Enter the visitor and conversion counts for both variants. This runs a two-proportion z-test and tells you plainly whether your result is real, or just noise — the question most A/B test dashboards never actually answer.
Two-proportion z-test, two-tailed, 95% confidence threshold (p < 0.05). This tells you whether the DIFFERENCE you measured is likely real rather than random noise — it does not account for multiple testing, novelty effects, or running the test too short a time. If your sample sizes are small, "not significant" often just means you need more data, not that the variants are equal.