# DV360 Display Guide ## Project Knowledge for Claude and ChatGPT **Version:** 1.0 **Purpose:** Give the DV360 Agent practical, evidence-led decision logic for planning, building, auditing, diagnosing, optimising, testing, measuring, and quality-assuring display campaigns. **Companion guides:** Use alongside the related operating and specialist guides in this knowledge library. --- # 1. Purpose This guide teaches the DV360 Agent to operate display as a deliberate media capability rather than a cheap-impression channel. Display can support: - Broad or incremental reach - Demand creation - Consideration - Qualified site engagement - Lead generation - Online sales - Retargeting - Customer retention - Creative sequencing - Contextual activation The correct display design depends on the advertiser's business outcome, audience, inventory, creative, measurement, risk, and budget. No single metric determines success. The DV360 Agent must not optimise display using click-through rate, cost per thousand impressions, viewability, or platform conversions in isolation. Display decisions must connect delivery and quality signals to qualified outcomes and the channel's assigned role. --- # 2. Agent Mandate When using this guide, the DV360 Agent must: 1. Start from the approved advertiser objective and display role. 2. Distinguish awareness, consideration, performance, retargeting, and retention jobs. 3. Validate Floodlights and the business source of truth before recommending bidding or scaling changes. 4. Assess placement, supply, audience, creative, frequency, destination, and measurement together. 5. Diagnose the primary constraint before prescribing an action. 6. Separate platform-attributed outcomes from incremental business impact. 7. Keep line-item architecture understandable and sufficiently funded. 8. Apply placement exclusions only with evidence and approval. 9. Use experiments when the cause of performance differences is uncertain. 10. Limit recommendations to the highest-impact actions. 11. Require human approval for live-account changes. 12. Escalate cross-platform budget and channel-role decisions to the appropriate human decision-owner. --- # 3. Display Roles and Success Measures ## 3.1 Awareness and reach Use when display must expose a defined audience to a proposition across suitable environments. Primary measures may include: - On-target or eligible reach - Incremental reach - Frequency distribution - Viewable reach where supported - Brand or search lift where measured Guardrails may include: - Cost - Viewability - Invalid traffic - Suitability - Publisher diversity Do not use attributed conversions as the only judgement of an awareness role. ## 3.2 Consideration Use when display must encourage qualified product exploration or progress toward an outcome. Primary measures may include: - Qualified sessions - Product or service exploration - Journey starts - Engaged visits - Relevant micro-conversions Guardrails may include: - Landing-page quality - Frequency - Inventory quality - Post-click and post-view reconciliation ## 3.3 Performance Use when display has a defensible path to leads, sales, applications, or another qualified action. Primary measures should come from: - CRM-qualified leads - Approved applications - Revenue - Margin - New-customer outcomes - Retained-value outcomes Floodlight conversions may support bidding, but must be reconciled with business results. ## 3.4 Retargeting Use to reconnect eligible users with a specific unfinished customer task. Retargeting requires: - Clear inclusion event - Clear completed-event suppression - Relevant recency - Message progression - Responsible frequency - Incrementality awareness Strong attributed performance does not automatically justify more retargeting budget. ## 3.5 Retention and cross-sell Use only when customer eligibility, data permission, proposition, and measurement are clear. Do not mix acquisition and existing-customer economics without disclosure. --- # 4. Required Inputs Before planning or auditing, organise available information into: ## Business - Product or service - Market - Objective - Qualified outcome - Revenue, margin, or lifetime-value logic - New versus existing customer priority ## Campaign - Advertiser - Campaign - Insertion order - Line item - Dates - Currency - Budget and pacing - Bid strategy and target ## Audience - Prospecting and retargeting definitions - First-party or marketplace sources - Exclusions - Membership duration - Expansion settings where available ## Placement and supply - Environment - Exchange or SSP - Deal - Publisher - Domain - App - Placement - Device - Position where available - Authorised-seller status where available ## Creative - Format - Dimensions - Concept - Message - Version - Language - Landing page - Approval status ## Quality - Measurability - Viewability - Invalid traffic - Suitability - Verification outcomes ## Measurement - Floodlight activities - Attribution windows - Click-through and view-through outcomes - Analytics - CRM - Revenue or qualified outcomes - Conversion lag State missing data and its decision impact. --- # 5. Campaign Architecture ## 5.1 Campaign level Use campaigns for meaningful business, market, flight, or governance boundaries. Avoid separating campaigns merely for reporting convenience when the same outcome, budget, and controls apply. ## 5.2 Insertion-order level Use insertion orders to manage a coherent budget, objective, pacing strategy, and optimisation design. Separate insertion orders when: - Business objectives differ - Budgets require independent control - Markets or currencies differ - Guaranteed commitments require protection - Measurement or governance differs ## 5.3 Line-item level Use line items for meaningful differences in: - Format - Inventory source - Deal or buying route - Audience role - Bidding strategy - Frequency - Creative experience - Test cell Do not create one line item per minor segment, creative, domain, or reporting label without a decision reason. ## 5.4 Fragmentation test Before adding another line item, ask: - Does this require separate budget or control? - Does this represent a distinct hypothesis? - Will it receive enough opportunity and budget to learn? - Can results be interpreted? - Is frequency managed coherently? If the answer is no, consolidate. --- # 6. Placement Strategy A placement is not simply a site name. Placement decisions may involve publisher, domain, app, page type, ad slot, position, environment, device, exchange, deal, and size. ## 6.1 Open-market placement discovery Open-market delivery can reveal useful publishers and contexts. Use it with: - Authorised-seller controls - Brand-suitability controls - Verification - Domain and app reporting - Frequency management - Outcome-quality analysis Do not treat all long-tail inventory as poor. Do not leave material long-tail spend unexplained. ## 6.2 Allowlist strategy Use an allowlist when: - The advertiser needs strict environment control - A defined publisher set is central to the plan - Policy or regulatory risk requires it - A test needs a controlled placement universe Risks: - Reduced scale - Higher cost - Concentrated frequency - Manual maintenance - Accidental exclusion of useful inventory Decision rule: use an allowlist only when the control benefit justifies the scale and operational trade-off. ## 6.3 Blocklist strategy Use blocklists for documented unsuitable, invalid, irrelevant, fraudulent, or operationally problematic supply. Rules: - Record the evidence and reason. - Apply exclusions at the narrowest effective level. - Distinguish domain, app, placement, exchange, and seller causes. - Review inherited or global lists for relevance. - Define an exception and reinstatement process. Do not exclude a publisher solely because CTR is low. ## 6.4 Site and app evaluation Assess: - Content and user experience - Ad density - Viewability and measurability - Invalid traffic - Suitability - Placement position - Audience relevance - Reach contribution - Qualified outcomes - Supply path Separate web and app because measurement, identifiers, environment, and quality can differ. ## 6.5 Position and page environment Where data exists, review above-the-fold, below-the-fold, in-content, sticky, interstitial, or other eligible placement types. Do not assume one position is universally superior. Evaluate attention opportunity, disruption, suitability, cost, and outcomes. ## 6.6 Deal placement Use the PMP Guide when a deal provides specific publishers, sections, formats, positions, or audiences. "Premium" is not a sufficient placement definition. --- # 7. Placement Decision Rules ## Rule 1: Do not optimise on CPM alone Low CPM may reflect broad efficient access or weak-quality inventory. Check viewability, invalid traffic, user experience, frequency, and business outcomes. ## Rule 2: Do not optimise on CTR alone A high CTR may reflect relevance, curiosity, accidental clicks, intrusive formats, or low-quality environments. Validate landing engagement and qualified outcomes. ## Rule 3: Do not optimise on viewability alone Higher viewability may improve exposure opportunity while reducing scale and raising cost. It does not prove message attention or business impact. ## Rule 4: Exclude with evidence Exclude a placement when evidence shows persistent material risk or weak value and when alternative causes have been considered. ## Rule 5: Validate concentration High spend concentration can be deliberate and valuable or accidental and risky. Review reach, frequency, inventory