Layer 1 copy

As a startup, there are opportunity costs everywhere you look. Using Sift Science allows us to not waste resources on fraud. Instead, we can focus on building the best product possible for our diners.

Ian LaveyDirector of Operations

Overview

  • On-demand food-ordering platform

  • Growing rapidly into new cities

Challenge

  • Expanding into new markets introduced fraud

  • Battling chargebacks and users gaming their rewards program

Solution

  • Automation to streamline the fraud review process

  • Lowering chargebacks with real-time machine learning

Results

  • 85% reduction in chargeback rate

  • 8-9x return on investment

Overview

Simple Ordering for Restaurants and Diners

EatStreet is an online food-ordering platform that empowers restaurants to easily accept orders via web, mobile, or social media. Founded as a local service for Madison, Wisconsin in 2010, EatStreet now works with more than 15,000 restaurants across 49 states.

In addition to being listed on the EatStreet website, restaurants using the service can benefit from a custom website, pay-for-performance marketing, Facebook integration, and customer support for their online ordering. Diners can easily see which local restaurants offer takeout and delivery, order for free, and get rewarded with coupons.

Challenge

New Markets Invite New Fraud Challenges

During its first four years of business, EatStreet rarely suffered from fraud. However, when the company started expanding aggressively in 2014, a troubling trend emerged. More markets meant more fraud – and more chargebacks. Some fraudsters were using stolen credit card information to place orders, and others were gaming the site’s rewards program by using coupons and then canceling orders.

The fees started to add up, and EatStreet’s Director of Operations, Ian Lavey, began looking for a tool that would apply intelligent fraud-prevention barriers, without turning away legitimate diners.

“The last thing we wanted to do was make it more difficult to order. We wanted a solution that would combat fraud but wouldn’t hurt the experience for our diners.”

EatStreet assigned a single employee, Ashley Fueger, to dedicate part of her day to analyzing new orders, with some customer support members helping out by gathering more information as necessary. However, they realized that the on-demand nature of their business model – people expect their food to arrive within 30-60 minutes – meant that there was very little time to manually review orders. As their fraud problem became more pressing, Ashley discovered and championed Sift Science for a real-time, automated fraud solution.

Solution

Real-Time Fraud Prevention

EatStreet used Sift Science’s simple, straightforward API documentation to get up and running, and the real-time machine learning technology quickly began stopping fraudsters. EatStreet’s chargeback rate decreased by 70% within the first month. And by the second month, the results were even stronger: 85% lower than before they’d implemented Sift. EatStreet’s chargeback rate has stayed at that same low level ever since.

Sift Science’s ability to support automation helped Ian and his team streamline their fraud workflow, relying on Sift Science’s machine learning to auto-ban users. Then, they use the Sift Science Console to review flagged orders, looking at each one’s Sift Score, getting insights on why they’ve been flagged, and labeling users to make the machine learning model more accurate. EatStreet uses Sift Science as a one-stop-shop for order review.

“The fact that we haven’t had to scale up an entire fraud department has been a huge advantage. Now we can focus on creating a better experience for diners instead of fighting fraud.”

Results

8-9x Return on Investment

EatStreet’s dramatic improvement in chargeback rate – a 85% reduction with Sift Science – speaks for itself. Since implementing Sift Science, EatStreet has seen an 8-9x return on investment, compared with what they project they would be spending on fraud without Sift if early trends had continued.

Automating fraud prevention also means the team can work smarter and focus on improving their core product. Now, the employee who handles fraud can spend much less time reviewing individual orders and can dedicate her time to growing EatStreet’s business.

“As a startup, there’s opportunity costs everywhere you look. Using Sift Science allows us to not waste resources on fraud. Instead, we can focus on building the best product possible for our diners.”