Should you take advantage of Meta Advantage+ and let AI rule your campaigns or stick with manual set up? Find out which is best for your business.
When it comes to Facebook advertising, AI is nothing new. Meta has been using AI for some time now. But with Meta Advantage+, it’s putting more focus on AI-driven campaigns. Meta has been not only encouraging users to adopt these AI features into their advertising strategies but also making it extremely difficult not to use these features and to set up campaigns and ad sets the old-fashioned way.
So the question is, do Advantage+ settings work better than a manually created campaign by a marketing expert? What exactly is this “AI” that we’re being told is the next big thing? In this article, I will answer these questions and more.
The Meta Advantage+ suite is the core of all Facebook ads AI capabilities. Meta Advantage+ tools and features are AI-powered solutions designed to execute specific Facebook ad campaign strategies and tactics more efficiently. I wrote about the pros and cons of using Meta Advantage+ here.
The largest takeaway, from my perspective, is that you are handing over the keys to your paid advertising to a machine-learning tool with limited visibility into how that thing really works. The results my clients have experienced from these new AI-driven tools have been less than stellar. Allow me to elaborate in the next section.
💡 Could using Meta Advantage+ help or hurt your business’s Facebook ad account? Find out, and see how your current strategy is working with our free Facebook Ads Grader!
I have a client who drives a lot of business through Meta Ads and is always looking for ways in which we can expand and continue to scale. The option to begin incorporating Advantage+ campaign settings was an obvious next step as the entire platform is actively encouraging you to use these features.
Instead of using the typical Facebook ad targeting options such as lookalike audiences or manually targeted interests and behaviors, we figured perhaps AI could identify users that were just like our customers better than we (or these antiquated methods) could.
First of all, if you are creating a new campaign in Meta Ads and you select this option, you will not be able to edit specific placements.
However, if you do choose this option for a new campaign, you are saddled with all of the shiny Advantage+ features Meta has to offer. My client and I chose this option for the sake of experimentation and to determine if Advantage+ features were responsible for higher unqualified rates. The results? 130 unqualified leads within less than a week.
But, hey, $8.48 per junk lead is not bad in this economy!
In all seriousness, let’s break down this test. We launched this campaign under “recommended” settings and defined our general targeting parameters like we always do: age, gender, location, etc. Now with Advantage+ audiences, you provide inputs to the system that essentially gives the AI (machine learning) guidance on who to target. If you are familiar with Google’s Performance Max campaign type this works in a very similar way.
We did this test with one audience that contained interest and behavioral audience inputs:
This is the audience that we were able to acquire 130 unqualified leads from. Now we also tried one with custom audience inputs of a customer list. This resulted in another two leads which were unqualified and thrown out. For context, both of these Advantage+ audiences we constructed had a plethora of exclusions to assist the machine learning in identifying who we wanted to go after.
After these initial results, I decided that perhaps the inclusion of the audience network led to the influx of unqualified leads coming through. For the second attempt, I recreated the campaign and made the settings “manual” to remove that placement. From there, I attempted to construct the Advantage+ audience again. We ran this campaign for the remainder of the month, and it drove another 12 leads, two of which had been qualified as sales opportunities. This seemed like progress at first, until we discovered through conversations with the sales team that neither of the two leads showed up to the scheduled demo. Plus, another lead reached out directly saying that he was getting the client’s ads all throughout Facebook and Instagram but that he accidentally submitted and he was not their target audience.
So we gave Meta’s AI several inputs that should have allowed it to find the correct people that are a good fit for the product and services that we are selling. But it didn’t. Not only did it fail to construct an audience comparable to that of just a basic lookalike audience, it wasn’t even close. For whatever reason the “AI” seems to flood our system with fake and junk leads every time we attempt to use Advantage+ settings. But why?
❓ Not sure how to gauge your manual or Meta Advantage+ campaign performance? Compare your current metrics, like cost per lead, cost per click, and more to averages in your industry using our latest Facebook Ads Benchmarks report!
Meta’s Advantage+ is essentially a black box. We don’t know how or why it is making decisions, but we do know that it has the goal of driving leads (or clicks, whichever your intended goal is) at the highest volume for the lowest cost.
Now this isn’t entirely different than how automated bidding works in both Google and Meta, however, the difference is that you specified that audience and then the machine learning works within the audience that you created or specified managing the bidding aspect and which people in the audience to serve ads to and when.
With something like Advantage+, the machine learning is making the decision on who to target in all of Meta’s database of users and is using your suggestions to guide it but it doesn’t understand (the way a human would) the subtle nuances that define an audience and make them a qualified lead or customer.
My takeaway is that, on their own, the current state of these Meta Advantage+ products is not ready to be serviceable to businesses with complex marketing and sales funnels—unless the user can directly interact with it and give it more prompts like “these leads are unqualified, this is what would make them better, etc.”
As of right now, there is just too much complexity to businesses who have marketing funnels or nuanced products and services for Advantage+ to work as effectively as other methods.
What I have found Advantage+ to be very good at is targeting options for ecommerce businesses. This makes a lot of sense though—sales-oriented businesses are looking for people within a demographic that have particular buying behaviors. The AI can predict this much better than someone who is going to talk to sales after submitting a form. The buying process is much simpler and straightforward.
I do believe that AI products in marketing will reach a point where they will be the better alternative but it’s just not there yet. This is why it concerns me that companies like Meta and Google are essentially forcing you to use them by making it increasingly difficult to not use them. This could be to allow these systems to learn and improve, or it could be due to overconfidence in their effectiveness for everyone.
If their marketing product is the easiest to use and you can just let an AI take over, perhaps they believe more people will advertise.
One thing is for certain though, there is much less visibility into how everything is working even for the most experienced marketer using the platform. This means that Google and Meta (and others) can spend your advertising dollars for you and not really own up to how that is being spent. My advice, for now, is that if you’re a business with any complexity at all to how your marketing or sales funnel is built, switch back to the original audience options and thank me later.
You don’t have to navigate platform developments like Meta Advantage+ all alone! See how our solutions can help you use AI the right way to reach your marketing goals.
