A “split test,” or an “A/B test,” can often be one of your best friends when you’re trying to figure out what works, and what doesn’t. Here’s how it works:
- The page you’re looking at right now has one type of content, one link, and one title. In a split test, there could be two or more versions of this article. The differences in those articles could be something like the link title that leads to it, the subject of the article, the type of image used in the article’s header, and other variables. Then users can distinguish which version is randomized.
- Over time, results are added up to see which version of the site gives you the effect that you’re after. There are different metrics that you can use to normally measure a site’s performance, and with split testing, it’s applied to both versions. For example, some popular online blogs and link aggregators will test several versions of a link to a story with different headlines to see which version actually gets more traffic. The winner is usually selected and used. In some cases, the process happens the same day when an article was posted.
- The changes aren’t always dramatic. One version of a page with a call to action and an information form for visitors to fill out could have the email address and phone number fields reversed. A slightly different background color could be used, or a different video presentation may play on the page upon loading. When you’re fine tuning a website, these more subtle differences can be informative for building a stronger strategy on how to engage your visitors and convert them into buyers.
Whether or not you should split test usually depends on how much you want to invest into your traffic and your marketing as a whole. The answer, in most cases, will be that it’s a good idea to do it at least once. The entire process can be very educational, and you may be able to avoid some serious missteps in the future.
Which Elements Get Split Tested the Most?
- Visuals, including graphics, videos, and choice of color.
- Layouts, including side bars, headers, forms, and buttons.
- Text, including the actual content, any calls to action that your page has, and descriptions. Even language differences can be split tested, if your primary demographic is bilingual.
- Flow, which covers how visitors interact with the site, click on links, where they go, and if they go to where you need them to.
Split testing can also be useful if you want to promote a new product or deal that your business is offering, because they can be done quickly if prepared ahead of time. You can split test a site in as quickly as a day, as previously mentioned, but the larger the differences between pages, the more time you should take to further narrow down your A/B selection for a result you want.