How much traffic you need to trust a website test

Key takeaways

Trusting a website test needs far more traffic than most owners expect. To detect a lift from 2% to 2.5% conversion with standard confidence, you need roughly 14,000 visitors per version. Most small business pages do not get that in a reasonable time, so they should fix clear problems directly and compare periods carefully instead of running A/B tests.

• About 14,000 visitors per version to detect 2% to 2.5% reliably.
• Smaller lifts need far more; bigger lifts need less.
• Most small business pages cannot reach that in a few weeks.
• Use before-and-after comparisons and micro-conversions instead.

Why does sample size matter?

Conversion numbers bounce around naturally. With 30 enquiries a month, a 'good' month and a 'bad' month can differ by a third with nothing changed. A test needs enough visitors that a real difference stands out from that noise. Without it, you will declare winners that are really luck, and make decisions that quietly hurt the site.

• The difference between a good website and a high-converting website
• A/B testing for small business websites

How much traffic do you need?

The required sample depends on your current conversion rate and the size of lift you want to detect. Free tools such as Evan Miller's sample size calculator give the figure for standard settings (95% confidence, 80% power). These approximate examples show the scale.

• 2% baseline, detect a 25% relative lift (to 2.5%): about 14,000 visitors per version.
• 2% baseline, detect a 50% relative lift (to 3%): about 3,800 visitors per version.
• 5% baseline, detect a 20% relative lift (to 6%): about 8,000 visitors per version.
• Two versions means double the total traffic.

• Low-traffic website? Why conversion matters more than visits

What should low-traffic sites do instead?

Fix clear problems without testing, because a broken form or missing price does not need proof. For judgement calls, compare periods: the eight weeks after a change against the same eight weeks last year and the eight weeks before. Watch micro-conversions for earlier signals, and change one area at a time so you know what caused any movement.

• Fix obvious problems directly.
• Compare equal periods, allowing for seasonality.
• Change one area at a time.
• Track micro-conversions for earlier signals.
• Use user testing to understand why, not just whether.

• Micro-conversions: the small steps that predict enquiries
• User testing for small business websites

What are the common testing mistakes?

The most common mistake is stopping a test as soon as one version looks ahead. Early leads often vanish. Others include testing several changes at once, running tests across seasonal swings, and measuring clicks rather than enquiries. Google also advises keeping tests temporary and using proper redirects or canonical tags so search is not affected.

• Website conversion myths that waste budget

When is testing worth it for a small business?

Testing is worth it on your busiest pages, such as a homepage or main paid landing page with tens of thousands of visits a month, and for close decisions where either version could plausibly win. It also works for bold changes, where the expected difference is large enough to detect with less traffic.

What does this look like in practice?

A UK tradesperson ran an A/B test on a page with 600 visits a month and declared a winner after two weeks. The 'winning' version was rolled out and enquiries fell the following month. With that traffic, the test would have needed well over a year to be reliable. We switched to fixing clear issues and comparing quarters, and enquiries recovered.

• See client work
• Website Design & Development for UK businesses

Testing checklist

• Check the page's monthly traffic and conversion rate.
• Use a sample size calculator before starting.
• Only test if you can reach the sample in 4 to 8 weeks.
• Decide the sample and end date in advance; do not stop early.
• Measure enquiries, not clicks.
• Otherwise, fix clear issues and compare equal periods.

Next step

Testing is powerful when the traffic supports it and misleading when it does not. We can tell you which of your pages can be tested and how to measure the rest.