A/B test
Admin · Experiment
Light vs Night
Each new visitor is randomly shown one version and keeps it on return visits. Sign-ups and purchases are credited to the version they saw. Bots and preview links are not counted.
RunningSince 16 Nov, 12:15 pm · 1 visitors · 50% of new visitors get Night
Version A
Light
- Visitors
- 0
- Sign-ups
- 0 0%
- Started checkout
- 0
- Purchases
- 0 0%
- Revenue
- ₹0
- Per visitor
- ₹0.00
Version B
Night
- Visitors
- 1
- Sign-ups
- 0 0%
- Started checkout
- 0
- Purchases
- 0 0%
- Revenue
- ₹0
- Per visitor
- ₹0.00
What the numbers say
Sign-ups
Collecting data — needs at least 30 visitors in each version before comparing sign-ups.
Purchases
Collecting data — needs at least 30 visitors in each version before comparing purchases.
Confidence comes from a two-proportion z-test. Wait for 95% before declaring a winner — and give it at least a full week so weekday and weekend visitors are both counted.