Explain household-income reporting before changing reach
Google Ads household-income audits for agencies: evaluate reported value without guessing customer affordability
Read income bands, market coverage and current inclusion correctly. Separate the advertising report from personal finances, and build a client decision from actual campaign evidence.
Establish the agency's income-review question
A household-income audit should answer a business question before it produces a recommendation. An agency might be investigating an inherited exclusion, explaining a change in recorded value or checking whether a premium-product campaign reaches the intended market. Those questions need different evidence and can lead to different outcomes.
Write down the chosen period, campaign population and success measure. Highest conversion value and highest return are separate rankings. The largest income-band row can contribute most of the sales value while another produces a stronger return on its smaller cost. Establish which comparison the client actually requested.
Connect the review to an agreed commercial brief. The client may value new customers, repeat orders, enquiries or a specific product range. An account-wide income table cannot decide that preference. Use the existing conversation when the objective is already clear instead of asking the client to repeat settled information.
The final answer can be an explanation rather than an edit. It can also be a request for more precise campaign evidence, retention of intentional reach or a supported proposal. Merely displaying uneven demographic results does not establish that an exclusion or audience change would improve the business.
Make the agency's decision deliverable explicit. A short statement such as investigate the existing restriction on the seasonal campaign gives the analyst a useful boundary. It prevents a reporting question from expanding into a wholesale targeting redesign that the client did not request and the evidence cannot justify.
Explain the household-income bands
The current audit presents Top 10%, 11–20%, 21–30%, 31–40%, 41–50%, Lower 50% and Unknown. These are source reporting categories. They are not currency amounts, salary certificates or individual customer records. Keep their labels intact when describing the table to the client.
Google's income-range enum reference defines percentage-range categories and an undetermined range. The application maps those enum values to its readable labels. It does not receive an exact household income alongside each performance row.
Lower 50% covers a broader labelled range than any of the individual ten-percentage-point bands above it. Do not present seven equally sized slices of the population merely because the table contains seven rows. Unknown is also an information category, not another income interval of the same width.
A comparison of raw conversion totals therefore describes recorded advertising outcomes, not equal-sized samples of households. More value in one band does not establish greater demand per household across the market. The report does not provide the complete population denominator needed for that wider claim.
Do not add invented monetary thresholds to simplify the explanation. A familiar local salary figure might sound helpful, but it changes the meaning of the report without supporting evidence. If a client needs market-income statistics, treat that as a separately sourced research question rather than filling blanks in the Ads table.
Understand what the current reader collects
Google's income-range view reference describes an ad-group-level performance resource. The current ppctor collection requests income type, the criterion's negative flag, impressions, clicks, cost, conversions and conversion value. That provides aggregate advertising evidence rather than a list of identifiable households.
The account reader folds returned rows by income category. Its campaign counterpart preserves campaign identity before aggregating categories. Choose the appropriate scope for the question. An account summary can identify an investigation, but its figures cannot simply be renamed as the results of one campaign.
Income is collected alongside age, gender and parental status through four separate demographic queries. Concurrent retrieval does not join those dimensions into one population. The existence of separate income and age totals does not establish how a particular age group performed within a particular income band.
The income audit finding is informational and has no direct universal fix. It does not assign a targeting defect merely because one band records a weak result. A client-facing interpretation should focus on the actual table and decision rather than manufacturing an issue score from the existence of demographic information.
Keep the provider fields and the agency's interpretation distinct. The returned cost is a measurement input; a statement about commercial suitability is an analytical conclusion. Show the evidence supporting that conclusion and the remaining scope limitations instead of presenting both as though the API supplied the same certainty.
Keep market eligibility explicit
Google's demographic-targeting guidance describes estimates, Unknown users and household-income availability for particular markets and campaign types. Availability is not a universal account entitlement. The agency needs to preserve applicable market and channel conditions when explaining an absent income report.
Do not assume every demographic dimension has the same geographic coverage. A campaign might have useful age or gender information while income evidence remains unavailable for the relevant population. Another account's populated table does not demonstrate that the client's selected market must return the same type of data.
