A/B Testing: The What and the Why & the How
Before we get into understanding what A/B testing means, let’s understand why we should learn and know about this concept. A/B testing is a method which helps marketers and developers gain insights into the behavior of the visitors and increase conversion rates. It is a very useful method and technique that hasn’t been as popular as SEO and Web Analytics mainly because people have not completely understood what this method is and how to use it adequately .
This blog is meant to guide you through the theory of A/B Testing.
What is A/B Testing ?
A/B testing, also known as split testing or bucket testing. As the name suggests, it is the process of experimenting or of comparing two web pages or apps against one another in order to find out which one works better. It’s a method which helps you to understand if any changes made on your web page or app will help you improve your conversion rates.
This method involves making two variations of the same web page and showing them to the predetermined percentage of visitors. So basically, there are two versions. One is the original web page which has the existing design, referred to as the “control.” and the second version where changes are made, referred to as the “challenger”.
An A/B test is the most common type of landing page test to conduct.
Let’s take an example in consideration here and understand the entire process.
XYZ company is a client who has a travel business and provides hotel bookings, car rentals, air fares etc. over the Internet through a website and a mobile app. The challenge faced here is that the click-through rates and conversion rates for their airfare page are relatively low. Hence, the main aim is to redesign the airfare page to increase click-through and conversion rates for the mobile app.
What to test?
Your choice of what to test depends on what your goals are. Let’s take an example: If your goal is to increase the number of sign-ups, then a good suggestion would be to test the length of the signup form etc. The aim here is to see what is preventing your visitors from signing up. Is the form too long? Are visitors concerned about privacy? All of these questions can be answered one by one by testing the appropriate website elements.
For the above example, these are the certain elements which can be tested:
- Headline or the description of the product
- Layout of the website
- Images on the landing page
- Amount and size of text on the website
- Testimonials
- Call to Action buttons and many more
How to conduct an A/B test?
In order to successfully conduct an A/B Test, you should follow the following framework:
- Identify Goals:
The very first step includes identifying your conversion goals. These are the metrics that will help you determine if the variation is more successful than the original version. Goals can be anything ranging from increasing sign ups to product purchases etc. Here, the goal is to increase the conversion rates over the mobile web. - Collect Data:
The analytics of your website will provide you with the information that will help give you an insight about the performance of your website. It helps identify the areas which bring you high traffic or low conversion rates or drop off rates that can be reformed. - Generate Hypothesis:
Once you’ve collected data and identified your goals, you can generate a hypothesis which helps you define a purpose for developing a new version. Your hypothesis can contain the elements that are needed to be added or removed from the website to make it better. For the above example, the hypothesis could be to redesign the entire air fare mobile webpage to provide a better virtual experience to their end customers which can help increase the conversion rates. - Generate Variations:
Make multiple variations of the web page with the desired changes. It can be either changing the colour of the web page, shifting the call button from one position to another, making new customised navigation tools or anything that you want to change. Here, variations of the page can be made by adding different images or colour codes or navigation tools etc. which make the user experience good. - Run an Experiment:
Kickstart your experiment and wait for your visitors to respond. The visitors will randomly be directed to the control or the variation webpage. Their interaction with each of these versions will be looked into and observed for further analysis. - Analyze the results:
The interactions of the visitors with the pages, will help you figure out which one is working better and if there are any further changes that need to be made.
Whatever is your experiment’s result, use them for future tests and continue to optimize your website or app.
There are certain tools which you can use to conduct your A/B testing successfully, viz:
1) Google Analytics Experiments
2) Optimizely
3) Marketizator
4) Unbounce
And many more. These are popular plugins that help in optimizing your website’s conversion rates and they help in streamlining the A/B testing process. You can create various online surveys and banners to increase the audience satisfaction.
Successful A/B Testing snippets –
A good designer or a developer is one who creates new designs or codes and tests them before putting it out to the end visitor to avoid mishaps. Just how you wouldn’t release a product in the market without testing it with a few buyers before, similarly, a new or edited version of a web page cannot go out without testing. A data driven design can be really effective in order to attract more visitors to your website.
Let’s take a look at these few examples that show the success of A/B Testing and how they did it.
-
Revolve –
Revolve is a virtual high end fashion brand which designer apparels, shoes and accessories for both men and women. Since they are a virtual store holder, their main aim is to provide the ultimate digital experience to their customers.
Revolve managed to do so by revamping their website and also revamping their mobile site aesthetic. They wanted to fine tune the experience of their mobile users and hence, their team ran a test to increase the number of shoppers or visitors on their website. They created a hypothesis which stated that directing the mobile web users to a splash screen which included heavy promotion, then that would motivate the customers to make purchases from the mobile app and hence increase the number of downloads.
As a result of this, Revolve found out that the variation page with the splash screen worked wonders and their page saw an overwhelming increase in the downloads of mobile apps through the web page directly.
2) Trulia –
Trulia is an online hub which makes it easy for home buyers to look for homes over the internet. Their main challenge was low conversions and engagement metrics. They started observing that their conversion rates are dropping below average, leading to lower engagement rates.
This led them to trying out and putting a new call-to-action button and an infinite scroll. Hence, they added a new call to action button that would submit the lead information directly on the search result and help reduce and remove unnecessary friction and increase the conversion rates. They also decided to add infinite scroll so that the customer can keep viewing properties at a stretch. As a result, there was a 60% increase in pageviews per visit and also a 15% increase in leads per visit.
Hence, from a designers point of view, A/B testing can help you to prepare versions of the website with designs that would help the visitors to stay on the page for longer and help generate more leads. This can be done by changing the colour pattern of the website, making navigation tools for the ease of the visitor to go from one page to another etc.
As web developer, A/B Testing can help you try out new drop down menus to enable customers to look for more offers and a better experience, create clickable links to direct customers from one page to another etc
A/B Testing is one of the many ways to increase the conversion rates on websites and is a simple one to conduct. Let us know in the comments below if this was useful!
