Search Neutrality

Contrary to some reports, the Commission’s June 2017 Prohibition Decision in the Google Search case did not leave it to Google to propose a remedy.  Instead, the Commission laid out a detailed set of requirements[1] for a remedy based on the principle of non-discriminatory, equal treatment for all services: Google’s own and those of its competitors.  The particular implementation of this principle-based remedy was, however, left up to Google.

Towards the end of last month, Google presented the European Commission with an outline of how it is proposing to comply with the Commission’s stated remedy requirements.  And we understand that, since then, Google has started to meet privately with various comparison shopping services throughout Europe (presumably under NDA) to brief them on its auction-based proposal.

That’s right.  Auction-based.  Like Google’s three previous remedy proposals under Commissioner Almunia—all of which were resoundingly rejected precisely because they were auction-based (see our deconstructions of Google’s first, second, and third commitment proposals to see why).

In our March 2017 Remedy Paper, we set out the various options available to Google for implementing a solution that would be both effective and compliant with the Commission’s stated requirements for an equal-treatment remedy.  It is no coincidence that none of these options involved a paid auction.  We will reserve final judgement until we have seen the details of Google’s proposal, but it is difficult to imagine how Google could devise an auction mechanism that would not inflict significant additional consumer harm, both by further restricting competition and by aiding and abetting Google in its long-term goal to substitute unprofitable, relevance-based natural search results with highly-profitable, pay-for-placement[2] advertisements.  It is also difficult to imagine how Google could devise an auction that Google’s own comparison shopping service could meaningfully participate in (without full-blown structural separation), or that would not result in Google’s competitors being compelled to bid away the majority of their profits to Google.  In other words, it is difficult to imagine how Google could devise an auction-based remedy that does not fall far short of the Commission’s stipulated requirements for an equal treatment remedy.

Unless Google is volunteering to break up its general- and specialised-search businesses, the inclusion of Google’s comparison shopping competitors into a new or existing pay-for-placement auction would simply create an additional anti-competitive barrier—one that would formalise the transformation of free, relevance-based traffic into paid, pay-for-placement traffic for all services but Google’s own.  This would only serve to escalate the inability of these services to compete against Google’s services in the face of Google’s immensely powerful search manipulation practices, particularly as, in an anti-competitive double-whammy, the company these competitors would be forced to hand their profits to would also be Google.

It is also difficult to imagine how Google hopes to persuade its users to accept what would be such a clearly retrograde step in the evolution of Google’s search service.  We suggest that the primary reason that users have so far tolerated Google’s incremental shifts away from relevance-based search results towards pay-for-placement advertisements is not because they like or approve of these changes, but because they have so far been largely unaware of them (see our December 2016 Paper[3] for some of the history, context, and consumer harm resulting from Google’s progressive blurring of the lines between relevance-based search results and pay-for-placement advertisements). Of course, the same is true of Google’s illegal search manipulation practices; in our experience, Google’s users are still largely unaware of these immensely harmful anti-competitive practices.

In the months following the Commission’s April 2015 Statement of Objections, Google began to develop a new line of argument—one based on the pretence that the Commission’s formal anti-trust charges are about ads rather than search results and on the false notion that these two things are interchangeable.  It now appears that Google is doubling down on this strategy.  Indeed, through Google’s own writings and those of its extensive network of influencers[4], Google has tried to create the false impression that an ad-based remedy is the only option—even going so far as to wrongly suggest that the Commission is specifically demanding such an ad-based remedy.  For example, on 4 September 2017, Google’s outside counsel for the EC’s investigation, Maurits Dolmans, presented the Commission’s emphatically non-prescriptive, search-result-based remedy as though it was explicitly and exclusively about ads (emphasis added)[5]:

“[The Commission] say Google should treat its rivals the same way it treats its own ad business.  They want…the rivals’ ads to show up in Google’s own product ads, and then the rival would get paid.  In other words, some kind of non-discrimination rule.”

