Preview
What it is.
Statsig lets a small team run experiments properly without a purchase order. The free Developer plan carries 2M events per month, unlimited flag and config checks, plus 50,000 session replays per month.
Most teams meet it through feature flags and stay for the experiment engine. Flags are unmetered, so you can gate every release behind one and only pay attention to the event count.
What you get.
- A/B tests with the statistics handled, so the debate is about what to test.
- Unlimited flag and config checks, which makes gating every release free.
- 2 million events a month for the analytics behind the tests.
- 50,000 session replays a month, included at no extra cost.
How to set it up.
Install the SDK and put your next release behind a flag before you try any analytics.
Set the flag to a percentage rollout, so the first users are a sample and not everyone.
Define the metric you care about before you start the experiment, since choosing afterwards invalidates it.
Let the test run to the sample size it asks for, even when an early result looks obvious.
Keep an eye on the 2 million event allowance if you log every page view as an event.
Pricing Plans
Developer
$0
2 million events a month, unlimited flag checks, 50,000 session replays
Use cases
Gate a risky launch
Wrap the change in a flag, since flag checks are not metered.
Run a test without buying anything
Design the experiment and let the platform handle the statistics.
Kill switch for a release
Put the release behind a config you can turn off without a deploy.
Best for
Your first A/B test
The statistics are handled, so the argument is about what to test and not about whether the result is significant.
Gating a risky release
Unlimited flag checks mean you can wrap every launch in a flag without watching a meter.
Small teams doing growth work
Analytics, flags, plus experiments in one place beats stitching three vendors together at this size.
Open the tool
Opens on Google in a new tab.