I’ve found it tough to get my head around all the arguments in the recent EU judgement against Google. I find that writing helps me get my thoughts together and work out what I really think, so here goes. Let’s find out whether I agree with the ruling or not…
First – the background – you only really need to read two things to get the gist of the complaint:
The EU’s press release announcing the record EUR2.4 billion fine is surprisingly accessible and readable, and links out to a bunch of useful resources
Google’s response is relatively short and seemingly light on content, but actually frames the key points of the counter-argument well when you know what you’re looking for – see below!
Definitions
In a case like this that relies on complex technical subjects and also areas of the law (in this case competition law) with which most of us are not familiar, I think it’s useful to make sure we are all talking about the same things. In fact, this is probably my biggest criticism of the EU’s press release (not the judgement – just the communication). It would have benefited greatly from making sure we are all talking about the same things. Here are my best explanations of the key elements you need to understand:
Search (sometimes “general search”) – at its simplest, this is the process of typing a query, and receiving links to pages that satisfy your need. In practice, it has extended to other inputs (e.g. voice) and other outputs (e.g. answers, rich results etc.). We’ve written extensively about changes in the search market
Organic search – the results of searches that appear in an order determined by the search engine based on their quality and relevance as well as their likelihood of satisfying the searcher’s intent. Organic (or “natural”) search results are not advertising results, no money changes hands, and there is no way to pay for inclusion or for a better ranking
Paid search (also “Pay Per Click” or PPC) – adverts sold by Google and other search engines allowing advertisers to pay to appear next to search results based on the search word or phrase (and other variables)
Comparison shopping engines (CSEs) – websites where you can search for a product type, category or brand and then compare different products and/or different retailers – typically by sorting, filtering, and applying facets. The business model is typically either for retailers/brands to pay directly for inclusion, to pay for the clicks they receive, or for the engine to receive affiliate payouts when a searcher buys from the destination site
Google Shopping – Google’s own comparison shopping engine – where you can sort and filter products, apply facets, and click out to retailers’ sites – typically accessed via the “shopping” link at the top of a search results page
Product Listing Ads (PLAs) – a form of paid search whereby adverts for individual products (complete with rich information in the form of photos and prices) appear above, or to the side of the search results. Individual PLAs are presented in the same form as product links in Google Shopping, and the data comes from the same sources, but PLAs in the search results do not constitute a comparison shopping engine to my mind – there is no filtering and there are no facets
Want more advice like this in your inbox? Join the monthly newsletter.
//
A challenging definition – what markets are we talking about?
In Google’s response, they highlight Amazon and eBay as examples of sites that have grown during the time in question (despite Google’s alleged anti-competitive behaviour), and which offer much of the user benefit of comparison shopping engines. The user experience is clearly similar, while the key difference is that you can actually check out and buy on both sites. In some cases on Amazon, you are buying directly from Amazon – i.e. they are also the retailer – and in many cases on both eBay and Amazon, they are functioning as a marketplace so you enter your payment details on their site, but you are buying from a third party. This raises two difficult and related questions for me – questions which the EU has not answered to my satisfaction:
Must comparison shopping engines send users off to another site in order to purchase? If not, and Amazon and eBay are examples of comparison shopping engines, then I think the case is much harder to make – certainly some CSEs have fared poorly during the time period in question, but some (most notably Amazon) have thrived
If comparison shopping engines are narrowly defined, is it really a separate “market”? The EU’s case relies on a finding that Google is using its dominance in one market (general search) to crush competition in another market. I’m not even convinced that “product search” is actually a separate market to “general search” (I’m inclined to think users see everything they do from the main Google search box as one thing), but if you define comparison shopping so narrowly that it’s a different market to both “search” and “searchable marketplaces” (or whatever you label Amazon and eBay as) then the sub-divisions stop making sense to consumers in my opinion
Going even further down this rabbit-hole, I would love to have an expert (a competition lawyer?) explain to me how the markets are delineated and where these different features/businesses fall – all of which satisfy some of the same user intent:
General search
Product search
Product comparison engines (facets, filters, etc)
Visual search (e.g. Pinterest)
Product recommendation sites (e.g. The Wirecutter)
The easy points of agreement
I don’t think it is hard to make the case that Google meets the criteria to be considered a monopoly in “search” (almost any way you define it) in Europe. Google themselves make essentially no effort to rebut this, so let’s allow this strut of the EU’s argument.
Google has a comparison shopping engine – in the form of Google Shopping (previously Google Product Search and originally Froogle). You can see this in action by going to the “Shopping” tab at google.co.uk and searching for a product (the example in Google’s own post is [puma shoes]). You then get the opportunity to compare products by a range of metrics and facets, with the links going out to places to buy the individual products (retailers, manufacturers, brands).
The problems with the EU’s case
In addition to the definitional problem I highlighted above (and that Google presses on heavily in their response), I think there is another problem with the EU’s case in the specific market of shopping (note that the ultimate EU case is much broader and covers many other verticals – more on that below).
The EU’s ultimate finding is:
Google abused its market dominance as a search engine by promoting its own comparison shopping service in its search results, and demoting those of competitors — Commissioner Margrethe Vestager
I have issues with both parts of that:
Where does Google promote “its own comparison shopping service in its search results”?
