Testing creatives on Facebook

Testing creatives on Facebook

Ekaterina Kostyuchenko

To make your campaigns more profitable and to spare your ad budget at the same time, you should split test different approaches and ad sets against each other to find the most effective ones. Let us talk about split testing, Facebook Ads features, dynamic creatives, and much more.

What you can test


1. Focus on the features of the item you promote. Highlighting various sides of the item may cause the audience to react differently.

2. Change the background and color scheme. It is difficult to predict which color palette will turn out the most attractive. Do not be afraid to experiment and do not restrict yourself to the standard color palette common in your company.

3. Try different graphic styles. For example, compare the way your audience reacts to a vector picture against a realistic image.

4. Depending on your product, try to play on the audiences' feelings or appeal to their common sense instead. Creatives for B2B services usually use rational style of argumentation. If you want to provoke an impulse purchase, try to evoke certain emotions in your audience.


1. Describe the specific properties of your product. Research your audience to find out what exactly they want to know about your product.

2. Add information about discounts and special offers.

3. Try out different calls to action (registration button, learn more button, go to website, etc.)

4. As with images, compare rational and emotional style creatives.

How to test creatives

There are many factors that can influence the way you test creatives: product, budget, ad placement, your strategy in itself.

We recommend allocating about 30% of the ad budget to test campaigns. If you set aside a small sum, it may not be sufficient to gather enough data to go on with the campaign launch. On the other hand, if you spend too much on testing, you will have next to nothing left for the actual campaign featuring the best creative.

There are different ways to test creatives:

  1. Test them all at the same time. You target the same audience with several ads, specify the time interval or cap the number of impressions. CTR of the ads will be your identifier in this case.
  2. Test them one by one. You target multiple ads at your audience, but you run them one after the other. If you have many creatives, run the test campaigns several hours apart.

These approaches require your immediate participation. Moreover, there are factors that may affect the reliability of test data. These two methods will work for a small budget or if you do not have enough conversions to automate the testing process.

Split testing (A/B testing)

In A/B testing you split the audience in two even groups and the new half-audiences do not overlap. A/B testing allows you to test more creatives, and the selection of users is more objective. To obtain reliable data, you should pour enough budget into testing and be patient.

Here are some variables that you can test: creative, audience, impression optimization, placements, set of items, personality variables.

87.1% of American marketers use Facebook. 44% of users stated that their purchasing behavior is influenced by the content they see on Facebook which sports more than 80 million online stores. 26% of the Facebook users tend to finalize the purchase after clicking on an ad.

Below you will find some variables for A/B testing: 

Creative. Compare different images, copy, or creative types. This way you can determine which version of the text is more effective when used with the same image or video or find out what performs better: a video or a carousel. You can manage creative variables at the ad level in Ads Manager.

Audience. Compare your campaign performance with different audiences or different demographic settings. For instance, you can configure different regions while setting up an audience or compare custom audiences with lookalike ones. You can manage audience variables at the ad set level in Ads Manager.

Impression optimization. Compare the campaigns with budget optimization and the ones without it. Thus, you can track how optimization influences campaign performance.

Placements. See how the placement affects the reach of your campaigns. This will help you define whether it is better to select placements manually or automatically.

Set of items. Compare different sets of items specified in Ads Manager. For example, split test two sets against each other and pick the better performing one. Set variables can be managed at the ad set level in Ads Manager.

Personality variables. You can use multiple variables for A/B testing if you need to test out complex hypotheses with a similar conversion rate. You can tweak a campaign, an ad set, or a separate ad using several variables.

You will have access to different variables depending on the way you set up the A/B testing. Fox example, if you create a copy of an ad, you can only test different creatives within the ad. If you copy a campaign or use the "Experiments" feature, you will have access to other variables for testing.

We recommend running first split tests with a single variable to better understand its influence on the overall performance. Below are several types of A/B tests depending on the variable you want to work with.

Going deeper into the Facebook Ads creative testing

Automated Ads

If you create an ad set with two or more creatives, the algorithm selects the most effective creative based on the statistics it has collected. Facebook Ads evaluates performance based on the in-built platform metrics.

Dynamic creatives

With dynamic creatives, you can run the same ad in different variations. The algorithm combines titles, texts, and images, and then chooses the best performing creatives and shows them to users.

How to launch dynamic creatives:

  1. Enable dynamic creatives when setting up ad campaigns. You will see the usual interface for creating ads.
  2. Add several options for images and text.
  3. Facebook Ads will combine them and launch campaigns.

This method allows you to quickly launch campaigns and test hypotheses, though you will not be able to stop one of the campaigns as the algorithm manages them all automatically. But you will get granular statistics on all the components of your ads.

Evaluating test results

Use metrics and the collected data for your analysis. For example, look at CTR.

If there is not enough data on the conversions, make CTR your main focus. The higher it is, the better the audience likes your creative. A low CTR indicates that the audience does not like or comprehend the ad, thus showing little interest in it. The appeal of your ad heavily influences the cost per click and the number of clicks. The average CTR is different for any given ad network and it can vary greatly for each specific product or service. 

You need to keep in mind that the CTR is never a fixed value, you have to determine it for yourself with testing.

How to maximize the performance of your creatives

It is quite difficult to predict the exact reaction to your creatives. That is why it is important to test your ads. Experiment with the components: titles, images, ad copy, or call-to-action buttons.

To collect enough data and make it significant, follow these tips:

  1. Pick a large audience for the test campaigns. This provides a sufficient sample for testing.
  2. Allocate enough money for the tests, otherwise, you will have to spend several weeks just testing.

In-built ad features on social networks allow us to test out hypotheses and determine the best creatives without employing a specialist in analytics.

It is important to prepare for testing at the stage of developing the creatives. This preparation helps you determine the elements that need to be tested for the success of your future campaigns.