Google Ads household-income audits for consultants: explain percentile coverage and justify exact changes

Choose the commercial question

A household-income audit should begin with the client's intended decision. They may want a report of primary outcomes, an explanation of a Mixed status, creative ideas for a confirmed offer or a change to a particular criterion. These intentions require different evidence. A consultant should not assume that every income question is a request to restrict advertising.

Establish the account or campaign, dates and outcome measure when they are absent. If those facts are already supplied, retain them through follow-up messages. “Compare consultation CPA by income band for this campaign” is a defined performance question; “exclude this band in that group” introduces a separate live effect requiring a concrete proposal.

Agree how the answer will be used. A client presentation may need only a bounded table and a coverage explanation. A live criterion change needs actual current state and supported capability checks. A creative brief can often use the confirmed business offer without a provider request. This division keeps the work proportionate.

Write the deliverable and completion condition in the engagement note. A useful comparison can be complete even when no edit is justified. A draft policy can remain pending owner input. A verified mutation closes its configuration task, while commercial evaluation may remain scheduled. These are distinct milestones rather than interchangeable claims of optimisation.

Understand the reported percentile bands

The current audit presents Top 10%, 11–20%, 21–30%, 31–40%, 41–50%, Lower 50% and Unknown in a consistent order. These labels represent the provider's income-range categories. They are not currency amounts or salary values disclosed by the account.

Google's income-range reference defines the percentage ranges and an undetermined category. Its greater-than-ninety range maps to the audit's Top 10% label. Read the definition before interpreting the percentage as a conversion rate or a share of the client's buyers.

The category name and its performance measure are separate fields. A Top 10% row can have any returned conversion rate; the label itself does not mean ten percent of visitors converted. Likewise, Lower 50% does not mean that half the campaign's cost belongs to that row.

Keep that distinction visible in the client report. Use the band name as the grouping label and the actual cost, outcomes and ratio as the observed measures. A readable explanation prevents the consultant from turning a percentage in a category name into a numerical business conclusion it does not support.

Separate a category from a customer's finances

A reported band is not proof of an individual customer's salary, savings or ability to afford the offer. The report supplies aggregate advertising metrics associated with the provider category. It does not give the consultant a financial profile of identifiable customers.

Google's demographic guidance explains supplied and estimated information, country-dependent income availability and people whose demographics are not identified. That classification mechanism should remain separate from claims about the business's actual purchasers.

Do not infer buying intent from a band name. A person can purchase on behalf of somebody else, research a future need or respond to an offer for reasons the table does not establish. Those are potential business contexts, not measured properties of every reported category member.

A consultant can still use the comparison for a valid decision. Describe the actual observation, selected population and measure. Avoid extending it into a claim that a whole income category cannot be a valuable customer or that an advertising platform has verified the financial situation of each person reached.

Check country and campaign eligibility

Retain eligibility notes when preparing a recurring client review. A later campaign expansion can introduce activity that the earlier income report did not represent. That does not make the earlier scoped answer incorrect, but it can change the relevance of its denominator to the new business question. Revisit actual coverage when the account changes materially rather than carrying a permanent assumption that one previously available report surface covers every future market and campaign.

Household-income availability depends on the relevant provider context. The current guidance identifies eligible countries and campaign options rather than promising universal coverage. Check the actual campaign and returned evidence when that eligibility matters to the client's request.

A missing income table is not automatically proof that the account has no advertising activity. The selected population can be outside the report surface, genuinely empty or unavailable because a query failed. Establish which state occurred before presenting a confident all-zero conclusion.

Mixed-country advertising can complicate a broad report. A returned income population may not represent all activity in every targeted country. Keep the scope explanation attached to the result. Do not allocate unsupported activity into bands merely to make the total match a wider financial report.

If the owner asks for a specific country and band, resolve the supported reporting grain and current capability. Do not assume that a general account demographic card already supplies every geographic intersection. Actual available fields should guide the read, and any needed additional evidence should be obtained before claiming a complete answer.

Retain Unknown in the evidence

Unknown is a legitimate returned category, not the same state as a provider error. It represents unestablished demographic classification. The audit table retains it and includes its returned metrics in the reported total, so a consultant should not drop it solely to make a neat identified-band distribution.

