How to Tell If Your Google PMAX Campaign Is Actually Working in 2026

PMAX will spend your budget across every Google network it has access to. Whether that spending is working comes down to two things: whether your conversions are set up correctly, and whether you know how to read where the money is actually going. Channel performance is the report that shows you both.

This post covers how to read that report, what healthy spend distribution looks like, and what to do when it doesn't.

What Does PMAX Actually Optimize For?

https://www.youtube.com/watch?v=ktl7ZvH5Q24

Performance Max has one job: get conversions. That's it. It uses Google's AI to bid and place ads across Search, Shopping, YouTube, Display, Discovery, Gmail, and Maps — wherever it thinks it can find someone likely to convert.

The problem is that PMAX defines a conversion as whatever you've told it a conversion is. If your conversion setup is loose (tracking button clicks and page views alongside form submissions and phone calls, PMAX will optimize for all of it equally. It doesn't know which actions actually matter to your business. You have to tell it.

Think of it like a compass. If the compass isn't calibrated correctly, it doesn't matter how powerful the engine is; you’ll be heading in the wrong direction. Conversion setup is the calibration. Get it wrong, and PMAX will spend your budget efficiently toward the wrong outcomes.

Pretty much every account has some version of this problem. The fix starts with knowing what you actually want PMAX to go after.

The Two Things to Check Before Anything Else

1. Are your conversions set up correctly? Primary conversions should reflect actual lead actions — form submissions, phone calls, chat initiations. Soft actions like button clicks and page views belong in secondary, if at all. If PMAX is chasing the wrong signals, no amount of budget optimization will fix it.

2. Where is your money going? Open channel performance and check the spend distribution. Search and Shopping should dominate. If significant budget is going to Display or Discovery and you're not seeing results from those networks, your conversion setup may be pulling PMAX in the wrong direction.

These two checks won't tell you everything, but they'll tell you whether the campaign is fundamentally pointed in the right direction.

Full Build vs. Vehicle Ad: What's the Actual Difference?

A vehicle ad (VA) is a PMAX campaign that runs on the feed only. No headlines, no descriptions, no creative assets, just the inventory feed. It shows primarily on the shopping network, occasionally on display, but it has no access to search terms the way a full build does.

A full build PMAX campaign includes all the assets: headlines, long headlines, descriptions, images, videos, sitelinks. Because it has all the ingredients, it's eligible to serve across all of Google's networks, not just shopping.

While a VA is a type of PMAX, not every PMAX is a VA. When manufacturers or agencies talk about "running PMAX," they may mean very different things. Knowing which one you're running matters before you try to interpret channel performance data, because the two campaigns distribute spend very differently.

Primary vs. Secondary Conversions: How to Set Them Up and Why It Matters

Not all conversions are equal, and PMAX needs to know the difference.

Primary conversions are the hard actions you want Google to actively optimize toward, such as form submissions, click-to-calls, and chat initiations. These are the events you've decided represent a real lead. When you mark something as a primary conversion, you're telling PMAX: go get more of this.

Secondary conversions are softer actions you want to monitor but don't want Google chasing, such as a credit application button click, a VDP interaction, and a trade-in tool engagement. These go in the "all conversions" column rather than the "conversions" column, which means Google watches them but doesn't bid on them.

Secondary conversions serve a useful purpose beyond just keeping them out of the primary signal. They act as supporting data. If your form submissions are up but VDP interactions are flat, that's worth investigating. If both are moving in the same direction, your primary data is more reliable.

This setup also matters beyond PMAX. If your GA4 account is linked to Google Ads, the events you mark as conversions feed Google's smart bidding across all campaigns. Getting this right upstream makes everything else more reliable downstream.

What Is PMAX Channel Performance and Where Do You Find It?

Channel performance is inside your PMAX campaign under Insights. It's currently in beta, but it's usable and worth checking regularly.

The report shows spend distribution across all eight of Google's networks: Search, Shopping, Search Partners, YouTube, Display, Maps, Gmail, and Discover. For each network, it shows impressions, interactions, and results — where results are the conversion actions you've defined.

