What Is An AB-test?
- ad versions
- ad propositions
- landing pages
- PPC campaign settings
- email subject lines
Due to the many opportunities for measuring effectiveness, an AB-test is a widely used within PPC marketing. In addition, they are also often used within email marketing, usability and (offline) direct marketing.
Best practices for AB-testing
Best practices are:
- Determine which KPI you want to use for your AB-test.
- Write a AB-testing plan with the components and versions you want to test
- Make sure that various AB-tests don’t interfere with each other (or keep this within acceptable limits).
- Always test different ad versions within PPC campaigns, to determine which one is the most effective.
- AB-test various landing pages, using tools as Visual Website Optimizer,
Google’s Website Optimizer(note: this tool has been retired and replaced with a new function in Google Analytics), or Adobe SiteCatalyst’s Test & Target.
- Make sure your findings are valid using a chi-square test.
- Use the findings from online campaigns for offline media.
- Test everything, always and continuously, to keep improving your marketing campaigns!
If your website receives sufficient traffic, or your AB-test has a high conversion rate, you can use multivariate testing in order to increase your AB-testing capacity.
AB testing tools
AB testing tools help to determine whether a test is valid. A test is valid if, based on the available data, one component outperforms the other. The outcome of such a test is usually measured in a percentage. Within marketing, a 95% threshold is commonly used to declare a winner. This means that one version has a change of at least 95% to provide better results than the other version(s).
SEObook provides a good (and free!) AB-testing tool. You can find it here.