HKUST Business Review

39 HKUST Business Review In other words, when sellers already have access to reviews, additional visibility into highly granular customer data gives them so much information that platforms opt to reduce access to customer data and increase prices to account for sellers’ ability to price more opportunistically. As a result, manufacturers also raise their prices and are discouraged from investing in quality. This, in turn, means that customers end up buying lower-quality products at higher prices. For example, imagine that Acme Inc. sells widgets on Amazon. If Amazon provides Acme with some customer data, Acme can use that data to build widgets with the features that customers seem to like the best, and they can also adjust their wholesale prices to reflect their customers’ price sensitivities, thus boosting both their own profits and the quality of the widgets that platforms and customers receive. But if Acme has access to much more extensive customer data, then they will be able to adjust their wholesale prices much more opportunistically, for example by increasing prices in response to positive reviews or decreasing them in response to negative ones, which cuts into Amazon’s margins. Anticipating this effect, Amazon may choose to reduce the amount or quality of the customer data it shares with Acme, limiting Acme’s ability to improve its product quality and pricing in response to data. To be sure, this should not be misconstrued as suggesting that transparency is always harmful. However, it does illustrate the pitfalls of excessive transparency. Platforms that publish product reviews may think of those reviews as largely benefiting customers, but they also function as an information source for sellers, helping them to estimate demand. As a result, the additional benefit of getting private platform data diminishes, while the potential for manufacturers to renegotiate prices based on that data grows. In other words, from the platform’s viewpoint, sharing additional customer information becomes mostly downside, so the platform is incentivized to reduce access to that information. Unfortunately, without that data, manufacturers face greater uncertainty about customer preferences, which they hedge against by lowering production costs (often by lowering quality) and increasing wholesale prices. 1 Mitigating the Risk of Too Much Transparency In most cases, it’s not realistic or advisable for platforms to limit manufacturers’ visibility into customer reviews. After all, as Bezos has aptly noted, these reviews are the lifeblood of many platforms — and there’s not generally a way to ensure that only customers (but not sellers) can see them. That said, there are two levers that platforms can pull to navigate these dynamics more effectively: Shade customer information First and foremost, my analysis demonstrates that more data is not always better. To the contrary, sharing slightly less accurate or less frequent data through a strategy I call “information shading” can increase product quality and overall efficiency. This involves intentionally limiting the precision, timeliness, or scope of the data shared with partners. For instance, a platform might provide weekly rather than daily performance metrics, aggregate data by region instead of city or zip code, or delay access to certain conversion rate analytics. Real-world platforms are already exploring various approaches to shading customer information. Alibaba’s Taobao and Tmall, for example, offer tiered analytics packages, where sellers have to pay more for more-granular data. JD.com only shares traffic and conversion reports and keeps deeper behavioral analytics private, offering suppliers enough insight to improve products while maintaining its own strategic position.