How many visitors does a test need before its result means anything? Enter your current conversion rate, the smallest lift worth detecting and your daily traffic to get the sample size per variant and the test duration.
Free, no sign-up. All maths runs in your browser; nothing is sent anywhere.
Choosing the inputs
Minimum detectable effect — the smallest relative lift you’d act on. Halving it roughly quadruples the sample size, so be honest about what’s worth detecting.
Power — the chance the test detects the effect if it’s real. 80% is standard; 90% costs about a third more traffic.
Confidence — 95% means a 5% false-positive rate when there’s no real effect. With more than two variants the calculator splits that budget across comparisons (Bonferroni).
Run whole weeks and don’t stop early when results look good — stopping on a lucky day is the most common way tests mislead.
Methods. Conversion rates: two-sided pooled two-proportion z-test (equivalent to a 2×2 chi-square), Wald interval for the difference, delta-method interval for the relative lift. Averages: Welch’s t-test and Cohen’s d; Mann–Whitney U on raw values. Sample size: normal approximation for two proportions. All checked against SciPy.
Know when a metric really changed — without running the numbers by hand.
QuantumLayers connects to your database, analytics or spreadsheets, runs significance tests like these across every segment, and monitors your key metrics so you’re emailed when something statistically shifts.