Preliminary thoughts on Google’s proposed commitments
As long anticipated, here are some comments on the proposed commitments in the Google case (I graciously granted myself an extension, like the one other third parties have received; it actually is convenient because I can comment on others’ comments as well).
Four caveats are in order:
- The views expressed below are written against the background of the Commission’s concerns as set out in the press release and the Q&A doc. accompanying the market testing of Google’s proposal. The relevant question to keep in mind is whether the proposed commitments –in their current form- are apt to address the concerns identified by the Commission in its preliminary assessment, not whether they are apt to lead to candy world for satisfy the wishes of all third parties.
- My views are necessarily incomplete and they’re also work in progress. I’ve only read the limited publicly available information and have not had access to any confidential info or documents that might be contained in the case-file. Moreover, I have allocated two flights time to draft this (and I should ideally also do some billable work, you see), so I’ll (i) update and improve this document on the basis of any new thoughts or possible feedback and (ii) refine my thoughts for a forthcoming piece on Oxford’s Journal of Competition Law and Practice
- My views are mine (sounds like a tautology, but don’t always take this for granted in our area of work…); some of my colleagues and clients may well have different opinions.
- I haven’t worked nor for Google nor for any of the 17 complainants.
In case I haven’t yet got you tired before even starting, here is a methodological explanation. This will be a five-pronged analysis; I will very succinctly summarize (i) DG Comp’s concerns; (ii) my take on the substantive concerns; (iii) the content of the proposed commitments; (iv) third-party criticism of the proposal (notably that read here, here, here or here) (I actually read some favorable comments as well); and (v) my take on the proposed commitments. And this for each of the four concerns flagged by the Commission (although only the two first ones raise interesting issues).
The structure will make this post longer. In order not to cram the page, click if interested.
Commitments v (positive or negative) Decision
The first question to be asked is whether this is a case well-suited to be solved by resorting to Art. 9 commitments. We have hosted quite a few discussions on this point, and on the wider issue of the use or commitment decisions as an enforcement tool (see here for my take). As I said then, I’ve mixed feelings. On the one hand, I understand that the role of a competition authority is to effectively and efficiently eliminate restrictions on competition; from this viewpoint a negotiated solution would offer many obvious advantages. On the other hand, I don’t see DG Comp like just any other agency or judge enforcing competition law. I would argue that DG Comp has the most able people to effectively face, address and resolve the sort of complex issues raised in this case. If they –with scarce but excellent resources- don’t do it, others –with few and not so prepared staff- would probably do it worse.
Favorable treatment of Google’s specialized/vertical search sites
DG Comp’s concern: According to the Q&A document accompanying the launching of the market test, the Commission’s concern is that
“Google prominently displays links to its own specialised search services within its web search results and does not inform users of this favourable treatment. Due to the favourable treatment of Google’s own services, consumers are more likely to not make use of potentially more relevant competing services. First, users are not aware of the promotion of Google’s offer within the search results. Second, competitors’ results that are potentially more relevant are less visible and even sometimes not directly visible to users – they are more difficult for the user to find, for instance because the user has to scroll down the screen to see them or has to go to a subsequent search results web page.
The Commission is concerned that this practice unduly diverts traffic away from Google’s competitors in specialised search towards Google’s own specialised search services. It therefore reduces the ability of consumers to find a potentially more relevant choice of specialised search services”
My take on the concern: As developed in my presentation on the case and in a number of previous posts, I’ve never been a big fan of the not-yet articulated theory of harm underlying the concerns on search. Are they based on previously taxonomized theories of harm –refusal to supply, discrimination, tying- or on a new one? Contrary to some, I wouldn’t object to the latter (the idea that there’s a closed list of categories of abuse sounds very counterintuitive to me) but it would only be good to know more about it.
The Commission seems to acknowledge that Google’s Universal Search is an efficiency-generating practice that in itself raises no competition issues (the FTC recently concluded that it was “plausibly justified as an innovation that improved Google’s product and the experience of its users” and observed that other search engines present their results in the same way).
The Commission appears to accept the argument that Universal Search provides users a response to what they are looking for (thereby constituting not a different service, but rather a direct reply to increasingly sophisticated users’ queries; this is a crucial philosophical difference among the parties to the case); its concerns are limited to the display of the Google Specialised Results Link.