You will not find Google Shopping pages ranking in Google’s organic search results. Nor will you find links to Google Shopping in paid search links. There is only one link to Google’s own CSE on any of their search results page – on shopping queries it is here:
And on other searches it is in the “more” menu:
Now, it’s possible that even this level of cross-promotion (simply linking to the Google Shopping product from their menu) is too much, and maybe they should remove it (see below) but this is too small to warrant the huge fine in my opinion. (There is actually one more way of getting there – if you click the right-arrow by the PLAs multiple times to scroll through all the products on offer, you eventually get to a link that takes you to the Google Shopping results page. I’m willing to bet that the percentage of people who actually do this is miniscule and Google could remove it with essentially no impact on Google Shopping).
I believe the EU has a problem with the individual products listed at the top of the search results in the first image above – but that is very clearly not comparison shopping functionality (it is very similar to the links to individual products in the organic results below) – it’s simply what paid search looks like on commercial product queries. It is also very clearly not links to Google’s own CSE – those links go out to retailers’ sites – not to Google Shopping. The data comes from the same source – that is all.
“…abused its dominance…by demoting…competitors”
As I pointed out above, Google Shopping does not appear anywhere in the organic search results. To the extent that they treat Google Shopping differently to a third party CSE, they treat it worse – ciao.co.uk (one of the complainants does appear somewhere in the organic results, while Google Shopping appears nowhere). [In reality, they actually treat it identically – if Ciao were to block Google’s crawlers the way Google Shopping does, they would also rank nowhere in the organic results].
Given that none of the organic results are Google properties for any of these product queries, any “demoting” of a specific competitor involves the promoting of another. For every Ciao that loses rankings, there must be a retailer, marketplace, or other CSE that gains rankings.
This argument is problematic in other verticals – where there are links to Google properties in the search results, and those links do push down links to competitors – but it holds strong in comparison shopping as far as I can see.
The competitors are objectively poor in comparison shopping
If you rule marketplaces and retailers like eBay and Amazon out of the “comparison shopping engine” market, then the general quality of these sites is low. The complainants in particular – foundem.co.uk and ciao.co.uk are both very much worse user experiences than either regular Google search or Google Shopping. This SearchEngineLand article breaks Foundem down nicely, while Ciao is still online and you can go and see for yourself:
Slow loading, intrusive irrelevant banner advertising, broken links, missing images, and irrelevant reviews:
“Advantages: Great price, good quality, useful pockets” — review on the Puma shoes results page (emphasis mine).
I know that EU competition law focuses on the impact on competitors rather than the impact on consumers as US competition law does, but we should also step back a moment and look at the fact that these sites complaining about unfair treatment are objectively worse than Google’s offering.
[Note that this is not the case in other verticals – travel and financial services, in particular, are sticky areas for Google – see below.]
A ballsy response from Google
I hesitate to say that this would be my recommended course of action if I were advising Google, but a direction I would love to see them take is as follows:
Leave PLAs as they are – if comparison shopping is a separate market to “general search” in which Google has a monopoly, then PLAs definitely fall in the general search part rather than the comparison shopping part. They are integrated into the results a searcher receives when they perform a search that starts at the Google homepage, and there is no comparison functionality – it simply links to products
Remove the shopping link in the top menu – this is the one area I can see that they have favoured their comparison shopping engine (Google Shopping) over others (e.g. ciao.co.uk – one of the complainants) who cannot get their homepage linked from the top menu
Open up Google Shopping pages to their own search index – i.e. enable pages like the result you find when you search [Puma shoes] on the Google Shopping tab to be indexed and appear in the regular organic search results (to be clear, this does not happen at the moment – Google keeps these pages explicitly out of the main search index). Doing this will increase competition in the general search results for the complainants, but it clarifies that Google is treating their comparison shopping engine (Google Shopping) exactly on a level playing field with competitors such as ciao.co.uk and paves the way for them to treat (all) comparison shopping engines as harshly as they like in regular search [I’m not the first to think of this – see Danny Sullivan’s excellent article from the beginning of this case]
The reality: this is really bad for Google
The EU has started with what I think is the weakest of the verticals, and I think there are strong arguments that Google has not abused their market power in the specific ways this case claims.
But the EU has ruled against Google. In the weakest case against them there is.
The EU has shown here that they are prepared to take action to defend businesses offering worse user experiences against integrated changes Google makes to their core search engine. Shopping is an arguable case, but there are many more verticals where this precedent opens the way for future large fines:
Travel (especially flight search)
Maps / local
Jobs
Financial products
Images
Video
Apps
News
If the EU insists on treating each of these verticals the same way they have treated shopping, then Google is facing many huge fines – quite aside from the AdSense and Android cases (each of which could be a big deal in its own right). If the EU is prepared to take a lower-quality competitor in each case and say competitiveness has been harmed by Google’s inclusion of a great user experience directly in the general search results, then each of these is at least as egregious as the comparison shopping case.
Even worse for Google, some of these other competitors are not lower-quality. In travel and financial products, in particular, there are some spectacularly good sites offering great UX. The defence in those areas will not be as easy for Google.
What does real regulation of Google look like?
The case for some increased regulation of Google (and other tech giants) is growing. I recently wrote another article breaking down my objections to a New York Times article calling for a break-up.
I certainly don’t have all the answers, but I did get drawn into a bit of speculation about what I thought effective regulation could look like in the comments on that post. To be clear, I don’t expect to see an unbundling, but I was interested to see the same thought experiment applied to Amazon the other day.
For more on the subject, I recommend this week’s public Stratechery article and the other articles linked from it. Stratechery remains my top paywalled recommendation – easily worth the $100 / year in my opinion.