Facebook Ads: Split Testing 101
By Mike Hauben
“Split testing” is one of the first things you should do for your jewelry store or brand’s Facebook campaigns. Let’s learn a bit more about this fundamental technique and how we can implement it.
The advancements in digital marketing have taken an enormous amount of guesswork away for business owners. Tracking results from a billboard, for example, is very difficult. Online techniques, however, enable us to know exactly which campaigns are driving traffic and revenue.
One of the simplest and most effective ways to test which campaigns are working best is by Split Testing. Split Testing lets us test different variables in our ads to see which combinations of words, text, and targeting are working the best.
This is a strategy we should use in all our online marketing – not just Facebook Ads. We can test different landing pages or email subject lines, for example, to see which ones convert the most customers.
For now we’ll focus on split testing Facebook Ads.
Setting It Up
We’ll begin by going into our Business Manager account and beginning the standard process for creating an ad. For the sake of this article, I’ll choose Traffic as my marketing objective.
You’ll notice at the bottom you’re offered the option to create a Split Test (see the red box in the below graphic).
If the “Create Split Test” button isn’t clickable, it may be because the “Campaign Budget Optimization” is switched on (you can’t do both). Switch that off and you’ll be able to switch the Create Split Test button on.
After clicking this button “on”, you’ll see a new prompt asking to choose the Variable. This is where you’ll decide what exactly it is that you’re testing.
Choosing A Variable
There are 4 different things you can test.
Creative – You can test to see which graphics and creative elements are most effective.
Delivery Optimization – You can test to see which Delivery method – aka marketing objective – is most effective (Link Clicks, Impressions, etc)
Audience – You can test which audiences are reaching your objectives best.
Placement – You can test which placements are getting the best results (Instagram, Facebook news feed, etc)
Choose whichever one you’d like to test. I think a really important benchmark to learn about your store is to first test Delivery Optimization. This will help you learn which types of campaigns work best for different types of offers you regularly run.
Now after clicking to proceed to the next screen, your ad creation process may look a little different than normal. Thats because depending on the Variable you chose, Facebook will now give you different prompts. Choosing to test for best Delivery Optimization, for example, will give you a new screen at the end to choose which Delivery Optimization methods you’d like to test for, like this:
However, if you chose to test the Creative in your ad, you’ll instead be asked to create two different ads.
Simply follow the prompts and finish creating your Facebook ads, and click to run your ads.
You’ll now notice a copy of each ad (one per variable) in your Ad dashboard. I recently ran a split test to test the best Delivery method for a business. Here’s what it looks like in my dashboard:
Very clearly we can see that the ad optimized for Link Clicks and Engagements both drive similar amounts of clicks for a similar budget. You’ll also notice that the ad optimized for Impressions significantly under performed those two – only 7 clicks across 25,720 impressions! This really shows you that optimizing for the right goal is incredibly important. If our goal was for general awareness of our product, Impressions may have been great because so many people have seen that ad. However, we’re trying to drive sales in this case, so we want to focus on link clicks. I’ll now eliminate the Impressions campaign and focus on a campaign for clicks.