top of page

Facebook Experiments Section: Science and Creativity Collide

Running an ad on Facebook costs money. And I don’t know about you, but when I spend money, I like to make sure it is well spent. Well, Facebook agrees.

When running an ad on Facebook, businesses have the option to run different types of experiments on their ads. The purpose here is to allow businesses to make informed decisions about which ads are most effective.

Types of Testing

  • Lets you compare two or more ad campaigns to see which one is performing best on a cost per result or cost per conversion lift basis.

  • Exposure is evenly split and randomized between each version to determine a winner.

  • If you choose to measure cost per conversion lift, it will also include a holdout to compare the incremental conversion lift of each.

  • Measures the incremental conversion lift of a specific ad campaign or all of your active advertising.

  • Your audience will be randomized and divided into groups that will or will not have the opportunity to see your Facebook advertising to determine the difference your advertising makes on conversion activity, such as purchases, leads or app installs.

  • Lets you use a brand survey to measure what incremental effect your advertising has on brand awareness, perception or recall.

  • Your audience will be randomized and divided into groups that will or will not have the opportunity to see your Facebook advertising.

  • Then your audience is surveyed with poll questions you can choose

  • Certain budget minimums apply.

  • Lets you use an existing ad campaign as a template to see how campaign budget optimization affects your cost per result performance.

  • The test will copy your ad campaign and automatically create an A/B test, one with campaign budget optimization turned on, and one with optimization turned off, so you can easily compare performance between each strategy.”

What’s New

This week Facebook launched a new feature to help advertisers manage these tests. There is now a section called Experiments in Ads Manager. This is a central location for all of a user's ads and results. Facebook says “with a more streamlined interface in Ads Manager, it’s easier to choose the right test for a business goal, create tests, scale what’s working, integrate insights into marketing strategies and help a business grow."

The Experiments Section adds no new functionality; it is meant to increase the ease with which users can sort data and to understand the context of each test. While all of this information was previously available, it is easier to access now.

Why test

Testing adds a level of science to your ad campaigns. Team members will make suggestions based on experience and training, but that cannot be the end of it. People have been known to be wrong. Jan on the team says let’s use puppies. Run a test to see if puppies perform better!

Using tests and using the new Experiments Section does not cost any money; however, Facebook does say “but Experiments can affect the overall cost per impression (CPM) of ad campaigns in your test depending your test type and bidding strategy. This is because tests that measure cost per conversion lift or brand lift include holdouts, or people in your intended audiences that are intentionally withheld from seeing your ads. This is necessary to understand the incremental value of your Facebook advertising against your other organic and paid marketing efforts.”

When to use testing

Running a test can take a lot of time, so knowing when to test and when not to test will save time and ultimately, money.

Here are some ideas of what is a good to test:

  • Imagery

  • Headlines

  • Hard sells vs soft sells

  • Call to Actions

  • Company wide messaging

  • Generic v persona specific ads

  • Content Length

Most importantly, anything you test has a hypothesis so your results are clear.


bottom of page