The metric-led headline considers known bands separately. Its comparison can therefore have a different denominator from the table's Unknown-inclusive total. Explain that distinction when the owner compares the two statements instead of assuming one must be wrong.

If the user explicitly asks for identified bands only, honour that population and label it. Do not describe its total as all reported income activity. If they ask for all returned categories, retain Unknown alongside the named bands.

Excluding Unknown is its own reach decision. It does not establish that only unwanted customers would be removed. The consultant should resolve the business intention and exact target before preparing a supported change, keeping incomplete classification separate from demonstrated irrelevance or poor commercial value.

Read the actual income-report surface

The reader uses the income-range view, selecting the criterion's income type and negative flag alongside impressions, clicks, cost, conversions and conversion value. Account results aggregate by income type. Campaign results include campaign identity and aggregate the corresponding group rows within each campaign.

The returned negative flags drive inclusion status. They are not inferred from impressions or conversions. This prevents a quiet row from automatically being labelled excluded and preserves the possibility that different parents have different criterion settings.

The four demographic views are independent reads. A rejected income query propagates rather than being converted into a successful empty object. That distinction lets the report-loading boundary preserve unavailable evidence instead of allowing the income section to disappear as though it had been checked.

A focused conversation should select the relevant income fields and population through the shared reporting framework. It does not need every demographic or creative field to answer a single comparison. Broad audit composition and narrow conversational reads can use the same definitions while serving different scopes.

Explain the account's Reported total

The account demographic card labels its aggregate Reported total because the ad-group demographic surface does not include Performance Max and App activity. Other dimension and country eligibility can further affect coverage. This total should not be advertised as complete account spend.

The actual queries impose the selected period without an enabled-campaign predicate. A source comment describing enabled campaigns is not evidence of a filter absent from the query. Keep historical report activity separate from a current enabled-only inventory.

When the owner compares the table with the account financial total, reconcile the populations first. Different coverage can produce a legitimate difference. If scope and dates actually match but the discrepancy remains, obtain the relevant focused evidence rather than dismissing it or inventing an allocation.

Useful bounded reporting does not require pretending coverage is universal. The consultant can explain which campaign types and returned income rows contribute, then use an appropriate broader total for a separate account question. Clear denominator language protects both the comparison and any later decision based on it.

Find the campaign and offer behind the pattern

An account band combines participating campaigns and offers. Its observed ratio can reflect a different mixture from another band. Use campaign context to identify what contributes to the pattern before proposing a change for a particular offer.

The campaign income result still aggregates returned group rows. It does not prove that every group has the same performance, negative flag or intended commercial audience. A group-specific mutation requires the actual current criterion under that group.

The demographic card is informational and does not create a scored issue simply because bands differ. The consultant should separate observed comparison, interpretation and client decision rather than presenting a weaker ratio as an automatic setting defect.

A campaign outside this report population should not receive invented income metrics. If the owner wants a different supported source, discover and read it. Do not silently substitute audience ingredients or a financial hypothesis for actual returned performance data.

Interpret Included, Excluded and Mixed

Included, Excluded and Mixed describe the flags retained in the aggregated rows. Mixed occurs when both included and negative criteria contribute to the selected band population. It is not automatically an error in one resource.

Different parents can intentionally use different settings. A consultant should locate those actual parents when the owner asks why the status is Mixed. Do not choose the first group in a response as the target for a campaign-wide intention.

The historical metrics and current criterion question remain separate. An excluded row can have performance associated with an earlier part of the period. The flag is not derived from the outcome total, and the total does not certify current inclusion in every group.

For a live change, read the selected current criterion and ownership. The demographic view's flags are useful reporting evidence, but not automatically an exhaustive inventory of every current income criterion. A complete-sounding targeting statement needs the appropriate current scope rather than a convenient aggregate label.

Keep zero activity separate from current exclusion

The canonical table can retain a band with zero displayed metrics even when its current status is not established in the returned rows. A dash preserves that missing configuration fact. Do not convert a stable row layout into proof that every band's criterion was independently read.

