Gatekeepers’ Tollbooths for Market Access: How to Safeguard Unbiased Intermediation

Gatekeepers may seek to maximize revenues by replacing any unbiased, purely relevance-based intermediation of business users with a biased, payment-based intermediation. To this end, search engines, app stores or online marketplaces may exploit an entrenched intermediation power by making the outcome of their matching increasingly dependent on the amount of money that a business user is paying the platform for the intermediation, rather than on objective criteria to determine the relevance of offerings. Using the example of the systematic replacement of “organic” search results with paid results in Google Search, the article outlines under which circumstances a shift to a biased intermediation constitutes a market failure. The article concludes that an excessive substitution of unbiased intermediation by a gatekeeper may amount to an unjustified private “toll” to access markets that reduces competition and innovation. The article then outlines how regulation can ensure a sufficient level of purely relevance-based intermediation of gatekeepers.

By Thomas Höppner¹

 

I. INTRODUCTION

It is the central function of commercial online intermediaries to best match the offers of their business users with the demand of their end users, as expressed by the latter’s search query to the platform.²

Where faced with competition, intermediaries strive to enhance the quality and hence the attractiveness of their service by indexing, ranking and displaying (“matching”) all available business offers purely on the basis of their genuine relevance to the search query entered by an end user on the platform. However, the situation changes once the respective intermediation market has tipped in favor of a particular platform that thereby emerges as a gatekeeper to the single-homing users of its service. The firmer both user groups (end users and business users) are locked into one intermediary in lack of alternatives that generate similarly strong indirect positive network effects, the stronger the incentive and ability of such gatekeeper to successively exchange a purely relevance-based intermediation with a biased matching that best serves its own commercial interests.

Such biased intermediation can be found where the intermediary systematically presents its own content or services as the most relevant results or where it makes the display and ranking of business users dependent on their payments or the granting of other commercial advantages to the intermediary. As the most prominent example, over the last years “Google has pushed the organic links further and further down the results page and featured more search advertising results and Google’s own vertical or specialised search offerings.”³

This article outlines why such exchange of unbiased with biased intermediation is harmful and how specific regulation should therefore ensure that online intermediaries serving as gatekeepers provide their users with a sufficient level of unbiased, purely relevance-based search results.

 

II. BACKGROUND: SIGNIFICANT INTERMEDIATION POWER OF UNASSAILABLE GATEKEEPERS

A. Intermediation Power

The Internet is vast. For an end user to find the right website or software application (app) with the relevant content or commercial offering to satisfy its respective needs costs time and effort. To reduce such transaction costs, every end user is dependent on online intermediaries to find his or her way round the depths of the Internet. Conversely, every business wishing to offer any product or service on its website or through an app, depends on being found by end users on the intermediaries that they use (general or specialized search engines, app stores, marketplaces, etc.).

This dependency goes even deeper. Intermediaries have taught end users that the most relevant results appear at the top of a results interface or are particularly highlighted. Due to a “saliency bias”⁴ end users are more likely to focus on items or information that are more prominently presented and ignore those that are less so – even if they are in fact more relevant to their query. Their primary intention is to find (some) relevant content as quickly as possible. For this purpose, end users often blindly trust the algorithmic judgments of the intermediary they are accustomed to and click on pre-installed choices or the results at the top.⁵ The majority of users are not (and have no time to be) concerned about the precise mechanisms underlying the results suggested by their chosen intermediary. They click on the most prominent top results even if those are labelled as “advertisement” or any other bias is revealed in accompanying information on the platform. For example, “[v]arious studies over the last 16 years have shown that users are often unaware of or unable to distinguish Google’s paid ads from natural search results.”⁶ One study found that even in 2018, two centuries after the introduction of paid search, over 59 percent of end users were not aware of the difference between natural (“organic”) results and paid ads on Google, and only 34 percent of those who did recognize paid ads said they would deliberately avoid clicking on them.⁷

Conversely, this saliency bias means that every business user wishing to get access to an intermediary’s end users, is not just dependent on being found somewhere on that platform but on being found in those attractive formats on the results interface that attract most end user attention. If a business user is not to be found there, it will lose out vis-á-vis its direct rivals that do appear in those top positions. Hence, intermediaries can determine the winners and losers on the markets for the intermediated business users. This renders the matching by intermediaries the central battle grounds for competition online. In turn, the power to determine who wins these battles enables intermediaries to play-off business users against each other to extract maximum benefits from them.

