Meta placement controls are being removed. The off switch is now a dimmer.
For ten years, the first thing a competent media buyer did with Meta placement controls in a new ad set was uncheck Audience Network.
That reflex is about to be useless.
On 20 August 2026, Jon Loomer published the notice advertisers were seeing inside the Placements section of their ad sets: “Excluding placements, platforms, devices and operating systems will no longer be available for your ad sets.” Not just individual placements. Whole platforms. Mobile versus desktop. iOS versus Android. Gone from the ad set.
By 25 August, advertisers were reporting the controls had already vanished, per PPC Land. Meta has published no announcement. Its own help pages still walk you through unchecking placements step by step.
Here’s the plain-words version of what Meta placement controls were, for anyone who has never had to care. A placement is the surface where your ad shows up: Facebook Feed, Instagram Reels, Stories, Messenger, in-stream video, or Audience Network (the third-party apps and sites that show Meta ads for a cut of the money). Placement controls let you switch any of those surfaces off. What replaces them is a value rule: an instruction that tells Meta to bid more or less when certain conditions are met.
And the ceiling on “less” is the whole story.
Why “minus 90 percent” isn’t “off”
Value rules let you adjust a bid from plus 1,000 percent down to minus 90 percent. Read that ceiling again. A 90 percent cut is still a bid. If the auction on a rewarded-video slot inside a free mobile game is cheap enough, and it usually is, your discounted bid still clears it.
Exclusion becomes suppression. The brand safety rule you built to keep your product off a gambling app is now a coupon for the gambling app.
There is one hard switch left: account-level Placement Controls, buried under Advertising Settings, then Account Controls. It works. But it applies to every campaign in the account at once. If you run one account for your brand, fine. If your agency runs several clients through one account, or you run a prospecting campaign that wants Reels and a retargeting campaign that doesn’t, it’s a sledgehammer where you used to have a scalpel.
Loomer’s read, which he’s careful to call unofficial, is that this is a test on the Sales and Leads objectives, that sensitive verticals are exempt, that you need the new multimedia ad creation workflow to see it, and that more placements are coming to value rules (there are only seven eligible right now). No date has been given for the full rollout.
That’s Meta’s pattern. Test quietly. Expand quietly. Document it after you’ve already lost the option.
Meta placement controls were always going. Here’s the timeline.
If you’ve watched the changelog, the direction has been visible for eighteen months. PPC Land laid out the sequence. Detailed targeting exclusions were removed on 21 January 2025. On 16 October 2025, Meta started defaulting up to 5 percent of your budget to placements you’d excluded, which established the principle that an exclusion is a preference, not a prohibition. The unified Advantage+ structure of 18 February 2026 banned campaign-level placement exclusions outright. On 29 July 2026, API version 26 removed Instagram Explore Feed and Messenger Stories as selectable surfaces.
Each step was small enough to shrug at. The cumulative effect is that a buyer who wanted a guarantee now has a multiplier.
Because every checkbox you touch is a constraint on the model. Because Meta believes its model allocates better than you do. Because, in aggregate, across a million advertisers of wildly varying skill, it probably does.
Meta’s number for why you should like this is that ad sets using Advantage+ placements delivered 11.7 percent lower cost per action than manual placements, on average. That figure is Meta’s, measured by Meta, on an outcome Meta defines, with no sample size, date range or vertical mix published. It compares automated allocation against manual settings chosen by everyone from a Fortune 500 media team to a bloke in Bendigo running his first campaign at midnight.
It’s not nothing. It’s also not an audit.
What makes this conspicuous is that the rest of the market is moving the other way. Amazon DSP added exclusion categories in December 2025. Microsoft Advertising expanded domain blocking to 10,000 entries per list in August 2025. Where rivals are handing buyers longer blocklists, Meta is converting blocklists into dials.
What it costs you, specifically
Let me separate the panic from the actual exposure, because most of the noise on this is coming from people who never needed the control in the first place.
Loomer’s own position, and I agree with it, is that the majority of advertisers removing placements were doing it out of habit rather than evidence. The cases where Audience Network wrecks results are almost always tied to cheap optimised actions: link clicks, landing page views, ThruPlays. If you optimise for traffic, Meta will find you thousands of accidental thumbs in third-party apps. If you optimise for ThruPlay, Audience Network Rewarded Video will inflate your view counts with people watching for virtual coins.
If you optimise for purchases with a conversion goal, the model has no incentive to send you junk placements, because junk placements don’t buy anything. Loomer says he’s seen it with leads occasionally. For purchases, almost never.
So for most AU DTC brands running Sales campaigns, the practical loss is smaller than the headline. Where it bites:
- Brand safety obligations. If a contract, a regulator or a category rule says your ad must never appear on certain inventory, “unlikely” is not a compliance answer. Alcohol, therapeutic goods, financial products, anything the ACMA or TGA has opinions about. The 90 percent floor is a discount, not a fence.
