Required trials 0 Waiting for inputs

Baseline arm 0 Qualified trials

Variant arm 0 Qualified trials

Calendar read 0 wk Enrollment plus trial lag

Inputs

Planning mode Baseline trial-to-paid rate

12%

Target conversion rate

16%

Margin of error

±3pp

Confidence Power Traffic to variant

Qualified trials per week Trial-to-paid lag Already enrolled trials Minimum runtime

Readout Notes

Needs inputs

Context notes

What It Estimates

The A/B mode estimates trials needed to detect a target trial-to-paid conversion lift with a two-sided normal approximation. The estimate mode sizes one cohort for a selected confidence interval width.

Why Trial Lag Matters

A trial can enter the test today and convert later. Calendar readout adds the entered trial-to-paid lag after enough qualified trials are enrolled.

Before Reading A Test

Keep the cohort definition fixed, avoid mixing traffic sources with different intent, and decide the target lift before looking at results.