PostHog for Product Analytics: Setup Guide
PostHog is an open-source product analytics platform that bundles analytics, session replay, feature flags and experiments. Startups use it to see what users do after sign-up, where they drop off and which features keep them.
What is PostHog used for?
Understanding behaviour inside your product: activation funnels, retention, feature use, session replays of real users, and feature flags and A/B tests to change the product safely. Its free tier includes 1 million product analytics events, 5,000 session replays and 1 million feature flag requests a month, with one year of data retention; paid plans retain up to seven years (PostHog pricing). How it fits beside GA4 is in the analytics stack.
| Step | Users | Conversion |
|---|---|---|
| sign_up | 1,200 | 100% |
| project_created | 720 | 60% |
| invite_sent | 290 | 24% |
| purchase | 96 | 8% |
How do you set up PostHog?
- Create a project in the US or EU cloud, based on where your users are.
- Install the JavaScript snippet or SDK in the app.
- Call identify after login so events join to the user.
- Send key product events from your event tracking plan, ideally from the backend.
- Group users by company if you sell B2B, so you can see account-level use.
Which reports should you build first?
- An activation funnel from sign-up to the first value moment.
- Weekly retention of new users.
- Feature use by plan.
- Session replays filtered to users who dropped off in onboarding.
These connect to the activation work in free trial activation and product-market fit signals.
Should you use feature flags and experiments?
Yes, once you ship often. Flags let you release to a few users first and roll back without a deploy. Experiments compare versions on a real metric such as activation. Keep experiments to changes with enough traffic to read a result. Respect consent when you record sessions; see consent and privacy.
Measure activation, not visits.
Traffic tells you marketing worked. Activation tells you the product did.
What should you watch for?
- Autocapture creates many events; define the ones that matter as actions.
- Mask inputs in session replay so sensitive data isn’t recorded.
- Set billing limits so a traffic spike doesn’t surprise you.
- Report product metrics next to marketing ones in dashboards.
Frequently asked questions
What is PostHog?
An open-source product analytics platform with funnels, retention, session replay, feature flags and experiments.
Is PostHog free?
It has a free tier with 1 million analytics events and 5,000 session replays a month.
Should I use PostHog or Google Analytics?
Often both: GA4 for marketing traffic and campaigns, PostHog for in-product behaviour.
Can PostHog be used for B2B SaaS?
Yes. Group analytics lets you see usage by company as well as by user.
Does PostHog have an EU cloud?
Yes. You can choose US or EU hosting when you create a project.
How long does PostHog keep data?
One year on the free plan and up to seven years on paid plans.
Sources & further reading
Analytics tools change quickly. These sources let you check the dated facts directly. Screens shown are mock-ups for a fictional SaaS company.