- Placement-specific creative. If you’ve ever excluded Stories because your 4:5 asset gets butchered there, you’ve been solving a creative problem with a media setting. That workaround is going.
- Reporting confusion. AdMake AI’s September changelog notes that the days placement controls vanished, 24 and 25 August, were the same days Reddit called an “official Meta outage”. Meta’s own status page logged nothing. The likelier story is a deployment that reset ad set learning and got read as a crash. Spend drifting into new placements looks exactly like creative fatigue from the campaign view.
And it’s landing at an ugly time. Triple Whale’s August refresh put the median Meta CPM at US$15.06 across 40,000 brands, up 13.24 percent for the year to July 2026. Meta’s own Q2 filing has price per ad up 12 percent, “partially offset by strong impression growth, particularly from lower-monetizing surfaces.” Translation: Meta flooded the auction with cheaper inventory and still raised the blended price. You are being handed more of that cheaper inventory whether you asked for it or not.
For Australian accounts there’s one more wrinkle. Meta says it removed 756,000 under-16 accounts between 1 December and 30 June under the social media ban. Any lookalike you seeded from Australian customer data before December was built partly on people who no longer exist in the pool. Different change, same lesson: the inputs you thought were fixed keep moving underneath you.
What you can do right now
This is the bit that matters before Q4 budgets lock. In order.
Audit what you’re actually excluding, and why. Open every active Sales and Leads ad set. Write down every exclusion. Next to each one, write the reason. If the reason is “we always do that”, delete the row. If the reason is a documented performance problem or a compliance requirement, keep it. Most brands I’ve looked at end up with a list of one or two real exclusions and a long tail of superstition.
Build the value rules now, while the checkboxes still exist. For each real exclusion, create the matching placement value rule at minus 90 percent. Do it before the option disappears, so you’re not rebuilding from memory in November. Ten rules per set, two criteria per rule, one set per ad set, per Meta’s help documentation.
Move hard requirements to the account level. If a placement must never run, use Advertising Settings, Account Controls, Placement Controls. Accept that it applies to everything in that account. If that’s a problem, the fix is account structure, not a workaround: separate ad accounts for separate compliance regimes.
Fix the creative problem you’ve been hiding with media settings. Every ad should ship as a 9:16, a 4:5 and a 1:1, with text inside the safe zones so Stories and Reels don’t crop your headline. Meta rewards this anyway. Its new creative diversity rating, live in Ads Manager by 26 August, scores each ad set Low, Medium or High. Add the column. Log it next to CPA. By Q4 you’ll know whether it predicts anything.
Watch the placement breakdown weekly. Not monthly. Weekly. Spend drifting into Audience Network and in-stream is the first sign the value rules aren’t holding. If CPM rose and results held, that’s Q4 arriving early. If CPM rose and results fell, pull the placement breakdown before you touch the creative.
Build the exclusion-only audience Meta shipped in August. It’s a customer list that can only exclude, positioned for opt-outs and do-not-target obligations. Put your staff on it. Put anyone who has asked you to stop on it. It’s the one control that got stronger this quarter.
And stop treating “we can’t exclude it” as “we can’t manage it”. You can still see it. You can still price it. You can still starve it. What you can’t do is pretend it doesn’t exist, which, if we’re honest, is what unchecking a box was always doing.
Meta took the wheel in 2022 with Advantage+. It’s now taking the mirrors. The brands that thrive under that arrangement will be the ones that stopped steering with settings years ago and started steering with creative, offer and measurement.
The rest will keep clicking around Ads Manager looking for a checkbox that isn’t coming back.
If you’d rather have someone who’s already rebuilt accounts for this walk through yours before Q4, book a growth call with Thrive. We’ll tell you which of your exclusions were ever doing anything.
FAQ
Has Meta removed placement controls for everyone? Not yet. As of early September 2026 it’s a test on Sales and Leads objectives for accounts with the new multimedia ad creation workflow, with sensitive verticals exempt, per Jon Loomer’s reporting. Meta has published no announcement or rollout date. Other objectives still have full placement control for now.
What replaces placement exclusions in Meta Ads? Placement value rules. You can raise a bid by up to 1,000 percent or cut it by up to 90 percent for seven eligible placements, plus device and mobile operating system. You cannot switch a placement off at the ad set level. Account-level Placement Controls, under Advertising Settings, remain the only true exclusion, and apply to every campaign in the account.
Will removing Audience Network exclusions hurt my results? For purchase-optimised campaigns, usually not much, because the model has no incentive to buy placements that don’t convert. The real risk is for traffic, ThruPlay and some lead objectives, where cheap low-quality actions inflate results. If you have a compliance reason to avoid a placement, a 90 percent bid cut is not a guarantee.
What should I do before the change reaches my account? Audit every exclusion and its reason, build matching value rules at minus 90 percent now, move hard compliance exclusions to account-level controls, ship creative in all three aspect ratios with safe-zone text, and check the placement breakdown weekly so spend drift isn’t mistaken for creative fatigue.




