# Meta Reporting & Insights Framework v2.0 ## Enterprise Operating System Guide **Purpose:** Govern how Meta campaign data is validated, interpreted, converted into business insights, communicated to stakeholders, and translated into accountable action. **Companion guides:** Meta Setup & Activation, Measurement, Reach & Frequency, Optimization, Attribution & Incrementality, Testing & Experimentation, Creative Strategy, and Troubleshooting & Root Cause. **Core rule:** Metrics describe performance. Insights explain what changed, why it matters, and what decision should follow. Reporting is incomplete until evidence, business impact, action, risk, ownership, and confidence are visible. --- # 1. Enterprise Reporting Mandate The Meta Agent must answer: 1. What happened? 2. Against which target, baseline, forecast, or comparison? 3. Is the underlying measurement reliable? 4. Which drivers explain the movement? 5. What is known, inferred, assumed, or unresolved? 6. What is the business impact? 7. What should happen next? 8. What risk accompanies the recommendation? 9. Who owns and approves the action? 10. When will the result be reviewed? The Agent must not convert a metric movement directly into a recommendation without validating measurement and investigating root cause. --- # 2. Reporting Principles - Metrics are not insights. - Every material claim requires evidence. - Facts, calculations, interpretations, assumptions, and recommendations must remain distinct. - The reporting source depends on the business question. - Meta-attributed outcomes are not automatically incremental outcomes. - Platform revenue is not automatically net revenue, margin, or profit. - Average frequency does not describe exposure distribution. - A high-performing segment may not be scalable. - Every recommendation requires an owner, risk, review trigger, and rollback condition where material. - Reports should focus attention on decisions, not maximize the number of metrics. - Comparisons require aligned definitions, periods, currencies, events, and attribution settings. - Never invent root causes, benchmarks, lift, or confidence. --- # 3. Reporting Architecture ## Data layer Includes Meta delivery, reach, frequency, events, attributed outcomes, analytics, CRM, commerce, finance, experiments, and approved external context. ## Validation layer Checks event quality, deduplication, attribution, currency, timezone, latency, qualification, refunds, cancellations, and report scope. ## Diagnostic layer Identifies mathematical drivers and segments movement by audience, creative, placement, market, device, time, journey stage, and customer type. ## Intelligence layer Converts validated findings into explanations, implications, risks, and opportunities. ## Decision layer Defines action, owner, approval, review timing, guardrails, and rollback. ## Learning layer Stores insights, decisions, outcomes, and reusable learning in enterprise registries. --- # 4. Decision Tree 1: Reporting Mode Selection ```text Who is the audience? Executive or client leadership → Business outcome and decision view Account or strategy team → Business, channel, and learning view Activation or trading team → Diagnostic and action view Measurement team → Source, methodology, and reconciliation view Finance or commercial team → Validated value, cost, margin, and risk view What decision is required? No decision → Provide concise status and material risks Decision required → Apply full insight and recommendation governance ``` --- # 5. Source-of-Truth Framework ## Meta Use for platform delivery, supported reach and frequency, campaign attribution, and operational diagnostics. ## Analytics Use for approved website or app behaviour, sessions, engagement, funnel progression, and analytics attribution. ## CRM Use for created, accepted, qualified, opportunity, and closed lead stages. ## Commerce or booking system Use for confirmed orders, bookings, subscriptions, availability, refunds, and cancellations. ## Finance Use for approved net revenue, margin, profit, and commercial reconciliation. ## Controlled studies Use for scoped incremental or lift conclusions. No source is universally superior. Select the source appropriate to the question and explain relevant differences. --- # 6. Decision Tree 2: Source Selection ```text Need campaign delivery or Meta credit? → Meta Need site or app behaviour? → Approved analytics source Need qualified lead or customer outcome? → CRM or commerce Need net revenue, margin, or profit? → Finance-approved source Need causal media impact? → Approved incrementality study Sources disagree? → Reconcile before drawing a business conclusion ``` --- # 7. KPI Hierarchy ## Tier 1: Business outcomes - Qualified leads - Customers - Net revenue - Margin - Profit - Customer value - Retention ## Tier 2: Marketing outcomes - Validated conversions - Qualified CPA - Attributed ROAS - Incremental conversions - Incremental return - Brand or consideration movement ## Tier 3: Journey outcomes - Landing-page views - Sessions - Product views - Cart or form starts - Completed forms - Message starts ## Tier 4: Delivery and quality - Spend - Impressions - Reach - Frequency - CPM - CTR - Video consumption - Placement delivery Higher tiers should govern business conclusions. Lower tiers explain delivery and diagnostic movement. --- # 8. Decision Tree 3: KPI Selection ```text Is a validated business outcome available? Yes → Use as primary KPI No ↓ Is a validated marketing outcome available? Yes → Use with limitation and improvement plan No ↓ Use the closest validated journey proxy ↓ Do not present delivery metrics as business success ``` --- # 9. Insight Taxonomy Classify each insight so readers understand its role. ## Performance insight Explains movement against target, baseline, forecast, or previous period. ## Diagnostic insight Identifies a supported root cause or contributing factor. ## Strategic insight Changes audience, creative, campaign, market, or channel strategy. ## Commercial insight Relates to revenue, margin, customer value, or investment efficiency. ## Risk insight Highlights measurement, delivery, brand, policy, customer, or commercial exposure. ## Attribution insight Explains configured credit without causal language. ## Incrementality insight Reports causal evidence from an approved method. ## Experimentation insight Translates a controlled test into a rollout, rejection, or retest decision. --- # 10. Executive Narrative Engine Every material narrative should follow this sequence: ## Situation State the business context, objective, period, and comparison. ## Change State what moved and by how much when validated data is available. ## Driver State the supported root cause or leading hypothesis. Separate causal evidence from correlation. ## Impact Explain the effect on the business target, customer, revenue, risk, or opportunity. ## Action Provide the narrowest useful recommendation. ## Risk State what could go wrong, what uncertainty remains, and the risk of doing nothing. ## Owner Assign responsibility and accountable approval. ## Review State the evidence and trigger required to validate or reverse the action. ### Narrative template ```markdown Situation: Change: Evidence: Driver: Business Impact: Recommended Action: Risk and Trade-off: Responsible Owner: Accountable Approver: Review Trigger: Rollback Condition: Confidence: ``` --- # 11. Decision Tree 4: Insight Construction ```text Material finding exists? No → Do not manufacture an insight Yes ↓ Measurement validated? No → Report data-quality issue first Yes ↓ Root cause supported? No → Label as hypothesis and request evidence or test Yes ↓ Business impact identified? No → Keep as operational observation Yes ↓ Action, risk, owner, and review defined? No → Insight is incomplete Yes → Publish after QA ``` --- # 12. Insight Quality Scorecard Assess every material insight against: - **Accuracy:** Data and definitions are validated. - **Evidence:** The conclusion is supported by relevant analysis. - **Relevance:** The insight addresses the business objective. - **Causality discipline:** Correlation and causation are distinguished. - **Actionability:** A practical decision can follow. - **Commercial value:** Business impact is clear. - **Risk visibility:** Trade-offs and uncertainty are stated. - **Ownership:** Action and approval are assigned. - **Timeliness:** The insight remains useful when delivered. - **Reusability:** Learning can inform future work. Use qualitative classifications such as strong, adequate, provisional, or reject unless an approved scoring system exists. --- # 13. Executive Reporting Framework Executive reports should prioritize: 1. Business outcome 2. Material variance 3. Root cause or evidence gap 4. Attribution and incrementality distinction 5. Decision required 6. Risk and opportunity 7. Owner and review date Avoid exhaustive metric tables in the executive narrative. Place detailed diagnostics in an appendix or operational section. --- # 14. Operational Reporting Framework Operational reporting should support action across: - Measurement health - Budget and pacing - Audience and overlap - Creative and delivery concentration - Placement and device - Reach and frequency - Conversion journey - Attribution - Qualified outcomes - Changes made and results Classify actions as: - Now - Next - Test - Monitor - Escalate --- # 15. Client and Leadership Reporting Translate platform language into business language. Instead of: > CPM increased and frequency moved higher. Use: > Media costs increased while reach growth slowed, concentrating more exposure among people already reached. This raises saturation risk and reduces the likely value of additional spend unless the audience, creative, or placement opportunity expands. Do not remove necessary methodology caveats in the name of simplicity. --- # 16. Attribution and Incrementality Reporting Report attribution and incrementality separately. ## Attribution view - Event - Setting - Interaction types - Windows - Attributed result - Limitations ## Incrementality view - Method - Control and treatment - Primary outcome - Result and uncertainty - Practical significance - Scope of inference ## Business-value view - Qualified outcome - Revenue basis - Margin or profit where available - Customer status --- # 17. Decision Tree 5: Attribution or Incrementality Narrative ```text Does the claim concern credit? → Use attribution language Does the claim concern customer paths? → Use descriptive journey language Does the claim concern caused outcomes? → Require incrementality evidence Does the claim concern profit? → Require finance-approved value No causal evidence? → Do not use causal language ``` --- # 18. Reach and Frequency Reporting Include: - Reach unit and source - Average frequency - Frequency distribution where supported - Audience coverage - Saturation indicators - Overlap or duplication - Added or incremental reach method - Cost per reached or additional person - Quality and business outcomes Never use average frequency alone. Never add reach across periods or channels without valid deduplication. --- # 19. Creative Insight Framework Assess: - Concept - Hook - Offer - Format - Placement compatibility - Delivery share - Reach and frequency - Engagement quality - Conversion journey - Qualified outcome - Scalability A creative with high CTR but weak business outcomes is not automatically a winner. Separate concept learning from execution learning. --- # 20. Decision Tree 6: Creative Diagnosis ```text Creative performance changed? ↓ Measurement and delivery comparable? No → Do not declare a winner Yes ↓ CTR change only? → Check landing and qualified outcome Frequency rising and reach flat? → Fatigue hypothesis Delivery concentrated? → Allocation or eligibility issue Business outcome improved and scalable? → Controlled rollout ``` --- # 21. Audience and Placement Insights ## Audience Report scale, overlap, reach, frequency, customer quality, attribution, incrementality, and marginal headroom. ## Placement Report delivery role, format suitability, reach contribution, cost, customer journey, business outcome, and quality. Do not remove an audience or placement solely from one weak metric without assessing role and trade-offs. --- # 22. Optimization and Experiment Reporting For each material optimization, document: - Problem - Root cause - Change - Expected effect - Actual result - Guardrails - Rollback status - Learning For each experiment, document: - Business question - Hypothesis - Control - Treatment - Primary outcome - Result class - Practical significance - Decision - Rollout or retest status --- # 23. Decision Tree 7: Experiment Reporting ```text Experiment complete and measurement mature? No → Report as in progress Yes ↓ Result valid? No → Classify invalid Yes ↓ Positive, negative, neutral, or inconclusive? ↓ Commercially meaningful and scalable? Yes → Controlled rollout No → Reject, maintain, or retest based on decision value ``` --- # 24. Budget Allocation and Marginal Reporting Budget recommendations must consider: - Qualified business outcome - Attribution - Incrementality - Marginal return - Reach and saturation - Creative capacity - Market opportunity - Strategic role - Commercial commitment - Measurement confidence Average CPA or ROAS does not answer where the next dollar should go. --- # 25. Decision Tree 8: Budget Reallocation ```text Reallocation proposed? ↓ Outcomes comparable and qualified? No → Reconcile first Yes ↓ Incremental and marginal evidence available? No → Stage change or test Yes ↓ Strategic, reach, learning, or market role protected? No → Redesign allocation Yes → Approve with guardrails and rollback ``` --- # 26. Forecast and Variance Framework Compare planned, forecast, and actual outcomes using aligned definitions. Diagnose variance across: - Budget - Pacing - CPM - Reach - Frequency - Conversion rate - Conversion volume - Order or lead value - Customer quality - Market or product availability - Attribution and measurement A forecast miss can result from an invalid assumption, implementation issue, external change, or performance movement. --- # 27. Decision Tree 9: Forecast Variance ```text Material variance detected? ↓ Definitions and data aligned? No → Correct comparison Yes ↓ Volume, cost, response rate, or value driver? ↓ Internal, external, or measurement cause? ↓ Temporary or persistent? ↓ Reforecast, correct, test, or escalate ``` --- # 28. Recommendation Governance Every material recommendation must include: - Current state - Proposed action - Evidence - Expected effect - Business impact - Risk and trade-off - Responsible owner - Accountable approver - QA reviewer where needed - Review trigger - Rollback condition - Confidence Do not publish vague recommendations such as "optimize audiences" or "refresh creative." Specify the decision and validation method. --- # 29. Decision Tree 10: Recommendation Validation ```text Action specific? No → Rewrite Evidence supports action? No → Label as test Business impact clear? No → Keep as operational task Risk stated? No → Add trade-off Owner and approval assigned? No → Recommendation incomplete Review and rollback defined? No → Add governance before publication ``` --- # 30. Confidence Framework ## High Validated, directly relevant evidence supports the conclusion. Definitions align and material uncertainty is limited. ## Medium Useful evidence supports the conclusion, but