Performance Max was pitched as the campaign type you set up once and let run. Five years into its existence, that pitch has aged badly. PMax only performs when someone is actively steering it, and left alone it burns budget on exactly the impressions that don’t convert. The setup producing the most consistent results for DTC and ecommerce accounts right now is a Performance Max hybrid strategy — PMax running alongside Standard Shopping, each doing the job it’s actually good at, rather than “PMax instead of everything else.”

If your account is one PMax campaign covering the entire catalog against a ROAS target you set three months ago and haven’t touched since, the rest of this is worth reading closely.

What the data actually says about PMax on its own

A 2024 Optmyzr study across 24,702 Performance Max campaigns found that 82% of advertisers were already running PMax alongside other campaign types — and that PMax consistently underperformed those other campaigns whenever they competed for the same traffic. That’s a telling gap between how PMax is positioned and how it behaves inside a real account.

The reach is genuinely unmatched: Search, Shopping, YouTube, Display, Gmail, Discover, and Maps from a single campaign. But that reach has always come with a visibility and control tradeoff that frustrated ecommerce advertisers from launch. Google has closed some of that gap — campaign-level negative keywords rolled out in late 2024 and early 2025, channel performance reporting now shows which properties are actually driving conversions, and search theme inputs per asset group doubled from 25 to 50. The “PMax is a black box” complaint is harder to defend in 2026 than it was in 2022. It still isn’t a campaign you can walk away from.

Why running both campaign types beats running either alone

The logic behind the Performance Max hybrid strategy is simple: Standard Shopping gives you control and clean data visibility, PMax gives you reach and automated discovery. Combine them and each covers the other’s weak spot.

Google’s priority rules changed at the end of 2024 too, moving from automatic PMax prioritization to an ad-rank model — the highest ad rank now wins regardless of campaign type, which makes running both side by side more viable than before. In practice, Standard Shopping carries your core, known-intent traffic, while PMax handles full-funnel discovery across the channels Standard Shopping can’t touch.

The account structure that’s produced the strongest results for ecommerce clients looks like this:

• Standard Shopping campaigns covering top-revenue SKUs and product categories, running tROAS targets with manual bid-management levers
• A Performance Max campaign focused specifically on new customer acquisition, built around lookalike and in-market audience signals
• Brand exclusions applied inside PMax, so it doesn’t cannibalize branded search traffic your dedicated branded Search campaign should be handling
• Campaign-level negative keywords filtering out low-intent terms like “free,” “cheap,” and competitor brand names where the impression cost isn’t worth the cannibalization risk

The reason this structure works comes down to conversion density. Spread budget and conversions too thin across too many campaigns and the machine learning never gets enough signal to learn properly. The target is enough segmentation to be strategic — not so much that the algorithm starves. For a broader look at whether separating campaigns beats consolidating them entirely, see is Performance Max actually better than running separate campaigns.

The product feed is still the biggest lever most accounts ignore

Most advertisers optimizing PMax spend their time on campaign settings, when the bigger opportunity usually sits in the product feed. PMax pulls heavily from Merchant Center to serve Shopping placements, so feed quality directly caps what the algorithm has to work with. Weak titles, generic descriptions, and missing attributes produce weak output no matter how well the campaign itself is structured.

Product titles should mirror the terms buyers actually search, not internal naming conventions. Descriptions should describe what the product does, not repeat a marketing tagline or packaging copy — skip the jargon entirely.

Margin management belongs here too. Google’s algorithm has no inherent preference for your most profitable SKUs over ones that simply drive volume — it optimizes for conversions, not margin. That means actively excluding low-margin SKUs from PMax, or using product-level asset group segmentation to control where budget flows. For catalogs of any real size, this is ongoing maintenance, not a setup task you finish once.

