• English
    • Español
  • All you need to know about A/B testing

    3 November, 2017

    The A/B tests consist of producing two versions of the same element and launching it to see how the users behave and which they prefer. It is widely used to optimize email marketing strategies or to improve the effectiveness of a landing page. You can also use it to detect different problems of a website, for example, finding out what elements cause a high bounce rate; why users do not subscribe to the newsletters and if it is related to elements such as the design, size and type of sources, excess of information, among other problems.



    With the A / B tests on the landing page you can check the effectiveness, for example, a button to join the red newsletter and a green button. And thanks to the statistics provided by both, it will be possible to evaluate which one works best and why. You can change all kinds of elements:

    – The colors, both the background and the different elements

    – The text or the location of the button

    – The size and location of the images

    – The structure of the contact form, the whole web or email

    However, you can only make two versions so you can compare them correctly and easily identify the error or the element that works the worst. If you are willing to modify more than one element, it would be a multivariate test, not an A / B test, which is a bit more complicated.



    If you are going to run an email marketing campaign and want to do A / B tests to see which mail would work best, you have to create two emails and modify the information or the design. The main elements that you can modify in an email are:

    – Name and email of the sender

    – Subject or title of the campaign

    – Header image

    – Text – Images

    – CTA or calls to action

    – Landing page

    Keep in mind that you can change almost all the elements of an email from the free version of any email marketing service platform, but if you want to modify the structure of the content, style and design you should hire the version of payment.



     Try that the two versions have a significant enough difference to be able to decide precisely what you should do.

     Before testing, establish the parameters that will serve to measure the effectiveness of the versions- such as conversion rates, number of subscriptions to the newsletter – as well as the period of time you will use to observe the data.

     Do all the A / B tests that you consider necessary, do not settle for the first one even if you have managed to improve the conversion ratios.

     Do not forget to also randomly distribute the versions on those platforms where you are advertising. For example, if you are investing advertising on Facebook, what better way to know how each of the versions could work.


    Share This: