How To Use Experiment Tool on Facebook to Create an A/B Test

Why are we using it?

Hello everyone, I am Aylin from the PPC department!

Firstly, I want to start by explaining why we are using it. Experts estimate that an average person sees around 10k ads in a day. Since there are a lot of ads, our ads need to be easily specified through others to get conversions. Because people don’t use social media or websites to see ads so basically, we try to get attention from someone who is not looking for an ad or deal. And that’s the reason why we are using A/B testing. With A/B testing, we determine the best version of our ads and publish them to get the best results possible.

If we briefly count the benefits of A/B testing we can say:

• It helps to find the version that we can get better results at a lower cost.
• We can explore and know our target audience.
• It gives us an idea about the content that our target audience is engaged in.
• We can find out which strategy will work, which strategy that we should stick with and which we can just leave. 

Examples of variables that we can test with the A/B test:

• Target audience
• Conversion event
• Creative
• Ad text
• Placement
• Product sets
• Budget

What is an A/B test?

A/B test is a way to measure how the changes that you think to do would affect your campaign. If you’re thinking about an ad text change or you think that your campaign would perform better if you change your audience, but you don’t want to take a risk because it already meets the goals you set for your campaign, then A/B testing is the best way to not taking the risk of changing the whole campaign but creating another version of it and just test it. 

How Does It Work?

When you create an A/B test, basically you create a test campaign based on the campaign you want to test. Then users start to see both of your campaigns, and you can see the results on the Facebook Ads Manager panel to see how it goes. Now, I’ll briefly explain how you can create one. 

Step 1:

When you are on the campaigns page on your ads manager, you can find it under the campaigns tab. 

But, you can also find it on all tools bar as “Experiments”

Step 2:

After you click the “Experiments” button, it’ll lead you to:

You need to click the home button to see this page. 

Step 3:

Click the “Get Started” button under the A/B test section.

Then it’ll lead you to this page:

Step 4:

After you land the page above, you need to decide on your test details. For example, what will you test? Are you going to test campaigns, ad sets, or campaign groups? After you decide on the details, you can schedule your test, give your test a name and determine the winning version of your test. A winning version of your test means that for example if you want a lower cost per result, you can select the key metric as cost per result to determine the winning version of your test. 

Step 5:

After you determine all your test details, at the end of the page you’ll see a section which says estimated test power. It gives you estimations about your test. If everything is okay, you can click the “Review Test Details” button to see your test details briefly.

Step 6:

After you check all your test details, you can publish your test. And if you want to see how it goes, you can click the “Tests” button below the “Home” button. 

 There is a video that explains it thoroughly. You can check our Youtube account and watch it to understand deeply.

How should you read it? w/examples

If you watched the video that I linked, you saw 2 examples, but I’ll show you other examples to make it clear for you. 

1. We run a DPA campaign for one of our clients in the beauty industry, and we want to test the audience because we thought that our ads don’t show up to the best audience possible. Then we tested 2 audiences. One of them was website visitors, and the other was cart abandoners. After we tested two different audiences for 1 week and then the results told us that we could get more conversions with less cost per result and less amount spent.

    

2. Here is another example with an audience test. We tested a new user audience by excluding some specific audiences to see if we reached the best audience possible by excluding specific audiences or not. In the end, test results showed us we need to exclude that specific audience to get the best results.

    

   Briefly, A/B testing is a great way to decide what you’re going to do and it provides you with useful learning about your campaigns and brands.  In some cases, it is basically a lifesaver. For example, if we take a look at the examples above we’ll realize in the 1st example the test results made us realize which audience is the best for our client’s DPA campaign. This is a great learning experience because now you know that you can use that audience because it’ll perform better.Thank you for reading, and I hope to meet you again to share our knowledge.