This is something I’ve speculated on for a while. Are audience suggestions really useful?
I have two very clear examples now…
Example #1
One example is a test I ran a few months ago using Advantage+ Audience, with or without recommendations. When I use them, I provide custom audiences that match my audience.
When I compared the results, I actually reached these audiences more without offering any advice.
Example #2
Now we have a new example, this time related to age ranges…
I recommend using Advantage+ Audience when optimizing for conversions (especially purchases). If you’re optimizing for leads, you’ll need to monitor closely to spot problems. Well, I ran into this problem recently.
When using Advantage+ Audience, you can choose to provide audience suggestions.
When running ad sets optimized for prospects who didn’t take advantage of the recommendations, I found that broken down by age, Meta spent 70% of its budget on those over 55.
I’d happily spend money on people of all ages, but I find these are low quality leads.
Since I didn’t use audience suggestions, I decided to try it again with a suggested age range of 25 to 44.
Keep in mind that when using Advantage+ Audience, you cannot set a maximum age audience control. Therefore, Meta can still provide services beyond that scope.
Well, this advice doesn’t actually do anything. Meta still spends over 78% of its budget on people outside the recommended age range.
Why does this happen
In theory, it makes sense for this to happen. The focus of Meta is to get you as many targeted actions as possible within your budget. Audience controls can limit the capabilities of algorithms. While this makes sense when talking about expanding your audience beyond lookalike audiences or detailed targeting, it can be problematic here.
I’ve discussed before how Meta needs to solve this problem with Advantage+ Audience. Sometimes age or gender needs to be an audience control or it won’t work.
One final point. I wouldn’t use this as a reason to avoid Advantage+ Audience when optimizing leads in all situations. I even ran a comparison test a few months ago that showed that Advantage+ Audience without suggestions generated more overall and high-quality leads than detailed targeting and lookalike audiences. This is something you need to monitor.
What do you think?
