AdvertisingContent MarketingDigital MarketingWebsites

A/B Landing Page Testing with Google Content Experiments

By June 11, 2013 2 Comments

Online marketing pros grow accustomed to a daily “the new best thing is…” email ritual – some new tool, gadget, or idea that’s set to propel Internet marketing to the next level. Some of these are great, some are just okay, and some just don’t pan out. To be honest, it’s exhausting to try to stay atop all of the new possibilities. With so many ideas hitting daily, even the good ones can get lost amongst the online chatter.

So today, I’m going to take a new look at a not-so-new tool that could really change your approach to landing page optimization (if it hasn’t already): Google Content Experiments. Content Experiments was released about a year ago as an updated version of Google’s older Website Optimizer.

Some background… While the need for some landing page changes is obvious, it’s generally wise to let data drive the decision. Maybe you’re thinking of replacing an existing landing page offer to increase sales. Without testing, which offer is stronger is just a matter of speculation. So it’s time to test.CY 6-13

Traditional landing page testing involves analyzing historical landing page data, replacing the page, and then comparing the two sets of results. So in our example, we’d replace landing page A with landing page B, then use Analytics to compare landing page A’s prior results to the new results produced by landing page B.

But there’s a problem: changing market conditions and seasonal factors are just two of the variables that can skew results and lead to false conclusions about the better offer. Content Experiments to the rescue.

Content Experiments enhances the traditional switch-and-test method with a real-time, side-by-side experiment. It’s an easy process. Tell Experiments:

• the experiment objective (improving sales, conversions, or other performance metrics),
• the percent of traffic you’d like to include
• the time period for the experiment
• the degree of confidence required of the conclusion (the higher the degree of confidence, the longer the test will need to run before reaching a conclusion)
• the URL of the original page and it’s variations (up to 5)

Then add provided experiment code to one page on your website, and begin the Experiment. Content Experiments handles it from there.

At any time, you can check the status of your experiment and see how the variations are performing. When the experiment has concluded (either by reaching its time limit or coming to a conclusion with the specified degree of confidence), you can make the recommended solution permanent on the website, knowing that your choice of landing page was based on data in a side-by-side experiment.

I know it sounds strange to most people, but I love working with this stuff. If you’re ready to improve the success of your landing pages, I’d recommend experimenting, and letting data drive your decisions.

Join the discussion 2 Comments

Leave a Reply