terms and conditions, and to provide a reasonable notice period to allow business users to adequately

233 Such as for example the effectiveness of the technical means used for informing business users about mediation and redress possibilities. On the need to tackle those technical problems in the area of consumer dispute settlement see Report from the Commission to the Council and the European Parliament on the functioning of the European Online Dispute Resolution platform established under Regulation (EU) No 524/2013 on online dispute resolution for consumer disputes, forthcoming (in ISC), p. 5: Online traders are under an obligation to include a link to the ODR platform on their website. However, a scraping of more than 20 000 web shops across the EU conducted by the Commission to check traders' compliance showed that there is scope for improvement.

prepare for the anticipated changes. Online platforms would remain free to legally design solutions that effectively result in business users being provided with a period during which they can continue to trade subject to the existing terms and conditions.

  • 2. 
    Legal obligation to provide relevant business users with an actionable statement of reasons upon delisting of their accounts or of individual products and services.
  • 3. 
    Legal obligation to state in contractual terms and conditions the general criteria of the ranking mechanism on the platforms, as well as conditions for use of any mechanism that allows business users to influence their prominence against remuneration. It will be explained that all these obligations shall be without prejudice to the protection of trade secrets under Directive (EU) 2016/943 and that, therefore, the required description will be at a general level, while accurately reflecting the norm for the relevant online intermediation service, based on historic data. In addition, the Commission could work with the relevant online platform and online general search engine providers to explore practical tools for meaningful transparency, to improve the verifiability of this paid-for prominence within the specific e-commerce area, to potentially develop industry standards and proactively run audits; and it could monitor the functioning of the wider digital advertising space.
  • 4. 
    Legal obligation to clearly explain to business users whether platforms' competing services or goods (e.g. own apps, owned retailer operations or other services of the same type as the business users') enjoy any preferential treatment (e.g. pre-installation of apps) as compared to that reserved to business users. The description would be at the general level rather than at the level of individual products or goods offered through those services in order to ensure proportionality. This will include any specific measures or behaviour concerning access to personal or non-personal data, ranking, remuneration for the use of the platform or ancillary services (e.g. the provision of delivery services or payment facilities).
  • 5. 
    Legal obligation to clarify transparently the data policy regarding business users.
  • 6. 
    Legal obligation to unequivocally state as part of platforms' general commercial proposition to business users whether they demand – contractually or otherwise – best prices and/or product selections to be offered to their market places. This will be combined with a legal obligation for online platforms to make easily available to the general public unambiguous explanations as to the relevant commercial, legal or any other considerations underpinning the use of such most-favoured-nation clauses. For this to be proportionate, it will be limited to a targeted and understandable description of key elements such as the need to prevent freeriding in light of the size of the platform.

The preferred Option 2c adds a targeted transparency obligation for providers of online general search engines to the above transparency obligations for providers of online intermediation services. Providers of online general search engines would under this option be obliged to make a description of the main ranking parameters used to operate their search ranking mechanism available to the general public. The obligation would build on real-life examples of meaningful transparency (cf. Annex 7.3), and defines a legal standard on the basis of industry best practice. The transparency obligations shall be without prejudice to the protection of trade secrets under Directive (EU) 2016/943 and, therefore, the required description will be at a general level, while accurately reflecting the norm for the relevant online general search engines, based on historic data.

8.1.2 Improved internal, external and judicial redress available to business users

Redress possibilities for business users will be improved by a combination of measures, listed below. The resulting structural redress and some of the legal obligations tackling individual harmful trading practices will be directly mutually reinforcing. 234

  • 1. 
    Encouragement for platform operators to set up external organisations that can provide industry-specific mediators independent mediators, which should – if industry does in fact set them up – comply with certain effectiveness principles to be spelled out in law. These mediators would provide an additional pool of mediators in addition to those that already exist.
  • 2. 
    Legal obligation for platform operators to put in place an effective and accessible internal complaint-handling mechanism. The functioning of such internal complaint handling systems would be monitored for effectiveness, and subject to further recommendations or industry codes-of-conducts.
  • 3. 
    Legal obligation to issue annual public reports on the effectiveness of the internal appeals mechanisms, which will be designed to limit costs for the platforms concerned. Rather than implying any continuous publication of data (feeds), the reporting obligation would take the form of high-level reports published only annually and these reports would moreover cover only a limited number of elements such as the total number of complaints received, the subject matter of the complaints, the time period needed to process the complaints and the decision taken. Online platforms can largely automate data collection and reporting.
  • 4. 
    Legal obligation for platform operators to list in their contractual terms and conditions existing mediators who meet certain objective quality criteria, and with whom they are willing to engage. This shall include any platform-specific, independent mediators. Platform operators will in addition be subject to a legal obligation to act in good faith towards attempts by their business users to engage with such mediation process in the EU.
  • 5. 
    Formulating the legal transparency and redress obligations as mandatory rules to the greatest extent possible (cf. Annex 1.4 in this respect) as a key element to improve the chances of enforcement of the proposed targeted legal obligations in EU courts notwithstanding the exclusive choice of law and forum clauses included in the contracts between online platforms and their business users that frequently designate non-EU courts.

