Your Guide to Split Testing Ads in Facebook

When I started selling lemonade as a kid, I couldn’t figure out which garnish my customers would prefer the most.

Could it be mint, rosemary, or basil?

I remember getting the, “I’m-not-sure-either” answer when I asked my parents about it. And so, the solution I came up with was to offer 1 type of garnish per day, for the entire week, and see which garnish would get more sales.

Interestingly enough, the end results showed that my customers prefer not having any garnish in their lemonades.

Pretty anticlimactic, huh? 🙂

But wait a minute, why am I talking to you about my experience as a kid selling lemonades? What does it have to do with use selling shirts, you might?

My answer is… everything!

In case you did not notice, I was split testing while I was selling the lemonades.

Of course, I didn’t know at that time that I was split testing either. It just made sense for me to do it so I could figure out what kind of lemonade my customers would want.

And so the question becomes, are you doing the same thing with your Facebook ads? Or, are you running your ad campaigns blindly with your fingers crossed, hoping that you’re showing the right ad designs to the right audience, with the right elements on your marketing materials?

If you answered in the affirmative, then allow me to help you change all that. Then allow me to share with you how to split test your ads, so you can stay clear from the guessing game and be able to optimize your ads for better conversion.

But before anything else…

What exactly is split testing?

Split testing or A/B testing is a strategy where two versions of an ad, landing page, or an app (among others) are tested against each other to figure out which one will give you the best results.

Split testing is important because it gives you a better understanding of how your audience behaves towards your marketing materials. Of course, the more you understand your audience, the better you can optimize your marketing materials to improve its conversion.

The most common elements to split test.

While you can run split tests on virtually any part of your Facebook ad, these are the ones that would tend to give you the most impact (in no particular order):

  • Call-to-action
  • The target audience
  • Header
  • Image
  • Ad placement
  • Copy/text

Facebook’s ad campaign structure

I’d like to talk about structuring split tests since it can be quite confusing if you are new to it.

However, before we can even talk about it, you first need to have an idea of how Facebook’s ad campaign structure works. Comprehending this will give you a clearer picture of how to go about structuring your split tests.

Facebook’s campaign structure has 3 levels:

– Campaigns

This is the level where you can add your overall campaign budget, and the objective of your campaign, among other things.

– Ad sets

If there are two elements that would cross my mind when thinking about the Ad set level, those would be, “defining the target audience” and “the budget for the specific audience.”

Of course, the ad set level has other functions as well. However, the two that I mentioned are the highlights.

– Ads

What I’d like to highlight on this level is the word, “design.”

These are the visible parts of your ad.

Facebook Ad with break down its parts

Allow me to give you a bit more information on each part.

Text – Facebook will display 500 characters in this area. However, you might want to use Power Editor instead, since Facebook will allow you more text.

Image – It’s best to use high-resolution images so you can avoid your ad image looking fuzzy on devices with large screens.

Remember that the suggested size for your image is 1,200 x 628 pixels, and that Facebook will decline ads with texts that are 20% more than the image.

You can use Facebook’s Grid Tool to see how much text your image has.

Headline – Facebook will display 1- 2 lines on your headline. Your headline needs to have the punch as much as possible since it is one of the most visible parts of your ad.

Facebook will display 1 – 2 lines on your headline. You need to keep your headlines as punchy as it can be since it is one of the most visible parts of your ad.

News Feed Link Description – Facebook will display around 2 – 3 lines in this section. A good practice that most seasoned marketers use is to add their CTA here.

* Important note – the numbers I shared with you are ads based on websites/links conversions, and how they would display on a desktop news feed. If you’d like a guide that details Facebook’s design and display specs,  you can check this out.

Website URL – Even if your URL has extensions, the only thing that will appear in this section is the domain name. That being said, even if you add here, the only thing that will appear in the ad is TEESPRING.COM.

Call-to-Action – The ad I showed above doesn’t have a call-to-action button. However, this is how the button would look like.

Facebook Call To Action

These are the CTA options that you can choose from:

  • Apply Now
  • Book Now
  • Contact Us
  • Download
  • Learn More
  • Shop Now
  • Sign Up
  • Watch More

Now that we have a better idea of how each level works let’s proceed to structure your split tests.

How do I structure my split tests?

As you can imagine, there are a couple of ways to structure your split tests.

Others use this pattern:


  • Adset: Target audience – Firefighters in California | Budget $20
    • Ad 1 – Header 1 – $4
    • Ad 2 – Header 2 – $4
    • Ad 3 – Header 3 – $4
    • Ad 4 – Header 4 – $4
    • Ad 5 – Header 5 – $4
  • Adset: Target audience – Firefighters in New York | Budget $20
    • Ad 1 – Header 1 – $4
    • Ad 2 – Header 2 – $4
    • Ad 3 – Header 3 – $4
    • Ad 4 – Header 4 – $4
    • Ad 5 – Header 5 – $4

In the split test above, we’re looking to target firefighters in different states, and the element that we’re trying to compare against are the headers.

That being said, everything in the ad design is the same, the only thing that differs are the header texts.

Once we’ve figured out which type of header is giving us the best results, then we can move on to split testing other elements.

