A/B-testing for dummies

Have you ever asked yourself what A/B testing is all about? We have listed the most important things for you.

What is A/B testing?

You can simply say that two versions of a page are tested to find out which one is the most popular with the public. For example, we at Marsmedia Webdesign Amsterdam can test whether people believe more or less in our expertise, are inclined to contact us, start a chat session or call us when we mention the word WordPress on our site. For more information about why we use WordPress, please read our previous blog post.

On Optimizely it says the following: “A/B-testing (also known as ‘split testing’ or ‘bucket testing’) is a method of comparing two versions of a web page or app to determine which of the two performs better. A/B testing is essential in an experiment in which two or more variations of the same page are randomly displayed to users, and where the statistics are then used to determine which version offers a higher conversion rate”.

Why would a web designer use Amsterdam A/B testing?

Websites always have a certain purpose, something they strive for or focus on. For example, webshops want their visitors to buy products, software developers want potential customers to download a trial version and then buy the programme, and news sites want their readers to click on the advertisements or purchase a paid subscription. Testing and analysis are therefore in the heart of Marsmedia Webdesign Amsterdam. We not only try to deliver a visually attractive website, but also to take care of the site afterwards. What does the website contribute to the conversion of leads to customers? How much will the new site yield for the company? We believe it is important to pay attention to the figures, the statistics and try to understand them. A/B-testing ensures that website optimisation is no longer guesswork and enables informed decisions that ensure that motto is no longer “we think” but “we know”. Even the smallest change on a landing page or Webdesign Amsterdam can have a big impact on the leads, revenue and profit that the page in question generates.

How A/B testing works

1) Analyse: work with a research tool such as Google Analytics to analyse your website and identify the pain points in the conversion process. For example, you can identify the pages with the highest bounce rate (see below). Let’s assume that your home page has an unusually high bounce rate.

2) Investigate the behaviour of your visitors by using tools such as Heatmaps, User Sessions, Form Analysis & On-Page Surveys, and find out why you fail to convert leads. For example: “The CTA (Call-to-Action) button is not prominently displayed on the home page”.

3) Suppose: use the data you’ve acquired to build a hypothesis or strategy that will increase your conversion. For example: “Enlarging the CTA button will put it in the spotlight and convert more leads”.

4) Test the assumption: design a second version of the current page and use it in the A/B test together with the original page to take the test to the sum. For example: “A/B test your original home page together with another version equipped with a larger CTA button”. Determine the test period based on the number of monthly visitors, your current conversion and the expected result.

5) Conclusion: analyse the results of the A/B test and find out which version performed best. If there is a clear winner, you should no longer doubt which page to implement. If the test is not decisive, go back to step three and work out a new assumption.

Stumbling points at Webdesign Amsterdam A/B-testing

There are also a number of pain points with regard to A/B testing. You would think: “The more you test, the better” but unfortunately that is not the case. It is important to make a distinction between a lot of short term experiments or long term testing. The former gives you quick results, but the latter gives you more insight. If you only focus on speed, you spend less time building up the experiment and you might miss the ball. Another problem that often occurs is that when you try to test a host of different ideas at the same time, you put your ability to validate a single idea at risk and take advantage of it.

Another myth is that only the effect on conversion counts. A/B-testing, for example, can also have a major impact on the user flow, something that is highly valued at Marsmedia Webdesign Amsterdam. Secondary objectives (number of clicks, number of visits, filled in fields, etc.) can also mean something. It all depends on what exactly you want to focus on. All these secondary insights can help you come up with a number of follow-up experiments, identify pain points and gain a better understanding of how visitors navigate your site. And don’t just test any idea, because a lot of time and preparation precedes even the smallest experiments. Read it on the next article if you are interested in an example.

What is the bounce rate?

What is a ‘bounce’? And what is the bounce rate? Marsmedia Webdesign Amsterdam decided to take a look at the support pages of Google Analytics and found the following: “A bounce is a session on a single page of your site. In Analytics, a bounce is specifically seen as a session where a single request is sent to the Analytics server, for example, when a visitor opens a page on your site and then leaves again without sending any other request to the Analytics server during that same session.

The bounce rate, on the other hand, is the number of those bounce sessions divided by the total number of sessions, i.e., the percentage of all sessions on your site where users viewed only one page and sent only one request to the Analytics server”.

Marsmedia Webdesign Amsterdam hopes you have learned something from this article! Don’t hesitate to contact us, we are happy to help you.