How to do A / B testing in 2021?

If you haven’t encountered the term A / B testing yet, it’s a good idea to introduce it right from the start. It is a method that is widely used in marketing, which most often tests software and web applications, but there are no limits to its use.

In order to maximize user comfort during the usage of an app, and especially to increase conversions. The result, which this testing should significantly contribute to, should be the creation of a product in which users can intuitively perform all the actions you require – for example, click on advertising banners to increase your advertising revenue or open the inquiry form as often as possible, which in turn increases the chances of ordering the goods or services you offer. Remember the sales funnel?

Specifically, this method consists in showing one part of the users of the test program one variant of the program (A), while the other one presents a different variant (B). Two basic rules are followed – on one hand, this testing must be performed simultaneously so that the results are not distorted due to the time lag between the individual tests. At the same time, you only need to test a few functional or design elements at a time, ideally, if you can afford it with respect to time and money, always test only one element. So the point of A / B testing is definitely not to show people two completely different products and ask them to choose the better one but instead optimize the components leading to the purchase.

Why to still test like this in 2021?

The main reason for A / B testing is the fact that only because of it, you can truely get to know your customers. You will test what works for them and what does not. Yes, while there are general guidelines for which marketing tools work for people and which don’t, they are general and may not apply to your specific customer groups. In short, each user base is unique and has different requirements and preferences. Get to know them!

By giving your users the opportunity to co-decide on the final form of your product through A / B tests, you will relatively easily find a variant that will suit them best. And you will then be able to look forward to much higher conversions and profits, even by a few tens of percent! At the same time, you can use A / B tests to resolve disputes in the marketing department, where you will meet two different variants of newsletters or advertising texts – simply use the A / B test to let your customers decide which of the variants is more suitable.

Another reason in favor of this type of testing is that it can be fully deployed in normal operation, so you do not incur additional costs otherwise associated with testing outside normal operation. It will also be pleasing to the fact that this testing is designed to be refined with each subsequent iteration, while at the same time reducing the influence of external factors on the results. In short, the higher the iteration (repetition) of this test, the better the final product can be expected and well your customers are the testers themselves!

What can be tested with A / B testing?

As mentioned at the beginning, this test can be used for virtually anything. However, there are certain situations in which A / B testing is a little more suitable than ever. Specifically, we have selected three areas for you in which A / B testing makes the most sense in 2021. Namely:

A / B banner testing

Banner ads still form a significant part of the income of most websites, so their appearance should not be underestimated. It is therefore necessary to prepare several variants of banners and then, ideally, have them tested using A / B tests, and always place the winning one, or the one with the highest rate of clicks, on the website. By placing the winning banner on the website, the work should definitely not end for you yet, you need to start creating a new banner immediately, which is then included in the A / B testing against the winner of the previous one – only then will you be able to find it after some time. really the most suitable variant. And no by banners we do not mean the typical disturbing ads but it can be e.g. category banners leading to the promoted product category in your online shop!

A / B newsletter tests

The advantage of A / B testing of newsletters is the fact that you can use an existing customer database for it. So if you already have some recipients of your newsletters, you can send small parts of them to two different variants of one newsletter and see what conversion rate they will credit to your account. After evaluating the results, all you have to do is send the remaining customers the variant that was more successful in the A / B test.

A / B testing of PPC ads

PPC ads are without a doubt the most frequently chosen way to let the world know about your services today. However, all you need to do is release an ad that will attract and click on people – and A / B testing will help you with that. So create two ad variations and activate both for the same audience at the same time, allowing a small number of users to see them for a while. You can repeat this process several times. Over time, evaluate the results and select the more effective ad (i.e. with a higher conversion rate) and then release it to the world.

When does A / B testing make sense to me?

It follows from the logic of the matter that the larger the database of potential testers (users, newsletter subscribers, etc.) of a company, the more it pays for A / B testing, because on a larger sample it is possible to achieve even more accurate results. It is not for nothing that giants such as Google, Microsoft, Amazon or Netflix commonly use this testing in their projects. However, with the increasing improvement of tools for A / B testing, the way to A / B testing is opening up for smaller companies and individuals.

Thanks to more effective tools for A / B testing, practically anyone can throw themselves into it today and it can be used in any project, except for complete beginners and the smallest ones – despite the otherwise high return on investment of this method of testing, it might not pay off for them. Simply put, the cost of procuring testers and testing itself is likely to outweigh its financial benefits.

In short, A / B testing definitely makes sense for all companies and individuals focused on marketing or website development, with an already established user base. They will also welcome the main advantages of this tool, which, in addition to the high informative value of the tests, also include the ease of running them and the ease of evaluating the results.

What tools to use for A / B testing?

Of course, there are a number of tools that enable A / B testing, but we have selected the 3 best and most popular for you, which will help you find the ideal layout of your website, text and appearance of your PPC ad or the most appropriate form of your distribution. newsletter.


As the name of this platform suggests, it is a tool that aims to optimize your final product through experimental tests, which include A / B testing. Optimizely is designed not only for product managers, but also for marketers, developers and analysts – it can make it easier for all of them to find this best possible solution. Due to its higher price (it is always individual, but ranges from EUR 1,500 per month upwards), it is practically exclusively suitable for larger companies.

There are two variants of this tool. The first is purely focused on optimizing and customizing websites. In addition to the ability to compare two variants of a website at once and find out which suits your visitors more, there is also a visual editor, which can be used to make the necessary changes to the page without the need for knowledge of coding. The second of the packages then has a much wider range of uses – it can be used to optimize, for example, applications for mobile phones, while allowing you to run an unlimited number of concurrent tests.

Visual website optimizer (VWO)

It is the current market leader in A / B testing tools. This is due, among other things, to a much more affordable price compared to the competition, the package with professional testing tools costs $ 199 per month (approximately EUR 180), and there is also the opportunity to try VWO completely free of charge. However, to maximize the number of conversions, it’s a good idea to use the VWO Testing package in conjunction with VWO Insight, as they combine not only the conversion rates for the variations, but also how visitors behaved on the site and where they spent most of their time.

As in the case of Optimizely, VWO also includes a visual editor, with which you can create, for example, a visually attractive pop-up window. Of course, there is also a code editor. The number of interactive and graphically successful graphs with test results, which will make it easier for you to choose the optimal variant, is definitely worth a compliment.

Zoho PageSense

This tool is an ideal tool for beginning marketers or web developers who would like to get into the secrets of A / B testing, because it is extremely cheap – the basic package will cost you only $ 17 per month (approximately EUR 15)! There is also a 30-day free trial.

As for the functions, despite its low price, it is definitely not a source. There is an unlimited number of executable experiments at once, as well as a visual editor or heatmap monitoring the behavior of users on the web. And there are even more advanced features for better targeting and segmentation.

[The original article was written in English. To other languages it was automatically translated by Google Translate]

Leave a Reply

Your email address will not be published. Required fields are marked *