Precision Prospects Embraces Multivariate Testing
We broke ground this week on our newest project here at Precision Prospects: The Lab, a flexible landing page delivery system that utilizes multivariate testing methods to provide us with the data needed to enhance our landing page conversion rates. When we were first brainstorming this project, we just envisioned a simple A/B Testing scenario but then we started thinking big — Instead of merely testing two different options, why not test thousands at a time? Of course this option would take a little more work, but we concluded that the resulting data would be worth the effort.
So the basic function of The Lab becomes the delivery of dynamic landing pages, each having a given number of variables (x), and each variable having a given number of possibilities (y). The result will be that we will, in effect, have x^y different landing pages. That number can get pretty big in a hurry. For example, if we test 5 variables with 5 options each, we’ll have 3,125 possible landing pages. That’s far too many to analyze as separate entities so we will instead be looking for trends that hold true across different pages. For example, if x is set to a, the page performs better than when x is set to b. We can take this even further and learn how different variables affect conversion across multiple campaigns.
Where this gets interesting is from the design standpoint. Traditionally, designers have tinkered away at our landing pages playing with this and that until the whole thing looks perfect. But when you’re talking about thousands of different pages, that becomes impossible. Some of the random combinations (maybe even many of the random combinations) might look bad. What is going to be the cost of losing that “human element” in the design of a perfect landing page? To be honest with you, we don’t know yet. That’s the scary part. We’re definitely going to keep an eye out for any particularly poor performing variable possibilities and remove them from the rotation as soon as possible. The other thing we are going to do is to start small. Instead of throwing a bunch of hugely different variables into the mix right of the bat, we’re going to start with a few minor adjustments to a standard landing page.
So, why go through all this work? Because, at the end of the day, what’s important isn’t how good our design team feels about the landing pages, but instead, how well our landing pages are converting. With The Lab’s multivariate testing methods, we are going to take the control away from the designers and give it to the potential leads because they’re the ones who really count.
Posted by Robert James Reese in Development | 0 Comments
We broke ground this week on our newest project here at Precision Prospects: The Lab, a flexible landing page delivery system that utilizes multivariate testing methods to provide us with the data needed to enhance our landing page conversion rates. When we were first brainstorming this project, we just envisioned a simple A/B Testing scenario but then we started thinking big — Instead of merely testing two different options, why not test thousands at a time? Of course this option would take a little more work, but we concluded that the resulting data would be worth the effort.
So the basic function of The Lab becomes the delivery of dynamic landing pages, each having a given number of variables (x), and each variable having a given number of possibilities (y). The result will be that we will, in effect, have x^y different landing pages. That number can get pretty big in a hurry. For example, if we test 5 variables with 5 options each, we’ll have 3,125 possible landing pages. That’s far too many to analyze as separate entities so we will instead be looking for trends that hold true across different pages. For example, if x is set to a, the page performs better than when x is set to b. We can take this even further and learn how different variables affect conversion across multiple campaigns.
Where this gets interesting is from the design standpoint. Traditionally, designers have tinkered away at our landing pages playing with this and that until the whole thing looks perfect. But when you’re talking about thousands of different pages, that becomes impossible. Some of the random combinations (maybe even many of the random combinations) might look bad. What is going to be the cost of losing that “human element” in the design of a perfect landing page? To be honest with you, we don’t know yet. That’s the scary part. We’re definitely going to keep an eye out for any particularly poor performing variable possibilities and remove them from the rotation as soon as possible. The other thing we are going to do is to start small. Instead of throwing a bunch of hugely different variables into the mix right of the bat, we’re going to start with a few minor adjustments to a standard landing page.
So, why go through all this work? Because, at the end of the day, what’s important isn’t how good our design team feels about the landing pages, but instead, how well our landing pages are converting. With The Lab’s multivariate testing methods, we are going to take the control away from the designers and give it to the potential leads because they’re the ones who really count.