Split Test Calculator Form
Enter traffic, conversions, and revenue for both variants. The calculator measures conversion lift, significance, confidence interval, and revenue impact.
Performance Chart
The chart compares conversion rate and revenue per visitor across both variants.
Example Data Table
| Metric | Variant A | Variant B | Interpretation |
|---|---|---|---|
| Visitors | 12,000 | 11,850 | Both variants received comparable traffic. |
| Conversions | 420 | 485 | B generated more orders during the test. |
| Revenue | $29,400 | $37,345 | B produced stronger sample revenue. |
| Conversion Rate | 3.50% | 4.09% | B converted better than A. |
| Revenue Per Visitor | $2.45 | $3.15 | B monetized traffic more effectively. |
| Monthly Rollout Visitors | 50,000 | Used for scaling conversion and revenue lift. | |
Formula Used
1) Conversion Rate
CR = Conversions / Visitors
2) Absolute Difference
Difference = CR(B) - CR(A)
3) Relative Uplift
Uplift = (CR(B) - CR(A)) / CR(A)
4) Pooled Conversion Rate
Pooled Rate = (Conv(A) + Conv(B)) / (Visitors(A) + Visitors(B))
5) Standard Error for Hypothesis Test
SE = sqrt(Pooled Rate × (1 - Pooled Rate) × (1/Visitors(A) + 1/Visitors(B)))
6) Z-Score
Z = (CR(B) - CR(A)) / SE
7) Two-Tailed P-Value
The calculator uses the normal distribution to convert the z-score into a p-value.
8) Confidence Interval for Difference
Difference ± Zcritical × Unpooled SE
9) Revenue Per Visitor
RPV = Revenue / Visitors
10) Projected Monthly Revenue Lift
Monthly Revenue Lift = Monthly Rollout Visitors × (RPV(B) - RPV(A))
How to Use This Calculator
- Enter total visitors, conversions, and revenue for Variant A.
- Enter the same values for Variant B.
- Choose the confidence level that fits your testing standard.
- Add projected monthly rollout visitors to estimate scaled impact.
- Click the calculate button to show results above the form.
- Review conversion lift, p-value, confidence interval, and revenue metrics.
- Use the chart to compare conversion efficiency and monetization.
- Download the results as CSV or PDF for reporting and sharing.
Frequently Asked Questions
1) What does this split test calculator measure?
It compares two ecommerce variants using conversion rate, relative uplift, p-value, confidence interval, revenue per visitor, and projected monthly impact. This gives both statistical and commercial insight before launching a winner.
2) Why is p-value important in A/B testing?
The p-value estimates how likely your observed difference happened by random chance. A smaller p-value means stronger evidence that the conversion gap is real rather than accidental variation.
3) What does confidence level change?
The confidence level changes the decision threshold and the z-critical value used for the interval. Higher confidence demands stronger evidence before declaring a statistically reliable winner.
4) Why include revenue and not just conversions?
A variant can win on conversion rate but still produce weaker business results. Revenue helps measure revenue per visitor, average order value effects, and projected monthly lift after rollout.
5) What does “no clear winner yet” mean?
It means the current uplift is not statistically strong enough at your chosen confidence level. The better-looking version may still be promising, but more traffic or time is usually needed.
6) Should visitors always be evenly split?
Even splits are common, but the calculator can still compare uneven traffic volumes. It uses the actual visitor counts from each variant when computing significance and interval estimates.
7) Can I use this for email or landing page tests?
Yes. Although designed for ecommerce, the math also works for email experiments, pricing tests, product pages, signup funnels, and most conversion-focused A/B comparisons.
8) What is the best metric to focus on?
Start with conversion rate and p-value for reliability, then check revenue per visitor and projected revenue lift for business impact. The best decision balances significance and profit.