2020 Issue

38 Family of Measures: A Method to Examine and Improve Metric Systems Dr. David Wetzel — INCOSE Figure 1 Periodic Table of Business Metrics A common image of metrics on posters and the internet is of a triangle with “Good,” “Fast,” and “Cheap” on each side and the saying “Pick Two” in the middle. The implication is that if you pursue any two of them, then the third one suffers. For example, if you want good and fast, then it costs more. The biggest issue with this triangle-view of metrics is that “good” usually means yields, defects per million, or failure rates, which are all productivity rate metrics and not quality metrics. We need a more robust model to untangle quality metrics from productivity, timeliness, and financial metrics. We call this model the “Family of Measures.” We have used this model over the past three decades to help organizations evaluate existing metric systems; differentiate order winners from market entry metrics; evaluate trends, articles, books, and claims; illustrate the Deming chain reaction and Taguchi loss function; explain why reliability metrics are quality met- rics; create truly balanced scorecards; and even illustrate history. The sincere hope is that the Family of Measures re- veals fresh insights into metric development and gives you an additional tool to make better data-based decisions. Four Distinct Metrics Organizational performance metrics are like the periodic table in chemistry (Figure 1). There are base elements, or metrics: productivity, finan - cial, quality, and timeliness. These building blocks can be used as standalone metrics or combined to form com- pounds (i.e., rates and percentages). There are two easy ways to identify and separate the four metrics. Each metric answers a different question and has a different unit-of- measure. For example, if the unit-of-measure is currency (e.g., dollars, yen, euro), then the metric type is financial. If the unit-of-measure is an index (e.g., customer satisfaction, process capability index), then the metric type is quality. If the unit-of-measure is time (e.g., seconds, days, years), then the metric type is timeliness. If the unit-of-measure is units, (e.g., counts, °F, grams), then the metric type is productiv- ity. The second way to differentiate the four metrics is to understand what question is being asked. Financial met- rics ask, “How much?” Productivity measures ask, “How many?” Timeliness metrics ask, “How fast?” Quality metrics ask, “How well?” All four base metrics can be either counts (attributes) or measures (variable). Table 1 is a summary of the two ways to differentiate the four metrics, including examples and definitions. Of the four metrics, financial and timeliness metrics are the most intuitive. We have been counting money and telling time since grade school. It is the other two, productivity and quality, that give organizations trouble. Two interre- lated problems exist. First, there is the common practice of defining quality with productivity, financial, or timeliness metrics. These substitutions and misconceptions hinder the creation of true quality metrics. Especially, they hinder the development of leading indicators of process quality that would improve variability, stability, and capability.