{"skill":{"slug":"campaign-retrospective-analyst","displayName":"Ads Campaign Review","summary":"Run retrospective analysis for campaigns on Meta (Facebook/Instagram), Google Ads, TikTok Ads, YouTube Ads, Amazon Ads, and DSP/programmatic channels.","description":"---\nname: campaign-retrospective-analyst\ndescription: Run retrospective analysis for campaigns on Meta (Facebook/Instagram), Google Ads, TikTok Ads, YouTube Ads, Amazon Ads, and DSP/programmatic channels.\n---\n\n# Ads Campaign Review\n\n## Purpose\nCore mission:\n- root-cause analysis, lesson extraction, next-cycle design\n\nThis skill is specialized for advertising workflows and should output actionable plans rather than generic advice.\n\n## When To Trigger\nUse this skill when the user asks for:\n- ad execution guidance tied to business outcomes\n- growth decisions involving revenue, roas, cpa, or budget efficiency\n- platform-level actions for: Meta (Facebook/Instagram), Google Ads, TikTok Ads, YouTube Ads, Amazon Ads, DSP/programmatic\n- this specific capability: root-cause analysis, lesson extraction, next-cycle design\n\nHigh-signal keywords:\n- ads, advertising, campaign, growth, revenue, profit\n- roas, cpa, roi, budget, bidding, traffic, conversion, funnel\n- meta, googleads, tiktokads, youtubeads, amazonads, shopifyads, dsp\n\n## Input Contract\nRequired:\n- question_or_report_goal\n- metric_scope: KPI, dimensions, and date range\n- data_source_scope\n\nOptional:\n- attribution_window\n- benchmark_reference\n- dashboard_filters\n- confidence_threshold\n\n## Output Contract\n1. Metric Definition Clarification\n2. Query Plan\n3. Result Summary\n4. Interpretation and Caveats\n5. Decision Recommendation\n\n## Workflow\n1. Disambiguate metric definitions and time window.\n2. Build query slices by platform, funnel, and audience.\n3. Compute trend deltas and variance drivers.\n4. Summarize findings with confidence level.\n5. Propose concrete next actions.\n\n## Decision Rules\n- If metric definitions conflict, lock one canonical definition before analysis.\n- If sample size is small, mark result as directional not conclusive.\n- If attribution changes materially alter result, show both views.\n\n## Platform Notes\nPrimary scope:\n- Meta (Facebook/Instagram), Google Ads, TikTok Ads, YouTube Ads, Amazon Ads, DSP/programmatic\n\nPlatform behavior guidance:\n- Keep recommendations channel-aware; do not collapse all channels into one generic plan.\n- For Meta and TikTok Ads, prioritize creative testing cadence.\n- For Google Ads and Amazon Ads, prioritize demand-capture and query/listing intent.\n- For DSP/programmatic, prioritize audience control and frequency governance.\n\n## Constraints And Guardrails\n- Never fabricate metrics or policy outcomes.\n- Separate observed facts from assumptions.\n- Use measurable language for each proposed action.\n- Include at least one rollback or stop-loss condition when spend risk exists.\n\n## Failure Handling And Escalation\n- If critical inputs are missing, ask for only the minimum required fields.\n- If platform constraints conflict, show trade-offs and a safe default.\n- If confidence is low, mark it explicitly and provide a validation checklist.\n- If high-risk issues appear (policy, billing, tracking breakage), escalate with a structured handoff payload.\n\n## Code Examples\n### Query Spec Example\n\n    metric: roas\n    dimensions: [platform, campaign]\n    date_range: last_30d\n\n### Result Schema\n\n    {\n      \"platform\": \"Meta\",\n      \"spend\": 12000,\n      \"revenue\": 42000,\n      \"roas\": 3.5\n    }\n\n## Examples\n### Example 1: Daily report automation\nInput:\n- Need 9AM daily summary for key campaigns\n- KPI: spend, cpa, roas\n\nOutput focus:\n- report schema\n- anomaly highlights\n- top next actions\n\n### Example 2: Attribution window comparison\nInput:\n- 1d click vs 7d click disagreement\n- Decision needed for budget shift\n\nOutput focus:\n- side-by-side metric table\n- interpretation caveats\n- decision recommendation\n\n### Example 3: Traffic structure diagnosis\nInput:\n- Revenue flat but traffic rising\n- Suspected quality decline\n\nOutput focus:\n- source mix decomposition\n- quality signal changes\n- corrective action plan\n\n## Quality Checklist\n- [ ] Required sections are complete and non-empty\n- [ ] Trigger keywords include at least 3 registry terms\n- [ ] Input and output contracts are operationally testable\n- [ ] Workflow and decision rules are capability-specific\n- [ ] Platform references are explicit and concrete\n- [ ] At least 3 practical examples are included\n","tags":{"latest":"1.0.0"},"stats":{"comments":0,"downloads":573,"installsAllTime":0,"installsCurrent":0,"stars":1,"versions":1},"createdAt":1772503995076,"updatedAt":1778491695332},"latestVersion":{"version":"1.0.0","createdAt":1772503995076,"changelog":"Initial public release of campaign-retrospective-analyst\n\n- Supports retrospective analysis for ad campaigns on Meta, Google Ads, TikTok Ads, YouTube Ads, Amazon Ads, and DSP/programmatic platforms.\n- Specialized to deliver actionable insights: root-cause analysis, lesson extraction, and next-cycle design.\n- Provides a structured workflow with clear input/output contracts, decision rules, and confidence handling.\n- Includes platform-specific guidance and recommendations tailored to channel and campaign objectives.\n- Offers practical examples, code/query formats, and an operational quality checklist for reliability.","license":null},"metadata":null,"owner":{"handle":"danyangliu-sandwichlab","userId":"s177ps4rekejas69j3g9mecnzx885pdp","displayName":"danyangliu","image":"https://avatars.githubusercontent.com/u/184733968?v=4"},"moderation":null}