The book on running Performance Max campaigns intentionally rather than letting Google's automation do whatever it wants with your budget. Covers the structural moves: the asset-group design that influences how Google's machine learning attributes performance, the audience signals that actually inform the model (versus the ones that get ignored), the conversion-tracking setup that's required for PMax to even work, and the optimisation patterns that have emerged from real campaigns at scale. Built for the paid-search operator who's noticed PMax is the default now and wants the operator-tier read on how to use it well.

The structured monitoring rhythm for any active PMax campaign. Walks through the daily check (the metrics that need eyes daily, not the dashboard noise that doesn't), the weekly check (asset performance, audience signal effectiveness), the monthly check (campaign-level decisions about what to scale, hold, or kill), and the alert thresholds that prevent surprises. Run on every active campaign. Most operators monitor too much daily and too little weekly; the document calibrates the right cadence per metric.

The structured check before scaling spend on any PMax campaign — the operational discipline that prevents the most common failure pattern (doubling budget on a campaign that hasn't actually validated the unit economics). Walks through the unit-economics sanity check (real CAC including post-purchase costs), the audience signal verification (does the campaign know who it's converting), the asset performance check (which assets are actually getting served), and the scale-step calibration (the right percentage increase per period that prevents Google's model from getting confused). Built for the operator scaling spend deliberately.

The pre-launch check for any new PMax campaign. Walks through the conversion tracking verification (the absolute prerequisite — PMax doesn't work without good conversion data), the audience signal preparation, the asset library completeness (Google needs all the slots filled to optimise well), the budget calibration that gives the campaign enough learning runway, and the campaign-structure decision (one campaign or multiple). Run before every launch. Most underperforming PMax campaigns have a structural setup problem that no in-flight optimisation can fix.

The detailed framework for creating PMax assets that earn impression share rather than getting deprioritised by the algorithm. Covers the asset-quality criteria Google's auction actually rewards, the asset variety that keeps the optimisation algorithm from boxing you into one creative direction, the platform-specific calibration (assets perform differently on YouTube versus Display versus Discover), and the iteration protocol for replacing underperforming assets without disrupting the campaign. Built for the marketer producing assets at the cadence PMax needs.

The detailed launch sequence for a PMax campaign. Covers the pre-launch verification (conversion tracking, audience signals, asset readiness), the launch-day decisions (budget pacing, learning-period protection), the first-week protocol (what to monitor, what to leave alone — most operators meddle too early and disrupt the model), the second-week optimisation, and the transition to ongoing operations. Worked example through one full launch. Pair with the troubleshooting guide for the issues that emerge in the first month.

The structured diagnosis for PMax campaigns that aren't performing. Walks through the symptom-to-cause mapping (low impression share, high CPA, asset rejection patterns, attribution discrepancies), the verification steps that distinguish a real problem from normal variance, the intervention hierarchy (which fixes to try first, in what order), and the escape protocol when a campaign should be killed and rebuilt rather than salvaged. Built for the operator running multiple PMax campaigns and tired of the trial-and-error approach to underperformance.

Eight daily emails that take an operator from 'I've heard PMax exists' to running a deliberate PMax campaign. Day 1: the model behind PMax (so the optimisation choices stop being mysterious). Day 2: conversion tracking setup. Day 3: audience signal preparation. Day 4: asset creation. Day 5: campaign structure decision. Day 6: launch. Day 7: monitoring rhythm. Day 8: optimisation and scaling. Built for the paid-search operator new to PMax or running it badly and ready for the structural reset.

Working prompts for the AI-assisted parts of PMax management: the campaign post-mortem from raw performance data, the asset-rejection analyser (paste the rejection reason, get the structural fix), the audience-signal optimiser (input current performance, get suggested signal adjustments), the budget-pacing scenario planner. Each prompt comes with input/output format. Tested across Claude and ChatGPT. The value is in turning PMax analysis from a 'someone who knows PMax has to look at this' problem into something a small team can do well.