B. Gatekeepers’ Significant Intermediation Power Over Single-homing End Users

From a policy perspective, the power of intermediaries to determine market outcomes is less of an issue where there is strong competition amongst intermediaries and end users engage with several platforms, i.e. multi-home. In general, the success of multi-sided intermediaries with inherently strong indirect network effects depends on the quality of the matching of their user groups. The better the matching, the more attractive the intermediary becomes for both end users and business users. Therefore, in a competitive environment, intermediaries have an economic interest to fully focus on providing the most relevant intermediation service possible by unbiasedly identifying and displaying the most relevant content for the respective end user needs at the lowest transaction costs for all user groups. Particularly if the intermediation service is provided for free, end users will turn to the platform with the expected highest intermediation quality, which in turn depends on the size of the platform’s content base and its matching abilities. If consumers find a platform’s intermediation quality to be inferior due to any biases, they may switch to alternative platforms (if such exist). Thus, in a competitive environment intermediaries’ commercial interests are generally in line with the interests of the intermediated user groups and competition concerns are limited.⁸

Matters change dramatically, however, once an intermediary gained significant intermediation power because at least one user group became dependent on it. The larger the number of users on each platform side, the less relevant the individual user becomes for the platform. Intermediaries require sufficient content from business users to cater for the various requests of end users (product offers, price information, images, reviews etc.). But once an intermediary has amassed a sufficient minimum portfolio of content, it typically no longer requires any particular content from any particular business user to satisfy the end user’s demand. That is because, most end users turning to intermediaries do not know which content (such as offers) is available and would best fit their respective demand. While they expect that an intermediary makes some relevant content available to them, they cannot assess whether the suggested list of results indeed contains the most relevant sources. Thus, where an intermediary has a sufficient portfolio of content, it does no longer depend on having a particular businesses user “on board.” There is no mutual dependency. This in turn drastically shifts the bargaining power to the advantage of the intermediary: The more single-homing end users an intermediary has bundled, the more business users have to make their content available to the platform to reach its unique user base but the less dependent the intermediary becomes on individual business users. Since it is in the full discretion of the intermediary whether and on which conditions it enables business users to reach the intermediary’s accumulated end user base, it is the intermediary that pulls all the strings.

As a result, with the growth of an intermediary, its bargaining power towards business users increases while the business users’ flexibility to leave the platform diminishes. They become increasingly dependent, up to the point that the platform is an unavoidable trading partner. At this point, the intermediary serves as a “gatekeeper” by controlling the access to and the business decisions of a unique user group (its single-homing end users). As a consequence, the more established an intermediary becomes and the less competition it is facing, the stronger its incentives and ability to use its intermediation power to maximize rent from its user groups, resulting in a stronger need for regulation.

 

III. GATEKEEPERS’ EXCHANGE OF UNBIASED WITH BIASED RESULTS AS EMANATION OF INTERMEDIATION POWER

A. Need for a Holistic View on Gatekeepers’ Significant Intermediation Power

There are several ways in which a platform acting as a gatekeeper may exploit its significant intermediation power.

The most prominent and broadly discussed option is “self-preferencing,” where a vertically integrated intermediary promotes its own downstream service vis-á-vis competing business users when matching their offers to the demand of end users.

A gatekeeper may also exploit its intermediation power by making the matching of a business user dependent on this user granting additional commercial advantages. For example, gatekeepers have made the ranking of business users dependent on them using ancillary services,⁹ refraining from granting better prices and conditions on other platforms,¹⁰ providing additional data to the platform¹¹ or granting free licenses for the use of their content¹² that were neither required nor proportionate for the actual intermediation service at stake but benefitted the gatekeeper elsewhere.

