Multivariate testing (also shortened to MVT testing) is a thorough, scientific way of comparing different combinations of elements and content on a particular page to see which configuration yields the highest conversion rate.

Multivariate tests are an important part of Impression’s CRO (conversion rate optimisation) service. Along with our other tools and techniques, these tests help us to analyse how users are interacting with your site and make changes that will ultimately lead to more sales for your business.

What is a multivariate test?

Multivariate tests are conducted to allow us to see how different combinations of elements and content affects the conversion rate of a websites. Each version of the page is shown to a segment of the site’s visitors and we track engagement with each one. We use industry leading software from VWO to carry out our multivariate tests and to track them precisely.

Multivariate testing vs A/B testing

Alongside multivariate tests, we also carry out A/B tests (split tests), where only two different versions of a page, normally with just one difference, are compared. With multivariate tests, the number of versions is equal to the number of variations of elements multiplied together.

If you’re testing two headings and two images, you have four variations. But if you’re testing three headings, two images and four call to action buttons, you have 24 variations to test (3 x 2 x 4). Multivariate tests, therefore, take longer than A/B tests, as we have to split the audience down into smaller segments, meaning it takes longer to get a statistically significant result.

While both forms of testing require a good level of traffic to your website, multivariate testing will often have a much higher minimum traffic threshold than split testing. The advantage of multivariate testing, however, is that more changes can be tested in combination, which often allows us to pinpoint the most impactful elements with a high level of accuracy, thereby making changes that will have the biggest possible impact on your conversion rate.

Quick Comparison


Multivariate tests

Number of page variations depends on number of changes tested.

Used to identify which elements are most impactful.

Scientific, quantifiable testing method.

Leads to changes that have a proven impact on conversion rate.

A/B tests

Normally only 2 versions of a page are tested (can be up to 4).

Used to pinpoint the optimal version of a particular element.

Scientific, quantifiable testing method.

Leads to changes that have a proven impact on conversion rate.

Talk to us about multivariate testing

Your digital strategy should be having a direct impact on your business bottom line. If you’d like to find out how we can help you achieve better results, get in touch today! Simply fill in the form below and a member of our team will be in touch. Alternatively, give us a call using the number at the top of the page.

Your data is secure. The connection to this website is encrypted.

Thank you for your enquiry - our team will be in touch as soon as possible.