Customer Satisfaction, Advertising Competition, and Platform Performance

We study how customer satisfaction affects platforms' advertising competition and their equilibrium market shares and profits. Consider a market with multiple competing (online) platforms, all of which decide advertising outlay to maximize their own average profit. A customer chooses which platform to visit depending on her experience of the last purchase and on the platforms' advertising outlays. After purchasing from a platform, with certain probability (satisfaction probability), the customer would be satisfied. A satisfied customer is more likely to repurchase from the platform than if she is dissatisfied. There are two types of customers in the market: loyal and switcher. Conditional on either satisfaction outcome, the loyal is more likely to repurchase from the last-visited platform than the switcher. In the duopoly, we show that lower a platform's customer satisfaction or a higher proportion of switcher may alleviate the advertising competition. Consequently, a platform does not necessarily get a higher profit when its customer satisfaction is higher; and the platform with higher (lower) customer satisfaction benefits from a larger proportion of loyal (switcher). We then study the case with more than two competing platforms. Under the symmetric market, the equilibrium advertising outlay first increases and then decreases as the number of platforms increases. Every platform spends more on advertising and gets a lower profit when customer satisfaction improves. The equilibrium profit of platform is always increasing in the proportion of switcher. Under the fragmented market, we find that similar to the duopoly, the platforms with higher (lower) customer satisfaction benefit more when facing the loyal (switcher) than when facing the switcher (loyal). Several extensions are considered: two-period impact of purchasing experience, an outside option for customers, endogenous satisfaction probabilities, and asymmetric information between platforms.