A country-targeting setting and the existence of an income row are different facts. Targeting can establish the intended geographic scope, but it does not turn an unsupported or unreported demographic population into a measured one. Retain that distinction when the client asks why the expected bands are missing.
For a multi-market agency account, avoid comparing categories as though each represented an identical monetary threshold. The current row contains the provider's category, not a verified salary conversion across currencies. A useful commercial comparison may require separate market and campaign evidence before an aggregated income label can support a decision.
Keep the answer proportional to the question. A brief coverage statement usually explains an unavailable result more clearly than a long provider-country catalogue. If the client wants an exact eligibility check, verify the relevant market and campaign conditions rather than quoting a remembered universal list.
Define the report population
The account demographic card describes demographic-reportable campaigns and labels its aggregate Reported total. Its scope excludes Performance Max and App spending from this ad-group collection. A complete-looking income table therefore does not automatically describe every cost or conversion in the Google Ads account.
Additional eligibility and returned-row conditions can narrow that population. The absence of a campaign's spending from the demographic total needs a coverage explanation before it becomes an allegation of missing money. Reconcile like populations before treating differences between account and income totals as a calculation failure.
The current account query applies the requested date condition but does not explicitly restrict campaigns to enabled status. Historical activity can therefore remain relevant even when today's campaign configuration differs. Avoid saying enabled campaigns only unless the actual scoped request and returned evidence establish that boundary.
If the client requests one campaign's household-income results, use campaign evidence. If it requests the account, keep the broader reportable population visible. The same category labels appear in both views, but identical headings do not make the metrics interchangeable between these two scopes.
Preserve this population description when exporting or handing over the review. A spreadsheet titled entire account income performance can mislead another analyst even when the displayed sums are calculated correctly. The agency should simplify presentation without enlarging the underlying coverage claim.
Separate historical performance and current settings
The selected reporting period describes when the measured interactions occurred. The criterion flag describes the retrieved configuration. Those facts can coexist without contradiction: a currently excluded income band may still have historical cost from a time when it was eligible or from a different returned child configuration.
Do not infer the date of an exclusion from the first visible change in performance. An agency needs actual change evidence before connecting the configuration to a particular day. A falling conversion line alone cannot establish who changed the setting or why the account's reach changed.
Use the account's reporting dates consistently across the supporting observations. A recent demographic table and an older product report may answer different questions. If the comparison periods differ, state the difference and obtain matching evidence when the decision depends on a direct comparison.
Completed reporting dates do not guarantee that every conversion is final. Retain a reasonable measurement interpretation when recent outcomes are still developing. The useful question is whether the evidence is sufficiently settled for the requested decision, not whether a date label makes the result permanently unchangeable.
A takeover review should record the current setting and the selected historical period separately. This gives the next account manager a reproducible baseline. It also prevents a client from interpreting a present configuration screenshot as proof that the same restriction applied throughout the entire performance window.
Retain the Unknown population
Unknown means the report does not assign the interaction to one of the displayed identified income bands. It is not a low-income label, a fraud category or proof that the people lack commercial value. Do not silently move these results into Lower 50% to make the table look more complete.
Unknown can contain measured cost, conversions and conversion value. That makes it commercially relevant even when household income is not identified. A client asking for the report's overall return needs those outcomes retained in the inclusive total rather than removed from the denominator without explanation.
A known-band comparison is still useful when clearly described as that subset. It answers how the identified categories compare with one another. It does not answer the complete report's return, and the agency should not replace the inclusive total with a more attractive known-only ratio.
An exclusion of Unknown would concern eligibility, not a correction of missing demographic information. It would not reveal the incomes of the remaining people or repair measurement. Any supported proposal needs a commercial reason, an exact scope and the connected owner's approval before a live change.
When the client asks why Unknown is large, explain what is actually established. The collection shows an unidentified share within its reportable population. It does not contain a household-level investigation that can explain every interaction. Avoid attributing that share to a guessed audience identity or customer behaviour.
Interpret Included, Excluded and Mixed
The current aggregation uses the returned negative flags to produce Included, Excluded or Mixed. A category with both positive and negative indications becomes Mixed. That is a configuration observation across the retrieved rows, not a statement that a certain percentage of its cost was excluded.
Mixed can signal different underlying choices that deserve investigation. The account category alone does not identify the exact campaign and ad group responsible for every flag. Obtain the necessary child evidence before proposing a correction; do not guess the parent relationship from the aggregated row.
Recorded impressions do not override an Excluded label, and zero impressions do not prove exclusion. Performance and current settings describe different parts of the evidence. Use the actual configuration flag when explaining current eligibility rather than inventing status from the amount of traffic.
The current explanatory exclusion note is triggered by purely excluded rows, not every Mixed row. Its absence therefore does not establish that all underlying criteria are included. Read the table's status alongside the note instead of using a missing sentence as a universal clearance signal.
A client decision should name the real target after discovery. An account-wide income band is not necessarily one editable criterion. When the intended effect concerns a particular child, preserve its campaign and ad-group context so the owner can recognise the proposed scope before approving execution.
Distinguish missing rows and measured zero
The current table renders a canonical set of income labels. A category absent from the returned map can receive fallback zero metrics and a dash status. That presentation keeps a stable layout, but it does not establish that Google returned a measured zero for the missing category.
The distinction matters when diagnosing an exclusion. A fallback row cannot prove the campaign had no eligible people in that band. Nor can it establish that a target was intentionally removed. A stable set of headings is a display convention, not a complete inventory of measured observations.
A successful response with no rows and a failed request are also different outcomes. Preserve which occurred before offering an explanation. Neither should be translated into a confident claim that every household-income band generated zero traffic across the client's account.
Unexpected returned category keys create another reconciliation question. Actual aggregation can include evidence beyond the familiar displayed labels. Check the retained population when a visible table does not explain the total rather than immediately forcing the difference into Unknown or altering the source measurements.
Use clear client language for these boundaries. The report did not provide an observation for this band is more accurate than nobody in this band converted when only a fallback row exists. Precision here prevents an attractive table from becoming an unsupported targeting recommendation.
Reconcile totals and highlighted leaders
The current report total aggregates actual returned rows, including Unknown. The summary's leading identified category is selected after Unknown is filtered out. Those two features describe different populations. A highlighted band is not necessarily the largest contributor across every row included in the Reported total.
The summary also uses absolute conversion value or conversion count, according to the account presentation. Most value does not mean best return. Most conversions does not mean lowest acquisition cost. The agency should read the relevant ratio before translating a volume highlight into an efficiency recommendation.
For a client who wants the most efficient band, state the chosen metric and compare valid denominators. For a client who wants the largest contribution, keep the volume ranking. Both are legitimate questions, but neither should be answered with a label intended for the other.
Reconcile sums from the actual rows before publishing an apparently inconsistent headline. If the visible canonical categories do not account for every returned key, explain or inspect that boundary. Do not manufacture a balancing category merely to force the presented table to resemble a separate account report.
The useful summary names both the observation and its scope. For example, an identified band supplied the most recorded value within the reviewed demographic population. That wording keeps the commercial fact understandable while avoiding an unsupported statement about the highest-value households in the entire customer base.
Calculate meaningful aggregate ratios
Calculate return from summed conversion value divided by summed cost for the chosen population. Do not average the individual bands' return percentages to represent the total. Bands carry different costs, so a simple mean can give a small row as much influence as the largest spending category.
Similarly, acquisition cost comes from total cost divided by total conversions. A row with no conversions does not have a zero acquisition cost merely because the denominator is missing. The current presentation uses a dash where the required ratio cannot be computed meaningfully.
Preserve fractional recorded conversions in calculations. Display rounding should make the result readable without turning every contribution into a whole-number event. A rounded table can differ slightly from adding its displayed numbers; calculate with the underlying retained metrics and explain material differences if necessary.
The audit presents return as a percentage. A value-to-cost ratio of three therefore appears as 300%, not 3%. Keep that convention consistent in the narrative. If another surface uses a multiple, make the unit clear instead of allowing the same performance to look like two incompatible results.
Use the connected account's verified currency in a client deliverable. A compact money symbol in a rendering is not evidence of the currency itself. The agency should not compare market amounts or produce a financial recommendation from an assumed currency based only on the table's visual formatting.
Assess commerce value and margins
Conversion value is recorded advertising value under the account's measurement configuration. It is not automatically net margin, collected cash or a settled financial statement. Household-income reporting inherits that same definition. A high-return band can still purchase products with different margins from another band's orders.
For a premium catalogue, investigate which offers and campaigns contribute the observed value. A higher-priced product can raise recorded value without proving that an income category is inherently more profitable. The current income table does not contain the full order and margin evidence needed to settle that question.
Distinguish acquisition from retention where the client's objective requires it. A category's strong value might reflect repeat customers, a promotion or branded demand. The demographic aggregate alone cannot identify all those causes. Use separately available evidence rather than attaching a new-customer interpretation to every conversion.
Discuss any commercial assumption with its owner. The media agency can explain the advertising metrics, while the client may need to supply margin, fulfilment or product-priority information. This division keeps the analysis actionable without pretending the demographic API includes confidential business facts that were never provided.
An appropriate conclusion can retain broad reach while requesting a product-level investigation. That is a concrete analytical outcome, not indecision. It follows the evidence when the reported income distribution is clear but the commercial reasons for an exclusion have not been established.
Assess lead quantity and commercial quality
A lead-generation account commonly evaluates conversion rate and acquisition cost in this audit. Those metrics describe recorded actions. They do not independently establish whether an enquiry was qualified, affordable to serve or likely to become a customer merely because it belongs to a particular reported income band.
Keep the client's lead definition attached to the analysis. A call, form submission and later offline outcome may represent different stages. The income report's aggregate conversion number needs its actual measurement scope understood before the agency presents it as a count of unique prospects or completed sales.
When the client supplies CRM quality evidence, preserve that source and population separately. A reliable connection can support another analysis, but similar totals do not prove that every lead was matched. Avoid retroactively assigning household-income categories to CRM records from an aggregate advertising table.
A cheap lead category is not automatically the best commercial target. It may contain the wrong enquiry mix, or the client may value a more expensive category's downstream results. Obtain the relevant business evidence instead of equating the lowest reported acquisition cost with the greatest customer value.
The agency can still provide a useful immediate answer: explain recorded lead cost, identify the strongest evidence and name what remains needed for the commercial decision. This is clearer than inventing a quality score or recommending that the client exclude a broad income category from a shallow comparison.
Check the conversion definition
The collection requests the selected conversions and conversion-value fields. It does not independently convert them into a verified purchase-only total or a unique-customer count. The agency needs the current account and campaign measurement context before assigning a more specific commercial label.
Where the client asks about primary conversions, retain the actual primary and campaign-goal semantics of the reporting request. Do not replace a broader aggregate with one preferred action without disclosure. A named conversion question needs evidence for that action rather than a guessed share of the income total.
Lead types should remain leads for the client's aggregate comparison unless the request selects a specific event or action. Calls and offline outcomes can also contribute recorded conversions and value under their configuration. Their presence does not permit the assistant to infer a separate household-level purchasing identity.
A configuration review and a deduplication investigation are different tasks. The income table does not diagnose whether multiple actions recorded the same business outcome. If the client raises that issue, examine the relevant measurement implementation rather than silently subtracting contributions from the demographic aggregate.
State the chosen definition once in the report and carry it into the conclusion. This prevents an agency from describing the first paragraph as conversions, the chart as purchases and the recommendation as customers when the available data supports only the original recorded-action measure.
Investigate product and campaign mix
An account-wide income result combines the campaigns that contribute to the reportable population. Those campaigns can have different products, geographic reach, creative, budgets and conversion settings. A weak aggregate row does not establish that its income category caused the weaker result.
