DataTerminology.comNEW
Also known as: Performance Indicators
Quantifiable measures used to track performance, outcomes, or system behavior consistently over time.
DESCRIPTION
Metrics translate business goals and system behavior into measurable, reproducible numbers. In analytics, they are defined over modeled data with clear grain, filters, and aggregation logic. In observability, metrics expose time-series measures for SLI/SLO tracking and alerting. Good metrics are unambiguous, governed, and versioned to avoid ‘multiple versions of the truth.’ Metric stores and semantic layers centralize definitions for BI and AI products. Clear ownership and documentation ensure trust and reuse across dashboards and experiments.
ALTERNATIVE DEFINITIONS
Wikipedia
A performance indicator or key performance indicator (KPI) is a type of performance measurement.
Highlights KPIs as a class of performance measurements.
Google Analytics Help
Metrics are quantitative measurements.
Concise definition emphasizing quantitative nature in analytics.
RELATED TERMS
Key Performance Indicator Semantic Layer Objectives and Key Results Measure Definition Analytics Engineering Observability Service Level Indicator Model Data Dashboard
COMPETING TERMS
RELATIONS
- Narrower: Business Metrics Data Change Activity Metrics Data Ingestion and Consumption Metrics Data Steward Coverage Metrics Evaluation Metrics Fairness Metrics Operational Metrics Performance Metrics Strategic Metrics
KNOWLEDGE AREAS
Data Science Data Warehousing & BI Data Products Observability & Telemetry
TARGET AUDIENCE
Data Analyst BI Specialist Analytics Engineer Data Scientist AI Product Manager Chief Data Officer
EXTERNAL SOURCES
- Wikipedia - Performance indicator
- OpenTelemetry - Metrics signal concepts and types
- dbt Labs - Metric definitions in a semantic layer
- Google Cloud - Cloud Monitoring metrics model
TOP QUESTIONS
What makes Metrics trustworthy across reports?
A single semantic definition with governed calculation, grain, and filters ensures consistency. Versioning and clear lineage prevent drift and misinterpretation.
Category: Requirements
How do Metrics differ from Dimensions?
Metrics are aggregated quantitative measures, while dimensions are categorical attributes used to slice and group those measures.
Category: Misunderstanding
Why are Metrics important in A/B testing?
They operationalize success criteria and enable statistical evaluation. Well-defined primary and guardrail metrics reduce p-hacking and misreads.
Category: Relevancy
What are common mistakes defining Metrics?
Ambiguous definitions, mixing grains, and hidden filters cause inconsistent results. Over-aggregating can mask important variation.
Category: Mistakes
How should Metrics be documented?
Document purpose, formula, grain, filters, owners, and validation tests. Include lineage from sources to downstream dashboards.
Category: Management
What is currently a hot topic around Metrics?
Universal semantic layers and metric stores that serve BI and AI workloads are gaining traction to eliminate metric drift.
Category: Topic
How are Metrics monitored in production systems?
Observability pipelines collect, aggregate, and alert on time-series metrics with SLIs/SLOs, using systems like Prometheus and OpenTelemetry.
Category: Management
What is the risk of poorly governed Metrics?
Conflicting numbers erode trust, drive wrong decisions, and waste time reconciling dashboards.
Category: Risks
Where to start learning about Metrics modeling?
Study dimensional modeling, dbt metric docs, and semantic layer guides. Practice by defining a small, governed KPI set.
Category: Requirements