Meta stopped asking. Advantage+ is now the architecture.
Through 2025, Advantage+ Shopping Campaigns were a recommendation. In 2026, they are the default campaign structure Meta pushes every advertiser toward 1. The March 2026 updates consolidated budget controls, expanded existing customer budget caps inside ASC, added new performance impact estimates for Advantage+ Audience, and updated creative signals for Advantage+ Creative. All of it points in one direction: less manual control, more decisions made by the machine.
Meta's internal data says the tradeoff is worth it: a 22% increase in ROAS and up to 32% lower CPA in ecommerce verticals compared to manual setups 2. Those numbers are real. They are also incomplete.
What the 22% number does not tell you
The lift comes from Meta's broader audience matching and its ability to test creative combinations at a speed no human media buyer can match. When the system has strong creative to work with, it finds pockets of demand that manual targeting would have missed. That is genuine.
But the tradeoff is visibility. Inside Advantage+, you cannot see exactly which audiences are performing, cannot exclude with the granularity you once had, and cannot run the kind of structured tests that tell you why something worked. The system tells you what converted. It does not tell you who saw the ad, where they came from, or which audience segment drove the result. For operators who need to understand their funnel, not just see a ROAS number, that is a real loss.
The other gap is attribution. Meta's new attribution reporting now splits conversions into click-through and "engage-through" categories. Engage-through is Meta's term for people who saw your ad, did not click, and converted later. The number is useful directionally but difficult to validate against third-party data. Operators who trust Meta's self-reported numbers without cross-referencing will overcount.
Creative is now the only lever you control
An AppsFlyer report found that 70 to 80% of Meta ad performance is now driven by creative quality rather than budget or targeting configuration 3. That number has been climbing for three years, and the Advantage+ consolidation accelerates it.
This is the practical consequence of the firm's thesis: the platform has automated the targeting layer, so the only variable left in your hands is what the ad looks like and where the click goes. Agencies that still sell audience segmentation expertise on Meta are selling the layer Meta has taken away.
The work that moves the number now is producing creative at the cadence the platform rewards, testing hooks and offers against each other systematically, and iterating on the landing page the click arrives at. If your Advantage+ campaign is running three static images from six months ago, the automation will optimize delivery of mediocre creative to a broad audience. That is not a performance strategy. That is efficient waste.
What to do with this
Accept the architecture. Fighting Advantage+ by trying to reconstruct manual campaigns with granular audience exclusions is a losing game; Meta is deprecating those controls one update at a time.
Instead, move your hours to creative production and testing. Run enough creative variants that the system has material to test against. Meta's own guidance says ASC needs 7 to 14 days to exit the learning phase and another 7 to 14 days to optimize delivery, so a minimum evaluation period is 21 days per test cycle 2.
Use the existing customer budget cap to protect prospecting spend. This is the single most important control inside ASC; without it, the system will retarget warm audiences because they convert cheaply, and your prospecting pipeline dries up.
Cross-reference Meta's reported conversions with your own backend data. Engage-through attribution inflates reported ROAS for most accounts. If you are making budget decisions on Meta's self-reported numbers alone, you are probably overallocating.
The discomfort is the point
Losing control feels bad. But the question is whether the control you lost was producing results or producing comfort. For most accounts we audit, the answer is comfort. The elaborate audience structures and manual exclusions were not outperforming Meta's automation; they were giving the operator the feeling of precision without the math to back it up. Advantage+ strips that comfort away and replaces it with a simpler question: is your creative and your landing page strong enough to win when the targeting is out of your hands?