paid media.you can no longer switch a placement off. only price it out.
Meta has finished retiring the manual placement selector. What remains are value rules with a hard floor of -90%. Brand safety stops being a list of unchecked boxes and becomes a production decision.
in three lines
- Since 25 August 2026, the manual placement selector no longer appears in ad sets. This is not an interface glitch: the rollout is complete.
- Value rules are the replacement. They move a placement's value between +1,000% and -90%. That -90% is a hard floor: suppression is available, exclusion is not.
- Meta justifies the change with an 11.7% lower cost per acquisition on automatic placements. That is vendor data on a vendor product, but the selector is not coming back.
what changed, with dates
On 25 August advertisers started opening their ad sets and finding no placement selector. It was not broken and it had not moved behind a new toggle: it was gone. Meta had been withdrawing it in stages since 20 August, and that week the withdrawal was completed.
This did not happen overnight. Instagram Explore Feed had already been removed as a placement at API level in version 26.0 of the Marketing API on 29 July. Messenger positions were emptied out, and on 27 August Messenger Stories stopped existing as an available placement. Each step read at the time like a minor adjustment. Together they are something else.
The figure Meta uses to justify the move is an 11.7% lower cost per acquisition on automatic placements. That is a vendor number about a vendor product and should be read as such. But the argument is not the point: the selector is gone, and that part is not up for debate.
exclusion is no longer a lever
Value rules are what replaced the selector. A rule shifts the value the system assigns to a placement across a range that runs from +1,000% to -90%. That -90% is a hard floor, and the whole story sits there: you can price a placement out until it is effectively unwinnable, and you can suppress it down to 10% of its value. Switch it off, you cannot.
For anyone buying media, that is a change of nature, not of interface. An exclusion was a guarantee: uncheck a placement and the ad did not appear there. A value rule is a probability. Minus 90% is not zero. It means rarely, almost never, unless the inventory gets cheap enough that the system decides it is worth entering anyway.
And brand safety had been solved with guarantees. The conversation with a client about where we do not want to show up ended in a list of unchecked boxes and a screenshot attached to the meeting notes. That screenshot can no longer be taken.
It is worth being clear about what has not changed, so the point does not get oversold: value rules remain a useful and reasonably precise tool. Seeing the real impression split by placement, pushing up the inventory that performs and penalising the inventory that does not, is all still available and still moves results. Exactly one thing was lost, and it is specific: certainty. In a sensitive category, the gap between almost never and never is not an optimisation nuance, it is a contract clause.
creative was the targeting, now it is the brand safety too
None of this is a product whim. Andromeda, the retrieval engine Meta consolidated through 2025, preselects which ads get to compete for each impression by reading the creative itself: image, video, copy. Every genuinely distinct asset gets its own entity ID. Interests and audiences carry less weight than they did three years ago; the creative carries more.
From the engine's point of view, a manual placement selector is a constraint that removes options without supplying information in return. What this withdrawal says is that Meta would rather read the signal in the creative than take a manual instruction that contradicts it.
The corollary is uncomfortable but workable: if the creative determines where the ad runs, then control over where it appears is exercised in production. A nine-second vertical with its hook in the first three goes to Reels and Stories because of its own shape, not because of a checkbox. A dense horizontal will not win in Reels no matter how hard we push it. The lever still exists; it changed departments, moving from the settings screen to the production table.
regulated categories and an exception worth not planning around
When the withdrawal started on 20 August, health and wellness and the special ad categories were left outside the change. That is a real, dated fact, and it is also a rollout-stage fact rather than a commitment. Building a regulated advertiser's plan on a deployment exception is betting that Meta forgets to finish what it started.
In those verticals the problem compounds. Meta already restricts Conversions API use in health and wellness, so the engine receives fewer conversion signals and compensates by demanding more creative diversity. Less signal below, less control above: the only variable left on the advertiser's side is how many distinct concepts ship each month.
That number already had a known floor. The threshold Andromeda rewards is 10 to 15 distinct creative concepts per month, where distinct means a different hook, a different person, a different format, a different narrative. Not a recolour, not a headline test. Anyone selling four assets a month was already below the threshold; now they have lost the exclusion that used to work as a safety net.
An exclusion was a guarantee. A value rule is a probability. Minus 90% is not zero, and that difference is the whole brand safety conversation from here on.
what changes in day-to-day operations
the brand safety brief now precedes the ad set
If control is exercised through creative, the decision about where we do and do not want to appear belongs in the production brief, not in the settings screen. It is a conversation with the creative team, not with the media buyer.
value rules are budget, not filters
They are useful for discouraging inventory that performs poorly or sits badly with the brand, and they should be documented for what they are: a bid modifier with a -90% floor and no guarantee of exclusion. Promising a client their ad will never appear somewhere is no longer something we can do.
the august baseline is worth money
Save the placement-level report for August before reading September. A better cost per result next month may simply be the removal of placements that were dragging the average down, not a campaign win. Without a baseline, that reading is impossible.
production is the lever that remains
Creative volume has stopped being an efficiency recommendation and become the one place where the advertiser still decides. Anyone who cannot sustain 10 to 15 concepts a month is no longer choosing their performance or their context.
questions we get
Can I still stop an ad from appearing in a specific placement?
Not with platform controls. The manual selector is gone and the value rule floor is -90%, which lowers the probability of winning that impression without taking it to zero. The closest thing to an exclusion left is a creative the engine does not consider a fit for that inventory.
So is there any point in setting a rule to -90%?
Yes, as statistical suppression. It sharply reduces exposure to a placement that performs poorly or that we would rather avoid, and the effect shows up in the impression split. What it does not give you is the binary guarantee the checkbox used to give, and that distinction should be put in writing with the client.
Is the 11.7% improvement trustworthy?
It is the figure Meta uses to justify the change, and it comes from Meta. No independent audit has been published. It is useful for understanding why the platform made the decision, not as a performance projection for a specific account.
where this comes from
- PPC Land, 27 August 2026
- PPC Land, 26 August 2026
- Jon Loomer Digital, 20 August 2026
- Common Thread Co, 27 August 2026
- Meta Marketing API v26.0, 29 July 2026
- Engineering at Meta, 2 December 2024
Sources are cited as text, with outlet and date. We don't link addresses we haven't verified.
and what it takes
The note explains the problem; the services explain the work.
the other notes
running on agents
Plenty gets said about agencies using AI, and very little about how one is actually run. This is the method from the inside: how knowledge is organised, where the human sits, and what breaks when nobody looks after it.
read the note → 05 · August 28, 2026reading year-over-year comparisons
Global CPM is up 32% since 2024 and US inventory is at an all-time high. A campaign can look worse in this year's report without having gotten worse for a single day.
read the note → 04 · August 28, 2026originality and reach
Instagram no longer rewards originality out of principle. It measures it. Reposting loses reach in ways you can verify, the penalty has moved beyond video, and the hashtag no longer makes up for anything. Here is what counts.
read the note → 03 · August 28, 2026measuring without e-commerce
Most real work doesn't end in a cart. It ends on a shelf, in a store, in a doctor's appointment. That can be measured. What can't be done is inventing the metric at month end.
read the note → 01 · August 28, 2026creative volume on meta
Meta split the decision of which ad you see into two stages. The first one reads the creative, not the targeting. That is where creative volume stopped being a production luxury.
read the note →