As an even more obvious option, gatekeepers may exploit an entrenched intermediation power by making the outcome of their matching increasingly dependent on the amount of money that the business user is paying the platform for the intermediation, rather than on objective relevance criteria. This is the case, where instead of selecting, ranking and displaying all business users on the basis of their genuine relevance, a gatekeeper starts to take into account how much a respective business user is willing to pay the intermediary for any matching or where the intermediary increases the relevance of such payment factor vis-á-vis other matching factors. For example, a general search engine may successively exchange “organic” search results with “paid” results. An app store may present apps more prominently dependent on their bids to be displayed or the revenue share they are willing to grant and an online marketplace may rank equivalent offers in the order of the commission that the merchant has agreed to pay for any concluded transaction.

All of the above cases have in common, that the gatekeeper replaces its purely relevance-based (unbiased) intermediation service for a (biased) intermediation service, considering factors that are unrelated to the actual relevance of the intermediated business offers and hence (should be) irrelevant for the crucial matching that the intermediation service provides.

Over the years, competition authorities¹³ and legislators¹⁴ have opined that a gatekeeper must not take its own affiliation with downstream businesses into consideration, i.e. that is shall not favor own services. There is also a growing understanding that gatekeepers must not make their matching depending on the granting of unrelated advantages such as the provision of data or free licenses for content.¹⁵ However, thus far, little attention has been paid to the third emanation of intermediation power: the direct exchange of a purely relevance-based intermediation service with a payment-based intermediation.

The case of Google may best exemplify the need for a more holistic regulatory approach.

B. The Example of Google Search

Launched in 1998, Google has become the world’s most popular general search engine. Its success stems from the high quality of its “organic” search results which are ranked on the basis of unbiased criteria for the relevance of a web page, in particular its “PageRank.”¹⁶ Google praised itself for neither showing own content nor accepting any payments to influence its search results.¹⁷

In 2000, Google launched AdWords, which allowed website operators to pay for keyword-based ads. First, only some ads appeared to the right of Google’s organic search results. However, since 2004, the year in which the market “tipped” towards Google, the search engine has significantly changed the design of its results pages in several ways, most notably by (1) increasing the number and the size of ads placed above organic search results, (2) blurring the distinction between how ads and organic listings are presented and (3) placing own content between ads at the top and organic results at the bottom.¹⁸

As a result of Google reducing the space (real estate) devoted to organic listings, in case of queries with a commercial intent, users now hardly find any organic search result on a standard results page and (are made to) click more on ads or Google’s own offerings, even though such results are less relevant to their query. The trend is especially pronounced on mobile devices, which make up more than half of all searches now. For queries that suggest a commercial intent, the first organic listing on mobile devices now often only appears on the bottom of the third search results screen.¹⁹ The rate of clicks on such results “fell by more than 30% between January 2016 and June 2019, while paid click-through rates over the same period of time more than tripled”.²⁰ That is despite an estimated average price increase for search ads of about 5 percent year-on-year in the U.S., and even more in countries with higher market shares.²¹

Despite all criticism, to further increase the share of clicks on paid search results, in late 2020 Google started testing larger gaps (white space) between search results²² as well as taller shopping ads²³ – in order to effectively push organic results further out of end users’ sight.

Even when end users clearly indicate to which website they wish to be navigated to by entering the trademark of particular businesses (e.g. “tripadvisor”), Google will only present the organic link to this site below any paid search results that third-party websites may purchase for such navigational keywords (e.g. ads for competing travel sites). As affected companies have reported, this effectively means that businesses are extorted “to pay for the privilege that consumers who search for our trademark brand name can find us because if we don’t they will sell our brand name as misdirection to our competitors.”²⁴

All of this contrasts with Google’s own policies. Google has a dedicated algorithm that looks at the layout of a website and demotes those that “load the top with ads” or “make it hard to find the actual original content on the page.”²⁵ According to a Google blog post of 2012, “[i]f you click on a website and the part of the website you see first either doesn’t have a lot of visible content above-the-fold or dedicates a large fraction of the site’s initial screen real estate to ads, that’s not a very good user experience. Such sites may not rank as highly going forward.”²⁶ Similarly, while generally business users can buy ads for trademarks of their competitors, this is not possible for Google brands.²⁷

 

IV. THE CASE FOR ENSURING A SUFFICIENT LEVEL OF UNBIASED INTERMEDIATION

The following arguments are calling for a regulatory limitation of such practices by digital gatekeepers.

