a b split test resized 600As a marketer, A/B tests and multivariate testing are a great way for your business to optimize web pages and generate more leads and conversions. They allow you to figure out which elements of your marketing are working and which ones can be improved upon.

Here are 5 best practices to keep in mind when conducting A/B testing on your website:

1. Keep It Simple; Then Get More Advanced

Although some people are really good at A/B testing, there is always some skill involved with testing some of your website’s more complex features. Instead, start your A/B testing with something simple, like moving your registration form to the left of the page instead of the right. Then, as you start to understand the A/B testing process, you can start conducting more sophisticated experiments. For example, WikiJob decided to conduct an A/B test by adding something simple -- testimonials -- to one of its pages to see if they increased conversions. As a result, the web page including testimonials performed 34% better in sales than the web page without them.

describe the image

2. Understand That a Minor Change Can Lead to a Significant Result

Something as simple as changing the text in the bullet points on your web page can increase your conversions. People sometimes think that big, sweeping changes need to occur in an A/B test. The truth is, the details of the page are just as important. Remember that something simple can still drive big improvements.

3. Test 1 Variable on a Page, But Don't Limit Yourself to one Variable

In order to see if a feature on a page is working effectively or not, you have to isolate it in your A/B test. Test one item at a time, but remember that your web pages are also made up of a number of other features. You don't have to limit yourself just to testing color background or text size. Think also about your images, videos, language, bullet points, and headlines, to name a few.

4. Always Be Testing

Your first A/B test may have been a huge success, helping you discover a new way to make your web page more effective. However, there’s always room for more optimization on your website. Try conducting an A/B test on another feature of that same page. For example, you can test for headlines, body copy, color schemes, images, adding features, etc. Then move onto another page of your site, and do some testing there. There’s a pretty good chance you can still increase conversion rates and leads elsewhere, too.

5. Don't Give Up, Even if a Test Isn't Significant

You may decide to conduct an A/B test on the headline for a web page but see no statistically significant result to persuade you to run with one page over the other. Don’t think your A/B test failed. Use the failed data to help you figure out a new iteration on your new test. For example, consider testing new headlines, and see if that makes a difference. If not, the headline may have no bearing, but there may be another feature of your page you can adjust to increase leads. For example, try switching up your call-to-action button, and see if that makes a significant difference.

What other lessons do you think marketers should keep in mind when doing A/B testing?

New Call-to-action

Originally published Sep 26, 2011 1:00:00 PM, updated June 28 2019