Data Brokers Co-Opetition

Abstract

Data brokers collect, manage, and sell customer data. We propose a simple model, in which data brokers sell data to downstream firms. We characterise the optimal strategy of data brokers and highlight the role played by the data structure for co-opetition. If data are "sub-additive", with the combined value lower than the sum of the values of the two datasets, data brokers share data and sell them jointly. When data are "additive" or "supra-additive", with the combined value equal to or greater than the sum of the two datasets, data brokers compete. Results are robust to several extensions.

Competing data intermediaries

“…Lemma A1 generalizes Proposition 18 ofBergemann, Bonatti, and Gan (2019) in that the equilibrium payoff profile in the downstream market is shown to be unique even if D k ⊂ D j for some k and j = k Gu, Madio, and Reggiani (2019). assume K = 2 and consider not only submodularity but also supermodularity.C The RAND Corporation 2021.…”

“…2 Recent work also studies how firms monetize data. Gu, Madio, and Reggiani (2019) study data brokers' incentives to merge data. I abstract away from contracting between intermediaries but consider data collection in the upstream market.…”

“…In practice there seem to be counterexamples-for example, online advertising intermediaries sell the same targeting data to advertisers via their ad networks. In this respect, we should interpret Proposition 1 as an approximation: As in models of product differentiation, intermediaries will have an incentive to differentiate themselves in terms of what data they hold (Gartner, 2016;Madio, Gu, and Reggiani, 2019;Gu, Madio, and Reggiani, 2019). 11 The incentive to differentiate and the nonrivlary of data discourage intermediaries from paying for data their rivals already hold.…”

“…Although empirical evidence on data markets is sparse, an industry expertGartner (2016) describes that "data brokers tend to specialize in certain industries in order to gain a competitive advantage. "Gu, Madio, and Reggiani (2019) describe that different data brokers may be particularly strong in collecting different kinds of consumer information.12 If the consumer chooses one of the "carriers" that will make take it or leave it offers to sell data to intermediaries, competing carriers offer to pay to the consumer in equilibrium. The consumer's single homing is natural if carriers apply compensation as discounts to positive net prices of their services.C The RAND Corporation 2021.…”

Plateformes numériques, algorithmes et discrimination

“…Les opérateurs peuvent également acquérir des données auprès de tiers, réseaux sociaux ou de data brokers, c'est-à-dire de courtiers en données (McSweeny et O'Dea, 2017). De surcroît les plateformes peuvent combiner les données issues du monde en ligne avec des données relatives au comportement horsligne des consommateurs, par exemple à partir d'accords avec des grands distributeurs ou des acteurs du secteur des cartes bancaires (Gu et al, 2019).…”