Investigate the offer before restricting its audience
Google Ads household-income audits for freelancers: interpret percentile groups without guessing who can buy
Check report coverage, country eligibility, Unknown and criterion status. Connect income-group performance to a defined business question, then prepare supported changes with clear targets and owner approval.
Start with the owner's commercial question
A household-income audit should begin with the decision the client needs to make. A freelancer may be asked whether an expensive product reaches suitable buyers, whether an inherited exclusion is sensible or whether the reported groups contribute differently to purchases. Those questions require different evidence. Write the question in business terms before choosing a threshold. The table can describe reported advertising outcomes, but an income label alone cannot decide which customer deserves an offer or which audience will produce the best margin. The purpose of the review is to turn the available evidence into an appropriately scoped next action.
Ask what prompted the concern. The owner may be reacting to a high acquisition cost, a sales conversation or a previous agency's audience recommendation. Preserve that context without accepting the explanation as proven. A campaign selling a premium item could have a weak landing page, a poorly defined conversion or a promotion that changes the normal order mix. Income segmentation is one possible analytical lens. A freelancer should establish which observation the report can test and which part of the concern needs separate commercial or measurement evidence from the business.
Agree on the deliverable before discussing implementation. A report, an offer hypothesis and a live criterion change are separate tasks. The client can request an explanation without authorising an exclusion. If the engagement includes edits, the eventual proposal should name the campaign or ad group, state the actual effect and require confirmation before execution. This keeps the review useful for an owner who wants clarity first. It also gives the freelancer an explicit record of the question, the evidence supplied and the decision that was approved.
Read the seven reported groups
The current income table uses Top 10%, 11–20%, 21–30%, 31–40%, 41–50%, Lower 50% and Unknown. Read those as the provider's household-income groups. They are not seven salary amounts, seven price sensitivities or seven customer personas. The lower half is one broad group, while the upper half is split into smaller bands. That uneven grouping matters when comparing volume. A larger contribution from Lower 50% does not establish stronger efficiency, just as a smaller top band does not establish that the campaign should avoid it.
Keep the labels intact in a client report. Renaming Top 10% to wealthy buyers or Lower 50% to customers who cannot afford the product adds a business claim that the returned data does not support. A household classification also differs from the spending decision made by an individual purchaser. The person may buy for someone else, save for a particular purchase or act on behalf of a business. The freelancer can investigate those possibilities through the offer and client knowledge without assigning motives to everyone in a reported group.
A useful explanation names the group, period, measure and reporting population together. For example, discuss the purchase value reported for a percentile band in a defined campaign, rather than announce a universal affluent-customer result. Do not translate a percentile into a dollar cutoff without a separately established source and scope. The card itself supplies no verified salary boundaries. Preserving that limitation makes the analysis more precise and prevents the owner from treating a report classification as a personalised financial assessment of the people who saw the ads.
Check country and campaign eligibility
Google documents household-income options by campaign type and supported country. The eligible lists are not a promise of worldwide reporting. Its current guidance lists countries for Search, Display, Video and Demand Gen, with differences between the lists. Check the actual requested country and channel before expecting a table or offering a targeting edit. Google's demographic-targeting documentation is the reference for that provider eligibility. A freelancer should preserve the distinction between an available feature and a dimension that produces rows for this particular account and period.
Do not infer eligibility from the account's billing address, interface language or currency. The client's advertising activity may target another market, and the relevant campaigns may use different channels. Establish the country and campaign context that answers the request. If the business serves several markets, report which selection was reviewed. A country selector should reflect the intended scope rather than silently stand in for every location associated with the account. The owner should understand whether the result describes one market, a campaign population or a broader returned aggregation.
An unavailable income breakdown needs a scoped explanation. It does not establish that the business has no customers in a particular band or that every campaign is misconfigured. State whether the relevant read failed, whether the chosen scope lacks rows or whether provider eligibility prevents the requested analysis. A freelancer can then suggest a different available measure for the business question. Preserve useful campaign and conversion evidence without inventing the missing income field. The next step should address the actual limitation rather than trigger an unnecessary account-wide audience rebuild.
