- Social marketplace for short-term vacation rentals
- Fighting both fake property listings and fraudulent transactions
- Fake listings were harming customer experience
- No dedicated in-house team for handling fraud
- Automated fraud detection for catching fake listings to maintain an excellent customer experience
- Saved valuable time and money by applying machine learning to credit card fraud
- Saw useful results immediately after setup
- Constant improvements in accuracy for preventing fraud
Take this article with youDownload now
Enabling an authentic travel experience
Travelmob (acquired by HomeAway) was created to help global travelers find unique places to stay across Asia Pacific, from Bangkok to Melbourne. Hosts list their rooms and properties, and guests use the Travelmob site or mobile app to book them.
When they contacted Sift, the team at Travelmob was considering different approaches to handling the growing number of fake listings on its site that were negatively impacting the customer experience, as well as credit card fraud that was resulting in costly chargebacks.
Preserving customers' trust
As a social marketplace for travellers, Travelmob’s customer experience is key – it means a positive brand experience, repeat business, new users, and continued growth. That’s why the company was particularly worried about stopping a growing trend of bad users posing as legitimate hosts, posting photos of properties they didn’t own, and trying to con unsuspecting guests into making their payment offsite.
Jan Hecking, Principal Software Engineer at Travelmob, began evaluating options for tackling both their fake listings and credit card fraud. Without a team dedicated to handling fraud, the responsibility for all types of fraud prevention was shouldered across multiple teams.
Travelmob began by manually reviewing new listings and booking requests, but it soon became clear that this approach just wouldn’t scale – plus, without a structured process, fraud was slipping through the cracks. Travelmob realized that building dedicated internal tools for fighting fraud would require time and resources that they couldn’t spare, and that anything they created internally couldn’t adequately address the complexity of fraud.
A flexible, scalable fraud-fighting tool
After perusing Sift’s easy-to-use REST API, it only took a few hours for Jan and his team to get Sift’s machine-learning fraud solution up and running, and they were fully integrated within a week.
Travelmob soon observed that the signals Sift was using to evaluate fraud were sophisticated and useful. The Travelmob team tackles fraud efficiently by using Sift Scores to know which bookings are truly high-risk and deserve their attention for manual review. And then they dive into advanced tools and rich insights in the Sift Console – like Network Visualizations – that help them connect the dots between different users and locations so they can make smart decisions about who to block.
While the company initially used Sift to catch fake listings, that experience was so successful that they also applied the machine learning solution to their credit card fraud problem – halting chargebacks and saving them cash.
A single solution for all types of fraud
Once they were integrated with Sift, Travelmob started seeing results immediately. But Jan and the rest of the Travelmob team have been especially impressed by how Sift’s fraud detection grows increasingly accurate as Travelmob continues to send more data and feedback.
Through the Sift Events API, it was easy for the Travelmob team to record and send new data to Sift – in fact, Jan noticed that after doing this their fraud detection would grow remarkably more accurate in just two days. Also, Sift’s machine learning was incredibly flexible, allowing Jan to customize their fraud detection by sending data unique to Travelmob.
The flexibility of Sift has made it a one-stop shop for Travelmob to catch and prevent all types of fraud that threatened their business.
“It was easy to get started with Sift, and then we continuously evolved how we used it.”