Break down the investigation where actual scoped evidence is available. A seasonal campaign may explain an apparent account-wide shift, while an evergreen campaign continues to perform as expected. The agency should identify that concrete context before recommending an account-spanning targeting response.
Product pricing can also influence the mix. A campaign promoting one expensive item differs from another promoting an entry-level range. Recorded value from those campaigns does not create a clean experiment comparing equal offers across income bands. Preserve the offer differences in the client explanation.
Creative hypotheses remain useful, but describe them as hypotheses. The agency might propose a clearer product-benefit message or a different offer presentation. The existing income table does not establish which exact headline a category prefers, and it should not be presented as a creative testing result.
Use the smallest additional read that resolves the actual question. Repeatedly collecting every account report does not improve an answer when the missing fact concerns one campaign's current setting. A focused investigation helps the agency stay efficient while retaining enough evidence to justify the next decision.
Avoid unsupported financial assumptions
The reported category does not expose an individual's salary, savings, creditworthiness or willingness to pay. The agency should not turn an aggregate income label into a personal affordability assessment. The current fields do not support identifying which specific customer can purchase a particular product.
Household circumstances and purchasing decisions are also different concepts. A person can buy for someone else, use a business budget or prioritise a product despite the campaign's demographic estimate. These possibilities illustrate why the report is not a complete explanation of customer behaviour; they are not evidence about any particular user.
A premium positioning brief can inform creative and commercial objectives without becoming a rule that every lower labelled band must be excluded. The client should evaluate actual campaign outcomes and intentional reach. A category name by itself does not establish a waste amount or an unsuitable customer population.
Keep language respectful and evidential. Say the reviewed category recorded a certain return, rather than describing people as unvaluable or unable to buy. This produces a more accurate client report and avoids substituting a stereotype for the measurements actually returned.
If the business genuinely needs another financial analysis, define its lawful source and decision separately. Do not treat the absence of personal financial information in this API as a blank the assistant should fill by inference. The advertising review can remain commercially useful within its real evidence boundary.
Decide whether the evidence supports intervention
Consider the amount and stability of the evidence before recommending a reach change. A category with a few interactions can produce an extreme ratio that changes materially with one later conversion. The table provides observed metrics, not a built-in guarantee that every difference will persist.
Keep the cost exposure attached to the ratio. A weak-return row with very little cost has a different practical significance from one consuming a large share of the report. The client may need an explanation, a watch item or a campaign investigation rather than the same exclusion response in every case.
Compare the chosen period with the business calendar. Promotions, stock changes and delayed outcomes can affect the observed distribution. Check actual context where available; do not assert that an income category caused a performance shift simply because it appears in the same period.
A decision can be supported without claiming mathematical certainty. The agency should identify the evidence, intended effect and review condition. It should not invent a universal sample threshold or statistical test that the current audit never calculated merely to make its recommendation sound authoritative.
Record why no edit was recommended when that is the outcome. Retaining reach because the observed difference is explained by campaign mix is a useful client decision. The next analyst can then revisit it when relevant facts change instead of repeating the same unsupported exclusion proposal.
Respect the available segmentation
Separate age, gender, income and parental-status tables contain separate marginal observations. They cannot be multiplied or intersected to invent performance for affluent parents of a particular age. A combined claim requires a returned joint population, not arithmetic assembled from unrelated aggregate rows.
The current income collection also does not expose every product, query or customer dimension within the same row. A request for a narrower comparison needs a compatible supported report. If that evidence is unavailable, state the precise limitation instead of fabricating a segmented result from account proportions.
Listing-group exclusions describe another structure altogether. Product eligibility and household-income criteria are not interchangeable audience concepts. An agency investigating an excluded product needs the correct campaign and group tree, while this table describes demographic performance and current criterion indications.
Similarly, an audience configuration containing an income ingredient does not establish the historical income distribution of every impression it influenced. Settings and performance belong to different evidence surfaces. Keep them distinct when a client asks whether a recently created audience explains a result.
A useful assistant can still suggest the next compatible investigation. It should preserve the original question, identify the unavailable joint evidence and offer the supported scope that would help. This continues the conversation without treating an API boundary as permission to invent data.
