dS≥0 and Corporations
The second law of thermodynamics was discovered around 1860 according to various publications and attributed to Rudolf Clausius (1822-1888) who is considered the founder of Thermodynamics. The law has several versions, which mean the same: In a closed system, the amount of disorder can only grow or remain constant, unless there is an intervention from outside the system. That intervention is in the form of energy that enters the system from outside.
Although some of the initial assumptions leading to this law turned out to be incorrect as years passed and research progressed, the second law remained true and is one of the axioms of thermodynamics. Its implications are far reaching. To name a few examples: explaining the "arrow of time", why perpetual motion is impossible, what is the maximum energy efficiency of machines and how the universe may end.
Disorder or randomness is measured by a quantity called Entropy (also termed by Clausius), which also measures the number of microscopic states within the system that corresponds with a given macroscopic state, or the amount of "information" that is "contained" in a system.
Complex systems have many microscopic states, which can lead to the same macroscopic situation, thus - spontaneous reordering the system back to its initial microscopic state or to any other microscopic state can be considered highly unlikely. To achieve a desired microscopic state other than the one the system is at, one must invest a lot of energy or effort into the system.
A common example given in physics lessons is of a bucket of clear water, to which an experimenter drips a single drop of colour. This represents the initial macroscopic state of the system. This state is of low entropy as the system has evident order - areas with- and areas without colour.
Now, if the system is left alone, after a while the drop naturally dissolves in the water, or the experimenter stirs the water to get the colour to spread and tint the water. This is another macroscopic state that is much less ordered (high entropy) than the initial state, since the molecules of colour spread randomly between water molecules in almost uncountable arrangements.
The theory says that it is improbable that the system spontaneously returns from the later macroscopic (high entropy) state to the initial (low entropy) macroscopic state without investment of energy from outside the system. In fact - for the system described in this example, science and engineering has not yet devised a practical way, even with energy investment, to return to the ordered low entropy state.
So, why am I telling you all this, business leaders?
As a manager and technical leader in large organizations I noticed, over the years, that companies too have "entropy". As the organization grows and takes on more challenges, commits to more products and expands across different geographies and segments of its business, it gains degrees of freedom, or microscopic states that tend to have complex interactions with other microscopic states.
With time, more people join, and structures of management form to operate them, leaders are grown and direct the effort and interfaces are built between divisions to regulate the interplay.
However, disorder and entropy continue to grow, sometimes in the shadows (hidden from management attention) and sometimes in the light (but still unsurmountable), making the company less focused, less efficient, and less successful.
Good managements notice that and invest energy to bring new order and resolve these inefficiencies, and for a while they are successful, but as new growth happens - entropy grows as well, trumping management efforts.
Eventually, there is no more energy to invest and no known process to improve the interplay between the "molecules" of the organization - its microscopic states that combine to a macroscopic success or failure of the company.
At that point - an inevitable decline occurs, stemming not from a single, identifiable source or root cause, but everywhere in the organization, because of the combined entropy and "complicatedness" of the organization.
This situation resembles our water and paint droplet example's final macroscopic state. It will require a huge amount of energy and a miraculous plan to restore order and reduce entropy, for growth to be possible again. It is simply not possible to continue to grow as a single system anymore. Realization is painful and usually long overdue.
领英推荐
So, all is lost... or is it?
Luckily, a company is not a closed, indivisible system, so there are things we can do to reduce the disorder. When complexity and entropy is already extremely high and reducing it directly to obtain business efficiency is an unsurmountable task, we may opt for other options:
We can split the company to smaller businesses and force them to work independently, converting the mother company to a holding structure. We can bring new managements through external acquisitions and build those smaller businesses around them, in order not to carry unnecessary baggage from the old organization. Identifying an efficient business in a similar field and acquiring it, then slowly growing it by transitioning departments into the new organization. This will effectively split the mother company and form a smaller, effective organization.
Another thing we can do is to identify interfaces between teams within the company and try to reduce them. Redundant interfaces generate complex rules of engagements and more opportunities for entropy to emerge. Whenever team A collaborates with team C through team B - that is an opportunity to reduce complexity and eliminate two interfaces A-B and B-C.
Finally, the internal organizational structure may hint for more opportunities to reduce complexity. For example, if each project in the company had its own architecture team it is likely that complexities and specific traits have evolved in the architecture organization of each project. Changing the organizational structure from vertical to a horizontal architecture team may reset this type of complexity and force projects to reuse architectures.
Conversely, if the company operated for too long in a horizontal organizational structure, moving some or all its horizontal organizations to vertical, product specific organizations may also reset the complexity.
What is missing for the application of these ideas?
There are some loose ends in this theory that if tied, will ease its application in real businesses. For example, how to measure the disorder or entropy in the organization? can metrics be developed or is it a subjective assessment done by management and influenced by the personality and style of the leaders? Some will focus on efficiency; others would focus on the thinking processes and so on. When dealing with organizations and people, the strict definitions of physics may give way to softer understandings.
Another point worth considering: Can we detect trends in the organization early on and discover paths by which entropy grows more slowly and the organization is kept efficient for longer time and at larger scales? (in analogy to the Carnot cycle in thermodynamics - which is the most efficient heat transfer process possible), and so on...
Summary and final notes
I will not argue that the analogy described here is not perfect. A company is not a closed physical system. Organizational or operational complexity and entropy are not the same. The takeaway from this discussion is that monitoring complexity and disorder in a company can be worthwhile. Also, the realization that sometimes, calming the water and reducing complexity and disorder directly could be tough. When the levels of "entropy" in the company becomes extremely high and efficiency plummets - calming it down with traditional management "energy" may not suffice and more drastic action may be in order.
Notes, acknowledgements and disclaimers: