Facebook Advertising Split Test Case Study

Facebook Ads Manager has some great functionality for advertisers, including the ability to run split tests.  According to Facebook’s documentation “Split Testing helps advertisers understand how different aspects of their ads affect campaign performance. Similar to A/B testing, Split Testing lets you test different versions of your ads so you can see what works best and improve future campaigns.”

This has proved invaluable for the Facebook ads I run for our clients.  In the case below, the conversion goal was for consumers in a specific metro area to go to the web site’s dealer locator page and find a local dealer. The client is a building product manufacturer with a B2B business model, so this is the closest we can get to a sale from the site.

Two split tests were run.

The first test called “Audience split test” compared cost per click (CPC) and click through rate (CTR) of serving ads to consumers vs. builders.   After spending $800 over 2 weeks, the consumer audience click through rate had only a .06% advantage over the builder audience, but their cost per click was $.10 cheaper at $1.06 giving consumers the edge.

The second split test is called “Delivery Type” and was a 3 way test comparing the delivery type’s effect on conversion rates (find-a-dealer page visits on the client’s web site) delivering to consumers (the winner of the previous test) as well as testing 6 different ad images for each delivery type.

The types included:

  1. delivering to users Facebook deemed more likely to go to the find-a-dealer page on the web site (conversions)
  2. delivering to users Facebook deemed more likely to click on the Facebook ad
  3. delivering the Facebook ad as many times possible to as many users as possible within the target area

The first delivery type (conversions) was quickly the clear winner after only 24 hours, as was one of the six images tested.  The two losing delivery types were shut down and the conversions type was allowed to continue to run.  The losing 5 images were also stopped from running.  After $785 was spent on the winning combo over 13 days the conversion delivery type yielded 625 ad clicks and a CTR of 2.5% at $1.15 per click.


Summary – The winning combo was to deliver a specific lifestyle image to consumers who were likely to go to the find-a-dealer page. It yielded 161 conversions at an average of $4.88 per conversion.   During this time, this ad combo also yielded 1613 people visiting a product page.

Now that we know what works, and what doesn’t, I am currently running a remarketing campaign to serve this ad creative specifically to that group who looked at product, but didn’t look for a dealer in order to try to capture that audience who showed interest, but that didn’t take the next step.

I you have any questions or comments, feel free to let me know in the comment section below.  Thanks!

Also published on Medium.

Leave a Reply

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