The Commission seems to be concerned that Google deprives vertical sites of traffic. On the contrary, I intuitively think that foreclosure is difficult to prove (again, I’ve not seen the case file, so I might well be wrong). Google is not the best engine for everything; other vertical sites have what an extinct antitrust boutique used to call “the advantage of focus”. When I look for books I do it on Amazon; when I look for hotels or restaurants I check Booking or Tripadvisor; and when I look for flights (or rather the poor Susana who suffers my plane reservations) we do it on specialized sites, and so on. I would actually intuitively think that Google directs lots of traffic to vertical sites that they otherwise wouldn’t receive.
That a company favors its own products is, in my opinion, not an issue per se. More often than not these choices are the natural outcome of a healthy competitive process. If you ask me, the key question lies in to what extent consumer choice is preserved, and in to what extent rivals are able to viably compete.
And here there is a sliding scale of options: choice can be made impossible by technical or legal lock-in; it can be made difficult by technical or legal lock-in; or it can be unlikely because of technical issues and/or those + consumer inertia [by the way, it’s somehow odd to have heard the Commission and Microsoft jointly argue in Court the other day -in Microsoft/Skype– that when “competition is one download away” there can never be competition concerns and to compare that stance with the one held in this case…] [ Note that I’ve included/promoted the previous hyperlink because it cites me; it was not a neutral and organic choice ].
If this is so, it seems logical to have a sliding scale of remedies: whereas some problems could be addressed by forcing the exercise of choice (e.g. the ballot screen in Microsoft II) or by favoring/offering such choice (like the one at issue here, which seems like an effective, intrusive –but not excessively so- alternative). The nature of remedy proposed in this case further confirms the underlying sui generis theory of harm, somewhere between refusal to supply and discrimination theories.
Google proposes that every time a query entered into its search books leads to a Google Specialised Results Link it will visibly display a label explaining that the link is not part of the organic –neutral- search results . In other words, it will make sure that users are not misled and can distinguish Google’s specialized services from its standard results.
If the Google Specialised Results Link leads to specialized results from which Google makes money (a) the label will indicate where alternative rival links can be found; and (b) Google will visibly display 3 pertinent rival specialized search services.
There are different scenarios depending on whether Google monetizes its vertical site indirectly (e.g. via pay per click or display advertising such as in local/maps) (in which case the 3 competing links will be organically selected and displayed at no charge to the pages they lead to, and the specialized results link will be distinguished by means of graphical features; example here), or directly (through Paid Placement; e.g. products) (when the Specialised Results Link leads to a page showing at least one paid result the 3 competing links will be auctioned and the specialized results link will be distinguished and placed in a separate frame than other results; example here)
[In mobile contexts the three links will appear after clicking on an “other sites” hyperlink; see here]
The 3 rival links are to be selected and ranked in each case pursuant to a mechanism set out in Annex I to the commitments proposal:
In essence, there are three steps: (i) any specialized search site –not horizontal ones like Bing- would be able to apply to be considered for a pool of eligible Rival Vertical Search Sites; (ii) Google will review whether the applications meet the conditions in Annex I (it has to be a “rival vertical search engine”, not just any web page; it needs to have been properly incorporated, have a minimum traffic, not incur in defined “harmful practices”; it cannot required log-in or downloading; they must lead directly to search pagers that respond to users’ queries, and they have to ensure an overall user experience –in the latter case the Monitoring Trustee will have a say- and will create a set of “vertical pools” (one for each category of Google Specialised Results Links); (iii) the 3 vertical sites to be displayed –which could opt out at any moment- will then be selected by virtue of (a) their Web Search Rank –their neutral rank in organic search results- when the pool is for free rival links; (b) an auction mechanism –usual Varian highest losing bid mechanism- for price-per click payments (subject to a maximum budget) when the pool is for paid rival links.
Third parties have voiced out that they see the proposed commitments as a “non-starter”. This is partly so because their concerns seem to relate fundamentally to Google’s natural search results, to AdWords listings and the very concept of Universal Search, which do not appear to concern the Commission.
Given the number of complainants (17 formal ones) and the number of Google’s active competitors, a myriad critics have been voiced out. Here’s a non-exhaustive list:
It seems that critical third parties would want a proposal setting out non-discrimination principles; that is, that all sites get exactly the same treatment as Google’s own services. They argue that non-discrimination means that Google should apply exactly the same crawling, indexing, ranking, display and penalty algorithms to its own sites.
Some third parties –the competing vertical sites that would benefit from this- seem to argue that the proposed labeling and the new display would not affect users’ clicking behavior. FairSearch and others seem to have argued that they need guarantees that they will not be excluded from the market, and that a certain minimum click-threshold would be appropriate.
There has also been criticism targeting the auctioning of the three competing links for domains that are directly monetized. Some third parties argue that Google will not only commandeer traffic, but also charge for it. Some also claim that the bidding process should be more granular (that instead of allowing for blanket bids across entire domains, there should also be bids for specific keywords), and that the envisaged 2-4 weeks bidding cycle is inadequate.
