Learning-by-doing in Data Markets

Abstract

Data brokers learn from the large amounts of data they treat, lower their operational costs and gain efficiency, improving their ability to supply information to industries. When competing firms can buy consumer-related information from those brokers, we derive the implications of learning-by-doing from data on the firms’ and brokers’ profits. We show that learning effects can harm a data broker when firms anticipate how providing their data lowers the treatment cost of the broker, allowing it to sell information to rival firms. This strategic anticipation can deter firms from buying information from the broker, if doing so implies that competitors also remain uninformed, and all the more so when the broker enjoys a strong learning effect. When brokers compete, a broker with weak learning effects can dominate its competitors, and even achieve monopoly profits when competing with more efficient brokers.