Asset groups: where most accounts leave performance unclaimed

Thin asset groups are one of the most common underperformance patterns in PMax accounts. The algorithm assembles ads by combining headlines, descriptions, images, and video — limited or generic inputs produce a limited, generic output. A few adjustments consistently move results:

• Separate asset groups by product category or audience segment; a single asset group per campaign is rarely enough segmentation
• Include at least one video asset — Google’s algorithm favors campaigns that have video, and Asset Studio now generates video directly inside Google Ads using Imagen 4 and Veo 3, removing the production barrier that used to make this optional
• Lifestyle imagery showing the product in real use consistently beats plain product photography in upper-funnel placements like YouTube and Discover
• Headlines should cover functional benefits and emotional payoff, not just spec sheets

Channel context matters here too — a single creative rarely works everywhere. What lands on YouTube pre-roll doesn’t land the same way in a Gmail ad or a Discover placement. PMax will handle distribution on its own; the quality of what you feed it sets the ceiling on what it can do.

Audience signals guide the algorithm — they don’t restrict it

Audience signals are one of the most misunderstood parts of PMax. Most advertisers configure them once and move on without fully grasping what they do. Signals aren’t targeting — they’re guidance. You’re telling Google what a great customer looks like so it can find more of them; the algorithm isn’t confined to that audience, it’s using it as a starting point.

Build signals to give Google the best possible examples of your highest-value customers, not to narrow reach. Prioritize your customer match list of past purchasers first, layer in website visitors with meaningful engagement second, and fill out the rest with in-market audiences, which add breadth but are less precise on their own. Resist tightening the ROAS target too early — setting an aggressive target before the algorithm has enough data can cut total conversion volume by as much as 50%. Give the signals room to work before pulling any levers.

What the reporting is actually telling you

PMax reporting has improved meaningfully, but it still takes some interpretation. A few things worth checking regularly:

• Search Terms Report now lives at the campaign level instead of the asset group level, exposing more data — but search and Shopping traffic are blended, so a given term may reflect performance from both formats at once.
• Channel Performance Reporting flags structural problems — if most of your spend is landing in Display with almost nothing from Shopping, that’s a signal your feed or asset groups need attention.
• Asset Group Segmentation shows which creative combinations are actually driving conversion value, which makes it straightforward to lean into what’s working and cut what isn’t.

If you haven’t run an Uplift experiment yet, put it on the calendar. It tests the actual incremental contribution of PMax against everything else running in the account — the closest thing to a real answer on whether it’s working at all.

When PMax is the wrong call

Performance Max generally needs a minimum of 30 conversions in the trailing 30 days to optimize effectively. Below that, the algorithm doesn’t have enough signal and results stay inconsistent. If your account isn’t there yet, Standard Shopping with tROAS is the more predictable path — build conversion history first, then layer in PMax once the data density supports it.

Google’s own guidance recommends Maximize Conversion Value with a target ROAS for advertisers tracking values who want to drive as much value as possible, and for ecommerce specifically, revenue-first bidding tends to outperform pure conversion-volume optimization. Brands with niche catalogs where query-level visibility matters, or where creative control needs to stay tight, will still get more reliable and interpretable data from Standard Shopping. The hybrid approach only pays off when both campaigns are actively managed, not set and forgotten.

Frequently asked questions

Should every ecommerce account run PMax alongside Standard Shopping?

Only once the account has enough conversion volume to support both — generally at least 30 conversions in 30 days. Below that, Standard Shopping alone builds the conversion history PMax needs to optimize properly.

Does the product feed really matter more than campaign settings?

Yes, in most underperforming accounts. PMax pulls Shopping placements directly from Merchant Center data, so weak titles, descriptions, or missing attributes cap performance regardless of how the campaign itself is configured.

How tight should audience signals be?

Loose. Signals are a starting point for Google’s algorithm, not a hard boundary — build them around your highest-value customers and let the system expand from there rather than trying to constrain reach.

How do I know if PMax is actually adding incremental value?

Run a Google Ads Uplift experiment. It isolates PMax’s actual incremental contribution against the rest of the account, which is more reliable than comparing reported conversions across campaign types.

The bottom line

Nobody getting strong results from Performance Max in 2026 is treating it as automation that runs itself — that’s true of every campaign type we manage, not just PMax. The campaign amplifies whatever it’s given: solid strategy and clean inputs produce strong results, and weak feed data with no structure sends the budget straight to impressions that were never going to convert.