Performance Max Audit: A 12-Point Review to Take Back Control

Performance Max is usually audited in one of two ways.

Either it is treated as untouchable, on the basis that automation knows best and interference causes harm. Or it is blamed for everything, on the basis that nobody can see inside it.

Both positions are now out of date.

The controls and reporting available in Performance Max have expanded considerably. Campaign-level negative keywords, brand exclusions, account-level placement exclusions, network-segmented placement reporting, channel performance reporting, budget pacing views and first-party audience exclusions all exist.

The problem in most accounts is not that Performance Max cannot be audited. It is that nobody has audited it since the day it launched.

This guide covers twelve checks, in the order that usually matters commercially.

Before You Start: What Are You Actually Measuring?

Performance Max reports a blended result across Search, Shopping, Display, YouTube, Discover, Gmail and Maps.

That single ROAS figure can conceal:

Branded search demand the business already owned Remarketing conversions that would have happened anyway A small number of products absorbing most of the budget Display and video spend producing view-through credit that cannot be validated

The first job of a Performance Max audit is to work out how much of the reported result is genuine incremental acquisition, and how much is credit for demand the campaign did not create.

1. Conversion Tracking and Goal Configuration

Everything downstream depends on this.

Check:

Which conversion actions are set as primary for the campaign Whether any actions are double counted Whether conversion values reflect profit-relevant revenue, excluding VAT and delivery where appropriate Whether low-value actions such as newsletter signups are influencing bidding Attribution model and conversion window Whether refunds and cancellations are reflected Whether the campaign uses campaign-specific goals or inherits account defaults

If a campaign is optimising towards the wrong action, no structural change will fix it.

2. Brand Containment

This is the single most common source of overstated Performance Max performance.

Left unrestricted, Performance Max will capture branded queries. Branded traffic converts at a high rate and a low cost, which inflates campaign ROAS and makes the campaign look like an acquisition engine when much of it is demand harvesting.

Check:

Whether brand exclusions are configured at all Whether your own brand terms are excluded, and whether misspellings and variations are covered Whether competitor brand lists are being used deliberately or ignored For retail accounts, whether the Search-only exclusion option is appropriate, allowing Shopping ads to continue serving on brand queries while text ads do not Whether the branded Search campaign has lost impression share since Performance Max launched What proportion of Performance Max conversions come from branded queries

If brand is not contained, every other number in the audit is distorted.

3. Negative Keywords and Search Term Quality

Campaign-level negative keywords and account-level negative lists now apply to Performance Max.

Check:

Whether any negatives are in place Whether search term data has ever been reviewed Spend against informational, job-seeking or irrelevant queries Whether the campaign is matching to products or services you do not offer Whether negatives are so aggressive that valuable queries are blocked

Review search terms on a fixed schedule, not once at launch. This is the highest-frequency maintenance task in a Performance Max campaign.

4. Channel and Placement Distribution

Channel performance reporting shows how spend and results split across Search, Shopping, Display, YouTube, Discover, Gmail and Maps. The placement report, segmented by network, shows where ads actually served.

Check:

The proportion of budget going to each channel Whether the split is stable or has shifted suddenly Whether Display and video spend is proportionate to their contribution Placement quality, including apps and Search Partner inventory Whether account-level placement exclusions are in use Whether content suitability and inventory type settings match brand requirements

A campaign that reports well overall can still be routing a meaningful share of budget to inventory the business would never buy deliberately.

5. Merchant Center Feed Health

For retail accounts, the feed is the campaign.

Check:

Disapprovals, warnings and suspended items Missing GTINs, brand fields and product attributes Title quality against actual search demand Description, image and category completeness Availability handling and stock synchronisation frequency Whether promotions and sale price attributes are used How many products have received no impressions in the period

Feed work is unglamorous and routinely the highest-impact finding in a Performance Max audit.

6. Product Segmentation and Margin Exposure

Performance Max will optimise towards revenue unless told otherwise.

