What it is.
You start by getting usage data for each feature. Either install the snippet and let autocapture record views, clicks, and inputs, then group them into an action under Data Management, or add capture calls by hand around the code that runs the feature.
Then you build one dashboard. A Trends insight per feature shows active users over time, and a retention insight per feature shows how many of those users come back. Because they sit on one dashboard, you can compare features against each other.
The caveat PostHog raises is effort. A video view is easier than a long form submission, so an event that takes more work happens less often, and that alone can read as weaker retention.
What you get.
- A retention number per feature, comparable across features on one screen.
- Active user trends for each feature alongside its retention.
- An answer to which feature to invest in next, and which one nobody misses.
- A repeatable definition of feature usage, stored as an action you can reuse.
- A retention view PostHog also frames as a way to read product-market fit.
How to set it up.
Install the PostHog snippet or JavaScript library so autocapture starts recording views, clicks, and inputs.
Open Data Management and create an action per feature, combining the events that mean somebody used it.
Weigh the effort each feature takes before you compare, since a one-click action will always fire more than a form.
Click Dashboards, then New dashboard, and name it after the features you're comparing.
Add a Trends insight per feature to plot active users over time.
Add a retention insight per feature, then read them side by side on the dashboard.
Use cases
Compare features side by side
Read retention per feature on one screen instead of one report each.
Find what keeps people
Look at which features hold users past the first weeks.
Argue for a deprecation
Take the low-retention features into the roadmap conversation.
Best for
Deciding what to build next
Feature requests tell you what people ask for. Retention per feature tells you what they come back to.
Cutting features
A feature with usage and no retention is a maintenance cost with no return.
Reading product-market fit
PostHog frames retention as the honest version of this, and per-feature retention shows where the fit is.