Your guide to the Google Shopping decision

The Google Shopping case has shaken the competition law community to its core. Here is your guide to what the Commission has disclosed and what influential commentators have had to say about it.

First, let’s deal with the information that has come out of the Commission. The case concerns a leveraging of dominance from the search engine market to price-comparison services. Regarding dominance, the Commission relies on high market shares in all EEA markets (mostly of 90 percent) and on indirect network effects as the main barrier to entry (the more users who use Google’s browser the more attractive it is for advertisers and the more data Google has to improve search results and targeted advertising).

The comparison-shopping market, where the harm is alleged to have occurred, has been separated from merchant websites such as eBay and Amazon. The main difference is that comparison sites do not offer the option to buy the product of interest as the two mentioned companies do. In any case, the Commission argues that if one includes merchant platforms, Google’s conduct has caused harm in a substantial part of this joint market.

Now to the more complex point. The ruling on the conduct has been described by the Commission, in my opinion, in very confusing terms. The Commission insists that it is not meddling with how Google designs its results page or writes its algorithm. The conduct, however, concerns precisely how the company presents its own comparison-shopping results (at the top of the page with pictures of the products and their price) as well as the demotion of competitor services. The issue here is indeed the product design.

Regarding the harm in the market, the Commission states it has evidence that traffic to rival comparison sites has decreased significantly and that this can be traced directly to the demotion to subsequent pages of the results.

The commentary on the decision has concentrated on the harm part and the conduct. Some do argue that Google’s dominance is overstated but the harshest criticism has been directed at the conduct-harm analysis of the Commission. I will give you a review on five articles written by the following five influential commentators: Alfonso Lamadrid, Nicolas Petit, Alden Abbott, Geoffrey Manne, and Pinar Akman. As you will see, my conclusion will be in most cases that we have to wait to see the decision to make a final judgment on the merits of the case. It might not be the most exciting assertion but I think is the more sensible one. Let me explain to you why.

Product improvement or exclusionary conduct?

One of the arguments made in all the articles is that we are in the presence of a product improvement that the Commission has wrongly labeled as anticompetitive. Nicolas Petit used a more cautious language, though. I do agree that the universal search results, as the informative boxes in the results page are called, cannot be considered harmful to consumers. They provide valuable information in an improved format. The mentioned articles, however, overlook at least two important points. First, the boxes could include non-proprietary information, that is, content from other comparison-shopping sites. If such content is more relevant, then there could be harm to consumers if Google systematically favors its content. Second, the conduct that has been punished, as stated by the Commission, is not only the display of the box but also the simultaneous demotion of competing websites to the fourth page in the search results. This second part has been completely bypassed by the mentioned articles. It will be an important issue to make an assessment of the decision. There could be one hundred thousand business justifications for the demotions so I guess we’ll have to wait and see what the decision and Google said in more detail about it.

In this respect Pinar Akman makes a slight distinction. She argues that Google’s favoring of its own content in the universal search results (the box) can be seen as advertisement of its own products, which then makes the Commission decision incomprehensible. This is perhaps one of the best business justifications that Google could offer. However, one weakness of this argument is that across categories of queries Google does use third party content in the box—content from Wikipedia, for example.

Lots and lots of choices

Another popular criticism of the decision is that consumers have lots of choices when it comes to this type of content (comparing product prices). That is certainly true. Consumers can compare prices not only on the websites that Google allegedly excluded but also on Amazon and eBay, who in addition offer the option of buying the product through their own website (which may arguably make them a better choice than googling the product). All the scholars mentioned before make this point. What they do not address is the Commission’s statement that even taking these choices into account Google has significantly distorted competition because comparison-shopping websites are still closer competitors in the relevant market that was defined. My guess is that no one commented on this because everyone considers it a wild assertion.

What I would have to see to make an assessment is data on how consumers look for product information. The main issue would be to see if they favor comparison-shopping sites, and whether these are mainly reached through search engines (where Google is alleged to be dominant). Geoffrey Manne points that in the US more than half of product searches start at Amazon. If that were the case in the EU as well, then the Commission’s position would be weaker. We will have to wait and see the information on which the Commission based its conclusions.

