How To Do Facebook Split Test

How To Use Facebook Split Testing To Drive Results

When creating a campaign on Facebook do you find yourself having questions regarding which is the best audience? Or what is the best goal for optimisation? If you should invest more in Instagram or Facebook? Well, the answer to all these questions is here: Facebook Split Testing (also known as A/B testing).

Testing different versions of your ad sets and ad can help you refine future campaigns. Using split testing, you can compare ad groups, audiences and ads.

Furthermore, you can analyse which ones have had the most results and implement them in your strategy for your next campaigns.

If you are confused about how best to use this feature, read on!

What is split testing on Facebook?

Facebook Split Testing

The A / B Test (or split test) is one of the most effective strategies of digital marketing. Many companies use this type of test to make important decisions for their business.

On Facebook, you can use the A / B Test to benchmark data, devices, audiences, and other types of results you need to make decisions.

It is ideal for use in any type of business and you will receive all the data at the end of the test in your email.

Imagine that you are about to launch a new product on the market, a fitness clothing line. However, you have doubts on which audience would be the most ideal to target with your ads.

The first audience group is made of women of a certain age who enjoy boxing. Additionally, the second group also includes women of the same age. But, the women in this audience enjoy to practice Jiu Jitsu

Both target audiences are well suited to buy your product. However, you want to be more assertive and reach the audience that will be the most interested in your new clothing line.

The solution to finding the best suited one? The A / B test.

You may find that only those who like boxing have enjoyed your products, and women who practice Jiu Jitsu have not even clicked on your ad. Best of all? You’ll have the right answer with real, consolidated data.

Facebook Split Testing

The first audience group is made of women of a certain age who enjoy boxing. Additionally, the second group also includes women of the same age. But, the women in this audience enjoy to practice Jiu Jitsu

Both target audiences are well suited to buy your product. However, you want to be more assertive and reach the audience that will be the most interested in your new clothing line.

The solution to finding the best suited one? The A / B test.

You may find that only those who like boxing have enjoyed your products, and women who practice Jiu Jitsu have not even clicked on your ad. Best of all? You’ll have the right answer with real, consolidated data.

Basically, by A / B testing you’ll know which is the best way to go, where you should invest your money, invest your team’s time and also where you will be most successful.

What you need to know to do Split Testing on Facebook

To start a split test on Facebook, you need to follow some rules and understand how it works.

Available Testing Variables

Each tested ad set has a distinct difference called variable. Facebook will duplicate the ads and will only change the variable you choose.

To achieve the most accurate A / B test results, you can test only one variable at a time. For example, if you test two different audiences against each other, you will not be able to simultaneously test two placement optimizations. This is because you would not know for sure which change is affecting performance.

The variables available for the Facebook Split Test are:

Facebook Split Test Variables

Audience Split Testing

The A/B test divides the audience into non-overlapping groups different on locations, gender, age, or group of interests.

Facebook distributes these variables randomly among different ad sets.

This random distribution helps to ensure that the test is performed fairly, as other factors will not distort the results of group comparison. It also ensures that each ad

Objectives available for Split Testing

Facebook A / B tests are compatible with the following business objectives:

facebook split testing objectives

If the goal of your ad campaign is any other that does not fit within the above, you will not be able to create the test.

Quantity of ads

Another essential rule to succeed is the number of ads you test.

When you are doing A / B testing, it is very important to know that the less ads you compare, the better, as the intention is to compare the result of one with the other.

That is why in the Facebook Split Test it is possible to test two or even three ads within a single campaign. This means that you can create an A / B Test or an A / B / C Test.

Testing time

Any test requires a minimum of time to be accomplished, right? Facebook also works like that.

What is recommended by Facebook is that your test campaign lasts for at least 3 days and at most 14 days. The reasons for this is that in less than 3 days, you may not have enough data to know which ad performed best.

Testing over 14 days is unnecessary because within that period you can already know which ad performed best. This way, you can apply all your available budget for your campaign to an ad that has shown better results.

How to Create Split Testing on Facebook

To start creating your A / B Test on Facebook, simply follow the step-by-step instructions below.

Step 1

The first step is to choose what the ad goal will be.

Create Facebook Split Test

Step 2

To make A / B test ads on Facebook you need to mark the switches that appears in the image below.

Then, create the name of your campaign and click “Continue.”

Facebook Create Split Test

Step 3

Now, you’ll start choosing all the settings for your campaign.

As usual, settings will change according to the ad goal you’ve chosen.

In this screen you can follow a summary of the split test on the right side of the page and choose the variable of your ad.

Facebook Split Testing Variables

This variable means what you’re really going to test, for example: if you want to know which audience is best for your ad, you should choose the audience variable.

Today Facebook has 4 variables: creative, delivery optimisation, audience, and placement. Each one has different settings.

Creative:

If you have questions about your ad’s creative, such as: which image to use or which ad format brings the most, you can select this variable.

Facebook Split Testing Ad Set
Delivery optimisation

You can use this type of variable to test different types of optimization in a campaign.

For example:

Let’s say you have a video that has achieved a lot of reach and engagement, both in the number of clicks and views. If you want to test if it’s better to be watched or to generate conversions, you can choose to test both optimisations: Video Views and Conversions.

Facebook Split Testing delivery

Audience

If you have different audiences for your business or are in doubt about the targeting you are considering investing in, testing two or even three different audiences can help you.

Facebook Split Testing Audience

Placement

In placement, you can customize the social networks and devices you want your ads to appear on.

For example:

You can choose to compare Facebook and Instagram. Or just Facebook on a tablet or computer and make the necessary combinations to understand where you can reach more people effectively.

facebook split testing placement

Step 4

Lastly, you will choose the budget and set the period that your campaign will stay on air.

Then, you can click continue, choose the images and text of your ad and finalize the test settings just like any other campaign!

What to do after Split Testing?

After you pass the testing period, Facebook will send you an e-mail with all the data from your test results.

Then, you can see which ads had:

With all that data in hand, you’ll know which ad strategy works or does not work for your business. You can then invest your time and money in ads that will really bring results to your business.

Get more than what Facebook allows you to test

We believe split testing on Facebook will certainly help you get closer to what you want for you and your business. It will guide you through the initial steps you need to take towards the desired RESULT.

However, this is just the beginning of the road. If you want to dig deeper into this topic and have access to a (more advanced) companion material, I recommend you have a look at Filed, and how we can overcome expectations through allying Artificial Intelligence with your Facebook data, and having complete control over each small to big element of your campaign.

Split testing with Filed can take you to another level. Do you want to know more? Request a 30 day Free Trial now!