While this paper covers all aspects of the Digital Markets Act, our twelve-years (and counting) involvement in the Google Search (Comparison Shopping) case means that our analysis has a particular emphasis on the DMAs implications for Google’s ongoing search manipulation practices.
Last Friday, November 4, Google Executive Chairman Eric Schmidt submitted his written response to questions posed by the Senate Antitrust Subcommittee, following his testimony before the Committee in September.
Foundem’s Video deconstructing key elements of Mr Schmidt’s testimony before the Senate Antitrust Subcommittee in September 2011 (See here for the original full-length version of this video)
In a traditional bricks and mortar world, Google’s anti-competitive practices would be obvious to all. But in the unfamiliar and seemingly impenetrable world of Internet search, these practices are often predicated on Google’s ability to bamboozle, confuse, and obfuscate.
Behind closed doors, Google’s confuse-and-conquer strategy has been unravelling for over a year. Now, Mr Schmidt’s public testimony and written answers to the Senate Antitrust Subcommittee provide the first public glimpse of the extent to which Google’s paper-thin defence crumbles in the face of informed scrutiny.
A few of Mr Schmidt’s written answers seem gratuitously evasive. Many are markedly at odds with Google’s existing documents and prior statements. And some are just plain wrong.
The significant inconsistencies and inaccuracies in Mr Schmidt’s testimony at the hearing have only been amplified by those in his written answers. Taken together, the two misleading and often contradictory accounts seem to reveal a company beginning to sense that it has already lost the argument.
We suggest that our recent video (above), which deconstructs several key elements of Mr Schmidt’s testimony, is essential viewing for anyone wanting a better insight into some of the breathtaking semantic acrobatics of Mr Schmidt’s testimony and written responses.
The following are just two examples of Mr Schmidt’s problematic written answers following the hearing:
Is Google Product Search a Price Comparison Service?
During the hearing, Senator Lee referred extensively to a Foundem study examining the comparative Google rankings of the US’s leading price comparison services, including Google’s own Google Product Search. In the hearing, Mr Schmidt ducked several of Senator Lee’s questions about the troubling findings of this study with the bizarre claim that Google Product Search is not a price comparison service.
In our blog post immediately following the hearing we demonstrated the fallacy of this claim with evidence that included a screenshot of Google’s own description of Google Product Search as a “price comparison service”.
Senator Lee’s post-hearing follow-up question:
On September 28, 2011, a search query on Google for “UK product search” returned Google Product Search as the first result, described as “Google’s UK price comparison service.”
Is Google Product Search a price comparison service?
Mr Schmidt’s written response:
Google product search is a type of thematic search that allows consumers to compare prices and see which websites are selling a particular product.
In other words, “Yes.” As Mr Schmidt will be aware, all price comparison services can be described as “a type of thematic search that allows consumers to compare prices and see which websites are selling a particular product.”
It is difficult to understand how Mr Schmidt could have been so confused about this straightforward point during his original testimony. Google’s systematic favouring of its own price comparison service through Universal Search is an issue that has been central to formal antitrust investigations on both sides of the Atlantic for well over a year.
Does Google Favour its Own Services?
Although Mr Schmidt now concedes that Google Product Search is in fact a price comparison service, his written answers duck many of the same questions again – but this time by claiming that Google Product Search and other specialised Google services are not in fact “services”.
Senator Lee’s post-hearing follow-up question:
Are Google products and services subject to the same search-ranking algorithmic process as all other organic search results?
Mr Schmidt’s written response:
…what is crucial to understand is that thematic search results are not separate “products and services” from Google…Because of this, the question of whether we “favor” our “products and services” is based on an inaccurate premise. These universal search results are our search service—they are not some separate “Google content” that can be “favored.”…
Even if this were true, it would make no difference to the correct answer to the question. Whether Google Product Search, Google Maps, Google News, Google Places, Google Travel, Google Finance, and so on, are considered “services”, “products”, “off-shoots”, “sub-sections”, “sub-categories”, “thematic collections of search results”, “bananas”, or “hoojamaflips” makes no difference; as long as Google owns them, and profits from driving traffic to their pages, then what Google now chooses to call them is entirely irrelevant.
But of course, it isn’t true. Because these Google services are in fact “services”, no one should be surprised to learn that Google has consistently and repeatedly referred to them as such in all manner of places and circumstances. For example:
“The Google Product Search service is provided AS IS and Google expressly disclaims to the fullest extent permitted by law all express, implied, and statutory warranties regarding the information included therein. Under no circumstances shall Google be liable to any user on account of their use, misuse, or reliance on the Google Product Search service.“
Even Search Engine Land (which many have accused of being fervently pro-Google) was moved to take issue with Mr Schmidt’s claim:
“Google most certainly does have separate products and services, despite what Schmidt repeatedly told Congress today. It’s disingenuous, at best, for him to claim that YouTube, Google Maps, Google News and other Google products that appear as universal search results aren’t actually separate products. They are.”
