A/B Testing
A/B testing is a method of comparing two versions of a web page, email, or other marketing asset to determine which one performs better against a defined conversion goal.
A/B testing (also called split testing) is an experiment where you show two different versions of something to different segments of your audience at the same time and measure which version drives more conversions. In a typical A/B test, you create a "control" (version A, the current design) and a "variant" (version B, with one specific change). Traffic is randomly split between the two versions, usually 50/50.
After enough data is collected, statistical analysis determines which version performed better with a confidence level of at least 95%. The key to effective A/B testing is changing only one variable at a time. If you change the headline AND the button color AND the form layout simultaneously, you won't know which change caused the improvement.
In practice
- An e-commerce store tested product page layouts: the original with images on the left vs. a variant with a larger image carousel. The carousel version increased add-to-cart rate by 11%
- A B2B company tested two landing page headlines: "Grow Your Business" vs. "Get 3x More Leads in 90 Days." The specific, quantified headline increased form submissions by 38%
- An email campaign tested two subject lines: "Our Biggest Sale Ever" vs. "50% Off Everything, This Weekend Only." The specific offer subject line had a 22% higher open rate
- A nonprofit tested placing their donation amount options in ascending ($10-$25-$50-$100) vs. descending ($100-$50-$25-$10) order. The descending order increased average donation by 15%
Not sure how this applies to you?
A senior engineer walks your site, your search presence and the way enquiries reach you, then writes down the three things worth fixing first. You keep the document either way.