Pew Pew Laser Blog

When to Use a Split (or AB) Test.

Hey marketeers; here's a little blog on page optimization and testing. Are you wondering whether to do a split page test or a multivariate test? These are the situations where split testing is appropriate:

• A split test is the best choice for campaigns with a smallish volume of conversions - about 100 conversions or less per month. If you have more volume than that, you're wasting time by not testing all of the elements that a multivariate test would get you. You never know what will drive conversions unless you test.
• Page layout tests (multiple pages with the same content, but different layout and styling) are also perfect for split tests. Be sure to keep the content exactly the same between all the candidate pages, or you won't know which part of the winning layout was actually influencing the conversion rate.

All other situations really do call for multivariate testing. You can get a lot done by using multivariate (not full-factorial) testing.

Remember that any testing campaign should be run for a minimum of two weeks. Doing so is a critical step to smooth out traffic spikes due to the season and statistical anomalies.

Permalink