A known zero conversion count can matter to a defined performance question. An unavailable selected outcome is different. Missing values must not become the zeros that make a poor-performance filter match conveniently.

A ratio can also be undefined because its denominator is zero. CPA needs the selected conversion count and ROAS needs cost. The table's dash should not be ranked as the lowest cost per acquisition or as an established amount of wasted spend.

When the owner asks whether the band is allowed now, obtain the actual criterion evidence. When they ask whether it generated outcomes, use the selected reporting period. Answering the correct question prevents quiet historical activity from becoming an unjustified live exclusion.

Preserve the requested reporting period

Last thirty days is a rolling interval, while last month is a calendar period. Preserve the user-selected dates through follow-up reporting and chart generation. A cached audit default should not silently replace an explicit conversational request.

Resolve account time context and treatment of the current partial day when relevant. Different end dates can change a short-period ratio. State the actual window clearly enough that the client can compare it with another report without guessing which days were included.

Current criteria do not reconstruct every historical setting. A band excluded late in the window may still have earlier activity. If timing matters, obtain the relevant history and keep its retention and coverage boundaries explicit.

The consultant change-history guide explains why a bounded event sample is not a universal record of every prior state. Combine useful history and performance while avoiding unsupported claims about the cause of each outcome.

Select the relevant primary outcome

The ordinary income audit uses Ads conversions and conversion value. These are not automatically purchase-only measures. Eligible primary leads and calls can contribute assigned values too. Keep the selected definition explicit before describing a band as generating revenue.

If the owner asks for purchases, select the corresponding scope. If they ask for a named consultation conversion, preserve that actual action rather than substituting all conversions. Primary intentions should not silently include secondary observations.

A business-type default can help choose CPA or ROAS for presentation, but explicit user instructions remain the actual request. The framework should use the requested supported measure and ask for missing business correspondence where necessary.

Calls and offline conversions can remain valid primary outcomes when they fit the client's question. Their presence does not make them equivalent to completed sales. Distinguish recorded action, assigned value and actual commercial result when explaining what the report establishes.

Calculate ratios from summed inputs

ROAS divides selected conversion value by cost, while CPA divides cost by selected conversions. Calculate the reported total from summed base metrics before calculating its ratio. Averaging income-band ratios creates a different measure and weights unequal populations incorrectly.

Keep the Unknown-inclusive table total separate from a known-band headline comparison. Both can be useful when their definitions are clear. They should not be described as identical totals merely because the displayed ratio names are the same.

Use numeric precision for calculations and threshold conditions, then simplify the displayed result to at most two decimal places. Comparing rounded strings can select the wrong side of a strict boundary. Additional visible decimals do not make a small observation more reliable.

Weekly or monthly aggregation needs the same base inputs. Sum the relevant cost and outcomes for each period, then calculate its ratio. A reduced chart point count should not quietly change ROAS into an average of daily ratios.

Read commercial scale alongside efficiency

Show cost and selected outcome counts beside a ratio when they matter to the decision. A small band can have an extreme observed result without carrying most of the opportunity. Another band can be material because of its volume even with a less dramatic difference.

Clarify what the owner means by worst. Lowest return, highest CPA and largest amount of low-return cost are distinct criteria. Do not identify one target under a ratio rule and justify another under volume without explaining the changed intention.

The consultant can agree an engagement-specific review condition, but should not advertise it as a universal provider guarantee. If the available evidence is too limited for the proposed change, name the further observation or outcome definition needed.

A decision to wait should have a scope, measure and review date. That makes it actionable rather than a vague statement that more data might help. It also preserves the existing comparison as useful evidence without forcing an immediate mutation simply to demonstrate progress.

Review the offer without affordability assumptions

Make the offer investigation concrete enough for the owner to answer. Identify the actual promise, selected destination and observed point of concern. A vague instruction to make the offer fit a band is difficult to verify and can encourage unsupported assumptions. A request to confirm whether the page explains the consultation price and booking process has an observable outcome. Keep that page task separate from the advertising criterion and from any later commercial comparison.

The client's offer is a better starting point than an assumed ability to pay. Ask which product or service the campaign promises and how the business defines a useful outcome. A reported percentile does not independently establish whether a particular person can afford or wants that offer.

Different offers in one account can justify separate interpretation. A professional workshop, a routine service appointment and a premium product can have different commercial paths. Use confirmed business context rather than assigning the same purpose to every campaign in an aggregated band.

If a destination or message is suspected to be unclear, identify a concrete observation task. The income ratio alone does not prove that the page is broken or that the headline needs an invented financial claim. Obtain actual creative or website evidence where the question depends on it.

A consultant can help the owner draft an offer brief conversationally. Keep hypothetical customer motivations labelled as planning ideas. They should not become measured statements about everyone in a provider income category or an implicit authorisation for live targeting changes.

Preserve exact filters and logical operations

An income question can combine band, cost, outcome and efficiency conditions. Greater than differs from greater than or equal to. AND requires all specified conditions; OR admits the defined alternatives. Preserve that logic through the shared report framework.

Apply the condition at the requested grain. An account band ratio does not establish that every campaign or group meets the same threshold. A campaign aggregate is not an exact editable criterion selection.

Include Unknown according to the requested population. Missing outcomes remain unavailable rather than becoming zero values that satisfy a low-performance rule. This principle matters both to a read answer and to any later target selection based on it.

A short displayed preview is not a complete count. When the owner requests all matches, establish the actual retained collection and supported coverage. Do not report the first ten visible examples as every matching resource merely because Slack output is deliberately concise.

Require actual evidence for combined populations

Separate income, age, gender and parental tables do not establish their joint intersections. Multiplying marginal proportions does not create observed performance for a particular combined audience. Use an actual supported combined report if that is the user's question.

The same caution applies to geography. A country total and an income total do not independently establish each country's income-band performance. Discover the actual supported fields and obtain the relevant additional read rather than manufacturing a cross-tabulation.

Daily detail needs actual date-segmented income evidence. A period band total and a daily campaign total cannot be joined by proportionally spreading the band across days. That would invent observations and can make an unsupported trend look convincingly smooth.

A configured audience definition is different again. Its ingredients describe an intended capability, not proof of historical performance for every intersection. Keep configuration, marginal reporting and joint observations clearly labelled so richer requests remain grounded in actual data.

Define any Analytics comparison clearly

A GA4 comparison needs the relevant property, traffic scope, dates and outcome definition. Its site-side population is not automatically the same as the Ads demographic-reportable population. Similar headings do not prove identical denominators or classifications.

For leads, select the corresponding event with the owner. The assistant can inspect actual returned names and suggest a likely candidate, but business correspondence needs confirmation. A generic submission event can include more than the named Ads consultation conversion.

Ads conversion value and GA4 purchase revenue can be useful labelled comparative series when requested. They need not be equal to be informative. Preserve their definitions and actual returned dates rather than refusing a useful comparison solely because the measures differ.

All-traffic GA4 revenue divided by advertising cost is blended ROAS or MER. It answers a broader business question. Do not label it as an attributed income-band return without actual matching evidence. The report should make the selected source and scope understandable to the owner.

Keep creative requests conversational

The assistant can suggest headlines, descriptions or a clearer offer explanation from confirmed context without calling a provider for every creative request. Ideas should be presented as ideas, rather than as measured income-band preferences.

A request for more premium-sounding copy does not authorise excluding a lower income range. Creative preparation and live targeting are separate effects. If the owner wants both, preserve both intentions and prepare each supported action clearly.

Keep factual promises confirmed. Do not invent financing, discounts, eligibility or product guarantees merely to make a financial theme sound specific. The owner can select wording that accurately reflects the actual offer.

The consultant segmentation guide explains how offer, intent and destination can inform structure. That planning work can be valuable without pretending a provider category reveals the motives or finances of each customer.

Discover the supported criterion operation

Check preservation requirements before selecting a replacement sequence. A new criterion can have a new identity and a different historical association from the removed object. The tool should retain the user's desired current settings through the actual supported operation rather than assume all omitted properties survive. If a required fact cannot be established, obtain it before confirmation. The owner is approving a concrete effect, not authorising the planner to discover an unspecified implementation after the click.

