SEO testing can cause a lot of questions and confusion. Even the definitions change based on who you’re talking to!
We’ll start by saying there are two types of testing in digital marketing:
- A/B Testing, and
- SEO Split Testing
You may see some people use them interchangeably, but they are not the same thing.
Standard A/B testing involves creating and deploying multiple versions of the same landing page at the same time.
Different users are sent to different versions of the same page so the SEO or digital marketing team can measure things like conversion rate, for example. If one version has a significantly better conversion rate then, a test winner could be declared.
The version of the page the user lands on is really up to chance, whether they use Google search or follow a URL directly.
Googlebot can also be sent to an alternate version of the same page. It’s the content or presentation of the page that changes, not the URL.
Recommended Reading: 9 SEO Testing Ideas to Increase KPIs
If the difference between the two versions is significant, it can have a negative impact to SEO, as the content can be seen as cloaked or duplicate.
SEO split testing, on the other hand, involves grouping statistically similar pages and running a test on those groups (as opposed to on an individual page).
These are unique pages with unique URLs.
Here’s a graphic to illustrate the difference:
(A/B testing creates multiple versions of the same landing page.)
These groups need to be statistically significant — they shouldn't be random groups of pages.
SEO split testing is encouraged and does not damage your SEO program (SEO is in the name, after all!).
For example, you can test title tags and meta descriptions to see if length impacts click-through rate from the search engine results page (SERP), or experiment with keyword placement.
Follow along with our blog on how to run a successful SEO test if this type of testing is your goal.
If you're interested in A/B testing specifically, stick around! This blog answers the question, How does A/B testing affect SEO?
Does A/B Testing Hurt SEO?
A/B testing does not hurt your SEO, as long as you’re mindful of a few guidelines and considerations, per Google:
- No cloaking
- Use rel=“canonical”
- Use 302 redirects
- Be mindful of the test’s length
Another point to keep in mind is duplicate content.
We’ll explain all of these points down below, but if you follow them accordingly then testing shouldn’t impact your SEO (things like rankings).
While small changes – such as the size, color, or placement of a button or call to action – can have a large impact on user experience, these types of changes typically have a negligible impact on a web page’s search ranking.” — Google Optimize
How to Do A/B Testing the Right Way
#1. No Cloaking
We know that with user A/B testing, users are split and provided different versions of the same page.
Cookies ensure that those users land on the same version of the page on any subsequent visits.
As search bots access this page, they too are split — but cannot be cookied. As a result, they get multiple versions of the same page on subsequent visits. This creates the appearance of cloaking, where they detect the non-visible version hidden by the user testing mechanism.
Some testing vendors may tell you that this is not a problem due to the short duration of the test, but it is risky.
Google has its own testing tool called Google Optimize. They say on their resource hub that …
Google encourages constructive testing and does not view the ethical use of testing tools such as Optimize to constitute cloaking.”
#2. Use Rel="Canonical"
In the unfortunate event where a search bot finds two versions of the same page, the canonical tag helps resolve the issue at a later time.
Recommended Reading: 301 Redirects vs. Rel=Canonical Tags
#3. 302 Redirects
Since no test should last forever — as we’ll see below — it’s important to show Google that any redirect isn’t permanent.
This is where the 302 redirect comes in.
Indicating that the redirect is temporary works to prevent any loss in page rank for the original page.
#4. Test Length
Don’t let any test overstay its welcome.
If search engines see the test content and the original content, it can cause confusion — and this can result in lower rank. The longer a test runs, the bigger the issue this can become.
#5. Duplicate Content
Some tests result in cloning content and changing it slightly for the test.
It’s usually hidden from the user, but search engines can see it in the code and may flag it as duplicate content.
The short answer is: A/B testing doesn’t always hurt SEO if you take certain precautions.
You can still run a test to see what best engages or converts users, just don’t purposely confuse Googlebot.
<<Editor's note: this post was originally published in October 2018 and has since been updated.>>