Handle unavailable demographic collections
The four demographic queries currently share a Promise.all collection boundary. If one rejects, the whole demographic collection rejects. The outer report loader records the section failure. The agency should not claim that successful income data is independently retained through that failure when this collection does not provide that guarantee.
Loading, failed retrieval and a successful empty response need distinct explanations. An unfinished report is not a completed zero-result audit. An unavailable source is not proof that the client's targeting settings are wrong. Keep the actual read outcome visible when diagnosing a missing table.
A focused read retry can be appropriate for a temporary failure, but it does not justify silently repeating an uncertain live edit. Reporting recovery and mutation reconciliation are different operations. The agency should retain that distinction when an investigation later produces a proposed setting change.
Do not export placeholder zero rows as if they were verified results after a failed collection. The client may use the file outside the application, where the loading or failure context is no longer apparent. A concise unavailable explanation is more useful than a complete-looking but unsupported financial comparison.
When the source becomes available, resume the original question with the verified scope. There is no need to ask the client to recreate the whole commercial brief. Preserve the requested campaign, period and measure while obtaining the missing evidence required to finish the analysis.
Work through an invented agency review
Consider the invented agency Stone Harbour Media and its invented client Maple Court Gifts. The client asks whether the seasonal premium campaign should stop reaching the Lower 50% category. These names and the following amounts are illustrative; they are not a report from a connected customer account.
The analyst first confirms the selected period and the campaign's actual reportable population. The account table highlights a higher band by recorded value, but the client asked about one campaign and a potential restriction. The analyst obtains that campaign evidence before treating the account highlight as a targeting instruction.
Suppose identified categories together record 400 in cost and 1,200 in conversion value, while Unknown records 200 in cost and 200 in value. Identified-only return is 300%; inclusive reported return is 233.33%. Both calculations can be correct because their populations differ, and the report must label them accordingly.
The campaign's income status also shows Mixed for the category under discussion. That does not identify a single incorrect exclusion. The agency investigates the actual child configuration and finds an inherited difference that needs its original rationale checked. It does not infer that half the category is excluded from the word Mixed.
The agreed immediate outcome is to retain reach and clarify the commercial brief, then review any supported correction with the owner. No invented amount becomes a live recommendation. The example illustrates a disciplined decision sequence: define the question, retrieve the right population, reconcile the evidence and prepare only the justified next step.
Choose an appropriate practical response
An income review can end with several practical outcomes. The agency can explain the observed contribution, investigate a campaign, retain an intentional setting or prepare a supported change. It should choose the response that follows the evidence rather than treating every demographic table as an exclusion opportunity.
A creative suggestion can be entirely conversational. The assistant can draft a clearer benefit statement from the supplied product context without querying income performance or claiming that a particular band prefers it. Live metrics are needed for factual reporting, not for every brainstorming request.
The current audit does not expose a dedicated universal income fix in its convenience actions. A schema-mediated edit needs the actual resource, field and applicability established before it can be presented as executable. Do not promise a demographic operation merely because the account has a readable income report.
If the requested effect is unsupported, explain that specific boundary and retain the useful findings. A limitation in the editing surface does not make the reporting answer worthless. It also does not authorise substituting an unrelated audience operation that changes something different from the client's request.
For any proposed change, describe what remains unchanged as part of the actual scope when it matters to the decision. The owner should understand which campaign or group is affected and whether other existing criteria are retained. Precise effect descriptions prevent a broad discussion from becoming an accidental account-wide edit.
Distinguish audience signals and exclusions
The audience builder can include a household-income ingredient in a supported audience configuration. That capability is different from editing an existing demographic criterion. An income ingredient should not be described as removing all unwanted income bands from a campaign unless the actual operation establishes that effect.
Creating an audience alone leaves it unattached. Creating an audience signal can create and associate the audience with the selected asset group. Attaching an existing supported audience retains other existing signals. These are distinct operations and should appear distinctly in the owner's proposed action.
