How to A/B test a landing page when you have low traffic
Most service businesses do not have the traffic for a textbook A/B test. You can still learn what works.
The short answer. With low traffic, classic A/B tests rarely reach a clear result. Make bigger changes rather than small tweaks, test one thing at a time over a longer period, judge on enquiries and booked jobs, and use before and after comparisons carefully. Fix obvious problems first, because those do not need a test.
The problem
To prove a small difference between two versions you need a lot of visitors and a lot of conversions. A page with a few hundred visits a month and a handful of enquiries cannot do that.
Fix the obvious first
You do not need to test whether a broken form, a slow page or a missing phone number costs you enquiries. Fix those. Use the landing page checklist.
Test bigger changes
- A different offer or headline angle, not a different button colour.
- A shorter form against a longer one.
- A page with proof near the top against one without.
How to run a test with little traffic
- Change one meaningful thing.
- Run it for several weeks, or until you have a reasonable number of conversions in each version.
- Judge on enquiries and booked jobs, not clicks.
- Keep the traffic source steady, so a change in ads does not confuse the result.
- Write down what you changed and when.
Before and after, carefully
If you cannot split traffic, compare periods. Be careful, because seasons, ads and promotions change results too. Treat the result as a hint, not proof.
Get more of the right traffic
Sometimes the best test is better targeting. See fix your page or buy more ads.
A worked example (illustrative)
A clinic page gets 300 visits a month and about 9 enquiries. A button colour test would never reach a clear result. Instead it tests a new headline and a shorter form together as one bigger change for six weeks. Enquiries rise to 14 from similar traffic. It cannot say which change helped, but the page is better and it can test finer details later. Illustrative.
Common mistakes
- Testing tiny changes on tiny traffic.
- Running a test for a few days.
- Changing the ads during the test.
- Judging on clicks, not enquiries.
Common questions
How much traffic do I need for an A/B test?
Enough for a meaningful number of conversions in each version. As a rough guide, small sites should expect long tests or use bigger changes.
Can I test without special software?
Yes. Run versions one after the other, or split ad traffic between two pages.
Is it worth testing at all?
Yes, if you make meaningful changes and judge on enquiries and booked jobs, not clicks.