PMax Alpha Lets You Cut Partner Networks Without Showing What You'd Lose

Google launched an alpha test inside Performance Max that gives advertisers two independent checkboxes for the Search Partner Network and the Google Display Network, letting them exclude either, both, or neither at the campaign level. The feature was first surfaced by practitioners on LinkedIn in early May 2026, not by an official Google announcement, and it sits inside the campaign settings as a quiet "Alfa" flag. The catch: Google still does not show which of your existing PMax campaigns relied on partner-network delivery to hit their conversion targets, so the toggle ships without the audit trail to use it safely.

What the alpha actually unlocks

Until now, PMax has been a single envelope that distributes budget across Search, YouTube, Display, Discover, Gmail, Maps, and search partners with no direct control over any individual surface. The only way to influence delivery was through indirect tools, brand-safety lists, account-level placement exclusions, or the rather ugly "hold a feedback session with your Google rep" route that PPC Geeks documented as a temporary workaround last year.

The new alpha changes that, but only halfway. As PPC Land reported, the feature was spotted by Renaldo Spin and shared by Adriaan Dekker, and it presents two checkboxes at the campaign settings level. Search Partner Network is one toggle. Google Display Network is the other. YouTube, Gmail, Maps, and Discover are not part of the test, so this is not a full surface-by-surface controller. It is a partner-network override.

The "Alfa" label, as opposed to a standard beta, almost always means a small, hand-picked group of accounts. There is no opt-in form. There is no rollout date. If your account does not see the checkboxes, you wait. The PPC Land write-up is careful on this point: alpha tests at Google routinely die before reaching beta, and the only reliable signal of seriousness is when the toggles persist past the next product council review.

The 37% number that explains the demand

The reason this toggle matters is the same reason it took Google this long to ship it. Search Partner Network placements have been quietly underperforming Google Search proper for years, and the gap is not small. Intelligency Group's June 2025 analysis found Search Partner Network delivers a 37 percent lower return on ad spend versus Google Search, a delta also picked up by Search Engine Roundtable at the time.

That is a number every paid media director has been arguing with Google reps about since PMax bundled partner traffic by default. The standard rep response was usually some version of "trust the algorithm," and Google's own December 2025 marketing claim that advertisers with at least 5% Search Partner spend see an 11% lift in conversions, cited by ALM Corp, has always read as a defensive stat from the team that did not want this toggle to exist.

So now the toggle exists, kind of. The question working marketers actually need answered is the one Google still is not answering.

The audit trail Google left out

This is the part the alpha does not do. It does not tell you, before you flip the switch, which of your campaigns are sitting on conversions that came primarily through the Search Partner Network or through the Display Network. It just gives you the off-switch.

The visibility that did ship in February 2026 helps a little. Google added a Search Partners segment to PMax "When and where ads showed" reporting, as Search Engine Land covered, so you can now see which partner domains served impressions and how many. ALM Corp's analysis notes the same report shows network attribution per placement.

The thing it does not show is conversions per network. Impressions are not conversions, and the campaigns most at risk of a partner-network exclusion blowing up are the ones where partner traffic is contributing a disproportionate share of conversion volume relative to its impression share. From what I have seen in PMax accounts, these are usually mid-funnel-heavy lead-gen campaigns where the partner-network impressions look low quality on paper, but the algorithm is leaning on them for cheap, late-stage clicks.

This is the same data gap NMS covered last week when Microsoft Ads launched per-publisher conversion attribution for its own PMax-equivalent, a feature we walked through here. Microsoft is now telling advertisers exactly which URLs converted and at what cost. Google is telling advertisers which URLs got impressions and is still asking you to trust the bidding.

The 30-minute pre-flip audit

If your account is one of the alphas, or if the toggle ships to general availability later this year, here is the audit to run before you check either box.

Step 1, build the impression-share split. In PMax, go to the campaign, click "Insights and reports" then "When and where ads showed." Segment by the new Search Partners view. Export the last 30 days of impression data. You want a per-campaign percentage of impressions delivered through Search Partners. Anything over roughly 8% is the campaign you should not touch first.

Step 2, triangulate with conversion volume. Because Google does not give you per-network conversion data inside PMax, you need a proxy. Pull the campaign's overall conversion-per-impression rate before and after the February 10, 2026 parked-domain removal. ALM Corp documented that Google permanently removed parked domains from Search Partner inventory on that date. If your campaign's conversion rate barely moved across that boundary, partner traffic is probably not load-bearing. If it dropped 5% or more, the partner network is doing real work and an exclusion is going to hurt.

Step 3, isolate one campaign for the experiment. Pick one mid-tier PMax campaign, ideally a non-flagship spend tier, where the partner share is between 5% and 10%, and toggle the exclusion. Run for two weeks against a control campaign that did not flip. Look for the gap in CPA, not in raw conversions, because volume on the partner-excluded campaign will mechanically drop. The variable that matters is whether the conversions you keep are cheaper.

This is the workflow most agencies should already have built around the 70/20/10 budget allocation discipline NMS wrote up in our PMax spreadsheet guide, and it is the kind of test that does not work without the budget hygiene to set up a clean comparison group.

Why the alpha probably gets gated

I think this toggle ships to general availability in some form by Q3 2026, but with one important caveat. Google is unlikely to leave the Search Partner exclusion fully unrestricted, because the December 2025 marketing pitch about the 11% conversion lift would look extremely awkward against the 37% lower ROAS finding from Intelligency once advertisers actually compare both numbers in their own accounts.

The most likely outcome is that the toggle ships, but only for accounts above a certain spend threshold or only after an account audit, similar to how Google handled the rollout of cross-account negative keyword lists. Possibly with a forced 14-day notice period before the exclusion takes effect, which is the kind of friction Google introduced when it allowed video exclusions in the first place.

And here is where it gets uncomfortable. Working marketers have wanted this toggle for two years. Google's product team has resisted it because partner-network delivery quietly subsidizes the conversion-volume narrative they use to sell PMax against Microsoft and Meta. The alpha existing at all is a concession, not a generosity. Read it that way.

The signal worth watching next

Whether the alpha graduates, and whether it brings per-network conversion reporting with it, will tell you more about Google's posture on PMax transparency than any blog post. If the toggle ships without the conversion audit trail, that is Google saying the visibility was the concession and the toggle was the trade. If it ships with it, that is Google saying advertisers won the argument.

For now, my read is that the alpha is real, the audit trail is missing on purpose, and the safest thing to do if you get access is to run the 30-minute split and write down which campaigns you flipped and what happened. The data you generate is the only thing you will have when the next PMax change lands and your rep does not remember this conversation. From what I have seen of these alpha graduations, the accounts that document early get the loudest voice in the beta feedback loop, and the rest just inherit whatever defaults Google decides on.

Honestly, I would test the Search Partner toggle first and leave the Display one alone for a quarter. Display inside PMax is doing more work than the placement reports suggest, and the surface-by-surface conversion math is even murkier there.

Notice Me Senpai Editorial