(Think Gladiator...but with ideas instead of Russell Crowe.)
But before you jump in, we've detailed a process for determining which pages to A/B test in the first place, along with some suggestions on actual tests to run.
Where to start:
We use HubSpot to compare our most-visited offers against our highest-converting offers and then look for red flags.
In this case, a red flag would be a page that is driving a ton of traffic, but converting very few leads.
Essentially, a lack of conversions on a highly visited page signals to us that we're not doing enough on this page to persuade them to take action. These are the pages you want to be test because they have the greatest potential in terms of improvement.
Testing pages that are already seeing a lot of traction will help you to reveal more valuable, reliable insights compared to a smaller sample size.
Think about it. Let's say you're preparing to give a speech for an upcoming conference and you want to practice in front of a few of your coworkers. You could run it by 5 and they might all happen to love it, or you could try it out in front of 20 and find that the reception is a bit more divided.
What to test:
You've nailed down which pages you want to test, now what?
Before you start testing elements at random, there's a bit of research that can be done to help you make more informed-decisions.
CrazyEgg is a heat-map analytics platform that provides users with a visual account of how people are behaving on any given page. With insight in regards to where people are clicking, or how much of the page they're actually seeing, you can begin to identify specific parts of the page that need work.
For example, let's say that you notice a significant drop off in views below the fold. If your form lives there, you might want to consider testing the placement by finding a way to incorporate it above the fold.
Looking for more ideas on what to test? Check out these suggestions:
1) Content length
When QuickSprout's Neil Patel tested two versions of his homepage (1,292 words against 488 words) he was surprised to find that the longer variation not only converted 7.6% better than the shortened variation, but also produced more qualified leads.
While serpIQ found that the average content length for a page that ranks in the top 10 (for any keyword) on Google contains at least 2,000 words, consider experimenting with length on your pages to see what works. You may find that your audience is more inclined to consume more digestible content of specific pages.
2) Button text
When approaching button text, CopyHacker's Joanna Wiebe suggest that you start by putting yourself in the visitor's shoes.
Rather than default to "Submit" or "Sign Up", Wiebe suggests that you aim to explicitly state what it is that your visitors want.
"I want to _______"
Your button text should aim to fill in the blank.
With that said, consider testing your current button text against this approach to see what your visitor's respond best to.
3) Number of form fields
Visitors don't want to be tasked with a never-ending form in order to gain access to your content.
To help reduce some friction, consider experimenting with the number of required form fields you ask for. While you certainly don't want to sacrifice the quality of the lead by reducing the information down too far, often times there is an unnecessary step (or two) that can be cut out.