No duty to cooperate with rivals

A related point to the previous two is that if Google is not the only or most important choice for consumers, then there is no duty to cooperate with rivals. The Commission’s argument is that the company has to apply the same algorithm to its own content, even when it chooses what to display in the box. Lamadrid argues that since Google has not been labeled as an essential facility, such an obligation is unwarranted for. He uses a supermarket analogy to portrait how, in his view, the Commission’s decision makes little sense. One should not expect a grocery store to not treat its proprietary brands favorably over those of competitors and such a conduct does not harm competition because consumers can go somewhere else. Abbott uses the same analogy. What both miss is the inadequacy of comparing a supermarket with a search engine. The latter is a multi-sided platform to which different economic principles apply in order to establish dominance. Google has different incentives in promoting its own content than a supermarket does in placing its own brands on a strategic shelf position. The latter makes some profit when it sells other brands. Google does not make a cent when providing a user with product-price information. Google’s interest is in increasing its traffic to monetize the other sides of the market (vendors who may be charged a fee per click on the product or advertisers who wish to target a given audience).

Although the supermarket analogy is wrong, what is still true is that the Commission has omitted the use of the term essential facility. That makes the case, as Lamadrid points out, unprecedented. As a matter of EU law this makes the Commission’s position peculiar but is not, in my view, fatal to it. From an exclusively EU law perspective the case is, as Petit points out, on strong grounds because the standard for exclusion is too low. As a matter of economic policy, the law as well as the case may be wrong. This point is closely related to the conclusion I reached in the previous point regarding the choices consumers have. If the Commission failed to prove harm beyond the exclusion of a small group of firms then it might have caused a great deal of harm from a consumer-welfare perspective. Receiving a pat on the back from the General Court and the Court of Justice would not change that.

Hard to trace rival’s traffic loss to Google’s conduct

One point that Pinar Akman and Alfonso Lamadrid make in their respective articles is that it is hard to trace the loss of traffic from the affected comparison-shopping sites to Google’s conduct. In addition, they argue that there are more plausible explanations for that (for example, changes in consumer preferences). Although it is true that there may be many reasons why these websites experienced fewer visits, there are statistical tools that can be used to elucidate the causal flow. The Commission may have had all the information it needed at its disposal to make such an inference. We’ll know this in due time.

The fine

It is the factor that has come as the biggest surprise. As Lamadrid and Petit point out in their articles, the Commission attempted to settle in the past, which implies a consideration that the case was not eligible for a fine (recital 13 of Regulation 1/2003). Both of them put too much weight on this consideration. First, the recital does not prohibit the Commission from pursuing the case if a settlement cannot be reached. It is a longshot, at best, to argue that once the Commission attempts to solve the case via commitments it falls into a trap. What kind of leverage would it have in negotiating if the prospect of a future fine is removed from the picture? As Jones and Sufrin point out, on the other side of the coin, the Commission has accepted commitments when the allegations concerned serious violations to article 102 of the TFEU.[1]

As an additional point of surprise, Akman points out that the decision goes against previous practice. The Commission does not usually impose a fine when it is unfamiliar with the conduct. As a matter of law this is of no relevance. As a matter of policy it might be the sensible approach and it is indeed surprising that the Commission decided to set a world record under such circumstances and in a market where the existence of such a magnitude of harm does sound a bit odd. 2.4 billion euros is a lot of money. I am looking forward to see where this magic number came from.

Google is one of the most innovative companies on earth

Akman and Petit point out that Google has the 4th largest R&D expenditure in the world. From a legal perspective, it doesn’t matter how good a monopoly is if it indeed committed a competition law violation. From an economic policy perspective it is of some importance only if the fines and order have the ability to reduce Google’s incentives to innovate. In the case at hand, this could be true if indeed the decision sends the wrong signal, even if it is in accordance with EU case law. If there is not enough evidence in the decision to support the competitive harm, for example, then companies in the EU will be left under an uncertainty cloud. Google may have technical reasons to prefer its own content where comparison-shopping is concerned, and if the Commission did not give them due weight, innovation in Internet platforms may indeed be affected.


The commentary that I have analyzed has some good criticism of the Commission decision, at least based on what has been disclosed so far. However, it cannot be stressed enough that we can only make our final judgment once the decision itself is published. The main point will be that of harm. It is difficult, though not impossible, to prove that Google’s conduct is indeed welfare reducing, as Geoffrey Maine has pointed out in his criticism to a study regarding other types of search queries (local coffee shops). Other than that, I have nothing more to say than I am looking very much forward to see the decision in full.

[1] Jones, A. & Sufrin B. (2014). EU Competition Law: Text, Cases, and Materials, p. 982. Oxford University Press.

Tagged , ,

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: