Upload a CSV, pick your binary outcome and two groups — and get each rate with an honest confidence interval, the difference with its own interval, and a clear significance verdict with an exact-test fallback for small samples. Free.
Free analyses run on up to 10,000 rows. Larger files are randomly sampled to that size — sign up to analyze your full dataset.
Comparing the two rates...
Sent to — per-group rates with Wilson confidence intervals, the significance tests with cross-checks, the difference with its Newcombe interval, R code, and AI insights.
Analyze another fileEach group's success rate carries a Wilson score 95% confidence interval, which behaves honestly for small samples and rates near 0% or 100%. The rates are compared with a two-sample proportion z-test on the pooled standard error, cross-checked with the continuity-corrected chi-square test, and the absolute difference carries a Newcombe hybrid 95% interval built from the two Wilson intervals. If any cell of the 2x2 table (successes or failures in either group) has fewer than 5 observations, the normal approximation is unreliable and the verdict comes from Fisher's exact test instead — with the switch disclosed. The success level is detected robustly (1/TRUE/yes-style tokens preferred; otherwise the rarer level, flagged), and the report always states which level counted as success.
Use it when you have a two-valued outcome and two groups and the question is whether the rates genuinely differ — A/B tests, cohort comparisons, defect or churn rates.
Not for 3+ groups (use a many-group categorical test), numeric outcomes (use a group comparison of means), paired designs, or proving the rates are the SAME (that is an equivalence question with its own tool).
Built for: Marketers, product managers, and analysts comparing conversion-style rates between two groups
Typical data source: Any spreadsheet or CSV with a two-valued outcome column and a two-level group column
One binary outcome plus a two-level group column. For example, conversions by ad variant:
Minimum 10 rows · Best with 200-100,000 rows with two groups of comparable size
Standard-library analysis: did the conversion rate really differ between two groups? Map a binary outcome (0/1, yes/no, TRUE/FALSE, converted/not) and a two-level group column and get each group's rate with a Wilson 95% confidence interval, the two-sample proportion z-test with a continuity-corrected chi-square cross-check, the absolute difference with a Newcombe hybrid confidence interval, the relative change as computed, and an automatic Fisher's exact fallback — clearly disclosed — whenever any cell count is too small for the z-test.
The two rates side by side with Wilson 95% error bars — how far apart they sit and how much the intervals overlap.
The z-test, the chi-square cross-check, and (when cells are small) the Fisher's exact test that carries the verdict.
The raw counts behind the rates: n, successes, failures, each rate's interval, and the relative change as computed.
Full disclosure: success-level detection, why Wilson/Newcombe intervals, the cell-count guard, and the verdict rule.
Plain-English interpretation — what the numbers mean, what's significant, and what to do next.
Did the conversion rate really differ?
Map converted-or-not as the outcome and the variant as the group. You get each variant's rate with an honest confidence interval, the difference with its own interval, and a clear verdict on whether the gap is bigger than chance.
See our FAQ for details on pricing, data privacy, and how the analysis works. Every report includes a Methodology section showing the statistical test, assumptions checked, and diagnostics run.
Run any analysis on your own data — validated R analyses, interactive reports, AI insights, and PDF export.
Try Free — No Credit CardTell us what went wrong, in your own words. We capture the page you're on automatically, so no need to describe where you are.