BAZ Experimentation Engine
A/B Test Significance
Two-proportion z-test (control vs variant) — lift %, z-score, p-value, and a call on significance at 95%. One test run to significance beats a hundred opinions.
Your numbers
Your result — computed live
Lift
20.00%
- p-value
- 5.38%
- Verdict
- Not yet
Not significant yet (p=0.0538). Keep gathering data — don't call it early.
z = (p̂_v − p̂_c) ÷ √(p̄(1−p̄)(1/n_c + 1/n_v)) · p two-tailed
The canon
One test, run to significance, beats a hundred opinions — but only if you don't peek.
The benchmark
Call a winner only at p<0.05. Peeking kills tests — decide the sample size before you start.
Want the full fix-it report?
You already have your number above — free, no email needed. For the saved computation, the action playbook, and a follow-up from BAZ, drop your email.
One email with the report. No sequence, no spam. We earn the next contact.
This calculator runs the BAZ Experimentation Engine — the same algorithm BAZ runs for paying clients.