Stop The Split Tests! You're Ruining Your Ad!

Austin Distel      Thursday, November 2, 2017

Share this page with a friend

"Why Split Tests Are Ruining Your Ads?"

 

This young entrepreneur asked me one of the best questions I’ve gotten on a podcast recently. “Austin, what is something you used to believe that you have changed your mind on?”

 

Great question, right?

 

I had to take a moment and think. There were a few different directions I could have taken it. Should I talk about my beliefs on work-life balance? Maybe my change of mindset from cash today to now lifetime value?

 

No - while those are important, his audience will want something actionable. And since we talked a lot about Facebook advertising, I’ll focus on that.

 

Okay, I got it! This one was a big eye-opener for me...

 

I respond to his question with, “I used to believe successful Facebook ads  were the result from a ton of meticulous split testing. Now I believe that’s the worst thing you can do for your campaign.”

 

Let me explain. My team and I have spent $483,948 on ads in the last two years (I just checked my business manager as I’m writing this).

 

The way we used to run ads was a 4-step process.

 

1. Research like crazy on the audience

 

2. Create several variations of copy, design, call to actions, & descriptions

 

3. Split test all the ad variations against the same audience for a few weeks

 

4. Determine which creative had the highest click through rate (CTR) and scale

 

During that process, here’s something interesting we discovered…

 

Which part of an ad creative has the highest impact on CTR?

 

 

From running all those split tests of copy, video/photo, design, call to actions, and descriptions here’s the average percentage of variance for each section of the ad.

 

Screen Shot 2017-10-01 at 5.29.25 PM.png

 

Pretty nifty, yeah?

 

While running all these split tests gave us access to these insights, it actually was not our biggest “Ah-ha!” from the experiment.

 

It came when we realized every time we ran a new creative, we lost all the comments, likes and shares. Even if you correct a spelling error in your ad, Facebook will see it as a new ad and remove all the social proof.

 

On top of that, our budget at the time was $100 a day with only $5 to each ad. With each ad only spending a few bucks, it took forever to build up any comments, likes and shares.

 

So we ran another test...or should I say, we didn’t run another test.

 

For the launch of our new campaign, we decided no more split testing the ad creative.

 

We’d give it our best shot on which creative would work for our audiences and let it ride. Allow the social proof to build up and see how that affects the ad’s performance.

facebook-emoji-reactions.gif

 

 

What we discovered was amazing. As social proof began to compound, the ad started to convert better and better. Over the course of a few weeks, our relevance score went up, CPM we down dramatically (over 30%), and more leads were coming in for less cost.

 

As I think more about why Facebook rewarded us for having social proof on our ads, it became obvious. We sometime forget that Facebook at its core cares more about its users than it does its advertisers. They want to ensure people on Facebook enjoy what they see.

 

So, when your ad develops a lot of engagement quickly Facebook thinks, “Oh look, people like this piece of content. Let’s show it to more people, and since this person is creating value for our users we’ll reward them by decreasing the price.”

 

Now let me make this clear, I’m not saying I’ve stopped split testing everything. We only avoid split testing our ad creative. I’m still constantly split testing audiences and sending them all to the exact same ad.

 

 

Here’s how you do that.

 

Step 1. Build your campaign as you usually would. Choose your campaign objective, audience, placement and budget, then creative.

Screen Shot 2017-10-01 at 7.40.25 PM.png

 

 

 

Step 2. Go to your ad creative in business manager and preview it. It’ll open in a new tab with your ad. Grab the last group of numbers at the end of the URL - that’s your Post ID.

 

Screen Shot 2017-10-01 at 19.31.23.png

 

Step 3. Make a new audience or placement that you want to test. When you get to the ad creative, instead of building a new ad click “Use Existing Post” and then “Enter Post ID”. Paste your other ad’s ID here, and now your new audience will see the same ad as your first audience (and racking up all the social proof).

 

Listen to your intuition, give it your best shot on what you think your audience will like, launch your ad and let it ride.

 

Articles By The Author


Comments