Facebook ad placements: Advantage+ or manual?

The short answer on Facebook ad placements: in 2026 your default should be Advantage+ placements, and manual selection is the exception you switch on when you hold concrete evidence. Automated delivery moves the same budget across 19+ surfaces at auction time, so it typically finds cheaper impressions; a pinned list narrows the search space and slows the learning phase. The real gain is not in switching a surface off but in feeding each surface the right creative and reading the breakdown correctly.

What are Facebook ad placements and which surfaces do they cover?

Facebook ad placements are the screen positions where an ad can be delivered. They fall into four families: Facebook (feed, Marketplace, video feeds, right column, Reels), Instagram (feed, Explore, Reels, Stories, profile feed), Messenger (inbox, Stories) and Audience Network (native banner, rewarded video). The exact count shifts release to release; in practice you work with a list of 19 to 22 surfaces. Each one has its own aspect ratio, its own viewing behaviour and its own cost band.

Advantage+ placements or manual selection?

Advantage+ placements open your budget to every eligible surface and leave distribution to the auction. Manual placements mean you pin the list yourself. Industry benchmarks put automated delivery roughly 10-20% ahead on cost per result, especially on small and mid budgets, because a narrowed list has to buy the same outcome from more expensive inventory. We covered the campaign and audience side of the Advantage+ family in when to use Meta Advantage+; here we stay on the placement layer.

• Advantage+ placements: broad inventory, lower CPM, faster learning; in exchange you give up surface level control.
• Manual placements: full control and predictable formats; in exchange a narrower auction and typically a higher CPM.
• Hybrid: Advantage+ stays on and exactly one proven surface is excluded. In practice this is the balance that works most often.
• The false dilemma: 'should I turn off Audience Network?' changes the look of the report in most accounts, not the ROAS.

How do you read placement level performance correctly?

The placement breakdown is the report that splits one ad set's cost and conversions across surfaces. Reading it well takes three rules: a sufficient data threshold, a closed attribution window and the right order of metrics. In practice, waiting until a surface has accumulated at least 50 conversions or 1,000 clicks before you judge it visibly lowers the risk of reacting to noise.

1. In Ads Manager open the breakdown by 'Placement' and 'Device', and select the last 14-28 days so the attribution window has closed.
2. Look at spend share first, result count second. You cannot make a statistical call about a surface that takes 2% of spend.
3. Read cost together with CPM, CTR and conversion rate. A high CPA usually comes from creative that does not fit the surface, not from the surface itself.
4. Compare the same surface across two different creative families. Test the format before you blame the placement.
5. Write the decision down: at what threshold, on what data, for how long. Otherwise you will have the same argument again in three weeks.

How should you adapt creative to each placement?

Most of the placement debate is really a creative debate. A surface underperforms mostly because the asset entered it with the wrong aspect ratio: a 1:1 image gets cropped at the edges in Stories, text overflows and the first second is lost. Platform documentation points to a minimum set of two variants: 4:5 for feed, 9:16 for Stories and Reels. The full size table lives in ad creative sizes 2026.

• 4:5 vertical feed: the ratio that occupies the most screen space in Facebook and Instagram feeds; in most accounts this variant carries the core of impressions.
• 9:16 full screen: for Reels, Stories and Explore. Respect the safe area, roughly the bottom 20% overlaps with interface elements.
• 1:1 square: a safe fallback for mixed surfaces such as Marketplace, right column and Audience Network.
• Placement asset customization: inside the same ad you can define a different crop, a different text layer and a different destination URL per surface. It is the most direct way to fix a problem without closing the surface.

When should you exclude a placement?

Exclusion is a narrow tool for cases where evidence exists. In practice three reasons hold up: brand safety (you cannot audit the publisher list on Audience Network end to end), a format constraint (vertical surfaces are wasted if you only produce horizontal video) and regulatory limits. 'Its CPA looked high' is not a reason on its own; rule out creative and the data threshold first.

1. Threshold check: has the surface accumulated at least 50 results or 1,000 clicks?
2. Creative check: was the correct aspect ratio actually live on that surface, or was automatic cropping doing the work?
3. Contribution check: a surface can support view through conversions even when it wins no last click. Price the reach loss before you close it.
4. Decision: exclude it and clear the 5% box, then watch CPM and total result volume for 14 days.
5. Rollback rule: if blended CPM rises more than 10% or result volume drops, put the placement back.

Turning a placement off does not make inventory cheaper; it only shrinks the set of options the algorithm can choose from.Ads Sensor Team

How often should you revisit placement decisions?

Placement settings are not set and forget, but weekly tinkering also does damage. The cadence that works in practice: open the breakdown when creative is refreshed and once a month, then question the exclusion list from scratch once a quarter. When Meta adds new surfaces, old exclusion lists quietly narrow your reach, and in 2026 that is the most common source of hidden loss. If you cannot remember why a setting is there, it probably no longer applies.

Tracking this by hand becomes impossible once you run several accounts. Ads Sensor merges your Meta Ads and Google Ads accounts into one panel, turns campaign and ad set level deviations into prioritised actions with the reasoning attached, applies the recommendations you approve through the platform APIs and then keeps the before and after tracking itself. The early beta sign up form is here.