# DV360 QA Checklist ## Project Knowledge for Claude and ChatGPT **Purpose:** Enterprise checklist for pre-build, pre-launch, post-launch, in-flight, change-control, reporting, measurement, deal, supply, audience, creative, and closure QA in Display & Video 360. **Operating rule:** QA is evidence and accountability, not a tick-box exercise. Every material issue requires an owner, evidence, severity, action, approval status, review trigger, and rollback condition. --- # 1. QA Roles ## Builder Creates or changes the campaign setup and records implementation evidence. ## Reviewer Independently checks the setup against the approved plan, brief, policies, and knowledge guides. ## Approver Accepts material risk and authorises launch or change. ## Measurement owner Validates Floodlights, analytics, CRM, attribution, and studies. ## Commercial owner Validates budget, fees, deals, commitments, currency, and billing treatment. ## Rule Where risk is material, the builder should not be the sole reviewer and approver. --- # 2. Severity Levels ## Critical Do not launch or immediately pause the affected scope where permitted. Examples: - Wrong advertiser, market, budget, destination, or conversion - Unapproved creative or expired rights - Missing mandatory privacy or suitability control - Guaranteed-deal commitment at risk - Material tracking failure ## High Correct before launch or through an approved urgent change. ## Medium Correct promptly and monitor documented impact. ## Low Housekeeping or improvement that does not materially threaten delivery, compliance, measurement, or customer experience. The Agent must not invent severity. Apply the advertiser's approved risk policy. --- # 3. QA Evidence Record For every review, capture: - QA ID - Advertiser - Campaign - Insertion order - Line item - Market - Reviewer - Review time - Source brief and plan - Checklist version - Evidence links or screenshots - Issues - Severity - Owner - Due date - Resolution evidence - Approver - Launch or change decision --- # 4. Brief and Plan QA - Business objective is clear. - DV360 role is clear. - Primary outcome is approved. - Market, language, audience, product, and flight are defined. - Budget, currency, cost basis, and fees are defined. - Formats and environments are defined. - Supply, deals, and publisher requirements are documented. - Brand safety, suitability, invalid-traffic, and viewability rules are approved. - Frequency strategy is documented. - Measurement source, Floodlights, attribution, and reporting are documented. - Creative, landing page, rights, and localisation are approved. - Test design is documented where applicable. - Owners and approvers are named. Do not build from an ambiguous or conflicting plan without documenting the decision owner. --- # 5. Advertiser and Partner QA - Correct partner and advertiser selected. - Advertiser name and ID recorded. - Currency correct. - Timezone correct. - Billing profile and commercial setup confirmed by the owner. - Floodlight configuration correct. - Campaign Manager 360 connection correct where used. - Linked accounts and permissions appropriate. - Default targeting and advertiser-level controls reviewed. - Data access follows least-privilege requirements. - Market and privacy requirements reviewed. Currency and timezone errors can be difficult to reverse. Treat them as critical setup checks. --- # 6. Campaign QA - Campaign name follows the approved convention. - Correct advertiser and objective. - Start and end dates correct. - Market and product scope correct. - Campaign-level budget logic documented. - Brand-safety and suitability controls applied. - Verification settings approved. - Frequency level and inheritance understood. - Reporting labels and integration codes correct. - No obsolete or copied settings remain from a duplicated campaign. --- # 7. Insertion Order QA - Name and objective correct. - Flight dates correct. - Budget amount and currency correct. - Budget type correct. - Pacing type and period correct. - Performance goal or KPI correct. - Automatic budget allocation setting intentional where available. - Frequency cap correct and not duplicated unintentionally. - Bid strategy compatible with objective and line items. - Line-item budgets and commitments reconcile to the insertion order. - Guaranteed or reserved commitments protected. - No unintended inactive, archived, or duplicated line items. --- # 8. Line Item QA - Correct line-item type. - Name follows convention. - Correct insertion order. - Flight dates and timezone correct. - Budget and pacing correct. - Bid strategy, bid, and target correct. - Optimisation event correct. - Frequency cap correct. - Geography and language correct. - Device and environment correct. - Inventory source correct. - Audience and exclusions correct. - Content and technology targeting correct. - Brand-safety and suitability controls correct. - Verification and viewability controls correct. - Creatives assigned and compatible. - Floodlight activities selected correctly. - Status is correct for the launch stage. --- # 