A. Prisoners’ Dilemma and Barriers to Entry for Business Users

Exchanging a relevance-based intermediation with a biased (paid) intermediation effectively forces business users to pay the intermediary an ever-increasing price for the ability to reach end users even though these end users only turned to the intermediary to access the business users’ content.

Where payments to the intermediary influence the ranking, business users have no other option than to take part in the bidding and to increase their bids in accordance with any increased competition from other bidders in order to remain findable online. That becomes necessary even for those business users that are indeed offering the most relevant content for the respective queries of the end users. Instead of competing for users by offering high-quality services that should lead to better organic listings, these businesses must now compete for users based upon how much money they pay the gatekeeper.²⁸ “This raises their costs, reduces their competitiveness, and limits their incentive and ability to invest in innovations that could be attractive to users.”²⁹

An analysis of the economic literature on the welfare effects of sponsored ranking confirms this: “[B]oth the theoretical and the empirical literature have highlighted that the platform may be able to expropriate a significant part of the content providers’ gross surplus through a sponsored ranking. Generally, the more it pays to be prominent, the more likely the platform is able to expropriate the content providers’ surplus, and thus the more likely that it is ultimately the platform that gains the most. In reverse, content providers may end up playing a prisoners’ dilemma where no one can commit not to bid for a top position in the sponsored ranking, but ultimately all content providers are worse off by doing so.”³⁰

B. Lower Quality, Higher Prices and Less Innovation for End Users

End users turn to intermediaries to find the most relevant content as quickly as possible. As Google explained elsewhere, if “it’s difficult to find the actual content, [users] aren’t happy with the experience. Rather than scrolling down the page past a slew of ads, users want to see content right away.”³¹ Accordingly, exchanging unbiased intermediation results – the content that end users are using an intermediary for – with paid results (ads) reduces the user experience and degrades the quality of a service. It is apparent, for instance, that if a consumer searches for a particular business by entering its trademark brand and the intermediary instead shows paid results of rivals that bid on this trademark term, the consumer is not finding what it is looking for but is effectively misled.

In addition, (excessive) paid intermediation by gatekeepers increases overall prices for the products of business users. If platforms base their ranking (also) on the listing fees they can attain from content providers, ultimately those business users that can best monetize a click or download will be able to bid the highest price and therefore appear at the top. Those businesses, however, are not necessarily the ones that competition policy should promote because typically the ability to monetize a click stems from the highest profit margin due to the highest prices. Moreover, a sponsored ranking always favors the companies with the “deepest pockets,” which are typically incumbent dominant firms.

Economists therefore found that “[a]ll prices tend to be higher than in the absence of a sponsored ranking and consumers’ surplus is lower than in the case without sponsored ranking, if [the platform] is able to expropriate the benefits of prominence from the content providers.”³² Conversely, an organic ranking provides a business user with the highest incentives to invest in quality, precisely because it will compete for the top-ranking position only in terms of quality.³³

C. Higher Transaction Costs for all User Groups

Exchanging relevance-based intermediation with biased intermediation increases transaction costs for all user groups.

In general, intermediation services reduce decision complexity for end users by matching their expressed demand with the most relevant offerings. This reduces transaction costs and increase the average quality of a business user’s offerings that an end user accesses.³⁴

However, the more an intermediation process becomes influenced by biased, paid-for intermediation, the less relevant the results will be and the more time the end user will need to invest in strategies to avoid or even game the techniques used to manipulate the user through biased results. For instance, the end user will spend time hiding his or her personal data (for targeted ad results) or check price recommendations several times a day or via separate devices, in order to ensure that he or she is not falling victim to biased results.³⁵ This creates unnecessary expenditures and hence externalities.

D. Dominance-related Inability of User Groups to Escape Exploitation

Some have argued that excessive advertising or paid intermediation more generally may drive end users to switch to other platforms, which would solve the policy concerns.

