Sift > 实例探究 > How dbrand automated chargeback prevention

How dbrand automated chargeback prevention

公司规模

200-1,000

地区

America

国家

Canada

产品

Sift

技术栈

Machine Learning
API

实施规模

Enterprise-wide Deployment

影响指标

Cost Savings
Productivity Improvements

技术

分析与建模 - 机器学习

适用行业

零售

适用功能

销售与市场营销

用例

欺诈识别

服务

数据科学服务

关于客户

Dbrand is a leader in the custom skin market, offering shoppers the opportunity to personalize their countless gadgets with unique, customizable, and precision-fitted vinyl wraps. All of their products are developed and manufactured in Toronto. Dbrand's distinctive and industry-changing virtual skin building interface puts the creative power in the hands of consumers. As their business grew, they faced an increasing number of fraudsters on their site, leading to costly chargebacks and hours of manual review. They needed a smarter, more scalable solution to mitigate the impact of fraud on their bottom line and brand.

登录后查看完整内容

挑战

As dbrand's business grew, so did the number of fraudsters creeping onto their site. The majority of the fraud they experienced was from bad users purchasing goods using stolen credit cards. The resulting chargebacks were costly, not only due to the high-quality product that was lost, the sale that was refunded, or the bank-levied chargeback fees, but also the hours of manual review and headaches that the fraud caused. Even as their chargeback rate reached a high of 2.18% in a single month and 4 customer service employees became dedicated fraud management experts, fraudsters continued to slip past their defenses. To mitigate the impact of fraud on their bottom line and brand, dbrand sought a smarter and more scalable solution.

登录后查看完整内容

解决方案

After researching fraud management solutions, dbrand CEO Adam Ijaz was disappointed to find that many required ongoing manual review and hand-holding. In search of a vendor that could reduce their workload by growing efficiency, Adam discovered Sift, drawn by the product’s machine learning and automation features. Full integration took a week, and was extremely simple with Sift’s easy API and extensive documentation. With just one month of training, dbrand’s custom machine learning algorithms were catching fraud unique to the business, identifying returning and new fraudsters alike.

登录后查看完整内容

运营影响

Adam’s team saw accurate and actionable results within 3 months of integrating with Sift.
By using Sift Scores and the features that support automating fraud review within dbrand’s existing order management system, the team saved 200 hours a month in fraud investigation.
Now, dbrand dedicates just 1 hour every month to fraud management, reviewing the system parameters and ensuring that results remain accurate.
The fraud management team has since returned to their customer service roles, and zero people deal with fraud full-time; their system is so accurate that it’s in large part fully automated.

登录后查看完整内容

数量效益