assumptions, coverage, comparability, or causality remain limited. ## Low The conclusion is preliminary because critical data, definitions, scale, or control evidence is missing. Confidence applies to the conclusion, not the importance of the issue. --- # 31. Decision Tree 11: Confidence Assignment ```text Measurement validated? No → Low Root cause directly supported? No → Low or Medium Sources and definitions aligned? No → Medium at best Causal claim supported by valid experiment? No → Do not claim causality Material unknowns limited? Yes → High may be appropriate ``` --- # 32. Publication and Escalation Reports with material financial, customer, measurement, policy, legal, privacy, or reputational implications require appropriate accountable review. Escalate when: - The report could trigger material investment movement. - Measurement integrity is unresolved. - A causal claim lacks valid evidence. - A policy or customer-risk issue exists. - A recommendation is difficult to reverse. - Sources materially conflict. --- # 33. Decision Tree 12: Publication Approval ```text QA complete? No → Do not publish Material business impact? Yes → Accountable leadership review Source conflict unresolved? Yes → Publish limitation or block conclusion Recommendation actionable and governed? No → Return for revision All required approvals complete? Yes → Publish and register ``` --- # 34. Reporting QA Checklist ```markdown # Meta Reporting QA - Report ID: - Audience and purpose: - Business question: - Campaign, market, and period: - Comparison basis: - Primary source of truth: - Meta event and attribution setting: - Reporting timezone and currency: - Conversion and qualification lag applied: - Deduplication and event quality validated: - CRM, commerce, or finance reconciled: - Reach and frequency methodology stated: - Attribution and incrementality separated: - Facts and assumptions separated: - Root cause supported or labelled hypothesis: - Business impact included: - Recommendation specific: - Risk and trade-off included: - Owner and approver assigned: - Review trigger and rollback included: - Confidence included: - Sources and limitations visible: - Status: Approved / Conditional / Blocked ``` --- # 35. Reporting RACI ```markdown | Activity | Responsible | Accountable | Consulted | Informed | |---|---|---|---|---| | Define reporting objective | | | | | | Validate Meta data | | | | | | Validate analytics and CRM | | | | | | Validate finance outcomes | | | | | | Analyse reach and frequency | | | | | | Analyse attribution | | | | | | Analyse incrementality | | | | | | Conduct root-cause analysis | | | | | | Draft insights | | | | | | Approve recommendations | | | | | | Complete reporting QA | | | | | | Approve publication | | | | | | Distribute report | | | | | | Track actions | | | | | | Update registries | | | | | ``` One accountable owner is required for report accuracy and one for material business decisions. --- # 36. Reporting Registry ```markdown | Report ID | Status | Market | Campaign | Period | Audience | Owner | Approver | Publication Date | Next Review | |---|---|---|---|---|---|---|---|---|---| | No registered reports | Unavailable | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | ``` ## Full record ```markdown - Business question: - Source hierarchy: - Measurement limitations: - Material findings: - Decisions required: - Approved recommendations: - Risks: - Owners: - Review triggers: - Related insight IDs: - Related experiment IDs: - Previous report: ``` --- # 37. Insight Registry ```markdown | Insight ID | Type | Finding | Root Cause or Hypothesis | Business Impact | Action | Owner | Confidence | Status | Review Date | |---|---|---|---|---|---|---|---|---|---| | No registered insights | Unavailable | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | ``` Statuses can include proposed, approved, active, validated, rejected, superseded, or closed. Do not retain contradictory insights without documenting why one supersedes another. --- # 38. Reporting Audit Framework Audit: - Business question and audience - Source-of-truth decisions - Metric definitions - Event, attribution, currency, and timezone - CRM and finance reconciliation - Reach and frequency methodology - Attribution and incrementality language - Root-cause quality - Insight quality - Recommendation specificity - Risks, ownership, and review dates - Registry completeness - Action follow-through - Historical consistency Classify findings: - Retain - Correct - Restate - Reclassify as observational - Withdraw - Escalate --- # 39. Insight Pattern Library This library contains **150 diagnostic patterns**. Patterns are prompts for investigation, not automatic conclusions. Validate measurement, definitions, context, business outcomes, and causality before using any pattern in a report. ## Cost, efficiency, and conversion 1. **CPA up + CPM up + CVR stable:** Auction, audience, placement, seasonality, or inventory cost pressure is the leading hypothesis. 2. **CPA up + CPM stable + CVR down:** Audience, creative, offer, landing experience, or event quality requires diagnosis. 3. **CPA up + spend flat + conversions down:** Response deterioration, event loss, or conversion lag is likely driving the change. 4. **CPA down + qualified-lead rate down:** Apparent efficiency may be driven by weaker lead quality. 5. **CPA down + volume down:** Efficiency may reflect reduced scale rather than better performance. 6. **CPA down + volume up + quality stable:** The campaign may have a credible scaling opportunity, subject to marginal validation. 7. **CPL down + sales acceptance down:** Lead-generation efficiency is not translating into sales value. 8. **Conversions up + spend stable:** Conversion efficiency improved; validate event accuracy and customer quality. 9. **Conversions flat + spend up:** Marginal efficiency is deteriorating or reporting lag is unresolved. 10. **Conversions down + spend down proportionally:** The primary driver may be investment reduction rather than campaign deterioration. 11. **CVR down + landing-page views stable:** The issue likely begins after landing, such as offer, form, checkout, or product availability. 12. **CTR up + CVR down:** The ad may attract clicks that do not match the conversion promise. 13. **CTR down + CVR stable:** Click engagement weakened, but business conversion efficiency among clickers remained stable. 14. **CPM down + CPA up:** Cheaper delivery is not producing proportionate qualified response. 15. **CPM up + CPA stable:** Higher media cost is being offset by stronger response or value. ## Revenue, value, and profitability 16. **ROAS up + net revenue flat:** Platform attribution may have increased without total commercial growth. 17. **ROAS up + margin down:** Product mix, discounting, refunds, fulfillment, or existing-customer concentration may reduce profit. 18. **ROAS down + conversions stable:** Average order value, revenue accuracy, or product mix may be weakening. 19. **ROAS down + finance revenue stable:** The movement may be attribution or platform-value reclassification rather than business decline. 20. **Revenue up + orders flat:** Average order value increased; validate product mix and margin. 21. **Orders up + revenue flat:** Average order value declined or low-value products dominate. 22. **Attributed revenue up + refunds up:** Reported growth may overstate retained commercial value. 23. **Attributed revenue up + new-customer share down:** Efficiency may be increasingly dependent on existing customers. 24. **New-customer revenue up + blended ROAS down:** Acquisition may be expanding at a higher but strategically acceptable cost. 25. **Profit up + ROAS flat:** Margin or product-mix improvement may be creating value not visible in platform ROAS. 26. **Gross revenue up + net revenue down:** Refunds, cancellations, tax, fees, or fulfillment issues require reconciliation. 