A habit-tracking app obsessed over signups while their active base stayed flat. New users in, same users out. The churn was invisible in topline numbers.
The retention-curve playbook:
— Build a weekly cohort retention report in GA4 Explorations
— Read across each cohort row to find where the curve flattens vs falls off a cliff
— Identify the specific week the steepest drop happens
— Segment the churned cohort by their onboarding completion
— Interview or session-replay five users who left at that week
The cliff was week two: retention fell from 44% to 19%. Almost everyone who churned had skipped setting their first reminder during onboarding.
They made reminder setup mandatory in onboarding. Week-two retention rose from 19% to 31%.
Lesson: a flat active-user count is two opposing forces in balance. The retention curve shows you exactly which week to fight for.
The Pixel Diary
@thepixeldiaries
How to read a retention curve to find the week users quit
This post was published in the Telegram channel The Pixel Diary. Subscribe here: @thepixeldiaries.