The editing framework uses local schema-derived tools with bounded discovery. The reviewed definitions include criterion creation and removal, and mark income-range and negative fields as creation-only. They should not be advertised as a universal in-place toggle.

A serving-status change does not automatically reverse a negative flag. Inspect the actual criterion and choose the operation matching the owner's intention. If removal and creation are required, describe those effects honestly rather than calling the sequence a rename.

Schema discovery supplies candidate capabilities. Current ownership, eligibility, preservation requirements and provider validation still determine whether the concrete request is valid. A listed field is not a guarantee that every campaign accepts the same configuration.

Generic definitions should make the framework adaptable without narrowing it to one income phrase. They can preserve field behavior, exact resource relationships and verification needs. The final proposal should explain the selected real effect rather than exposing a technical payload as a substitute for readable targeting intent.

Resolve exact parents and current settings

A reported band is not an editable resource identity. Resolve the actual current criterion and applicable campaign or group parent. Similar names and identical bands can appear in many parents, so exact scope remains essential.

Keep canonical references behind the scenes and show understandable names with correct relationships. A group should not be shown as the parent of another group. An account aggregate should not become a group-specific target merely because it supplied the first metric.

Local dependencies establish structure without a provider request just to learn that a criterion has a parent. Focused reads establish actual ownership and current configuration. Reuse relevant context while retaining the live checks required for execution.

For a multi-parent intention, identify every selected resource and current state before confirmation. Mixed status can require different operations in different groups. Do not execute one arbitrary criterion and claim the whole campaign has been changed.

Separate audience ingredients from exclusions

An audience definition containing income ingredients is not the same resource as a demographic exclusion criterion. A Performance Max audience signal also does not make PMax activity appear in the ordinary ad-group income report.

The catalogue has audience creation and attachment actions with their own account or asset-group targets. Creating an audience does not automatically attach it everywhere. An audience signal should not be described as a hard exclusion without the actual supported effect.

Resolve whether the owner wants a report, a signal or a targeting restriction. Do not route every income-related request through the most familiar action. Shared dependency definitions can establish valid structure while current reads establish the actual selected resources.

Separate prerequisite effects should remain visible in the plan. The framework can order supported dependencies, but cannot invent readiness or collapse distinct operations into a vague completed setup. The owner must see the concrete intention that confirmation will authorise.

Follow an illustrative consultant review

Imagine an invented consultant, Nora Ash, reviewing the fictional Spruce Design Appointments account. The owner asks for primary consultation CPA by income band in one Search campaign. Nora establishes the dates, reads the relevant evidence and retains Unknown in the table alongside base cost and outcome counts.

A known band has a different observed ratio, but the account also contains another offer with a different population. Nora keeps the campaign result separate from the broader summary. Mixed status prompts a focused current-group question rather than a blanket account exclusion.

The owner first requests clearer headlines for the consultation offer. Nora can brainstorm from the confirmed business promise without another metrics call. If a later targeting change is requested, she resolves the actual criterion and supported operation before presenting precise effects for confirmation.

The engagement records the observed pattern, creative work and any verified configuration separately. Subsequent evaluation uses stable measures and actual exposure dates. This example illustrates the process; it is not a customer result, an affordability diagnosis or a universal recommendation that a particular band should never receive the offer.

Reuse evidence and keep calls focused

A follow-up asking for a chart of the returned rows should reuse those rows when only presentation changes. Do not fetch another broad account inventory merely to redraw known income data. Retain the source, period and outcome definitions attached to it.

A new measure or date range can justify another focused read. A live edit can require current criterion state even when performance evidence is established. Select the fields genuinely needed rather than requesting every demographic, keyword and creative attribute by default.

Known dependencies and live evidence have different jobs. The former explains resource relationships; the latter establishes the actual account state. Optimisation means avoiding needless discovery while keeping current prerequisite checks that make the proposal correct.

Maintain context through short replies. Resolve ambiguity if several groups remain plausible, but reuse an exact established parent when the owner says that group. A generic reporting and editing framework should preserve that continuity across income and other account conversations.

