Meta has recently released new recommendations for the ideal Facebook ads audience size, and it might surprise you!
This is good news because Meta usually isn’t clear about its recommendations, but this one is clear. I also think this represents a significant change in Facebook ad targeting, and we will need to adjust our approach as advertisers to accommodate this and get the best possible results. Meta is constantly changing, and you either keep up or fall behind. Let’s get started on these changes!
What Meta is Saying Now
The new meta recommendations are somewhat different from what I’ve recommended before. They are in the documentation, trying to convince advertisers to use broad targeting. This documentation shows why Meta recommends broad targeting—like achieving a lower cost per action. Of course, everyone who advertises wants to achieve a lower cost per action, so this is an exciting possibility. Facebook has been pushing broad targeting for a while now. Broad targeting means not including detailed targeting in your ad set—like interests, demographics, or lookalikes that narrow your audience. Instead, you use only a geographic targeting option and leave everything else open. It’s actually a very straightforward way to use the platform. Then, you will see on this page that Meta wants you to set your audience size to 2 million or more. This substantially increased from best practice even a few years ago.What I Used to Say About Audience Size
With very few exceptions, I used to say that you should always stay within an audience size of 500,000 people. However, I’ve been more concerned about having an audience that is too small rather than too large, so I’ve often targeted much larger audiences by default. However, Facebook is increasing the size recommendation now. 2 million is a much larger audience size than anything I’ve recommended as a minimum before. I think that reflects the revolution in machine learning that we are seeing now.Why Facebook Is Increasing Its Recommendation to 2 Million
Facebook’s machine learning has improved over the years. With a larger audience, you can help their delivery system advertise flexibly and work out who within that audience is the best and what time to advertise with fewer restrictions. You are putting fewer restrictions on Meta and allowing the machine learning the most freedom possible to get you the best possible results. This can help Facebook deliver a lower cost per action for your campaign over time. Larger audiences also typically result in more campaign stability. Another reason for this recommendation is that Meta isn’t included on its page, but you must also consider it. If all advertisers went to broad targeting, Meta would have far more ability to fill all the open spots in its ad inventory. Remember, ads are how Facebook makes its money, and the more ad slots that are filled, the more money the platform makes. When advertisers are more specific with their placement, Meta can’t fill all those ad spots, making less money. As an advertiser, you must always consider the motivations of the platform you are working on. Sometimes, Meta makes better suggestions for Meta, and you have to watch for that. Still, in this case, the suggestion for larger audiences is better for both Meta and the advertisers.Another Change Facebook Has Made to Audience Size Reporting
There is another change that Facebook has made to audience size reporting that you should also be aware of. Facebook used to estimate the audience size based on everyone on the platform. They estimate audience size based on who has seen an ad in the last 30 days. I like this measurement better because it gives you a much more accurate view of the audience size for your specific ad. After all, plenty of inactive people are on Facebook, and you can’t sell them anything if they aren’t looking at your ad.Make Sure Your Audience Makes Sense of Your Business
What I Recommend Going Forward
We’ve tested broad audiences a lot in my agency, and we’ve seen some very good results. However, not all businesses and ad accounts are the same, so I recommend broadening your audience where possible and testing new approaches from now on. To illustrate this, I recently surveyed people about the best targeting option for their business, and the results were interesting. Open targeting (broad targeting) won most of the time, but not by 90% or even close to that. The breakdown was:- 45% Open
- 34% Detailed
- 21% Lookalike