Case Study: Riskified cuts transaction response times 75% and scales reliably with New Relic

Riskified Takes on the Speed and Scale of E-Commerce Fraud With Insights From New Relic

Riskified is a New York–based fraud‑prevention company that uses machine learning to approve or decline e‑commerce transactions and even guarantees chargebacks for incorrectly approved orders. Because accuracy and latency directly affect millions in revenue and the experience of major global brands, Riskified needed to maintain sub‑second decision times and handle massive, sudden traffic spikes without service degradation.

To solve this, Riskified instrumented its platform end‑to‑end with New Relic (APM, Infrastructure, Synthetics, and custom Insights dashboards), retuned autoscaling based on the right metrics, and stress‑tested ahead of peak events. The result: a 75% reduction in average transaction response time (from ~800 ms to ~200 ms), rapid mitigation of a 100x traffic spike in about 20 minutes, better cost control during Black Friday, and stronger DevOps ownership across the company.

Riskified

Assaf Feldman

Co-Founder and Chief Technology Officer