How to Build an Effective Data Metric System: Classification and Naming Rules

When building a metric system, the first step is to define metric categories and naming constraints so that each metric is self‑descriptive and communication cost is reduced.

Metrics are divided into atomic metrics and derived metrics. An atomic metric is a measurement tied to a single business event and cannot be further split; it combines a business process with a measurement, e.g., payment amount.

Atomic metric = business process + measurement

Derived metric = time period + modifier(s) + atomic metric

The article explains related concepts:

Business domain: a higher‑level business partition used for very large systems.

Business process: an indivisible business event such as order, payment, refund.

Modifier type and modifier: abstract classifications belonging to a domain, e.g., access terminal type (PC, wireless) under the log domain.

Time period: the statistical window, e.g., last 30 days, natural week, up to today.

Dimension: the environment of a measurement, forming a set of attributes like geography or time, belonging to a data domain.

Derived metrics are formed by taking an atomic metric and optionally adding one or more modifiers and a time period. For example, “payment amount of overseas buyers in the last day” is a derived metric where the atomic metric is payment amount, the modifier is “overseas buyer”, and the time period is “last day”.

The methodology is based on a PPT from DAMA China, which provides a practical, ground‑level guide to constructing a data metric system.