Spinning up
Live
+312AI citations this month
Press⌘Kto search
Spinning up
CRO
The single most-cheated CRO metric. Most agencies declare 'wins' on tests with 200 conversions per arm, well below statistical significance. This tool computes the real sample size for your baseline CVR, desired lift, and confidence level so you don't ship false positives.
Confidence level
Statistical power
Required sample
Per arm (control / variant)
53,148
conversions × 2 arms
Total visitors needed
106,296
combined sample
Days to reach significance
54 days
@ 2,000 visitors/day
Too slow — lift expectation too small for traffic
Why this matters
Most agencies declare A/B test "wins" on 200-conversion samples — well below statistical significance. That ships false positives that get rolled back 60 days later when CVR drifts back to baseline. Xpand uses this formula on every CRO test to avoid shipping noise as signal.
Want the operating spine, not the calculator?
Senior strategist on the call. We pull up your stack live, show you 3 quick wins and 3 structural issues, and tell you honestly whether we're the right fit.