Does A/B Testing Hurt My SEO Rankings?

Tweet about this on TwitterShare on FacebookShare on LinkedInGoogle+Email to someone

The short answer is yes A/B testing can hurt SEO rankings, but if you keep in mind and follow best practices below, you have nothing to worry about. Back in 2012, Google stated “…your tests having little or no impact on your site in search results.”

The recommendations Google states to follow are:

  • No cloaking This may sound ominous but simply means don’t treat the Googlebot differently than your visitors. For example: don’t create tests that only target the Googlebot in order to test different ways of increasing your search rankings.
  • Use rel=“canonical”
    This recommendation would only apply if you are testing different URLs against each other. For example if you are testing page /home.html vs. /home2.html. Most modern methods of testing dynamically split traffic to different test versions at the page level using JavaScript so this is not an issue.
  • Use 302s, not 301s
    This recommendation is in relation to another seldom used testing method where you would send all of your traffic to /home.html but then redirect half of the traffic to /home2.html for example. In this case you would want to use the 302 temporary redirect not the 301 permanent redirect.
  • Only run the experiment as long as necessary
    This means do not keep running a test after statistical significance has been achieved. Similar to the first recommendation, Google may see running a test indefinitely as a way of deceiving or treating the Googlebot differently. If you are doing testing for the sake of increasing your sites performance you wouldn’t want to keep running a winning/losing test anyways!

blog-does-testing-hurt-seo

The long and short of it…

Depending on your strategy and where you are testing you may not even need to worry about these issues/recommendations. For example, much of the testing we perform is on PPC (Pay Per Click) specific landing pages that serve a very specific goal – because of this we don’t even want those pages ranking or competing with the client’s corporate SEO efforts.

For the testing that we perform on our clients’ corporate sites, we are using a third party testing platform that significantly improves our testing efficiency and analysis capabilities. The platform also uses best practices for dealing with search engines so we do not have to manually set canonical urls, redirects etc. or worry about this every time we run tests.

When evaluating the testing methods / tools you want to use, make sure it is clear how this will be addressed. Will you manually have to include code to each test to ensure compliance? How does a potential third party tool run tests and address issues like cloaking?

My final thought – Don’t let SEO implications scare you away from testing! Testing is invaluable and an integral part of successful online marketing. With the proper knowledge you don’t have anything to worry about regarding SEO rankings.

Comments are closed.