Identify the reportable population
The account reader obtains income rows from the income-range reporting view. It requests the criterion type and negative flag alongside impressions, clicks, cost, conversions and conversion value for the selected dates. The account card calls its total Reported total. Its caveat describes demographic-reportable activity and explicitly excludes Performance Max and App spend from this ad-group demographic-view coverage. Treat the table as the population it actually measures. A strong income result in the returned rows is not automatically the same result for all advertising spending in the account.
The current query applies a date condition but does not filter campaigns by their current enabled status. A descriptive code comment should not change that fact. A campaign paused today can still contribute historical activity during the requested period. A freelancer should explain that distinction when the owner sees a familiar paused campaign in a follow-up analysis. Current eligibility, historical delivery and report coverage answer different questions. If the owner requests only a specific currently enabled population, that population must be established and labelled through the appropriate scoped reporting path.
Coverage can differ between demographic dimensions. A returned age or gender table does not guarantee an equivalent income table, and a broad campaign-type caveat is not evidence that every campaign contributes to every dimension. Preserve the returned scope when preparing a client summary. If the owner compares the income total with the ordinary account spend total, explain the difference before drawing a conclusion about missing money. The report can be internally useful even when it is not a complete account ledger. That is a boundary to communicate, not an excuse to disguise the totals as interchangeable.
Keep inferred income separate from customer knowledge
Google explains that demographic information can be identified or estimated and cannot be established for everyone. The reported household-income group is therefore an advertising classification, not a verified statement of an individual customer's finances. A freelancer should use it to describe returned performance and supported targeting behaviour. It does not authorise claims about a person's salary, savings, creditworthiness or willingness to pay. The client may possess separate business evidence about actual purchasers, but that evidence needs its own definition and should not be silently merged with an inferred group.
Be particularly careful when the owner presents a familiar customer story. One buyer's occupation or postcode does not validate the meaning of every row in the report. Likewise, a few memorable expensive orders do not establish the financial circumstances of all customers who converted. The freelancer can ask what the business knows about order value, repeat purchases and completed sales. Those facts help define a commercial outcome. Keep the distinction between observed business records and the advertising provider's classifications visible throughout the analysis and eventual recommendation.
The practical consequence is modest language. Describe what the returned band contributed, how its measured efficiency compares and what remains unexplained. Avoid turning the audit into a financial portrait of the audience. A thoughtful recommendation might investigate a payment-message hypothesis or an acquisition-cost concern within one campaign. That can be valuable without asserting why each person bought. Maintaining this discipline also improves future handoffs: the next freelancer can inspect the same data and assumptions instead of inheriting an unsupported audience stereotype disguised as an established account fact.
Retain the meaning of Unknown
Unknown represents activity for which the provider has not established the relevant classification. It is not another income percentile and should not be assigned to the lower or upper bands to make a chart look complete. Keep its reported cost, conversions and value visible when reconciling the table. The current total includes it, while the headline comparison of known groups excludes it. A freelancer should explain those different populations when the owner asks why a highlighted band does not account for the same share suggested by the total.
Hiding Unknown in an analytical view and excluding Unknown from targeting are different actions. The first changes what the owner sees; the second can change who is eligible to receive ads. Do not allow a request for a cleaner comparison to become a live reach restriction. Google cautions that excluding Unknown can remove people the advertiser may want to reach. The freelancer should keep that provider limitation connected to the specific proposed action. A report can focus on known groups while still stating that unclassified activity remains part of the campaign's outcomes.
When Unknown is commercially important, investigate it as a reporting population rather than invent a demographic explanation. Compare the same requested conversion measure and period, review the scope and retain the difference between volume and efficiency. The owner may decide to keep the current settings while exploring another campaign-level hypothesis. That is a legitimate outcome. An income audit does not need an exclusion to demonstrate usefulness. Its value can be establishing that the available classifications are too incomplete to justify a proposed restriction, while identifying a better-supported next task.
Interpret Included, Excluded and Mixed
The income renderer derives status from returned criterion flags. It does not infer exclusion because a group has no impressions. Included means positive criterion evidence was returned for that aggregated group; Excluded reflects negative evidence; Mixed means both kinds occurred among the combined rows. These labels describe the returned settings evidence at the table's scope. They are not customer behaviour labels. A freelancer should review the actual campaign and ad-group context before turning an account-wide status into a proposed edit.
