Enter visitors and conversions for the control and the variant. You get the p-value and the probability that the variant beats the control.
Below 0.05 is the usual bar for "significant at 95%". It measures how unlikely the difference would be if the versions were equal.
The probability that the variant really converts better. 95% is a common bar and the one Stellar Autopilot uses to keep a winner.
How much higher or lower the variant converts than the control. A large number on a small sample is usually noise.
A result only counts once each version has its planned visitors and the test has run whole weeks.
Not started yet? Plan the test with the A/B test sample size calculator. For what real tests looked like, see the A/B testing benchmarks.