maxcdn-logo-extra-large-672x235A/B Testing is essential for data-driven marketers who want to understand what works well for their audiences. We chatted with Kris Kayyal, growth marketer at MaxCDN, who had some killer insights on using A/B testing tools to drive key business goals.

What are some things marketers need to consider before they start A/B testing?

Kris: You want to start with a hypothesis. This is your take on why you think your page isn’t performing, or customers aren’t converting, at the rates they should be. Maybe you think the homepage isn’t converting because it’s too busy, or people are abandoning your checkout because the process is too long. Whatever it is, you want to have an initial hypothesis when you’re thinking of what tests to run. Otherwise, you’ll be going in and blindly testing anything.

You can put together a hypothesis by gathering qualitative information, like feedback from customers about why they abandoned, or quantitative information, like data showing high bounce rates on your homepage.

The next step is to list out all your A/B testing ideas in a sheet, then rate them based on how easy they are to implement (effort) and which will give the biggest payoff (reward). This will help you pick tests that give the biggest reward for the least bit of effort.

When you’re ready to begin, first start with just one test. There are many articles out there that discuss how long a test should be run, but as a general rule, the bigger the change you’re implementing, the quicker the test will probably run.

What type of A/B tests have you run for MaxCDN that have been particularly impactful?

Kris: Some of the best A/B tests we’ve run on the MaxCDN site are headline tests, where we compare completely different page designs and different funnels. Really, you can test anything and everything: what works for our business may be very different from what works for yours. As long as you start with a data-based assumption, you’ll find where your leverage points are.

What type of business results have you seen from A/B testing?

Kris: A/B testing has helped us to drive increases in key business areas: leads, sales, revenue, and average order value. Some simple tests gave us a nice 20% boost on sales when we removed navigation tabs from the header and footer of our site. These results led us to the hypothesis that simpler is better, which we’ve since used to further implement testing and changes in other areas. So far, our tests show that it’s working.

What KPIs are most important when monitoring the success of your A/B tests?

Kris: You need to monitor all your KPIs when doing A/B tests, because one metric might go up while another goes down. For example, in one of our tests we compared two different headlines. One headline gave us more enterprise leads, while the other gave us more sales and revenue. If we had only measured for sales conversions and revenue, we would have missed the effect the alternate headline had on driving enterprise leads. For us, the most important KPIs to monitor are number of leads, quality of leads, total sales, and total revenue. We want to know how many people continue down the conversion funnel, and what affect our different strategies have on that.


Want more great insights on A/B Testing? Check out our webinar on demand: Testing Success: Top tips for unlocking revenue.