Why does eBay still use seller ratings?

On peer-to-peer platforms, trust and safety is of paramount importance. After all, they are intended as safe spaces on which certain categories of transactions can be made. But trust and safety isn’t perfect anywhere, and particularly not on marketplaces. Uber recently was in the news for one of their drivers having a criminal record without them being aware of it. Thumbtack has thorough automated background checks, but that is a defensible competitive advantage for them, rather than a de rigueur feature for the industry.

We recently met with a startup who has automated the seller fraud detection process for inventory on their site, by using machine learning on certain sets of behaviors and language that a given seller uses. In the event that a seller fails in certain among these criteria, they are prompted to take a CAPTCHA and other fraud-deterring steps, to ensure the integrity of the products that are sold on the site. Lyft has more stringent driver vetting processes than the livery industry, but they don’t have as strict a process on the passenger side. If a driver has a bad ride, the passenger is automatically blocked from that driver permanently. If it’s a bad actor, we assume they eventually kick them off, but we haven’t seen evidence of it (and, hopefully, neither have they). All of this is to say, technology can enable a bevy of automating processes which can prevent fraud, take steps to ensure user identity but also product integrity. And yet the backbone of eBay’s service is still, to date, human reviewers, not so different from Yelp.


Truth is, at least as it relates to product integrity, because they don’t want to. In 2008, eBay lost a lawsuit against LVMH, who sued them because of the sale of counterfeit goods on the eBay platform. According to details of the suit, more than 75% of the LVMH goods sold on eBay were fraud or fake. Tiffany’s has said that 80% of their goods sold on eBay are counterfeit. The pieces eBay has put in place now to create trust and safety on the platform around product integrity were done reluctantly, however. The bad goods were driving revenues on the platform, so why get in the way of that?

Interesting example of where being an “amoral” peer-to-peer platform can be dangerous, when the revenue leads the strategy…. 

Originally posted at blog.kanyi.me.