Present readable reports and charts

Choose chart axes that represent real dimensions. A bar comparison across income bands can explain a period result; a line chart needs actual dated observations if it is intended as a trend. Do not allocate period totals across dates simply to satisfy the requested shape. The shared metric library should provide the same defined base values to tables and charts so a presentation change does not create a second numerical interpretation of the report.

A table should show readable band names, actual status and selected commercial measures. Keep raw IDs, internal provider timestamps and undefined parents out of the client-facing explanation. Reported total should not become All account spend.

A requested native chart can combine up to three compatible metrics or segments with clear labels. More than three needs the readability-limit explanation and selection. A band distribution and a daily trend require different actual dimensions rather than an invented continuous axis.

The ordinary native chart limit is twenty points per series. Longer periods can use stated weekly or monthly aggregation unless daily detail is explicit. Longer explicit daily requests need split native charts with the point-limit explanation rather than silent fragmentation or an unrequested image.

Use readable dates and actual week ranges, omitting the year for current-year daily labels. Display at most two decimal places and calculate grouped ratios from base metrics. Place the user mention before the feedback text and controls so feedback remains the final part of the reply.

Show exact effects before confirmation

The proposal should identify the campaign or group, selected band and exact supported effect. If implementation removes one criterion and creates another, show those consequences honestly. “Optimise income targeting” is too vague for live approval.

One confirmation can cover a supported sequence in clear step sections. If the full request needs narrower proposals for safe review, explain that requirement before execution. Do not execute an arbitrary subset or omit a requested effect because another action was easier to prepare. The approved scope should remain clear to the client.

State “No changes have been made yet. This proposal expires in 15 minutes.” Confirm and Execute applies to the displayed scope. A report, a creative suggestion or a prior unrelated approval does not authorise this mutation.

If the parent or band changes during review, prepare an updated proposal. The owner should be able to compare the original intention and concrete effect without deciphering provider resource syntax. Clear target relationships are part of the approval boundary.

Verify configuration and reconcile uncertainty

Immediately show prominent change-in-progress feedback and inactive controls after confirmation, without a user mention in the interim state. The owner should see that the click registered while current checks and request validation run.

Validation checks the prepared request, while verification checks resulting resources, fields and relationships. A submitted mutation is not automatically a fully verified result. Each requested effect needs the evidence appropriate to its actual operation.

If the write may have happened but readback is incomplete, reconcile the affected state before retrying. Blind replay can duplicate criteria or repeat a completed removal. Report established effects and unresolved verification separately.

When checks pass, make executed and verified prominent and describe the actual configuration outcome. This does not verify people's finances, future delivery, website behavior or commercial improvement. Those remain separate questions that the consultant should retain in the engagement.

Leave ownership and an evaluation plan

Set a practical review date and condition. The client might revisit after a completed observation period, a confirmed offer change or resolution of a measurement question. Name the selected measure and responsible person instead of promising a particular improvement. If later evidence remains incomplete, retain the verified configuration milestone while leaving the commercial question open. This keeps a successful write from being confused with an unfinished evaluation and prevents an uncertain result from triggering blind re-execution.

Compare the final result with the original proposal. A group-specific operation should not become an account-wide claim, and a criterion replacement should not be described as a campaign pause. Report each established effect in a multi-step request and identify any unresolved step. A visible success banner helps the owner notice completion, while its readable target and action rows preserve what actually changed.

Record the original question, actual dates, demographic-report coverage and selected outcome definition. Include the observed comparison, current-state evidence and supported effects approved and verified. A successor should understand the decision without reconstructing every conversational reply.

Assign creative, website and measurement tasks to their actual owners with observable completion conditions. A criterion readback cannot close a landing-page investigation or demonstrate that a brainstormed advertisement was uploaded.

Preserve justified non-action and its review condition. Unknown, Mixed and different observed ratios are not automatically faults that must be removed. A documented decision to retain current settings can be the correct consulting outcome.

The consultant audit guide provides broader context. Useful household-income analysis connects bounded evidence to the client's commercial question, prepares exact supported changes where justified and keeps configuration success separate from personal assumptions and financial proof.