Glossary entry
What is A/B Testing?
A/B testing splits traffic between two or more variants of an ad, landing page, or email and measures which version produces the better outcome. Done right it converts opinion into evidence and compounds revenue 5 to 30% per quarter.
What to test
Headlines first (biggest impact), then offers, then visuals, then CTA copy, then layout. Subtle font and color tests usually waste traffic. Big swings (different value proposition, different lead magnet) move the needle.
How to design a clean test
One variable at a time, equal traffic split, sample size pre-calculated for your target lift, run until statistical significance (95% confidence usually). Stopping early because 'one is winning' is the most common error in the wild.
Sample size matters
A 5% lift on 100 visitors is noise. A 5% lift on 5,000 visitors is a number you can act on. Lower-traffic sites need either bigger expected lifts or longer test windows.
How Unled Network helps
Managed Campaigns runs continuous creative tests on every account. Creative Production generates the variants. Lead Generation tests funnel and form sequences.
Related Unled Network services
Frequently asked
How long should a test run?
Until you reach statistical significance, but at least one full week to capture day-of-week variance.
A/B or multivariate?
A/B for big questions (which offer wins). Multivariate when you have lots of traffic and want to test interactions.
How many variants at once?
Two to four for ad creative. Two for landing pages so you have enough sample per cell.
What if both variants tie?
Pick the one cheaper to maintain and move on to the next test. Time spent fighting ties is time not testing bigger ideas.