Metrics ​

Indicators or metrics allow measuring changes in the states of objects, evaluating the importance of a particular signal. Metrics describe quantitative changes that have occurred with a particular object of interest. Metrics are calculated when their measurement is simple and inexpensive and brings more value than the effort and resources required for their calculations.

Metrics can be both important and unimportant. On the one hand, they allow tracking signals about problems and reacting in a timely manner. For example, tracking a sales drop and analyzing the reasons behind it. On the other hand, agents often start focusing on numbers and forget about the real world behind them. For example, out of habit, they start "curing" sales drop by hiring new salespeople, when in reality, sales are falling due to significant market changes (competitors have introduced a new product to the market, successfully capturing company's market share). Moreover, how managers look at metrics and make decisions based on them often doesn't align with how engineers act to meet the metrics. For instance, managers set a goal to stay within budget during project implementation, assuming that engineers and engineering teams will learn to estimate the budget more accurately on their own. What happens in reality: teams start inflating the budget and overestimating the figures to avoid going over the budget. Or when it comes to creating plans: a department sets its own plans in a way that makes it convenient for the department to execute the plan, even though the primary focus should be the benefit for the company. These are just some examples of metrics that are tracked with good intentions but ultimately lead to the formation of destructive behaviors by employees^[1] without realizing it^[2].

To make metrics useful, it is important to first focus on qualitative explanations: what piece of the physical world stands behind the numbers, why the metric is being measured, whether it is suitable for evaluating action results and intervention. A well-chosen metric is one that both engineers and managers can explain "in plain language," without numbers, and describe what can affect the measurements and how the metric can be misused when making decisions.

Good metrics can be tailored to your situation and may not be universally accepted. For example, often in work planning, attempts are made to optimize "speed in process," i.e., the speed of completing individual tasks. But in reality, what needs improvement is "speed to outcome" -- the speed at which batches of goods / other value or benefit to the customer are delivered. It may seem that the difference is almost non-existent, but only until one starts to dig deeper. And then it turns out that "fast delivery of value" by the company is possible when the speed of working on constraints increases. Therefore, there is no point in investing a significant amount of money and effort in speeding up organizational links that are not constraints (except in cases where the amount of work is reduced for everyone). Another example: from the perspective of "speed in process," it is better to skim through a textbook or read it quickly. However, if your goal is not to "entertain" but to "learn," it is better to practice slow reading with notes. With this approach to learning, the student slows down, the speed of material reading decreases ("speed in process"). However, at the same time, the student more qualitatively assimilates the material and immediately applies it to life -- ultimately, they can gain mastery faster. That is, "speed in process" decreases, but "speed to outcome" increases.

Do not blindly rely on numbers. If you find that you cannot explain the physical meaning of introducing a metric, then you should understand what and why you are measuring.