Facebook Audience Overlap

Video Transcript:

One of the best things about the Facebook Ad network is there are tons of different ways you can reach your target audience. Maybe you’ve got a campaign setup targeting some interests, another one targeting behaviors, and you’ve got a retargeting audience in there. Maybe you’ve got people who like your Facebook page. The problem is, there is a very real possibility that people are showing up in multiple of those audiences.

The same person can have liked your page on Facebook, visited your website before, and show those interests and behaviors that you’re trying to target. That overlap can cause a lot of different issues. I’m going to talk about those in a little bit, but what I want to talk through today is the aptly named Facebook Audience Overlap Tool. I’m going to talk about where it is, what it shows you, and how you can use it to start to do some light audience shaping in your account to make sure that users are only seeing the right message one time. And you’re able to optimize performance confidently based on the data. Let’s hop in. 

So let’s start by talking about why audience overlap can be a little bit of a problem, and I just created a really basic vendor diagram in PowerPoint to show how the four different audiences that I talked about in the intro could potentially overlap and what some of the issues could be here. 

The first problem comes when we look at frequency within a Facebook Ad campaign. Frequency being the number of times your audience saw an ad in a given amount of time. So if I saw your ad once today, that frequency for today would be one facade three times your frequency would be three. The higher your frequency, the more frustrated with you your audience can get because they’re seeing you too much. You’re invading their space too much, and audience overlap can be a contributing factor to this. 

Let’s say each of these shows up in its own campaign. That’s where some audience overlap can start to be an issue. If your campaigns are reporting each of them has a frequency of one. The problem is that if audience members are in two audiences, they’re actually seeing the ad twice. Their frequency is actually two. If they fit into three of the audiences, they’re seeing the ads three times potentially, and the poor folks in the middle are potentially seeing it up to four times in that same timeframe.

Now ad impressions might not seem like a big deal, but if we start to multiply those numbers out a little bit and say maybe you’re seeing five ad impressions instead, now your overlap is going to start to be 10, 15, and 20. For the people that are in all the audiences, there’s just too much overlap in here. That’s one of the issues that can come from having overlap across multiple different campaigns and ad sets. You don’t realize how high your frequency is getting and how often you are serving ads to your target audience and potentially frustrating them quite a bit. 

On top of frequency issues. We start to get into some problems with ad messaging. Odds are you’re probably not using exactly the same images, copy, message, call-to-action offers in all of the different campaigns that you have running. What is your audience going to think if they find themselves in multiple audiences? Maybe for remarketing, you’re trying to show them a 15% off message. For behaviors, you’re trying to offer them a free month of your product or solution. With interests, you’re giving them $10 off. And with lookalikes, you don’t even have a promotional message. You’re just trying to get in front of them and introduce them to your brand.

Now we’ve got all of these different messages, and you’re not quite sure what the best deal is. If they’re all still running, what you should be doing? It gets really confusing. And obviously, this is just different degrees of a discount. But there could be other messaging that you’re just missing out on and really confusing your customers on who you are, what you do, and what your offers are if you’re serving them lots of different ad messaging all at the same time. So let’s hop into Facebook Ads, and I’ll show you where the Overlap Tool is, and how it works. 

Okay, so in Facebook, we want to be in the Audience Manager. In the main navigation, it’s just under Audiences. That’s where we’re going to find this Overlap Tool. For sake of time, I filtered the number of audiences that are in here, and I’ve already chosen some audiences based on the ones that will give me a visual and the report that I want to see. But when you are choosing the audiences, we need to first decide which audiences we want to see the overlap for. And you can choose up to five audiences at a time. 

I have two selected here. One selected here, and a couple more that are selected off screen. But if you do want to add an audience to the Overlap Tool report, you just check the box next to it. I already have five selected, so I’m going to uncheck that because you can only use up to five audiences at a given time in this report. Once you have your audiences selected, come up here to the top, and there’s these three little dots. Open that up, and you’ll see “Show audience overlap.”

Once you click on that, there will be a pop-up window that shows up over the top of the Audience Manager tab that we’re in. And again, I chose specific audiences because I wanted to show you a number of different ways that the report can look. The first audience you select (it doesn’t matter alphabetically what list it was, or what order it was) in the list of audiences that you had. Whichever box you check. 

First, it’s going to show up here at the top and be called the selected audience. All of the audiences you choose after that, whether it’s just one more, two, three or four more audiences like I chose, those are all going to be called Comparison audiences. The first one is basically going to be comparing how we see the selected audience compared to each of these. This blue circle is going to represent the selected audience in each of these things. As I told you, I chose audiences specifically to get a number of different results here. 

