Bloomberg — 李玲芳副教授最新研究报道：For Better Online Product Reviews, Pay the Reviewers

Many online product reviews aren't especially helpful. Reviews can be manipulated, biased, or even intentionally ridiculous, so it's hard to judge their accuracy. The online review system also puts newer merchants at a disadvantage because they begin with no reviews at all. The Chinese auction site Taobao has an unusual solution to these problems: It lets merchants pay customers to leave detailed transaction feedback.

That's a risky-sounding strategy. Certainly, if merchants were allowed to reward customers for giving glowing reviews, the feedback system would quickly become uninformative. Likewise, allowing merchants to penalize negative reviews would lead to bias, as well as fiascos. It's already so easy to buy positive reviews that mischief-makers have been able to drive traffic to a nonexistent business. But Taobao's system does not let merchants choose which customers to reward - it only lets them choose whether to reward them.

Under Taobao's "rebate-for-feedback" mechanism, merchants can commit to pay rebates to all customers who write detailed reviews - positive or negative. (The average cash rebate amounts to about 20 cents.) Commitments are listed publicly, so potential customers can see when they are available.

And it seems to work. A new study by economists Lingfang Li, Steven Tadelis and Xiaolan Zhou suggests that the biggest benefit of Taobao's system is that it can help high-quality newcomers gain footholds.

High-quality merchants have the most to gain by offering rewards for feedback: If you're Arnold's Drive-In (with the world's best jukebox), then you want to encourage reviews - your customers leave happy every day. But if you're more like the dysfunctional Dunder Mifflin Paper Company, then you definitely do not want to encourage reviews (and might even wish you could pay your customers not to write them)!

Because encouraging reviews helps high-quality merchants the most, those who choose to offer rewards for feedback are more likely to be high-quality. Consequently, if you're a merchant entering the market without any reviews at all, then you can signal high quality by offering rewards for feedback. Customers will flock to you because they can see that you expect their experiences to be positive. (And if instead you happen to be a low-quality merchant offering rewards for reviews so as to pose as high-quality, you will rapidly buy yourself lots of negative reviews.)

If online platforms were to allow merchants to promise rewards for reviews, the overall review pool would expand and new entrants might find it easier to get started. But of course, there are caveats.

One is that allowing rewards for reviews might increase the power of established merchants, who have more resources to support review buying. But this seems unlikely to overshadow the signaling value for new entrants. Once a merchant has acquired a number of ratings, it has less need for more of them - and thus has less willingness to pay for them. And indeed, Li, Tadelis and Zhou found that established merchants were less likely than entrants to offer rebates for feedback. (Moreover, there is some evidence that online reviews do not significantly affect traffic to especially well-known merchants one way or another.)

Customers are increasingly choosing merchants based on ratings, and established players already have lots of rating information available. That gives incumbents an advantage that a Taobao-style reward-for-feedback system might begin to correct.

Scott Duke Kominers is a junior fellow in economics at the Harvard University Society of Fellows.

Scott Duke Kominers