ENDPOINTS, CRITERIA, PARAMETERS, AND MEASUREMENTS (ECPM)
“You cannot manage what you cannot measure!” All management gurus must repeat this incantation thrice a course. Often, it remains the only thing their disciples remember after the enlightenment seance. Yet, the advice is meaningless and even harmful if not part of a planning-controlling system.
Such a system must connect measurements (metrics), parameters, and criteria. Criteria, in turn, will be defined by the desired endpoints. In this order: endpoints-criteria-parameters-measurements (ECPMs).
If you do not understand what your ECPMs are, what they do, and how they relate to one another and your work, any measurements you do will naturally degenerate into meaningless metrics. {MEANINGLESS METRICS}
So, what are these elements? Let’s start with a criterion. A criterion is not a parameter, yet a criterion only becomes meaningful in relation to some parameter. Also, a criterion is not a metric or measurement, although a particular value of a metric or measurement will be a criterion.
Its meaning comes straight from its etymology—Ancient Greek “κριτήριον” is “means for judging, standard.” We use criteria to judge—success, failure, slow, fast, important, not important, etc.—and to decide on the action—start, stop, buy, sell, join, leave, etc. Criteria help us decide and judge, and as such, they can and often will be subjective. {SUBJECTIVITY}
Thus, to protect yourself from misunderstanding, always define criteria at the start of any endeavor: what you and others (if it is a group effort) will consider success, what will trigger a stop, etc. Next, let’s define measurement.
Although they have different meanings, the terms metric and measure overlap. We use “measure” for more concrete or objective parameters and “metric” for more abstract, higher-level, or somewhat subjective parameters. Yet, we will combine them here to mean gauging the parameter for decision-making or judgment and use the term “measurement” in Stan Stevens’s way. {His paper} We will not use “metric,” only referring to the meaningless ones.
A measurement can only imply judgment when a relevant criterion exists. For instance, a 71km/h measurement cannot inform decisions or judgments. Is it fast or slow, allowed or prohibited? We can only tell if we have a criterion.
If the criterion (e.g., speed limit) is a 50km/h maximum, the vehicle moves too fast. And if another criterion exists—exceeding the speed limit by more than 20 km/h is a traffic crime—then driving at this speed will be too fast and can be considered a traffic crime. Yet, if the Pendolino train travels at 71km/h, we think it is slow, so we want to “pendal it on” because we are using a different criterion of what is fast.
And suppose a new acquaintance tells us he develops this speed while jogging. In that case, we consider him naïve, ignorant, crazy, or a liar, depending on the context and the additional data it provides. Yet, the measurement can stay the same—71km/h.
It is worth noting that any measurement must be agreed upon beforehand and so must be units of measurement.
If I say, “The car moves at 1.18 takhus,” will it sound meaningful? No. Although 1.18 is how many kilometers per minute a vehicle moving at 71km/h covers. However, we rarely use km per minute, and although “takhus” is “quick” in Ancient Greek, there is no such a unit of measurement. We discussed the case of the Mars Climate Orbiter and discussed what happens when we have not agreed on units of measurement.
What is a parameter—“P?” It is a characteristic of a system, some characteristic. In the example above, it is speed. Different systems will have different sets of parameters. Vehicles that move across the land will have speed expressed in units of distance covered during an hour; systems that process information will have speed in units processed per second. And on it goes.
Understanding and measuring our systems’ essential parameters and developing criteria is important.
To make all these criteria, parameters, and measurements meaningful, we must have an endpoint—what do we want to achieve with our actions? In the case of speed limits, we want fewer road traffic accidents.
The endpoints of a Ph.D. project may differ, and consequently, so will other members of the ECPM family. Suppose all that you want is the title to move on to work in the industry. In that case, speed will be the most critical parameter, so the criteria for the journals you will aim to publish in and the quality of data and methods will be less strict than if you want to make a career in research.
We can summarize that any work needs the ECPM, and only through the system can you track your progress, estimate its direction, and decide what to change or adjust. When you have none, the problems are closer than you think. They are already in your closet with knives pulled out when you use meaningless metrics.
{MAP YOUR DREAM, DAILY INCREMENTS, ENDPOINT TIERS, SMALL WINS, THE 5 EXCHANGES, SAMPLING RATE, TRACK PERFORMANCE}