Multivariate Testing (MVT)
Amadesa utilizes the ‘full factorial' model. In general terms, each element and variation is randomly shown with each other element and variation, collecting information about the results. Amadesa believes that full factorial testing is more beneficial than 'partial factorial' or 'Taguchi-based modeling' for two core reasons:
- The amount of traffic needed for a full factorial test and a partial or fractional factorail test is exactly the same.
- As the test continues to run, the data becomes more conclusive, leading to a greater interpretation of how two elements interact with each other.
Additionally, Amadesa's engine automatically removes the worst performing variations inside of the test, thereby reducing the total amount of time needed to reach significance in determining success.
Multivariate testing is slightly more advanced, but allows for the opportunity to understand interactions between various banners/images/content across two or more 'elements' on a page. This helps to understand which areas on a webpage tend to have a greater tendency towards influencing conversion or event completion.
You can run a Multivariate test on virtually any page that receives traffic. Usually, Multivariate tests are run on homepages, landing pages and other marketing pages, information pages, category and product pages, and shopping cart pages. With Amadesa's site optimization platform, you can test both http and https pages (unsecure and secure).
Multivariate testing is a concept that allows the marketer or web designer to test several different elements on a page at the same time and gain insightful information as to which element(s) have the greatest influence on the conversion rates of the test. Where a/b testing is usually specific to a single element or broadly applied to an entire page, Multivariate testing provides more details when one page outperforms another or several elements are tested together.
