Whatever the industry, statistics matter. Whether you run a small business or a major corporation, statistics will help you determine your base audience and their wants. Today, most businesses employ A/B testing to gather such data, which is a way of comparing two versions of a variable within a marketing campaign to see which one performs better.
Major companies such as Netflix and Amazon are known to fall back on the test for a number of website features. In fact, there aren’t many major companies with an online presence who aren’t using it. But the benefits aren’t limited to companies on the Fortune 500. Virtually anything can be tested, as long as the four essential components of a successful test are taken into account.
Determining Goals
Scientific experiments are not done without a purpose of some kind. A/B testing can be used for a number of purposes. The first step is choosing what you are trying to learn about the variable. If it’s part of your campaign’s message, you may try to test if the message is being transmitted clearly enough.
If you are curious about optimized pricing, then your goal may be to gauge customer response to a newly priced item. The most useful objectives used in A/B testing are DUMB—doable, understandable, manageable and beneficial—that enable you to test with a clear intent.
Measuring Success
With your goals well-stated, it’s vital to find your key performance indicators (KPIs). KPIs are metrics that will help you understand how your business is performing against your objectives. If your objective is to sell more T-shirts through your website, then one KPI would be the number of shirts sold. The idea is to keep your goal as specific as possible. In more complex versions of A/B testing, you are able to test for multiple outcomes. However, for beginners, it’s advisable to keep it simple.
Evaluating Your Baseline
No test can properly begin without understanding where things currently stand. A fast, easy way to establish your baseline is by using an app with analytics functionality. Google Analytics is the frontrunner, though there are many others. The more data you have, the more accurate your test will be.
Testing and Analyzing the Results
The final part of any A/B testing project is the results. This can best be explained through an example: Humana, the health insurance giant, tested two different versions of a banner for their homepage. Their goal was to increase click-through rate and their KPI was an increase in sales. The first version of the banner was drowning in text. Readers would have to trudge through a full paragraph of copy before arriving at a call to action. The second was much cleaner: a bold headline followed by a sentence and a very clear, large call to action button. The cleaner version saw a 433 per cent increase in clicks. It was extremely easy for the company to decide which version to go with.
Kenny Hedges | Contributing Writer