What we were asked to solve.
A regional SaaS product was seeing 8–9% monthly churn with no clear pattern. Product had hunches, sales had different hunches, and no one had the data to tell them apart. Every “save the customer” call was a shot in the dark.
Our approach.
We built an event‑level warehouse on their product usage, billing, and support data, then modelled churn against every variable that mattered. The signals we found — and the interventions we designed around them — reshaped how the whole company thought about retention.
- Event pipeline capturing every meaningful product action
- Churn model surfacing at‑risk accounts 30 days before cancellation
- Playbook of interventions tied to each churn signal
- Weekly retention dashboard for CS, product, and leadership
The outcome.
Monthly churn fell from 8.4% to 5.6% within two quarters — a 33% reduction — and Customer Success now starts every week with a ranked list of accounts to save.
33%
Reduction in monthly churn
30 days
Early warning window
6 mo
To full impact