Empowering
real-time
fitness
analytics
Nightly batch reporting could not answer live questions
Studio performance data landed in a warehouse overnight. Operators were making staffing and class-mix decisions against numbers that were up to twenty hours old, and the reporting stack could not absorb new devices without a rebuild.
A streaming ingestion layer and a managed operating model
We stood up a streaming ingestion path in front of the existing warehouse, moved transformation into the pipeline, and handed operations to a pod that owns the on-call rota.
Live numbers, and headroom for the next device generation
Operators moved from next-day to same-minute reporting. Device onboarding became configuration rather than engineering, and the platform held its SLA through peak-season load.
Class mix and staffing decisions made on the day
The reporting change altered how the operations team works: staffing and class-mix decisions are now made against the current day rather than the previous one, and the analytics roadmap is no longer gated on pipeline work.