split testing

Split Testing: Factors to Consider from an SEO Perspective

July 9, 2019 - 4  min reading time - by Evaldas Mockus
Home > Technical SEO > Split Testing: Factors to Consider for SEO

Incorporating split testing (A/B) in your ecommerce store can help take your sales and customer engagements to levels that you may not have experienced before.

A/B testing helps you determine which protocols are most relevant to your customers so that you can continuously duplicate and improve conversion rates, brand engagement, customer satisfaction and return on investment.

Here’s the problem. If your company is only focusing on the split test, it can negatively impact your Google rankings. That’s why it is vital to factor SEO into your split testing endeavors.

SEO Considerations with Split Testing for Ecommerce

Thankfully, Google encourages ecommerce stores to provide the best user experience for their customers. That means that A/B testing is encouraged and will help your search engine results if done correctly. Plus, Google offers its own testing platform, Google Optimize, for website designers to test many aspects of their web design.

Google also understands that some A/B testing can hurt your search engine results. In response, they provide a handy list of best practices so that your store can end up in as many searches as their algorithm will allow.

Two general principles to consider are:

1. Don’t discourage Google bots from seeing your tests.

Cloaking your test — showing one set of content to Google and a different set of content to humans — can result in having your page removed from Google search completely. Not something you want to happen.

2. Keep your test pages similar to each other.

If you create a test, and the pages are entirely different, it will be difficult for you to ascertain why one page performed better than another. Also, it might confuse Google into thinking you are attempting to manipulate the system. This behavior could also result in getting your page removed from Google search results.

A/B Testing and URL Redirects

One way to perform a split test is with a URL redirect. This type of test relies on redirecting a customer to a different web address, and it works exceptionally well if you are testing a new design, landing page or product page.

To avoid having the redirected page removed from Google search results, follow these tips.

1. Don’t use a robots.txt file to block Google from crawling your page.

Do not use the “disallow” or “noindex” command on your test page in an attempt to discourage bots from indexing your page. Instead, practice the following protocols, and you won’t have to worry about the alternative version showing up in Google search results.

2. Use the rel=”canonical” link attribute with the test page.

This attribute allows Google to understand that you are creating a test. When Google crawls your “test” page (with the canonical attribute), the bots will not show the test page in the search results. Instead, the original URL page will show up as usual.

3. Use temporary redirects so that test pages don’t get indexed.

A 302 redirect and a JavaScript redirect are temporary redirects that will assure the test pages are not seen as permanent fixtures on your store or blog.

4. Implement the beneficial changes.

Once you’ve completed your A/B test, incorporate any favorable changes on the permanent page (A). Add a permanent redirect (301) from the test page (B) that leads back to A. If you plan to do further testing, use a temporary redirect (302).

[Case Study] Improving rankings, organic visits and sales with log files analysis

In the beginning of 2017, the team of TutorFair.com asked for Omi Sido’ SEO services to help them. Their website was struggling with rankings and organic visits.

Split Test with JavaScript Overlay Protocols

JavaScript may be the simplest and most effective way to engage in a split test. With JavaScript testing, there is no need to change your URLs or create redirects. The overlay does all the work.

Even better, JavaScript testing does not adversely affect your search engine optimization attempts. However, there are still a few safety tips to employ so that Google doesn’t think you are engaging in deceptive practices.

  • Don’t change the overall structure or content of the page.
  • Don’t modify those parts of the page that are causing the page to rank.
  • Keep your testing periods relatively short to avoid the perception of deception.

If you have pages that aren’t generating a bunch of traffic, you have more freedom to break these rules. The same applies to pages that are not indexed. These pages might include shopping carts, order pages and order confirmations.

Mobile SEO Considerations

There are some unique challenges with A/B testing since Google has begun implementing Mobile First Indexing.

Regardless of the device that your customer is using to browse your store, Google will index your site based on the mobile content.

Some pertinent elements of search engine optimization may get clumsy for the user. The temptation may be to improve the user experience by removing web content that uses a lot of space on the phone or doesn’t appear visually pleasing. However, this could cause your search engine results (and traffic) to be negatively affected.

Mobile search engine optimization, and how it relates to A/B testing, is an area you will want to keep your eye on in marketing trends.

A/B Testing with email marketing tools

Choose an email marketing tool that offers a personalized A/B testing experience that does not affect your search engine optimization.

With an email split test, you can find out what moves your customers to engage with your message. Like with page titles when ranking in search engine results, a simple tweak of a headline can have a dramatic effect on your open rates, which will ultimately affect your sales.

Wrapping up

We hope this information was helpful and that you walk away with some great things to avoid to keep your search engine rankings up when engaging in A/B testing in your ecommerce store.

If we haven’t mentioned it yet: we definitely recommend that you test, test, test — until you can’t improve more. The truth is, there is always room for improvement in ecommerce — especially if you are looking for revenue growth.

Evaldas Mockus See all their articles
Evaldas Mockus is an Experienced Search Engine Optimization (SEO) Specialist with a demonstrated history of working in the information technology and Saas companies. Currently, he is associated with OmniSend, an e-commerce marketing automation platform built for growing e-commerce businesses.
Related subjects: