Conversions. These numbers can often represent the be-all, end-all for business owners.
You could implement recommended design and marketing tips for your industry all day long. You could also also fine-tune these tips to work specifically for you. This is where A/B testing comes in.
What is A/B Testing?
In complicated terms, it’s one of the most commonly used randomized controlled experiment. In layman’s terms, it’s a test of comparison to simply see which version of something works better.
This can mean generate more conversions, click-throughs, opens, reads, views, and many other important metrics. Generally, businesses A/B test to find which changes in elements generate more conversions.
The “A/B” refers to the two versions of an element that are being compared: Option A vs. Option B. The differences between the two options should not be extreme, but specific and impactful. It should also be testing a single element or factor, not an entire design scheme as that would be considered split-testing.
How to Create an A/B Test
You only need a few steps to generate an A/B test. Once you’ve nailed down what you want to work on, the hardest part is reviewing data numbers before and after your test. You can even follow these steps over and over to perfect each element of your website or marketing package.
- Pick one thing you want to test.
Get specific. Wanting to change the colours of your design is not one specific thing. The colour of your call to action button is a good example for a small, specific detail that can be tested.
You can choose which element to test through data you’ve collected already (pick a low-converting page to work on), or find something that’s high-impact that you would like to test (like an email subject headline).
- Set a measurable goal.
Choose a metric that is significant for you. Is it conversions? Click-throughs? Views? Reducing exit or bounce rates may be significant for you. Maybe you are hoping a different colour will increase click-throughs on your call-to-action button by a small margin.
It helps to review where you (and your data) stand before you start A/B testing. Google Analytics and Google Optimize are tools that shows you some metrics for free, which makes that a good starting place to return to after you’ve completed your A/B test.
- Choose two versions. Option A, and Option B.
Compare the industry recommendation with another idea you have, you might be surprised at the results! You can also compare two competing recommendations, such as longer copy or shorter copy, to see which works better for your business specifically. Every audience will have slightly different preferences, so don’t assume what works for a different business will automatically work for you.
“I don’t know the rules of grammar. If you’re trying to persuade people to do something, or buy something, it seems to me you should use their language.” – Exec. David Ogilvy (regarded as the father of advertising)
- Implement versions in the same audience.
For this to be an A/B test, the only thing that should be different between the two versions of something is the one element you’re changing. The audience should remain the same, the time factors and frequency of views should be the same as much as possible. Don’t forget to have your tools set to collect the data from your test.
If you include the coloured call to action button in an email, send one option to one half of your email list, and the other option to the other half.
- Consider results.
After some time, you should have some feedback or data collected. Compare the metrics of the element from the baseline (where you were at the beginning) with what has occurred after your test. What changed? See if there were other metrics that changed other than the one you were measuring. You can then choose whether to continue with more A/B testing on the same or different elements to further boost your metrics.
Where You Can Start A/B Testing
When considering what you’re going to change, there are different elements that affect people differently. A/B testing allows you to explore what changes in these elements affect your data and metrics. What kinds of things affect people? Colour psychology, emotional depth, voice, and availability are just a few.
The most popular places to use A/B testing are marketing and website design. Where in these spaces can you test? The list is endless:
- Calls to action
- Subject headlines
- Button colours
- Copy or content length
- Ad titles
- Gated content
- Hero images
- Image vs. video
- Site navigation
- Form fields
- Email personalization
You don’t have to limit yourself to A/B testing only in the digital realm. See what happens if you compare two different elevator pitches, or two versions of a company Christmas card. What works better for your business and employees?
A/B testing can go as far or in-depth as you choose. Hopefully, these ideas will get you started testing all on your own.