Below, we have made the adjustments that would be necessary to render Mr Dolmans’ statement an accurate representation of the Commission’s requirements[6] (changes highlighted in bold):

[The Commission] say Google should treat rival comparison shopping services the same way it treats its own comparison shopping service.  They want…the rivals’ comparison shopping services to show up in Google’s search result pages in exactly the same way as Google’s own comparison shopping service, and subject to exactly the same processes and methods determining the positioning and display, including the same relevance standards, ranking algorithms, penalty mechanisms and their respective conditions, parameters and signals.  In other words, some kind of non-discrimination rule.

In addition to the many manifest shortcomings of any ad/auction-based remedy, it is important to bear in mind that Google only introduced the pay-for-placement business model underpinning all of its ad-based assumptions and arguments in February 2013—at least 7 years after the introduction of Google’s anti-competitive practices, 3 years after the start of the EC’s investigation, and 11 months after the commencement of “settlement” negotiations with Commissioner Almunia.

An important part of Google’s remedy strategy seems to be to divert attention away from the obvious, straightforward, and infinitely more effective alternative.  Rather than preserving Google’s illegal anti-competitive practices and devising a way of unnaturally “shoe-horning” competitors’ listings into them, the more obvious solution is to simply end them.  Google could cease and desist its anti-competitive Universal Search and penalty practices[7] and instead entrust the selection and ranking of all appropriate specialised services to Google’s core crawling, indexing, and ranking algorithms (albeit while borrowing some of the visual and other enhancements currently reserved for Google’s own products)[8].  As we have explained previously[9], this option would be relatively straightforward to implement and monitor, and it would immediately act to reinstate the comprehensive, relevance-based search results on which Google forged its formidable reputation and market position (and which many of Google’s users still assume they are getting when they use Google).

The incentives behind Google’s apparent attempts to secure an ineffective and non-compliant auction-based remedy are clear: any remedy that restores Google’s search results to a level-playing field for commercially-oriented search terms would inevitably result in a significant short- to medium-term decline in Google’s revenues and growth.  Comparison shopping is a lucrative business model, and competing on the merits is considerably more difficult than relying on anti-competitive search manipulation practices that quietly commandeer the lion’s share of traffic and revenues.

While we have yet to see details of Google’s proposal, it seems unlikely that Google could have devised an auction-based remedy that does not fall far short of the equal treatment standard stipulated by the Prohibition Decision.  It is therefore unlikely that the Commission would consider any such auction-based remedy to be compliant.

We are concerned, however, that in sectors such as comparison shopping—where Google’s anti-competitive practices are so entrenched and extreme that Google’s competitors have now been running on fumes for several years—there might be an understandable temptation to seize on any proposal that seems to offer a rapid lifeline to survival, rather than holding out for a truly effective remedy.  Presumably, the reason Google is privately briefing various comparison shopping services is because it badly needs a number of them to sign-up to the deal before Google’s end-of-September implementation deadline.  Perhaps Google is hoping that, if enough services can be persuaded to participate, this will help to create the impression of a viable solution.  As an added temptation for participants and a bonus for Google, if the number of initial participants is small enough, participating services might enjoy a brief period of artificially inflated profits, resulting from what would be an abnormally under-subscribed auction.

It is also possible that Google might seek to tempt the Commission into tacitly accepting an auction-based “compromise” of a remedy by volunteering to apply it more broadly—perhaps across all of the other verticals the Commission is still actively investigating, such as travel and local search.  Of course, it would be easy for Google to make such an offer if it was confident that the remedy would be ineffective, or even, as in this case, a substantial source of additional revenue for Google.

From what we know so far about Google’s auction-based proposal, it is hardly surprising that Google has not yet applied for interim measures[10].  It would be difficult for Google to argue that a “remedy” likely to net it substantial additional revenue would cause it grave and irreparable harm.