This is something I’ve speculated for a while now. Do audience suggestions actually do anything?

I’ve now had two very clear examples of this…

Example #1

One example was a test I ran several months back of using Advantage+ Audience with and without suggestions. When I used them, I provided custom audiences that matched my audience segments.

When I compared results, I actually reached those audience segments more when not providing any suggestions at all.

Example #2

And now we have a new example, this time related to age range…

I recommend using Advantage+ Audience when optimizing for conversions, but especially purchases. If you optimize for leads, you need to monitor it closely to watch for problems. Well, I ran into such a problem recently.

When using Advantage+ Audience, you have the option of providing audience suggestions.

When running an ad set optimized for leads that didn’t utilize suggestions, I found using a breakdown by age that Meta spent 70% of my budget on people over 55.

I’m fine spending money on all age brackets, but I was finding these were low quality leads.

Since I wasn’t using audience suggestions, I decided to try again with a suggested age range of 25 to 44.

Keep in mind that you can’t set an audience control for maximum age when using Advantage+ Audience. As a result, Meta can still deliver beyond that range.

Well, that suggestion did virtually nothing. Meta still spent more than 78% of my budget on people outside of the suggested age range.

Why This Happens

In theory, it makes sense why this happens. Meta’s focus is on getting you as many of your goal action as possible within your budget. Audience controls can restrict the algorithm’s ability to do that. While that makes sense when talking about expanding an audience beyond a lookalike audience or detailed targeting, it can be problematic here.

I’ve talked before about how Meta needs to fix this with Advantage+ Audience. There are times when age or gender need to be an audience control or it becomes unusable.

One final point. I wouldn’t use this as a reason to avoid Advantage+ Audience when optimizing for leads in all cases. I even ran a split test a few months ago that showed Advantage+ Audience without suggestions generated more total and high quality leads than detailed targeting and lookalikes. It’s something you need to monitor.

What do you think?