Landing Page A/B Test Template for Product Growth

The minimum is 7 days to capture full weekly traffic patterns. The actual duration depends on your traffic volume and minimum detectable effect. Use a sample size calculator: input your baseline conversion rate, desired MDE, and confidence level (usually 95%). Divide the required per-variant sample by your daily traffic to get the duration. For most B2B SaaS pages with 500-2,000 daily visitors, expect 2-4 weeks per test.

It depends on your traffic and business impact. A 5% relative improvement on a page that gets 100,000 monthly visitors has a different business impact than a 50% improvement on a page with 1,000 visitors. Generally, design your tests to detect a 10-20% relative improvement. Smaller effects require much larger sample sizes. If your traffic cannot support detecting a 20% improvement within 4 weeks, focus on bigger, bolder changes rather than subtle tweaks.

If your mobile and desktop conversion rates differ by more than 30%, run separate tests. A headline that works on desktop (where users have more screen context) may fail on mobile (where only the first 5 words are visible above the fold). At minimum, segment your results by device during analysis even if you run a single test. The [Mobile Performance Template](/templates/mobile-performance-template) covers device-specific optimization in more depth.

An inconclusive test means the change you made does not meaningfully affect conversion. That is a valid finding. Document it, archive the test, and move to a different variable. Do not keep tweaking the same element hoping for significance. Common reasons for inconclusive results: the change was too subtle, the page element you tested is not a conversion bottleneck, or your traffic was insufficient for the effect size. Review your [funnel analysis](/templates/funnel-analysis-template) to confirm you are testing the right page.