Welcome to the ABCs of A/B testing. If you’ve spent any time in proximity to marketing, you’ve probably heard the term thrown around. Today we’re going to break it down at the most fundamental level, explore what it means for social media, and talk about how your brand can use it to improve performance.
What is A/B Testing?
A/B testing, also known as split testing, is much like the experiments we did in science class—except here, we’re testing the effectiveness of marketing materials. Think of it as a psychological experiment! Marketing is expensive, so before throwing your whole budget behind a campaign, you want it to be as effective as possible. This goes for traditional marketing and social media alike. By testing small variations in your advertising plan on two smaller segments of your audience (A and B), you can see what they react best to and gain valuable insights for your current campaign and even future ones.
With traditional marketing, one might conduct a focus group or study, where researchers show audience members ads and ask for their opinion/sentiment. While this can be effective, there’s a catch—their reaction to an ad may be different in the real world. On social media platforms however, we have the ability to serve ads to the A and B audiences without them even knowing they’re “test subjects.” Many networks have this capability, but Facebook (which includes Instagram) offers the most comprehensive options.
Common Variations to Test
When it comes time to run a test, there’s probably more than one element you’re curious about. Just like science class, testing multiple variations at once can lead to “confounding variables.” You won’t truly know what worked unless you test each element one at a time. Here are some of the most common variations to consider testing on social media:
Does your audience respond better to 6-second video or 15-second video? How about 30-second video? If you’re measuring by video completion rate, 6-second may seem like the way to go, but if your 30-second video is really compelling and people don’t click away, you have more time to persuade them.
If you want customers to take an action (“shop now” for example) you’ll want to test the placement of your CTA. If you’re running a video ad is the audience more likely to click when the CTA is at the beginning of the video, or the end? The middle? Both? The answer can depend on how compelling and informative your video is.
Post Copy Length
On mobile, Facebook will display 140 characters before the copy flows below the “Continue Reading” fold. Some may say that’s still too long. Some may say you can make it as long as you want if the most important info is above the fold. The only way to know what works best for your brand is to test!
Brands contain multitudes. There’s more than one way to talk about them! For example, a skincare brand may have acne-fighting properties AND is made with free-trade ingredients. Maybe one of those messages individually will resonate more strongly with a certain audience.
Does your audience prefer photography or graphic design? A lot of text on the ad or minimal text? Bright colors or muted shades? There’s so much to test when it comes to design, but remember not to stray so far from your brand guidelines that your presence isn’t cohesive.
All platforms have different formats/placement, but Facebook definitely has the most. You can run a video, an image, a carousel, a side banner ad, a messenger ad… the list goes on. Odds are, there isn’t ONE format that works best for your audience. But there may be one that works best for a particular message. For example, if you have a more complicated message, video may work better than static imagery.
Facebook gives you the ability to select an “objective” when running an ad. Generally, you should choose the one that most closely matches the goal of your ad. If you want comments, Facebook will serve your ad to people who engage more than the average user. If you want video views, it’s the same story. But there are some nuances and objectives that could overlap, so it can’t hurt to test.
Some brands know exactly who their audience is, but many are still testing the waters to see who resonates with their product the most. Facebook allows you to serve ads to different demographics to help you hone in.
Testing isn’t perfect. Here are some things to keep in mind when conducting tests and evaluating your results.
1. Your test results don’t necessarily speak for your whole audience.
They couldn’t possibly, because most likely, your whole audience isn’t on Facebook. It’s tempting to make assumptions about how ads will perform in non-social media placements based on social media tests, but be cautious. In the same vein, if you only spend a small amount of money to do a small test, the test may not have a large enough sample to be accurate.
2. Keep an eye on the algorithm.
When setting up a test, Facebook’s algorithm can sometimes favor one ad over another, and it’s always based on what is actually performing better. When conducting a test, check in to make sure each variation is spending around the same amount of money.
3. Make sure your audiences are large enough.
Believe it or not, Facebook can run out of people to test your ads on if the audience and objective are too specific, rendering the results skewed. There may only BE so many grandmas in Iowa with pet lizards who click on ad links often.
4. Set the standard for success beforehand.
If your brand has run ads in the past, historical data is the best benchmark. Sometimes, BOTH ads in your test will be below the benchmark, even if one outperforms the other. The good news is, there are so many other variables to test.
The world of A/B testing is highly complex, but these basics will get you started on the path to smarter social media ads. And if all of this makes your brain hurt, hit us up. Our Likeable ads team can help!