What is multivariate testing?
(By the way, to get articles like this free in your inbox, subscribe to our newsletter—the link opens in an overlay).
If you’re serious about online profits, you need to be carrying out split tests on your webpages. The results from multivariate testing and A/B split testing can be amazing; increases in revenue of over 100% are not uncommon.
What is a split test?
In a simple split test, different visitors see different versions of one of your pages. For example, some visitors may see the existing version of a page, whereas others see a new version. Split-testing software “keeps score” of which version generates the most conversions. (A conversion can be whatever you want it to be—orders placed, leads submitted, etc.)
Once you have enough data, the software announces which page was the winner. You then promote that page to be the new “champion.”
The test described above would be called an A/B test, because two page versions were tested, Version A and Version B. A test with three page versions would be called an A/B/C test. “A/B/n” is often used as shorthand for a test that has n different page versions.
What is multivariate testing?
Depending on how frequently you get conversions, it can take weeks or months—or longer—to collect enough data to declare a winner. You may get frustrated waiting for one test to finish before you run the next one. If so, multivariate testing may help. It allows you to carry out several A/B/n split tests on different page elements simultaneously. For example, you may choose to test two different headlines whilst testing two different calls to action.
If you haven’t heard of multivariate testing, maybe you know it by another name, such as
- MVT testing.
- Taguchi method (which is a type of multivariate testing).
- The name of one of the software tools, such as Visual Website Optimizer, Optimizely, Adobe Test&Target, Sitespect, and Google Analytics Content Experiments (which is free).
Most of the software tools allow you to carry out both multivariate tests and A/B/n split tests.
Here are some great resources to help you get the best results from multivariate testing software:
- Avinash Kaushik’s book Web Analytics—An Hour A Day contains some solid advice on carrying out split testing. See pages 237–262. Here is one of Avinash’s blog posts about experimentation and testing.
- Marketing Sherpa’s Landing Page Handbook is an intelligently written overview of the subject.
- This article includes an overview of multivariate testing and a list of 108 ways to use it.
- Jonathan Mendez created this good and thorough guide to multivariate testing.
- Our comparison of A/B testing tools should help you to identify the software that’s most suitable for your needs.
How much did you like this article?
What’s your goal today?
1. Hire us to grow your company
We’ve generated hundreds of millions for our clients, using our unique CRE Methodology™. To discover how we can help grow your business:
- Read our case studies, client success stories, and video testimonials.
- Learn about us, and our unique values, beliefs and quirks.
- Visit our “Services” page to see the process by which we assess whether we’re a good fit for each other.
- Claim your FREE website strategy session with one of our renowned experts.
2. Learn how to do conversion
Download a free copy of our Amazon #1 best-selling book, Making Websites Win, recommended by Google, Facebook, Microsoft, Moz, Econsultancy, and many more industry leaders. You’ll also be subscribed to our email newsletter and notified whenever we publish new articles or have something interesting to share.
Browse hundreds of articles, containing an amazing number of useful tools and techniques. Many readers tell us they have doubled their sales by following the advice in these articles.
3. Join our team (we’re actively hiring!)
If you want to join our team—or discover why our team members love working with us—then see our “Careers” page.
4. Contact us
We help businesses worldwide, so get in touch!
© 2021 Conversion Rate Experts Limited. All rights reserved.