quality, and outcome contribution. ## Rule 6: Test comparable placements Control audience, creative, market, dates, frequency, and bidding before attributing differences to placement. ## Rule 7: Protect unique reach Do not remove a route or publisher if it provides valuable incremental reach that is not captured by conversion-only analysis. ## Rule 8: Separate supply from publisher Poor results through one exchange do not automatically prove the publisher is poor. Different routes may expose different inventory, floors, placements, or quality. --- # 8. Creative System Creative must convert the audience and placement opportunity into a clear customer message. ## 8.1 Creative hierarchy ### Concept A materially different customer motivation, tension, need, use case, proposition, or proof. ### Execution A different visual story, product demonstration, layout, animation, format, or content treatment. ### Element A different headline, call to action, image, colour, offer, or product emphasis. Prioritise concept and execution learning before producing excessive element-level variants. ## 8.2 Creative brief Every display creative brief should define: - Audience and customer stage - Customer problem - Product truth - Proposition - Proof - Offer - Mandatory claim - Prohibited claim - Message hierarchy - Call to action - Landing page - Format and dimensions - Language and localisation - Usage rights - Measurement hypothesis ## 8.3 Format roles ### Static image Useful for simple, reliable delivery. Requires strong hierarchy and concise communication. ### HTML5 display Supports animation and richer sequencing within fixed-size or supported responsive executions. Official DV360 documentation describes HTML5 creatives as ads built with JavaScript, HTML, and CSS and distinguishes standard display from richer interactive formats. ### Responsive HTML5 Can adapt a single creative to multiple supported inventory sizes. Official documentation states that serving sizes still need to be selected so the creative can match eligible inventory. ### Rich media May support interaction, expansion, galleries, video, or other advanced behaviour. Use when the interaction serves the customer task and measurement is defined. ### Dynamic or data-driven display Use when message, product, or offer variation can be governed by approved data and clear rules. Automation does not remove the need for concept, brand, legal, feed, and destination QA. --- # 9. Creative Decision Rules ## Rule 1: Message before motion Animation cannot compensate for an unclear proposition. ## Rule 2: Design for the actual size Do not shrink a large-format layout into a small banner. Define what must survive at every size. ## Rule 3: Branding must be recognisable The correct level and timing depend on the objective and creative idea, but the advertiser should not pay for exposure that cannot be attributed to the brand. ## Rule 4: One clear next step The call to action should match the destination and the customer's stage. ## Rule 5: Maintain landing continuity Creative promise, product, offer, language, and visual cues should continue on the landing page. ## Rule 6: Format coverage affects eligible supply Missing common or strategically important sizes can restrict access and distort placement mix. Do not invent a universal mandatory size list. Use inventory forecasts, publisher requirements, and current platform specifications. ## Rule 7: Do not use platform labels as the verdict Creative approval, serving status, or an automated rating does not prove business effectiveness. ## Rule 8: Separate concept from placement effects A creative may appear weak because it served in poorer environments or to a different audience. Control confounders before declaring a winner. ## Rule 9: Measure qualified response CTR and interaction rate are diagnostics. Evaluate qualified visits, conversions, revenue, or brand impact according to the role. ## Rule 10: Localisation is more than translation Review proposition, imagery, examples, reading direction, offer, product availability, regulation, and cultural meaning. --- # 10. Creative Fatigue and Sequencing ## 10.1 Fatigue indicators Look for a combination of: - Ageing creative - Repeated exposure - Declining qualified response - Rising CPA or falling conversion rate - Reduced incremental reach - Weakening engagement - Limited concept rotation Frequency alone does not prove fatigue. ## 10.2 Alternative explanations - Audience saturation - Placement mix change - Seasonality - Offer weakness - Landing-page change - Measurement issue - Competitor activity ## 10.3 Sequencing Use sequencing when message progression matches the customer journey. Example stages: 1. Introduce the problem or category. 2. Present the proposition. 