Data-driven Platform Envelopment with Privacy-Policy Tying

Digital platforms may absorb competitors through envelopment acquisitions, to build ecosystems. Daniele Condorelli discusses how some recent envelopment attacks may result from the desire to acquire valuable sources of data rather than products to maintain a position of market power.

*Daniele Condorelli is an associate professor of economics at the University of Warwick.*

Big-tech firms, Google and Facebook featuring prominently among them, have not only fostered the development of large ecosystems of firms producing many interoperable products, but also enormously expanded their own scope, through entry and acquisitions.

Platform envelopment is a theory that explains such growth and expansion in scope (see Eisenman, Parker and Val Alstyne, 2011). A platform that is dominant in its own market performs an envelopment attack when it enters a different platform market and, by leveraging its large user base and engaging in product-tying, it exploits those network effects that would have otherwise protected the incumbent in that second market. Platform envelopment is especially compelling to explain entry and acquisitions in markets for products that are complementary or belong to the same supply chain. In fact, to the extent that firms selling products that are bought in conjunction with those of the dominant firm build market power, they necessarily erode the power of the dominant firm. There is, after all, only one monopoly profit to be made in these cases.

Data-driven Envelopment

Especially in recent years, we have witnessed a different form of platform envelopment, with entry taking place often in markets unrelated to the primary one. While there might be several reasons for conglomeration (see Bourreau and de Streel, 2019, Condorelli and Padilla, 2020a), in research work with Jorge Padilla we suggest that some of these recent envelopment attacks may be explained by an intention to acquire valuable complementary sources of data (see Condorelli and Padilla, 2020b). There is tying of data, not necessarily tying of products. This motivation may be especially relevant for advertising platforms, which rely on data to provide services to their user-side and targeting capabilities to advertisers. To name a few moves that may be data-driven, consider Facebook’s acquisitions of Instagram and WhatsApp, or Google’s acquisitions of YouTube, DoubleClick and AdMob; Google’s entry into comparison shopping and the tying of Android with Google Search and Google Play Store.

Privacy-policy Tying

When data-driven envelopment takes place, data combination is normally implemented by means of a practice we call privacy-policy tying, which consists in conditioning the provision of services to the subscription of a privacy-policy that allows bundling of user data across unrelated services. Most conglomerate platforms appear to engage in privacy-policy tying. To give an idea, Facebook officially collects information from

[…] Facebook (including the Facebook mobile app and in-app browser), Messenger, Instagram (including apps like Direct and Boomerang), […] and any other features, apps, technologies, software, products, or services offered by Facebook Inc. or Facebook Ireland Limited under our Data Policy. The Facebook Products also include Facebook Business Tools, which are tools used by website owners and publishers, app developers, business partners (including advertisers) and their customers to support business services and exchange information with Facebook, such as social plugins (like the “Like” or “Share” button) and our SDKs and APIs.
Excerpt from Facebook’s privacy-policy.

Data advantage and foreclosure

While it is conceivable that bundling of user data may produce efficiencies and benefit consumers, in our research we explain how data-driven envelopment by means of privacy-policy tying can also have exclusionary effects, and entrench the monopoly power of dominant platforms. In particular, through privacy-policy tying, the dominant platform leverages its market power across markets, in such a way as to create an insurmountable data advantage that cannot be replicated by competitors, which may lead to their exit or prevent their entry in the primary market.

Regardless of whether or not consumers benefit from data combination in the short run (e.g. if the improvement in the quality of advertising more than compensates for the loss of privacy and the potential for harmful price discrimination) if the end result is monopoly, then consumers are invariably worse off. Ultimately, this happens because consumers are unable to internalise their future losses and object to privacy-policy tying. First, as many researchers have showed, consumers pay little attention to privacy settings despite actually attaching high value to their privacy (see for example Economides and Lianos, 2019). Second, to make predictions for all consumers, platforms only need to collect data about a representative sample of them. This generates a classic externality that is hard to internalize when accepting a privacy policy (see Acemoglu, Makhdoumi, Malekian and Ozdaglar, 2019 and Bergemann, Bonatti and Gan, 2019).

Remedies

While, of course, foreclosure of rivals is not always the goal of entry or acquisitions, nor will it always take place when data are combined, it makes sense for regulators to remain vigilant, both during merger review and when looking at data combination practices by dominant firms. Arguably, such practices could in some cases be treated as an abuse of dominance, especially when services are provided conditional on consumers consenting to allow use indiscriminate use of their data. One natural remedy consists in preventing firms who have dominant position in some data-intensive markets from combining data, especially for advertising purposes, unless it can be proved that such combination generates substantial direct efficiencies for consumers.

Instead, as we argue in our paper, we think measures such as making data tradeable and shareable by firms, or even enforcing portability, may end up backfiring. To the extent that frictions to acquire data are reduced for the dominant firm, these measures may actually facilitate the construction of an insurmountable data advantage, which is harmful to consumers if it results in exclusion or entrenchment of monopoly.

Daniele Condorelli is an invited speaker at the panel on Competition Economics in Digital Ecosystems held 24 February 2021 during the OECD Virtual Competition Open Day. Register here to join us for this discussion and find more materials on the topic here.