Mixed is especially important in accounts with different offers or inherited structures. It can represent an intentional variation between groups, rather than a broken setting that needs to be made uniform. The table has compressed several resource occurrences into one band. Ask which campaign or ad group the owner wants reviewed and obtain that configuration. An edit needs an exact resource and effect. The percentile row itself is not a single mutable object, so a blanket instruction to fix Mixed would conceal the actual targets and potentially override valid campaign decisions.
Current status and selected-period metrics also have different meanings. Historical activity can remain visible after a setting changes, and an excluded row is not proof that its displayed performance was generated under one unchanged configuration throughout the period. A freelancer should check relevant change timing when that distinction affects the recommendation. The owner-facing explanation can remain simple: identify the current state, the historical window and the evidence used for the next decision. Do not promise a complete eligibility history from a compact status label.
Separate zero, missing and unavailable
The table renders the canonical income bands even when a particular key is absent from the returned map. Missing numeric values appear as zero and the missing status appears as a dash. A freelancer should not read that fallback as proof that the provider evaluated a confirmed included audience and measured no demand. Establish whether an actual row was returned. The visual consistency of the table is useful for navigation, but the interpretation must preserve the difference between a returned zero and a placeholder for absent evidence.
The account demographic reader requests four dimensions concurrently. If one query rejects, the combined demographics read rejects; the surrounding report handling can then show that the demographic evidence could not be checked. That is different from a successful response containing no income rows. Other independently collected reports may remain useful. A freelancer should state the unavailable read precisely and avoid declaring that the account has no income performance simply because collection failed. A failure belongs in the evidence record, not in a fabricated zero-value result.
Use proportionate follow-up. If a failed demographic read prevents the requested conclusion, retry through the established read path or investigate the actual failure. If a successful empty scope is expected because of eligibility, explain that boundary. Do not call every unrelated tool merely to fill the card. The owner needs an honest account of what is known and what task remains. Keeping zero, missing and unavailable separate also prevents a bad live edit: there is no basis to exclude a band just because its display was populated by a fallback.
Reconcile totals and summaries
The aggregate row sums the returned metric bases across the income map, including Unknown. Ratios should be understood from those summed bases, rather than as an average of the visible groups' percentages. A freelancer should reconcile cost, conversions and value first, then explain the requested ratio. If an unexpected provider key contributes to the aggregate but is not one of the canonical displayed bands, the total's underlying population still matters. Do not force an artificial reconciliation by deleting an unexplained contribution from the source data.
The compact summary uses the canonical known groups and highlights absolute value or conversion contribution according to account type. A group that produced the most value is not automatically the group with the best ROAS. A group that produced the most conversions is not automatically the group with the lowest CPA. The freelancer should answer the owner's exact question rather than promote the highlighted contribution into an efficiency recommendation. The necessary base metrics are available for a more deliberate comparison when the requested population and outcome are established.
The summary's exclusion note also has a narrow implementation meaning: it checks for a canonical group with an explicitly excluded status. Mixed is not the same condition. Absence of that note should therefore not be treated as a certificate that every underlying criterion is included everywhere. Use the table to identify questions and the actual target's configuration to answer them. This keeps the concise card helpful without asking it to prove more than it contains. The freelancer can preserve readability and still state the exact limit of the evidence.
Avoid false comparisons across markets
Percentile labels are relative income groupings, not a common monetary scale. A freelancer comparing campaigns in different countries should not assume the same label establishes equal purchasing power or an identical customer proposition. The table supplies no verified cross-country salary conversion. Start with the market each campaign serves, the offer currency and the business conditions. If the question concerns one market, scope the analysis accordingly. If it concerns several markets, keep the difference visible instead of describing a pooled band as a universal premium audience.
Campaign mix can further complicate the comparison. One country's campaign may advertise a subscription while another promotes a one-time package, with different budgets and conversion maturity. A higher reported value in one band may reflect those choices. The freelancer should compare like business questions before explaining demographic differences. This does not require an enormous exploratory account export. Obtain the market and campaign context relevant to the owner's decision, then identify which metrics are genuinely comparable. A small well-defined comparison is often more useful than a broad ranking with hidden differences.
Retain the account's established currency for financial interpretation. The compact demographic display uses a dollar symbol, which is not independent proof of the account's currency. A freelancer should check the actual account context before combining figures or presenting a financial conclusion. Do not add monetary values from different accounts as if they were one spend ledger. Country scope, currency scope and income classification are separate concerns. Stating each clearly allows the owner to understand a cross-market review without being given an invented conversion of demographic bands into actual household wealth.
