Performance Max in 2026: What's Changed and What Actually Works Performance Max en 2026: qué ha cambiado y qué funciona realmente

What's Actually New in PMax 2026

Performance Max has received more meaningful updates in the past 12 months than in its first two years of existence. Google has responded — partially — to advertiser demands for transparency and control. Here's what's genuinely changed:

Campaign-Level Negative Keywords

This was the most requested feature since PMax launched. You can now add negative keywords at the campaign level, preventing your PMax campaign from appearing on brand competitor terms, irrelevant search categories, and any queries that consistently waste budget. This is a significant improvement — but the implementation still has limits. You can't add negatives at the asset group level, and the negative keyword list isn't shared with your account-level negative lists automatically.

URL Exclusions

You can now exclude specific URLs from your campaign's reach, preventing your ads from appearing on specific websites, apps, or YouTube channels. This addresses one of the long-standing transparency complaints — budget appearing on content that damages brand perception.

Search Term Insights

Google has expanded the Search Term Insights report within PMax, now showing categorized search themes rather than just a vague "search" breakdown. You still don't get exact match data, but the category-level reporting makes it possible to understand roughly what search intent is driving performance — and which categories to consider adding as negatives.

Asset Group Performance Labels

Asset groups now have more granular performance labels (Learning, Limited, Good, Best) that update more frequently. This makes it easier to diagnose which asset groups are underperforming and need creative refresh.

What Still Doesn't Work

Transparency advocates and advanced advertisers will still find significant gaps:

• No ad group level reporting — PMax remains a black box at the placement level. You can't see what percentage of spend goes to Search vs Shopping vs Display vs YouTube vs Gmail vs Maps. This makes optimization decisions frustratingly data-poor.
• No frequency capping for Display/YouTube — Users can see your display and video ads dozens of times with no control on your end.
• Limited audience exclusions — You can't exclude specific audiences at scale the way you can in standard campaigns.
• Asset group bidding is impossible — You can't allocate budget differently between asset groups. The algorithm decides all allocation.

PMax vs Standard Shopping: When to Use Each

Use Performance Max when: You have 30+ conversions per month, your product catalog is diverse, you want to reach customers across all Google properties, and you have strong creative assets (images, videos, copy).

Use Standard Shopping when: You need granular bidding control per product, you're testing new products with unknown conversion rates, your margins vary significantly across SKUs, or you want full visibility into search query data.

The hybrid approach: Many sophisticated advertisers run Standard Shopping for their highest-margin or highest-volume product lines (where they want control) and PMax for discovery and long-tail categories (where the algorithm's reach is valuable).

Optimal Asset Group Structure

Asset group structure is the primary lever you have for guiding the PMax algorithm. The recommended framework:

• 5–8 asset groups maximum per campaign — More than this dilutes performance data, preventing any individual group from accumulating enough conversions to optimize effectively.
• Theme by product category or audience intent — Not by product attribute. "Running shoes - performance seekers" is a better theme than "Running shoes - blue".
• Each asset group needs its own audience signals — Don't use the same signals across all groups. Differentiate by customer list, custom intent, or in-market segment.
• Align landing pages with asset group themes — A high-performance asset group loses its impact if it sends traffic to a generic homepage. Match the URL expansion settings to keep traffic relevant.

Audience Signals: What They Actually Do

This is one of the most misunderstood aspects of PMax. Audience signals are not targeting — they're suggestions. You're telling Google: "These people are likely to convert, so use them as a reference." The algorithm will still show your ads to people outside these audiences if it predicts they'll convert.

The most effective audience signals to use:

1. Customer match lists — Your existing customer emails. This is the highest-quality signal you can give the algorithm.
2. Website visitors — Especially product page visitors and cart abandoners. Strong conversion intent signals.
3. Custom intent audiences — Built from search queries. Tells the algorithm to look for people searching terms related to your product.
4. In-market segments — Less powerful than the above but useful as a secondary signal, especially for awareness asset groups.

Creative Requirements: What You Actually Need

PMax is a creative-hungry campaign type. Providing only the minimum required assets is a strategic mistake — the algorithm performs significantly better with more creative variety to test and optimize.

• Images: 15+ recommended — Cover all required sizes: square (1:1), landscape (1.91:1), and portrait (4:5). Include lifestyle images, product images, and in-context shots.
• Videos: Required (not optional) — Google will auto-generate video from your other assets if you don't provide one, and auto-generated video consistently underperforms human-created video. Provide at least one 15-second and one 30-second video.
• Text: 15 headlines, 5 descriptions — Write these with clear differentiation. Avoid redundancy — if the algorithm combines headline 1 and headline 3, they should make sense together and not repeat the same claim.

Decision Matrix: Should You Use PMax?

Use this framework to decide whether PMax is right for your situation:

When to Pause PMax

There are clear signals that PMax is hurting your account and should be paused or restructured:

• ROAS has been below target for 3+ consecutive weeks despite budget and creative optimization
• Standard Search campaign impression shares have dropped significantly since PMax was launched (cannibalization signal)
• Brand search volume appears in PMax reports despite brand exclusions
• All asset groups are labeled "Learning" after 4+ weeks of running (insufficient conversion data)