The first time I had to create a landing page, I spent endless hours editing and tweaking. “Was the copy the most persuasive it could be? Does my form have too many fields? Does it have too few? Is my layout drawing the visitors eye to the call to action?”
Even today, my process has not streamlined much more and the torrent of doubts and second guesses is as present as ever. With so many variables at play, I suppose those worries go with the territory and I would venture a guess that I am not alone in my landing page process.
Any inbound marketer that has ever been tasked with designing a landing page has had these questions and a hundred others swirling around in their head before they push a page live. With so many variables affecting conversion, the opportunities to second guess and endlessly tweak are virtually limitless.
Thankfully, marketers have the ability to automate this endless tweaking to determine the best possible layout for maximum impact. The process is called Multivariate testing.
What is Multivariate Testing?
Similar to A/B testing, multivariate tests automate the process of determining the optimal mix of multiple variables by randomly assigning visitors to different variations of the same page and tracking which iteration performs the best. Although the process takes more time than A/B testing, multivariate testing returns far more accurate results. See the diagram below for a more in-depth look at the process.
The Multivariate Testing Process
Services like Pardot make the process of multivariate testing extremely straightforward. Below are the three basic steps required for effective multivariate testing.
Identify Your Variables
Obviously your multivariate tests would take an unreasonable amount of time if you test every variation you can think of. It is important to narrow your test down to a few key variables. Some of the most common areas for variations include:
- Call to Action
Create Page Variations
Each variable needs to have its own iteration of the landing page. Users will be sent randomly to a given iteration upon visiting the page. Visitors are cookied so that if they were to return to the same page again, they would see the version originally assigned to them. Return visitors will never be left wondering why your page keeps changing.
It is important to keep your multivariate test live for a long enough period of time to receive a substantial sample size. Once your page has received enough visitors, you will be able to view the results of your tests. Different services record different information, but Pardot allows you to track views, conversions, conversion rate, and will automatically determine the best page for you.
You may go through several stages of testing, manipulating different elements at different times to pinpoint the best headline, the most attractive layout, and the most compelling call to action. Though testing may take time, the payoff can be huge.
Good stuff as usual, Matt.
Thank you so much Tripper. That means a lot!