Choose a defensible reporting period
The period should fit the campaign's decision cycle. A very short interval may contain little outcome evidence, while a long interval can merge several offers and configurations. Establish the actual start and end dates in the account's reporting context. If the owner asks for the last thirty days, preserve that request rather than silently substitute last month. A freelancer should also identify whether the current partial day belongs in the analysis. The meaning of the window should be explicit before any comparison becomes a recommendation about audience eligibility.
Review changes that affect interpretation. A price promotion, a conversion-definition change or a new destination can alter the measured outcome independently of income settings. Campaigns paused today may have delivered during the window, and recent changes may not have enough mature results to evaluate. The freelancer should record the events that matter without turning every audit into a complete history export. This makes the next conversation easier: when the owner asks why a group changed, there is an established period and a relevant set of contextual changes to examine.
Comparison windows need the same scope and measure. A known-group-only result should not be compared with a previous period's total including Unknown without explanation. Likewise, a campaign-specific view should not be compared with an account-wide view as if the difference were a trend. A freelancer can label a deliberate scope change, but should not hide it inside a percentage. If the available data cannot support the requested comparison, state what is missing and request the appropriate read. Date precision is part of the analysis, not decorative information at the bottom.
Define the conversion outcome
The current income reader requests the standard conversions and conversion-value metrics. Do not relabel them as purchase-only outcomes simply because the account sells products. Establish which primary actions contribute to the requested measure. If the owner asks about all primary conversions, retain that definition. If the question concerns purchases, leads, calls or one named action, select and label the appropriate supported reporting scope. An income comparison becomes misleading when different business outcomes are silently counted as though they all represent the same sale.
Revenue language deserves the same care. Conversion value can include values assigned to non-purchase primary actions. The freelancer should identify whether the question concerns standard primary conversion value or a specifically defined purchase measure. Do not use the demographic table to repair deduplication choices that the owner has not requested. Explain the selected measure and keep the intended primary-versus-secondary boundary. This allows a campaign decision to be assessed against the owner's actual outcome, rather than an improvised revenue definition chosen because a value column was available.
A comparison with Analytics requires a matching scope. For leads, establish the corresponding site event instead of treating every site as if it uses the same event name. For total site revenue divided by advertising cost, describe the requested measure as blended ROAS or MER. The income audit does not itself return a cross-source demographic join or an individual-customer match. A freelancer can use another supported report to investigate a defined business question, but should not present two differently scoped sources as interchangeable evidence about the same classified users.
Compare volume and efficiency separately
Contribution answers how much of the returned outcome a band produced. Efficiency answers how that outcome relates to its measured cost or other base. Both can matter, but they are different questions. A freelancer should show the base metrics needed to interpret the selected ratio. A band with a small cost and one valuable conversion can have an impressive ratio without representing a dependable source of scale. A band with the largest value can still have a less attractive return. Avoid ranking audiences by whichever number happens to make the clearest headline.
For a commerce account, the current card includes value and ROAS; for other accounts it includes conversion rate and CPA. Read the labels and understand how they are displayed. A percentage-formatted ROAS should not be copied into another context as if it were an unlabelled multiplier. Undefined ratios with missing denominators need a dash or a clear explanation, not an invented zero-efficiency verdict. Conversion values can also be fractional in reporting. Presentation should remain readable without pretending the rounded number establishes an integer count of unique purchasers.
Discuss scale carefully. Restricting a campaign to a band that performed well historically does not prove that its past efficiency will hold after the population and delivery conditions change. A freelancer should frame a proposed restriction as a scoped decision with an observation plan. Retain the possibility that the better next step is improving an offer or clarifying measurement. The audit's purpose is to choose a justified action, not to promise that the largest ratio can be multiplied into unlimited profitable growth.
Bring actual business economics into the decision
The advertising report does not contain the full economics of the business. Gross conversion value may differ from margin, realised revenue or the value of a completed sale. A freelancer should ask which business measure matters for the current decision. If the client has established return rates, fulfilment costs or repeat-purchase considerations, use them as clearly labelled business context. Do not infer those facts from household income. A reported band can be commercially interesting without proving that its buyers are profitable after every cost is considered.
