Search interesting materials

Thursday, February 22, 2018

CCI's order against Google: infant steps or a coming-of-age moment?

by Smriti Parsheera.

The Competition Commission of India (CCI) recently concluded its six year long investigation into allegations of abuse of dominance by Google in India. It found that Google had utilised its dominance in general web search services to limit user choice (specifically in the context of Google flight search) and impose restrictions on its search syndication partners. It also noted a third violation relating to Google's historic practice of setting fixed positions for a particular category of search results (referred to as "universal results") that were sourced from its other verticals like images, videos and maps. A few months ago, the European Commission had also levied a penalty of Euro 2.42 billion on Google for offering preferential treatment to its comparison shopping service and demoting rival services in its search results.

The consequences for Google in India? CCI has directed Google to (i) desist from assigning fixed positions to universal results; (ii) add a disclaimer while presenting its commercial flight results; (iii) not enforce unreasonable restrictions on syndication partners; and (iv) pay a penalty of Rs. 1.36 billion (USD 21.1 million) for its anti-competitive conduct. To put this in perspective, the penalty translates to less than 0.02% of USD 110.9 billion, Google's worldwide revenues in 2017.

Are these the uncertain first steps of an infant CCI venturing into technology-ville or the coming-of-age moment of a regulator that has learnt to balance innovation and competition in the digital era? This post attempts to answer this question by tracing CCI's analysis on issues of universal results, Google flights and restrictive agreements, highlighting some concerns and summarising the takeaways from this decision.

Allegations against Google

CCI's order arises from two separate cases filed by online matchmaking portal Matrimony.com and Consumer Unity & Trust Society against Google Inc. and Google India Private Limited. In the course of the investigation, Google Ireland Limited, a key contracting entity for Google's advertising agreements, was also added as a party.

The informants alleged an abuse of dominance by Google in the general web search and search advertising markets in India, with the following specific claims.

  • Search bias: Google was using its dominance in the search engine business to promote its own results like videos (YouTube), news (Google News) and maps (Google Maps) in its search results.
  • Unfair advertising terms: Google has the largest number of search users which makes it an unavoidable partner for all advertisers who want to target their ads at those users. The informants alleged that Google was using this position to impose unfair and discriminatory conditions on its AdWords customers (advertisers who bid on keywords for ads to be displayed in Google's search results).
  • Denial of access: Google was using its dominance in the search and search advertising markets to impose unfair conditions that restricted its partners from contracting with other competing search engines. As a result, it was denying access to the market to competing businesses.

As per the scheme of the Competition Act, 2002 (Act), the complaints were first examined by the CCI to assess their prima facie merit. Finding a prima facie case of abuse of dominance by Google, CCI referred the matter to its investigative arm, the office of Director General (DG), for detailed investigation. The DG's office took about three years to complete its assessment and submitted an investigation report to CCI in March, 2015. It found Google to be guilty on all the counts raised by the informants in addition to a few others that were discovered by it in the investigation process.

CCI's reasoning and verdict

An abuse of dominance case under Section 4 of the Act requires CCI to establish that the entity in question held a dominant position in a defined relevant market and had abused that dominance through activities like imposing unfair or discriminatory conditions on others, limiting the supply of goods or services in the market, or using its dominance in one market to protect its position in another.

Relevant market analysis

CCI agreed with the findings of the DG that Google was operating in the following relevant markets: (a) market for online general web search services in India and (b) market for online search advertising services in India.

Google contested both these market definitions, arguing instead for a broader relevant market -- the wider the market the lesser the chances of Google being found to be dominant in it. Regarding general web search, Google claimed that it competes with all possible sources of information that can answer a user's specific query (about people, places, recipes, etc.) and there is no separate market for "general web search". This however ignores the fact that there are billions of webpages on the Internet, a majority of which are not known to users. Users therefore commonly rely on search engines to identify new sources of information and even access relatively well known ones. For instance, Alexa's web traffic analytics shows that 66.6 percent of the traffic to Wikipedia, the fifth most popular website in the world, comes through search engines.

