Enter each variant's numbers and find out whether the difference is real or just noise — statistical confidence, uplift and a plain-English verdict. Two-proportion z-test, in your browser.
Variant A
Rate: 5.00%
Variant B
Rate: 6.50%
Not significant yet at 95%
85.0% confidence · p = 0.1496
Rate A
5.00%
Rate B
6.50%
Relative uplift
+30.0%
Confidence
85.0%
The difference could still be random noise. There's a 15.0% chance you'd see this gap even if the variants were truly equal — keep the test running to gather more data, or the true effect may just be too small to matter.
Two-proportion z-test (two-tailed). Significance ≠ business impact — a tiny but “significant” lift on huge traffic may not be worth shipping. Avoid peeking and stopping the moment it goes green.