A/B testing is one of the best ways to determine which marketing, promotional, and online strategies work best for your business. It can (and should) be used to test everything from search ad copy to banner ad graphics to website calls to action to newsletter subject lines. There’s virtually no limit to what can be tested and the more you test (subject to sample size limitations) the more insights you’ll gain into what works and what doesn’t for your particular business.

How to run an A/B test:

First, figure out if you want to run an on-site or off-site test. On-site tests refer to any sales-related systems of your website including forms, landing pages, buttons and other calls to action. Off-site tests are basically any other marketing outside of your website. Off-site tests can involve broad tests such comparisons in performance between channels of marketing such as social media ads vs. traditional banner ads or paid search performance on Google vs. Bing. Off-site testing can get pretty granular including the color of buttons in your banner campaigns, banner ad sizes, the length of email subject lines, and performance differences between various demographic groups or personas. In the case of ad testing, if you can prove that one format performs better than another that format becomes the “control” and that winner gets the lion’s share of ad spend. Same with email. You can test subject line copy, length, urgency, personalization, etc. Differences in open rate and post-click performance will help you to determine winning subject line attributes to use in the future. Lay out all the variables you’d like to test and narrow them down to a few that you think will be the most impactful performance drivers you are going to test. For example, if you’ve decided to drill into your email performance, some attributes to test in addition to the subject line are the headline, copy, design, length and calls to action. In-depth testing takes time and each variable needs to be tested independently in an A/B test, but the results are usually worth it.

What Should I Test?

Just because you can test just about anything and everything when it comes to your digital marketing effort, that doesn’t necessarily mean you should. A better strategy is to focus on components that you think will have the biggest impact on your results. Some elements worth testing on a website for example, are headings, content, downloads and calls to action. Paid search ads are simpler and easier to test. Start by testing the title, copy, offer, call to action and/or vanity URL.

How Long Will It Take?

Whenever conducting a test, it’s important to allocate enough time to capture enough data so that your results are conclusive. How much time really depends on what you are testing. For example, testing for length of time on site takes far fewer visits than testing for lead- or sales-conversions. Don’t test too much at once too fast or you risk not knowing what is impacting performance. At AdEdge, we take a minimum of one month to run a test or longer for smaller sample sizes.

Can I Test Multiple Things At Once?

The simple answer to this question is, yes… but that will turn your A/B test into a multivariate test which is much more complex. If you’re new to testing, start by keeping it simple with one A/B test at a time. A/B testing refers to testing one attribute at a time. For example, if you want to test four different ad headlines, that’s actually considered an A/B test. However, if you want to test various headlines and calls to action, that is considered a multivariate test since it involves more than one ad attribute.

If you would like some advice on which digital marketing channels are likely to perform best in your industry or if you’d like some insights on the best attributes to test and whether to use A/B testing or multivariate testing, feel free to schedule a free consultation with one of our digital marketing experts.