Unfortunately, Mr Schmidt repeatedly restates this claim to avoid answering several of Senator Lee’s related, post-hearing questions. For example:
Senator Lee:
Does Google display Google [Product Search] results within its natural search results without any label identifying them as Google results or as otherwise distinct from true “search results”?
Mr Schmidt:
As stated in my response to Question 1, universal search results are not separate “products and services”; they are our “true” results.
And
Senator Lee:
What has Google done to let its users know that its natural search algorithm gives preference to Google’s own products and services?
Mr Schmidt:
As described in my response to Question 1 above, I believe that the premise of this question is incorrect.
We suspect that, in time, Google will reflect on this period with a sense of shame and regret.
If Google can justify promoting its own services through Universal Search while demoting its competitors for the lack-of-original-content that is a defining characteristic of all search services, then why does it not simply stand up and say so?
We are quite certain this is a debate Google will ultimately lose. But it is a debate that should have started two years ago. It is time for Google to come clean and start arguing its position based on facts rather than on forever shifting fiction.
Following on from our earlier posts around last month’s US Senate Antitrust hearing into Google, this video, comprising annotated highlights of the hearing, is intended to provide some background and insights into the more important inaccuracies and inconsistencies of Mr Schmidt’s testimony:
Some of Senator Lee’s questioning in yesterday’s hearing centred around a study conducted by Foundem in April this year. This study of Google’s US search rankings for hundreds of product- and product-price-comparison-related search terms was a follow-up to Foundem’s earlier study of Google’s UK search results (included in Foundem’s submission to the FCC and in Foundem’s EU Antitrust Complaint filed last year). The graphical output of this study demonstrates the breathtaking extent to which Google’s Universal Search mechanism consistently places Google’s own price comparison service at or near the top of nearly all product- and product-price-comparison related searches.
Foundem’s contention – which is based on Google’s own descriptions of its Universal Search mechanism – is that Google uses different algorithms and relevance signals to rank its own services than it does to rank everyone else’s. Because Google applies a different standard to its own services than to everyone else’s, Google can systematically favour its own services to whatever extent it chooses.
Many of Mr Schmidt’s answers to Senator Lee’s probing questions on this subject seemed to hinge on the surprising assertion that Mr Schmidt does not consider Google Product Search to be a price comparison service:
“There’s a conflation of two different things going on in this study…there’s a difference between sites that do product comparison and sites that offer products themselves. Google Product Search is about getting you to a product, and so we tend to look for the product as opposed to the product comparison in this particular case. Which is why the product is more highly ranked than the result of a product comparison site. If you did the same study with all of the other product sites you would find a very different result.
…Things like the companies that are mentioned there are price comparison shopping; they are different animals if you will. They are important, but they do different things. Google Product Search is about searching for specific products. In that sense Product Search does something similar to what PriceGrabber, Nextag and Shopper does, which is why the confusion exists. … It’s an apples to oranges comparison.”
Eric Schmidt, September 21 2011
In our earlier blog post we pointed out that this assertion raises an obvious question: namely, if Google Product Search isn’t a price comparison service, why does Google consistently insert this service at or near the top of the vast majority of Google’s price-comparison-related search results?
But more importantly, Google Product Search is a price comparison service. The following screenshot (taken this morning) shows that Google itself describes Google Product Search as a price comparison service (as does Wikipedia and ComScore):
Mr Schmidt also downplayed the anticompetitive impact of Google’s preferential treatment for Google Product Search (and other services) by suggesting that the inserted links were merely links to vendor sites, not to Google’s own service. But this is not the case.
The following screenshot, for example, shows a Universal Search result for Google Product Search similar to the one featured in the screenshot from yesterday’s hearing. In this example, and in the one from yesterday’s hearing, all four of the featured links are links to Google’s own Product Search service, not to vendors:
The Senate Antitrust Sub-Committee used one of Foundem’s scattergram graphs in its hearing earlier today. Senator Lee asked Mr Schmidt a few questions around the data. We believe that Mr Schmidt’s answers may have been misleading in several ways. For example:
Mr Schmidt suggested that the data was comparing "apples to oranges". Mr Schmidt implied that Google Product Search is some kind of "product search", not some kind of price comparison service. This is strange. Most people would consider Google Product Search to be a price comparison service. We wonder what Mr Schmidt would consider to be Google Product Search’s competitors, if not price comparison services?
