Layer 1 copy

After reviewing several services, Sift had the best solutions in terms of automation, machine learning, and extensive integration documentation.

Brandon RobbinsProduct Manager

Overview

  • Native commerce platform for online publishers, communities, and brands

  • 200 million monthly users across 750+ sites

Challenge

  • Battling credit card fraud on digital goods

  • Legacy rules-based system let fraud through and couldn't scale

Solution

  • Machine learning blocks fraud that their old solution couldn't catch

  • Sift Science Formulas & Actions to automate fraud management

Results

  • 25% drop in chargeback rate, saving $2,000 per month

  • While orders have grown by 30%, manual review time has not increased

  • Almost 5X ROI with Sift Science

  • Estimated 25% decrease in fraudulent chargebacks using Formulas & Actions

Overview

Creating new revenue streams for publishers

StackCommerce is the leading native commerce platform for online publishers, communities, and brands. They power deal stores for the world's top tech and lifestyle publishers by offering curated product recommendations tailored to each client's audience.

A fast-growing business in a thriving market, StackCommerce has more than 1,500 vendors offering products and services to over 200 million monthly users across more than 750 publishers' websites. As part of their service, StackCommerce handles fraud management for any orders placed on their platform.

Challenge

Rules don't scale, and fraud slips through

The main type of fraud StackCommerce deals with involves purchases made using stolen credit cards – and the most time-consuming and impactful type of fraud comes from the loss of digital goods that are distributed instantly. When this fraud occurs, it not only hurts cardholders, but also the merchants. StackCommerce needed to stop these transactions as quickly as possible, and they sought a solution that could prevent them in the first place.

Before Sift Science, StackCommerce was using a legacy, rules-based solution that didn't include any machine learning. As the company's order volume grew, they discovered the shortcomings of rules-based systems: they don't learn and they don't scale. The team found themselves reviewing hundreds – or even thousands – of orders per day, and fraud review became unmanageable. As a response to the increasing volume, StackCommerce began mass approving orders, which in turn increased disputes. There were times when their support queue was so backed up, they'd have to spend a day or more getting caught up.

Solution

Machine learning offers accuracy and efficiency

StackCommerce began looking for a tool they could confidently rely on to prevent fraud, and which also had automation capabilities. After extensive online research – and a recommendation by their payment gateway, Stripe – they landed on Sift Science.

The interface and APIs are extremely intuitive. I've now implemented several 3rd party services and Sift Science's API and integration / web interface is by far one of the best I have seen.

Using Sift Science's extensive online documentation, they were able to get up and running in less than two weeks. The team saw accurate results immediately, but the results were even more striking after they trained their machine model by labeling users.

The global model worked pretty well for us out of the box, but after 30 days, we saw bad users that were consistently getting through our old system being blocked (scored high) by Sift.

The StackCommerce team uses Lists to efficiently manage their fraud review process, making instant decisions or flagging orders for additional verification. They also use Sift Sciences automation tools – Formulas and Actions – to automation fraud decisions, saving even more of the team's precious time.

Results

Saving valuable money and time

With Sift Science, StackCommerce has reduced their chargeback loss rate by 25%, saving more than $2,000 per month on chargeback fees. Not only that, but despite a 30% increase in monthly order volume since implementing Sift Science, the StackCommerce team hasn't had to hire additional staff to manage fraud. In fact, they are now down to a single employee spending no more than two hours per day on manual review.

We saw useful results right away. Immediately, I was able to control and easily track the number of orders we were manually reviewing.

The insights provided by the Sift Science console have also helped StackCommerce learn more about what types of deals attract fraudsters. The Formulas and Actions integration has led to even more accurate and powerful fraud prevention, resulting in a 25% decrease in fraudulent chargebacks so far. Compared with the siloed nature of the rules they previously relied on, StackCommerce now leverages Sift Science's machine learning technology to accurately visualize the links between related users, using thousands of potential fraud signals to predict and prevent fraud.