Check:

Spend and revenue distribution by product, category and price band Whether a small group of SKUs consumes most of the budget Whether low-margin or clearance lines are being scaled Whether high-margin products receive any meaningful exposure Whether asset groups or listing groups separate products by commercial priority Whether supplemental feeds are used to pass margin or priority signals

Segmentation should exist because it gives control you will use, not because it looks tidier.

7. Asset Group Structure and Search Themes

Check:

How many asset groups exist and what logic separates them Whether each asset group has enough data to perform Whether asset groups mirror the site or product structure sensibly Whether search themes are specific and commercially relevant Whether search themes duplicate existing Search campaign coverage Whether any asset group is limited by low asset quality or missing assets

One asset group covering an entire catalogue is a common structural weakness. So is twenty asset groups splitting the data so thinly that none of them learn.

8. Asset and Creative Coverage

Check:

Headline and description quantity and variety Image coverage across required aspect ratios Whether video assets are present or auto-generated Asset performance ratings and unaddressed low performers Whether creative reflects current propositions, pricing and offers When assets were last refreshed Whether structured asset testing has been run

Weak creative limits what the automation can do, regardless of how well the campaign is configured.

9. Audience Signals and Exclusions

Audience signals guide the system rather than restrict it, but they still matter.

Check:

Whether first-party customer lists are uploaded and current Whether audience signals reflect the target customer or a generic remarketing list Whether first-party audience exclusions are used to focus spend on acquisition rather than existing customers Whether new customer acquisition goals or high-value new customer settings are configured and appropriate Audience reporting by demographic for obvious mismatches

For lead generation especially, a campaign spending heavily against existing customers will report well and grow nothing.

10. Final URL Expansion and Landing Pages

Check:

Whether final URL expansion is on, and whether that is intentional Whether page feeds and URL exclusions bound it appropriately Which pages are actually receiving traffic Whether traffic is landing on blog posts, out-of-stock pages or generic category pages Mobile load speed on the highest-traffic destinations Message match between asset copy and landing page

On large or messy sites, unbounded URL expansion is a quiet and significant source of waste.

11. Budget, Bidding and Pacing

Check:

Whether the target ROAS or CPA is realistic against historical performance How frequently targets have been changed, and by how much Whether performance shifts coincide with target changes Budget pacing and projected end-of-month spend Whether the campaign is limited by budget while performing profitably Whether seasonality adjustments have been used appropriately

Frequent large target changes destabilise Performance Max more than almost anything else, and they rarely appear in a checklist audit.

12. Overlap With Search and Shopping

Performance Max does not operate in isolation.

Check:

Whether Search campaigns cover the same queries How the ranking rules between campaign types affect which one serves Whether standard Shopping campaigns are competing for the same products Whether remarketing audiences overlap Whether impression share in Search campaigns has changed since Performance Max launched Whether the overall account structure gives a clear role to each campaign type

The question is not which campaign performs better in the interface. It is which structure produces the best total commercial result.

Ranking the Findings

An audit that produces twelve observations has produced nothing.

Score each finding on:

Impact — estimated commercial value if resolved Effort — time, complexity and dependencies Confidence — how reliable the data behind it is Reversibility — how easily it can be undone

Start with high-value, low-effort, reversible items. Exclusions and settings changes can be undone in minutes. Restructures cannot.

What to Fix First

Week 1 Brand exclusions, conversion goal configuration, negative keywords, obvious placement exclusions.

Weeks 2 to 4 Feed corrections, product segmentation, asset group structure where control is genuinely needed, asset coverage gaps.

Weeks 5 to 8 Audience signals and exclusions, URL expansion boundaries, landing page improvements, target adjustments and structured asset testing.

Change one significant thing at a time and allow the campaign to settle. Performance Max reacts poorly to simultaneous changes, and simultaneous changes cannot be measured anyway.

Common Performance Max Audit Mistakes

Accepting blended ROAS at face value Brand and remarketing credit inflate it in most accounts.