Granting to associations or representative bodies that have a legitimate interest in representing business users the right to seek action in court to remedy or prevent economic harm arising as a result of general non-compliance with the proposed legal obligations for online platforms. Annex 1 explains the compatibility with the Commission's collective redress Recommendation of 2013. Although the principal aim of the Initiative is to improve bilateral conflict resolution rather than judicial redress in a first step, granting representative organisations legal standing is a key element to convince EU Courts to enforce the proposed targeted legal obligations in EU courts notwithstanding the choice of law and forum clauses included in the contracts between online platforms and their business users that designate non-EU courts. This approach builds on caselaw of the CJEU that explains that actions by representative bodies, which are in the collective interest (as opposed to merely in the interest of a group of specifically identified individuals) are not subject to the jurisdictional provisions included in any private contracts. Therefore, such

234 In this respect, the legal obligations of transparency regarding (i) changes to terms and conditions, and (ii) delisting are particularly relevant. While in the first place aimed to address the two harmful practices, these obligations would increase the ability of business users to challenge platforms' unilateral actions by providing them with the tools needed to verify this behaviour against the contract. The contractual framework itself will also be clearer, as the possible reasons for delisting will have to be spelled out upfront. These two obligations thus facilitate businesses' access to redress.

actions are more likely to be capable of being brought before the court of the Member State where the alleged (future) harm will occur. 235 This provision therefore also helps to address the fear of retaliation, as it will enable representative bodies to act in the collective interest of business users, who may prefer to remain anonymous, to ensure the effectiveness of the proposed legal obligations. While firmly based on existing case-law of the CJEU, the effectiveness of this specific provision on legal standing for representative bodies will benefit from any further horizontal action the Commission may take as part of its follow-up to the 2013 Recommendation on Collective Redress (2013/396/EU) and accompanying study on its implementation. 236 This study has recognised that collective redress mechanisms for consumer matters for injunctive relief are in place in all EU Member States, and that deficiencies remain only in respect of collective redress – the latter not being part of this proposal. It is also complementary to the pending revision of the Injunctions Directive, as part of which several options to protect the collective interests of consumers are being explored. Although providing standing to business organisations (in addition to consumer organisations) to strengthen this enforcement is one of the options considered, this will indeed not provide the required enforceability of the proposed P2B rules which aim to protect individual businesses rather than consumers. This option may nonetheless work in synergy with this proposed revision and the improvements that shall flow from it and lead to an increase in the number of business associations broadly monitoring law compliance in the field of online platforms, in particular on the issue of ranking transparency where consumer protection rules (the Unfair Commercial Practices Directive) complement the business-facing transparency measure included in the Preferred Option. In addition, under the preferred option 2c, associations or representative bodies that have a legitimate interest in representing businesses whose websites are indexed by online general search engines have the right to seek action in court to enforce – exclusively – the legal transparency obligation on ranking in online general search.

8.2 Scope of application of the Preferred Option

The Preferred Option would apply to platforms falling in the scope of this Impact Assessment, as specified in Section 1.3, including online search engines. A detailed explanation of the legal definition is provided in Annex 1.12 (cf. also annex 1.4).

The geographical scope of the Preferred Option would be based on the contractual relations of online platforms with EU business users. Online platform companies would be covered regardless of whether they are established in the EU, as long as they intermediate between EU business users and consumers located in the EU. Similarly, online search engines would be covered regardless of whether they are established in the EU, as long as they allow users located in the EU to perform searches of websites of business users established in the EU. This would prevent online platforms from excluding relations that are capable of producing effects in the EU from the scope of the intervention by simply diverting this to a .com domain name operated from outside the EU. 237 Evidence suggests, however, that virtually all important non-EU platforms have an EU establishment 238 , which reduces the risk of avoiding application significantly. Furthermore, the proposed provision on legal standing granting associations or representative

235 Henkel ( Case C-167/00 of 1 October 2002) as well as Amazon (C-191/15 of 28 July 2016). The same principles in relation to domicile of the defendant referred to in footnote 240 would also apply. 236 Report from the Commission on the implementation of the Commission Recommendation of 11 June 2013 on common principles for injunctive and compensatory collective redress mechanisms in the Member States concerning violations of rights granted under Union Law (2013/396/EU), forthcoming, 237 The approach towards the geographical scope in this initiative takes inspiration from the Court of Justice's interpretation of the geographic scope of the EU competition rules under public international law in the recent Intel judgement (Case C‑ 413/14, Intel v. European Commission, ECLI:EU:C:2017:632, paras. 40 et seq.). 238 Cf. Ernst & Young, study (forthcoming).

bodies having a legitimate interest in representing business users the right to seek action in court to remedy or prevent economic harm arising as a result of general non-compliance with the proposed legal obligations will further reduce this risk.

