Website Optimizer: Landing page testing for Dummies
Testing is one of our core principles at Site Traffic Control; we test everything from ad headlines to follow up email sequences to what kind of hot sauce is best on our breakfast sandwiches (the answer is Frank’s RedHot).
If you’re running a paid search marketing campaign, you know how easy it is to test different ads. AdWords (Google) and adCenter (Bing & Yahoo) allow you to come up with as many headline and ad text variations as you want, and they’ll run the ads in head-to-head competitions. A few days/weeks later, you open up your PPC account, check out the ads’ click-through and conversion rates and you pick a winner.
But what’s an easy yet effective way to test different landing pages? If you want to run a true A/B test for your landing page, you need to look outside of AdWords and adCenter. Recreating the same ads with a different destination URL won’t give you a true traffic split, and other factors such as ad position and quality score can significantly affect conversions.
To run a proper A/B landing page test, check out Website Optimizer, a free testing and optimization tool from Google.
Once you create an account you’ll need to identify three web pages:
- Your original landing page
- Your test page
- Your thank-you/conversion page
Website Optimizer will output a short snippet of code for each page. Open up the page’s HTML file and paste the code immediately after the opening tag. (If you need help with the code, shoot me an email and I’ll help you out.) Website Optimizer will automatically validate the code before launching the experiment.
Once the experiment is live, traffic will be split evenly between the two (or more) landing pages. This applies not only to paid search traffic, but also email marketing, organic traffic and any other type of page visitor.
The tool will typically need around 100 conversions to declare a winner, so you’ll have to be patient. Once the winning landing page is determined, you’ll have the option to automatically send all traffic to the higher converting page, or run a follow up test.
Initially, try testing landing pages that are as different as possible. You can always narrow down things such as headlines and images in future tests.