Curiously, Google is also reported to have argued that "[b]ecause search is free, Google has no trading relationship with the users of its search service, and so the basis for establishing dominance is absent". CCI rejected this argument, noting that "it is not only flawed but altogether ignores the role of big data in the digital economy". In the multi-sided platform operated by Google, users offer their "eyeballs" and data in exchange for Google's "free" services, which are in turn monetised by Google through advertising revenues. To claim that only services that are directly paid for by users can constitute a relevant market would render large parts of the digital ecosystem outside the purview of competition laws, an outcome that is neither legally tenable nor socially desirable.

Assessing Google's dominance

The determination of dominant position depends on a number of factors, market share being one of them. CCI's order notes that Google has maintained a high market share in both the relevant markets but does not disclose the exact figures (these are blacked out as confidential information). Publicly available information from statcounter, however, clearly shows that in the period since 2010, Google has consistently held over 96 percent of the market share among search engines in India.

CCI also looks at other factors beyond market share. On the search side, it refers to Google's head start in crawling and indexing the web and resulting scale advantages. As explained by Matt Turck, Google benefits from significant "data network effects" -- "the more people search, the more data they provide, enabling Google to constantly refine and improve its core performance, as well as personalize the user experience". As we discuss in this paper on Competition Issues in India's Online Economy, multi-sided platforms like Google are also characterised by strong indirect network effects. CCI uses a similar logic to observe that Google's stronghold in general web search supplements its dominant position in the market for online search advertising, resulting in situation were advertisers are left with little countervailing powers over Google. In summary, CCI notes that this leads to a situation where "[t]he structure of the market is both indicative of and conducive to Google's dominance".

Abuse of dominance

Next, CCI turned to examine the specific allegations relating to the abuse of its dominant position by Google. This is also the point where the Commission significantly digresses from the findings made by its investigation unit. Unlike the DG's report, which found Google to be in violation of the Act on almost all the grounds examined by it, CCI limits its findings to the following three grounds.

  1. Fixed positions for universal results:

    Findings: Google's search results pages often contain certain "universal results" that are sourced from its other search verticals (See Figure 1 for an example). As per Google, these universal results compete with other generic blue links for the most favourable position on the results page based on their relevance. However, it also admitted that when the service was initially introduced, the display of universal results was limited to certain fixed (1st, 4th or 10th) positions as its systems were not advanced enough to determine the relevant position for such results. CCI dismisses this argument to hold that Google's historic practice of adopting a fixed position for such results was unfair and misleading to its customers who were led to believe that the responses were being ranked solely on the basis of their relevance.

    Consequences: Since Google has already discontinued this practice, CCI limits itself to issuing a desist order directing Google not to resort to such position fixing in the future.

  2. Figure 1: Universal image results in response to search term "Kullu"

  3. Commercial unit for flight results

    Findings: Google also places various "commercial units" in its search results. This refers to a demarcated ad space for displaying sponsored results relating to shopping, hotels and flights. CCI focused in particular on Google's flight unit (See Figure 2 for an example). It noted that Google provides a prominent placement to its flights unit in general search results with a link that takes the user to Google's own specialised flight search service. It found that this practice results in either pushing down or pushing out other competing vertical search services with the result of misleading users and denying them the opportunity to access the other websites.

    Consequences: CCI directed Google to display a disclaimer in the commercial flight unit box indicating clearly that clicking on the relevant link would lead to Google's flights page and not the results of any other third party service provider.

  4. Figure 2: Google Flight Unit in response to search term "Delhi to Kullu flights"

  5. Restrictions in syndication agreements

    Findings: In addition to the search and advertising services offered on Google's own website, it also enters into syndication agreements with other websites to offer its search and advertising services to them. These agreements can either the take the form of standard online contacts or directly negotiated agreements. CCI found that Google imposes certain unreasonable restrictions on its negotiated search intermediation partners -- websites that enter into negotiated agreements to use Google's services on their websites are restricted from implementing any search technologies that are "same or substantially similar" to those of Google. CCI found that this restricts Google's partners from using the services of competing search engines. It thus "creates conditions for extending and preserving Google's dominance in search intermediation".

