Web page testing has become main stream amongst online marketers. A/B and multivariate testing try to find the optimal version, the one that converts the most visitors. But oftentimes, achieving results requires more than just testing. You need to know which pages and which elements, if optimized, have the highest potential to improve your overall website conversion rates.
Your usual suspects are the homepage and direct conversion pages. However, don’t bank on it without looking at the data. It’s possible that the biggest potential for improvement lies elsewhere. We at Seperia had cases where the most important pages were internal landing pages, which got the most relevant organic traffic. So start with analyzing your popular pages, and work from there. One useful analysis is which pages serve a significant role in your conversion funnels. Those pages are already instrumental in driving people to convert – can they be improved and drive more visitors down the funnel? It might be useful to segment your visitors to visits with conversions and visits without conversions. Look at the top visited pages in each segment and see if there are major discrepancies between the top pages in each. This will allow you to identify better or worse performing pages within your “high traffic” group of pages.
You can test the original layout of the page vs. any number of other layouts, and you can test the same layout with different content (text, image, video, call to action). Theoretically, you have an infinite number of versions to compose, but in the real world you have a limited number of visitors to test them on, and limited resources. So the task is to narrow them down to the versions most likely to perform. So, should you test the template first, or try to optimize the elements in the existing template? While there’s no text book answer for this, I would recommend in most cases focusing first on messaging. Try to find an optimized messaging first, and then work to optimize your layout.
Templates are usually used for pages that are informational: product/service info, articles, etc. Most likely, individual information pages are not the most important to the conversion process. Aggregated, however, they might have big potential for driving more conversions, with an optimized template and call to action. Most testing tools enable testing of multiple pages using the same template.
One thing that we’ve come to learn about testing and optimizing is that it is an iterative cycle. Once you learn that attribute A converts better than attribute B, the goal becomes to refine the way of presenting the winning attribute – i.e. continue testing. Realizing the testing and optimization is a continuous process, it’s important to plan a systematic testing strategy from the beginning. For instance, a test might tell us that a specific version beats all the alternatives in the test. However, if it is a well-designed experiment, it does not stop there, but will also tell us what specific feature made the difference, and what should be included in follow-up tests, for continuous improvement of the website’s performance.