A/B Split Testing Your Advertising

How do you know if your advertising is effective? It might look good and sound good, but that’s only your graphics designer’s opinion, and he or she doesn’t necessarily speak for your target market. How is your readership going to react? Does your imagery make sense to your customers, or only to you? Is your message connecting with your audience, or does it fail to grab their attention?

The only way to know for certain is by testing your message.

Sometimes what might seem like a good advertising graphic or email may perform poorly, and you may have no explanation as to why it failed. Was it the graphic image you used? Or what is the subject line, Or the text? Maybe it was the time of day it was sent? It could be any of a number of reasons.

Your advertising vendor should be able to provide you with a way to A/B split test your advertisement, whether it’s an ad banner or an email, and give you reports on the performance of each ad.

How A/B Split Testing Works

The premise is simple; send a random sample of viewers one version of your advertisement, and another random sample a second version. The two versions should have no more than two major differences, one would be preferable. This way you can gauge the response to the alternating elements. Try two different subject lines in an email, change the gender of the model in the graphic, use different text colors, or different call-to-action language. Essentially, whichever element you want to determine has a greater reaction from your audience is the one you want to test.

Understanding Results

I had an email that I A/B tested with a 5-to-1 open rate, but the clickthrough rate once the email was opened was closer to 2:3. What did that result tell me? The subject line in the first email delivered a much more meaningful message to readers. But when people read the email, they didn’t connect with the message inside. For subsequent emails, we used the email body from the second with the subject line from the first.

Each of these elements can be measured and tested for their appeal to your target audience. By determining which advertising elements are successful, you can easily manipulate the ad to maximize its effectiveness.

The most common type is A/B split testing. By broadcasting two ads with one distinct difference to two similar demographic groups, you can see which advertisement received a better response. Any of the elements above can be A/B tested to determine which approach is more appropriate for your market.