cyberneticlibrary

Spec dashboards for BI teams

dashboard-briefskillsetup L2327
mohitagw15856/pm-claude-skills
What it does

Specify dashboards for business questions with implementation-ready blueprint

Best for

Spec'd dashboards reduce back-and-forth with BI team and prevent dashboards that answer nobody's question; one document replaces three meetings.

Inputs
  • · Business question dashboard should answer
  • · Primary audience and refresh cadence
  • · Available data sources (Postgres, BigQuery, Mixpanel, Salesforce, etc.)
Outputs
  • · Dashboard Brief (MD or PDF) with: KPI cards, charts with specifications, filters/controls, layout diagram, data requirements, access & ownership
  • · Implementation-ready spec for BI engineer (no follow-up meeting needed)
Requires
  • · BI tool (Looker, Metabase, Tableau, Power BI, Grafana, or custom)
Preconditions
  • · Business question clearly stated (not vague like 'show us the data')
  • · Primary audience identified (exec vs. product team vs. ops)
  • · Data sources available and documented
Failure modes
  • · Dashboard with 20+ KPI cards (information overload, no narrative)
  • · Charts without 'key insight' — shows data but doesn't surface signal
  • · Missing derived fields or joins — BI engineer has to build data infrastructure
  • · Filters too many (dashboard becomes paralyzed by options)
Trust signals
  • · Business question framed as narrative (not 'show me metrics')
  • · KPI cards limited to 3-6 (avoids noise)
  • · Every chart has a 'key insight to surface' (data with purpose, not data dump)
  • · Layout described in plain English (row structure, full-width vs. side-by-side)
  • · Data requirements section flags missing fields or complex joins