Marketing
What is A/B Test?
A/B testing splits traffic between two versions (A and B) to measure which converts better, with statistical significance.
A/B testing is the standard method for optimization: change ONE variable at a time (headline, CTA, button color), serve half the visitors to each version, measure conversion.
Common trap: stopping the test early (no significance) and making the wrong call. Rule: predefine minimum sample size (power analysis calculator) and minimum duration (≥ 1 week to capture day/hour).
Frequently asked questions
How many visitors do I need for an A/B test?
Depends on baseline and expected effect size. Use a sample-size calculator; typically thousands of visitors per variation.
Can I test 3 variations at once?
Yes (A/B/n test), but each needs more traffic to reach significance. At low volume, prefer sequential A/B.