Find out if your split test results have reached 95% statistical accuracy
A split test must reach a minimum of 95% statistical accuracy to be reliable, that way you know the winning test variation is not simply outperforming the other variations due to luck or accident.
Enter your A/B or multivariate test results in to the below fields to determine the statistical accuracy of your split tests.
The winning variation is the at a % confidence level.
You need 95% or more to trust the test results.