27. **Purchase value spikes + order count stable:** Validate event values, currency, quantities, and duplicate order IDs. 28. **Purchase count spikes + finance orders stable:** Duplicate or misfired Purchase events are a leading measurement risk. 29. **AOV up + conversion rate down:** The campaign may be reaching fewer but higher-value customers. 30. **Margin stable + media cost up:** Incremental profitability may be under pressure even if revenue holds. ## Reach, frequency, and saturation 31. **Spend up + reach flat + frequency up:** Audience saturation, overlap, or limited eligible inventory is a leading hypothesis. 32. **Reach up + frequency stable + qualified outcomes up:** Broader delivery may be adding useful scale. 33. **Reach up + outcomes flat:** Audience relevance, creative, or the customer journey may be weak. 34. **Reach up + viewability down:** Expansion may be accessing lower-quality inventory or placements. 35. **Reach down + CPM up:** Cost pressure, a narrower audience, or inventory restrictions may be limiting coverage. 36. **Reach down + frequency up:** Delivery is concentrating among fewer people. 37. **Reach flat + frequency flat + spend down:** Lower investment may have preserved coverage through cheaper or more efficient delivery. 38. **Frequency up + CTR down:** Creative fatigue or audience exhaustion is a hypothesis, not a proven cause. 39. **Frequency up + CVR stable:** Additional repetition has not yet changed conversion efficiency; assess marginal value. 40. **Frequency up + CPA up:** Repeated exposure may be adding cost without sufficient response. 41. **Frequency down + reach up:** Budget may be spreading across a broader population. 42. **Frequency down + brand lift down:** Reduced reinforcement may be relevant, but causal interpretation requires study design review. 43. **Average frequency stable + complaints up:** Heavy-exposure pockets may be hidden by the average. 44. **Added reach down + total reach up:** Expansion may be increasingly duplicative relative to the stated comparison. 45. **Cost per reached person up + CPM stable:** Reach growth is slowing even though impression cost is stable. ## Audience and customer mix 46. **Broad audience CPA down + quality stable:** Consolidation or broader scaling may be viable. 47. **Audience expansion CPA down + qualified rate down:** Expansion may be trading quality for volume. 48. **Lookalike CPA up + source-list age up:** Seed freshness or relevance may be deteriorating. 49. **Prospecting attribution weak + Conversion Lift positive:** Attribution may understate demand creation. 50. **Remarketing attribution strong + Conversion Lift weak:** Existing demand may be driving reported efficiency. 51. **Existing customers dominate revenue:** Acquisition effectiveness may be overstated. 52. **New-customer share up + CPA up:** Higher acquisition cost may reflect genuine expansion into new demand. 53. **Audience overlap up + reach flat:** Campaigns may be competing for the same people. 54. **Retargeting pool down + frequency up:** The pool may be shrinking while spend remains too high. 55. **CRM audience match down + delivery down:** Signal coverage may be constraining audience eligibility. 56. **Interest audience CTR up + qualified leads flat:** Engagement is not translating into deeper value. 57. **Broad audience reach up + qualified CPA stable:** The platform may be finding additional relevant scale. 58. **Customer suppression removed + ROAS up:** Reported efficiency may be increasingly driven by existing customers. 59. **High-value audience volume down + total conversions up:** Growth may be shifting toward lower-value customers. 60. **Audience size up + spend flat + frequency down:** Expanded eligibility is distributing the same budget more broadly. ## Creative and message 61. **CTR down + qualified outcome stable:** Avoid unnecessary creative changes when business performance remains healthy. 62. **CTR up + qualified outcome down:** Message promise, click quality, or destination continuity may be weak. 63. **One creative dominates delivery + no controlled test:** Delivery share cannot establish creative superiority. 64. **New concept improves qualified outcome + frequency stable:** The concept may be stronger and eligible for controlled rollout. 65. **Creative winner fails after scale:** Audience, placement, frequency, market mix, or implementation may have changed. 66. **Video completion up + site engagement flat:** Content consumption did not translate into deeper journey activity. 67. **Short-form CTR up + lead quality down:** The execution may win attention without attracting qualified intent. 68. **Static CPA down + incremental lift unknown:** Operational efficiency improved, but causal superiority remains unproven. 69. **Creative refresh increases CPM + improves CVR:** Higher cost may be justified by stronger customer response. 70. **Creative refresh increases reach + lowers frequency:** More eligible assets may be broadening delivery. 71. **Hook rate up + completion down:** The opening attracts attention but later content may not sustain it. 72. **Completion stable + CPA up:** The problem likely sits outside video consumption, such as audience, cost, or destination. 73. **Offer creative wins + margin down:** The offer may drive volume through economically harmful discounting. 74. **Localized creative CTR up + CVR down:** Translation or landing-page continuity may not support the localized promise. 75. **Partnership ad engagement up + sales flat:** Social proof may increase interaction without changing commercial outcomes. ## Placement, device, and delivery 76. **Reels share up + CPA up:** Check creative suitability, audience mix, click quality, and whether placement shift caused the change. 77. **Stories reach up + landing-page views flat:** Swipe or click behaviour may not be translating into completed page loads. 78. **Feed CPM up + qualified rate up:** Premium cost may be offset by stronger customer quality. 79. **Automatic placements CPA down + quality stable:** Broader placement allocation may be improving efficiency. 80. **Manual placements delivery down + CPM up:** Restrictions may be reducing eligible inventory and competition efficiency. 81. **Mobile clicks up + desktop conversions flat:** Cross-device journey or mobile destination quality requires review. 82. **Android conversions down + iOS stable:** Operating-system-specific app, site, tracking, or product issues may exist. 83. **Placement reach up + viewability down:** Scale may be coming from lower-opportunity-to-see environments. 84. **Audience Network volume up + CRM quality down:** Placement mix may be contributing lower-quality leads, subject to controlled validation. 85. **In-stream completion up + reach flat:** More video completion may be occurring among people already reached. 86. **Device CPM down + CPA up:** Cheaper device delivery is not producing proportionate outcomes. 87. **Placement excluded + reach down + CPA stable:** The exclusion reduced scale without improving efficiency. 88. **Placement excluded + CPA down + frequency up:** Efficiency improved but delivery may be concentrating elsewhere. 89. **Creative not eligible for key placement + spend down:** Asset coverage may be limiting delivery opportunity. 90. **Delivery shifts market + blended CPA improves:** Mix change may explain account-level improvement rather than within-market gains. ## Measurement, signal, and attribution 91. **Meta conversions up + CRM sales flat:** Attribution, duplicates, qualification, or channel substitution requires review. 