Excellent user experience, but not for fraudsters

Technology Category

Analytics & Modeling - Machine Learning

Applicable Industries

Telecommunications

Applicable Functions

Business Operation
Sales & Marketing

Use Cases

Fraud Detection

Services

Data Science Services

The Challenge

SEOClerks, a marketplace for SEO and other web-related services, was facing a significant challenge with fraud. Their approach to fraud prevention was largely reactionary, with fraudulent accounts being banned after a chargeback was received. However, these users would often return and create new accounts to continue their fraudulent activities. Despite having an IP-based fraud-detection tool, SEOClerks was still experiencing various types of fraudulent activity, including money laundering, referral fraud, account abuse, and friendly fraud. The main issue was money laundering using stolen credit card or PayPal information. They were unable to identify clear relationships between multiple bad users, and their existing fraud tool didn't provide any intelligence for spotting fraud rings or repeat abusers.

About The Customer

SEOClerks is a unique marketplace that connects buyers and sellers of SEO and other web-related services. The company was founded in 2011 and has hundreds of thousands of registered users. With a global user base, SEOClerks is always looking for new ways to expand their marketplace. This includes offering a lifetime affiliate program and other incentives. They have also expanded the services offered on the marketplace beyond SEO, allowing the community to request or offer any type of service.

The Solution

SEOClerks turned to Sift, a fraud detection vendor suggested by PayPal, to tackle their 4% chargeback rate. They integrated Sift within two days and immediately began seeing results, identifying previously undetected fraud rings that their previous tool had missed. Sift became an integral part of SEOClerks' daily workflow. With machine learning-based intelligence from the Sift Console, the SEOClerks team could easily uncover hidden links between fraudulent buyers and sellers. They used Lists based on custom criteria to quickly analyze high-risk users and decide whether they should be approved or blocked. Sift's machine learning detects repeat offenders immediately, so SEOClerks can automatically ban them, preventing them from placing orders or messaging other users. SEOClerks also uses Sift to dispute chargebacks involving friendly fraud by presenting evidence from Sift to show that the purchase was made by the rightful owner of the account.

Operational Impact

SEOClerks has seen sales on their platform rise, buoyed by enhanced trust that users have in the marketplace.
With fraudsters and scammers prohibited from creating accounts, good users are having an even better experience with the SEOClerks community.
SEOClerks has been able to train their machine learning model to recognize good, loyal customers.

Quantitative Benefit

SEOClerks’ fraud rate – previously at 4% – has declined significantly.
The amount of time the SEOClerks team dedicates to fraud management has also shrunk significantly, “minimized to minutes versus hours.”
On average, they’re manually reviewing 70% fewer orders than before they used Sift.