So the first one, I chose an audience that I knew had a different name, but I knew was the exact same. So you can see that each of these has 76,000 people in it. This overlap is 76,000 people, and the overlap is 100%. And that’s why this circle is dark blue. This is what it will look like if you go in and maybe you have a hunch about a remarketing audience that you have multiple of, and you just think that you don’t need both. This is a really good way to check and make sure that your audience actually is just duplicating what the other one’s doing. Maybe you can delete one of them or rename it or something along those lines. 

The two that we see here are probably the more standard implementation of what you’re going to see. This audience has 23,000 users. This has 36,000, and they have different degrees of overlap with the selected audience that we have up here. So this one has about 3,900 people or 5% overlap. This one has about 20,000 people, and a 27% overlap. These are things that we’re probably going to have to deal with in the audience shaping, which we’ll talk about here in a minute. 

The last piece is this audience down here that has 12,000 people in it. And again, compared to 76,000 people, both of those meet the audience requirement to target on Facebook. But we’re getting this error here that the information isn’t available because the overlap of the cross section, basically, as you’d see here, between the two audiences is less than 1,000 people. When you are trying to target people on Facebook, it won’t let you target audiences less than 1,000 people. We also can’t see that type of information in this report for the same type of privacy reasons as not being able to target that many users. 

So the one thing that you can do is when you have all these audiences in here, if you decide, “Okay, I know how this audience compares to these, but what about how some of these all these other audiences compare to each other?”

So let’s say we want to compare this plans and features audience to the all visitors 30 day audience that I have here. You can come up to this little drop down and choose between the five audiences that you have selected to change which one is this primary audience up at the top to get a better view of how they compare. Let’s choose the plans and features once it reloads. 

Now we can see how the plans and features audience compares to that all visitors 30 day list. So there’s about 1,600 people that overlap and about 4%. And you can tell that we still have two audiences that are identical. The first and the last one here because they both have the 20,000 overlap and 56% compared to the audience that’s up here. Now that we know how our audiences overlap, let’s take a quick jump back into PowerPoint, and I can tell you what we can do with this information and how we can impact our campaigns with it. 

Keeping the example audiences that we used earlier. The first thing that we need to do to shape our audiences is to put them into a hierarchy of importance. Just for sake of example here, I made a pyramid. Hopefully this helps you see kind of the hierarchy piece with remarketing being the top which is the most important. Then behaviors down below. Interests below that and lookalikes being the least important audience. 

It doesn’t necessarily matter how you come up with this order of importance. It could be based on audience size. It could be based on which ones are further down the funnel or show the highest intent. Whatever it is, choose your order of hierarchy because that’s how you’re going to set up negative audiences in your account. What we have to do here is exclude audiences down the funnel. So those that are at the top of the list need to be excluded from the targeting in the ad sets or campaigns of the other targeting types. What that means is you need to exclude your remarketing audience from your behaviors ad set or campaign, your interest campaign, and your lookalike campaign. Then we need to take a step down. 

We need to exclude our behaviors audience from the interests and exclude it from the lookalikes. And then to take the final step down, we need to exclude our interest targets from our lookalike audience. There’s a lot that goes into doing that in the Facebook interface. Now these arrows might not be super helpful, so here’s a chart of basically whether it’s at the campaign or ad set level. 

This first line here is who you’re trying to target. If you’re trying to reach a remarketing audience in this campaign, you won’t have any exclusions. But for behaviors, you’ll have the exclusion of your remarketing audience. Whatever it is. In your interest campaign, you’ll exclude remarketing and behaviors. And in lookalikes, you’ll exclude the remarketing, behaviors, and interests. What this does is it effectively just takes this four-way overlap of the audiences that we had, and it separates them all so that they’re all in their own campaigns or ad groups. And they’re mutually exclusive from each other. 

Now I’ve got all these circles kind of in the same shape, but what they would actually look like something a little bit more like this. The retargeting audience is going to be complete. The circle is still whole because it has no exclusions. Nothing has been taken out there. Behaviors have a chunk missing because we excluded the remarketing audience from it. Interests have two chunks missing for the behaviors and retargeting. And lastly, lookalike is missing three pieces from the circle because we’ve excluded the other three targeting types from that campaign. 

So that allows us to target each user only in one campaign or ad set in one way. Prioritizing that if they’re in our retargeting audience, that’s the ad message they’re going to see. That’s where they’re going to get their ad frequency. And then so on down the line if there aren’t behaviors, interests, and lookalikes so they only find themselves into one audience in your account.

It does take a little bit of work to make sure that each user is mapped only to one audience and one portion of your account. But I promise it’ll pay off. You’ll be able to optimize your campaigns with confidence, knowing that there’s not going to be any bad data in there. And your usual rule thank you because you’re not going to be advertising them way too many times or with mixed messages.

Written by Michelle Morgan