92. **Meta conversions down + CRM sales stable:** Signal loss or attribution changes may be reducing observed platform credit. 93. **Pixel events up + server events flat:** Browser-side change or duplication requires investigation. 94. **Server events up + source-system outcomes flat:** Validate payload, triggers, event IDs, and duplicates. 95. **Event Match Quality down + attributed conversions down:** Signal degradation may contribute, but direct causality should not be assumed. 96. **Events Manager count up + Ads Manager conversions flat:** Event receipt increased without additional attributed campaign outcomes. 97. **View attribution up + click attribution flat:** Exposure-based credit is growing; assess frequency and incrementality. 98. **Attribution window shortened + reported conversions down:** Treat the change as credit reclassification until business outcomes move. 99. **Attribution window widened + ROAS up:** More conversions qualified for credit; incremental value has not been proven. 100. **Direct traffic up + Meta spend up:** Potential influence is observational and requires controlled evidence. 101. **CRM sales up + Meta attribution flat:** Meta may be under-credited, or another factor may drive growth. 102. **Finance revenue down + Meta revenue stable:** Platform reporting may not reflect refunds, cancellations, or net-value changes. 103. **Duplicate-event diagnostics rise + ROAS rises:** Performance improvement may be measurement inflation. 104. **Consent rate down + tracked conversions down:** Observed decline may partly reflect lower measurement coverage. 105. **Reporting latency up + recent performance down:** The newest period may be incomplete rather than truly weaker. ## Incrementality and experimentation 106. **Attribution strong + lift weak:** Media may be receiving credit for outcomes that would have occurred anyway. 107. **Attribution weak + lift strong:** Configured credit may miss causal value. 108. **Brand Lift positive + sales flat:** Perception moved, but conversion conditions or lag may prevent immediate sales movement. 109. **Conversion Lift positive + incremental ROAS weak:** The campaign causes outcomes, but commercial value may not justify cost. 110. **Lift positive + high uncertainty:** Use staged rollout or retest rather than treating the point estimate as certain. 111. **Lift neutral + operational CPA improves:** Efficiency changed, but causal superiority is unproven. 112. **Test winner positive + practical impact small:** Statistical evidence may not justify implementation cost or complexity. 113. **Test winner scales + effect reverses:** Test conditions may not represent scaled delivery. 114. **Holdout contamination up + measured lift down:** Control exposure may dilute the estimated treatment effect. 115. **Geo pre-period fit weak + estimated lift high:** The causal estimate may be unreliable despite an attractive result. 116. **Incremental reach up + incremental conversions flat:** The newly reached population may be less responsive or the message ineffective. 117. **Frequency treatment up + attribution up + lift flat:** More exposure may be increasing credit opportunities without additional outcomes. 118. **Creative test CTR win + qualified outcome neutral:** The creative changes engagement but not the business decision metric. 119. **Audience test CPA win + new-customer share down:** The winner may capture easier existing demand. 120. **Retest reproduces direction + smaller effect:** The effect may be real but less commercially meaningful than the first estimate. ## Budget, pacing, and marginal investment 121. **Spend below plan + delivery constrained:** Review audience, placements, event, bid, creative, approvals, and account status. 122. **Spend below plan + CPA below target:** Efficiency is healthy, but volume opportunity may be constrained. 123. **Spend above plan + outcome on target:** Financial-control risk remains even when performance is acceptable. 124. **Budget up + marginal CPA up:** The next investment is less efficient than the historical average. 125. **Budget up + reach up + qualified CPA stable:** Controlled scale may remain viable. 126. **Budget shifted to lowest CPA + total sales flat:** Reallocation may have concentrated demand capture rather than created growth. 127. **Campaign budget optimizes away from strategic market:** Relative signals may conflict with market commitments or long-term strategy. 128. **Cost control tightens + volume collapses:** Nominal efficiency may be purchased through lost qualified scale. 129. **Cost control relaxes + volume up + quality down:** Additional delivery may be entering weaker opportunities. 130. **Daily pacing catches up + CPA spikes:** Aggressive late-period spending may reduce marginal quality. 131. **Spend flat + CPM up + impressions down:** Higher auction cost is reducing delivered volume. 132. **Spend flat + reach up + frequency down:** Delivery may be spreading more efficiently across new people. 133. **Marginal ROAS below break-even + blended ROAS healthy:** Historical averages may conceal poor next-dollar value. 134. **Market budget down + total CPA improves:** Mix shift may explain improvement without any market becoming more efficient. 135. **Test budget underdelivers + result inconclusive:** The study may lack sufficient treatment delivery for a decision. ## Forecasting, market, and operations 136. **Forecast miss + media delivery on plan:** Conversion, value, demand, product, or external assumptions may be wrong. 137. **Forecast miss + spend below plan:** Delivery shortfall is a direct contributor and requires root-cause diagnosis. 138. **Forecast beat + lead quality down:** Volume outperformance may conceal weaker business value. 139. **Forecast beat + margin stable:** Upside may be commercially valid, subject to measurement and incrementality checks. 140. **Market CPA up + all channels weaken:** External demand or market conditions may be contributing. 141. **One market improves + currency changed:** Reconcile currency before claiming performance improvement. 142. **Sales down + product unavailable:** Media may be unable to convert demand because of a business constraint. 143. **Clicks stable + site sessions down:** Page load, redirects, consent, or analytics implementation may be failing. 144. **Lead response time up + qualification down:** Operational follow-up may be reducing sales value. 145. **Campaign approved + no delivery:** Check schedule, billing, account status, audience, bid, creative review, and event. 146. **Report published + owner missing:** The recommendation is not operationally complete. 147. **Recommendation implemented + no review trigger:** The organization cannot determine whether to continue or roll back. 148. **Market disruption + performance change:** Rebaseline or use controlled analysis before attributing the movement to media. 149. **Promotion begins + conversion rate up:** Separate promotional demand from media optimization impact. 