cyberneticlibrary

Design and run A/B tests

ab-testingskillsetup L224
Infrasity-Labs/dev-gtm-claude-skills
What it does

Design and analyze A/B experiments

Best for

When designing statistically valid experiments with pre-committed sample sizes and rigor.

Inputs
  • · Baseline metric
  • · Hypothesis
  • · Expected lift percentage
Outputs
  • · Sample size calculation
  • · Test design document
Preconditions
  • · Baseline conversion rate known
  • · Traffic volume sufficient for sample size
Failure modes
  • · Peeking at results early causes false positives
  • · Sample size insufficient before stopping
Trust signals
  • · Sample size tables with baseline/lift combinations
  • · Evan Miller + Optimizely calculators linked
  • · Peeking problem explicitly named