Is Investment in Data Analytics Always Profitable? The Case of Third-Party-Online-Promotion Marketplace

A third-party-online-promotion marketplace allows merchants to reach out to consumers who are uninformed about merchants' offering, and charges commission from them. We consider a market that has two types of merchants who are heterogeneous in the proportion of consumers who are informed about their offering. The merchant type is private knowledge. By making an investment in data analytics capability, the marketplace receives signals generated by the system that classifies merchants as per their types, though with some probability of misclassification. The marketplace offers a contract to merchants based on the classification signal and offers a specific commission rate to each type of merchant. We refer to this as signal-based strategy. Alternatively, the marketplace adopts a pooling strategy of offering one commission rate to all merchants. We show that the relative strength and weakness of the signal-based strategy depend on the merchant type distribution and the investment cost of improving the signal accuracy rate. We find that the marketplace is better off by not investing in data analytics capability when one merchant type dominates the market. Moreover, we show that the signal-based strategy does not just lead to transfer of surplus to marketplace, but under some conditions, it could lead to an increase in surplus for not only the marketplace but also for merchants and consumers. This is so because the marketplace's signal-based strategy has a cascade effect on consumers through the merchants' optimal discount rate strategy. We conclude by identifying the conditions for a win-win-win situation