AB Testing best practices
As a marketer, you know that measuring and analyzing data is important. Without data, you have no way to gauge whether or not your efforts are paying off for your business.
A/B testing is a simple yet effective tool you can (and should) use to determine which campaigns are the most effective at driving conversions.
What is A/B Testing
A/B testing allows you to test two versions of your marketing asset - whether it's an ad, an email campaign, or a landing page - to see which version outperforms the other and effectively drives conversions. This is a simple yet poerful tool that marketers can utilize to optimize their efforts and ultimately improve conversion rates.
A/B testing isn't a "one-and-done" process; rather, it's a practice that should be used on an ongoing basis to continuously evaluate and tweak your campaigns. You should always be striving for improvement in your campaigns to stay successful in the long run.
How to Run an A/B Test
To begin A/B testing, you need to determine a few different things:
Control group and treatment. Your control group is the standard version of your 'test subject' that you would normally use. Your treatment is your variation.
How long you need to run your test. This will vary on a case by case basis, but it will depend on:
Your current conversion rate
How much of a difference in conversion rate you need to detect
Your current Web traffic
You'll want to make sure you give your customers enough time to 'solve' your test in order to get an accurate picture of your data. Some tests may require more time than others; for example, testing two different subject lines of an email may require a shorter waiting period than testing two different versions of a paid ad.
How many visitors to test. Again, this will vary depending on what you're testing, but you should test enough visitors to detect a noticeable difference in results so that you can make a decision.
Once you've established the foundations of your A/B split test, it's time to begin testing. Here are some basic A/B testing best practices to keep in mind while running your test.
Think big. When you first begin testing, test two different versions of your landing page, PPC ad, or other digital deliverables to quickly see which variation yields the higher conversion rate.
Then, focus on the details. After you understand the big picture of what works better, you can start fine-tuning. However, be cautious and only test one variable at a time - changing too many things at once can muddle the results. Test small changes, such as form placement on a landing page or two different versions of a CTA, to get an accurate idea of what resonates with your customers. Then, continue making adjustments as necessary.
Don't sweat the small stuff. Use your judgment when A/B testing. For example, making a minuscule change like 'Buy our book of 29 organizing best practice' or 'Order our book of 29 organizing best practices' will probably have little impact on the overall success of your campaign. However, testing a landing page with an image of the book against a version without an image may yield a more accurate representation of your audience's behavior.
A/B Test example
Here is an example of an A/B test in action:
Let's say you have a landing page you want to test and the objective of the landing page is to get website visitors to fill out a form to sign up for your e-newsletter.
To test the landing page, you'll first want to create two different versions of the same page. Make them different enough to distinguish the results of your test - you can move the location of the form on the page, change the page's title, and even change the body copy entirely. But you should only change one thing at a time. Version A (or your existing landing page) will be your control, while version B will be your treatment.
Next, you'll want to determine your audience and how long you want to run your test. Again, make sure you have enough time and a large enough audience to accurately determine your A/B test significance.
Once you determine the scope of your test and create two different versions of your landing page to be tested, you're ready to launch your test. Make sure your landing pages to be tested, you're ready to launch your test. Make sure your landing page is getting enough traffic to collect sufficient data - a page that receives little to no traffic isn't worth testing!
After you've gathered significant data to determine a winning page, don't stop there. Continue testing different elements of the page to improve your conversion rate - your call-to-action, different colors of buttons, etc. Be sure to make gradual changes to the page and only test one element at a time to get an accurate picture of what caused the results to change - for example, first change and test the form placement, then the CTA, and so on.
What to A/B test
Now that you have an idea of how to properly run an A/B test, your head is probably running wild with all the possibilities...or perhaps you're overwhelmed and unsure of where to begin. If you're wondering what exactly to A/B test, the sky's the limit. Test everything - nothing is too perfect to not be tested. Some examples of what you can (and should) A/B test include:
Email subject lines
CTAs
Layouts
Colors and fonts
Navigation
Images
Content length