5 articles in this category
A worked scenario for designing a trustworthy A/B test: pre-commit sample size, avoid peeking, check sample ratio mismatch, and choose the right stati...
I've seen too many teams call a test early and ship a dud. Here's how to design an A/B test that actually gives you a clear answer—without the statist...
Peeking at your A/B test inflates false positives to 30%. I argue for sequential testing methods that let you monitor without ruining your results.
Peeking at your A/B test before it's done inflates false positives by up to 5x. Pre-commit to a sample size or use sequential testing to get honest re...
You're checking results daily and your A/B tests keep lying. Here's how to stop peeking, set sample sizes, and get valid answers.