Advantage+ Audience without Suggestions and Algorithmic Remarketing

You may recall that a short while back I ran a test to see how much algorithmic remarketing happens when using Advantage+ Audience. After getting shocking results that didn’t line up with what I heard others were getting, I tried again. And the results have been much different (and encouraging!).

Meta says that if you don’t provide audience suggestions with Advantage+ Audience, its AI will learn from your prior conversions, pixel data, and interactions with previous ads to determine your audience.

Advantage+ Audience

Is this true? Does Meta prioritize remarketing if you don’t provide audience suggestions?

We can prove it. Let’s dive in…

The Setup

I’m running a manual Sales campaign…

Campaign Objective

…using Advantage+ Audience without audience suggestions.

Advantage+ Audience

This was the best way to test because manual Sales campaigns now allow for the use of audience segments for reporting.

Audience Segments Sales Campaigns

I defined audience segments for Engaged Audience and Existing Customers in my ad account settings.

Audience Segments

Distribution Results

And here’s how the distribution of my budget looks so far.

Audience Segments
  • Total Spend (to date): $272.37
  • Engaged Audience: $84.46
  • Existing Customers: $12.07
  • New Audience: $175.84

That means that just over 35% of my budget was spent on remarketing.

What It Means

This is actually amazing. It’s one ad set and I haven’t provided any targeting inputs. And yet, Meta’s AI has automatically spent 35% of my budget on remarketing and 65% on prospecting.

It’s also interesting that the CPM is about twice as much for remarketing than for the new audience.

Audience Segments

Not surprising, but just some insight behind some of the data that may drive this distribution.

Overall, several factors will impact this:

  • Budget
  • Time elapsed (likely less remarketing over time)
  • Size of Engaged Audience and Existing Customers audience segments
  • Cost to get the desired action from remarketing vs. prospecting audience
  • Optimized event (?)

The last one certainly has an impact, but how much isn’t clear. In some cases, it could be huge based on my very different results. The low distribution of algorithmic remarketing I saw was when using a custom event as my conversion event. This time, I used a registration.

If you are skeptical of Advantage+ Audience because you’re concerned the wrong people will see your ads, this is solid evidence that it’s doing its job.

This test is just getting started. I’m also testing Advantage+ Audience with audience suggestions (that mimic my audience segments) and custom audiences with Advantage Custom Audience turned on.

If this is something that you want to test, make sure to check out my blog post with background on how to set it up.

More to come!