How Facebook Ads A/B Testing is Done to Maximize Ad Performance


Facebook and the billions of users who frequent the platform daily have been a boon for the marketing world. When the Facebook Ad platform launched, marketers heralded the social giant’s targeting capabilities as the next big thing in marketing. Aside from Google Ads, no other platform had the same level of audience targeting and engagement that Facebook Ads did and it didn’t take long for people to notice.

But that was the past. Today, everyone is using Facebook ads. Rightfully so, it’s a great advertising platform. So great in fact that Facebook captures nearly 20% of all digital advertising spend in the US. So if everyone is using Facebook Ads, how can your ad campaigns standout against the competition?

If you’ve been working in marketing for any length of time you’ve likely heard of A/B testing. Especially in the advertising world, A/B testing is a necessity if you plan to maximize the performance of your ad campaigns.

Thankfully A/B testing has never been easier to perform. I’m talking doubling, even tripling your Facebook Ad’s click-through-rate.

In this article I’ll break down everything surrounding Facebook Ads A/B testing. First we’ll quickly run through the concept of A/B testing. Then I’ll give you five variables you can test today. I’ll finish with how to interpret your ad’s performance with Wishpond’s ad tool (though the rules apply for any ad tool). I’ve made sure to include examples to help you along so let’s get started.

What is A/B Split Testing?

A/B testing is a strategy in marketing in which two versions, A and B, (the Control and the Treatment) are tested against each other. The goal is to identify changes that increase the chance of the what you want to occur, occurring.

It’s used commonly for webpages, landing pages, marketing emails, and advertising. The best practices remain essentially the same across the board though:

  • Change the placement and formatting of objects on a page to get a user’s eye to where you want it
  • Change colors to ensure primary objects stand out
  • Change images to be eye-catching or elicit emotion: encourage engagement
  • Change text to be appealing to the reader, and encourage a desired outcome

With advertisements, we get to add one more, changing your target audience.

Statistical Significance:

I won’t get into this too much, but need to mention it, as a statistically insignificant change can be seen as incredibly exciting, and we wouldn’t want you wasting your ad budget on a mistake.

  • Before you invest your hard-earned ad budget on your ‘better-performing’ ad, run the results through this free A/B Test Calculator to be sure.
  • Unless you’re seeing a confidence level of 95% or more, you’re going to need to run the ad again.
  • A lack of statistical confidence can be caused by a small sample size (too few people viewed your ad)
  • Or not enough of a difference between the two ad’s performances (conversions or click-through-rates) to see a statistically significant result.

Our Hypothetical Facebook Ad

For this article, my “A”, or Control, advertisement will be for a spa weekend sweepstakes for two. This hypothetical ad would drive traffic to a contest landing page. This kind of full-service interaction, between your Facebook Ad and any kind of landing page (E-Commerce, SaaS, whatever), is in beta at the moment, and Wishpond will be releasing it in the next couple weeks. Needless to say we’re excited.

Anyway, here’s the hypothetical ad – our Control for the following A/B tests:

Below I’ll give four examples of how to create different versions of this ad and discuss targeting strategies as well as how to meaningfully interpret your Facebook Ad testing performance. Let’s optimize your ads.

1. A/B Testing your Image

Images are the most important part of our Facebook Ad because they are what catch the eye of users.

Facebook Ad Rule-of-Thumb: Engage with an image, convert with a value proposition.

Facebook Ads compete with user’s friends and family on the News Feed. Your image has to be more eye-catching than a picture of a user’s nephew covered in spaghetti sauce.

So let’s say we ran the control advertisement above for two days, weren’t getting a lot of clicks, but did see solid conversion. This could mean people were liking our title, offer, and landing page, but few were being attracted to our ad in the first place. So let’s make the image a bit more eye-catching:

A little more red, a little more skin – let’s see how it does. To A/B Test this image we run it for the same amount of time as our Control with the same other variables (landing page, copy, title, etc). We check its performance after two days and determine if we’ve seen an increase in our CTR.

For more information about optimizing your Facebook’s ad image. Read my recent article 6 Facebook Ad Image Best Practices that will Send your Click-Through Rate to the Moon.

2. A/B Testing your Title

After your image, your title is the most important factor in catching the eye of your audience.

In our Control advertisement we’ve gone with the value proposition title. Value propositions are one of the chief ways marketing professionals convince someone their ad, marketing email, or landing page is worth engaging with.

Many companies will use their logo for their title, which is fine if your logo is recognizable – otherwise it won’t grab the eye like you want it to.

Another title option would be a great CTA like ‘Enter to Win!’ or ‘Lose 20 pounds’.


Top Tip from Kevin Spidel: Use Pinterest to find images. They are already socially tested. The images with higher re-pins tend to get great CTR on Facebook ads.

