Intent

Describe a measurement, its context, and its source, and the instrumentation needed to collect such measurements.

Motivation

Business intelligence supports business decisions. Business intelligence offers governors and managers the information they need to guide and move a business in the directions established by its business policies. Business intelligence measures and quantifies those qualities of concern to a business, especially relative to business products, activities, and value(s).

Applicability

Use quantity description when

• you discover a business quality concern with its attendant metric(s), and thus
• you need to describe the measurement(s) that relate to the concern.

Considerations – Basic and Derived Measurements

Having discovered an important business quality concern, some questions naturally arise:

• How do we measure quality (and value) in a meaningful way?
• How do we collect and record the measurements?
• How are the measurements evaluated, summarized, and reported?

Some measurements are derived from other, more basic measurements. Where such derivation occurs, we want to know the formula used and the basic measures that produce the (usually more complex) derived measure. Consider the following from a sample quantity card (see linked card).

hazard type storage capacity = ( drum storage limit – drum count ) per hazard type

A building’s storage capacity is derived from three other terms: a storage limit, a drum count, and a hazard type. Each of these in turn has a source. Thus, summary measures (like storage capacity) are derived from other measures.

Considerations – Instrumentation

Some basic measurements come from a source instrument: some (sometimes physical) means of sensing, collecting, recording, and reporting a measurement. However, the source of a measurement may sometimes be a person, or a source of governance, such as a regulation, a license, etc.

Instrument design and deployment are beyond the scope of this consideration. However, here are some basic observations:

• Measurement always has some attendant cost(s) associated with it.
• Human involvement usually drives up measurement costs significantly.
• So, it’s usually best if measurement can be automated.
• In either case, the cost of measurement needs to be addressed.

In addition:

• Instruments need to produce appropriate values (with relevant metrics).
• Measurement invariably includes some frequency and scale of error (noise).
• Instruments usually need calibration (and human oversight).
• Calibration needs to happen with some regularity (frequency).
• Instrument (sensor) replacement may be required (when a failure occurs).

Consequences

It’s also best to establish a baseline and some goal(s) with respect to further improvement. Sometimes, it’s useful to establish a limit, i.e. the point of diminishing returns. This is usually achieved when the cost of further improvement nears or exceeds the cost of measurement. But, the cost of measurement may be a “cost of doing business”. Measurement may be a hard requirement, as when imposed by an external regulatory agency.

If measurements are a relatively new area for you, it’s good to get a proper introduction to measurement theory. Scott Whitmire included an extensive introduction to measurement theory in his book Object-Oriented Design Measurement. While this book focuses primarily on applying measurement theory to object-oriented designs, the ideas about measurement can be applied to a much wider range of practice areas.