---
name: marketing-kpi-review
description: Reviews a marketing KPI dataset and metric definitions for completeness, consistency, trend, target variance, segment differences, and decision usefulness. Use for marketing dashboards, KPI reviews, monthly reporting, funnel analysis, or metric-definition QA.
license: Apache-2.0
metadata:
  adlass.categories: "marketing-content/marketing-analytics, marketing-content/campaigns-email"
  adlass.industries: ""
  adlass.tags: "marketing-kpi, dashboard-review, metrics, funnel, variance, analytics"
  adlass.adaptation: "mapping"
  adlass.source: "original"
  adlass.version: "1"
---

# Marketing KPI review

## Purpose

Turn marketing metrics into a reliable decision brief by checking definitions, reconciling totals, identifying material movement, and separating evidence from interpretation.

## Scope

The skill covers metric QA, period comparison, funnel movement, segment patterns, target variance, data gaps, and commentary.

**Excluded:** changing source data, designing tracking, and claiming causes the data cannot establish.

## Data basis

- KPI table and metric definitions supplied as inputs.
- Targets, prior periods, campaigns and segment dimensions in scope when available.

## Result

A KPI findings sheet and management commentary brief.

## Quality criteria

- Every metric has definition, period, source and denominator.
- Totals and rates reconcile or discrepancies are visible.
- Commentary distinguishes observation, hypothesis and action.
- Missing data and attribution limits are explicit.

## Instructions

Use the supplied definitions even when a common alternative exists. Compare like periods and cohorts. Never infer causality from a trend alone. Prioritize movements that change a decision, and retain the underlying values beside any commentary.

## Adapt before use

- Map KPI, target, period, segment and source fields.
- Define formulas, comparison periods and materiality rules.
- Add campaign context and reporting audience.
- Keep source facts, analysis, recommendations, and unresolved questions visibly separate.
- Preserve stable document, section, page, sentence, URL, or row identifiers in every finding.
- Record assumptions about audience, channel, period, approval status, and data coverage.
- Treat missing guidance as an adaptation gap rather than inventing a company rule.
- Prefer a precise, bounded recommendation over a broad generic rewrite.
- Use supplied terminology consistently across the document, sheet, and any derivative copy.
- Do not infer performance, customer behavior, product capability, or market response from wording alone.
- Make every unresolved issue actionable by naming the missing decision, evidence, owner, or source.
- Keep factual claims traceable even when the final copy is concise.
- Recheck that counts, names, dates, and identifiers match across all outputs.
- Mark hypotheses as hypotheses and preserve competing explanations when evidence is mixed.
- Use the stated acceptance criteria as the final completeness check.
