Check Google Ads inquiry counts Get the source data and prepare the review Documentation-checked component path; proposed workflow, not tested WHAT WAS CHECKED Public official documentation and supplied candidate artifacts only. No product login, account exports, client data, installation, credential access, external writes, advertising spend, template imports or live tests. Proposed prompts do not prove a feature supports every requested field. Missing required source data blocks execution. BEFORE YOU START - Existing authorized access to the correct source account and records; no new login, credentials, installation or permission grant is included. - Responsible lawyer/operations reviewer approves scope, definitions and any later action. - AI optional; use only approved minimized data and retain originals separately. DATA ACQUISITION LIMITS - Aggregate counts cannot prove real inquiries or deduplication. No production-event test performed. STOP IF - Missing event definition, implementation evidence, comparable windows or isolated approved test path. KEEP THE ANALYSIS COPY SEPARATE Keep untouched original exports/screenshots and the identifying source map in restricted firm storage. Build a separate minimized analysis file; do not hide raw client data on another tab. Include the complete Scope/coverage note in the AI-visible workbook or prompt. Use aliases and administrative metadata only. Unknown is not zero, false, completed or absent. INPUT FORMAT LIMIT All capitalized analysis tables and alias fields are editorial contracts. Native exports retain real vendor headers; mapping is explicit. No blank starter template is evidence of available data. NATIVE REVIEW INPUTS version: 1 status: Requires distinct native-input templates or an explicit reviewed mapping; existing kit is not drop-in compatible compatibleWithoutMapping: False existingKitFilesInspected: - file: conversion-actions-template.csv firstLine: action_id,source,event_definition,count_setting,primary_or_secondary,attribution_model,conversion_window,timezone - file: measurement-plan-template.txt firstLine: Approved inquiry definition: [required] requiredNativeInputs: - name: Actions kind: editorial_table proposedDistinctTemplate: native-actions-template.csv columns: - action_alias - event_definition - source - count_setting - primary_secondary - window - evidence_ref acquisitionRequirement: Populate only from the recipe’s documented export or explicit manual evidence capture. Existing blanks and synthetic fixtures do not supply facts. - name: AggregateCounts kind: editorial_table proposedDistinctTemplate: native-aggregatecounts-template.csv columns: - system - metric - period - count - scope_ref acquisitionRequirement: Populate only from the recipe’s documented export or explicit manual evidence capture. Existing blanks and synthetic fixtures do not supply facts. - name: ExistingTests kind: editorial_table proposedDistinctTemplate: native-existingtests-template.csv columns: - test_id - expected_event - observed_events - evidence_ref acquisitionRequirement: Populate only from the recipe’s documented export or explicit manual evidence capture. Existing blanks and synthetic fixtures do not supply facts. - name: Scope kind: approved_text_section proposedDistinctTemplate: native-scope-template.txt requiredContent: Full approved scope, source account aliases, date filters, date basis/timezone, as-of capture time, definitions, source-to-alias map provenance, matching rules, inspected/excluded records, permission/privacy limits, missing-data reasons, coverage and responsible reviewer. Include recipe-specific policies, versions and tests. Scope must be inside the AI workbook or pasted packet, not merely a separate unprovided file. templateFilesCreatedByThisTask: False requiredGate: Kit owner must add the named inputs or document field-level mapping before replacing an existing kit prompt. Preserve the old kit contract; never rename headers without acquiring missing evidence. existingFileMapping: - file: conversion-actions-template.csv mapping: Map action_id to verified action_alias; primary_or_secondary to primary_secondary; conversion_window to window. Keep attribution_model/timezone in Scope. Add actual settings evidence_ref; never pretend these are native export headers. - file: measurement-plan-template.txt mapping: Carry the complete approved inquiry, comparison, timezone and attribution definitions into Scope. It contains no actual AggregateCounts. - file: synthetic-test-plan-template.csv mapping: Expected test_id/action_ids/event_sequence/expected_business_inquiries describe a plan only. They cannot populate observed_events or claim tests ran. ExistingTests requires separate actual controlled-test evidence; otherwise mark absent/untested. - file: NEW native-aggregatecounts-template.csv mapping: Acquire real Ads/GA4 report metrics and their scopes as described. No existing starter file supplies these counts. OPEN THE INPUT MAPPING GUIDE NATIVE-INPUT-MAPPING.txt Use the distinct native input scaffolds listed below. Complete them from the acquired, reviewed evidence and include every required scope section in the AI-visible packet. Existing synthetic examples exercise the original review prompt only. They are not acquired evidence or a test of this native chain. Actions: native-actions-template.csv AggregateCounts: native-aggregatecounts-template.csv ExistingTests: native-existingtests-template.csv Scope: native-scope-template.txt STEP 1: Read conversion-action definitions and settings Tool: Google Ads Open: Goals > Conversions > Summary > selected conversion action > settings/details Access needed: Existing authorized access and relevant plan/region feature; exact live subscription and permissions not tested. No setup or persistent access changes included. AI role: not_an_ai_step Capture actual action name/source, primary/secondary status, Count, windows and the business-event definition. If Edit settings is needed to view Count, inspect without changing or saving. Record whether the attribution setting makes a control managed in Analytics. Expected output: Conversion-action settings ledger with screenshots. Use it next: Pair each action with its documented website/CRM event, not merely a friendly name. Check: One counts per ad click/action, not unique people across all clicks/actions. Changing it affects future reporting. Documentation limit: The source verifies the component feature; workflow prompts, schemas, manual captures, checks and handoffs are proposed editorial adaptations. Source documentation: https://support.google.com/google-ads/answer/3438531?hl=en https://support.google.com/google-ads/answer/3123169?hl=en https://support.google.com/google-ads/answer/2375435?hl=en STEP 2: Download the actual aggregate reports Tool: Google Ads and GA4 Open: Ads relevant statistics table > selected date/columns > Download > Excel CSV/XLSX; GA4 Reports > selected existing report > Share this report > Download File > Download CSV Access needed: Existing authorized access and relevant plan/region feature; exact live subscription and permissions not tested. No setup or persistent access changes included. AI role: not_an_ai_step Preserve original reports and date/filter/timezone context. Record whether Ads uses interaction-date or conversion-time columns and which action/metric the GA4 report shows; capture threshold/coverage warnings. Expected output: Native report exports with their own metric definitions. Use it next: Manually map only genuinely comparable action/metric definitions into AggregateCounts. Check: An Ads/GA4 total mismatch is not proof of duplicate firing. Attribution date, windows, imports and primary/secondary scopes can differ. Documentation limit: The source verifies the component feature; workflow prompts, schemas, manual captures, checks and handoffs are proposed editorial adaptations. Source documentation: https://support.google.com/google-ads/answer/2404176?hl=en https://support.google.com/analytics/answer/9317657?hl=en https://support.google.com/google-ads/answer/2375435?hl=en STEP 3: Acquire implementation evidence safely Tool: Existing tag configuration and an approved test record Open: GTM existing container/workspace configuration; future controlled test: Workspace > Preview > Tag Assistant Access needed: Existing authorized access and relevant plan/region feature; exact live subscription and permissions not tested. No setup or persistent access changes included. AI role: not_an_ai_step For this read-only recipe use already recorded configuration/test evidence, or return a test-plan blocker. A separately approved technician can later inspect event order in Tag Assistant using a synthetic isolated test path. No preview session or form submission is run here. Expected output: Existing tag-to-event map and evidence, or a synthetic test plan. Use it next: Provide only tag/event labels, counts and non-identifying IDs to analysis. Check: Preview is not guaranteed side-effect-free: tags/forms may still send data. Transaction IDs are not reported in Ads, so a report cannot prove deduplication; unique IDs must be verified in implementation/test evidence. Documentation limit: The source verifies the component feature; workflow prompts, schemas, manual captures, checks and handoffs are proposed editorial adaptations. Source documentation: https://support.google.com/tagmanager/answer/6107056?hl=en https://support.google.com/google-ads/answer/6386790?hl=en STEP 4: Make the checked, minimized handoff Tool: Excel / source-record reviewer Open: Restricted original workbook > separate analysis workbook > Scope worksheet Access needed: Existing authorized access and relevant plan/region feature; exact live subscription and