Kertújítók client meeting: strategy, exact setup and measurement dashboards

Present the six exact campaigns and both maintenance groups, 5,000 Ft/day reviewed allocation, service intent, landing routes, bidding, campaign status/end, network settings, locations/radii/exclusions and presence option, languages, schedules, devices/demographics and modifiers, audience/observation settings, keywords/match/negative structures, RSA/assets and conversion-goal/tracking configuration where applicable. Separate configured facts, design rationale, observed outcomes and proposed next steps. State when a category is not configured or not exposed rather than inventing it. Reconcile the recorded launch setup with current configuration. Show material changed/removed settings as dated history, keeping the current inventory distinct. Cite recorded campaign intent where available, and label rationale reconstructed from settings as inference.

Verification: Compare fresh API facts with launch-receipt.json, setup-audit.json and the recorded launch contract. Maintain one setting-category matrix linking source, campaign scope, screenshot, designed slide and exact appendix. Check displayed values and distinguish recorded rationale from inference. Do not expand into exhaustive removed-item history.

Use genuine Google Ads UI screenshots for every setting category presented under deck-content, readable in the deck with observation time and a plain caption. Pair each with an easier-to-overview designed view of the same data. Representative screenshots may cover identical setups if this is explicit and all real differences are represented. Never render or fabricate a Google Ads screenshot.

Verification: The coverage matrix names a genuine screenshot and paired same-data view for every presented category, or an exact provider/access gap. Map representative captures to campaigns whose values are verified identical and show every difference. Save original UI captures with account/campaign context, capture time and API snapshot reference. Compare visible text to API values, remove unrelated private account information, and never fabricate a provider capture. Recover access where possible. An unresolved screenshot gap leaves this criterion incomplete while independent work continues.

Deck is polished and legible for live presentation, in Hungarian, with keyboard/button slide navigation, overview/direct navigation, useful screenshot zoom and printing. No clipping of essential content at a representative desktop presentation viewport; mobile can read and navigate without horizontal document overflow.

Verification: Exercise navigation, direct slide links, screenshot zoom and print preview; inspect representative main-slide and dense-appendix screenshots at desktop and phone widths. Follow required independent visual QA on deployed client-facing HTML. No automatic user-facing preview/panel opening.

List every non-removed configured keyword for the six campaigns and their ad groups, positive/negative and campaign/shared/ad-group scope, exact text, match type and enabled/paused state. Include shared negative list contents and links where relevant. Preserve accents and exact text. Include every non-removed RSA for these groups with all headlines/descriptions in provider order, pinning, paths, final URLs, active/paused state and policy status. Include all linked applicable asset copy and targeting/bid/budget values without approximations. Removed history need not be exhaustively included.

Verification: Generate reference tables from fresh structured API data, not hand-copying. Programmatically compare unique IDs/counts and each text/value to the source export; no omissions, duplicate records or paraphrases in fields labelled exact. Report provider-omitted fields as unknown.

Each simplified slide identifies its campaign/ad-group and links to the corresponding exact appendix and dated evidence. Provider omissions and source conflicts are visible, and the latest native end/date status is explicit. Include the historical-to-current reconciliation where settings changed, with original and current observation dates.

Verification: Test anchors and filters, reconcile source conflicts, verify IDs/counts/copy against source. Store credential-free provider evidence privately and publish only client-relevant safe evidence. Do not expose private unrelated records, credentials or entire raw account exports on the bare public URL.

Provide multiple functional dashboard views with sensible campaign/service/date controls and appropriate KPI cards, trends, comparisons and tables. Cover impressions, clicks, CTR, spend, CPC, budget pacing, search terms/match quality, conversion actions, qualified leads and CPL, plus form/event integrity where supported. Each metric has a definition, reporting window, source, denominator and availability/proof state. Aggregate correctly across shared group/campaign measures without double counting.

Verification: Load the fresh observed snapshot and test filtering and reconciliation to source totals. Document source-to-metric mapping and formula checks. Controls that cannot be supported by available date grain are disabled or clearly scoped rather than pretending to filter. Read and test the reporting rules in resources.reporting-semantics: account clock/currency, unique budget pools, reported search-term coverage, exact conversion field/date basis and complete-query zero handling.

Use actual fresh reported data for current observations. Use explicit no-data/unknown/not-integrated states for unavailable leads, revenue, attribution and measurement proof. Never convert tracking clicks into accepted leads or make ROI/ROAS claims without measured revenue. If illustrative data is needed for demonstrating a future populated chart, keep it behind a separate clearly labelled demo mode and prevent blending with actual results.

Verification: Verify zero/null/unknown and divide-by-zero cases, campaign filters, selected periods and labels. Confirm no invented outcomes, hidden sample data or fake integrations. Distinguish Ads conversion actions from accepted/qualified leads and compare windows/timezone consistently. Preserve source query success/completeness, snapshot time and reporting grain in the local data manifest. Validate a successful zero-data case separately from failed or incomplete retrieval.