150. **Promotion ends + CPA up:** Post-promotion normalization may explain deterioration. --- # 40. Expanded Diagnosis Templates Each template follows: finding, evidence required, likely causes, business impact, action options, risks, and confidence guidance. ## 1. CPA Rising - **Finding:** Cost per attributed or qualified outcome increased. - **Evidence required:** Spend, CPM, conversion volume, CVR, event health, audience, creative, placement, landing journey, and qualification. - **Likely causes:** Cost inflation, lower response, saturation, weaker offer, tracking change, or lower-quality traffic. - **Business impact:** Reduced volume at fixed budget or higher investment needed to meet target. - **Action options:** Repair measurement, isolate driver, refresh creative, expand audience, improve journey, or reallocate marginal spend. - **Risks:** Shifting to a shallow event can manufacture a lower CPA. - **Confidence:** High only when the mathematical driver and segment are validated. ## 2. CPA Falling - **Finding:** Reported cost per outcome improved. - **Evidence required:** Qualified outcome, customer mix, event definition, spend, volume, margin, and incrementality. - **Likely causes:** Better delivery, stronger creative, cheaper inventory, lower-quality outcomes, or attribution change. - **Business impact:** Potential scale opportunity or misleading efficiency. - **Action options:** Validate quality and marginal headroom before scaling. - **Risks:** Efficiency may be driven by remarketing or existing demand. - **Confidence:** Medium until downstream quality is confirmed. ## 3. ROAS Falling - **Finding:** Attributed revenue grew slower than spend or declined. - **Evidence required:** Revenue accuracy, order value, conversions, refunds, product mix, customer status, and attribution. - **Likely causes:** Lower CVR, smaller orders, cost growth, value error, or saturation. - **Business impact:** Revenue efficiency and potentially profit at risk. - **Action options:** Correct value, improve journey, adjust product strategy, or reallocate based on marginal value. - **Risks:** Cutting prospecting may protect short-term ROAS while harming incremental demand. - **Confidence:** Depends on finance reconciliation. ## 4. ROAS Rising - **Finding:** Attributed revenue increased faster than spend. - **Evidence required:** Net revenue, margin, refunds, customer mix, and lift. - **Likely causes:** Better conversion, order value, remarketing concentration, attribution shift, or real incremental growth. - **Business impact:** Possible scale opportunity. - **Action options:** Validate profitability and incremental value, then stage scale. - **Risks:** Platform credit may increase without total business growth. - **Confidence:** Medium until commercial validation. ## 5. Reach Plateau - **Finding:** Unique reach growth slowed while investment continued. - **Evidence required:** Reach source, frequency distribution, audience size, overlap, CPM, placements, and period. - **Likely causes:** Saturation, narrow audience, restricted placements, creative ineligibility, or rising costs. - **Business impact:** Additional spend may create repetition rather than coverage. - **Action options:** Expand eligible opportunity or reallocate marginal budget. - **Risks:** Broadening may reduce relevance or quality. - **Confidence:** High when deduplication and distribution are valid. ## 6. Reach Expansion - **Finding:** More unique people were reached. - **Evidence required:** Reach methodology, cost, frequency, audience quality, incremental comparison, and outcomes. - **Likely causes:** Audience, placement, market, creative, or budget expansion. - **Business impact:** Greater awareness opportunity. - **Action options:** Maintain if quality and marginal results hold. - **Risks:** Total reach can rise without incremental business value. - **Confidence:** Medium without outcome evidence. ## 7. Frequency Inflation - **Finding:** Repeated exposure increased faster than reach. - **Evidence required:** Distribution, audience size, overlap, budget, creative delivery, and customer response. - **Likely causes:** Saturation, small retargeting pools, campaign overlap, or limited placements. - **Business impact:** Waste, fatigue, or customer irritation risk. - **Action options:** Coordinate caps, expand opportunity, rotate concepts, or reduce spend. - **Risks:** Overcorrection may remove useful repetition. - **Confidence:** Medium unless outcome-by-exposure is tested. ## 8. Frequency Decline - **Finding:** Average or distributed exposure declined. - **Evidence required:** Reach growth, campaign duration, audience expansion, spend, placements, and outcomes. - **Likely causes:** Wider audience, lower budget, new creative eligibility, or delivery disruption. - **Business impact:** Message reinforcement may weaken or coverage may improve. - **Action options:** Judge against strategy and validated outcomes. - **Risks:** Treating lower frequency as failure without objective context. - **Confidence:** Medium. ## 9. Audience Saturation - **Finding:** Reach slows, frequency rises, and marginal outcomes weaken. - **Evidence required:** Reach curve, distribution, audience pool, overlap, and marginal CPA or value. - **Likely causes:** Small eligible population, excessive budget, stale remarketing, or narrow controls. - **Business impact:** Diminishing return. - **Action options:** Expand, refresh, reduce, or test another audience. - **Risks:** Expansion can lower lead quality. - **Confidence:** High when several indicators align. ## 10. Audience Overlap - **Finding:** Campaigns or audience groups target materially similar populations. - **Evidence required:** Audience definitions, exclusions, reach duplication, frequency, and auction delivery. - **Likely causes:** Redundant segmentation or disconnected campaign ownership. - **Business impact:** Internal competition and frequency concentration. - **Action options:** Consolidate, exclude, coordinate, or preserve intentional test cells. - **Risks:** Removing overlap can reduce scale or disrupt tests. - **Confidence:** Depends on overlap measurement method. ## 11. Creative Fatigue - **Finding:** Creative response weakens as exposure accumulates. - **Evidence required:** Reach, frequency, delivery share, CTR, CVR, qualified outcome, and time. - **Likely causes:** Repetition, concept exhaustion, market change, or audience shift. - **Business impact:** Lower response and higher cost. - **Action options:** Test new concepts or sequences. - **Risks:** Frequent cosmetic swaps can disrupt learning without solving concept weakness. - **Confidence:** Medium unless frequency effect is controlled. ## 12. Creative Winner Fails After Rollout - **Finding:** Test winner underperforms at wider scale. - **Evidence required:** Audience, placement, market, frequency, budget, and implementation fidelity. - **Likely causes:** Changed delivery mix, regression, small test, or rollout error. - **Business impact:** Expected gains fail to materialize. - **Action options:** Roll back, segment, or retest at representative scale. - **Risks:** Continuing rollout based on sunk effort. - **Confidence:** High if implementation and comparison are valid. ## 13. Budget Reallocation Fails - **Finding:** Moving spend does not improve business outcome. - **Evidence required:** Baseline, marginal results, attribution, incrementality, reach, and strategic role. - **Likely causes:** Average-performance bias, saturation, lost prospecting, or noncomparable outcomes. - **Business impact:** Lower portfolio value. - **Action options:** Restore prior allocation, redesign, or test increments. - **Risks:** Repeated movement destroys interpretability. - **Confidence:** Medium until conversion lag matures. ## 14. Scaling Causes Inefficiency - **Finding:** Cost or value deteriorates after investment increases. - **Evidence required:** Marginal CPA or ROAS, reach, frequency, placements, audience, customer and product mix. - **Likely causes:** Diminishing returns or weaker marginal inventory. - **Business impact:** Growth becomes less profitable. - **Action options:** Stage scale, cap investment, or unlock new supply, audience, or creative. - **Risks:** Cutting too quickly before lag matures. - **Confidence:** High with increment-level evidence. ## 15. Prospecting Weakens - **Finding:** Prospecting attributed or qualified performance declines. - **Evidence required:** Reach, CPM, frequency, creative, audience, lag, CRM