Advantage+ vs manual placements

You duplicated a winning ad set, left placements on Advantage+, and a week later the spend is up but purchases are flat. You open the placement breakdown and there it is: 38% of the budget went to the Audience Network at a 0.6 ROAS, while Instagram Reels quietly ran 3.1. The forum advice is loud and split. Half say never trust automatic placements. Half say Meta knows best. Both are guessing, because neither has looked at your breakdown.

Short answer: Advantage+ placements usually win on cost when you optimize for a real conversion, because Meta rarely leans on the Audience Network for purchase events. Manual placements earn their keep only when your placement breakdown shows one network eating budget at a ROAS your feed placements would never tolerate. What the breakdown shows is what settles it.

The takeaways

• Advantage+ is the right default for conversion goals. For a Purchase or a qualified-lead event, Meta barely touches the Audience Network on its own, so the cheap-inventory fear mostly evaporates.
• The Audience Network burns budget when you optimize for clicks. Link Clicks or Landing Page Views tell Meta to buy the cheapest taps it can find, and the Audience Network is where they live.
• Rank placements by ROAS and revenue. A placement's impression share and click rate say nothing about whether it pays; one can hold a huge slice of both and still drag your account under.

Should you use Advantage+ or manual placements?

Start on Advantage+ if you optimize for a conversion event like Purchase or a downstream lead. Meta's delivery system spreads budget across Facebook, Instagram, Messenger and the Audience Network, and for conversion goals it rarely spends much on the low-quality inventory, because that inventory does not produce the event you asked for. Meta's own guidance says Advantage+ placements tend to come in cheaper per result, and for most conversion campaigns that holds up.

The catch is that "cheaper per result" assumes the result you optimize for is the one you actually care about. If your optimization event is loose (a landing page view, a link click), Advantage+ will happily buy a flood of cheap, shallow actions. So the setting is not the real decision. The optimization goal underneath it is.

Why does Advantage+ dump budget on the Audience Network?

Because you told it to chase cheap actions. Jon Loomer has made this point for years: when you optimize for Link Clicks or Landing Page Views, Meta funnels a large share of budget to the Audience Network, where taps are cheap and often accidental. Optimize for ThruPlay and the same thing happens with video, because that placement rewards passive watch time.

Flip to a conversion objective and the pattern usually reverses. When you optimize for Purchase, the Audience Network barely gets used, if at all, because it rarely delivers buyers. This is why "always turn off the Audience Network" is folklore, not a rule. It is great advice for a link-click campaign and pointless advice for a purchase campaign. The honest version is: match your worry to your optimization goal, then confirm it in the numbers.

How do you read the placement breakdown?

Open Ads Manager, apply a Breakdown by Placement, and put ROAS and cost per result next to spend. Now you can see where the money actually went and what it bought. You are looking for the placement that holds a big slice of spend at a ROAS well below your feed placements. That is the leak. A placement can carry a fat impression count and a flattering click rate and still be the one bleeding you, which is exactly why raw reach and CTR are the wrong columns to sort by.

One caution: give each placement enough conversions before you judge it. A placement with three purchases behind its ROAS is a coin flip. Read the breakdown weekly rather than hourly, and act on the placements that have earned a real sample.

When is manual placement worth the control?

Manual selection earns its effort in two cases. First, when the breakdown proves a placement is a persistent drain across several ad sets and refreshes, not a one-week wobble, cut it. That is a decision your own data earned. Second, when your creative is built for a single surface: a 9:16 hook made for Reels and Stories will underperform stretched into a right-column slot, so restricting placements protects the asset.

Outside those cases, manual placement mostly narrows the auction and raises your costs for no gain. The instinct to control everything feels safe, but it usually hands Meta a smaller pool to optimize inside. Let the system run wide, then trim from the breakdown only where the revenue tells you to.

Whatever you decide, the trap is judging a placement on whichever metric flatters it while the one that pays gets ignored. Reading the number that matters is the same discipline behind how Adscalr scores ads: a composite that weights real ROAS and revenue per install, so a lucky click rate on cheap inventory can't win the read. If you want the deeper diagnostic on one network specifically, the Audience Network audit walks through spotting invalid traffic before it drains a budget.