Google, Facebook, and Amazon are no platforms

EDIT: a more academic version of this article is now available
at the Oxford Journal of European Competition Law & Practice: link.

It’s been a while since I wanted to write about Ben Thompson’s aggregation theory because it has significant implications for antitrust law. Today is the day I’m finally doing it. Yay.

1- Platforms vs. aggregators

Before introducing the aggregation theory, I first need to explain why all tech giants are not platforms.

According to the Historical Larousse dictionary, the term platform first appears in the French language in 1434. It is used to refer to a “horizontal surface acting as a support.” One finds the idea of being a surface on top of which to build something in today’s Oxford Dictionary, which defines a platform as “a raised level surface, for example, one that equipment stands on or is operated from.” The second part of the definition completes the historical one: a platform can be used for the construction of a product or service, but also its operation.

By contrast, aggregators cannot be used for the purpose of building or operating anything. They consolidate all kinds of existing information (including products and services) and sort it out. Contrary to platforms, aggregators have (i) a direct relationship with the customer; (ii) zero marginal costs; (iii) and they are lowering customer acquisition costs as it scales.

2- And so what?

It follows that not all tech giants are platforms… at least not generally speaking. Of course, tech companies are often operating in a wide range of markets, and generalizations about them are usually wrong. To be more accurate, I should have named this article “Google, Facebook, and Amazon are not operating platforms for the most part.

It remains that Google’s search service, Facebook’s social network, and Amazon’s shopping site are not platforms, but aggregators. The same goes for Netflix, YouTube, Uber, Airbnb, and Booking. Conversely, Google Android and Facebook Business Tools serve as platforms. And of course, Microsoft and Apple are also great examples of companies operating a platform as developers build their apps and software based on the APIs and tools they provide.

3- Market power and practices

What is the purpose of the distinction between platforms and aggregators? Is it just a question of semantics? Well… no, it is not.

In terms of the relevant market, one may question whether a platform and an aggregator can compete. The general answer seems negative as they do not operate the same function. They are not substitutable, and they don’t even derive their revenues from the same source.

The distinction is also useful for market power analysis. More specifically, the platforms’ ability to abuse their power differs significantly from aggregators. It all comes down to the following: platforms can directly impact the product and services built on top of them, while aggregators only have a direct impact on how they display products and services.

It follows that the risk of exclusionary practices is much higher with platforms than it is with aggregators. Platforms can indeed exclude competitors using two methods: the contractual one, and the technical one (by altering the code and operation of third-party products and services). By contrast, aggregators can only exclude competitors by way of contract. For that reason, one may be surprised to find out that antitrust agencies have been solely interested in contractual exclusions. Predatory innovation – which I defined as the “alteration of one or more technical elements of a product to limit or eliminate competition” – is still under the radar. That’s what I call a significant legal loophole!

Conversely, aggregators are more likely to engage in exploitative practices. Content (products and services) producers may need to gain access to aggregators of critical size. These aggregators can use their power to enter abusive agreements. Not surprisingly, the European Commission has launched several investigations regarding such practices. In the United States, exploitative abuses per se legality has led agencies to bend existing rules to tackle related practices.

Finally, one could underline the dangers that exist when the same company controls both a platform and an aggregator in two related markets. For example, Amazon Web Services is a platform, while is an aggregator. Similarly, Apple OS and iOS are platforms, while the App Store is an aggregator. It confers dual market power. One may expect antitrust agencies to be interested in these situations. But one question still remains: what about consumers?

4- What about the consumer in all that?

The more the years go by, the more the debate on tech giants is reaching a dead-end. Or two.

Here’s the first. On the one hand, tech companies stress the overall benefit they bring to consumers (which I’d like to qualify as a macro-analysis). Platforms insist on the wide range of products they bring to consumers (choice), and aggregators emphasize how much more comfortable they make the life of their users in a world where information has become ultra-abundant (ease). When they are combined, it improves the consumer’s experience on both ends.

On the other hand, antitrust agencies insist on the existence of exclusionary or exploitative practices (which I qualify as micro-analysis). This double-level is the first dead-end: companies and agencies are engaged in two distinct dialogues, i.e., one macro and one micro. While they both stick to their level of analysis (at least for now), part of the doctrine has started calling for macro-measures based on micro-findings. That’s Lina Khan’s Amazon Antitrust Paradox: (alleged) dominance + (alleged) anticompetitive practice = let’s reform antitrust from the ground. That methodology surprises me, as I underlined last year in an article entitled “Antitrust Without Romance.”

The European Commission is creating a second dead-end. Regardless of whether it investigates practices implemented by aggregators or platforms, DG Comp infers damage to the consumer from a reduction in the choice being offered. In the Google Shopping case, for example, Margrethe Vestager argued that Google “denied European consumers a genuine choice of services and the full benefits of innovation.

Without even discussing the “paradox of choice” (see this article), let us recall that only platforms are primarily in the business of choice; aggregators are in the business of making life easier for their users (precisely by reducing choice). It follows that when it comes to aggregators, harm to consumers actually lies in making things more complex. Antitrust agencies could, for example, insist that the practice has eliminated information from the aggregator, complicating the life of the consumer willing to find more accurate information. By insisting on choice reduction in cases involving aggregators, antitrust agencies are preventing them from offering an adequate defense as they are being accused of something fair competition would not bring in the first place (namely, a choice increase).

In the end, the distinction between platform and aggregator is useful throughout the entire antitrust analysis: to determine the relevant market, to detect practices, and to analyze them. Needless to say, the distinction is also essential for the design of remedies. Let us hope that it will soon be integrated into antitrust rulings. I have myself long made the mistake of calling all tech giants “platforms.” I won’t do it again.

PS: I have nothing to disclose! Plus, this article is not
arguing for less enforcement, or more, but for a different one.

Dr. Thibault Schrepel


Citation: Thibault Schrepel, Google, Facebook, and Amazon are no platforms, CONCURRENTIALISTE (July 27, 2020)

Related Posts