Assuming the campaign cannot be controlled The available controls have expanded substantially. Most accounts simply are not using them.

Over-excluding early Aggressive exclusions applied during a learning period damage performance and produce misleading conclusions.

Restructuring before fixing the feed Feed quality usually outweighs structure in retail accounts.

Judging channel splits without context Display or video spend is not automatically waste. It becomes waste when it is disproportionate to its contribution and nobody has checked.

Rebuilding rather than repairing A rebuild discards learning history. It should be justified, not assumed.

When Performance Max Is Not the Answer

Performance Max is not universally appropriate.

It tends to struggle where:

Conversion volume is very low Lead quality varies enormously and is not fed back into the account The product range is tiny and easily covered by Search and Shopping Brand demand dominates and cannot be cleanly separated Strict placement or brand safety requirements exist The business needs granular control over specific products or markets

An audit should be willing to conclude that a different campaign type would serve the objective better.

How AI Can Support a Performance Max Audit

The difficulty with Performance Max is volume and fragmentation. Search terms, placements, channel splits, product-level performance, asset ratings and audience data all sit in different reports, and the leaks are individually small.

AI can process:

Search term relevance at scale Channel and placement spend distribution Product-level profitability and exposure Asset coverage gaps across asset groups Overlap between Performance Max and Search Ranking of findings by impact, effort and confidence

GoogleAdsAudits.com reviews account structure, targeting, bidding, creative, feed and channel mix, then produces a prioritised action plan with each recommendation positioned by impact, effort and confidence rather than presented as one undifferentiated list.

The analysis can be automated. The decision about what to exclude, what to restructure and what to leave alone still needs commercial judgement.

Final Thoughts

Performance Max rewards the accounts that manage it and punishes the ones that leave it alone.

A proper audit answers five questions:

How much of the reported result is genuinely incremental? Where is the budget actually going, by channel and by placement? Which products and pages are receiving the spend? Which controls are available and unused? Which fixes are worth the most, and which should come first?

Answer those, and Performance Max stops being a black box and starts being a campaign type like any other.

Build a personalised Google Ads action plan that identifies wasted spend, scalable opportunities and the fixes your team should prioritise next.

Frequently Asked Questions

Can Performance Max actually be audited?

Yes. Channel performance reporting, network-segmented placement reports, search term insights, asset performance data and budget reporting give considerably more visibility than the campaign type had at launch.

Should I always exclude my brand from Performance Max?

In most accounts with a dedicated branded Search campaign, yes. Brand traffic inside Performance Max inflates reported performance and obscures acquisition results. Retail accounts may prefer the option that restricts text ads while allowing Shopping ads to continue serving.

How many asset groups should a Performance Max campaign have?

Enough to give control you will actually use, and few enough that each has sufficient data to perform. Structure should follow commercial priority, such as margin, category or customer type.

Do negative keywords work in Performance Max?

Yes. Campaign-level negatives and account-level negative keyword lists apply, and search term review should be a routine task rather than a one-off.

How often should Performance Max be reviewed?

Search terms and pacing weekly or fortnightly depending on spend. Channel splits, placements, assets and product exposure monthly. A full audit at least annually or after any significant change.

Why did my branded Search campaign lose impression share after launching PMax?

Most likely because Performance Max is competing for the same branded queries. Brand exclusions are the usual remedy.

Is a low ROAS in Performance Max always a problem?

No. Once brand is excluded, the reported ROAS often falls because it no longer includes cheap branded conversions. That is a more honest number, not a worse campaign.

Should I rebuild an underperforming Performance Max campaign?

Rarely as a first step. Correct the tracking, contain brand, fix the feed and apply the available controls before considering a rebuild that discards learning history.

AUTHOR BIO

Jack Felstead is an award-winning Paid Media, AI and Digital Growth Consultant with 17 years of experience helping businesses improve customer acquisition, advertising efficiency, revenue and profit.