Lead generation needs similar discipline. A low CPA can reflect many recorded enquiries that do not become customers. The owner's sales process may provide separate evidence about completed outcomes, but the compact income card does not establish that quality by itself. Agree how the client wants to assess the result and avoid replacing the requested primary conversion count with an invented qualified-lead definition. If offline or named-action evidence is needed, use the appropriate available reporting path. Keep the resulting population and timing clear in the final recommendation.
Use commercial context to sharpen the hypothesis, not to stereotype the bands. If an instalment offer changes order completion, investigate the actual offer and outcome. If acquisition costs exceed the owner's target, compare the defined measure before discussing exclusions. The freelancer can explain which evidence supports the decision and which remains a business assumption. This makes the income audit useful even when the final recommendation concerns measurement, messaging or a narrower campaign test rather than a demographic setting.
Investigate price, payment and offer context
A premium product does not automatically require a top-income-only audience. The offer may include entry products, gifts, bundles or payment arrangements that change the purchasing decision. A freelancer should review what the campaign actually communicates and where it sends people. The income table cannot establish whether visitors understood delivery costs, payment terms or the differences between packages. Those are specific journey questions. Investigating them can reveal a practical task without turning a broad percentile classification into a claim about every visitor's ability to pay.
Creative work can remain conversational. The owner may ask for a clearer headline, a package explanation or a less aggressive payment message after reviewing the report. The assistant can propose ideas from supplied context without making unnecessary provider calls. Account-specific performance claims and live creative edits still need their appropriate tools. A freelancer should keep these stages separate: brainstorm a response to the hypothesis, select a useful version and prepare any supported live change for review. The initial income audit is context, not automatic permission to rewrite ads.
Do not change price or payment terms merely because a band has weaker measured performance. Those decisions belong to the business and may affect every customer, not just a targeting group. The freelancer can recommend a focused investigation and document the evidence needed. If a campaign message is changed, evaluate its actual destination and commercial outcome separately from the provider's configuration readback. This produces a more useful engagement than an exclusion chosen because the product is expensive and one band sounds less likely to purchase.
Review campaign mix before narrowing reach
An account-wide income aggregate can combine brand demand, prospecting and several product lines. A freelancer should not assume a band's overall result applies equally to each campaign. The owner may want to review a particular offer, while another campaign contributes most of the account's returned conversions. Resolve the campaign that matches the request. Then compare the appropriate measure and period within that population. This preserves the business question and avoids an account-wide restriction driven by activity from a different offer.
Mixed criteria can be meaningful in this context. One ad group may have a deliberately broad audience while another has an owner-approved limitation. The aggregate status does not explain those decisions, but it identifies where a scoped configuration read is needed. A freelancer should ask whether the client wants consistency or wants to preserve distinct campaign roles. Do not make every group identical because uniformity is easier to report. A useful audit records the reason for the difference and only proposes changes that address an actual current objective.
Keep channel boundaries visible. Performance Max uses different resource structures from the ad-group demographic views used by this card. An absent income-performance row is not evidence that no audience-related resource exists in the campaign. Equally, a resource that contains income ingredients is not proof of the same reporting breakdown or hard eligibility control. Select tools by the actual resource, relationship and intended effect. That prevents the freelancer from forcing one account report to answer every channel's reporting and editing questions.
Make the proposed test proportionate
A good test begins with a specific uncertainty. The freelancer may want to establish whether a campaign's exclusion still reflects the owner's offer strategy or whether an apparent efficiency difference persists under a consistent measure. State the hypothesis and the evidence that would change the decision. The report should support the test's scope rather than merely decorate it. A small number of outcomes may justify continued observation or a message investigation, while a mature defined comparison may support a more concrete proposal.
Avoid invented universal thresholds. The report does not supply a rule that every band with a particular CPA must be excluded after a fixed number of clicks. The appropriate decision depends on the owner's target, conversion meaning and consequences of reduced reach. A freelancer can agree a commercial criterion with the owner and preserve it in the review. If the evidence is insufficient, say precisely what observation is needed next. That is a useful conclusion when it prevents a confident but poorly founded setting change.
Choose one interpretable intervention where possible. Simultaneously changing income eligibility, creative and several campaign settings can make later outcomes difficult to explain. The app's limit on logical actions helps keep a proposal readable, but does not establish an experimental design by itself. The freelancer should decide which change answers the hypothesis and which tasks belong in another request. Document what remains unchanged because it is part of the test's context, then evaluate the intended outcome without attributing every later account movement to one demographic edit.
