Design an A/B Test

Design an A/B test for [WHAT YOU WANT TO TEST].

Context:
- What is being tested: [PAGE / FEATURE / EMAIL / ONBOARDING STEP / etc.]
- Hypothesis: [IF WE CHANGE X, WE EXPECT Y BECAUSE Z]
- Primary metric: [WHAT WILL DETERMINE A WINNER — conversion rate / click rate / sign-ups / revenue]
- Secondary metrics: [OTHER METRICS TO WATCH]
- Current baseline: [CURRENT PERFORMANCE OF THE PRIMARY METRIC]
- Traffic volume: [DAILY VISITORS OR USERS WHO SEE THIS]

Test design:
1. Variants — describe control and treatment(s) clearly
2. Sample size calculation — minimum traffic needed per variant
3. Test duration — how long to run it
4. Targeting — who should see this test (all users / specific segment)
5. Guardrail metrics — metrics that if they move negatively should stop the test
6. How to declare a winner — statistical significance threshold
7. What to do with the results — when to ship, iterate, or discard