Sample ratio mismatch (SRM) checker
Check whether your A/B test's traffic split is broken. Enter visitors per version and the split you expected; get a chi-square test in one click.
Don't trust this test's result. Look for redirects, bot filtering, slow-loading versions or assignment bugs, fix it, and rerun. Expected: 10,300 / 10,300.
How it works
If you split traffic 50/50 and got 10,000 visitors in A and 10,600 in B, is that bad luck or a bug? The checker runs a chi-square goodness-of-fit test on the counts.
A p-value below 0.001 means the split is almost certainly broken, and the test result can't be trusted however good it looks. Common causes: a redirect that loses some visitors in one version, bots filtered differently, a version that loads slower so its tracking fires less, or a bug in how visitors are assigned.
There's no single standard cutoff. This checker flags below 0.001; Lukas Vermeer's SRM Checker used 0.01. Both are far stricter than the usual 0.05 on purpose, because you run the check on every test and don't want false alarms.
Questions people ask
What is sample ratio mismatch?+
When the number of visitors in each version doesn't match the split you set, beyond what chance explains. It usually means a bug in assignment or tracking, and it makes the test result unreliable.
What p-value means SRM?+
There's no single standard. This checker flags it below 0.001 and Lukas Vermeer's SRM Checker used 0.01. Both are much stricter than 0.05, because the check runs on every test and a loose cutoff would raise false alarms.
What should I do if my test has SRM?+
Don't trust the result. Find the cause (redirects, bot filtering, slow loading, assignment bugs), fix it, and rerun the test.
Learn more
More free tools
Want the tests run for you?
AI agents read your analytics and payments, find where you lose the most money, build the fix and test it. Every test is judged on revenue, not clicks. Plans from $29 a month.