Sift > Case Studies > Creating a successful fraud solution from the ground up

Creating a successful fraud solution from the ground up

Company Size

11-200

Region

America

Country

United States

Product

Sift Formulas

Tech Stack

Machine Learning

Implementation Scale

Enterprise-wide Deployment

Impact Metrics

Cost Savings
Productivity Improvements

Technology Category

Analytics & Modeling - Machine Learning

Applicable Industries

Retail

Applicable Functions

Sales & Marketing

Use Cases

Fraud Detection

Services

Data Science Services

About The Customer

Wanelo is a shopping app built to connect people with merchants. Through the mobile-focused marketplace, consumers can connect, discover, and buy millions of fashion and lifestyle products directly from global sellers. Wanelo is where Generation Z shops, providing a unique shopping experience that is as much about community and conversation as it is about buying. What began as a social shopping site evolved into a marketplace last year and has seen the number of sellers grow 5x since its inception. The company is headquartered in San Francisco with 90% of its user base in the U.S., but Wanelo also has remote teams globally to support its marketplace around the clock. Currently, fraud falls under the Marketplace Operations team, which executes all manual order review and order disputes.

The Challenge

Wanelo, a global marketplace for Gen Z, was facing a significant challenge with spammers and scammers. As the company evolved from a social shopping site to a marketplace, it started experiencing payments fraud. The fraud showed up in the form of disputes, with both friendly and scammy customers demanding 'charge not authorized' chargebacks. Nearly 70% of their chargebacks could be attributed to friendly fraud, which was a unique challenge to address because such customers often look like good and valuable users – until they decide that they don’t want to pay. Wanelo’s job then was to convince the bank that the customer is committing chargeback fraud. As more fraudsters attempted bad activity and Wanelo’s chargeback rate crept up to 0.87% – including friendly fraud – Courtney Bode, Marketplace Operations Manager, turned to the system that had worked so effectively for the social side of the company.

The Solution

Courtney decided to apply Sift’s machine learning solution to their new challenge. With the launch of the Sift Formulas feature, the Wanelo team adopted this automation tool and used it as the foundation of their fraud prevention system. As existing Sift users, Wanelo turned to their Sift Account Manager to assist with reshaping their business needs of the solution. In about one week, a pair of engineers fully integrated the additional APIs necessary to connect Sift Formulas with Wanelo’s internal order management system. After training with Sift’s Solution Engineers and overhauling their label history, Wanelo was able to immediately see useful and reliable Sift Scores.

Operational Impact

The Wanelo fraud disputes team uses Sift on a daily basis.
Sift Formulas, Wanelo’s favorite feature, allows the fraud prevention team the ability to create and manage automation without needing engineering resources.
While Wanelo started with no experience in fraud, their workflows are now seamless and their chargeback rate is exceptionally low.
They can easily weed out the malicious fraudsters based on Sift Scores — and have done so successfully without a jump in false positives.
In fact, the Wanelo team has begun to leverage Sift’s findings in their analysis of suspicious users contributing to a 52% reduction in order decline rate.

Quantitative Benefit