9. Budget and Pacing QA - Approved total budget matches the plan. - Currency and cost basis match. - Insertion-order and line-item budgets reconcile. - Daily versus flight pacing is intentional. - Start and end dates support the pacing plan. - Front-loaded, even, or ahead pacing is approved. - Deal and guaranteed commitments can be fulfilled. - Test cells have protected budgets where required. - No duplicated budget exists across copied structures. - Fees, data, verification, and serving costs are understood. - Pacing alerts and review triggers are defined. Do not change pacing merely to spend the budget if quality or measurement is failing. --- # 10. Geography, Language, and Device QA - Included markets and subregions correct. - Excluded regions correct. - Location presence or location logic matches the plan. - Language targeting is intentional. - Device types correct. - Operating system, browser, carrier, or model restrictions are justified. - Advanced device restrictions do not unintentionally remove eligible supply. - CTV device treatment is explicit. - Mobile app and web environments are separated where needed. - Market-specific inventory and privacy constraints are documented. --- # 11. Audience QA - Audience source and owner recorded. - Prospecting, retargeting, retention, and suppression roles clear. - First-party data has approved use and consent. - Membership duration and freshness are appropriate. - Geography and market availability confirmed. - Audience size is viable. - Existing customers and completed converters suppressed where required. - Inclusion and exclusion conflicts checked. - Audience overlap assessed. - Expansion or optimised targeting setting intentional. - Strict retargeting boundaries protected. - Marketplace or publisher audience cost and provenance documented. - Sensitive or protected characteristics are not used discriminatorily. - Frequency risk assessed for small pools. --- # 12. Inventory and Supply QA - Open Auction, deal, publisher, exchange, and package roles documented. - Correct exchanges and inventory sources included. - Unapproved sources excluded. - Publisher, domain, app, channel, or programme scope correct. - ads.txt or app-ads.txt policy reviewed. - Seller and SupplyChain transparency reviewed where available. - Direct and reseller routes understood. - Duplicate paths reviewed. - Long-tail inventory governance defined. - App identifiers checked for CTV and app inventory. - Invalid-traffic and made-for-advertising risks assessed. - Allowlist and blocklist versions current. - Scale and frequency effects of restrictions assessed. --- # 13. Deal QA - Deal name and ID correct. - Publisher and seller correct. - Deal type correct. - Seat and advertiser access correct. - Deal accepted or negotiated where required. - Price, floor, currency, dates, and inventory correct. - Volume or commitment documented. - Unique versus duplicated access assessed. - Correct deal assigned to the line item. - Bid compatible with floor where auction-based. - Audience and targeting do not conflict with seller settings. - Creative meets publisher requirements. - Measurement and reporting available. - Cancellation, amendment, underdelivery, and makegood terms documented. - Commercial owner approved. Accepting a deal does not by itself activate delivery. Line-item assignment, budget, targeting, creative, and active status still require QA. --- # 14. Creative QA - Creative type matches the line item. - Asset name and ID recorded. - Format, dimensions, ratio, duration, and file quality correct. - Brand, product, offer, price, legal copy, and call to action correct. - Language and localisation correct. - Landing page and display URL correct. - Click, exit, QR, and follow-on actions tested. - Image cropping, text truncation, and rendering reviewed. - Video sound, captions, opening, and end frame checked. - CTV legibility and rights checked. - Audio voice, music, mix, companion, and rights checked. - Native component completeness and publisher rendering checked. - HTML5 click tags, exits, and responsive serving sizes checked. - Dynamic feed, default creative, and decision rules checked. - Talent, music, image, footage, territory, and expiry rights approved. - Platform, brand, legal, and publisher approval complete. --- # 15. Landing Page QA - Destination resolves without error. - Tracking parameters correct. - Market, language, product, offer, and currency match the ad. - Mobile and desktop experience appropriate. - Page speed and critical functionality checked. - Form, checkout, app-store, or product journey works. - Cross-domain tracking works where required. - Consent behaviour correct. - Confirmation page or event fires correctly. - Unavailable products and expired offers removed. - Redirects preserve tracking. - No test or staging destination remains. --- # 16. Floodlight and Measurement QA - Correct Floodlight configuration. - Activity name and ID correct. - Counter or Sales type correct. - Counting method matches the business event. - Attribution enabled where intended. - Tag fires on the