amazon-account-growth-operating-system

Amazon Account Growth Operating System

Purpose

Act as the decision-making and orchestration layer for proactive Amazon account growth. Do not merely summarize metrics. Decide what should happen next, in what order, under which guardrails, and how the account should learn from the result.

Optimize for profitable sales growth, BSR and organic momentum, retail readiness, and wasted-spend reduction. Protect current revenue before scaling new growth.

Upstream Skills

Use this skill after, or alongside, these upstream analyses when available:

• amazon-ads-performance-drop-diagnosis: Use as the downside input. Pull in ASINs, campaigns, keywords, targets, search terms, BSR movements, retail-readiness blockers, root-cause hypotheses, severity, confidence, and action gates tied to performance decline.
• amazon-growth-opportunity-finder: Use as the upside input. Pull in ASINs, campaigns, keywords, search terms, product targets, budget-capped winners, BSR momentum, profitable scaling candidates, harvest candidates, priority scores, and confidence.

Do not treat upstream outputs as generic summaries. Treat them as separate evidence layers with their own gates. Preserve the downside skill's actionability gates and the upside skill's evidence thresholds, incrementality checks, retail-readiness gates, and confidence labels.

If one upstream analysis is missing, continue with the available evidence, lower confidence, and state what cannot be concluded. If raw account data is provided instead of upstream summaries, perform the equivalent downside and upside pass before building the operating plan.

Upstream Execution And Conflict Resolution

When the user asks for an account operating plan and the upstream skills are available, run or apply them before finalizing this skill's output:

1. Use amazon-ads-performance-drop-diagnosis first when there is any decline, risk, performance break, BSR deterioration, conversion drop, TACoS increase, efficiency decline, inventory issue, or important campaign loss. Carry forward its data reliability gate, exact windows, break timeline, root-cause confidence, biggest losers, action gates, and verification plan.
2. Use amazon-growth-opportunity-finder next for upside. Carry forward its source map, report grains, join-key caveats, commercial-impact scoring, evidence thresholds, incrementality checks, retail-readiness gates, and action rows.
3. If this skill is used alone with raw data, internally perform both passes: a downside pass using the Performance Drop Diagnosis logic and an upside pass using the Growth Opportunity Finder logic. State that the upstream skills were not separately run if that is true.
4. If the upstream skills disagree, resolve the conflict in the operating plan rather than averaging them:
   • High-confidence Protect or action-gated recovery findings override scale recommendations until the risk is fixed or explicitly accepted.
   • A high-upside opportunity with a retail-readiness, margin, inventory, or Featured Offer / Buy Box blocker becomes Fix Before Scaling.
   • A profitable campaign tied mostly to branded or defensive traffic becomes Controlled Scale, Defensive Advertising, or Investigate unless total-sales, TACoS, and incrementality evidence support expansion.
   • A low-confidence downside finding should not block high-confidence, reversible growth, but it should add monitoring and approval requirements.
   • A high-confidence waste finding can fund growth only when the spend is isolated and not protecting rank, launch velocity, brand defense, or strategic market share.
5. Transfer upstream action gates directly:
   • Do not recommend bid, budget, negative, pause, relaunch, or structural execution if the Performance Drop Diagnosis action gate says the evidence is not action-safe.
   • Do not recommend exact harvesting, negatives, bid-downs, budget increases, placement changes, or rank-growth spend unless the Growth Opportunity Finder thresholds and readiness gates are met.
   • When gates are not met, use Investigate, Monitor Only, controlled test, or retail-readiness fix actions.