3. Provide product proof. 4. Address objections. 5. Present the action or offer. Sequencing requires sufficient audience scale, identity continuity, frequency governance, and a clear comparison with non-sequenced messaging. --- # 11. Viewability Framework Viewability represents the opportunity for an ad to be seen under the relevant measurement definition. It does not prove that the user noticed, understood, or acted because of the ad. ## 11.1 Key measures - Measurable impressions - Viewable impressions - Viewability rate - Cost per viewable impression or thousand viewable impressions where calculated - Reach and frequency among measurable delivery where available Always disclose the measurement provider and definition. ## 11.2 Predicted viewability targeting Official DV360 guidance states that line items can target inventory based on predicted viewability using Active View. It also warns that higher predicted-viewability settings can reduce eligible inventory, and that inventory that cannot be measured should not be targeted for viewability. ## 11.3 Viewability decision rules ### Rule 1: Set viewability as a role-based guardrail The appropriate requirement depends on the objective, format, inventory, market, cost, and advertiser risk. ### Rule 2: Do not maximise viewability blindly Stricter targeting can reduce supply, increase CPM, shift publisher mix, and concentrate frequency. ### Rule 3: Distinguish measured from predicted Predicted targeting is an eligibility input. Reported viewability is an observed measurement. Do not present them as the same thing. ### Rule 4: Investigate measurability A weak viewability rate and a weak measurability rate are different problems. ### Rule 5: Connect quality to outcomes Assess whether higher-viewability inventory improves reach quality, engagement, brand impact, or qualified outcomes. ### Rule 6: Control confounders A high-viewability line item may have different publishers, audiences, devices, formats, floors, and creative. ### Rule 7: Do not trade away mandatory suitability Viewability improvement does not override brand-safety, fraud, privacy, or suitability requirements. --- # 12. Viewability Diagnosis ## High viewability, weak outcomes Investigate: - Audience quality - Creative clarity - Placement context - Offer and destination - Frequency - Attribution - Whether optimisation selected viewable but less valuable supply ## Low viewability, strong outcomes Investigate: - Measurement coverage - Conversion concentration - Retargeting or existing demand - Post-click versus post-view outcomes - Whether the lower-viewability supply provides unique reach Do not ignore quality, but do not remove valuable inventory without analysis. ## Viewability falls suddenly Check: - Publisher and app mix - Exchange or deal mix - Device and format mix - Creative sizes - Measurability - Verification settings - Platform or tag changes - New inventory sources ## Viewability targeting causes underdelivery Check whether the threshold is limiting eligible supply. Review the advertiser's minimum requirement, quality objectives, reach, frequency, cost, and alternatives before changing it. --- # 13. Inventory Quality and Verification Review: - Viewability - Invalid traffic - Brand safety - Brand suitability - Geographic delivery - Domain and app transparency - Ad clutter and user experience - Attention signals where approved Verification metrics are controls and diagnostic inputs. A high quality score does not automatically prove business effectiveness. When platform and verification providers disagree: - Confirm definitions - Confirm measured population - Confirm date, timezone, device, and format - Confirm pre-bid versus post-bid logic - Identify tag or integration gaps - Avoid choosing the more favourable number without investigation --- # 14. Frequency Governance Frequency must reflect: - Objective - Audience size - Customer stage - Creative rotation - Flight length - Format - Placement environment - Conversion lag - Cross-channel exposure ## Decision rules - Do not impose one universal cap. - Retargeting usually requires tighter recency and relevance management. - Awareness may require enough repetition for message exposure, but additional frequency must still have a role. - Manage frequency across overlapping line items where possible. - Review frequency distribution, not only average frequency. - State limitations in cross-device and cross-platform frequency reporting. ## Frequency problem indicators - Reach remains flat while impressions grow - Small audiences dominate spend - Qualified outcomes flatten - Creative response weakens - Placement concentration rises - Users remain eligible after completing the outcome --- # 15. Bidding and Budget Decisions Before changing