Free CRO tool
Test Velocity Calculator
How many A/B tests can your traffic actually support? Find out how many statistically valid experiments you can run per month, quarter, and year, and what to do about it.
Your results
- Tests per year
- —
- Sequential tests on your traffic
- Tests per month
- —
- Tests per quarter
- —
- Required sample size
- —
- per variation
- Weeks to reach sample
- —
- at your traffic
Not sure what to test next? Run a free CRO audit
How test velocity is calculated
Test velocity is determined by dividing your available traffic by the sample size required for each test. The required sample size uses the two-proportion z-test formula at 95% confidence and 80% power, the industry standard for A/B testing.
Test velocity benchmarks
| Tests / year | Program tier | Recommendation |
|---|---|---|
| 50+ | World-class | Invest in a full CRO team; parallelize across multiple funnels |
| 24–50 | Excellent | High compound learning; prioritize bold hypotheses |
| 12–24 | Strong | Solid program; use a CRO audit to maximize test quality |
| 6–12 | Selective | Must prioritize carefully; use audit insights to avoid wasted tests |
| 2–6 | Limited | Direct implementation + 1–2 high-impact tests; complement with qualitative research |
| < 2 | Low traffic | Focus on qualitative CRO: user testing, session recordings, surveys |
How to increase your test velocity
- Increase your MDE. Raising your minimum detectable effect from 10% to 20% reduces the required sample size by about 75%, dramatically increasing velocity. The trade-off: you'll only detect bolder changes.
- Consolidate traffic. Rather than testing across multiple pages simultaneously, focus all tests on your single highest-traffic funnel step.
- Test higher in the funnel. Homepage and category page conversion rates are lower, but they receive far more traffic, so tests complete faster there than deep in the checkout flow.
- Drive more traffic. More traffic directly increases test velocity, and CRO improvements make additional paid traffic more profitable.
Frequently asked questions
What is a good A/B testing velocity?
12–24 tests per year is considered a strong, professional-grade CRO program. Top-tier teams at high-traffic companies run hundreds of tests simultaneously, but that requires dedicated engineering and analysis resources. For most sites, 12–24 well-chosen tests per year produces significant compound improvement over time.
Why does the calculator enforce a minimum test duration?
User behavior varies significantly by day of week. A test that only runs Monday–Wednesday misses weekend shopping patterns, which can dramatically skew results. Running tests for at least 1–2 full business cycles (2 weeks minimum for most sites) ensures your data reflects true behavior rather than day-of-week variation. 2 weeks is the standard; use 4 weeks for B2B sites with longer decision cycles.
Can I run multiple tests at once to increase velocity?
Yes, if you test on different pages or mutually exclusive user segments, concurrent tests don't interfere with each other. However, testing overlapping elements on the same page simultaneously creates interaction effects that make results uninterpretable. The calculator assumes sequential tests; if you have multiple high-traffic pages to test independently, your actual velocity could be 2–3× higher.
What should low-traffic sites do instead of A/B testing?
For sites under 5,000 monthly visitors, qualitative research provides much faster and richer insights: moderated user testing (5 sessions reveals most major issues), session recording analysis, user surveys, and heatmaps. A CRO audit synthesizes these methods into a prioritized action plan you can implement directly, no waiting for statistical significance.
The formula
- n
- Visitors per variation at 95% confidence (Zα/2 = 1.96) and 80% power (Zβ = 0.8416).
- weekly traffic
- Monthly visitors ÷ 4.33, split 50/50 between control and variation.
- ⌈ x ⌉, ⌊ x ⌋
- Rounded up and rounded down to a whole number.
Assumes a 50/50 traffic split and sequential (non-overlapping) tests on all of your traffic. Your preferred duration sets a minimum floor: tests never run shorter than selected.