TOOL · 01 · CRO
60 SECONDS · NO SIGNUP
A/B Sample Size.
Plug in your baseline conversion rate and the lift you want to detect. Get the sample size per variant and the days you'll need to run.
INPUTSSTEP 1 OF 2
%
PP ABSOLUTE
VISITS
RESULT · SAMPLE SIZESTEP 2 OF 2
PER VARIANT
8,158
5.00% → 6.00% (20% relative lift).
16K
TOTAL N
11
DAYS TO DETECT
+20.0%
MDE RELATIVE
METHOD
Two-tailed normal-approximation formula for two proportions, fixed-horizon test. Numbers assume independent observations and no peeking — if you're running sequential or Bayesian tests, your effective n is different.
FIELD NOTES · WEEKLY
The 7 most common A/B test failure modes.
A brief on why most CRO tests don't replicate, when sequential testing helps, why peeking wrecks your p-values, and how to set MDEs that aren't fantasy.
Need someone to actually run the program?
I run CRO sprints for DTC brands. Hypothesis stack, prioritization, build, readout. 4–8 clean tests per month.
—
CHAPTER · MORE