Assessing and Quantifying Local Network Effects in an Online Dating Market

Abstract

We empirically examine and quantify network effects on a large online dating platform in Brazil. We consider the effects of a seeding intervention by the platform operator, wherein it acquired its primary competitor and subsequently imported the competitor’s 150,000 user accounts over a 3-day period. The acquisition thus constitutes a large exogenous shock the composition of the acquiring platform’s user base. We estimate the effect of the shock on the rate of subsequent enrollments and exits amongst heterosexual users across 120 cities. Bearing in mind that the purchased users were exclusively heterosexual, we employ a difference-in-differences specification in which homosexual enrollment and exit rates serve as plausible controls. Our estimates indicate that the treatment increased the rates of both enrollment and exit, for both genders, with a net positive effect that translated to a 22% increase in short-term revenue for the platform. Further, we find that the response amongst male users was significantly stronger. When we consider that female participation was being fully subsidized by the acquiring platform, this result is consistent with the idea that subsidies and seeding strategies are substitutes, rather than complements. Finally, we explore nuances of the observed effects, quantifying local features. In particular, we show that the treatment effect varied significantly, depending on age differences and the degree of co-location between new and existing users.