Facebook Advertising: 6 Ways to Split Test Like a Pro [Infographic]

One of the biggest reasons that brands fail in their Facebook advertising efforts is that they do not split test. They create one ad or two ads or three ads… They don’t find success and they label it a failure.

The key to success with Facebook advertising is no different than the key to success in almost anything: Fail. Fail a lot. Find what works by trying out things that don’t work.

Sound crazy? It’s not. The truth is that most marketing efforts are great ideas that won’t work. If you put all of your eggs into that one basket — without testing and retesting — the odds are stacked against you.

That’s why you must split test. You must try out dozens of variations, fail like crazy and find what doesn’t fail. When you realize something that doesn’t work, stop spending money on it. When you discover something that works, ride it into the ground.

The following should act as your checklist of ways you need to split test your advertising.

1. Copy

You have 90 characters to convince a Facebook user to Like your page or buy your product. That’s hard. You will fail a lot.

In other words, you should split test the heck out of this. Create identical ads in every way, including targeting, but with one exception: Switch up the copy. Use a different call to action. Substitute words. Try a different tone. Use an exclamation mark in one but not the other.

How you say it is often more important than what you say.

2. Imagery

You’ve been told that imagery is the key to advertising and to success on social media. The problem is that you don’t know what images people will respond to best to make the desired action. And that imagery will be different from person to person and demographic to demographic.

Use different faces. Use an inanimate object. Flip the direction that the person is looking. Use wide and narrow images. Try one that’s black and white and one that’s color. Use a photo with text on it and one without.

3. Destination

You created a terrific ad that attracts the attention of Facebook users. It inspires clicks. The Click Through Rate is incredibly high. Yet… Those same users are abandoning your Page instead of sealing the deal.

Why is that? It’s often because your landing tab didn’t provide what you promised in your ad. So with the same ad, create multiple landing tabs and find which is most successful at fulfilling your business goals after the initial click.

4. CPC vs. CPM vs. Optimized

Sometimes you can save money by bidding CPM instead of CPC. Sometimes you’ll throw that money away.

Likewise, Optimized CPM and CPC targets your ads at people most likely to perform a specific action. That also comes with the prospect of much higher costs and bots.

What works best? It depends. Try each one within the same campaign. Whatever works best, run with it.

5. Demographic vs. Demographic

Your ad may be perfect, but it could be targeted at the wrong audience. Maybe your imagery, offer and copy will reach optimal efficiency when targeted at 40-50 year-old men, and you’re throwing money away when targeted at everyone else.

How will you figure out if this is the case? Split test like crazy. Select different imagery and copy that will appeal to a specific demographic. Or try out a generic ad, split tested to age and gender groups.

6. Detailed Targeting vs. No Targeting

Sometimes, you can over think it. If you target an ad too much, you can cut out a segment most likely to positively respond. And when you loosen up targeting, ad pricing typically goes down.

So when you create an ad that could potentially appeal to a large audience, try running it without any targeting at all and see how it does. Who knows? You may have a winner!

You Can Only Fail if You Don’t Try

Don’t tell me in one breath your Facebook advertising failed and in the next say that you only ran one or two ads. The Facebook ads didn’t fail, you did.

Do me a favor: Next time your advertising is struggling, stop. Try something else. Try a dozen “something elses.” Don’t throw money away, but keep trying something new.

Are there any other ways that you split test your ads? Let me know in the comments below!

Facebook Ad Split Testing Infographic

  • Pingback: Facebook Ads Guide Review Overview()

  • Pingback: Key to Facebook Advertising Success: Split Test Like a Pro | JonLoomer.com | Social Media Trends & News | Scoop.it()

  • AmytranterMKTG

    Do you split test within same campaign or divide into separate campaigns since FB let’s the more successful ad in a campaign be shown more frequently?

    • http://jonloomer.com/blog Jon Loomer

      Hi, Amy! I usually start within the same campaign, assuming the changes are all related to the same campaign. If I’m not getting any impressions from an ad, I’ll create a separate campaign from it, but that doesn’t tend to be necessary.

  • Pingback: Adverteren op Facebook: een compleet overzicht | Likeconomics.nl()

  • http://www.facebook.com/dianecossie Diane Cossie

    Jon I have only just stumbled across you and have just started FB ads having gone all out on my personal file and Twitter I know the importance of FB so will follow many of your great suggestions about split testing! Thanks for the great advice and your Ebook “liked” your page and look forward to sharing the journey with you.

    • http://jonloomer.com/blog Jon Loomer

      Awesome! Thanks so much, Diane, and welcome aboard!

  • Pingback: Adverteren op Facebook, hoe moet dat? - Tijl Vos()

  • Pingback: The Ultimate Facebook Marketer's Checklist for 2013 - JonLoomer.com()

  • Sven Teuber

    Very informative, thanks for the write-up. Can’t agree more that split testing any marketing effort is key. Also, so true that people just give up too easily. By the way how are people responding to the voice messaging to the right?

  • Michael

    Hi Jon, I’ve enjoyed reading your posts. I’ve put a lot of time into FB ads, yet I still wonder: how can we test ads scientifically — e.g., how can we evaluate if one image or headline is better than another — if Facebook refuses to show ads evenly? I believe Facebook determines much too quickly which ad in the campaign is the “winner”.

    I launched two campaigns with similar audiences and duplicate ads. In one campaign, it was clear which ad Facebook chose as the “winner”: one ad had gotten the most impressions. In the other campaign, this same ad got almost no impressions, and not one click. It didn’t even get enough impressions to disqualify it as a bad ad. Another ad went on to be the “winner”.

    Putting ads in their own campaign never seems to do the trick… FB loves having ads in a campaign compete against each other.

    I understand that Facebook only wants high-quality ads, but marketers need the experimental data! That’s how we make our ads more relevant.

    How have you dealt with this problem? How do you actually split test?

    • Kenneth

      Hi Michael, I was just going to ask / post something similar – I find that very bizarre how Facebook doesn’t (at least for a few days or a week) display ads evenly within a campaign. Google Ads does this, I’ve always wondered why FB doesn’t. The only way to truly split test ads seems to be by creating an entirely separate campaign for each ad.

    • Brandon_Becker

      Here’s my workaround:

      If I have an ad set with many ads, and I notice one ad is stealing all the traffic, I’ll just turn that ad off temporarily. Then usually a second ad will start stealing all the traffic, so I turn that one off. I repeat this until each ad has ~400-1,000 reach, then I can make some informed decisions about what’s working.

      This is your ‘testing’ phase, usually 10 – 20x less than your projected campaign spend (so you don’t ‘waste’ too much on failed ads). Then take the rest of your budget and blast into ‘production’ mode with your winners until they die.

      At least that’s what I do.

  • Pingback: How to Create a Saved Audience in Facebook Power Editor - JonLoomer.com()

  • Pingback: Facebook Ads Tip: How to Create a Dark or Unpublished Facebook Post - JonLoomer.com()

  • Daniel Chiteji

    How long do you give an Ad before you make the necessary changes to split test if you are not getting the results you desire?

  • Ramanvir sodhi

    Hi Jon, Can we do split testing on promoted page posts for engagement ads? Can I change the og:image and create different ads under one ad set, to analyse CTR?