Use caseMarketing agenciesFeature: A/B test1 min read
A/B test two landing pages with a single link
Your client can't choose between two versions of their offer page? Split the visitors of one link between both, then compare. No testing tool, no code.
An agency is preparing a campaign for a trade fair. The client has two versions of the sign-up page: one highlights the programme, the other the speakers. Everyone has an opinion, nobody has numbers.
The campaign link is already everywhere: emailing, LinkedIn, QR on flyers. That is exactly where you can test.
Splitting visits
In the link's Advanced tab, add one or more variants: other destinations, each with a weight. For example:
- main destination ("programme" page): 50%
- variant B ("speakers" page): 50%
On every visit, Bref draws the destination according to those weights. You can test up to four destinations in all. Redirect rules come first: a visitor sent elsewhere by a rule is not part of the test.
Reading the results
In the link's analytics, an A/B test block shows how many visits each destination received over the period. You check that the split is being respected.
To know which page converts best, measure sign-ups on the site itself. Add different UTMs to each destination (utm_content=programme, utm_content=speakers): your analytics tool then compares the sign-ups of each version.
Two precautions
- The draw happens on every visit, without a cookie. The same person coming back may land on the other version. That is the price of a test without tracking; in a campaign where most people click only once, the effect is small.
- Wait for volume. With a few dozen visits per version, a difference may be pure chance. Aim for several hundred visits per version before concluding.
After the test
When a version wins, remove the variant: 100% of visits go to the better page. The link printed on the flyers has not changed, and the CSV export keeps the last 30 days of visits, variant included, for your client report.
