The Brands Are Fed Up With Black-Box Ad Platforms. Now What?

The Take: Brand and agency executives are done pretending platform-side reporting is the same thing as independent measurement. Google's Performance Max, Meta's Advantage+ and The Trade Desk all faced open revolt this year over what buyers can and can't see, and the businesses protecting themselves are the ones auditing automation rather than trusting it outright.

There is a rebellion happening in paid media, and it isn't loud. It's the kind where senior buyers quietly start moving budgets and stop signing off on renewals without asking harder questions first.

At the Digiday Programmatic Marketing Summit, brand and agency executives spent an entire town hall session naming names. The Trade Desk, Google's Performance Max and Meta's Advantage+ all came up, and none of them came out well.

Digiday's reporting from the event, published in May, captured the frustration in granular detail. The Trade Desk was already under pressure heading in. Three major holding companies, Dentsu, Publicis and WPP, had advised clients back in March to pull spend away from the platform's OpenPath product specifically, citing transparency concerns. One attendee described testing what the platform's automated bidding layer, Kokai, actually did to cost: "Kokai did the complete opposite" of what automation is supposed to deliver, with costs rising 10-15% on smaller campaigns once it was switched on.

But the frustration with The Trade Desk was, in one buyer's own words, "almost quaint" next to what they described happening inside Performance Max and Advantage+. On PMax specifically, one buyer said geography reporting had changed so that traffic significantly over 50% of the time no longer matched the geos explicitly set in campaign settings, unless exclusion audiences were added manually on top. On Advantage+, another buyer described a campaign where audience expansion settings switched on two weeks into a flight without notification, quietly reaching 15 million people the team never approved.

Because platform-side reporting tells a buyer what the algorithm believes happened. It doesn't tell them what actually happened, and those two things increasingly diverge the more automation sits on top of a campaign.

The structural issue sits underneath every one of these platforms. Each attribution model is built to credit itself generously for conversions that might have happened through another channel entirely, or would have happened regardless of any ad at all. That isn't a bug unique to one vendor. It's the default incentive wired into any closed system where the same party runs the auction and grades its own result.

Three things, roughly in order of how quickly they pay off. Get an independent benchmark running. Check settings on a fixed schedule instead of assuming nothing has quietly changed. Push for placement-level detail even when the platform doesn't offer it up front.

For an Australian business spending real budget on paid media, the practical risk is concrete rather than theoretical. Missing placement-level detail means creative decisions get made on partial information. Settings that reset quietly mean acquisition modelling breaks before anyone notices performance already has. Attribution that blends across channels means cost per acquisition becomes a number quoted with more confidence than it has earned.

Yes. Across New Rebellion's benchmark scoring of Australian and New Zealand businesses, Acquisition Performance sits noticeably ahead of Data & Tracking on average, and the split is widest in industries leaning hardest on paid channels without measurement infrastructure to match.

The typical scored business lands at 62.9 out of 100 on Acquisition Performance against 58.5 out of 100 on Data & Tracking, a gap of roughly 4.3 points across the full dataset. At industry level the widest split belongs to Beauty Salons & Hair at 20.7 points, followed closely by Veterinary Clinics, Fashion & Apparel and Cosmetic & Aesthetics, industries where paid acquisition consistently runs ahead of the measurement discipline needed to audit it properly. That pattern isn't coincidence. A business unable to independently verify what a platform reports has no real way to catch the exact failures buyers described at the Digiday summit before those failures have already cost money.

Methodology note: figures drawn from New Rebellion's benchmark dataset of Australian and New Zealand businesses scored across six marketing dimensions (Digital Maturity, Acquisition Performance, Conversion Efficiency, Retention & Loyalty, Brand & Positioning, Data & Tracking), scored as at May 2026, based on 729 businesses with completed scoring status. Read industry-level figures as directional given that sample sizes vary by industry. Full methodology at How We Score.

The platforms aren't handing control back voluntarily. Every response so far has been more automation, more AI, more assurance that the machine knows best. The buyers at the Digiday summit weren't asking for automation to switch off. They were asking for the receipts underneath it.

The businesses holding up best in paid media over the next year won't be the ones rejecting automation outright. They'll be the ones who never fully surrendered the ability to check its work.

If your paid media reporting has started feeling like it's asking for trust rather than offering proof, put it next to the rest of your marketing data. NR Studio benchmarks acquisition and measurement maturity together against real Australian businesses, so a gap like this shows up before it costs anything.

Why did Dentsu, Publicis and WPP pull spend from The Trade Desk?

The three holding companies advised clients in March 2026 to move budgets away from The Trade Desk's OpenPath product specifically, citing transparency concerns. The move triggered a broader industry conversation about platform accountability that carried through to Performance Max and Advantage+.

What's the core transparency problem with Performance Max?

Buyers report limited visibility into exactly which placements their budget lands on. At least one buyer at the Digiday summit described geography reporting diverging from actual delivery by more than 50% in some campaigns, requiring manual reconciliation against Google Analytics to catch it.

How can a business check whether Meta's Advantage+ has changed its own settings?

Review campaign settings on a fixed schedule, at minimum weekly, rather than assuming settings set at launch stay unchanged. Buyers in the Digiday reporting described audience expansion switching on mid-flight with no notification at all.

Does automation in paid media always reduce cost per acquisition?

Not reliably. One buyer at the Digiday summit reported costs rising 10-15% on smaller campaigns after The Trade Desk's Kokai automation layer was enabled, the opposite of what automated bidding is meant to deliver. Independent measurement remains the only reliable way to confirm which direction a platform's automation is actually moving the numbers.