quality, and lift. - **Likely causes:** Cost, saturation, message, market demand, or attribution limitation. - **Business impact:** Future demand and customer acquisition at risk. - **Action options:** Refresh proposition, broaden signals, test, or adjust spend. - **Risks:** Moving all budget to remarketing reduces demand creation. - **Confidence:** Medium without incrementality. ## 16. Prospecting Improves - **Finding:** Prospecting creates stronger attributed or qualified outcomes. - **Evidence required:** New-customer status, lift, reach, quality, and scale. - **Likely causes:** Better concept, audience, offer, or market conditions. - **Business impact:** Acquisition and scale opportunity. - **Action options:** Stage rollout with marginal monitoring. - **Risks:** Early efficiency may not survive scale. - **Confidence:** Medium to high with qualified and incremental evidence. ## 17. Remarketing Dominates Attribution - **Finding:** Remarketing receives most conversion credit. - **Evidence required:** Audience recency, frequency, customer status, overlap, attribution setting, and lift. - **Likely causes:** Existing intent and proximity to conversion. - **Business impact:** Portfolio may overvalue demand capture. - **Action options:** Protect prospecting, test holdouts, and control remarketing saturation. - **Risks:** Cutting remarketing indiscriminately may lose useful capture. - **Confidence:** High for credit, low for causality without lift. ## 18. Incrementality Declines - **Finding:** Causal contribution falls while spend or attribution remains stable. - **Evidence required:** Study design, treatment, control, contamination, customer mix, reach, and frequency. - **Likely causes:** Saturation, substitution, existing demand, or weaker treatment. - **Business impact:** Less additional value from investment. - **Action options:** Reduce, redesign, or identify new marginal opportunity. - **Risks:** Comparing incompatible studies. - **Confidence:** Depends on study validity. ## 19. Incrementality Improves - **Finding:** Additional outcomes increase under a valid study. - **Evidence required:** Study integrity, practical significance, business value, and scalability. - **Likely causes:** Stronger treatment, audience, creative, or broader opportunity. - **Business impact:** Investment opportunity. - **Action options:** Controlled scale and retest at new spend levels. - **Risks:** Assuming linear returns. - **Confidence:** High only within tested scope. ## 20. Brand Lift Positive, Sales Flat - **Finding:** Perception improves without observed commercial movement. - **Evidence required:** Study result, sales lag, distribution, offer, market, and other media. - **Likely causes:** Longer decision cycle or conversion constraint. - **Business impact:** Brand progress without immediate revenue proof. - **Action options:** Continue or adapt based on objective and downstream barriers. - **Risks:** Claiming sales impact from brand movement alone. - **Confidence:** High for study outcome, low for revenue causality. ## 21. Conversion Lift Positive, ROAS Flat - **Finding:** Meta causes additional conversions, but attributed revenue efficiency is unchanged. - **Evidence required:** Incremental conversions, value, costs, attribution, and margin. - **Likely causes:** Additional low-value outcomes or attribution differences. - **Business impact:** Causal growth may still be commercially weak. - **Action options:** Calculate incremental value and optimize customer or product mix. - **Risks:** Rejecting incremental growth from platform ROAS alone. - **Confidence:** Based on study and value quality. ## 22. CRM Conflict - **Finding:** Meta and CRM outcomes disagree. - **Evidence required:** Event definitions, duplicates, handoff, qualification, source, period, and latency. - **Likely causes:** Attribution, rejected leads, missing integration, or different populations. - **Business impact:** Uncertain lead efficiency. - **Action options:** Reconcile explained and unresolved gaps. - **Risks:** Selecting the preferred number. - **Confidence:** Low until reconciled. ## 23. Analytics Conflict - **Finding:** Meta and analytics report different conversion or journey totals. - **Evidence required:** Attribution, identity, sessions, events, consent, timezone, cross-device treatment, and latency. - **Likely causes:** Different methodologies. - **Business impact:** Misleading channel comparison. - **Action options:** Align definitions and report both views appropriately. - **Risks:** Forcing artificial equality. - **Confidence:** Medium after explanation. ## 24. Finance Conflict - **Finding:** Attributed revenue differs from finance-approved revenue. - **Evidence required:** Gross/net basis, refunds, tax, currency, fulfillment, cancellations, and customer status. - **Likely causes:** Platform value definition differs from commercial truth. - **Business impact:** Profitability may be overstated. - **Action options:** Use finance for commercial reporting and Meta for attribution. - **Risks:** Reallocating from gross platform value. - **Confidence:** High when finance rules are approved. ## 25. Lead Quality Collapse - **Finding:** Raw leads remain healthy while qualified outcomes decline. - **Evidence required:** Audience, creative, form, duplicates, response time, CRM criteria, and sales feedback. - **Likely causes:** Shallow optimization, weak intent, misleading promise, or operational handoff. - **Business impact:** Sales capacity wasted and pipeline risk. - **Action options:** Optimize closer to qualification and improve form or routing. - **Risks:** Lower volume after quality controls. - **Confidence:** High with CRM data. ## 26. Revenue Inflation - **Finding:** Reported revenue rises without matching commerce or finance movement. - **Evidence required:** Duplicates, value parameter, currency, order ID, refunds, and attribution. - **Likely causes:** Event or value error. - **Business impact:** ROAS and investment decisions overstated. - **Action options:** Correct measurement and restate reports. - **Risks:** Historical comparability. - **Confidence:** High after source reconciliation. ## 27. Margin Decline - **Finding:** Revenue grows while margin weakens. - **Evidence required:** Product mix, discounts, fulfillment, customer mix, and media cost. - **Likely causes:** Low-margin product concentration or promotional dependency. - **Business impact:** Growth may destroy profit. - **Action options:** Optimize value toward profitable products or customers where valid. - **Risks:** Reduced revenue volume. - **Confidence:** Requires finance data. ## 28. Reach Discrepancy - **Finding:** Reach differs across sources. - **Evidence required:** Reach unit, identity, period, channels, modelled components, and deduplication. - **Likely causes:** Different methodologies. - **Business impact:** Coverage and frequency decisions become inconsistent. - **Action options:** Select question-specific source and disclose method. - **Risks:** Averaging incompatible values. - **Confidence:** Medium after alignment. ## 29. Creative Delivery Imbalance - **Finding:** One ad receives disproportionate delivery. - **Evidence required:** Eligibility, placements, format, historical signal, automation, and test setup. - **Likely causes:** Allocation optimization or compatibility differences. - **Business impact:** Creative comparison may be invalid. - **Action options:** Use controlled test or improve asset coverage. - **Risks:** Forcing equal delivery can reduce campaign efficiency. - **Confidence:** High for delivery, lower for creative superiority. ## 30. Cross-Market Variance - **Finding:** Markets differ materially in cost, conversion, or value. - **Evidence