As marketers, our job is about more than just thinking up beautiful campaigns and clever copy. We also have to be statisticians. That means testing, measuring and reporting the results of our marketing initiatives to validate both our time and clients’ money.

A/B testing is one tried and true marketing best practice you may not be taking advantage of. The concept is simple enough: A/B testing involves comparing two different versions of a web page, ad, button, email subject line, etc. to see which performs better. The execution, however, is a little more involved.

To get you started on a path to successful A/B testing, here are a few pointers:

Clarify and commit to one objective.

Much like a science experiment, you’re developing a hypothesis here. You’re asking yourself what you want the outcome to be (more conversions, higher click rates, more subscribes, etc.) and testing for that outcome. By focusing on one aspect of your asset—whether it be copy, such as email subject lines, or a design element such as form length or placement—you can be confident to some degree that the higher-performing variation resulted specifically from your tweak and not from any of the other factors having affecting that asset’s performance.

Create your variants.

Once you’ve established your objective, it’s time to put your ideas to work! Create one version of the asset and then create a second that takes a different approach. Ultimately, your goal is to find out which path to take with that asset in the future, so make sure you’re testing something you can put into practice later. For example, if you’re testing subject lines, you could make one straightforward and phrase the variant as a question. Your test would tell you whether your audience likes the direct approach or perhaps likes a little mystery before they click.

Determine your sample.

One major thing to keep in mind for A/B testing (or any type of statistical testing) is your sample size. Your sample is comprised of a select number of people that are acting on behalf of a much larger population. Typically, you would want at least 1000 people in your sample to really be able to attribute your result to one variant over the other, but if you don’t have that many in your database, that’s okay. Oh, and your sample should always be selected at random!

Decide on a timeline.

Setting the right timeframe for your campaign is definitely key to your success with A/B testing. You want to make sure your asset receives enough views or actions for you to be confident that your test was a success. Whatever you decide to test, do some research into how those assets have performed for you in the past, and set a timeframe accordingly.

A/B Test! (And find out the winner!)

Ready … GO! It’s time to go live with your variants. Once you test your assets to make sure everything’s in place and working properly, sit back and let your campaign do the hard work. Keep an eye on things to make sure everything’s going smoothly throughout the process, and once you’ve reached the end of your timeline, you should have a winning variant! Simple, right?

So much of what we do as marketers can feel like guesswork if we’re not thinking strategically about how our content resonates with online audiences. A/B testing is a simple way to gain valuable insight into your audience’s preferences that you might not have been able to capture in a form or survey.