How to test 1–4 thumbnail variations effectively, what metrics actually matter, how to interpret results, and the common mistakes that make A/B test data useless.
