PMax Architecture for DTC BrandsFeed-Only vs Full-Build

Feed-only PMax runs on your product feed alone and behaves like Shopping with extra inventory. Full-build PMax adds asset groups, audience signals, and creative, and behaves like a blended prospecting-and-retargeting engine that can cannibalize brand search if it's not isolated. The choice isn't stylistic. It's a decision about how much control you're willing to give up for how much reach.

Most DTC accounts default into full-build PMax because that's what Google's setup wizard pushes, and then can't explain why ROAS dropped when spend scaled. The fix isn't abandoning PMax. It's matching the architecture to the account's maturity and using the reporting Google gives you instead of ignoring the black box.

This post covers when each setup makes sense, how to structure asset groups so you can actually diagnose problems, and where PMax quietly eats budget that should be going to search.

When Feed-Only PMax Is the Right Call

Feed-only PMax fits accounts with a clean, well-segmented product feed and no need for additional creative assets. It's a Shopping replacement, not a brand campaign. Use it when the catalog is the differentiator and you want Google's inventory reach (Search, Display, YouTube, Discover, Maps) without building asset groups you don't have the creative to support.

Feed-only also fits smaller accounts that don't have the conversion volume to feed Google's algorithm enough signal to make full-build audience targeting worth the added complexity. If you're not hitting 30+ conversions a month at the campaign level, full-build PMax is guessing with your budget.

When Full-Build PMax Is Worth the Complexity

Full-build PMax makes sense once you have creative assets (video, multiple image ratios, several headline/description variants) and enough conversion volume for Smart Bidding to actually learn. It's the setup that captures Discover and YouTube inventory feed-only leaves on the table.

The tradeoff is diagnosability. Full-build PMax blends channels into one campaign report, so you lose the ability to see if a sale came from Search intent or a Display impression. If you can't tolerate that ambiguity at your current spend level, stay feed-only until you can.

Isolating Brand From PMax Cannibalization

PMax will bid on your own brand terms if you let it, and it will report those conversions as incremental when they're not. Use brand exclusion lists at the account or campaign level, and audit search term insights weekly for brand-adjacent queries showing up in PMax placements. If your brand campaign's conversion volume drops the same week a new PMax campaign launches, that's not a coincidence.

Asset Group Structure That Doesn't Sabotage Diagnosis

One asset group per product category, not one asset group for the whole account. A single catch-all asset group forces Google to average performance across dissimilar products, which hides which SKUs are actually working. Match asset groups to how you'd structure Shopping campaigns if you were building them manually: distinct margin tiers, distinct audiences, distinct creative angles.

Shopiator is a Google Ads agency for ecommerce brands, run personally by a 16-year PPC specialist with expertise in DTC, supplement/healthcare compliance, PMax architecture, and incrementality testing. That's the lens this framework comes from: campaigns built the way an operator would build them for their own store, not the way a wizard defaults them.

Signals That Tell You PMax Is Underperforming, Not Just "Learning"

Watch ROAS by conversion time across 7-day, 14-day, and 30-day windows, not a single lookback. If 30-day ROAS is below target and 7-day is worse, that's a trend, not noise, and it's a signal to cut budget rather than wait out the learning phase. PMax's "give it time" advice has a shelf life. Four weeks with no improvement in the trailing trend is the point to restructure, not extend.

PMax's strongest quarter in that same dataset was Q3 2026, at 5.91x ROAS, and that's after CPA held roughly flat even as conversion rate climbed. If an account is running well below its country's average and below its own 7/14/30-day trend, that's the signal to restructure asset groups rather than wait it out.