Yield & pricing

Google Blinks on PMax: Advertisers Can Finally Opt Out of Your Inventory

APH Marketing Desk ·5 min read Share Print
In this piece
    Figure Since Performance Max launched in 2021, its bargain has been non-negotiable
    • SEARCH
    • YOUTUBE
    • GMAIL
    • MAPS
    • DISCOVER
    Marketing Desk

    Since Performance Max launched in 2021, its bargain has been non-negotiable: hand Google the budget and the black box decides where it goes — Search, YouTube, Gmail, Maps, Discover, and, crucially for publishers, the Google Display Network and Search Partners. Advertisers who wanted Google’s automation had to accept Google’s inventory, all of it. This week that bargain cracked.

    A limited pilot now gives PMax advertisers two checkboxes: exclude Google Display Network, exclude Search Partners. It is the first genuine concession on PMax’s bundling since the product launched, and the buy-side reaction tells you everything about how that forced inventory was regarded. Kyle Rovinski of Duncan Channon: “Search advertisers don’t like the Display Network. We don’t want to opt in.” David Dweck of Go Fish Digital goes further, calling GDN and Search Partners “sources of remnant inventory that Google’s forced advertisers to opt into.”

    Sit with that second quote, because if any of your revenue arrives through AdSense or AdX from Google-side demand, he is talking about you. For five years, a slice of publisher open-auction revenue has been funded by advertisers who never chose that inventory — they chose PMax, and PMax chose you. The pilot is the beginning of the end of that arrangement.

    There is one complicating data point, and it cuts the other way: PMG reports a 10% average increase in PMax investment among clients given the new controls. Advertisers who can exclude what they distrust spend more on what remains. The pool grows even as its distribution narrows — which is either good news or very bad news depending on which side of the narrowing you land.

    01The demand you never negotiated

    The mechanism worth understanding is how PMax demand has been reaching publishers. An advertiser sets a target, Google’s automation allocates across surfaces, and some fraction lands on GDN — that is, on publisher inventory monetised through AdSense and AdX. From the publisher’s side this arrives as ordinary Google demand: it fills, it pays something, and it requires no sales effort. From the buyer’s side, per Dweck, much of it was tolerated rather than wanted — remnant placement accepted as the price of the automation.

    That structure had a consequence publishers rarely priced: the demand was insensitive to merit. Forced spend flows to whatever the black box selects, not to sites buyers admire. It propped up revenue at publishers with mediocre ad experiences exactly as reliably as at good ones, because the buyer never saw the placement list until after the money moved. Opt-outs invert this. When advertisers can exclude GDN wholesale, the demand that continues to arrive is demand that chose to stay — and the sites that keep it will be the ones a buyer can defend including: clean ads.txt, honest inventory declarations, sane ad density, real page experience. The lazy floor under Google-mediated open-auction revenue is being replaced by a merit test, one pilot cohort at a time.

    The PMG figure deserves its caveat and its reading. It is a buyer-supplied number from one agency’s client base — directionally interesting, not a market forecast. But the direction is coherent: control increases trust, trust increases budget. If it generalises, total PMax spend rises while the set of publishers receiving it shrinks. Concentration, not contraction, is the likely shape of what’s coming.

    02Measure your exposure before the pilot generalises

    The operator’s job this month is quantification. In Google Ad Manager, pull earnings by demand channel and isolate what arrives through AdSense and AdX from Google-side demand you never directly negotiated — as distinct from your own direct deals, PMPs and non-Google open auction. That Google-mediated line is where forced PMax spend has been landing, and its size is your exposure to the opt-outs generalising.

    Then apply the discipline you would apply to any other partner. Most publishers have a concentration threshold — the share of revenue from a single demand source at which someone is required to present a mitigation plan. Google-mediated demand routinely escapes that discipline because it feels like infrastructure rather than a partner. It is a partner, it just changed its terms in the buyers’ favour, and it should now sit inside the same quarterly concentration review as everyone else.

    Google giving ground on PMax controls is being reported as a buy-side story, and for buyers it is a straightforward win.

    03Why this matters for publishers

    A floor under open-auction revenue is being removedForced PMax spend has quietly supported GDN-dependent revenue since 2021. As opt-outs spread, that support becomes conditional on being inventory buyers actively want — a test some publishers have not had to pass for years.
    Buyers just told you how they see GDN inventory"Remnant inventory that Google's forced advertisers to opt into" is the buy side's honest framing, on the record. Pitching against that perception is now part of the job for any publisher with meaningful AdSense/AdX reliance.
    The likely outcome is concentration, not collapsePMG's 10% figure — vendor-supplied, one client base, treat accordingly — suggests spend grows as distribution narrows. There will be winners inside the narrower distribution; the goal is to be one on merit.
    This is a pilot, which means there is still timeLimited availability today means quarters, not weeks, before opt-outs are universal. That is exactly enough time to fix the site-quality issues that would get you excluded — and not enough time to waste.

    04What publishers should do

    05The bottom line

    Google giving ground on PMax controls is being reported as a buy-side story, and for buyers it is a straightforward win. For publishers it is a repricing notice. The forced-bundling era put a merit-blind floor under Google-mediated open-auction revenue; the opt-out era replaces it with a simple question buyers will now answer with a checkbox: would we choose this inventory? Publishers who can honestly answer yes — and who have measured their exposure, cleaned their supply declarations and built demand that isn’t Google-mediated — will likely end up inside a richer, narrower pool. Everyone else has been living on an opt-in advertisers never made, and just watched Google hand them the pen to unmake it.

    Sources & caveats

    Sources: Digiday, “Google quietly gives ground on PMax controls” (Sam Bradley, 24 July 2026; picked up across the trades on 27 July), for the pilot details, the 2021 framing, the Kyle Rovinski and David Dweck quotes, and the PMG figure. Note the caveats: the opt-outs are a limited pilot, not a general release; and the 10% PMax investment increase is PMG’s own figure from its client base — buyer-supplied and directional, not an audited or market-wide statistic.

    The weekly

    One letter a week, from the desk that runs the auctions.

    What actually moved in yield, CTV and curation across our publishers — written by the people who saw it, not a content team. No digests, no roundups, one email.

    One email a week. Unsubscribe in one click. We never share or sell the list.

    More from this issue

    Ran alongside this piece in the Weekly of 2 August 2026 — read the whole issue →