If you’ve hung around my corner of the internet long enough, you know I typically write blog posts to save you from the same ad disasters I’ve already learned from. I’m talking to all the scrappy business owners and overworked marketers out there who are just trying to get their Facebook ads to actually make money (instead of wasting it).
Facebook, I have found, is an incredible way to reach fresh eyeballs and rekindle old flames in your customer base. Your content introduces your brand. It’s also how you present your offer, spotlight how your product solves someone’s headaches, and start building a relationship that (fingers crossed) ends with a purchase.
Wait, before I get into the details: I keep a running collection of free game here on my blog — think breakdowns of my best ad tricks, case studies that spill all the tea and step-by-step guides so you don’t have to piece it together alone. I also have a “Live and Learn” section where I put my own mistakes on blast so you can avoid them. I write it because I love you ❤️😘
So let’s talk about the dark side of Meta’s Advantage+ Shopping campaigns.
Meta’s been pushing these hard — they come loaded with things like Advantage+ placements, budgets, audiences, creative enhancements and more. In theory, the more of these settings you allow, the more Meta’s machine learning gets to mix and match your creative assets, targeting who they think will buy, and finding the cheapest paths to conversion without you having to run fifty split tests.
Marketers like me love this, mostly because it means less babysitting — no need to manually tweak budgets across a dozen ad sets or stay up all night obsessing over which ad set performed 0.2% better. In the old days, we’d slice campaigns into a bunch of tiny ad sets and shift budgets around like Wall Street interns. With Advantage+, we mostly let Meta’s AI figure it out.
But as with everything in life, here’s the catch: sometimes the algorithm makes decisions so mind-boggling, it feels like it’s actively trying to sabotage you. Be careful!
A real-world example (so you don’t think I’m just yelling at clouds)
I recently ran an Advantage+ Shopping campaign for a client. It had four different ads — three videos and one animated gif. Over three days, it chewed through nearly $300. Pretty standard. But here’s where it went off the rails:
The algorithm placed the bulk of the spend into one video ad. Sounds great, right? Except that ad soaked up more than a third of the budget and had zero add-to-carts and zero purchases. Meanwhile, the animated gif, which got the least spend, was the only one pulling in purchases.
This is where trusting AI becomes dangerously close to mis-using your budget. Sure, Meta’s learning systems are powerful. They’ll test your creatives a thousand ways. But sometimes it pushes spend to an ad because it’s getting the most engagement — likes, comments or for some other random reason. It doesn’t matter. None of that pays the bills if nobody’s adding to cart or checking out.
So yes, lean on AI to simplify your campaigns. It’ll save you time and probably outperform your old manual ad set splits. But please, watch your metrics like a hawk. Otherwise, you’ll end up funding a popular IG post instead of a conversion machine.
TL;DR: Always be ready to cut losers fast
This is why the role of a smart human (like yours truly 😉) isn’t going anywhere. Advantage+ makes the grunt work easier, but it still needs oversight. When I noticed that losing video hogging spend with nothing to show for it, I killed it fast, especially if your client doesn’t have piles of cash to “let it ride.” You need to protect that budget like it’s the last $20 in your bank account.
I’ll keep watching this campaign over the next few days to see if shifting spend to our proven ads boosts results. Meanwhile, if you’ve got a head-scratcher campaign, or if your own Advantage+ is, well, not so advantageous, shoot me an email!
Ready to optimize your campaigns without losing control? The key is finding the sweet spot between automation and oversight—letting AI handle the grunt work while you focus on the metrics that actually matter.