correct action. - Duplicate firing tested. - Revenue, currency, order ID, product, and custom variables correct. - Test transactions identified. - Primary and secondary activities distinguished. - Line items use the intended optimisation activities. - Attribution model and windows documented. - Click-through and view-through reporting understood. - Analytics and CRM reconciliation plan agreed. - Conversion lag documented. - Consent and privacy review complete. - Incrementality method approved where required. Official Google documentation describes Floodlight as the shared Google Marketing Platform conversion system and documents Counter and Sales activities and their counting options. Current platform instructions must be verified before implementation. --- # 17. Brand Safety, Suitability, and Verification QA - Advertiser suitability profile approved. - Digital content labels correct. - Sensitive categories and exclusions correct. - Keyword, category, channel, site, and app controls reviewed. - Verification provider and integration correct where used. - Pre-bid and post-bid roles understood. - Invalid-traffic controls approved. - Viewability controls approved. - Geography verification reviewed. - Native, audio, CTV, and app measurement limitations documented. - Controls do not unintentionally remove required scale. - Exception process and approver documented. Do not invent universal thresholds. Use the advertiser's approved standards or the approved benchmark registry. --- # 18. Display QA - Static, HTML5, responsive, rich-media, or dynamic format correct. - Required dimensions cover intended inventory. - Creative remains legible at each size. - Backup assets present where required. - Viewability targeting is intentional. - Web and app inventory treatment clear. - Placement lists current. - Click and landing continuity validated. - Retargeting suppression and frequency checked. --- # 19. Video QA - YouTube versus third-party video structure correct. - Objective and subtype correct. - Ad groups configured intentionally where required. - Skippable, non-skippable, in-feed, in-stream, out-stream, Shorts, or sequence role correct. - View and completion definitions documented. - Duration and aspect ratio correct. - Sound, captions, claims, and landing page correct. - Reach and frequency strategy correct. - YouTube and partner settings intentional. - Third-party measurement compatibility verified. --- # 20. Connected TV QA - Device Type and CTV-compatible inventory correct. - YouTube CTV and third-party CTV roles separated where needed. - App names and identifiers reviewed. - Publisher, channel, programme, exchange, and deal documented. - Household, device, user, or modelled reach definition documented. - Frequency strategy accounts for fragmented supply. - Large-screen creative, sound, rights, and QR code checked. - Clickless response and cross-device measurement limitations documented. - Completion is not treated as proof of attention. --- # 21. Audio QA - Cross-exchange or YouTube audio workflow correct. - Streaming, podcast, radio-like, gaming, or other content type intentional. - Creative duration and file requirements verified. - Script, brand cue, voice, music, mix, call to action, and companion checked. - Device, audience, content, deal, and frequency settings correct. - Music and voice rights approved. - Completion and reach definitions documented. - Brand, search, analytics, CRM, or causal measurement agreed. --- # 22. Native QA - Native display or native video type correct. - Required images, logos, headlines, descriptions, advertiser name, call to action, display URL, and destination complete. - Square and landscape assets intentionally composed. - Text remains truthful if truncated. - Essential copy is not embedded where cropping can remove it. - Representative publisher renderings reviewed. - Tracking compatibility verified. - Viewability-coverage limitation documented. - Ad disclosure and advertiser identity clear. --- # 23. Test and Experiment QA - Decision and hypothesis documented. - Control and treatment defined. - One primary variable identified. - Primary outcome and guardrails approved. - Audience, inventory, creative, budget, and timing controls documented. - Sample or power logic documented. - Conversion lag included. - Contamination risks assessed. - Mid-test optimisation restrictions documented. - Analysis method and success rule defined before launch. - Owner and approver recorded. - Results will be labelled directional if design integrity is lost. --- # 24. Pre-Launch Sign-Off Before launch, confirm: - No unresolved critical issues. - High issues have approved disposition. - Budget, dates, market, and destination rechecked. - Measurement rechecked. - Creative approval rechecked. - Inventory and deal access rechecked. - Status changes are scheduled or controlled. - Builder and reviewer sign-off complete. - Approver sign-off complete. - Launch-monitoring owner available. - Rollback steps documented. --- # 25. Post-Launch Checks Immediately after activation and at the first meaningful