Two things to look for immediately:

Spend distribution. For most dealership campaigns, you want the majority of the budget (roughly 80–90%) to land on Search and Shopping. These are the highest-intent networks. Someone searching "Ford F-150 near me" is further down the purchase funnel than someone scrolling through a YouTube video. Search dominance in channel performance is generally a sign that your conversion setup is steering PMAX correctly.

Ads using product data vs. ads not using product data. The report distinguishes between placements that use your inventory feed and placements that use your uploaded creative assets. On the Display network, for example, PMAX may serve ads using your banners rather than your feed. This isn't necessarily a problem, but it's worth knowing, because it tells you whether your inventory is actually being shown or whether Google is falling back on generic creative.

How to Read PMAX Channel Performance Data

A maps-focused PMAX campaign is a useful case study because it shows how conversion setup and channel distribution connect.

If you want a PMAX campaign to skew toward Google Maps placements, targeting searches like "Jeep dealer near me" or "Mercedes service near me," you limit the conversions to maps-relevant actions: get directions, phone calls, store visits. By narrowing what PMAX optimizes for, you nudge it toward the network where those conversions occur.

In a campaign set up this way, you'd expect to see roughly 15–20% of spend going to Maps, the majority on Search, and minimal spend on YouTube and Display. If Maps spend is lower than expected, that's a signal that the conversion mix may need tightening.

One thing you'll also notice in maps-focused campaigns: branded search terms show up heavily. That's fine in this context. You want to be visible when someone searches your dealership name on Maps. The cost per click is low, and the intent is clear. The same branded terms would be a problem in a general conquest campaign, but campaign purpose matters when interpreting search term data.

Display and YouTube generating results on minimal spend is also worth noting. In the example from these campaigns, a $5 display spend attributed to a form submission and an $11 YouTube spend was attributed to conversions. These aren't dominant networks for dealerships, but they're not worthless either. Google knows when to serve an ad on a lower-intent network based on the conversion signals it's receiving. A small allocation to these networks, as part of a full build, appears to add value without requiring active management.

Why You Should Check Search Terms Alongside Channel Performance

Channel performance shows you where money went. Search terms show you what triggered it.

If 80% of your budget is going to Search, you need to know what searches are actually triggering your ads. Pull the search terms report alongside channel performance and look for two things: whether the terms are relevant to what you're trying to sell, and whether branded terms are appearing in campaigns where you don't want them.

Branded search terms showing up in a general new vehicle PMAX campaign is a common problem. It inflates conversion numbers because branded searches convert at a high rate, but those people were already looking for you. You're paying for traffic that would likely have found you anyway.

Search terms won't tell you everything. You can't filter channel performance by search term, and you can't isolate network spend by keyword directly from these reports. But the two reports together give you a much clearer picture than either one alone.

What PMAX Channel Performance Still Doesn't Show You

Channel performance is useful, but it has real gaps.

There's currently no filter in the channel distribution table for product data vs. non-product data. You can see the distinction in the summary view, but you can't sort or filter by it in the detailed breakdown. That would be a meaningful addition.

You also can't isolate network spend by asset group directly from this report. If you're running multiple asset groups targeting different inventory or audiences, you can't see how each one is distributing spend across networks.

To get around these gaps, use the Products tab inside PMAX to approximate shopping spend, and use Search Categories inside Insights to approximate search spend. Neither is precise, but together they help fill in what channel performance doesn't surface directly.

Is PMAX Still a Black Box in 2026?

Less so than it was, but it's not transparent.

The better description now is a frosted window. You can make out shapes, where the budget is going, which networks are producing results, and whether your conversions are firing correctly. You can't see the detail behind the glass. You can't see exactly which placements, which audiences, or which creative combinations are driving performance within each network.

Channel performance is the clearest view into PMAX that currently exists inside Google Ads. It doesn't answer every question, but it answers the two that matter most: are my conversions right, and is my money going where it should?

That's enough to make meaningful optimization decisions, which is more than was possible before this report existed.