Also, if Google Product Search is not a price comparison service, then it begs the important question: why did Google Product Search show up so consistently at or near the top of Google’s search results for the hundreds of product-price-comparison-related queries shown in Foundem’s graph? Why does Google’s Universal Search mechanism so consistently place Google Product Search at or near the top of most price-comparison-related searches, if Google Product Search is not a price comparison service? For example, hundreds of the data points shown in Foundem’s graph were for queries of the form: "compare prices [MAKE MODEL]" and "best prices [MAKE MODEL]".
Mr Schmidt also suggested that if a user clicked on these prominently placed Google Product Search links it would merely take them to a vendor site. This is not true in most cases. In the vast majority of cases a click on these links takes users to Google Product Search – Google’s own ad-supported service – not to a vendor.
The data in the above chart was collected on April 15 2011. Some examples of the 650 product- and product-price-comparison-related searches illustrated include: “Canon PowerShot S95”, “best prices Canon PowerShot S95”, “compare prices Canon PowerShot S95”, “Canon PowerShot S95 prices”, and so on for a wide variety of products.
For a more complete description of some of Foundem’s analysis and arguments around the anti-competitive power of Google’s Universal Search mechanism, see our comment to the FCC.
We note with interest Google’s recent announcement that it has changed its search algorithms to further target “sites that copy others’ content and sites with low levels of original content” (Matt Cutts, Google’s Head of Search Quality, 28 January 2011)
Is this simply a commendable attempt to punish spam and reward the authors of original content, or does it mark an escalation in Google’s ongoing disadvantaging of rival vertical search services?
Google recently used similar language to try to justify Foundem’s three year exclusion from Google’s search results:
Google says it "de-indexed" [Foundem] because much of its content – about 87% – was copied from other sites, which it says leads to automatic downgrading in its search results. (The Guardian, 30 November 2010)
But, as Google is well aware, copying, organising, and presenting the content of others is a defining characteristic of any search service, including Google’s own.
The following extract from Foundem’s submission to the European Commission speaks to the heart of this issue:
Google – the Godfather of Spam?
The Original Content Fallacy
Google tends to emphasise the value of content, while downplaying the value of service. Its “original content” mantras are convenient for Google, first, because Google requires 3rd party content to hang its ads on, and second, because these mantras help to foster the view that rival search services have little inherent value.
Accusing a search service of having little or no original content is like accusing a library of not writing its own books. While accurate, it is clearly missing the point of the valuable service that a library provides. Besides, the same accusation can be levelled at Google. But when KinderStart did just that in 2006, Google’s attorneys objected vigorously:
“According to KinderStart, Google’s search results function solely to link a user to third party websites and contain no original content. But Google’s search results are original content, expressing Google’s opinion of the relative significance of websites”, Google Attorneys, 2006[1].
Clearly, exactly the same argument applies to Foundem or any other legitimate search service. Moreover, by gathering all of the relevant information about a product, flight, house, or job from dozens of suppliers and presenting it all on one sortable and filterable page, any high-quality vertical search site clearly provides a valuable service, whether or not it authors any original content of its own.
Google’s relentless obsession with original content—whether feigned or genuine—has serious consequences:
· It inevitably leads to cases like Foundem’s, where the appeals of an unjustly penalised search service can be repeatedly ignored by a succession of Google employees—all indoctrinated with the false belief that only original authored content adds value.
· It encourages legitimate sites to overreach their area of expertise to write ‘fake’ reviews, ‘fake’ buying guides, and ‘fake’ blog posts, in the pursuit of this Google-mandated original content. To the uninitiated, this may sound like crazy talk (surely this fakery is rare), but the production of essentially ‘fake’ content and the acquisition of essentially ‘fake’ natural links has become the mainstay of mainstream SEO.
Google is Conflicted about Web Spam
Another unfortunate consequence of Google’s increasingly relentless focus on content over service is the proliferation of Web spam.
Thirty percent of Google’s revenues come from its AdSense programme[2], which extends the reach of its advertising network across millions of third party websites and has played a crucial role in Google’s meteoric growth. AdSense allows any site (spam included) to efficiently monetise its traffic by simply copying and pasting a few lines of JavaScript. This ease of integration, coupled with the fact that Google does little to vet its AdSense publishers, has made AdSense one of the driving forces behind the explosion of Web spam. So much so, that the acronym MFA (Made-for-AdSense) was coined to describe the millions of spam sites, comprising “original” but low-quality content, that are thrown up every year simply as vehicles for Google ads.
“If Google wants to organize the world’s information they shouldn’t fund a large portion of the world’s information pollution.” (Aaron Wall, November 14 2007[3])
It is important to realise that Google has a serious conflict of interest with these Made-for-AdSense spam sites, because it has a substantial stake in their revenues.
