Performance Max campaigns now account for over 80% of Google Shopping spend for most ecommerce advertisers. Google has been steadily funneling more inventory, more automation, and more control into PMax — and the brands that understand how to work with (not against) the algorithm are seeing 30–50% better ROAS than those running PMax on autopilot.

We manage PMax campaigns across dozens of ecommerce accounts. The difference between a well-structured PMax campaign and a default setup is often the difference between 3x and 7x return on ad spend. Here's what actually moves the needle.

How Performance Max Works in 2026

Performance Max is Google's AI-driven campaign type that serves ads across every Google property — Search, Shopping, Display, YouTube, Gmail, Discover, and Maps. You provide assets (images, headlines, descriptions, videos) and a product feed, and Google's algorithm decides what to show, where to show it, and to whom.

The "max" in Performance Max refers to maximum reach across Google's inventory. Unlike traditional Shopping campaigns where you control placements, PMax distributes your budget wherever the algorithm predicts conversions will happen. This sounds like a loss of control — and it partially is — but the algorithm's predictive capabilities have improved significantly since PMax launched.

For ecommerce brands, PMax is particularly powerful because it combines Shopping ads (driven by your product feed) with prospecting ads (driven by your creative assets) in a single campaign. A customer might see your Shopping listing on Google, then get retargeted with a YouTube video, then see a Display ad — all from one campaign.

Campaign Structure That Drives Results

The biggest strategic decision in PMax is campaign segmentation. You have three main approaches:

Single campaign (basic): One PMax campaign with all products. Simple to manage, but gives you zero control over budget allocation between product categories.

Category-segmented campaigns: Separate PMax campaigns for each major product category (e.g., one for shoes, one for accessories, one for apparel). This lets you set different budgets and ROAS targets per category.

Profit-tier campaigns: Group products by margin tier rather than category. High-margin products get one campaign with aggressive targets; low-margin products get another with conservative targets. This approach optimizes for profit, not just revenue.

We recommend profit-tier segmentation for most ecommerce brands doing over $500K in annual Google ad spend. The revenue difference between optimizing for ROAS vs. optimizing for profit margin can be 15–25% on the bottom line.

For smaller accounts (under $5K/month in spend), start with a single campaign. PMax needs conversion data to optimize, and splitting too thin starves the algorithm. Google recommends at least 30 conversions per campaign in a 30-day window for optimal learning.

Asset Groups: The Foundation of PMax

Asset groups are PMax's equivalent of ad groups. Each asset group contains a set of creative assets and an audience signal that tells Google who you're targeting.

What makes a strong asset group:

Audience signals are suggestions, not restrictions. They tell Google's algorithm where to start looking for customers, but PMax will expand beyond your signals if it finds converting audiences elsewhere. Use them to accelerate the learning phase:

Product Feed Optimization

For ecommerce PMax, your product feed is everything. Shopping placements typically drive 60–80% of PMax conversions, and Shopping ad performance is directly tied to feed quality.

Title optimization is the highest-leverage feed improvement. Your product titles should follow this structure: Brand + Product Type + Key Attribute + Size/Color. For example: "Nike Air Max 90 Running Shoes - Men's Black/White - Size 11."

Why this matters: Google matches Shopping ads to search queries based on your product title and description. A title like "Cool Shoes" won't match "men's black running shoes size 11." Specificity drives relevance, relevance drives click-through rate, and CTR drives Quality Score.

Other feed optimizations that move the needle:

If you're comparing PMax to other paid channels, our Meta Ads vs Google Ads breakdown covers how to allocate budget between platforms.

New 2026 PMax Features You Should Use

Google shipped several meaningful PMax updates in early 2026:

Campaign-level experiments: You can now A/B test PMax campaigns against each other with a proper 50/50 traffic split. Previously, testing PMax variations required running parallel campaigns that competed against each other. This is a significant improvement for performance marketing teams that want to validate structural changes.

Asset group-level reporting: Google now provides more granular performance data at the asset group level, including conversion attribution by asset group. This helps you understand which product groups and audience signals are driving results.

Negative keyword support (expanded): PMax now allows up to 1,000 campaign-level negative keywords. Use this to exclude branded terms (if you run separate brand campaigns), irrelevant search queries, and competitor terms you don't want to bid on.

AI-powered asset generation: Google can now generate headline and description variations based on your landing page content. These auto-generated assets are worth testing — some accounts see 10–15% CTR improvements from AI-optimized copy. But review them manually; the AI sometimes generates claims your brand wouldn't make.

Bidding Strategy and Budget Allocation

Two bidding strategies dominate PMax for ecommerce:

Maximize Conversion Value — tells Google to get the most revenue possible within your budget. Use this when you're scaling and want maximum top-line revenue.

Maximize Conversion Value with Target ROAS — adds a return-on-ad-spend target. Google will try to hit your ROAS goal, which may mean spending less than your full budget if the algorithm can't find conversions at your target efficiency.

Our recommendation: start with Maximize Conversion Value (no target) for the first 4–6 weeks. This gives the algorithm maximum flexibility to learn. Once you have 50+ conversions, switch to Target ROAS and set it 10–15% below your actual observed ROAS. Gradually increase the target as performance stabilizes.

Budget allocation rule of thumb: PMax campaigns need at least $50–$100/day to exit the learning phase within a reasonable timeframe. If your total Google budget is $3,000/month, run one PMax campaign rather than three campaigns at $1,000 each.

Common PMax Mistakes to Avoid

After auditing hundreds of PMax setups, these are the errors we see most often:

PMax is the default Google Ads campaign type for ecommerce in 2026 — but default settings produce default results. The brands winning with PMax are the ones treating it as a system that needs strategic inputs, not a "set it and forget it" solution.

If your PMax campaigns aren't hitting the ROAS you need, our performance marketing team can audit your setup and identify specific optimization opportunities.