Performance Max: The Guide Brands Actually Need

Performance Max campaigns sound like a marketer’s dream: one campaign, every Google channel, AI-optimised in real time. In practice, without the right inputs, they become a budget disposal mechanism — spending aggressively on low-intent placements while your actual buyers go elsewhere.

What Performance Max Actually Does

PMax is a campaign type that runs across all of Google’s inventory — Search, Shopping, Display, YouTube, Gmail and Maps — from a single campaign. Google’s AI allocates budget across channels based on what it predicts will convert.

The critical word is “predicts.” PMax uses your conversion signals, creative assets and audience inputs to build a model of who your best buyers are. The quality of that model determines everything. Give it weak signals and it optimises toward the wrong outcomes. Give it strong signals and it’s genuinely one of the most efficient campaign types available.

The Signal Problem

Google’s algorithm needs to understand who your best buyers are. It learns this from your conversion events, your audience lists and your creative performance. Most accounts have significant gaps in all three.

Conversion tracking gaps: If you’re only passing purchase events, Google has no visibility into which customers are actually valuable. Pass enhanced conversions with lifetime value data and the algorithm learns to find customers worth keeping — not just customers worth acquiring once.

Audience signal gaps: PMax performs dramatically better when you provide audience signals — customer match lists of your existing buyers, high-intent segments from GA4, lookalikes built on your best customers. Without these, Google starts from scratch and the learning phase burns more budget.

Creative gaps: PMax uses your creative assets to serve ads across every format and channel. If you provide one image and one headline, you’ve given Google nothing to test. The algorithm needs variety to find what resonates.

Asset Group Architecture

Most brands set up one asset group for everything. This is a mistake. Asset groups should mirror your product catalogue — grouped by category, audience type or margin tier.

A clothing brand should have separate asset groups for men’s, women’s, sale items and new arrivals. Each group gets headlines, descriptions and images specific to that category. This gives Google context to serve relevant creative to relevant audiences — and gives you the reporting visibility to identify which segments are performing.

Within each asset group, aim for: 5+ images in different formats and ratios, 3–5 headline variations testing different angles (price, outcome, urgency), and 2–3 long description variations.

Brand Exclusions and Search Themes

PMax will serve on branded search terms unless you explicitly exclude them. This is one of the most common and expensive mistakes — it means PMax is claiming credit for conversions that would have happened anyway through branded organic traffic.

Add your brand name and all variations as negative keywords at the account level. Then review the Search Terms report weekly (in the Insights tab) to catch any inappropriate placements early.

Bidding Strategy and Budget

Start with Maximise Conversions, no target CPA. Let PMax learn your conversion landscape for 4–6 weeks before adding constraints. The algorithm needs at least 50 conversions per month to optimise effectively — pushing CPA targets before that forces it into an under-delivery spiral.

Once you have sufficient data, layer in a target ROAS or CPA set roughly 20% above your recent actual performance. Gradually tighten as the campaign matures. Daily budget should be at least 15–20× your target CPA.

Reporting and Optimisation

PMax has limited native reporting compared to standard campaigns. Use the Insights page for audience segment and creative asset data, and supplement with GA4 — look at new vs returning customer ratios, session quality and LTV of PMax-acquired customers vs other channels.

Monthly: refresh low-performing creative assets, review search themes for relevance, update customer match lists. Quarterly: run an incrementality test by pausing PMax for 2 weeks in a lower-revenue period and measure the impact on total conversions.