09 Sep A/B Testing Magento product pages with Google Analytics
A/B testing is an essential method in eCommerce for increasing conversion rates. Follow our easy guide to setup an A/B test on your Magento product page. You can also easily apply the principles to the category pages, home page and landing pages.
It is essential that you only A/B test one element at a time, e.g. product photos, product description, fonts, price etc.
To perform this test you will need to be able to add code into the <head> tag of individual product pages. Please ask your developer if you do not know how to do this.
Step 1 – Duplicate the product you wish to test
Locate the product and then press the ‘Duplicate’ button.
Change the product SKU and URL key, ensuring you untick ‘create URL redirect’.
So that your test page will not be crawled for SEO purposes, add the following layout update to the ‘Design’ tab.
To ensure that the test product will only be accessible by the URL, set ‘Visibility’ to ‘Catalog’, then remove the test product from all categories so it won’t actually show in the catalog.
Add the following code to the <head> tag of the duplicated product.
<meta name=”robots” content=”noindex, nofollow”>
Make the change you wish to test and save the duplicated product.
Step 2 – Setup the test in Google Analytics
Login to Google Analytics and select ‘Behaviour’, and ‘Experiments’.
Select ‘Create experiment’. Name the test and select which metric you wish to measure – in our case eCommerce Revenue.
Click next step, and add in the two variation URL’s.
Click next step, and ‘manually insert code’.
Add the code within the <head> tag of the original product.
Click next step, and ‘start experiment’ to save.
Step 3 –Analysing Test Results
Leave the experiment running for a suitable period of time, i.e. if you have a high traffic site you will have enough data quicker than a low traffic site.
Once you go back into the experiment created, you will be able to see a graph of the performance. You can look at metrics such as revenue and conversion rate to identify the better performing product. The data can also be segmented to see if there are any significant differences for example with new V’s returning visitors, and mobile V’s desktop.
Hopefully you will find this guide useful, and you will now have some ideas of what to test out on your own Magento site.
Remember not to be complacent – what you may think is a good idea may not be the same in your customer’s eyes. The data never (or at least rarely) lies – so test, test and test again.