The Arbitrage Rebuild: How We Made Our Provider Scoring Smarter

Prelude’s SMS OTP routing system, known as arbitrage, selects providers for each message based on delivery performance and cost, using successful code verification rather than inconsistent provider-reported delivery metrics as its primary signal. Its original shared conversion-rate model became inaccurate when customers with different user behaviors, such as e-commerce checkout abandonment, fraud traffic, or sender ID filtering, distorted provider scores and caused routing decisions to shift unnecessarily. Attempts to cap large customers’ influence or build fully separate customer models highlighted a trade-off between sufficient data volume and customer-specific insight. The rebuilt system combines global and customer-level statistics, gradually increasing the influence of a customer’s own data as confidence grows and distributing traffic based on providers’ probability of being the best option. While routine conversion gains are expected to be modest, the new approach is designed to improve resilience during outages, filtering problems, and fraud incidents by adapting routes more rapidly and reducing manual intervention. Prelude is also developing methods to account for each customer’s baseline conversion behavior so provider performance can be evaluated more independently of differences in application flows and end-user behavior.

No tracked trend matches for this post yet.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.