8.3 Thresholds

On the basis of the analysis in Section 6 platforms qualifying as a small enterprise would be exempted from the internal redress mechanism. While available data does not allow drawing a clear line as to whether this threshold should apply to micro- or small enterprises, the proportionality principle would advocate for setting the internal mechanism-specific threshold at a higher level. This would prevent any disproportionate regulatory burden to be imposed on start-up and scale-up platforms. Either the turnover or the employee count of a platform can be considered to be acceptable metrics to determine the threshold. Indeed the Recommendation on the Definition of SMEs combines both factors in its definitions as a cumulative requirement. Turnover is an indicator of the number of transactions intermediated by a given platform, and once a platform exceeds a turnover of EUR 10 million (threshold above which an enterprise is considered as medium-sized) both the degree of intermediation can be assumed to be significant, while such platforms can then also be assumed to be able to absorb the additional cost of the measure. At the same time, employee count is frequently easier to measure, and platforms may already exercise relative bargaining power before generating significant turnover, as a general characteristic of the platform business model. As to the possibility to set an additional threshold exempting enterprises from the entire regulation, the analysis of pros and cons does not allow concluding on the need to add a horizontal exemption to the internal mechanism-specific one (cf. Section 6.1.2.3).

Given the fast moving nature of the platform environment, any threshold which would be set, may need to be reviewed to ensure that it continues capturing platforms displaying specificities underlying the problem identified. A review clause should therefore allow for a revision of the threshold if needed. The work carried out under the Observatory would allow monitoring the efficiency of the proposed threshold and adjusting it as appropriate.

8.4 Stakeholders' views 239

Business users are generally supportive of the intervention proposed under the Preferred Option and generally in favour of a stronger and co-regulatory intervention 240 . The main requests by business users focus on effective redress options, greater transparency of platforms' ranking practices, the prohibition of MFN clauses, and transparency in delisting processes. Some business users also ask for an access to data obligation requiring platforms to share certain data, e.g. customer names, addresses, telephone numbers, email addresses and provenance 241 . In business users' view, more transparent ranking criteria would give the possibility to business users to make informed and substantiated complaints towards the platforms. Platforms should instruct business users of the reasons having led to their de-ranking or the suspension. Business users stressed the need for legislation to set up a contact point or a support function to deal with errors in ranking algorithms.

Online platforms - most would agree that providing an explanation to a business user in case of delisting or take-down of an offer seems to be a reasonable legal obligation - provided their legal obligations to take down illegal content and cooperate with investigations are respected. Online

239 See also Annex 4.

240 As tested in a focus group with stakeholders.

241 Subject to prior consent of the consumer for the transfer of personal data to the business-user, and for the business-user to agree that they will use it purely for information purposes (i.e. not to circumvent the platform)

platforms do not see the value added of external dispute resolution, because they trust their own internal dispute resolution systems. While platforms do not generally see a problem with implementing notice periods for changes in terms and conditions, they are, however, not in favour of rigid notice periods.

Regarding transparency around rankings and data use, platforms would agree with high-level disclosure as in the Preferred Option, but warn of "gaming" and manipulation of algorithms subject to too much transparency. Generally platforms appear supportive of the idea of monitoring the platform economy, if such monitoring is not intrusive in their trade secret policies.

Online general search engines already provide substantial guidance on how to optimise the ranking, but warn of the ineffectiveness of disclosing algorithms, not least in light of many and frequent changes to search algorithms, as well as on the risks to manipulation of search results.

