Of late many have asked Google whether website testing affects the sites performance in terms of search results. Google is thrilled to know that you are indeed testing and asking these questions pertaining to A/Btesting and multivariate testing. Both types of testing are a great way to know if you are offering the thing to viewers that they are looking for. Before we go further let us see what this testing is all about.
Website testing is normally done by trying out various versions of your website so that you know exactly how customers react to each version. You normally use a software to collect data and check which version prompts users to take an action that you want them to take. This gives you a good idea of which version results in increased sales. Once you have zeroed in on the best version, which results in increased purchases, you update your website likewise.
A/B testing is done by creating various versions of your page. When viewers try to access your website you redirect them to several other pages with rearranged content and then based on their behavior on each page you decide which page with its content is best for you.
In Multivariate testing you typically use specialized software to change parts of your website like the cart button, pictures on the website, or text blocks. The software then analyses and determines which combination returns the results in terms of increased purchases and then you decide on the most effective version of your website.
Now the main issue that you are concerned about is how Googlebot sees all this different variations in content and whether it affects your ranking. Here are some guidelines that will clarify your doubts.
Make sure that you always serve original content whenever the user agent is Googlebot. Be judicious whenever you serve content and do not make any distinction between general users and Googlebots.
When you are running the A/B test make sure you use the rel=”canonical” link attribute on all your alternate URLs so that it is crystal clear that the original URL is the preferred version.
Make sure that you inform the users whenever you redirect them to a different URL. This can be done by using a 302 (temporary) redirect. You should under no circumstances use the 301 (permanent) redirect. This will tell search engines that the redirect is only temporary.
Do not run the tests for too long periods of time. Just run it till you have sufficient data to analyze. This will, however, depend on the conversion rates on your website and the amount of traffic you have on your website. Once the test is done update the website with the desired content and erase all the traces of the test. In case you do not remove the variants, Google will automatically assume that you are deliberately doing this to deceive their search engine.
These are some of the suggestions that you need to carefully think about when conducting tests on your website.
