8 min read
How to Audit PMax Channel Performance Across Client Accounts
Pull comparable channel data from a manager account, avoid misleading averages, and send each finding to the account it belongs to.
A manager account review of Performance Max channels exists to decide where the team's attention goes this week. The channel performance report is accessible from manager accounts, so the data is available in one place; the work is making it comparable and reading it against each client's own goal and history rather than against a portfolio average.
There is no healthy channel mix for a portfolio. A retailer, a local service business, and a software company should not be judged against one Shopping share, and Shopping is a product data segment in the report rather than a channel in any case. Compare each account with itself, group peers only when they genuinely share a business model, and turn every difference you find into a specific question for the person who runs that account.
Compare each account with itself first
Two clients with very different channel mixes may both be exactly where they should be. A catalog retailer will show heavy product ad delivery; a firm that wants a dozen qualified enquiries a month will not. Flagging whichever account sits furthest from the agency average finds the most unusual business, not the most troubled campaign.
Start with whether each account is delivering the result its client values, then look for meaningful movement against that account's own history. A familiar mix with falling qualified-lead volume deserves more attention than an unusual mix with healthy sales.
Peer comparisons earn their place when the accounts genuinely have things in common: business model, market, goal, use of product data, seasonality, rough scale. Say what the grouping is. A band that nobody can justify is a benchmark in costume.
Keep the distinction in client language too. Different from our other accounts is an observation. Wrong for this business is a conclusion, and it needs the account's own evidence.
Build a table that survives a second look
The channel performance report is accessible from the manager account, so start there rather than opening each child account in turn. Where a child account has filters or conversion settings the manager view does not reproduce, export from the child and combine, and say which rows came from where.
Keep the identifiers in the working data: account, campaign, currency, time zone, reporting period, filters, and the conversion actions being compared. The client summary can use readable names. The analysis needs a path back to its source.
Make sure every export is the same breakdown. Google's channel definition separates the seven serving surfaces from formats such as ads using product data. A row filtered to product data in one account next to an unfiltered channel row in another produces a difference that is not there.
Mark missing data as missing. A report that is unavailable, filtered differently, or not yet collected is not a zero, and the distinction matters most when a sheet or an agent is combining files without a person watching each one.
| Keep with the export | Why |
|---|---|
| Account, campaign, and reporting period | The reviewer can return to the exact evidence. |
| Currency and time zone | Spend and daily periods are not directly comparable across them. |
| Conversion actions and value definitions | A form submission is not a purchase or a qualified opportunity. |
| Channel, format, and other filters | Different subsets should not sit in the same column. |
Open the channel performance report from the manager account
- Campaigns
- Insights and reports
- Channel performance
Do the arithmetic before trusting the average
Take two hypothetical accounts in the same currency with genuinely comparable conversion actions. Account A spent $1,000 for 10 conversions, a cost per conversion of $100. Account B spent $9,000 for 300 conversions, a cost per conversion of $30. The average of the two figures is $65. The combined figure is $10,000 divided by 310, about $32.26.
Both numbers answer a question, and not the same one. The $65 describes a typical account. The $32.26 describes the portfolio's money. Label whichever you use; calling either one portfolio CPA without saying which invites the wrong decision.
Channel share has the same trap. An average of account percentages weights a $500 account as heavily as a $50,000 one; a spend-weighted share lets the big accounts dominate. Pick the measure the decision needs and keep the individual rows visible beside it.
Even correct arithmetic can hide the thing that matters. A large client's strong quarter can cover a small client's collapse. Read material changes account by account before presenting a reassuring total, and only add money across accounts after currencies are reconciled and only add outcomes that mean the same thing.
| Measure | Hypothetical result | What it describes |
|---|---|---|
| Average of account CPAs | $65 | A typical account; each account counts once |
| Combined CPA (total cost over total conversions) | $32.26 | The portfolio's money; large accounts dominate |
Prioritize the problem, not the strangest chart
Give priority to findings with a business consequence: a tracking failure, a product eligibility problem that removed a client's best sellers, a sustained fall in valuable conversions, a client-specific suitability concern. An unusual percentage can lead you to one of those. It is not one of those.
Open the child account and read the timeline. Check the conversion definition, the status diagnostics, campaign changes, and business events around the date the row moved. If several accounts moved together, look first for what they share, a promotion calendar, a reporting delay, an integration, before assuming Google changed something.
Leave each recommendation with a named person and a concrete question. Check whether the feed change removed our best-selling products from eligibility is actionable. Investigate channel quality sends someone back to the same chart.
Resist batch changes. Several accounts on the same exception list is a reason to look at each one, not a licence to adjust budgets or targets across all of them. Each proposed change needs an account-specific reason and a scope, and each one is approved on its own.
Questions and answers
Is the Performance Max channel report available from a manager account?
Yes. Google's Highlights of 2025 page says the channel performance report is accessible via manager accounts. Child-account filters and conversion settings still need to be reconciled before rows are combined.
Is there a healthy Shopping share for Performance Max accounts?
No. Shopping is the ads using product data segment, not a channel, and the right share depends on the business. Compare an account with its own history and goal, and use peer groups only when the businesses are genuinely alike.
Should I average the CPA of each client?
Only if a typical-account figure is what you mean to report. For the portfolio's money, divide comparable total cost by comparable total conversions, and label the result. Keep currencies, conversion definitions, and the individual rows visible.