amazon-keyword-traffic-analysis

ZooData — Amazon Keyword Intelligence

Respond in the user's language.

Start here

1. Classify the request: seed-keyword expansion, target-keyword analysis, product traffic analysis, or a single lookup.
   • Route keyword-centered questions about demand, market/SERP structure, trend, value, relevance, or targeting fit to target-keyword analysis. An ASIN may be supporting evidence without changing the keyword-centered subject.
   • Route ASIN-centered questions about traffic health, current traffic terms or sources, channel/term structure, changes, trends, anomalies, or causes to product traffic analysis.
   • Route a broad ASIN traffic analysis, overview, or health check directly to the product traffic health overview. Do not ask the user to choose between structure and change first.
   • For an ASIN × keyword request, route value/fit/relevance questions without movement or causal intent to target-keyword analysis; route visibility, placement, exposure, movement, anomaly, or causal questions to product traffic analysis.
   • If product traffic analysis identifies a term and the follow-up asks about its value, start target-keyword analysis and reuse compatible ASIN traffic evidence. If keyword analysis identifies a product-side movement question, start product traffic analysis and reuse compatible keyword evidence.
2. Read the local references/cli-contract.md, references/reference.md, and the relevant zoodata.py --help before selecting a tool. The shared contract owns CLI invocation and result handling; reference.md is the sole source for production endpoint availability, parameters, response fields, dates, batching, credits, and API capability boundaries.
3. Load references/output-rules.md for user-facing rendering and apply execution-guide.md § Final Output Gate on every rendering path. For a single lookup, also load only references/execution-guide.md sections Authority and routing, Execution mode, Structured Field Identity Gate, Interface Failure Stop Gate, Final Output Gate, HTTP Validation Rule, and Credential and Credit Failures; use output-rules.md § Quick Mode Output and do not load a scenario unless the follow-up broadens the request.
4. For every full-mode request, load the complete references/execution-guide.md, references/evidence-protocols.md, and the applicable scenario guide below. The guide is the sole scenario/stage and Gate contract; evidence protocols operate only inside its active stage. After every retrieval or tool result, apply its Interface Failure Stop Gate before selecting any next capability or command. Route to one applicable scenario, or multiple non-exclusive scenarios only when the guide permits combination:
   • references/scenarios-expand.md
   • references/scenarios-keyword-analysis.md
   • references/scenarios-product-traffic-analysis.md
5. For a causal, anomaly, or action question, additionally load references/diagnosis-action-protocols.md. Do not load it for a non-diagnostic stage merely because diagnosis is available.
6. After API retrieval, load only the field-semantic reference needed for the returned data:
   • references/metrics-market-profile.md for market-profile
   • references/metrics-trend-profile.md for trend-profile
   • references/serp-and-rollover.md for SERP or organicRolloverRate
   • references/traffic-observation-semantics.md for traffic-term lists, traffic timelines, or traffic-profile data
7. Before requesting or interpreting a seller artifact, load references/sqp-field-semantics.md. Treat it as the sole acquisition and field-semantics source for user-provided ABA-SQP or Amazon Ads data.