Split-Test Calculator

Plug in up to eight variants, keep tabs on Bayesian win probabilities, overlapping conversion-rate curves, and the real revenue upside before you make a decision.

Variants

Minimum of two, maximum of eight cards.

Variant A

Variant A

Variant B

Variant B

Variant C

Variant C

Probability Calculations

Conversion rate vs. probability of being the best variant.

Variant A

1,000 visitors · 4 conversions

0.40%

Conversion Rate

Probability Best

0.0%

Variant B

1,000 visitors · 5 conversions

0.50%

Conversion Rate

Probability Best

0.0%

Variant C

1,000 visitors · 6 conversions

0.60%

Conversion Rate

Probability Best

0.0%

95% confidence

Variant A still leads

Challengers need more traffic or better performance to unseat the baseline.

Variant AVariant BVariant C
Absolute lift vs. baseline
0.00%
Relative lift vs. baseline
0.00%
Incremental conversions
0
Incremental revenue
$0

Business Context

Translate results into revenue.

Statistical Settings

Dial in how strict you want to be.

Confidence level determines how sure you need to be before calling a winner. Higher confidence means fewer false positives but requires more data.

Data Quality

Minimum thresholds for reliable results.

Recommended Sample Size

4,000 visitors

You have 3,000 — need 1,000 more

Leader 5+ conv Total 10+ conv Sample size
Loading analytics...