However, while we can rely on such switching where genuine competition between intermediaries exists, no such self-healing can be relied upon where the market has irreversibly “tipped” and a gatekeeper emerged.

End users do not turn to intermediaries for advertising or paid results. They turn to the platform from which they expect the highest weighted content quality and price on average. One could assume that a rational end user that becomes aware of a high share of bias paid-for rankings start to consider alternatives. In particular if paid results are based upon an auction, a rational end user may be afraid that it is subjected to a “double marginalization,” where both, the business user and the intermediary seek to maximize revenues from the user’s attention. Accordingly, to prevent users from switching, in a competitive and transparent environment, every intermediary would have an incentive to ensure a sufficient degree of purely relevance-based intermediation.

However, to rely on a self-regulation of the market due to end users’ ability to switch the platform in case of bias intermediation is no longer justified once the market has tipped towards one intermediary. The more acquainted consumers got with a particular intermediary, the higher their status-quo bias to stay with this platform even if the quality degrades. Accordingly, if end users are faced with a poor quality of the intermediation results of their incumbent platform, they tend to invest into a more comprehensive analysis of the results available on that platform, e.g. by looking at more results or fine-tuning their query, before considering any switch to an alternative intermediary. For a switch, the expected performance of the alternative platform would have to be so much better in terms of relevance that it clearly outweighs the disadvantages of the incumbent’s biased results. This, however, can be a very high hurdle as the incumbent typically enjoys significant quality advantages from data-driven positive network effects. Moreover, the incumbent will be able to test at which precise share of biased intermediation end users typically commence switching. This allows the gatekeeper to fine-tune and adjust the balance between bias and unbiased intermediation to prevent any relevant switching.

E. Gatekeepers’ Degenerated “Advertising” Models as Unjustified Private Internet Tolls

In traditional media such as newspapers, television or radio, advertising serves a justifiable purpose. Ads make consumers aware of new products and businesses and may create a new demand. To potentially trigger a commercial interest of the consumer creates the value of the ads in such media.

The situation is quite different for ads that are used to form sponsored rankings of intermediary’s business users. In such cases, the ads are displayed in return for an already existing commercial interest and consumer demand: Where an end user turns to an intermediary and enters a commercially relevant keyword for a particular product or business that triggers paid results, such consumer is already determined to purchase such product or to visit the respective business. Hence, the ads in the sponsored ranking that the intermediary displays no longer contribute any added value in creating a new demand or in delivering the end user to the advertiser. The end user was already intending to turn to such “advertiser” and to buy the respective product before he or she even saw the sponsored result/ad. Hence, sponsored rankings do not fulfil the traditional purpose and justification of advertising to generate an additional interest or demand.

Instead, ads used to create sponsored rankings now effectively serve as a fee to access the intermediary’s end user base, a private “toll” for market access. The firm intention of the end user to purchase something is monetized, not by the supplier that the end user wishes to reach, but by the intermediary that is interposed between the end user and the supplier. Somewhat shockingly, the currently dominating auction model for paid results means that the more determined and the more likely a consumer is to purchase something when turning to the intermediary, the higher the “toll” that the intermediary will charge the suppliers that the consumer will ultimately turn to. That is because the stronger the commercial intent that the entered keyword suggests, the higher the average price that a business needs to bid to be considered as a top paid search result. In the most extreme case of paid results in return of navigational search queries for a particular business or one of its products, that business may have to outbid all its rivals (that bid on its trademark terms) to appear at the top, even though the end user was searching specifically and only for that business and the intermediary’s lists of paid search results does not provide a new “match” but only causes confusion.

By forcing businesses to increase bids in order to remain findable for end users, gatekeepers’ intermediation power allows them to extract a toll from virtually every business that hopes to participate in digital markets.³⁶ As seen in the example of navigational queries, this is the case even where the intermediary does not add any value in terms of matching offer and demand.

