How to Implement Page Tests

Written by Lemuel Galpo

November 6, 2014

In this webinar, Convert.com and Jana Fung (of MixRank) teamed up to talk about the scientific method behind conversion optimization. You can view the complete webinar here. We already had previous discussions about the scientific method behind conversion optimization. We talked about the types of tests we need to do, how to conduct successful tests, when to stop, and where to get ideas for new landing pages. Now, we will show you how to implement page tests by making your landing pages dynamic.

[Tweet “Before you start testing your landing page, it is very important to know what you are going to test”]

Preparation

It’s almost funny if it was not really sad, most people just start a test because they want to test, and will come up with test ideas in the visual editor. After determining what to test, you can then start creating variations of your page by changing a variable, like adding or removing, a button, a logo, or a text. You can even implement this in mobile. You can check how it looks like on your mobile. Once you verify the change, you can start testing. Convert.com can help you in that area. Just set who your targets are, who is your target audience; and the system picks up the audience. You can also set it so that only new visitors can see the change. You can also refine your settings by running it only in a single day; or only on a specific group of browsers. Convert.com has lots of options that you can try on. We can make it simple, or we can make it advanced. You can also set goals here, and there are a lot of default goals to choose from. Things like increasing engagement, decreasing bounce rate or new goals. You can even set multiple goals.

A/B Testing

Take this as an example, we have the original version, a no logo variation, an awesome variation, and a 2014 version (as discussed in the webinar); a total of 4 different versions. We then perform an A/B test on each of these versions. So what is A/B testing again?

Quoting from Nancy Medica:

“A/B tests allows you to measure which version of a landing page converts better. For this, you have to know what your conversion rate is.”

Going back to the example, the main goal of the test is to get people to search. We’d like to keep this test run for at least a week; 2 weeks is fine as well. Around 100% traffic is included. By using A/B testing, this means that each original variation gets 25% more traffic. Once you have installed the code on the site, you can now schedule tests. You can run it directly, or schedule one for the future. You can also keep a draft and save the test. Now you can basically setup a whole test and see the report. The report can show conversion rates, revenues, and product order items.

The idea here is that you can:

Setup a test really fast
Be able to do a couple of tests every month
And learn new things from those tests

Integration of other tools

You can also integrate Google Analytics so you can monitor things like revenue or discuss variables. You may also integrate Kissmetrics if you go to our Convert homepage. You can also see all the integration we filled with WordPress, Magento, and many of this are e-commerce and CMS system. So give it a try.

Trying out things based on the best practices we previously discussed is also a really good thing here. The importance of setting a baseline, knowing what the audience wants, and the creation of a hypothesis based on this assumptions is very, very important. If you want to learn more from this webinar, you can view the entire presentation here.