bidding or budget, validate: - Business outcome - Floodlight integrity - Conversion volume and quality - Attribution and lag - Audience and placement mix - Creative readiness - Inventory quality - Frequency - Available headroom ## Budget increase rule Recommend an increase only when the next investment has credible headroom and the likely marginal outcome is valuable. ## Bid increase rule Increase bids only when additional eligible inventory or deal access is expected to create value and the economics support the expected price. ## Target adjustment rule Do not loosen a target merely to deliver more or tighten it merely to improve average efficiency. Assess lost valuable volume and marginal returns. ## Underdelivery rule Diagnose budget, bids, targeting, audience, inventory, creative, policy, frequency, dates, and measurement before changing a single setting. --- # 16. Display Audit Workflow ## Stage 1: Define the decision Examples: - Why did qualified CPA rise? - Which placements should retain investment? - Is retargeting saturated? - Should viewability controls change? - Which creative concept should scale? ## Stage 2: Validate data - Scope - Dates - Currency - Timezone - Attribution - Floodlights - Analytics and CRM - Conversion lag - Recent changes ## Stage 3: Review delivery - Spend - Impressions - Pacing - Reach - Frequency - Bids and wins where relevant ## Stage 4: Review placement and supply - Exchange - Deal - Publisher - Domain - App - Placement - Device - Format - Long tail - Concentration ## Stage 5: Review audience - Prospecting versus retargeting - First-party and marketplace data - Suppression - Expansion - Overlap - Saturation ## Stage 6: Review creative - Concept - Execution - Size coverage - Format - Eligibility - Landing continuity - Fatigue ## Stage 7: Review quality - Measurability - Viewability - Invalid traffic - Suitability - Verification discrepancies ## Stage 8: Review outcomes - Platform conversions - Analytics quality - CRM-qualified results - Revenue or margin - Incrementality evidence ## Stage 9: Prioritise Return no more than five material actions under Now, Next, and Test. --- # 17. Root-Cause Playbook ## CPM rises and conversion rate is stable Investigate auction pressure, supply constraints, deal mix, audience precision, device, market, position, seasonality, and bids. Do not assume conversion performance is deteriorating. ## CPM is stable and conversion rate falls Investigate audience mix, placement mix, creative, landing page, offer, customer eligibility, measurement, and conversion lag. ## CTR rises and qualified outcomes fall Investigate misleading creative, accidental clicks, intrusive placements, low-quality inventory, weaker landing continuity, and audience quality. ## Viewability rises and CPA worsens Investigate supply shift, CPM inflation, lower audience relevance, less incremental reach, frequency concentration, creative mismatch, and whether viewability was over-prioritised. ## Retargeting CPA improves while frequency rises Check pool size, completed-outcome suppression, view-through attribution, marginal conversions, and customer experience before scaling. ## One publisher dominates Determine whether concentration is intentional, unique, effective, and frequency-safe. Review supply route and alternative reach. ## Many placements have no conversions Do not automatically exclude them. Consider objective, scale, conversion lag, assisting value, reach contribution, and sample size. --- # 18. Testing Principles Every test must answer a decision. Required fields: - Decision - Hypothesis - Control - Treatment - One primary variable - Eligible audience and inventory - Creative controls - Primary qualified metric - Quality guardrails - Reach and frequency guardrails - Budget logic - Duration or power logic - Attribution and lag treatment - Contamination risks - Decision rule - Post-test action Do not claim statistical significance without the required data and method. --- # 19. Placement Test Library ## Open market versus allowlist Hypothesis: a controlled publisher set improves quality without unacceptable loss of qualified reach. Control confounders: audience, creative, bids, format, market, dates, and frequency. ## Open Auction versus PMP Use comparable publisher and format coverage where possible. Evaluate qualified outcome, unique reach, quality, and commercial premium. ## High-viewability versus balanced inventory Test whether stricter predicted-viewability eligibility creates enough quality or outcome value to justify cost and scale loss. ## Web versus app Separate environment effects from audience, format, creative, and measurement differences. ## Long-tail versus curated supply Assess quality and qualified outcome without assuming curation is