Such Internet toll from business users is not justified. First and foremost, considering that end users are interested in the offerings of businesses (rather than their ads or any intermediation), it is the online business users that have attracted consumers to the Internet and there to the intermediaries in the first place. If it was not for high quality offerings of business users, there would be no business case for commercial intermediaries. More generally, the largest share of the costs for setting up the infrastructure enabling high-speed Internet services was borne by states or telecoms providers. In Germany alone, during 2014 and 2023 with the support of the state, telecoms providers committed to invest € 100 billion into glass fiber networks to bring broadband services to rural areas. Considering the previous investments into the infrastructure, by now a high three-digit billion figure will have been invested in the Internet architecture in Germany alone. However, while the costs for enabling Internet access are primarily borne by taxpayers or telecoms firms, the toll for using the Internet is now imposed by digital gatekeepers using sponsored rankings to artificially block consumers off business users.

 

V. HOW A SUFFICIENT LEVEL OF UNBIASED INTERMEDIATION COULD BE ENSURED

A. Alternative Approaches

All of the above shows that an excessive exchange of unbiased with biased results is ultimately an issue of intermediation power as a sub-category of market power. The issue therefore needs to be addressed by asymmetric gatekeeper regulation rather than horizontally applied obligations for all intermediaries.

There is no need to outright prohibit any form of advertising or paid intermediation for gatekeepers. Neither is there a need to regulate their prices for such intermediation. In general, auction mechanisms are a competitive tool to determine an adequate price. The issue is not how the prices for paid results are determined. The concerns arise from the replacement of relevance-based intermediation with biased intermediation which undermines competition, misleads consumers and degrades the overall quality of central intermediation services.

To secure a sufficient level of unbiased intermediation by digital gatekeepers could be approached in several ways. One approach would be a universal-service type obligation to secure a sufficient level of purely relevance-based intermediation. Since the Internet can be seen as a service of general interest, i.e. a service that a consumer requires on a daily basis, the provision of Internet access is covered by the universal service obligation of telecoms providers.³⁷ Internet access is of limited use, however, if consumers cannot find their way around it. Gatekeepers therefore often praise themselves for securing the functionality of the Internet.³⁸ It could therefore be acknowledged that an unhindered access to intermediation services that are indispensable to find, visit and use relevant Internet services is also of general interest. Accordingly, central Internet intermediaries could be obliged to secure the sufficient access of end users to purely relevance-based intermediation services.

An alternative approach would be to reflect the limitations that exist(ed) for traditional media as regards the amount of advertisement that they may serve as compared to the content they present.

Overall, the following obligations could be imposed on intermediaries acting as gatekeepers:

B. Cap on Capacity for Biased Results

It could be regulated that the space for paid-for results or results from the gatekeeper’s own services on an average intermediation results interface shall not exceed a maximum cap of (for instance) 25 percent of the visible space in comparison to the remaining, unpaid (purely relevance-based) “organic” results (for example, generic search results on Google, third-party products on Amazon, third-party apps on Apple’s or Google Android’s app store).

Limiting maximum ad space would not necessarily reduce intermediaries’ ad revenues because the limitation of ad “capacity” would go hand-in-hand with increased auction prices for the remaining ads. On the upside, it will, however, leave more space for a purely relevance-based intermediation, for consumers to “meet” the most relevant business user.

C. No Favoring of Biased Results

It could be regulated that relevance-based (“organic”) results must not be displayed less favorably than sponsored/paid-for results on the regulated intermediation interfaces. This would ensure that a gatekeeper may not force businesses to buy ads by making them visually more attractive than purely relevance-based results.

D. No “De-commercialization” of Organic Results

It could be regulated that for commercial queries a gatekeeper shall not reduce the quality/relevance of (free) organic results vis-á-vis biased (paid) results by demoting commercial offerings in organic results in favor of non-commercial offerings (such as Wikipedia articles, product descriptions or news articles).

E. No Paid Results in Case of Navigational Search Queries for Trademarked Brands

Where an end user has clearly indicated his or her intent to find or access a particular business, a digital gatekeeper, in particular search engines, should be obliged to display (at least) a free link to this business above any paid results. Alternatively, the gatekeeper could be prohibited from allowing companies to bid on keywords that include the trademark terms of competitors with a view to displaying ads above the organic results for such competitors.

