Stopping fraud, not legitimate customers
Steven Ou, Touch of Modern’s CTO, always strived for a low fraud rate. As a young company, Touch of Modern suffered very little fraud and so it was simpler for their team to just pay the costs of any chargebacks incurred and move on. Unfortunately, this system was neither practical nor scalable as the company grew, and fraud only continued to increase over time.
Touch of Modern’s customer service team has a very limited window to stop fraud—just the time between processing an order and shipping it. The team knew they needed a better solution for fraud, but they were hesitant. Fraud management solutions introduce the potential of hurting customer experience, which was and still is Touch of Modern’s highest priority.
Touch of Modern needed a fraud solution that could accurately stop fraud without blocking legitimate users. Additionally, Steven wanted a solution that would continue to scale with their business and integrate directly with Touch of Modern’s internal systems. Because Touch of Modern’s customers expected a quick turnaround time between processing and fulfillment, Steven realized that automation would also be necessary to ensure immediate decision-making and smooth processes.
We don’t like to make customers jump through hoops—you might catch more fraud, but you’re hurting good customers.
Steven Ou, CTO
A simple integration to uncover complex fraud
Steven discovered that Sift met all of his criteria for an ideal fraud solution, and he decided to get started before fraud could become a larger issue. He integrated Sift in a single day, and their personalized machine-learning model immediately proved its value by detecting fraud rings among Touch of Modern’s users. Steven considers Sift’s most powerful feature to be its ability to detect hidden and complex connections between users. Through the Network Visualization tool, Touch of Modern can easily prevent waves of fraud by unearthing hidden networks.
Additionally, Sift’s APIs provided the convenience and functionality that Steven needed. Touch of Modern pulled Sift Scores directly into its internal consoles using the Score API, which allowed Touch of Modern to quickly decide which orders to approve, block, or review, based on analysis of Sift’s machine-learning algorithms.
Stopping entire networks of fraud
By uncovering fraud rings through Sift’s Network Visualizations, Touch of Modern was able to prevent thousands of dollars in potential losses. As an example, if a single fraudulent $1,000 purchase was detected, the team would immediately catch that order and stop the accompanying financial loss. Touch of Modern could then use Sift to identify related users (even if the users didn’t have any suspicious activity yet)—which could lead to savings of over $100,000, all from stopping that initial fraudulent order.
Touch of Modern’s machine-learning model is specially customized, learning the unique fraud patterns and signals for its business. Sift’s API enables Touch of Modern to provide automatic feedback that improves its results. Every time their team marks a user as bad or notes a chargeback in their system, these updates instantly improve the model’s accuracy in catching fraud.
In the fast-moving world of e-commerce, retailers have to interpret and analyze a ton of data to make a decision on an order in a short window of time. With Sift, Touch of Modern’s team can literally see and understand the whole story with ease, allowing them to confidently find and stop fraud—all while giving legitimate customers the best experience possible.
The network visualization is incredible. Nobody else has anything like that.
Steven Ou, CTO