To fight fraud, Amazon now screens third-party sellers through video calls

Amazon is piloting a new system aimed at validating the identify of third-party sellers over video conferencing, the company announced on Sunday. The technology is a part of a series of seller-verification processes that Amazon uses to combat fraud on its platform, which the company claims stopped 2.5 million suspected bad actors from publishing their products to Amazon in 2019.

Earlier this year, Amazon began testing a process where seller verifications were handled in person. But due to the coronavirus outbreak and social distancing requirements, the company says it pivoted to live video conferencing in February.

The pilot program is now running in a number of markets, including the U.S., U.K., China and Japan. To date, more than 1,000 sellers have attempted to register an account through the pilot experience, Amazon says.

To vet the sellers, Amazon's team sets up a video call, then checks that the individual's ID matches the person and the documents they shared with their application. The Amazon associates also lean on third-party data sources for additional verification. In addition, the call may be used to provide the seller with information about problems with their registration and how to resolve them.

"Amazon is always innovating to improve the seller experience so honest entrepreneurs can seamlessly open a selling account and start a business, while also proactively blocking bad actors," an Amazon spokesperson said about the new initiative. "As we practice social distancing, we are testing a process that allows us to validate prospective sellers' identification via video conferencing. This pilot allows us to connect one-on-one with prospective sellers while making it even more difficult for fraudsters to hide," they said.

In addition to video conferencing, Amazon also uses a proprietary machine learning system to vet sellers before they're allowed online, it says. This system analyzes hundreds of different data point to identify potential risk, including verifying whether the account is related to another account that was previously removed from the marketplace, for example. The sellers' applications are also reviewed by trained investigators before being approved.

Seller verification is only one way Amazon has taken on fraud, however.

The issue continues to be a serious problem across online marketplaces, where sellers hawk counterfeit items and scam consumers. Some retailers, including Nike and Birkenstock, have found the hassles aren't worth the risk of dealing with Amazon, as a result.

While the retailer has long been accused of avoiding issues around fraud, it's more recently pledged to spend billions to address the problem and has inserted itself into legal battles with fraudulent sellers and counterfeiters in recent years.

For example, it filed three lawsuits in 2018 in partnership with fashion designer Vera Bradley and mobile accessories maker Otterbox over counterfeits. It has also sued sellers buying fake reviews and others involved in the fake review industry.

Last year, Amazon announced an initiative called Project Zero, which introduced a range of tools for brands to use to help Amazon fight fraud. The brands can opt to provide Amazon with their logos, trademarks and other key data, allowing the retailer to scan its billions of product listings to find suspected counterfeits more proactively.

Another tool, serialization, allows brands to include a unique code on their products during manufacturing, which can later be scanned to verify that a purchase is authentic. This tool, now known as Transparency, expanded to other markets last summer, including Europe, Canada and India.

But unlike these earlier efforts, seller verification aims to cut down on products being listed in the first place -not just removed once listings go live or stopping fraudulent products from being shipped to customers.