 

VI. CONCLUSIONS

The significant intermediation power of commercial gatekeeper platforms can be exploited in various ways. One emanation is to excessively exchange unbiased with biased intermediation with a view to maximizing profits from the matching of end users and business users. Where following the tipping of a market, the intermediated user groups have no option to escape any such manipulation by switching to alternative intermediaries, regulation needs to step in to ensure a sufficient level of unbiased intermediation.

---

1 Professor of Business Law at Technical University Wildau, partner at Hausfeld, Berlin. The author advises several clients on competition and regulatory matters relating to digital gatekeepers. The views expressed are those of the author.

2 Unless stated differently, the terminology in this article is based upon the definitions used in EU Commission’s Proposal for a regulation on contestable and fair markets in the digital sector (Digital Markets Act, (the “DMA”)). Contrary to its Article 2 (6), the term “intermediary” shall also include online search engines of any type.

3 USA et al. v. Google LLC, Complaint of October 20, 2020, District Court for the District of Columbia, Case 1:20-cv-03010, para. 170.

4 Cf. Pedro Bordalo, Nicola Gennaioli, & Andrei Shleifer “Salience theory of choice under risk” (2012) The Quarterly Journal of Economics 127(3), p. 1243-1285.

5 European Commission, Case AT.39740, 27/06/2017, Google Search (Shopping), Section 7.2.3.1.; Case AT.40099, 18/07/2018 – Google Android, Section 11.3.4.1. II.; Bing Pan et al. “In Google we trust: Users’ decisions on rank, position, and relevance” (2007) Journal of Computer-Mediated Communication, p. 12.

6 Richard Falconer “Google: Blurring the line between ‘natural’ and paid search results,” Sept. 19, 2019, https://weareyard.com/blog/google-blurring-the-line-between-natural-and-paid-search-results, with references to 8 studies.

7 Mark Jones “Two-thirds of people don’t know the difference between Google paid and organic search results, Marketing Tech News” Sept. 6, 2018, https://marketingtechnews.net/news/2018/sep/06/two-thirds-people-dont-know-difference-between-google-paid-and-organic-search-results/.

8 End users should still be protected against misleading or bias results. However, where genuine alternatives exist, transparency obligations to reveal any bias or conflicts of interest for intermediaries will suffice.

9 Cf. Italian Competition Authority, A 528, Amazon: investigation launched on possible abuse of a dominant position in online marketplaces and logistic services, Press release of Apr. 16, 2019, see also https://www.thelocal.it/20190417/amazon-italy-antitrust-investigation.

10 See Bloomberg “Hotel site accuses Booking, Expedia of EU antitrust breaches” Jun. 11, 2019 https://www.bloomberg.com/news/articles/2019-06-11/hotel-site-accuses-booking-expedia-of-breaking-eu-antitrust-law; in detail Daniel Mandrescu “The return of the MFN clauses – platform ranking as an enforcement mechanism for price parity” Jun. 29, 2019, https://www.lexxion.eu/coreblogpost/the-return-of-the-mfn-clauses-platform-ranking-as-an-enforcement-mechanism-for-price-parity/.

11 Cf. Natasha Lomas, “Travel startups cry foul over what Google’s doing with their data,” Aug. 14 2020, https://techcrunch.com/2020/08/14/travel-startups-cry-foul-over-what-googles-doing-with-their-data/.

12 French Competition Authority, Case 20-MC-01, Apr. 09 2020, SEPM and APIG v. Google, paras 234, 237; Almunia, SPEECH/13/1042, Dec. 19 2013, “Google creates a link between getting the right to use material from other sites on its specialised search services and the appearance that these sites have on Google’s general search results – a practice that allows Google to benefit from investments made by other firms. I have asked Google to sever this link to restore competitive incentives.”

13 European Commission, Google Search (Shopping) (n. 5); Competition Commission of India, Case Nos. 07 and 30 of 2012, 08/02/2018, Section 195.; Turkish Competition Authority, Case 20-10/119-69, Feb. 13 2020; USA et al. v. Google LLC, Complaint of October 20, 2020 (n. 3); Colorado et al. v. Google LLC, Complaint of December 17, 2020, District Court for the District of Columbia (to be consolidated with Case 1:20-cv-03010), para. 29, 168, 190 et seqq.