data point, check: - Correct entities are active. - Spend and impressions begin as expected. - No unexpected market, device, exchange, app, publisher, or placement delivery. - Creative renders correctly. - Clicks and destinations work. - Floodlights record correctly. - Audience and suppression behave as intended. - Deals are participating. - Pacing is viable. - Quality and suitability signals are available. - No severe frequency concentration appears. The precise timing should follow spend, risk, traffic, and conversion-lag conditions rather than a universal interval. --- # 26. In-Flight QA Review material changes in: - Pacing - Budget allocation - Bids and targets - Conversion volume and quality - Reach and frequency - Inventory concentration - Audience scale and overlap - Creative delivery and fatigue - Viewability, invalid traffic, and suitability - Deal delivery - Measurement discrepancies Do not turn routine optimisation into uncontrolled structural change. --- # 27. Change-Control QA For every material change, record: - Change requested - Reason - Evidence - Affected scope - Expected effect - Risk - Owner - Approver - Implementation time - Before state - After state - Review trigger - Rollback condition Recheck dependent settings after changes. A budget change may affect pacing; an audience change may affect scale and frequency; a creative change may affect eligibility; a Floodlight change may affect bidding. --- # 28. Reporting QA - Reporting period and timezone correct. - Currency and cost basis correct. - Filters and entity scope correct. - Metrics and dimensions match definitions. - Formula calculations validated. - Attribution model and windows stated. - Conversion activity and counting method stated. - Reach unit and methodology stated. - View and completion definitions stated. - Platform, analytics, CRM, and partner sources reconciled. - Benchmarks have approved status. - Assumptions labelled. - Insights separate fact from interpretation. - Recommendations include evidence, owner, guardrail, and review trigger. - No confidential or personal data included without permitted use. --- # 29. End-of-Campaign QA - Final spend reconciled. - Delivery and commitments reviewed. - Makegoods or discrepancies documented. - Final Floodlight and business outcomes reconciled after lag. - Attribution and measurement settings archived. - Experiment integrity and result documented. - Creative, audience, deal, and placement learnings recorded. - Rights-expired assets retired. - Obsolete lists and deals reviewed. - Final report approved. - Actions transferred to the next planning cycle. --- # 30. QA Issue Template ```markdown ## QA Issue: [ID] - Severity: - Advertiser: - Campaign / IO / Line item: - Market: - Finding: - Evidence: - Risk: - Required action: - Owner: - Approver: - Resolution evidence: - Status: - Review trigger: - Rollback condition: ``` --- # 31. Launch Sign-Off Template ```markdown # DV360 Launch Sign-Off - Advertiser: - Campaign: - Market: - Flight: - Budget and currency: - Builder: - Reviewer: - Measurement owner: - Commercial owner: - Approver: - Critical issues open: - High issues open: - Approved exceptions: - Launch decision: Approved / Approved with conditions / Blocked - Evidence location: - Monitoring owner: - Rollback plan: - Sign-off date: ``` --- # 32. QA Decision Rules - Block launch when a critical control is unresolved and no authorised exception exists. - Do not mark an item complete without evidence. - Do not accept "copied from previous campaign" as validation. - Recheck inherited settings after duplication. - Do not invent approval. - Do not relax privacy, policy, legal, suitability, or rights controls to spend budget. - Do not use benchmarks that are not approved. - Escalate commercial commitments and cross-channel decisions. - Prefer targeted correction over broad changes. - Preserve a change log and rollback path. --- # 33. Default Agent Output **Mode selected:** Pre-Build QA / Pre-Launch QA / Post-Launch QA / In-Flight QA / Change QA / Reporting QA / Closure QA ## Launch or change decision - Approved - Approved with conditions - Blocked ## Critical issues ## High issues ## Medium issues ## Low issues ## Evidence and unknowns ## Required actions ## Owners and approvers ## Review triggers ## Rollback conditions --- # 34. Source and Freshness Notes Official Google documentation describes Floodlight as the shared Google Marketing Platform conversion system and documents Counter and Sales activity types and counting methods. Current platform workflows must be verified before implementation. Core setup areas include Floodlight configuration, campaign setup, deal activation, creative assignment, reporting, and verification controls. Validating foundational tagging and reconciling ad-server, analytics, CRM, and sales sources are essential measurement practices. Platform interfaces, settings, policies, supported formats, verification integrations, and reporting fields can change. Use current official documentation and the advertiser's approved policies during every QA review.