Skip to content

Free tool · Test sizing

How long until you can read the test?

Not just the sample size: the weeks it takes to enroll it, let the metric mature and ship the change. Over about 10 weeks, a micro-tweak is the wrong test.

The share of users who complete the step now, e.g. 32% publish a schedule within 7 days.

Relative to today’s rate: +10% turns 32% into 35.2%.

Only the users who can see the change: accounts for B2B, orders for e-commerce.

Design

Advanced: confidence and power

Confidence (two-sided)

Power

ReadyReadable within 6 weeks.

6 weeks to read

Run a standard A/B test with a guardrail metric, and decide the stopping rule before you start.

  • Enrolling: 4.3 wk
  • Metric maturing: 1.0 wk
  • Launch lag: 0.0 wk
Users per arm
850
Users in total
1,700
Readable in 6 weeks at your traffic
A lift of about +18.4% or bigger.
Weeks to read, by the lift you want to detectHover or focus a point
Show as a table
LiftWeeks to read
+2%426
+5%69
+10%18
+15%9
+20%6
+30%3
+40%3
+50%2
+75%2
+100%2

Two-sided test on a conversion rate, n = 16·p(1−p)/δ² scaled for your confidence and power (Kohavi et al.), at least 2 full weeks to cover weekly cycles. The same math the Gamify It MCP uses.

Your product has its own version of this.

Find the stage you're stuck in, then make one move. Connecting Gamify It is free; Pro starts it.