Short vs Descriptive Category Headings A/B Test

Visitors from search or ads who leave right away, session recordings of people who read the heading and turn back before the products, or support questions about what the category covers signal a category heading that is out of step with what visitors expect. If wording is the cause, one heading sends visitors into the products with clearer expectations, so bounces fall and purchases rise. If exits come from the category's contents or from poorly targeted traffic, look at that first, and check before launch that both headings read without truncation on mobile.

Category-page purchase conversion rate

Compare purchases among eligible visitors assigned to the short or descriptive heading.

Also track

• Click-through rate (CTR)Check whether the first click rate changes.
• Scroll depthTrack how far visitors explore the page.
• Time on pageRead time alongside purchases; a longer visit is not automatically better.
• Bounce rateTrack exits that may follow an expectation mismatch.

• Copy approachCompare a short heading with a descriptive one that explains the same product and offer.
• Keep constantKeep supporting copy, CTA, category content and placement unchanged.
• ScopeRun the test in the predefined category area; do not also change product, campaign or blog headings.
• MobileCheck line breaks and truncation for both headings.
• InterpretationTie the result to the specific copy compared, not length alone.

1. 1A higher click-through rate or a longer time on page does not identify the better heading on its own; judge each heading by its category-page purchase conversion rate. A heading that raises the bounce rate may be setting an expectation the page does not meet; if purchases do not rise either, do not ship it.
2. 2If one heading wins clicks by teasing something the page does not deliver, do not count those clicks as success or ship that copy. These findings belong to the two headings you compared in this category area; test again before using it for other short or long headings, or for product and campaign headings.
3. 3A heading change may have only a small effect, and a small effect takes a large sample to see. On a low-traffic category, enter the lift you expect as the minimum detectable effect in the sample size calculator, and compare the sample it gives per variant with the category's traffic before launch. If the traffic falls short or the gap stays unclear, write down that this test could not choose between the two headings, and do not declare a winner.