LiveSponsors rebuilt its loyalty engine on SurrealDB Cloud and took queries from 20 seconds to 7 milliseconds
LiveSponsors runs a loyalty and rewards platform for media and entertainment operators. Balances, partner feeds and recommendations for 25 operators now run on SurrealDB Cloud, where a MySQL and MongoDB stack ran before.
Challenge
LiveSponsors' loyalty engine keeps a running balance for every member across 25 operators, ingests partner feeds and reconciles them daily. It ran on a hybrid of MySQL for the relational data and MongoDB for the documents, and the hybrid struggled as the platform expanded.
Daily reconciliation leaned on complex transformation and aggregation logic that had become a bottleneck. Core loyalty operations spiked to 20 seconds, far too slow for the real-time feedback members expect when they earn or spend.
Two database technologies also meant two of everything to run, and the infrastructure burden and technical debt grew with each operator added.
Solution
LiveSponsors consolidated the loyalty engine into SurrealDB Cloud, the managed service, so the team runs the product rather than the database. Relational and document access are unified in one engine.
Schema-optional tables take partner feeds as they arrive and evolve without downtime. SurrealQL's blend of SQL-style joins and document deep fetching replaced the transformation logic that had slowed reconciliation.
The multi-node replicated stack became one streamlined high-memory deployment. SurrealDB Studio and the REST endpoints sped up schema changes and the APIs on top of them.
Results
From 20 seconds to 7 milliseconds
Real-time loyalty logic that spiked to 20 seconds now answers in 7 milliseconds.
Lower infrastructure overhead
Complex replicated nodes were replaced by a single-node setup on SurrealDB Cloud.
Faster iteration
Studio and REST endpoints made schema changes and new APIs quick to ship.
Real-time intelligence
Balances are calculated on the fly and partner recommendations are made as members act.