The current association helper treats an already attached audience as a no-op. That avoids claiming a new change where the intended association already exists. It does not establish that the campaign's demographic exclusions have been modified or that historical income performance now covers every asset-group interaction.
Validate the actual audience ingredients and provider applicability during preparation. The presence of an income-range array in a builder is not proof that every enum or campaign context has already passed local validation. The proposal needs the supported real configuration, followed by the required live preflight and verification.
Explain the operation in normal client language. The owner needs to know whether an audience is being created, attached or retained, and where. Raw dimension structures are internal preparation details. A readable scope description can remain exact without exposing technical payloads or implying an unrelated eligibility restriction.
Prepare and verify an approved change
A live edit follows discovery and preparation, then an exact owner confirmation. The proposal should name the action and target, including the relevant campaign and parent group where needed. Reporting or discussing a weak income result does not itself authorise a write to Google Ads.
Present multiple logical actions as clear steps under one confirmation when the supported request permits them. If the complete instruction cannot be prepared within the safeguards, explain why it needs smaller reviewed requests before writing. Do not silently execute only a subset while describing the entire original request as completed. The client should see the actual scope of every proposed demographic effect.
After confirmation, show visible progress while the request runs. Preflight checks current settings and applicability; verification reads back the effect that was actually requested. These stages answer different questions. A provider acceptance alone is not the same as confirming the resulting configuration.
If dispatch succeeds but readback remains uncertain, say so accurately and reconcile the focused result. Do not automatically repeat the mutation merely because the final Slack message was delayed or the verification read failed. Replaying an ambiguous write can create a second unintended effect.
A verified setting change confirms the technical result, not future performance. The agency still needs a later review using an appropriate reporting window and actual business outcomes. Keep the execution status prominent and the subsequent commercial assessment separate so the client can understand both clearly.
Keep client reports readable and accurate
Start the client response with the requested conclusion and the key evidence. A concise explanation of the reportable population, chosen metric and relevant bands usually helps more than an unfiltered technical payload. Add deeper detail where it supports the decision rather than reproducing internal query structures.
Keep metrics readable, with no more than two displayed decimal places. Use clear currency and ratio labels and preserve calculations from unrounded underlying values. Do not create apparent precision by displaying long decimals that the client cannot use to judge a practical targeting decision.
Show asset names in client-facing edit outputs rather than raw resource IDs. Parent context should reflect the real hierarchy: an ad group belongs to a campaign, not another ad group. If discovery has not established that relationship, obtain it before writing a misleading proposal or execution summary.
A table, chart or export should use the same population and conversion definition as the narrative. Moving to a larger file does not permit Unknown to disappear or fallback rows to become measured observations. Keep the evidence boundaries consistent across Slack and any shared report.
Separate internal failure diagnostics from the client explanation. The user needs to know whether the data was unavailable or the execution was uncertain, but not every internal reference or provider stack. Clear outputs can remain truthful while withholding implementation details that do not help the client's decision.
Continue the conversation and hand over the review
The Slack assistant should preserve the campaign, period and metric already established in the thread. A follow-up asking why a band looks weak should continue that scope. It should not discard context and demand the same campaign name again when the retained conversation provides an unambiguous target.
Use fresh evidence when the question requires current settings or a new reporting population. Reuse compatible retained results when they already answer the follow-up. This reduces unnecessary API work without inventing freshness or treating a previous account-wide report as every possible campaign comparison.
End the agency handoff with the actual decision, its evidence and the next owner. Record retained reach, unresolved commercial facts or the exact verified edit as applicable. Another account manager should be able to continue the review without reconstructing which income bands, period and campaigns were discussed.
Keep client-approved rationale beside an existing exclusion when available. A future analyst may see weak recent performance and propose a different response, but it should first understand the intentional business context. Configuration history, current performance and owner intent complement one another; none replaces the other two.
A strong household-income audit leaves a narrower, clearer claim than a guess about customer affordability. It establishes what the report measured, how the actual settings relate to that evidence and what the agency will do next. That gives the client a useful decision while respecting the limits of the available information.