What makes one marketing email perform better than another? The color of the call-to-action button? The length of the subject line? Many small business owners rely on intuition to guide their email marketing strategy. With A/B testing, they can get hard data, a much better guide for business decisions.

A/B testing is a way to compare two versions of something to see which one does better.

Here’s a simple example: You send an almost identical email to two halves of your subscriber list, with one exception—half of subscribers get a red call-to-action button, half get a green one. You find 20% more subscribers clicked on the red button.

The benefits of such a test are clear. From now on, you can opt for the red button and enjoy the benefits. Over time, these small changes can make a big difference for your business.

How to run your first email A/B test

Many email marketing software platforms include A/B testing as an easy-to-use feature.

Decide what to test

Even the shortest small business marketing email is made up of many components. Each one can subtly influence your readers to click through or click away, engage or disengage. Here are some ideas of elements to test:

• Subject line length
• Sender name
• Font
• Images vs. no images
• Button color
• Offer vs. no offer
• Copy in bullets or paragraphs

Build each version

Actually performing your A/B test is simple with email marketing platforms like Mailchimp, Constant Contact and HubSpot. Follow the directions to build the “A” and “B.”

Create your two versions, incorporating the change you’re testing (like learning which button color performs better), keeping every other element such as font, subject line and sender name the same. This way, you know it’s your chosen element that’s changing your readers’ reactions. And you can save those other elements for another test you’ll do.

Send to two like sections of your audience

Once you’ve built the A and B versions of your email in your marketing software, you can identify a sample of your audience and send out the two versions.

Be sure to double-check the makeup of the group you select. Sending to two radically different sections of your audience—like primarily local versus international, or mobile users versus desktop—could interfere with results. You want the two audiences to have as much similarity as possible.

Evaluate the results

After a week or two, check your results. If there’s a marked difference—like the prior example, in which 20% more people clicked on the red button—you can change your marketing strategy accordingly.

Dos and don’ts of A/B testing

Taken step by step, A/B testing is pretty simple. But there are some things to keep in mind when putting it into practice.

DO: Consider statistical significance

One big disclaimer to A/B testing for email marketing: Sometimes, more people click on version A than version B for no particular reason. The bigger your subscriber list and bigger the difference between the two versions, the more confident you can be in your results.

Here’s an example. You send version A, with the red button, to 100 people and version B, with the green button, to 100 people. You find that 10% of people click through version A and 20% of people click through version B. You can be 98% sure version B actually works better—it’s not just a fluke.

But if just 15% of people click version B, you can only be 86% sure it was the button that made the difference. There’s a decent chance it’s just a coincidence.

You can play with this A/B testing statistical significance calculator to learn more.

Just remember, this is not an exact science. You’re dealing with human behavior and choices. In the hard sciences, statistical significance is key because the goal is to advance human knowledge, even cure diseases. In marketing, the goal is to bring you more business.

DON’T: Test more than one factor at a time

An A/B test is a single variable test. Multivariate tests are great, too, but they require some statistical finagling your small business may not have time or budget for. Sometimes tests like that are better left to large enterprises with expansive marketing teams.

DO: Take your results seriously

Be sure to actually use your results.

In test after test, if the subject line “John, Get Your 50% Off!” performed better than just “Get Your 50% Off,” use it.

With the right perspective, A/B testing can be a low-effort, high-reward way to see what works for your email marketing and how your email marketing can work for your business.