A/B Testing: Boosting Conversions Through Experimentation

What is A/B testing?

A/B testing refers to a randomised testing process in which two or more versions of any variable are presented to different groups of the audience simultaneously to determine the more successful model in terms of impact and business metrics. The variable could be a website, webpage, app etc.?

It is one of the elements of Conversion Rate Optimisation or CRO.?

‘A’ refers to the original or control design whereas ‘B’ is the new version. The version that drives your business metrics would be chosen for implementation.?

Let us understand this with the help of an example.?

Say you wished to improve your website performance, so you came up with two versions of your website. To test which one would be more effective, you decided to run an A/B test.?

Above, let’s assume ‘A’ and ‘B’ are two versions of a website or a webpage that you have tested. We can also see that ‘B’ performs better compared to ‘A’

Therefore, you would finalise ‘B’ and thus, implement it.?

A/B testing can offer insights such as:

• user behaviour
• engagement rate
• pain points
• satisfaction with website features (including new features, revamped page sections, etc.)

which are both quantitative and qualitative in nature.?

Why should you be A/B testing?

It enables you to resolve pain points in your design.

When a potential client visits or uses your website, webpage or app, they do so to satisfy a particular need or want. Often, they might find some difficulty in doing so. Your design might be hard to navigate or use. In this case, user experience worsens. Identifying key pain points and resolving them can help you improve upon this aspect.?

You can reduce bounce rate.

For the reasons given above, you can also ensure that a potential client stays on your website for longer, and who knows, may even become a paying customer.?

You can improve ROI from existing traffic.

ROI significantly depends upon user experience, making it a key focus point. When you create designs to cater to your client, even the smallest of changes have the potential to increase ROI.

You can achieve remarkable results with small changes.

When you test variations, you gain important data that allows you to pinpoint exactly what it is that you need to change from your existing design. There is no need for you to start from scratch again and redesign everything. As a result, your business metrics also greatly benefit.?

What should you be A/B testing?

Content

The very first thing that you should A/B test is the content which you put up. In your content, pay close attention to the following:

• Headings and subheadings
• Body
• Subject lines

The crucial factors to consider here would be formatting and writing style. Ensure that your content is understandable (don’t bombard it with technical jargon), clear, concise and engaging. All important points must be emphasised.?

Your main concern here would be customer satisfaction. Apart from the above criteria, find out what type of content they like - be it long, lengthy and detailed or short and sweet, and format it likewise.?

Design and layout

Good designs are clear and simple to understand. Having customer reviews up would also improve peoples’ trust and impression on your business. When you also create a sense of urgency (eg. by saying that certain items are selling quickly or by making certain items limited edition), you manage to psychologically convince people to make decisions faster. This can often lead to conversion!

Navigation

What you need to keep in mind here is that your navigation structure needs to be predictable as well as matching user expectations. By this we mean placing common navigational elements in their typical positions and? grouping certain items of a group? together. For example, you can place the ‘Return’ button in the upper left corner, or you can group trousers, capris and cargos under ‘Pants’ so that they don’t have to be looked up individually. Your navigation structure also needs to be fluid so that the audience can locate what they want instead of getting lost. This means that your mapping should be direct and logical.

CTA or Call To Action

This is your ultimate goal to ensure that someone makes a purchase or submits a survey or a search request. With A/B testing, you can play around with the formatting until you find what works for you.

Social Proof

This includes customer reviews and testimonials or recommendations from field experts, and can support or validate your claims. Based on the results of A/B testing, you can ascertain whether your website, web page or app may require social proof in the first place. If you think you need social proof, you need to further find out how many, and of what type you should put up. Their layouts and placements can also be experimented with.?

What are the different types of A/B testing out there?

Split testing

A completely new and different design is tested alongside the original. No changes are made to the control design.

Multivariate testing

Combinations of variations of multiple variations are tested out simultaneously to analyse performance.?

Multi Page testing

Variations of certain elements are tested over multiple pages.?

‘Redirect’ AB testing

At times, simply making changes to a page might be insufficient and you may wish to redirect your audience to a different page altogether. Here, you can simply replace the existing URL with the one that you wish for them to be redirected to.?

How do you decide what to A/B test?

Let us look at a simple yet logical process to find out:

Identification of issues

Using analytical tools and heat mapping data, identify what problems you might be facing.?

Choice of metric to measure

Based on your issues, choose a suitable metric to track performance.

Development of hypothesis

Based on performance data, decide what changes should be made or what other elements should be tested further.?

How do you run the actual A/B test?

• Create the variations that need to be tested.
• Determine the size and groups of the target audience that you will be testing on, equally.
• Run the A/B test.
• Evaluate performance.

What are some common mistakes that you should avoid?

• Not running the test for an appropriate amount of time
• Running tests at the wrong time
• Running tests on imbalanced or unequal sample sizes
• Testing too many versions or variations simultaneously
• Measuring more than one metric at once

These errors may render your analytical data incorrect and therefore,useless.

Is A/B testing successful?

A/B testing has produced wonderful results. Let us look at a few examples.

• Grene redesigned its mini cart and saw a 2x increase in overall purchase quantity.
• WorkZone increased its leads through its testimonials page by 34%.
• Zalora’s increased its checkout rate by 12.3% by optimising its product pages.

There are multiple other instances where A/B testing has significantly improved conversions and other business metrics.?

The ultimate goal of business is customer satisfaction and profit. When you can undertake an experiment that ensures both, why not give it a try?