Split Testing is one of the common forms of A / B testing. By comparing several versions of your web pages such as your landing pages or your home pages, the split test allows you to identify the one your internet users convert the most.
It is a very effective method to enhance advertising campaigns and increase conversion rates of business. Through these tests, you can experiment with ads from different variables, and find out what works best with each audience.
With this method, you can test and evaluate the most different elements of your advertising piece, such as images, title, positioning, copy, etc. You can also learn more about your audience and determine how best to target it to achieve effective results with a given set of ads.
Split Testing is, as its name implies, a division test: you have 2 versions of an element (A and B) and an analysis tool that gives you the real-time figures. To determine which version is the best, you run a test on both versions simultaneously. At the end of the test, you keep the version that increases your conversion rate the most.
This tool is similar to an experiment in biology class: on a plant, you pour 2 different products with the same goal: to grow the plant. Then you see the results and keep the one that has grown your plant as much as possible.
For a site, it's the same thing. So you have 2 versions of your website: A and B. A is the existing website and B is the new one. So you set up your split test and measure the audience (conversion rate, sales, bounce rate, etc.). Once the results are analyzed, you keep the version that transforms best.
Websites are full of links that direct people to different sections of the same or externally, sometimes we place them in a specific position and design intentionally but what we often do not know is if those elements are fulfilling their objective or if there is any way to improve them.
We have the metrics like Google Analytics where we can know if someone clicked on an item or not and that is complemented with the A / B Tests that is to create different versions of an element or a complete page that we will randomly show the users to know which one is performing better, measuring it in conversions such as clicks, sales made, subscriptions, and many more.
In order not to confuse the user each one will see a single version of the design that we are testing, but after let's say about 1,000 users are visiting the site we will have shown all the versions and measured which of them yielded better and then make a decision and use the best.
When the split test is launched, the traffic on your pages is randomly distributed on different versions of your pages. The performance of each is plotted and analyzed by the split testing tool to identify the most significant version that will generate the most conversions. The split test thus determines the version on which the submitted sample has the most converted.
To understand why your users do not convert, it's essential to ask what content can be a drag on conversion. Your intuition is not enough, what matters is the effect on your visitors.
By performing a split testing, you place your visitors at the heart of the decision process. The results obtained through split testing will allow you to know which versions of pages they interact with the most, and what information and design transform them into loyal customers?
The split test promises you qualitative feedback on the user experience to help you identify blocking areas and optimize your conversion tunnel.
With this type of tests, we simultaneously measure different versions instead of changing them from time to time to compare results (the old school).
You begin an AB test by deciding what you want to test, for example, the colour of the Add to Cart button on a product page.
You must then know how to evaluate performance. In this case, it can be the number of visitors who click on the button (we are talking here about micro-conversions ).
To perform the test, you display to two groups of equivalent users (randomly constituted when they visit the site) the different versions (the only difference being the colour of the button) and determine which has had the most significant impact on your performance indicators. In this example: Which button generated the most clicks?
In real life, many settings can lead a person to click. This is why randomization is essential in this context because it minimizes the chances that other factors will disturb the results.
If the concept seems simple, its implementation requires calculating the sample size necessary to achieve statistical significance.
The example of the colour of the button is a straightforward example. In fact, the tools of AB Testing allow you to modify many elements of your website and to study the impact of these changes.
The choice of test items depends on your goals. For example, if your goal is to increase the number of registrations on your site, you will test the following: length of the registration form, type of fields to fill, font size, etc. The purpose of this test is to determine what blocks the visitor to register. Is my form too long? Too much information is requested? Illegible texts? Will the green validation button make it better? All of these questions will be answered by testing the elements one by one.
Some examples of frequent tests:
Split Testing is a revolution for the web and these users. You are now the master of your content, and can quickly analyze the numbers to improve them.
Split Testing allows you to test different versions of a page hosted on various URLs, while AB Testing uses versions created from the graphical editor of the AB Testing software.
So in AB Testing, the visitor stays on the same page which is changed dynamically, while in Split Testing the visitor is transparently redirected to another page.
A typical AB test tests only two variations of a single-variable page. When testing multiple versions of the same page simultaneously, an A / B / n Test is performed. In each of the versions, only one element is modified compared to the version A. Imagine for example that the version B changes the button Call-to-Action while the version C changes the title of the page without touching the button.
AB Testing is used by many companies to improve the marketing performance of their website:
Just about every visible element of a website can be A / B tested:
Titles - They are one of the first and most direct forms of communication with the visitor.
Product pages - Your product pages must highlight your offers and allow the user to access the shopping cart with an appropriate Call-to-Action button easily.
Texts - Tell a story, tell your visitor how you feel ... words matter
Prices - You cannot offer different prices to different customers, but you can change the way they are displayed. Simple changes in the way you present your prices (using the effect of Serial Position ) can have a real impact on the average price of your basket.
Images - There is evidence that people pay more attention to images than to text: this is called the image's superiority effect. The way a visitor will react to an image is unpredictable; it is, therefore, necessary to collect data.
Colours - The colours Google is known to have tested the colours of its site
Forms - The fill rate forms very much more than the conversion rate, so optimization is the key to success here
You can also test the appearance of your buttons, customer testimonials, navigation menu, Social Proof elements, links, promotions, navigation, and layout.
There may be several tools to achieve this, the one which has been widely used is Google Website Optimizer (GWO) which is quite useful at the same time as complex at the beginning, and two types can be realized:
It serves to compare the performance of two or more different versions of a page, instead of changing an element, we change everything (or almost everything) in the design to know how the users behave.
It can be useful when we have planned complete redesigns of a site and we do not know if the users will feel lost or if they arrive easier to some element that we want, like the shopping cart or the register, or also to evaluate if the new design affects the performance of advertising due to new positions.
Also if we are announcing ourselves somewhere, we can send users randomly to the versions to know which one has better conversion where we change the background colours, the position of the elements, the contact form, etc.
For this experiment, we need a different URL for each version we make.
In this case, we are only going to play with some elements, as in the case of Highrise where they only changed the text of a button, so we identify which colour in the title calls the most attention, or in which way a button is better to register in the site.
All changes are recorded in the same address which is useful as in the cases of blogs where the pages and post have a fixed address, and we do not want to duplicate the content to do the tests.
In either of the two experiments, it is essential to know that we must have a goal that we want users to reach, for example, if they read our articles we want the goal or conversion to be to click on the feed button to subscribe, to complete a purchase or fill out a form.
Facebook, is not only is the largest social media world, bringing together 2 billion monthly users but is also one of the platforms with the best tools to run an advertising campaign. With Facebook Ads, it has become increasingly easier to serve ads to hyper segmentalized audiences, even on low budgets.
Recently, it became possible to perform division tests on this platform, with different campaign objectives and several variables. Want to learn simply and quickly how to take a split test on Facebook? Want to find out which variables are available, the best ways to conduct your analysis, and the most appropriate budget and time-period? Do not worry, we offer you the essential information and tips so you can start testing your ads right away!
Split Testing is a process similar to the experimental methods we all know. However, rather than merely observing the effects of a change on a variable, the goal of split testing is to achieve the most significant impact possible.
As with any other test, human errors prevent convincing results. So, if you're thinking about conducting split tests on your website, make sure you do not fall into common pitfalls.