The devil is in the details… And Facebook knows it! Perhaps that’s why the network has introduced Split Testing. Their objective? To allow you to test ads by playing with different variables in your campaign settings, in order to determine which strategies produce the best results. Please note: these are not A/B tests on creative ad content, but on delivery strategies. Split testing creates distinct, non-overlapping groups within your audience, who will be exposed to different variations of ad sets. It’s also worth noting that split tests are based on people, not cookies, and therefore collect results across multiple devices.
Facebook ads split test variables
Testing means testing variables! Advertisers can run split tests on three variables: target audience, broadcast optimization and placement. To ensure that results are easy to read, only one variable can be tested at a time.
-
Target audience
The Target Audience variable is used to compare ads across two or three distinct audiences (e.g. 18-30 vs. 31-43). Please note that split tests are only effective for saved audiences.
-
Diffusion optimization
The Delivery Optimizations variable lets you run a split test to compare different types of delivery optimization. For example, with the same audience, you can determine whether your ads are more effective at generating conversions on your website, or at generating clicks.
-
Investments
The Investments variable lets you use a distribution test to compare different investments. This makes it possible to analyze the effectiveness of automatic versus customized placements. You can test your ads on Facebook, Instagram and Audience Network, and determine how changes in variables affect campaign performance.
Objectives of Facebook ads split testing
For the time being, Facebook Ads split testing can only be used for three commercial purposes:
- Increase the number of conversions on your website (Conversions on a website)
- Encourage installations of your application (Application installations)
- Get leads for your company (Lead generation)
Budget and schedule for split testing Facebook ads
In order to be able to carry out meaningful tests, Facebook determines a minimum budget requirement based on the size of your splits. This budget is then divided among the ad sets. Test duration can vary, but it’s recommended to stay within a window of 3 to 14 days. The shorter the window, the less time is needed to determine the most effective ad, and the more useless it is, as the “winner” is usually known before that time.
Test results by breakdown
Facebook analyzes the performance of each set of ads in relation to the campaign objective, and determines the winning set. Once the test has been completed, the results are sent by e-mail, enabling the advertiser to adapt the campaign. Split testing is therefore a very useful tool for determining the best distribution strategies, allowing you to avoid relying solely on your intuition, which isn’t always the best advisor when it comes to Facebook advertising…
The experts at our Facebook Ads agency are at your disposal to support you in this process. The same goes for promoting your product catalog on Facebook.
