Glossary
What is cohort retention?
This means grouping customers by when they joined, then following each group to see how many are still coming. It is the honest way to tell whether a change worked, because it compares like with like instead of mixing new and old customers together.
The technique that stops a growing total from hiding a leaking bucket.
The problem it solves
Total active customers can rise every month while every individual group of joiners drops off faster than the last. Acquisition papers over retention. Cohorts pull them apart.
How to read one
Each row is a joining month, each column is months since joining. Reading down a column tells you whether recent cohorts are doing better or worse than older ones, which is the question worth asking after any change.
The practical use
Change your reward threshold, then compare the cohorts who joined after against those who joined before. That is a real answer, not an impression.