Google Ads Finally Lets You See Where PMax Sends Your Money
Google released Ads API v24.2 on June 24, 2026. It adds an ad_network_type segment to Performance Max placement reporting, so advertisers can finally see whether PMax spend landed on Search, Display, or partner networks instead of one blended number. The same release ships new fields that label AI-generated creative, the groundwork for EU AI Act disclosure rules that take effect August 2.
For years the standard complaint about Performance Max went like this: you hand Google a budget, it spreads the money across Search, Shopping, Display, YouTube, Gmail, and Discover, and the report you get back is one campaign-level row. You could see that it worked, or didn't, but not where. The v24.2 release notes quietly change that for anyone pulling data through the API.
The channel breakdown people have been asking for since PMax launched
The specific change is small on paper. You can now segment the performance_max_placement_view report by ad_network_type, which means Search, Display, and partner networks show up as separate lines. Search Engine Land's writeup framed it as deeper insight into where ads actually appear, which is accurate but undersells it a little. For anyone who builds their own reporting, this is the difference between guessing and knowing.
Here is why it matters in practice. A lot of PMax budget historically drifts toward Display and partner inventory because that inventory is cheap and abundant, and the algorithm will happily spend there to hit a volume target. That's fine when those placements convert. It's a quiet drain when they don't, and until now you mostly had to infer the problem from soft signals like rising impressions with flat conversions. With network-level segmentation you can pull the actual split. Some of this was already partly visible inside the Google Ads UI through the Performance Max channel report, but having it in the API is what lets agencies bake it into automated dashboards instead of screenshotting a Google Ads tab every Monday.
One caveat worth saying out loud: this is placement-level network reporting, not a full per-channel ROAS console with bidding controls attached. You still can't tell PMax to stop serving on Display. What you get is visibility, not a steering wheel. From what I've seen, visibility alone still changes how people argue with the platform, because now the argument comes with a number.
The AI labeling fields are the part nobody asked for but everybody gets
Bundled into the same release are two new structures, SyntheticContentInfo and SyntheticContentAttestation, attached to both Asset and Ad resources. In plain terms, these let you flag programmatically whether a piece of creative was AI-generated, and whether it was made fully automatically or reviewed by a human. The reference docs spell out both attestation dimensions.
The timing is not a coincidence. The EU AI Act's transparency provisions land August 2, 2026, and they require machine-readable disclosure for AI-generated content. Google is building the plumbing now so advertisers running creative through the API have somewhere to put that disclosure. There's a wrinkle: in v24.2 the advertiser attestation fields are read-only. You can see Google's attestation, but you can't write your own until v25 ships, which is scheduled for July. According to PPC Land's breakdown, there's also a grace period under the Digital Omnibus that pushes the hard machine-readable requirement for pre-existing systems to December 2, so most teams have a few months of runway before this becomes a compliance fire.
Whether you care about this depends entirely on how much AI creative you're already shipping. If your team is generating ad variants at scale, and a growing number are, this is the start of a labeling regime you'll be living inside by year end. We covered a version of this tension when Meta auto-enrolled REI into AI ads and produced a bike with two handlebars. The platforms want you generating creative automatically and disclosing that you did. Those two goals sit in slight tension, and the disclosure fields are how that tension gets formalized.
There's a real data point underneath the compliance language. The IAB's January 2026 disclosure framework, cited in the PPC Land coverage, found a 37-point sentiment gap between advertisers and consumers on AI-generated ads. Advertisers are comfortable with it. Consumers are noticeably less so. Labeling doesn't close that gap, it just makes it visible, and honestly I'm not sure most brands have thought through what happens when a disclosure badge sits next to their cleverest synthetic creative.
A testing feature that's easy to miss
Tucked into the same release is a new experiment type, COMPARE_CAMPAIGNS, that lets you run a campaign-mix test across as many as five arms. The practical use is finally being able to A/B Performance Max against a standard Search or Shopping campaign in a structured way, rather than eyeballing two campaigns and arguing about which one the algorithm starved. If you've ever wanted to prove to a skeptical client whether PMax is actually pulling its weight versus a manual campaign, this is the cleaner way to set that test up. It won't settle the debate on its own, but at least the comparison runs on the same traffic instead of two separate guesses.
The security change that should outlive the news cycle
The headline security feature in v24.2 is Multi-Party Approvals. Sensitive account actions, specifically user invitations and access-level changes, now require a second administrator to sign off. Search Engine Roundtable noted it was backported to older API versions too, which tells you Google considers it important enough to push everywhere rather than gate behind the newest release.
This one matters more than it looks. The most common way agency accounts get compromised isn't a clever exploit, it's a single admin account getting phished and an attacker quietly adding themselves as a user. A second-approver requirement breaks that chain. If you read our piece on the Gemini hijack numbers, you already know self-reported breaches were running well ahead of anything the platforms caught proactively. Multi-Party Approvals is the kind of unglamorous control that prevents the breach instead of explaining it afterward.
What to pull this week before you do anything else
If you manage Performance Max spend, the highest-value thing you can do in the next few days takes about fifteen minutes. Pull a performance_max_placement_view report segmented by ad_network_type for your last 30 days, and look at the share of spend and conversions going to Display and partner networks versus Search. If Display and partner are eating more than roughly 30 to 40 percent of spend while delivering a noticeably worse conversion rate, you've found budget that was hiding in plain sight. I'd treat that as a rough line, not a law, because the right split depends a lot on your vertical and how much of your demand is genuinely upper-funnel.
If you can't query the API directly, the channel performance report in the UI gets you most of the way there, just without the automation. Either way, the move is the same: get the split, then decide whether your asset groups and audience signals are actually pointing PMax where you want it, because the network mix is downstream of those inputs.
The bigger picture is that Google spent this release doing two things at once. It handed advertisers a reporting win they'd been requesting for years, and it started wiring in the disclosure machinery that regulators are about to require. The reporting is the part you'll use tomorrow. The disclosure fields are the part you'll be glad you understood when the August deadline stops being abstract. I don't think the teams who benefit most here are the ones with the biggest API teams. It's probably the ones who actually open the placement report and act on what it shows them, which is a smaller group than it should be.
Notice Me Senpai Editorial
