| name | measurement-plan |
| description | Design a measurement plan that links business or product questions to metrics, events, dimensions, data quality, and decision thresholds. |
Measurement Plan
Use before implementing analytics or evaluating a product/marketing initiative.
Procedure
- Start from the decisions the data must support, not from a list of available metrics.
- Define outcomes, leading indicators, guardrails, and relevant segments.
- Specify each metric's exact numerator, denominator, time window, unit, and source.
- Define events/properties or source fields needed to compute the metrics consistently.
- Identify data-quality risks: missing events, identity stitching, bot traffic, duplicate events, attribution ambiguity, and changing definitions.
- Define baselines and decision thresholds where appropriate.
- Document what the measurement cannot prove, especially causality.
Quality gate
Two analysts should be able to compute the same metric from the specification and understand what decision it informs.