Criticism is also targeted against the requirement that rival sites must already rank in the top hundred thousand worldwide sites for traffic, and against the limitation that rivals should only be allowed to apply for one or a closely related set of pools.
It has also been argued that the registration process for the Rival Links Pool would facilitate Google’s task if it were to decide to remove competing vertical sites from its natural search results.
Some third parties further argue that the proposal would not preclude Google from demoting competing sites (this relates mainly to algorithmic penalties targeting “lack of original content”, although this does not seem to have given rise to concerns on the part of the Commission.
Finally, discontent has been voiced out against the apparent limitation of the remedies to EEA websites.
My guess is that, if third parties are pragmatic they will most likely concentrate their efforts on attempting to obtain constraints on Google’s ability as to where and how to display universal search results (flight search, maps and others) on its page, as well as on the prominence of the 3 competing links.
My take on the commitments:
The key point here is that the Commission, like the FTC, does not take issue with the notion of universal search itself. The Commission seems to want to preserve Google’s ability to provide Universal Search results –which are seen as an innovative and efficiency generating feature- while (i) ensuring that users are “aware of the promotion of Google’s offer within the search results”; and (ii) increasing the visibility of competitors, thereby enhancing consumer choice.
In this sense, the inclusion of the explanatory label and the placing of rival links alongside Google’s specialized/vertical results would seem adequate and proportionate to address the two concerns identified by DG Comp.
In fact, the Commission has addressed a problem that other antitrust enforcers (namely the FTC) did not consider so serious to merit a remedy, and it has managed to obtain changes to Google’s search page, which is a striking concession that the FTC probably never thought was feasible. From DG Comp’s point of view, the commitments –as market tested- would constitute an enforcement success.
Whether the display of three rival links is liable to generate more clicks to competitors or not is an empirical matter on which I can’t comment. But I’m not sure it’s relevant: one thing is to ensure a level playing field and to give opportunities (which, in my mind, is a bit of what competition policy, and public policy for that matter, are about), but imposing equality –or ensuring that competitors benefit from a minimum fixed number of clicks- sounds like hardly justifiable interventionism. Commitments should be about furthering user choice, not about favoring particular choices.
Whether the competing links are prominent enough or not could possibly be relevant as some third parties’ claim, but I’m inclined to think that the remedy (an obligation to alter the user interface and visibly and permanently display rival links imposed upon a company which prides itself on its simple user interface) is intrusive enough as to mandate too detailed specifications as to the size and placement of those rival links. I nevertheless don’t rule out that the Commission might ask for more specific rules in this regard.
The fact that the display of 3 rival links is limited to commercially relevant result pages also seems adequate. I have more doubts regarding the minimum traffic threshold proposed by Google for vertical sites to be displayed as part of the 3 competing links. My doubts are, again, against the background of Comp’s concerns: If one assumes (leaving aside for now what the bases for the assumption are) that vertical sites are deprived of traffic because they have no visibility on Google, then setting a high threshold of pre-existing traffic could be regarded as a barrier to entry for newcomers and small competitors.
With regard to the criticism that the proposal does not address Google’s ranking and its ability to demote other sites, I would tend to argue that Google must enjoy unfettered freedom in picking the criteria to effect its rankings, since that is precisely the recipe for its business. Depriving it of the ability to rank sites is akin to depriving it from the ability to actually do what it does (quite well). Google has all the incentives to provide the most useful search rankings.
I’m not saying Google’s results are always fair (I’d love my parents’ hotel, my firm’s site or this blog to be ranked higher; just saying, in case an algorithm is listening…) but nor are any similar rankings (ask most firms in Brussels whether they are happy with their positioning in Chambers, or wine makers whether they feel well treated by Robert Parker; but neither Chambers nor Parker are- for good reasons- told what criteria to use). [While I agree with most of the arguments put forward by Google’s partisans in this respect, I’m not so persuaded by the argument that search engines are a form of free speech within the sense of the ECHR].
Whether the auction mechanism is right or not is also one of the main issues on the table. Given that rival links would be displayed next to paid-for search results (i.e. advertising, as opposed to organic results), their presence should logically reduce advertisers’ willingness to pay for appearing there. In my view, it would not make much sense to oblige Google to provide competitors display at no cost when it charges its clients for it.
Moreover, I don’t see how the auction mechanism could be regarded as “raising rivals costs”. It would essentially be offering them an additional channel of display in exchange for whatever price they are willing to pay. As argued before, I don’t think Google is regarded as an essential facility for competing vertical sites, but even if it were, it would still be entitled to charge a price for access to it.