    Consequences: CCI directed Google not to enforce the restrictive clauses in its negotiated direct search intermediation agreements with Indian partners.

As noted above, the DG's investigation had found Google to be guilty on many other counts. This included questions about Google's conduct in relation to its AdWords customers; its practice of allowing bidding on trademarks owned by competitors; and its arrangements with distributors like Apple and Mozilla to make Google the default search engine in their products. CCI, however, disagreed with the DG's findings on all these other counts.

Some questions and concerns

CCI's decision raises many important issues regarding Google's conduct in the search and search advertising markets. The design of ranking algorithms to provide preferential positions to certain types of results (without sufficient disclosures) and the imposition of unfair restrictions in commercial contracts are certainly critical issues that can have far reaching implications for online competition in India. Yet, despite agreeing with the principles behind these ends, it is hard to ignore some issues with the means adopted by CCI to reach them. This section focuses on the lack of sufficient evidence-based analysis, selective focus on flight search and CCI's own uncertainty about calculation of the penalty, all of which are factors that could expose the order to subsequent scrutiny.

Absence of robust data and evidence

The first issue, which has also been emphasised at length by the two dissenting members of CCI, relates to the absence of robust data and evidence to support the findings against Google. While the discussions in the order are sufficient to develop a strong intuition about Google's anti-competitive conduct, this intuition should ideally have been followed through with supporting data to build a water tight case. Specifically in the context of flight search, the dissenting members point to the absence of actual data about the traffic flows to the Google flight unit or to competing websites, the positions on what these websites actually appear on Google's results page and the impact that it has on consumer behaviour.

We can contrast this with the European Commission's approach in its similar case against Google. In a press release issued in June, 2017, the Commission announced that its order against Google was supported by evidence from various sources, including "(i) significant quantities of real-world data including 5.2 Terabytes of actual search results from Google (around 1.7 billion search queries); and (ii) experiments and surveys, analysing in particular the impact of visibility in search results on consumer behaviour and click-through rates". Based on this evidence, it was able to gauge the precise effects of Google's prominent placement of its comparison shopping service.

  • The traffic to Google's comparison shopping service increased 45-fold in the United Kingdom, 35-fold in Germany, 19-fold in France, 29-fold in the Netherlands, 17-fold in Spain and 14-fold in Italy.
  • There was a sudden drop of traffic to certain rival websites, to the tune of 85% in the United Kingdom, up to 92% in Germany and 80% in France.

In the present case, it is clear that the dissenting members are not disagreeing with the merits of the case against Google but the absence of sufficient data to make those claims. That being the case, the logical course for the Commission would have been to direct further investigation by the DG on these specific grounds or pursue an inquiry on its own. Both these courses were open to the Commission under Section 26(7) of the Act but their adoption would of course have meant a further delay in an already long pending decision.

Questions about the flight search analysis

The next set of issues revolves around CCI's focus on Google flight search and its prominent display on the search results page as a ground for abuse of dominance. The order examines the impact of the prominent real estate given to Google's flight unit vis-a-vis third-party travel sites like MakeMyTrip.com or Yatra.com and its misleading impact on users. However, as noted here, Google's flight service (at least at present) only offers users the option of comparing flight prices and not of directly making the bookings through Google. Therefore, the market in question is that of flight fare comparison websites and it would accordingly have been relevant to consider the impact on competing fare aggregation sites like "skyscanner" and "farecompare" instead of only those that provide flight booking services. Such an analysis would have also enabled CCI to examine a potential violation of Section 4(2)(e) of the Act, which relates to the use of dominant position in one market to protect its position in another.

