Facebook Ads: The Sweet Science of A/B Testing
Published: November 6, 2013
Author: Molly Shotwell
Today’s post is by James Scherer, social media expert for Wishpond.
Are you getting the best ROI possible from your Facebook ads? Have you thought about A/B testing but think it’s too complicated for your business?
This article will break down A/B testing into bite-size pieces. I’ll explain exactly what it is and give you the four variables you should be A/B testing to get the best possible return on your Facebook ad investment.
A/B Testing Defined
It’s used commonly for webpages, landing pages, marketing emails, and advertising.
Your Facebook Ad tool should allow you to create multiple ads within a single campaign. I recommend putting time and energy into your first couple of campaigns and testing what your target audience responds to.
Create four or five advertisement variations and see which one works best. This will save you time in the long run, as once you know what works best for your business you can optimize your future Facebook ads.
How it works for Facebook Ads.
-Run your control ad (“A”) for a set period of time to a set audience.
-Record the results of your control.
-Run your variation ad (“B”) for the same period of time with the same audience
-Record the results of your variation, keeping in mind that your results must be statistically significant. Run them through this A/B Testing Calculator to be sure.
-Compare your results and adjust future advertisements based on the conclusions you make.
Running your control ad to a different audience is also a form of A/B Testing. I’ll discuss demographic testing below.
Example Facebook Ad:
For this article, I’m going to use a hypothetical Facebook Ad for a $100 gift card to Acme Sunglasses. Here’s what the control advertisement looks like:
A/B Testing your Facebook Ad Image
Your ad image is what grabs the attention of Facebook users. It should be colorful, exciting, and encourage a second look. Let’s say that I’ve just recently read a great article on optimizing my Facebook Ad image, and know that the most engaging picture I can use is of a smiling woman. Let’s test it:
I also know that Facebook’s color scheme is blue and white, so the optimal photo for my Facebook ad would be a smiling woman on a red or orange background. These colors will stand out far more than my ocean backdrop does, so testing a color change would also be worthwhile.
A/B Testing your Facebook Ad Headline
Once your Facebook ad image gets the attention of a Facebook user, your ad headline should grip them. I like value propositions for my Facebook headlines, as communicating the value of engagement to a user quickly and easily is vital to an ad’s success.
Let’s see an option for a headline test:
The word “Free” has a tendency to grab the attention of Facebook users, and as soon as they see it out of the corner of their eye they snap back to see exactly what is being offered. If the free offer is desirable, they’re likely to read further.
A/B Testing your Facebook Ad Details
You’d be surprised at how important the small details can be. Kevin Spidel found that adjusting small variables within his Facebook ad doubled his CTR. It might sound inconsequential, but simply adding a border around our control image could increase your engagement:
Small details are easy to add to your Facebook Ad image. Microsoft Paint, or free online photo editing tools like PicMonkey.com work well enough.
Four small additions that could improve your CTR:
- A colorful red, green, or orange border (something that stands out against the blue and white of the Facebook News Feed)
- A subtle shadow around your image
- A colorful background around your logo
- Increasing the yellow/green gamma on your image
A/B Testing your Facebook Ad Demographic
How it works:
Facebook ads can be targeted more specifically than any other online ads. You can tap into timeline variables like relationship status and parenthood as well as precise interests like sports teams, health and well-being — even (I’m not joking here) recent car purchases. This is all alongside the straightforward targeting by age, gender and location.
How it could work:
Target just men with an image change:
Target parents with an image and body copy change:
Other Testing and Targeting Strategies
-Limit the ad to married men, run it during the Christmas season, and change the headline and copy to “Need a Christmas present? // Surprise her this year with designer shades. Enter to win a $100 gift card.”
-Limit the ad by ‘Connection’ and make it exclusive to your Facebook pages’ fans or friends of fans.
-Target by custom audience and target recent customers, lapsed customers, or users who have entered your contests in the past.
-Test different sunglasses. I know it’s tempting to think this small a factor wouldn’t matter, but small changes are the basis of A/B Testing. Here’s a real-world Facebook Ad example from BlueNile Jewelers:
Hopefully you now have the confidence to A/B test your Facebook Ads. You’d be surprised at how impactful a small change in demographics, image, text or detail can be. Remember to implement only those changes that had statistically significant effects on your Facebook Ad’s performance.
Have you had successes with A/B testing your Facebook Ads? Was there anything that surprised you? Start the conversation below.
–James Scherer is a social media expert and blogger for Wishpond. Wishpond makes it easy torun social contests and promotions on Facebook, Twitter, websites & mobile.