permissions not tested. No setup or persistent access changes included. AI role: not_an_ai_step Build Actions, AggregateCounts and ExistingTests with actual definitions. Scope includes attribution/date differences and untested repeat-submit scenarios. Keep untouched original exports/screenshots and the identifying source map in restricted firm storage. Build a separate minimized analysis file; do not hide raw client data on another tab. Include the complete Scope/coverage note in the AI-visible workbook or prompt. Use aliases and administrative metadata only. Unknown is not zero, false, completed or absent. Expected output: Actions(action_alias,event_definition,source,count_setting,primary_secondary,window,evidence_ref); AggregateCounts(system,metric,period,count,scope_ref); ExistingTests(test_id,expected_event,observed_events,evidence_ref) Use it next: Format each approved range as a named table. Carry source-row references and coverage into the analysis workbook. Check: Every field has an observed source or an explicit missing reason; staff assignments are marked editorial, not native. Documentation limit: The source verifies the component feature; workflow prompts, schemas, manual captures, checks and handoffs are proposed editorial adaptations. Source documentation: https://support.microsoft.com/en-us/excel/merge-queries-power-query https://support.google.com/google-ads/answer/3438531?hl=en STEP 5: Ask AI to inspect the checked inputs Tool: Excel Copilot, optional Open: Analysis workbook > Copilot icon at lower right > Chat only Access needed: Existing authorized access and relevant plan/region feature; exact live subscription and permissions not tested. No setup or persistent access changes included. AI role: downstream_application_ai Open the minimized workbook containing the named tables and Scope. Paste the exact prompt; if unavailable, perform the same rule checks manually. PASTE THIS PROMPT Use Actions, AggregateCounts, ExistingTests and Scope. First identify scope/date/attribution differences that prevent comparison. Separate confirmed wrong-event wiring from suspected repeated firing and untested overlaps between native/imported actions. Explain One versus Every without equating either with person-level deduplication. Return evidence, hypothesis, smallest synthetic test and risk of a proposed change. No missing transaction ID may be inferred from an Ads export. Do not modify goals, tags, bidding or upload personal data. Expected output: Measurement audit and controlled synthetic test plan, not a diagnosed cause without evidence. Use it next: Copy checked results into a new Review sheet, retaining source references and unresolved blockers. Check: For each confirmed defect require specific implementation/test evidence; no invented event sequence. Documentation limit: The source verifies the component feature; workflow prompts, schemas, manual captures, checks and handoffs are proposed editorial adaptations. Source documentation: https://support.microsoft.com/en-us/excel/copilot/get-started-with-copilot-in-excel STEP 6: Human review and final action Tool: Responsible firm reviewer Open: Internal Review worksheet / review document Access needed: Existing authorized access and relevant plan/region feature; exact live subscription and permissions not tested. No setup or persistent access changes included. AI role: not_an_ai_step Marketer and website owner review evidence and approve any isolated test or fix with rollback planning. Retain unknown causes where tests are absent. Expected output: Reviewed conversion-count audit. Use it next: Only a separately authorized operator edits source tags/goals; no production tests or changes performed. Check: Primary-goal changes can affect bidding. Never send real prospect data as a test. Documentation limit: The source verifies the component feature; workflow prompts, schemas, manual captures, checks and handoffs are proposed editorial adaptations. Source documentation: https://support.google.com/tagmanager/answer/6107056?hl=en https://support.google.com/google-ads/answer/3438531?hl=en ACCOUNT AND SOURCE CHECKS STILL NEEDED - The live account, entitlement, exact headers, records and field coverage have not been observed. USING THE FILES IN THIS KIT Use the exact files or source text named in the steps above. Native exports keep their own headers. The kit tables are editorial review worksheets; they are not a claim that a product exports those fields. Some routes use reviewed text directly or require additional original files. Never rename an unreviewed raw export to make it appear mapped. Keep expected-result files separate from inputs. The source account, actual data, prompts and end-to-end workflow were not tested. Any final communication, filing, permission change, payment or publication needs its own authorized human action. This guide performs none.