---
name: kpi-dashboard-spec-builder
description: Design the specification for a KPI dashboard: the right metrics, definitions, and layout for a given role. Use this skill whenever a user wants to design a dashboard, choose KPIs, define metrics, spec a reporting view, or says 'what KPIs should we track', 'design our dashboard', or 'we track too many metrics and none of them help'. Trigger whenever a role or business needs the right small set of metrics defined and laid out to drive decisions.
---

# KPI Dashboard Spec Builder

## What this does and why it matters
Most dashboards track everything measurable and drive no decisions, because a metric that does not change behavior is just decoration. This skill specifies a focused dashboard: the few metrics that actually matter for a role, precisely defined, laid out so the viewer sees status and can act. It produces the spec a builder or BI tool implements, not the chart itself.

## Inputs to gather
1. The audience and the decisions the dashboard should support.
2. The business or function it covers.
3. The data sources available.
4. The cadence (real-time, daily, weekly, monthly).

## Method

### 1. Start from decisions, not data
Ask what decisions the viewer makes and work backward to the metrics that inform them. A metric earns its place only if it would change an action. This inverts the usual "what can we measure" mistake.

### 2. Choose a focused set
A handful of primary KPIs, not thirty. Distinguish the few headline metrics from supporting diagnostics that explain them. Too many metrics means none get attention.

### 3. Define every metric precisely
For each: the exact definition, the formula, the source, the time window, and the target or benchmark. Ambiguous definitions ("active users" meaning three different things) are how dashboards lose trust. This is the most valuable part of the spec.

### 4. Pair metrics with context
Every number needs a comparison (target, prior period, trend) to be meaningful. Specify what each metric is shown against.

### 5. Lay it out by priority
Headline metrics first and largest, diagnostics below. Group logically so the viewer's eye lands on what matters.

### 6. Define leading vs lagging
Note which metrics are leading indicators (predict the future) versus lagging (report the past), since a good dashboard has both.

## Output format
ALWAYS use:

# Dashboard Spec: [Audience / Purpose]
## Decisions this dashboard supports
## Primary KPIs (metric | definition | formula | source | target | leading/lagging)
## Supporting diagnostics (same fields)
## Layout (priority order and grouping)
## Cadence and refresh
## Definitions glossary (exact meaning of each term)

## Anti-patterns to avoid
- Tracking everything measurable instead of what drives decisions.
- Vague metric definitions.
- Numbers with no comparison or target.
- All lagging metrics and no leading indicators.

## Example
A CEO dashboard spec starts from four recurring decisions, defines six primary KPIs with exact formulas and sources, pairs each with a target and trend, and lays headline revenue and cash above supporting funnel diagnostics.