Managing Variation: Why Entropy Matters

No management effort can ignore the existence of entropy and variation. What do we need to do about them?

What is variation?

It is beyond the scope of any blog to teach the fundamentals of Statistical Process Control; Don Wheeler and other experts have already done that majestically. The goal here is to examine the fundamental role of managing variation in any effort to manage an organization systemically, in particular using the Decalogue methodology.

If we examined the arrival times of employees to their work place every day, we would immediately notice that nobody ever manages to get to the office at exactly the same time. No matter how optimized their routine is, the arrival time is always different. The reason that prevents everybody from getting to work every day at exactly the same time is variation.

Why can’t any process be standardized so that no variation occurs? It is because in nature there is a ‘variable’ called entropy that accounts for the variation associated with every process. The 2^nd law of thermodynamics states that any “spontaneous” change in a “closed” system is accompanied by an overall increase in entropy. When water evaporates molecules are dispersed and tend to occupy the whole space, resulting in an increase of entropy. The entropy of the universe, for example, is always increasing.

Entropy is a measure of disorder, or randomness (variation) in a system. Any organization, or system, in its spontaneous evolution, is naturally affected by the increase of entropy. The day-by-day repetition of simple actions at our work place will never be the same because of the natural increase of entropy.

Variation and managers

Variation affects all aspects of our life, and all processes in an organization. It is of profound importance to managers who, in order to exert their role, must ensure a stable and predictable environment. Indeed, the essence of management is prediction. Let’s have a look at few key points we have to consider when we talk about variation.

Walter Shewhart was the first to have an intuition about this phenomenon, which is intrinsic to every process and system. We can define a process as a set of actions/activities that happen over time, following a rationale/procedure and aimed at a specified goal. Shopping for food at the market with our family on a Saturday is a process and so is the set of actions that get us to the office every day.

Let’s examine the set of actions that everybody goes through to get to work, and the relevant variations:

• How many minutes do people stay in bed after switching off the alarm? Perhaps they stayed up late the night before, so they want to stay in bed a few more minutes;
• Do they make the coffee or some other family member makes it?
• Is the bathroom free or do they have to wait to get in?
• How long do they stay in the shower and how long will they take to get dressed? Did they decide the night before what they were going to wear?
• Will the car start first time or will they have to let the engine warm up?
• How many red lights and how many green lights will they come across on our way to work?
• Will they find a parking spot near the entrance to the office?

The arrival time depends on how these simple actions, with their associated variation, are combined together.

The variation associated with each action is identified mathematically by its ‘variance’, whereas variation associated with the whole process is identified by the combination of the set of variances relevant to each action, which is the co-variance. As a matter of fact, processes inside an organization are highly interconnected and interdependent, and predicting the outcome of a sequence of events becomes very difficult.

In our next post we will lo0k at variation and processes.

This post is an extract from the book ‘Sechel: Logic, Language and Tools to Manage Any Organization as a Network’