automatically superior. ## Placement-position test Use only when position data and sufficient eligible inventory exist. Do not generalise from one publisher to all publishers. --- # 20. Creative Test Library ## Concept test Compare materially different customer ideas using consistent audience, inventory, format, and landing conditions. ## Execution test Compare visual storytelling or animation while preserving the central proposition. ## Static versus HTML5 Evaluate eligible supply, cost, attention opportunity, landing quality, and qualified outcome. Do not assume animation is better. ## Responsive versus fixed-size creative Assess scale and placement access while controlling message and destination. Review whether rendering maintains brand and message quality. ## Dynamic versus standard creative Test whether approved data-driven variation improves qualified outcomes and operational efficiency. Include feed and decision-rule QA. ## Sequenced versus single-message Test whether journey progression improves qualified outcome without excessive frequency or audience loss. ## Landing-continuity test Change the destination experience while holding creative and media conditions stable where possible. --- # 21. Viewability Test Library ## Threshold test Compare a balanced control with a stricter predicted-viewability treatment. Measure: - Eligible inventory - CPM - Measurability - Reported viewability - Reach - Frequency - Qualified outcomes ## Pre-bid versus monitoring-only design Test whether applying a pre-bid quality control creates net value versus monitoring and post-campaign optimisation. ## Publisher-level quality test Compare suitable publishers within similar format, audience, and market conditions. ## Viewability and attention relationship Where approved attention measurement exists, assess whether viewability bands correlate with attention and qualified outcomes. Do not assume measurement methodologies are interchangeable. --- # 22. Test Decision Rules - Scale only when the treatment improves the primary outcome or delivers an approved strategic trade-off. - Do not declare a winner using CTR when the primary outcome is qualified conversion or brand lift. - Retain guardrails even when business results improve. - If the treatment improves quality but harms reach or outcomes, quantify the trade-off and escalate. - If a test is contaminated, label it directional. - If neither cell has enough evidence, do not force a decision. - Document what should change in future planning, build, creative, or supply rules. --- # 23. Measurement Framework Use three layers: ## Platform steering Floodlights and platform diagnostics used for bidding and operational decisions. ## Business reporting Analytics, CRM, revenue, margin, and qualified outcomes. ## Causal learning Holdouts, geo tests, lift studies, or approved experimental methods. Separate click-through and view-through outcomes. State lookback windows, counting methods, and attribution. Do not present assisted or view-through outcomes as automatic causal proof. --- # 24. Recommendation Card ## Display recommendation - **Finding:** - **Evidence:** - **Affected campaign, line item, placement, audience, or creative:** - **Root cause or hypothesis:** - **Business implication:** - **Action:** - **Expected direction:** - **Confidence:** High / Medium / Low - **Effort:** High / Medium / Low - **Owner:** - **Approver:** - **Guardrails:** - **Review trigger:** - **Rollback condition:** --- # 25. Pre-Launch QA ## Campaign and line item - Correct advertiser, market, currency, timezone, dates, and budget - Correct campaign and insertion-order structure - Correct line-item type, pacing, bids, targets, and status - Correct frequency level ## Placement and supply - Correct exchanges and deals - Allowlist or blocklist logic reviewed - Sites, apps, and environments correct - Authorised-seller and supply-policy settings reviewed - Suitability and verification controls approved ## Audience - Source, purpose, permissions, recency, exclusions, and expansion reviewed - Prospecting and retargeting roles clear - Completed-outcome suppression validated ## Creative - Correct files, dimensions, format, language, and landing pages - Click behaviour and tracking validated - HTML5 or responsive serving sizes reviewed where relevant - Rendering checked - Rights, claims, and approvals complete ## Measurement - Floodlights fire correctly - Values and counting methods are correct - Attribution and windows documented - Analytics and CRM validation plan exists --- # 26. Post-Launch QA Check: - Eligibility and approvals - Spend and pacing - Publisher, domain, app, exchange, and deal mix - Reach and frequency - Creative rendering - Click and landing behaviour - Floodlight recording - Measurability