It is also worth noting that Affiliate marketing is one of the few viable alternatives to AdSense as a means of monetising Web sites. Most sites looking to monetise their traffic can choose between revenue-bearing AdSense ads or revenue-bearing Affiliate marketing links. Currently, both methods are used by legitimate and illegitimate sites alike.
Despite Made-for-AdSense spam sites outnumbering thin-affiliate spam sites by some considerable margin, Google harps on relentlessly about “thin-affiliate” sites but barely mentions the far more prevalent problem of Made-for-AdSense sites.
Google is Also Conflicted about the Quality of its Own Search Results
The vast majority of Google’s revenues come from users clicking on the sponsored links to the right and above Google’s search results[4]. But users don’t go to Google for its sponsored links.
Users only really look to sponsored links when their search results don’t have what they’re looking for. For any ostensibly free search engine, there is an inevitable tension between the need to produce good enough search results to attract users and the need to ensure they are bad enough that users regularly resort to clicking on the sponsored links to find what they are looking for.
If Google’s search results were perfect, its revenues would plummet. It is inconceivable that this inconvenient truth has not occurred to Google’s strategists. It is therefore sensible to question Google’s motives whenever it claims that preserving the quality of its search results is its top priority.
Google tries to characterise its business model as somehow purer than that of vertical search because it does not earn revenues directly from its organic results. The reality may be quite the opposite. There is something inherently peculiar about a business model that is characterised by the need to be good but not too good. And this, of course, is hugely exacerbated by the distinct and persistent lack of healthy competition among horizontal search engines.
- Foundem EU Submission, August 2010
Debunking Google’s Search-Within-Search Fallacy
“Google is a search engine. A search engine’s job is to point you to destination sites that have the information you are seeking, not to send you to other search engines”. (Danny Sullivan, The Incredible Stupidity Of Investigating Google For Acting Like A Search Engine, Search Engine Land, 30 November 2010[5])
This spurious argument is never far away when Google (or its unofficial spokespeople) are defending Google’s disadvantaging of rival vertical search services.
The following extract from Foundem’s submission to the European Commission addresses this issue:
Of course, horizontal search engine results like Google’s should not feature the search results of other horizontal search engines. It would be a recursive nightmare, for example, if Google featured Bing, which featured Google, which featured Bing, and so on. Fortunately, this scenario is as unlikely as it is undesirable.
…
The real question is whether horizontal search engines like Google should feature results from vertical search and price comparison services. The answer is clearly ‘yes’, first, because they can, and second, because users routinely search for these services on horizontal search engines. For a horizontal search engine to deliberately exclude these services would be to deliberately frustrate many of its users.
Indeed, by routinely inserting its own vertical search services into prominent positions in its Web results through Universal Search, Google now emphatically endorses the view that vertical search and price comparison services are often exactly what its users are looking for.
- Foundem EU Submission, August 2010
In conclusion, original authored content is not a legitimate requirement for a search service. Search services such as Google and Foundem are not intended to provide this kind of original content; they are intended to efficiently search and summarise the content of others. Yet, according to Google, it penalised Foundem for allegedly failing to meet this spurious requirement – a requirement that none of Google’s own search services fulfil. The notion that Google can penalise a search service for a lack of original authored content has very serious anti-competitive implications for the future of innovation in search.
Well, it’s the end of what has been a very long day on the receiving end of a mud-slinging barrage from what is perhaps the world’s mightiest PR machine.
There are a number of important inaccuracies in the story Google has been weaving, which we will address in due course. But, in the meantime anyone interested in the facts behind Foundem’s European Anti-Trust complaint might want to read the following:
Foundem’s FCC Comments (Describing how Google’s Universal Search mechanism poses an immediate threat to healthy competition and innovation)
One important point that probably shouldn’t wait is to correct the impression that Google has tried to create that Foundem’s European filing is concerned solely with Google’s Penalty algorithms. Either Google hasn’t read the entire complaint (it is quite long!), or it is being deliberately misleading. The complaint gives roughly equal billing to Google’s increasing use of arbitrary penalties (both in Search and in AdWords) and its increasingly aggressive deployment of Universal Search (Google’s mechanism for “blending”/bundling its own services into prominent positions within its web search results). Foundem’s position is that, coupled with Google’s overwhelming dominance in Search, these tools give Google an unparalleled and virtually unassailable competitive advantage. And this advantage reaches far beyond the confines of search. Universal Search allows Google to leverage its search engine monopoly into virtually any field it chooses: price comparison, books, online mapping, financial search, property search, job search, travel search, and so on.