In sum, most of the criticism I have read is in reality targeting DG Comp’s preliminary assessment rather than the proposed commitments. The Commission could nevertheless attempt to extract further concessions or rather more concrete ones from Google in relation to how rival links will be displayed.
DG Comp concern:
The press release states that the Commission has concerns about “[t]he use by Google without consent of original content from third party web sites in its own specialised web search services”.
My take on the concern:
Some time ago I specifically addressed this issue at lenght, so I refer to my previous post and to the most interesting discussion that it happened to trigger (available here).
In a nutshell, I concur with Commissioner’s Rosch’s opinion issued within the framework of the FTC’s proceedings that this is a matter for copyright, not competition law.
Proposed solution: Google proposes to cumulatively
(a) offer all websites the possibility to opt-out from the use of all of their content in Google’s specialized search services; those opting out would not be penalized in the display of organic- generic- search results. (Google acknowledges that there could be an indirect effect which is outside its control; any subsequent technology changes that could affect the ranking of opted-out sites would be discussed with the Monitoring Trustee). A similar solution was endorsed by the FTC
(b) offer all competing search web sites specialized on products and local search to black out categories of information (e.g. telephone numbers or ratings of restaurants) that cannot be indexed and used by Google); (the data at issue would be marked by means of an HTML component, and it could not amount to more than 10% of the readable text of any web page).
(c) newspapers would be able to choose on a web-page by web-page basis whether they wish to (a) fully exclude their content from appearing in Google News; (b) select a date after which particular articles should no longer be displayed on Google News; and (c) prevent the display of snippets of their articles (as opposed to the links to the articles) on Google News.
Newspapers’ choice would not affect their ranking within organic search results [subject to the same caveat pointed out in point (a) above].
Some third parties opine that there should be more and more specific individual opt-outs.
My take on the commitments:
Google must have had a hard time offering this. Accepting the notion that it cannot crawl through certain info might not have been easy for its engineers to swallow.
As stated, I’m particularly reluctant with regard to this theory of harm, but forgetting about the substantive analysis and focusing on the proposed solutions, they would prima facie seem to be adequate to resolve Comp’s concerns.
I don’t see how one could oblige Google to further offer more granular opt-outs. From the moment a vertical site has the possibility to opt-out, it is hard for that site to claim that its content is misappropriated. If they’d rather be in, but with new or better conditions that is because they benefit from being there in the first place!
Exclusivity arrangements with third party web sites
DG Comp’s concern– The press release indicates that the EC takes issue with Google’s “[a]greements that oblige third party web sites (“publishers”) to obtain all or most of their online search advertisements from Google”
My take on the concern: Exclusivity obligations could possibly be harmful in a two-sided market with the characteristics of search. To me this is the most (sole) obvious concern raised in the investigation. Can’t say much more because I have no specific knowledge about the existence and/or scope of any such obligations.
Proposed solution: Google would no longer include in its agreements with publishers any written or unwritten obligations that would legally or de facto require them to source online search advertisements exclusively from Google’s AdSense for Search. (Google would retain the ability to require that advertisers label the Ads visibly, identify Google as their source, identify the ads as sponsored, and display Google’s ads in a single continuous block). [Note that the FTC’s deal allowed for more flexibility in relation to exports of AdWords campaigns, but did not extend to AdSense].
Complainants’ concerns: I haven’t really read anything on this.
My take on the commitments: Eliminating the alleged restriction seems effective and sufficient.
Portability of advertising campaigns
DG Comp’s concern: The EC has explained that its last concern relates to [c]ontractual restrictions on the transferability of online search advertising campaigns to rival search advertising platforms and the management of such campaigns across Google’s Adwords and rival search advertising platforms”.
My take on the concern: I see this as an add-on to the concern related to exclusivity arrangements, so I refer to my take with regard to that concern.
Proposed solution: Google proposes to facilitate campaign portability for AdWords API Clients.
Complainants’ concerns: I’m not aware of any public criticism targeting this.
My take on the commitments: Eliminating the alleged restriction seems effective and sufficient.
Google would commit to appoint and remunerate a qualified and independent Monitoring Trustee (to be approved by the Commission) to monitor the compliance with the duties that would flow from the proposed commitments. Interestingly, the section of the commitment proposal devoted to setting out rules governing the appointment, replacement, functions and obligations of the trustee –and of Google vis à vis the Trustee- is the longest in the proposal.
I don’t have time to comment any major comments on this section on the proposal for the time being.