How EBTH achieved a streamlined fraud workflow and delightful user experience
0
Fraud analysts on-site
0
Extra security checks
OVERVIEW
Shop estate sales from anywhere in the world
Everything But the House (EBTH) brings the thrill of estate sale shopping online. Through a global marketplace and community of buyers and sellers, EBTH has transformed the traditional estate sale model while preserving its fundamental charm: the excitement of discovering something special.Featuring 150 sales a month and shipping to anywhere in the world, EBTH’s full service model—photography, cataloging, payment and delivery—makes downsizing easy for sellers. For buyers, the reach of an e-commerce platform and the starting bid of $1 for all items—from a purse to a Picasso—means there’s always something new to discover.
CHALLENGE
Preserving bidders’ trust
EBTH has always maintained a fairly low chargeback rate. Instead of looking to reduce their fees, the company was mainly concerned with the effect that fraudulent bids would have on the user experience, bidders’ trust in their service, and in the overall value of items sold on their website.
Examples of fraud included when users would bid on EBTH using stolen credit card info or without any real plans to complete their purchase. Bobby Uhlenbrock, Chief Technology Officer, explained that if a fraudster won an auction, they would have to relist the item, which would delay profits for the seller. As a result, the item might sell for a lower price the second time. If fraud continued, customers could start questioning the integrity of the site.
When EBTH first contacted Sift, weeding out these fraudulent bids was one of the responsibilities of a 10-person customer service team. They were using a tool that sent identifiable information about bidders to its servers, but there were two problems. First, the tool didn’t offer any proactive notifications, so employees would have to monitor far more auctions than necessary to pinpoint the small number with actual fraud. Moreover, the tool was reactionary, meaning that by the time it spotted fraud it was too late. So Bobby started investigating solutions that would detect and prevent fraud, instead of merely reporting it after the fact.
SOLUTION
A proactive, accurate fraud tool
One of the main selling points of Sift was its transparent, pay-as-you-go pricing, which didn’t lock EBTH into a rigid contract. The EBTH team was impressed by how simple it was to set up Sift and start getting useful results.
Now, EBTH uses Sift exclusively to manage their fraud review process. Using Sift’s webhooks, the customer service team gets proactively notified when the machine learning technology detects someone suspicious, so they can decide whether they want to investigate further.
In addition to webhooks, there are a couple of other features in the Sift Console that Bobby says makes the team’s process particularly smooth and efficient. One is the Developer tab, which shows all the events that have been sent from EBTH to Sift, allowing developers to dive in, understand the data, and decide what tweaks to make.
They also love the ability to easily set up a deep link between the Sift Console and their internal dashboard. This lets EBTH easily switch between reviewing a user in the Sift Console to looking at them in their internal pages, with a single click.
RESULTS
Preserving user experience and peace of mind
Since integrating Sift, EBTH has been able to streamline their fraud prevention workflow and dedicate the time they formerly spent reviewing false positives to other initiatives. Since Sift provides the customer service team with such useful, actionable data, EBTH hasn’t needed to add any employees who focus solely on fraud prevention.
Sift provides such high accuracy that EBTH has been able to tackle fraud without introducing steps that could potentially deter trusted bidders. Previously, EBTH had considered adding security checks to the bidding process so they could be more confident about a bidder’s identity and intentions. But thanks to Sift, those extra measures aren’t necessary.
