The mandate
Give every important metric one definition and one decision to inform.
Defines decision metrics, event contracts, dashboards, data quality checks, and review cadence.
The Analytics Agent defines a small decision-oriented metric system, event taxonomy, lineage, privacy constraints, validation, freshness, dashboards, and review cadence. Missing data remains different from zero.
What to bring
Bring the decisions, source systems, tracking gaps, and collection boundaries.
The plan starts with what people need to decide, then works backward to the smallest data contract that can support it.
- 01
The product, growth, customer, financial, or operating decisions measurement must support.
- 02
Systems of record, event sources, owners, identifiers, and known joins.
- 03
Existing metrics, events, dashboards, known gaps, disagreements, and data-quality evidence.
- 04
Purpose limitation, privacy, retention, access, residency, and collection boundaries.
- 05
The decision, metric, event, dashboard, or quality outcome the plan must enable.
What you receive
A measurement system whose definitions survive dashboards and handoffs.
Metrics, event contracts, and dashboard specifications stay connected to sources, owners, privacy, validation, freshness, and a recurring evidence review.
- 01
Company measurement plan
analytics/measurement-plan.mdReady when
Metrics map to decisions, owners, sources, and review cadence.
- 02
Product and marketing event taxonomy
analytics/events.jsonReady when
Names, properties, privacy rules, and validation are explicit.
- 03
Operating dashboard specification
analytics/dashboard.mdReady when
Definitions, segments, freshness, and alert thresholds are testable.
The method
Name the decision before naming the event.
The agent minimizes collection by defining the decision and canonical metric first, then specifying validation and review.
- 1
Name decisions
Identify owners, choices, review cadence, and the smallest metric set that can change action.
- 2
Define contracts
Specify metrics, events, properties, identities, lineage, privacy, and retention.
- 3
Validate
Create release checks, freshness expectations, quality tests, and alert thresholds.
- 4
Review
Design dashboards and evidence reviews that preserve definitions and expose missing data.
Signals & guardrails
More collection is not a substitute for a clear decision.
Inconsistent definitions, unowned metrics, missing lineage, excess personal data, or events shipped without validation stop the measurement plan.
Signs the work is useful
- Critical metric freshness
- Tracking validation pass rate
- Decisions with cited evidence
Reasons to pause
- Metrics have inconsistent definitions
- Collection exceeds privacy or retention needs
Approval boundary
This entrypoint declares no open-world effect. Applying its todos still requires an authenticated, scoped project run.
Technical appendix
The exact contract behind the profile.
Useful for operators who need to inspect the immutable Kit version, typed boundary, and verification surface.
- Kit
- company-suite@0.1.0
- Entrypoint
- plan-analytics
- Function
- plan-analytics
- Runtime
- python@3.12
- Input
- specialist-brief.schema.json
- Output
- specialist-plan.schema.json
Verification
specialist-plan-quality · specialist-plan-schema
Connected systems
Analytics