The sudden rise in your Meta Advantage+ campaign costs isn't a sign that the algorithm broke. It's a sign that it's working exactly as designed, and you're feeling the violent snap-back of an optimization process that you've ceded too much control to. The fundamental problem isn't Meta; it's the common misconception that Advantage+ Shopping Campaigns are a "set it and forget it" solution. They are incredibly powerful, but they are amplifiers. They take your inputs, good or bad, and find the fastest path to the goal you've set, and a trailing seven-day performance window is a powerful input. When that window includes a sale, a viral creative, or any other short-term anomaly, the algorithm learns to bet big on those conditions continuing. When they stop, it keeps betting big, and your costs skyrocket.
The game has changed entirely in the last 18 months. As Taylor Holiday and Richard Gaffin have detailed on the Ecommerce Playbook podcast, the rollout of what they call "Andromeda" is the biggest shift in Facebook ads since iOS14. This new underlying model, which powers Advantage+, consolidates targeting and automates placements far beyond what we've seen before. It's Meta's push to regain performance by using its vast internal data instead of relying on compromised user-level tracking. The practical result is that our old levers are gone. We can't obsess over audience nuance or placement tweaks. Our primary points of leverage now are creative, offer, and the economic constraints we give the algorithm, like cost caps and ROAS targets.
Retaking Control of the Algorithm
The common advice to "just feed it more creative" is correct but incomplete. As Andrew Faris mentioned on The Bottom Line: Ecommerce Tactics for Profitable Growth, having ten times the amount of creative is a massive advantage. But creative without constraints is just speeding up your path to burning cash. The consensus is often wrong in treating Advantage+ like a magic black box that you simply trust. You must manage the machine, and your primary tool for this is cost controls. Taylor Holiday and the Ecommerce Playbook crew constantly come back to this. You have to be willing to set a cost cap, and if the algorithm can't hit your number on a given day, you accept that it didn't spend the full budget because it was an unprofitable bet. As he noted in one episode, if a campaign's AOV comes in lower than expected, you tighten the cost control to maintain efficiency.
This is especially true after a period of high performance. On The Bottom Line, Taylor explained exactly what happens after a sale like Black Friday. Meta's algorithm, looking at the last seven days of incredible conversion rates, has no idea your sale has ended. It will continue to spend with wild aggression on the Tuesday after Cyber Monday unless you manually pull back by lowering your cost caps. This is likely what happened to you. Your "good week" was your Black Friday. You celebrated the results without realizing you were creating a data cliff that the algorithm was about to drive off.
Another critical piece people miss is incrementality. The on-platform ROAS that Meta reports is not the ground truth of your business. The team on Ecommerce Playbook breaks this down brilliantly in their episodes on the Profit Allocation Model (PAM). They've found that Meta prospecting campaigns optimized for 7-day click tend to under-report their true, incremental impact by about 20%. This means you need to apply a 1.2x multiplier to understand its real value. If you're managing directly to Meta's numbers without accounting for this, you're leaving growth on the table or, worse, cutting campaigns that are actually more profitable than they appear.
A 90-Day Plan for Profitable Scale
If I were in your shoes, here is what I would do over the next 30, 60, and 90 days to not just fix this problem, but build a durable system for profitable growth with Meta Ads Strategy.
First 30 Days: Stabilize and Diagnose First, I'd immediately apply a cost cap or a minimum ROAS target to that Advantage+ campaign. I'd set it at a number I know is profitable for my business, even if it dramatically cuts volume in the short term. The goal is to stop the bleeding. Second, I'd start an aggressive creative development cycle. My goal would be to get at least 3-5 new creative assets live within the week to give the algorithm new things to test. Third, I'd do a quick and dirty analysis of my incrementality. I'd look at my total store revenue and total ad spend (MER, or Marketing Efficiency Ratio) and compare its trendline to my Meta 7-day click ROAS. This gives me a starting point for understanding how much I can "trust" the platform's numbers.
Next 60 Days: Systematize and Test With the account stabilized, I'd build a real system for management. Drawing from what Matt Axline and Peter Hassan described on Ecommerce Playbook, this means a daily check-in with a simple question: "Did I hit my efficiency target yesterday?" If yes, I might give the campaign a slightly looser cost cap to pursue more volume. If no, I hold the line or tighten it. I would also formalize my creative testing. I would avoid common mistakes like trapping budget in perpetual "testing campaigns," a problem Cody from The Bottom Line often points out. Instead, I'd use a single Advantage+ campaign and continuously feed it new ads, letting Meta's algorithm suppress losers and amplify winners, while I maintain discipline with my overall cost cap.
Next 90 Days: Scale and Refine Now, with a disciplined daily process and a steady stream of new creative, I can focus on scaling intelligently. I have a clear, incrementality-informed efficiency target. I trust my creative process. Scaling no longer means just cranking up the budget. It means strategically adjusting my cost caps based on daily performance and letting Meta spend whatever it can within those profitable constraints. I'd also start exploring Meta's own conversion lift studies, as Taylor Holiday mentioned in the episode on Meta's Andromeda: The Biggest Shift in Facebook Ads Since iOS14, to get an even clearer picture of my true incrementality. This is how you move from being a passenger in your own ad account to being the pilot, using Meta's powerful engine to go exactly where you want to go.