Testing landing pages when you don't have much traffic
Hookpair team. Updated
Classic A/B testing advice assumes thousands of visitors a week. Most small stores have far fewer. You can still learn from tests, if you change what you test and how you read the results.
Test bigger differences
With little traffic, only large effects show up in the data. Testing a button colour won't tell you anything for months. Test changes that could plausibly move results a lot: a different promise in the headline, a page that matches the ad versus one that doesn't, an offer above the fold versus none.
Decide before you start
- 1Write down what you change and why you expect it to help.
- 2Pick one main number to judge it by, such as add-to-cart rate or revenue per visitor.
- 3Set how many visitors or how many days you will wait before reading the result.
- 4Decide what you'll do if the result is unclear: keep the cheaper version, or the one closer to the ad.
Writing this down first protects you from stopping the test the day it looks good.
Read results honestly
- A difference of a few orders is noise, not a winner.
- Look at revenue per visitor, not only conversion rate: a cheaper offer can convert more and earn less.
- Check that both versions got similar traffic from the same ads.
- Keep a log of what you tested and what you learned, including tests that showed nothing.
Hookpair's split links send each visitor to one page variant and show visits, clicks and orders per variant. The log reports data-quality warnings instead of declaring a winner. See the methodology.
Free tool: Message match check · Compare the ad's hook with the landing page headline: which words and offer numbers carry over, and which disappear after the click.
Questions
How many visitors do I need for an A/B test?
It depends on your current conversion rate and the size of the effect you hope to see. With small effects you need far more traffic than most small stores have, which is why bigger changes are the better tests.
Is it okay to just make the change without testing?
If the change fixes an obvious mismatch between ad and page and costs little, yes. Testing is for decisions where you genuinely don't know the answer.