Further, the order does not clarify as to why the flight search functionality is more problematic than similar commercial units displayed by Google in response to shopping or hotel related searches in India. In case of shopping results, the Commission makes a passing remark that "Google's display of Shopping Unit may not per se affect the ranking of free search results". In case of hotels, the order states that Google does not offer this feature in India even though a search for hotels on Google's India site does display a commercial unit, which has similar features to its flight search function -- it leads to another Google page that contains advertisements from hotels and hotel booking websites.

Penalty imposed by CCI

Relying on the Supreme Court's decision in the Excel Crop Care case, CCI decided to limit its penalty to Google's "relevant turnover" from India. For this purpose CCI sought information from Google regarding its revenues from different segments of its India operations, which the Commission notes was provided in an unsatisfactory manner. For instance, it is unclear from the order whether the information supplied by Google under the head "Relevant Turnover from Direct Sales in India" was based on (i) the income earned by all Google entities from end-users based in India; (ii) the income earned by Google India Private Limited from advertisers in India; or also (iii) the income earned by Google Ireland, Singapore and others from Indian advertisers.

Unfortunately, CCI acknowledges these infirmities and still goes on to determine the final penalty amount based on the unclear information furnished by Google. Given the criticality of this point, it would have been appropriate for CCI to seek more specific information from Google and, if required, provide further time for the same. It could also have used its statutory powers to elicit this information.

Conclusion

The Google case is an important development in India's competition jurisprudence on abuse of dominance in multi-sided technology markets. CCI's order acknowledges at the outset that "intervention in technology markets has to be carefully crafted lest it stifles innovation". It also highlights CCI's intent to refrain from interfering in specific product design elements unless the conduct in question is particularly egregious and an intervention becomes necessary to correct certain distortions.

Despite its well intentioned attempts to balance the interests of innovation, competition and consumer welfare, the decision falls short on some counts. Firstly, in an industry centered around click-through-rates, analytics and ranking measurements, the order primarily relies on qualitative and descriptive accounts to establish Google's violations. Secondly, CCI chooses to intervene in certain product design elements (universal results, commercial units) but not in others (like the AdWords ranking mechanism and trademarks bidding policy). While doing so, it fails to offer any broader guidance on the basis for demarcating general product design elements (that could also negatively impact competition) from particularly egregious conduct that merits competition intervention.

Yet, irrespective of the fate of this particular decision, actions such as these serve an important function in taming the conduct of big tech -- the threat of external regulation creates an impetus for better "self" regulation. In the past, Google amended its AdWords terms to make it easier for advertisers to simultaneously manage advertising campaigns on competing ad platforms. This was done through voluntary commitments offered by Google in relation to an inquiry by the Federal Trade Commission. Similarly, a press release by the European Commission notes that in the context of its anti-trust proceedings, Google had modified its direct AdSense contracts (with websites who use Google's services to display ads on their pages) to give its partners more freedom to display competing search ads. Recent moves by Facebook and others to control fake news on their platforms are also grounded in similar concerns.

Finally, the case lays important ground work for subsequent cases against Google and other dominant players in India's online ecosystem. As CCI's orders get challenged before the appellate tribunal and eventually the Supreme Court, we will see new jurisprudence around issues of competition in the digital economy. This will hopefully create a feedback loop for increased rigour and evidence-based analysis in future cases in this sector.

 

Smriti Parsheera is a technology policy researcher at the National Institute of Public Finance & Policy. She has previously worked as a researcher with the Competition Commission of India, including on the Google case. The views are personal.

No comments:

Post a Comment

Please note: Comments are moderated. Only civilised conversation is permitted on this blog. Criticism is perfectly okay; uncivilised language is not. We delete any comment which is spam, has personal attacks against anyone, or uses foul language. We delete any comment which does not contribute to the intellectual discussion about the blog article in question.

LaTeX mathematics works. This means that if you want to say $10 you have to say \$10.