Establish the exact supported edit
The income audit is informational. It does not provide a universal automatic income-performance fix. A freelancer should establish the exact operation through the current editing framework before offering execution. Google documents income-range criteria at campaign and ad-group levels, with support constrained by the relevant resource and campaign context. Google's criteria reference describes those provider levels. A type appearing in that reference is not a guarantee that every requested combination can be prepared by the current application.
Name the intended effect. Adding an exclusion, removing a criterion and modifying a supported bid setting are different operations. The freelancer should not substitute one because its tool is easier to call. Resolve the real campaign or ad group and verify the supported request shape. Demographic validation also imposes constraints on combinations; do not build a request that removes every eligible target or assumes Unknown-only targeting is an unrestricted escape route. The proposal needs to be accepted for its actual context before the owner is asked to approve a live write.
If the current framework cannot prepare the exact requested action, state that limitation clearly and preserve any completed analysis. An owner can still use a scoped recommendation in Google Ads. Do not describe a fabricated successful edit or route the request through an unrelated creation tool. The freelancer's responsibility is to connect business intent with a real supported operation. Provider capability, local preparation, approval and execution are distinct stages, and each should be established before the next is represented as complete.
Distinguish criteria from audience resources
The app can work with audience resources that contain demographic ingredients and with audience relationships used by asset groups. Those are not interchangeable with an income exclusion on a campaign or ad group. A freelancer should identify whether the owner wants a new audience, an attachment or a hard eligibility change. The same income phrase can occur in different resource types while carrying a different effect. Tool selection should follow that effect and the actual parent relationship rather than the superficial similarity of the ingredient's name.
The current audience builder checks meaningful supported dimensions, resolves ingredients and checks for a conflicting enabled name. Creating an audience without attachment has a different result from creating it with a supported signal relationship. The freelancer can keep technical resource details behind the scenes while stating the action plainly. The income performance row does not automatically identify the asset group where an audience should be attached. Resolve the parent's name from the user's intent and current account evidence before presenting the proposal.
An audience signal should not be described as a hard income boundary. Conversely, an exclusion request should not silently become signal creation. This distinction is essential when the discussion moves from the account's income report into Performance Max. The reporting scope and editing resource tree must remain explicit. A generic framework becomes useful by preserving resource types and dependencies, not by collapsing every audience-related task into a single demographic command. The freelancer should inspect that the prepared proposal expresses the owner's requested effect.
Follow an invented freelance review
Consider an invented freelancer, Owen Marsh, reviewing the invented Maple Room Lighting account. The owner sells several lighting packages and asks whether only the highest-income group should remain eligible. Owen establishes the offer, market, selected period and conversion-value definition. He explains that the income table covers returned demographic-reportable activity and includes Unknown in its total. He does not translate the bands into verified salaries or assume that a lighting purchase reveals the buyer's finances. The first deliverable is a defined review of the owner's proposal.
The owner clarifies that the concern is acquisition efficiency for one package campaign. Owen narrows the evidence to that campaign and reviews its current criteria rather than treating account Mixed as one setting. He compares the requested ratio with its underlying cost and outcome volume, then asks about the actual package message and payment terms. A creative explanation may be the appropriate next task. If the owner instead requests a supported eligibility change after reviewing the evidence, the assistant prepares that exact scoped operation for a readable confirmation.
Owen records any confirmed setting and separates its readback from the later commercial review. He keeps the original income population and measure available for a compatible follow-up. The case is an invented reasoning example, not a testimonial or a promise of improved returns. It illustrates how a freelancer can challenge an assumption constructively: define the business question, preserve the provider's classification limits and establish the real target before proposing a change. The owner receives a decision record that remains useful after the conversation ends.
Use conversation to refine the question
A useful Slack conversation can move from income contribution to campaign efficiency, then into offer ideas or a supported setting request. The assistant should retain the latest scope, period and measure. A freelancer should inspect that a follow-up such as this campaign or that group resolves to the intended resource. The framework should choose available fields and conditions for the request, rather than depend on a new phrase-specific tool each time the owner describes the same analytical need in different words.