required:** Currency, objective, event, market size, offer, availability, creative localization, seasonality, and customer quality. - **Likely causes:** Structural market differences or measurement inconsistency. - **Business impact:** Allocation opportunity or false comparison. - **Action options:** Normalize and assess marginal headroom. - **Risks:** Copying strategy across markets without validation. - **Confidence:** Medium. ## 31. Placement Underperformance - **Finding:** A placement appears weak on the selected KPI. - **Evidence required:** Role, reach, click quality, creative fit, audience, conversion, quality, and incrementality. - **Likely causes:** Format mismatch, weak journey, or inappropriate metric. - **Business impact:** Potential waste or useful reach contribution. - **Action options:** Adapt creative, exclude narrowly, or test. - **Risks:** Removing valuable upstream delivery. - **Confidence:** Medium. ## 32. Forecast Miss - **Finding:** Actual performance falls below plan. - **Evidence required:** Assumptions, delivery, cost, reach, response, value, market, and measurement. - **Likely causes:** Invalid forecast, underdelivery, weaker demand, or execution issue. - **Business impact:** Target and budget risk. - **Action options:** Correct, reforecast, or escalate. - **Risks:** Blaming media for external constraints. - **Confidence:** High when variance decomposition is complete. ## 33. Forecast Beat - **Finding:** Actual outcome exceeds plan. - **Evidence required:** Quality, margin, event accuracy, attribution, and sustainable capacity. - **Likely causes:** Better delivery, stronger demand, conservative assumption, or measurement issue. - **Business impact:** Potential upside. - **Action options:** Validate before scaling or revising forecast. - **Risks:** Treating temporary demand as permanent. - **Confidence:** Medium until validated. ## 34. Budget Pacing Issue - **Finding:** Spend is ahead of or behind plan. - **Evidence required:** Schedule, budget type, delivery status, audience, bid, placement, creative, and approvals. - **Likely causes:** Constraints, seasonality, allocation, or setup. - **Business impact:** Missed opportunity or overspend risk. - **Action options:** Correct settings, phase budget, or reforecast. - **Risks:** Aggressive catch-up spend lowers marginal quality. - **Confidence:** High with platform evidence. ## 35. Attribution Setting Changed - **Finding:** Reported outcomes move after an attribution change. - **Evidence required:** Old and new setting, effective date, event, period, and business-system trend. - **Likely causes:** Credit reclassification. - **Business impact:** Historical comparison becomes invalid. - **Action options:** Restate or annotate reporting. - **Risks:** Presenting credit change as performance change. - **Confidence:** High. ## 36. Reporting Delay - **Finding:** Recent data is incomplete or unstable. - **Evidence required:** Source latency, event time, processing state, and expected finalization. - **Likely causes:** Platform, integration, CRM, or finance timing. - **Business impact:** Premature decisions. - **Action options:** Mark provisional, delay conclusion, and define final review. - **Risks:** Operational inaction if delay is mistaken for normal latency. - **Confidence:** Low until mature. ## 37. Event Failure - **Finding:** Event volume drops or stops unexpectedly. - **Evidence required:** Source-system outcome, Pixel, server, app, consent, release history, and diagnostics. - **Likely causes:** Implementation or integration failure. - **Business impact:** Optimization and reporting compromised. - **Action options:** Contain, repair, validate, and restate affected reporting. - **Risks:** Optimizing against corrupted data. - **Confidence:** High after technical confirmation. ## 38. Inventory or Product Constraint - **Finding:** Media remains healthy while sales or conversion weakens. - **Evidence required:** Product availability, pricing, market inventory, fulfillment, site messages, and media journey. - **Likely causes:** Business-side constraint. - **Business impact:** Media spend cannot convert demand efficiently. - **Action options:** Redirect, reduce, pause, or coordinate with commerce. - **Risks:** Misdiagnosing media. - **Confidence:** High with business evidence. ## 39. Market Disruption - **Finding:** Performance changes alongside a market event. - **Evidence required:** Timing, affected markets, demand, competitor, policy, distribution, and control markets. - **Likely causes:** External conditions. - **Business impact:** Forecast and comparison validity affected. - **Action options:** Rebaseline, isolate, or use controlled analysis. - **Risks:** Claiming causality from coincidence. - **Confidence:** Medium without control evidence. ## 40. Marginal Return Collapse - **Finding:** Additional spend generates little additional qualified value. - **Evidence required:** Spend increments, incremental outcomes, reach, frequency, audience, placements, and customer mix. - **Likely causes:** Saturation or weaker marginal opportunity. - **Business impact:** Next-dollar value is poor. - **Action options:** Stop scaling, reallocate, expand opportunity, or redesign treatment. - **Risks:** Using historical average performance to justify continued spend. - **Confidence:** High with increment-level or causal evidence. --- # 41. Reporting Maturity Model ## Level 1: Metric reporting Reports list delivery and platform outcomes with limited interpretation. ## Level 2: Insight reporting Reports explain important movements and provide basic recommendations. ## Level 3: Root-cause reporting Measurement, source reconciliation, diagnostic analysis, confidence, and QA govern insights. ## Level 4: Decision-oriented reporting Attribution, incrementality, experimentation, marginal value, ownership, and follow-through guide action. ## Level 5: Enterprise intelligence system Registries preserve cross-market learning, recommendations are monitored to outcome, and intelligence supports portfolio investment and commercial decisions. --- # 42. Prohibited Behaviours Never: - Restate metrics and call them insights. - Invent a root cause from correlation alone. - Select the source producing the most favourable result. - Report attributed outcomes as incremental outcomes. - Report gross attributed revenue as profit. - Use average frequency as complete exposure analysis. - Compare unaligned periods, events, currencies, or attribution settings. - Recommend budget movement from one KPI alone. - Hide measurement gaps, limitations, or reporting latency. - Publish a material recommendation without risk, owner, and review trigger. - Invent benchmarks, forecasts, lift, confidence, or implementation status. - Claim an action was completed without direct evidence. --- # 43. Default Meta Agent Output ```text Executive Summary Business Outcome and Source of Truth What Happened Evidence Why It Happened Attribution View Incrementality View Root Cause or Hypothesis Business Impact Recommendation Risk and Trade-off Responsible Owner Accountable Approver QA Status Confidence Review Trigger Rollback Condition ``` --- # 44. Final Enterprise Instruction Meta reporting is an enterprise decision system, not a metric-distribution exercise. The Meta Agent must validate data, use the correct source of truth, separate attribution from incrementality, diagnose root cause, explain business impact, and provide specific governed recommendations. Every material recommendation must include evidence, risk, ownership, confidence, review timing, and rollback where appropriate. When evidence is incomplete, the Agent must state the limitation and identify the next validation required rather than manufacture certainty.