There’s a problem, however…

If you think that Facebook will simply divide your budget evenly — meaning, Facebook will allocate $4 per ad since there are 5 ads and your adset budget is $20 — you are gravely mistaken.

You see, Facebook has a tendency of unevenly distributing your allocated budget. Facebook will tend to allocate most of your budget to the ad that they think is performing best.

This results to your other ads being under-tested.

To avoid this from happening, you might want to use this structure instead:


  • Adset 1: Target audience – Firefighters in California | Budget $4
    • Ad 1 – Header 1
  • Adset 2: Target audience – Firefighters in California | Budget $4
    • Ad 1 – Header 2
  • Adset 3: Target audience – Firefighters in California | Budget $4
    • Ad 1 – Header 3
  • Adset 4: Target audience – Firefighters in California | Budget $4
    • Ad 1 – Header 4
  • Adset 5: Target audience – Firefighters in California | Budget $4
    • Ad 1 – Header 5

As you can probably imagine, with the structure above, none of your ad designs will be under-tested since each ad design will have an equal amount of budget. Considering how everything is the same across each ad design (minus the header), we’ll be able to tell accurately which header works best.

Split test samples.

To give you a better idea of the kind of value that you can get from split testing, let’s take a look at this Facebook ad campaign.

Facebook ad campaign from

* Note – I work with NestOffer.

The way NestOffer works is we buy homes for cash at a lower price. We then rehab the property and put it back on the market for more money.

Here’s what we did.

We ran two Facebook ads where everything is identical, except for the image. The first ad didn’t have any price tags attached to them while the second one has the tags.

Basic Facebook Ad featuring NestOffer

(Ad #1 – This ad data was sourced from Compass)

A/B tested Facebook Ad

(Ad # 2 – This ad data was sourced from Compass)

These are the kind of results that we had after running both ads:

  1. The first ad managed to get a better CTR (click-through rate) 6.31%.
  2. The second ad only got us a 2.84% CTR.
  3. However, the second ad seems to have generated more — let alone better — quality leads, despite its CTR being lower than the first one.

* Note – I obtained these data because I work directly with NestOffer.

We had a couple of hypothesis in mind as far as why the second ad managed to get us more leads:

  1. Most people on Facebook were looking at the regular old house on the first ad and saw the word “BUY.” They must’ve assumed we were selling the house. We managed to validate this idea since there were people who commented on the ad asking how much we were selling the property for.

The second ad, however, pushed the people to look at the image and the copy because of the red price tags. We got the price tag idea from our competitors, btw.

Our audience realized after reading the ad that we were a cash buyer of homes and not a seller of homes.

Because the image on our second ad provided a bit of context and pushed people to read, it became an overall win.  

There are a couple of lessons that we can learn from the split test that we ran:

  1. Just because an ad has a good CTR, doesn’t mean that it is optimized for generating leads. That being said, DO NOT base your judgment on whether or not to keep an ad purely on its CTR. You need to check whether it is bringing you quality leads.
  2. There is certainly value to be had when adding context to your ad’s elements (image, copy, header, etc…

As you can see, there are a plethora of benefits to be had when doing split testing. Not only did we get a better insight of how our audience reacted to our ads, but also, we’ve managed to determine which ad was getting us more leads (among others).

Learn from your competitors.

Before you even think about split testing, however, you can observe how your competitors are running their campaigns first. That way, you’ll be able to hypothesize the best route to running a successful Facebook ad — let alone a successful Teespring campaign.

There are a few benefits to be had from learning from your competitors:

  1. You’ll know what kind of CTA will tend to give you the best conversions.
  2. You’ll know what kind of hook, or angle to use when positioning your ads.
  3. You can check out their successful ad designs then build from it.
  4. You can get more ad ideas from looking at their audience’s comments.

Of course, I’m just scratching the surface with the points that I mentioned above.

And so the question becomes, “How do I spy on my competitors?”

That’s where we come in.

Just like how we copied our 2nd ad’s idea from our competitor — We Buy Ugly Houses, you can also do the same thing by running a quick search on your keywords using our software.

Here’s what we came up with when we typed in our keywords.

The Facebook ad we based our test on

When we saw this ad, we liked how the price tags almost immediately invoked a sense of curiosity to my team and me. And so, we decided to apply this element to our ad’s images as well.

There are a couple of things that you can learn from observing our competitor’s ad:

  1. They talked about a problem that their audience is commonly struggling with (fixing their house before they can sell it), making their ad relatable” to their audience.
  2. Their photo is catchy and has an element of curiosity about it.
  3. They have a clear call to action.
  4. They were clear about the kind of benefits that their audience can experience.

At this point, I hope you realize the value of observing your competitor’s ads. Not only does it help in pointing you in the right direction as far as what kind of ad design works, but it can also give you an idea of how your audience thinks by reading their ad’s comments.

Click here to sign up for our beta version

What’s next?

Have you been split testing your Facebook ads for your Teespring campaigns? What are some of the best practices that you can share with our audience? Please take the time to share them in the comments below. Cheers!


For more growth hacking tips, tricks and marketing guides check Rand’s marketing blog.

About the author

Rand Owens

Chief Growth Hacker at

Leave a Reply

Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

Copyright © 2014. Created by Meks. Powered by WordPress.