14 See Art. 6 para. 1 lit. (d) DMA-Proposal (n. 2); U.S. House of Representatives Subcommittee on Antitrust, Commercial and Administrative Law “Investigation of Competition in Digital Markets” (2020), Section VI.A.2.

15 See footnote 12 above; § 19a para. 2 lit. 7 of German Competition Act as of Jan 15, 2020.

16 Cf. Google Search Central Blog, “Evaluating page experience for a better web,” May 28, 2020, https://developers.google.com/search/blog/2020/05/evaluating-page-experience.

17 Cf. Larry Page, “Most portals show their own content above content elsewhere on the web. We feel that’s a conflict of interest, analogous to taking money for search results. Their search engine doesn’t necessarily provide the best results; it provides the portal’s results. Google conscientiously tries to stay away from that. We want to get you out of Google and to the right place as fast as possible. It’s a very different model.” Google Founders’ Interview with the Playboy Magazine (“The Google Guys”), Sept. 2004, https://www.sec.gov/Archives/edgar/data/1288776/000119312504139655/ds1a.htm#toc59330_25b.

18 US House of Representatives Subcommittee on Antitrust, Commercial and Administrative Law “Investigation of Competition in Digital Markets” (2020) p. 194-195.

19 Id. p. 197.

20 Id. p. 201.

21 Id. p. 196 (footnote 1186).

22 To this end, Google increased the front-size of headings of search results from 18 pixels in 2018, 20 pixels in 2019 to now 22 pixels.

23 Barry Schwartz, Google Search Testing Taller Shopping Ads, SearchEngineRoundetable, Dec. 10, 2020, https://www.seroundtable.com/google-taller-shopping-ads-30576.html.

24 Statement of David Hansson, Cofounder & Chief Tech. Officer, Basecamp in U.S. House of Representatives Subcommittee, (n. 18), p. 202.

25 Google Inside Search, Page layout algorithm improvement, Jan. 19, 2012, https://search.googleblog.com/2012/01/page-layout-algorithm-improvement.html.

26 Id.

27 Cf. Hansson, (n. 24).

28 U.S. House of Representatives Subcommittee, (n. 18), p. 201.

29 USA et al. v. Google LLC, Complaint of October 20, 2020, (n. 3), para. 170.

30 Jan Krämer & Daniel Schnurr “Is there a need for platform neutrality regulation in the EU?” (2018) Telecommunications Policy 42, p. 514, 525.

31 Google Inside Search, (n. 25).

32 Jan Krämer & Daniel Schnurr, (n. 30), summarising the findings of Mark Armstrong, Jidong Zhou “Paying from prominence” (2011) The Economic Journal, 121(556), F368-F395; Jidong Zhou “Ordered search in differentiated markets, International Journal of Industrial Organisation” (2011) 29(2), 243-362.

33 Compare Jan Krämer & Oliver Zierke “Paying for prominence: The effect of sponsored rankings on the incentives to invest in the quality of free content on dominant online platforms” (2020) https://ssrn.com/abstract=3584371 “Intuitively one would expect that an organic ranking provides content providers with the highest incentives to invest in quality. […] Indeed, this is what we denote as the competition effect, and we can show that it becomes stronger the more similar content providers are with respect to their ability to produce quality content.”

34 Id. p. 9.

35 Compare Ryan Calo “Digital Market Manipulation” (2014) 82 George Washington Law Review, p. 995, 1027.

36 Cf. Fiona Morton & David Dinielli “Roadmap for a Digital Advertising Monopolization Case Against Google” (2020), p. 30 regarding Google. The situation is similar for app stores and marketplaces.

37 Art. 4 para. 2, recital 8 of Directive 2002/22/EC of the European Parliament and of the Council of March 7, 2002 on universal service and users’ rights relating to electronic communications networks and services (Universal Service Directive).

38 Cf. German Federal Supreme Court, judgement of September 17, 2017, I-ZR 11/16 – Vorschaubilder III „In the interest of the information society, [search engines] ensure the functioning of the Internet. Without recourse to search services, a meaningful use of the unmanageable scope of information on the Internet would practically be precluded.“ (unofficial translated from German original).