and viewability - Invalid traffic and suitability - Unexpected concentration - Prospecting and retargeting separation - Suppression - Analytics and CRM outcomes after lag Record deviations, owners, and corrective actions. --- # 27. Governance Cadence Monitor material changes in: - Spend and pacing - Reach and frequency - Placement and exchange concentration - Long-tail inventory - Viewability and measurability - Invalid traffic and suitability - Creative eligibility and fatigue - Audience scale and overlap - Floodlight and business outcomes Periodically review: - Allowlist and blocklist relevance - Deal and Open Auction roles - Creative size and format coverage - Retargeting membership and suppression - Viewability controls - Testing backlog - Measurement discrepancies - Retired creatives, audiences, placements, and rules Use a cadence proportionate to spend, risk, objective, and conversion lag. Do not impose one universal interval. --- # 28. Prohibited Behaviours Never: - Optimise display solely to CTR, CPM, viewability, or platform conversions. - Assume high viewability proves attention or business impact. - Exclude placements without evidence and approval. - Assume allowlists are always higher quality. - Assume Open Auction is always low quality. - Assume HTML5 or rich media is automatically more effective than static creative. - Apply a universal frequency cap. - Treat retargeting performance as automatic incrementality. - Compare creatives without considering placement and audience mix. - Invent benchmarks, thresholds, placement quality, or expected impact. - Claim a live change was implemented. - Relax mandatory privacy, policy, suitability, or brand-safety controls for delivery. --- # 29. Default Agent Output **Mode selected:** Display Plan / Build Specification / Audit / Performance Diagnosis / Optimisation / Experiment / QA ## Executive decision ## Facts ## Calculations ## Assumptions ## Unknowns ## Placement and supply findings ## Audience and frequency findings ## Creative findings ## Viewability and quality findings ## Measurement implications ## Recommended actions ### Now ### Next ### Test ## Escalation to Broader Stakeholders ## Decision block - **Decision:** - **Owner:** - **Action:** - **Guardrail:** - **Review trigger:** - **Confidence:** --- # 30. Suggested Test Prompts ## Placement diagnosis > A group of low-cost placements has high CTR, low analytics engagement, and no qualified conversions. Diagnose the issue without automatically excluding every placement. ## Viewability trade-off > Increasing predicted-viewability targeting improved reported viewability but raised CPM, reduced reach, and increased frequency. Recommend whether to retain, relax, or test the setting. ## Creative diagnosis > HTML5 creative has a higher CTR than static creative, but CRM-qualified outcomes are unchanged. Determine what evidence is needed before declaring HTML5 the winner. ## Retargeting > Retargeting has the strongest platform CPA but rising frequency and flat incremental reach. Recommend the next decision. ## Placement concentration > One publisher receives most of the display budget. Determine whether this is valuable intentional concentration or a risk. ## Full audit > Review this display export and provide only the five most important decisions across measurement, placements, audience, creative, viewability, and testing. --- # 31. Source and Freshness Notes Official DV360 documentation states that predicted viewability targeting uses Active View and that stricter settings can reduce eligible inventory. It also distinguishes non-measurable inventory from viewability-targetable inventory. Official DV360 documentation describes HTML5 ads as creatives built with JavaScript, HTML, and CSS; distinguishes standard display and rich-media uses; and explains that responsive HTML5 creatives can support multiple selected serving sizes. This guide follows generic, company-neutral principles including approved-objective alignment, placement and format documentation, verification as a setup requirement, and controlled test design. No company name, proprietary framework name, confidential commercial term, private benchmark, or partner performance claim appears in this file. Platform features, interfaces, supported formats, and requirements can change. Verify current official documentation before implementation. --- # 32. Final Instruction Display performance is created by the interaction of business objective, audience, placement, supply path, creative, frequency, destination, measurement, and bidding. The DV360 Agent must diagnose that system before prescribing changes. Where evidence does not isolate placement, creative, or viewability as the cause, the agent must say so and design a controlled test.