Based on a questionnaire and a meeting with national authorities 242 , many national experts are of the view that addressing issues around terms and conditions is core to the entire P2B issue. They also consider that the proportionality of a transparency obligation would depend on the precise wording and on the size of the platform. On issues such as notice periods views diverge depending on the experience at national level ranging from no need to regulate to mandatory fixed notice terms. Some national experts are also of the view that terms and conditions should be simplified in order to make them transparent and user-friendly for businesses – similarly to B2C legislation. National experts supporting a legal transparency obligation find legitimate the requirement on platforms to provide a statement of reasons for delisting. National experts share the general view that delisting-related requirements should be aligned with illegal content/notice and action procedures. A transparency obligation on ranking criteria is overall considered proportionate and legitimate. The experts with more experience on ranking issues are supportive of measures solving the problems encountered in a timely manner, although some say that the issue should be left to commercial and competition law. A general preference to (i) opt for a transparency obligation covering ranking practices in general, and (ii) work towards identifying best practices in ranking captures the broad consensus view. On data, non-discrimination and MFNs experts cautioned that further reflection was needed. However, there is an overall agreement among experts on the importance of effective redress. Some concerns exist that internal complaints mechanisms could be more burdensome for SMEs. Some national experts are in favour of promoting existing best practices (possibly as part of a self-regulatory measure). Experts representing national authorities overall recognise the interest of the monitoring exercise. They are, however, generally opposed to the creation of a new body or European Agency created for that purpose.

Although the Preferred Option was not tested with consumer organisations, consumers are expected to be supportive of the Preferred Option despite the high-quality products/services they are currently benefitting from. Longer term competition- and choice-related considerations have been put forward by one consumer association in one of the Commission's workshops. A representative of this association has in particular argued in favour of some stricter nondiscrimination measures more in line with the telecommunications regulatory framework.

242 Views based on a questionnaire and a meeting with national authorities - Expert group on electronic commerce, established by Commission decision of 24 October 2005, OJ L282 (26/10/2005) p 20-21

9 M ONITORING AND E VALUATION OF I MPACTS

Monitoring of the evolution of impacts constitutes a key part of the intervention in this domain, as the online platform economy remains a dynamic, fast evolving area of the economy. The monitoring is therefore divided into two strategic parts, the EU Observatory of the Online Platform Economy, and the specific monitoring of the evolution of impacts related to the regulatory and self-regulatory components of the intervention.

9.1 EU Observatory of the Online Platform Economy

The European Commission will monitor market indicators for the online platforms environment through the EU Observatory, consisting of a group of independent external experts, supported by a dedicated Commission secretariat and by a study which will provide a website, as well as evidence and data gathering capacity.

This analysis includes data such as, but not limited to: number and types of businesses trading on online platforms, number and type of complaints handled through internal and mediators, number of cases successfully solved, the amount of time needed to resolve the case, the place of establishment, size of online platforms trading in the EU including turnover realised in the EU market as an online intermediary. Through the EU Observatory, the Commission will also monitor emerging challenges and opportunities for the EU in the wider digital platforms economy and online general search. This implies data and evidence gathering on matters such as access to data flows and their monetization opportunities controlled by platforms; transparency and accountability in the wider online advertising ecosystem; alleged discriminatory practices of platforms competing with their users; use and effect of MFN clauses including the justifications put forward by platforms or algorithmic decision-making in online platforms. In the context of search engines, data gathering and analysis will also cover issues such as: conditions for inclusion in and display of search results. This will include the use of third party content, including issues on access to data and the monetisation of the original content that this may raise and the information given to users when such content is displayed. It will also gather impacts from regulatory trends in the Member States or where relevant in third countries and on this basis prepare a set of evidence-based analytic papers to inform EU policy making. To this end, the Observatory will conduct data and evidence collection, collect opinions from a broader stakeholder base and experts and interact with and consider results of relevant research and studies, including European Commission funded studies and projects. Given the diverse nature of issues emerging in the Online Platform Economy, the Observatory would also liaise with relevant expert bodies at EU and national level to ensure holistic, multi-disciplinary outputs to inform EU policy-making.

The Commission will also analyse how to enhance the monitoring of the online platform economy through a dedicated study supporting the work of the EU Observatory, additional data sources and data collection, as well as collaboration with national statistical offices and Eurostat.

9.2 Specific indicators and operational objectives

Table 5 below summarises the specific objectives as well as a series of quantitative and qualitative indicators to be used for monitoring the effectiveness and efficiency of the Preferred Option. Importantly, the procedural monitoring will be accompanied by a monitoring of emerging practices on both the platforms' and the business users' side. Consequently, the impact of the Initiative will be assessed in the context of an evaluation exercise and activate, if so required, a review clause 3 years after entry into force of the adopted instrument.

Table 5 : Indicators of impact for the purpose of monitoring and evaluation


2.

Behandeld document

27 apr
'18
Proposal for a REGULATION OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL on promoting fairness and transparency for business users of online intermediation services
PROPOSAL
Secretary-General of the European Commission
8413/18
 
 
 

3.

Meer informatie

26 apr
'18
COM(2018)238 - Bevordering van billijkheid en transparantie voor zakelijke gebruikers van onlinetussenhandelsdiensten
25 mei
'16
COM(2016)288 - Online platforms en de digitale eengemaakte markt: kansen en uitdagingen voor Europa

Terug naar boven