Our process
Retention analysis should reduce uncertainty. Our process makes assumptions visible, keeps your team close to the evidence and ends with decisions you can put to work.
We define who must decide what, which user behaviour matters and what “retained” means in your business model. This prevents an elegant analysis answering the wrong question.
Before interpreting curves, we check identity logic, event coverage, date boundaries and cohort construction. Any limitation travels with the finding it affects.
We compare acquisition source, activation behaviour, customer type, geography and product use—selectively, not as a fishing expedition.
Findings become a short list of product, lifecycle or measurement actions, each tied to evidence and an observable result.
Working principle
We do not disappear into analysis and return with a deck. Short working sessions at each stage let product, growth and data owners challenge assumptions while they are still easy to change.
Plan a working session ↗