Choose a compact output that suits the question. A table can combine reported bands, status and base metrics. A chart can show a supported comparison or trend with a clearly named measure. Do not manufacture daily observations from a single period total. When a native chart needs a coarser interval or multiple panels because of its point limit, label that choice. The freelancer should ensure that the visual helps answer the decision and preserves Unknown and reporting coverage, rather than making a partial population appear complete.
Replies should be polite and concise, with financial and ratio values presented readably. Keep raw resource IDs, technical request bodies and internal data-fetch explanations out of ordinary client output. Use campaign and ad-group names when the hierarchy is relevant. Creative suggestions can use the supplied offer context, while claims about current settings and measured performance need the appropriate evidence. This balance makes the assistant useful without unnecessary API traffic and without allowing a conversational guess to become a factual account statement or a live edit.
Present readable actions for approval
The confirmation should show the exact target and action before the owner clicks. For an ad-group criterion, name the ad group and parent campaign on clearly separated lines. State which income group is affected and what change will occur. A percentile label followed by an ambiguous arrow is not enough. The freelancer should review the proposal as the client sees it, ensuring that it contains the requested scope and excludes unrelated changes. No live mutation should be represented as already made during this preparation stage.
A supported sequence can share one confirmation, with each step visibly separated. If the complete request cannot be prepared safely, explain the need for smaller reviewed requests before execution. This is not a reason to omit the last action silently. If the owner asks for an eligibility change and another supported campaign task, show both resolved steps. Local dependency reasoning should identify the necessary order without unnecessary provider reads for tasks that do not use a particular resource. The owner should review every actual effect of the accepted proposal.
The proposal tells the owner that no changes have been made yet and that it expires in fifteen minutes. The confirmation control initiates the prepared live request, followed by visible progress and a distinct result. Current settings are checked according to the execution flow so stale assumptions do not pass unnoticed. The freelancer should not treat an earlier report as permanent permission or permanent configuration evidence. A readable proposal connects analysis to authority: the owner approves the specific change that is about to be executed.
Separate execution from commercial evaluation
Preparation establishes whether a request can be built and checked; execution sends the approved live operation; readback establishes the requested configuration where the verifier supports it. These stages have different evidence. A freelancer should use the visible executed-and-verified result only for the setting that was actually confirmed. It does not establish that the audience has become profitable, that the destination works for every visitor or that policy review has accepted new creative. Those outcomes require their own appropriate checks and later observations.
If execution returns an uncertain or incomplete verification result, reconcile the exact target before requesting another write. The provider may have accepted a change even when the subsequent read failed. A blind replay can create unintended duplicate resources or obscure what happened. Preserve the operation result and use a focused read of the current setting. The owner-facing message should name the uncertainty plainly and identify the next verification task. It should not turn an ambiguous outcome into a confident failure or claim that no change occurred without evidence.
Commercial follow-up should reuse a defined measure and appropriate later period. Compare the intended population while acknowledging reach changes and other relevant events. The freelancer should evaluate whether the owner's hypothesis received support, rather than search for any improved number after the edit. A correctly verified income setting can still be a poor business decision; a sensible decision can also need more observation. Keeping configuration verification and commercial evaluation separate lets the owner assess both without being given a guaranteed uplift claim.
Leave a useful record for the next review
The handoff should preserve the market, reportable population, dates, outcome definition and reason for the recommendation. Include Unknown's treatment and the actual scope of any criterion evidence. If a live change was approved, record the named target, action and verification outcome separately from the analytical observation. This gives the owner a durable explanation that another freelancer can follow. It also prevents a future reviewer from interpreting an inherited exclusion as a universal business rule merely because the original conversation is no longer visible.
Record the next decision point in concrete terms. The owner may need a later compatible report, a measurement clarification or an offer discussion before another edit is justified. Do not leave a vague instruction to optimise the audience. Identify the evidence that will answer the question and who is responsible for supplying business context. A concise summary can be supported by a larger report when the data exceeds a readable Slack response, but additional rows should not replace a clear explanation of the population and measure.
A strong freelance income audit leaves the client with a clearer commercial decision and an honest account of the available evidence. It preserves relative percentile labels, country eligibility and Unknown instead of inventing personal finances. It connects a recommendation to actual campaign context and a supported approved operation, then separates configuration readback from later outcomes. The table is useful because its boundaries are understood. The freelancer's work is to make those boundaries practical so the next conversation begins from established facts and a specific business question.