Vovy sets up a PostHog experiment with a feature flag, ships both versions, and tells you when the result is real, not luck.
Run an A/B test on my signup button and tell me which one wins
How it works
- Pick one change and one goal: Vovy helps you choose a single change, like button copy, and one goal event, like signed_up, so the answer is clear.
- Create the experiment: In PostHog Experiments it creates a test with control and one variant, which makes a feature flag that splits visitors 50/50.
- Build the variant: Cursor wraps the button in a check of the flag and shows the new version only to the test group. You review the diff.
- Launch the test: Vovy deploys, confirms both versions show for different visitors, then launches the experiment in PostHog.
- Read the result honestly: When PostHog reaches significance, Vovy shows a card with each version's conversion rate, the winner, and whether to ship it.
What you provide
- PostHog installed
- Enough traffic, ideally hundreds of visitors a week
What you get
- A running experiment
- Feature flag controlling the split
- Winner with real confidence
- Ship or drop recommendation
FAQ
How long does a test take?
It depends on traffic. Low-traffic sites may need weeks. PostHog estimates the run time from your numbers before you start.
Can I test with very few visitors?
Only big changes will show a clear winner. With little traffic, talking to five users often teaches more.
Will people see the page flicker?
Loading flags early, or on the server, avoids the flash of the wrong version. Vovy sets that up.
Related tasks
All tasks