3. A/B Testing the Details

One of the amazing things with A/B Testing (and sometimes the most surprising) is how effective a small change can be. A border around your image, a slight shadow, the word ‘now’ within your body copy, or the subtlest nuance a graphic designer throws at an image and suddenly you’re seeing a 100% increase in click-through-rate.

You don’t need a graphic designer either. Microsoft Paint, or free online photo editor tools like Canva work well enough.

Details you can test:

  • 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

4. A/B Testing your Ad Body Copy

I said above that the image engages, and your title and copy convert.

What does that mean?

It means without good copy all you’ve got is someone to pay attention (which is a feat in itself on Facebook, but 100% immaterial if they’re not clicking through). The body copy of your ad is where you communicate the “why”.

The copy for our Control advertisement is already pretty good. But what if we tried a whole different approach and tone? What if we targeted parents and sold them on a weekend away from the kids?

In the image above you’re technically seeing a ‘multivariate’test as we’ve changed more than one variable within the Facebook ad. A true A/B test would be to change only the copy and keep the image of two people getting a massage. But Wishpond’s Ad Tool makes it so straightforward to change the text and image, as well as create multiple advertisements, that we might as well test multiple variables if we think it’ll improve performance.

5. A/B Testing your Target Audience

Keeping your Facebook Ad exactly the same but changing who sees it is a fantastic way to increase your conversions with little effort. Wishpond’s Ad Tool allows you to target by the basic demographics (age, gender and location) but also by precise interest, broad category interests and connections. Let alone enabling you to target by existing contact list.

Let’s look at a couple ways we can test changing the target audience of our Facebook Ad:

Version 1 (age):

A/B Testing an image of a middle-aged couple (instead of the young couple in our Control ad) would be worth trying, even before we test targeting two different age demographics.

But a lot of A/B Testing is common sense. It makes sense that an ad targeted at 35-65 with an image of a middle-aged couple would perform better than an ad with an image of a young, mid-twenties couple.

Version 2 (Broad Category):

Wishpond’s Ad tool allows you to target based on Broad Interest. Broad Interest targeting allows you to target your audience based on what they’ve included in their timelines (relationship status, political leanings, travel, birthdays, new job, recently moved, etc).

This is great for our spa ad in the ‘need a break?’ version above. With broad category targeting we can show our ad to only parents – ensuring we’re not wasting any of our ad budget. We can do the same by targeting only people in relationships (because a romantic spa day for two is useless if it’s targeted at a 64 year old widower).

Other Targeting Strategies:

  • Limit the spa ad to married men and change the copy to ‘Give her a break!’ // ‘Does your wife need a weekend off? Surprise her with a 500$ spa weekend. Enter now to win.
  • Limit the spa 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

Top Tip: A/B Testing your target audience is one of the primary reasons why this is for small businesses as well as huge corporations. You want to get the most out of your ad budget, and targeting the best possible audience is how you ensure you’re not wasting a cent of your hard-earned money.

Interpreting your A/B Test’s Performance

Unless you know which of your advertisement versions is performing best you’re wasting your time. A good analytics page is important, and knowing what you’re looking at is essential.

As you can see I’ve created three hypothetical advertisements to test. Advertisement A is our Control. Advertisement B is our ad with the change of image. Advertisement C is an ad targeted at the 35-65 year old demographic segment.

To find our Click-through-Rate we divide Clicks by Views. For Conversion Rate, we divide Conversions by Clicks.

All the results you see below are statistically significant with at least a 95% confidence level.

Ad A: Our Control ad had a click-through-rate of about .05% – very respectable for most ads, but not fantastic with contest advertisements. Our conversion rates were about 29%.

Ad B: The change in image has increased our Clicks substantially, giving us a click-through-rate of almost .09%. Awesome!

Our conversion rate stayed about the same at 28%.

Ad C: As you can see, our ad views have dropped off with the change in audience (targeting only 35-65s). This is because we’ve essentially cut our reach in half. Did it pay off?

Our click-through-rate was .07%. So yes, it did pay off – but not as substantially as I’d like. So maybe we go on to test the same demographic target with a middle-aged couple in the image and see what we get.

Our conversion rate was 35%.

What does this mean?

It means A/B Testing works. It means a simple change of image increased our click-through-rate by 80%, and another change of image and target audience doubled it.

It also means if I ramped up my Reach metric (to 10,000 views, for instance) I could be getting almost 300 conversions from one ad.

As far as conversion rates go, I like seeing this kind of consistency (around 30%). However, if I’ve worked this hard to attain a Facebook Ad click-through-rate of .1%, I would definitely take a look at A/B testing my landing page. You’ll have to wait a bit for that article, though.


Hopefully you now understand the importance of A/B testing your Facebook Ads and can do it for yourself. Hopefully you get how simple it can be, and how amazing the results.

Further reading:

Have you had any successes, or frustrations, with A/B testing your Facebook Ads? Wondering about integrating your Facebook Ad and landing page with one tool? Start the conversation below, or visit


Leave a Reply

Your email address will not be published. Required fields are marked *