Characteristics of a Measurement System

by Andrew D. ShamRao, Ph.D.

Metrics have become a very important aspect of organizational management. In an effort to know what is happening in their organization, leaders of companies mandate measure of all sorts. A general mandate for metrics usually results in a mass measurement effort that yields a lot of data and little useful information. The end result is a collection of haphazard measures instead of a measurement system.

In a measurement system for processes in which humans are involved, two general categories of metrics exist, results and behavioral measures. In most organizations, results are the primary focus. For example, typical HR measures include turnover, absenteeism, cost of benefits etc., all of which are result measures. Another important characteristic of a measurement system is that the result and behavioral measures are correlated and aligned with the mission of the organization. This prevents the collection of useless data.

A third characteristic is that the metrics are descriptive and functional, rather than just descriptive. For example, many organizations collect data on turnover at the level of counting loss of employees over a period of time. This information is descriptive of the problem. To make this data functional, periodic behavioral measures need to be collected also, such as those for assessing the organizational culture and the extent to which it promotes employee loyalty. By correlating the behavioral measure with the result measure, the metrics now constitute a system that yields data useful for decision-making and action.

While all measures provide some descriptive information, the critical determinant of the value of a metric is its usefulness. Where metrics related to human performance are concerned, a measure is most useful if it exists within a system of correlated result and behavioral measures. An assessment of a company’s existing metrics will reveal whether a measurement system exists, and if not, what result and behavioral measures are missing.

about the author