July 26, 2010

Marketers know that testing and tracking different elements of their campaigns is the best way to understand what works with their intended audience. Common methods for running controlled experiments range from simple A/B testing to sophisticated multivariate testing. You can test pretty much anything, particularly in the online world—from e-mail to video to mobile content to advertising. For the purposes of this piece, I’ll focus on testing Web content.

In A/B testing, one or more new versions of a page or single site element compete against the original (control) version to see which one is more persuasive. A/B testing is quite effective for answering certain types of marketing questions, including creative bake-offs where the decision has come down to two competing versions. But for more complex questions, often those that arise earlier in the testing cycle, multivariate testing is often the way to go.

Read more: http://www.businessweek.com/smallbiz/tips/archives/2010/07/when_multivar...

Add new comment

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.