Case Study: SendinBlue achieves scalable, automated fraud detection and faster account validation with Dataiku

How SendinBlue Uses Dataiku's Self Service Analytics to Scale Risk Team Productivity with Predictive Analytics

SendinBlue, a relationship marketing SaaS that sends over 30 million messages per day for more than 50,000 companies, faced a scaling challenge: manual validation of uploaded contact lists to prevent stolen or non-opted-in addresses was slow, costly, and hurt the user experience. To automate fraud detection and speed up account validation, SendinBlue partnered with Dataiku and adopted Dataiku Data Science Studio (DSS) to build a predictive analytics workflow.

Using Dataiku, SendinBlue ingested more than 1 billion historical email events, thousands of blocked accounts, and hundreds of fraud criteria to automatically classify new customers as “good,” “bad,” or “uncertain,” with downstream rules to block, validate, or route accounts for manual review. The Dataiku-powered solution was developed and deployed in under three months, saved the equivalent of a full-time role, dramatically shortened validation delays, now handles the majority of new accounts (up from 24% in initial rollout), and achieved an 83% precision rate on email classification—improving scalability and SendinBlue’s email reputation.

SendinBlue

Armand Thiberge-Sendinblue

CEO