A/B Test Design
Design a rigorous A/B test — hypothesis, metrics, sample size, variants — or an experimentation program.
Scenarios
Tags
Estimated duration
Usually 4-15 min
30-day retention before cleanup
Results, submitted inputs, and uploaded file metadata are kept for 30 days after execution, then cleaned up.
HistoryUse this before running a split test, so the result can actually be trusted. It turns a vague idea into a structured hypothesis, picks primary/secondary/guardrail metrics, plans sample size and duration from your own baselines, and specifies exactly what varies and what stays constant. Program mode adds an ICE-scored backlog and velocity math.
-
Prepare: product, the surface and change idea with the observation behind it, your current conversion numbers, and weekly eligible traffic.
-
You get: a test design ready to implement — hypothesis, metric stack, sample/duration plan, variant spec — or a program design with backlog and cadence.
-
Use it for: any comparison you plan to measure: pages, onboarding, pricing, emails.
Output sample
A/B test design preview What the result contains - Structured hypothesis - Metric stack with guardrails - Sample size and duration plan - Variant spec The final report adds evidence, reasoning, and recommended next actions.