Reading Time: 4 minutes
A/B Testing is a widely known concept among marketing and design practitioners. In short, it stands for the process of deploying two versions of something – a landing page, an email newsletter, or even a flyer – and seeing which one is more effective.
As a common practice, A/B testing has been used by many companies. Endless possibilities make it a thrilling and never-ending process. If you’re building a landing page, you can play with fonts, colors, texts, and page sections; if you are an ad campaign manager, you can have fun creating variations of the audience age and sex, for instance.
It is important to remember that the perfect output doesn’t exist, and it’s the first step for A/B testers. The ego must also be put aside when it comes the time when the solution you though best performs worst than its contenders.
For A/B Testing to be successful, marketeers typic set goals. To reach them, settle on a specific percentual improvement and work up until you attain it.
But how should we effectively deploy an A/B testing strategy?
Picking an A/B testing strategy
The first thing to do is to pick a strategy. You can choose from two main strategies.
1. Broad strategy: on the broad approach, you’ll test at least two fundamentally different versions of something. Imagine you are building two landing pages: landing page A will be completely different from landing page B (different colors, texts, fonts, and structure). Once you have the winner, the results will give you a direction that you’ll have to narrow. The next step is to start isolating colors, texts, and fonts so your tests become more specific.
2. Specific strategy: instead of testing completely different versions, you go for tiny iterations. If you seek to understand if other call-to-action (CTA) impacts your conversion rate, create an A/B test in which the only difference between version A and version B is that specific call-to-action.
The decision of which strategy to pick is up to you. Usually, the broad one is used in the early stages of a digital product or when you’re trying to establish a visual or communication direction.
For improving conversion rates, the specific strategy tends to show better results. Whenever possible, isolating tiny changes will allow you to learn which variation leads to an extremely particular outcome. Making rounds of specific tests will lead you towards a better variation.
What to A/B test? Where to start?
Anything can be A/B tested, but some things need it more than others. A website homepage, for instance, can be tested, but the relevant question is if it needs to be your priority.
To manage priorities, what we advise you to do is to gather information on your digital presence. Start by designing it using a mind mapping tool, such as Miro or using post its. Your digital presence can be composed of a website, blog, social media. Put them all down in post-its.
The next step is to connect the post-its, minding that each connection refers to traffic. If you have a website or landing page, where do users come from? Put it down on your mind map.
Lastly, add internal processes, such as emails that are sent after the conversion.
Once you have it all done, go through the results you have in each platform, and compare them with industry standards. If your outcomes aren’t so good, there you go. Here you have the thing you should have A/B test first.
When making the decision, also remember to understand your business priorities. Social media may be your Achilles heel, but if the landing page is the central spot of your presence, you might need to work on the latter first.
Setting A/B testing rounds
A/B Testing requires discipline. When conducting a test, you must be organized and be careful not to jump into hasty decisions.
When deciding on what to A/B test, it is common to see several things that can be tried and experimented with. Note them all down and avoid testing everything at once. The downfall of doing it is to get to the end of the test without any specific conclusion. Do a calendar instead, and test varieties in two weeks, depending on the people you reach.
Mind also the topic of external factors that may impact A/B Testing. Ideally, users to whom each version is presented must be chosen randomly to avoid any sort of influence. For a test to be successful, it is vital to have a significant number of people reached and pick dates that have no connection to a specific event or trend.
When comparing newsletters, one with a Christmas layout may perform better than the other, but that may only happen due to the holidays. If you’re aiming for long-term results, make sure you avoid external influences that impact your business.
A/B Testing as a mindset
Some salespeople argue that creating an A/B testing mindset will have a positive factor in the company’s revenue. I tend to agree, but I would also add that it will also impact your company culture, replacing the fear of failure and replacing it with a self-improving mindset.
After all, having an A/B testing mindset is nothing more than being critical about the job that is being done and come with ideas of what can be done differently. To get from thoughts to real actions, all we need is to put in practice. Some versions fail, but others might succeed.