Google Content Experiments: A/B Testing

Every online retailer that wants to increase sales must ensure that their design elements are being tested regularly. This can be achieved in a variety of ways, including paid, custom-coded and free options. Google’s Website Optimizer (GWO) was a free option that offered better results in the past. Google’s Website Optimizer (GWO) was discontinued in August 2012 and replaced by Content Experiments. Google claims that the move allows anyone to perform A/B testing, without any technical knowledge. It is also fully integrated with Google Analytics, while GWO had its own reporting areas outside of Analytics.

The move makes it much easier to setup an A/B test. There is less tracking code and the setup process is quicker. GWO supporters aren’t convinced Content Experiments is worth the effort beyond static landing pages. GWO works well with static and dynamic pages, but Content Experiments requires a unique URL landing page for every test variation. This is difficult with dynamically generated content. It makes the claim that “no technical skills required” seem to only apply to non-dynamic content. This makes Content Experiments impossible for ecommerce sites that are primarily composed of dynamic data driven content.

See our products:

  1. Magento 2 point of sale
  2. Shopify pos integration
  3. Bigcommerce pos app
  4. Woocommerce POS
  5. Netsuite POS
  6. Commercetools POS

Content Experiments is a valuable tool, even though it falls short in some areas. Let’s take a look at Content Experiments.


1) Content Experiments uses the Google Analytics tracking instead of placing multiple tracking code snippets on different sections of a page or site, as was the case for GWO. It’s as easy as adding a small tag in the original tracking script.

2) Content Experiments includes a wizard to help you set up new tests and guide you through the process.

It is easy to set up an experiment and takes only a few minutes if you are not creating the variant page. (Read more at cons.

Each test uses a conversion indicator from Google Analytics.

5) URL referral information does not disappear during redirection, even dynamic parameters.

6) Users will quickly see new experiment data with Google Analytics Premium, the premium version.


1) Multivariate testing is not an option.

2) Each experiment must have its own URL. This means that each test will require its own landing page. A product page can be created from one template. This dynamically inserts product IDs into the URL to display content. However, the page technically remains the exact same. To create a test variant, one would have to duplicate the URL. This defeats the purpose of dynamic content. Then, add to the URL some kind of test indicator to let Google know what to display.

Here’s an example:

This would require that the page’s programming code be modified in order to make it work on a dynamic website. It could also cause duplicate content issues with Google if Google finds the URL and indexes it. Google has a solution: make sure that the variant is properly tagged so that robots can’t index it. Technical? Technical?

3) Google Analytics’ Content Experiments goals mean that users will soon reach the maximum permitted.

4) At the moment, users can only test five variations. It is important to carefully consider the variants that are being tested. Many variations can be used in large volumes.

5) Google Analytics users who are not paying for it must wait longer before they can receive experiment data results. This could lead to lost sales and unnecessary expenses.

What online sellers can do.

There are many split testing tools that can be used to meet the needs of online retailers. You can make page changes without A/B testing software and just use Google Annotations for the date that the test began. This allows all traffic to view the same test, rather than splitting it. In which case one would see version A while the other would see version B. However, the trending reports in Google Analytics can provide very accurate and telling results.

This method of testing is best used to test only one item. It is not easy to identify which element(s) had a positive/negative impact when you test multiple elements simultaneously on a page.

Online retailers should test everything and not be satisfied with the results. Keep it simple, and remember to test thoroughly. It doesn’t matter what method you choose, and it doesn’t need to be complicated. Log into Google Analytics to learn more about Google Content Experiments.