The Slippery Slope of a Damaged Reputation
Cabral, Luís; Hortaçsu, A.
Publisher: Blackwell
Original document: The dynamics of seller reputation: Evidence from eBay
Year: 2010
Language: English
If you want to sell a laptop or get rid of a coin collection, chances are you’ll find a buyer on eBay. But if you are a regular eBay seller with an impeccable reputation, beware of your first disgruntled customer. Even for sellers with a long and perfect track record, negative feedback affects sales. Negative comments also tend to snowball after the first bad rating.
In “The Dynamics of Seller Reputation: Evidence from eBay,” Luis Cabral and Ali Hortaçsu, professors of Economics at IESE and the University of Chicago respectively, explain just how much reputation matters.
The authors tracked the eBay auctions of IBM ThinkPad notebook computers, collectible coins and Beanie Babies stuffed toys over a six-month period, studying how feedback affected selling patterns. They analyzed the impact of seller reputation, both on the buyer’s and seller’s behavior, and observed when and why a seller exited trading altogether.
The authors discovered that the first negative comment has significant impact on sales. For example, the average ThinkPad seller experiences a positive growth rate of 5 percent per week. However, this growth rate drops by 13 percent to a negative 8 percent after the seller receives the first negative feedback.
Moreover, the first negative feedback creates a snowball effect, with subsequent negative comments arriving 25 percent more rapidly than the first comment. The average ThinkPad seller, for instance, makes 93 sales before the first negative comment is received, but only 58 before the second negative comment arrives.
Negative Feedback Builds Momentum
The momentum of negative comments can be attributed to changes in buyer or seller behavior. Is it what’s known as herding, or conformist, behavior – where a buyer, after reading one negative comment, scrutinizes the seller more closely and is biased by a fellow buyer’s experience? Or does a seller become demoralized by the loss of reputation after having a previously immaculate record tainted by a bad comment?
The authors explore several possibilities on the buyer’s side, including the herding phenomenon. They review complaints that are subjective (e.g., “bad communication,” which is a matter of opinion and may be beyond the seller’s control) as well as complaints that are objective (e.g., “item never sent,” which would seem to implicate the seller more directly). However, they find no empirical evidence of any significant change in buyer feedback behavior following the first negative comment.
Hence, though the first negative comment may or may not be warranted by seller behavior, the authors find that the seller’s behavior backslides. In other words, the seller now feels less incentive to invest as much effort in future transactions – a case of moral hazard.
To illustrate this phenomenon, the authors cite one particular seller who had 22,755 positive comments but 11 negatives in the last 25 transactions. It is difficult to recover from this sort of downward spiral. The worse his record is, the more typical that a seller will exit eBay altogether.
Driving Away the Less Than Perfect?
Since 1996, eBay founder Pierre Omidyar has encouraged an open market with honest dealings through the feedback forum. In a note discussing the importance of feedback, he once explained that, while most people were honest, for those who were not, they could not hide on eBay. In fact, he said, “We’ll drive them away.”
Though this reputational mechanism was conceived with laudable motives, it does seem to be doing a bit of overkill. As the authors discovered, the behavioral reaction to the reception of negative feedback starts the seller down a slippery slope. From there, it is difficult for them to regain their footing and remain in the game.
The eBay honor system has done well in establishing the trust among strangers that makes it possible to generate billions in revenue. But there is always room for improvement. A good area for future research, say the authors, would be in creating a more efficient system, one that drives away the untrustworthy without inadvertently pushing out the less than perfect.
Top 