Entropy, management, and xkcd 927

Posted on Sat 07 May 2022 in blog • 6 min read

xkcd 927 is a modern internet classic that is frequently brought up in conversations to remind people that a proposal that they’re making will, while being intended to simplify things, actually make them more complicated.

Most people quote that strip to satirize or even ridicule the idea of introducing a “15th standard”, as if the natural order of things was simplification. Such people are frequently baffled by the amount of cruft and clutter that accumulates over time in an organization they work in, and some of them embark on a constant — perhaps career-long — quest of “streamlining,” “process optimization,” or “reducing technical debt.”

If you are one such person, please get ready for some bad news.

As far as we know, there are three fundamental theories that, combined, explain the universe as we know it: general relativity, quantum mechanics, and thermodynamics. Thermodynamics has a famous Second Law that can be stated in various ways — in one modern and simplified form, we say:

The total entropy of a system never decreases.

“Entropy,” in this context, is essentially the degree to which the system is disorderly. In effect, the Second Law states that any system can stay just as orderly as it is now, or it can become more disorderly, but in can never again become as orderly as it once was. The normal state of the world is that things keep getting more and more disorderly.

There are multiple classic examples of this: you can mix two paints in a bucket but cannot unmix them, you can open a container of gas in a vacuum chamber and the gas will disperse but never go back into the container, you can scramble and cook and egg but never return it to its original protein structure.

Sometimes the growth in entropy isn’t noticeable: you can of course pick up your cluttered desk and put everything neatly away in boxes or drawers (or the trash), and your office will look nice and clean and uncluttered thereafter. But, in the process you will have turned so much of your body’s energy into heat that the overall disorder in the “system” (consisting of the things in your office, the room, you, all the gas molecules in the air, and so forth) will have gone up quite considerably.

Now, I realize that not all laws of physics can be directly applied on a macro scale, that is, to organizations, families, or societies. For example, you’ll have to go through various mind-bends to imagine your life as a path through gazillions of Everettian many-worlds bifurcations. But I’d posit that the constant growth of entropy is indeed rather fundamental — after all, growth in entropy is one of our best definitions of the passage of time. Escaping the growth of entropy is literally just as impossible as stopping time.

What does that mean for each of us, individually? It means, bluntly speaking, that our lives get objectively and perpetually messier over time. I don’t know if you’re in a better or worse place than you were 10 or 20 years ago in your life, but I’m pretty sure that you’re in a more complicated place now. And many of us might probably want to go back to our less-complicated life from back then, but alas, backwards time travel (and hence entropy reduction, read: “a more orderly life”) is not an option.

Now as long as you’re just trying (and failing) to rewind disorder in your own life, then — as long as you live and work alone — that will probably not have a harmful effect on anyone. But it gets tricky when you’re applying the same thinking to living with a spouse, or in a family. Good luck trying to rewind your life with teenage offspring, for example, to the presumably simpler time when they were three month old babies that slept most of the day.

But let’s also talk about how this affects your work in a management position.

If you are a manager, it is your job to slow the growth of disorder in your part of the organization. You won’t be able to reduce disorder, and any attempt to do so pits you against a most fundamental law of physics. (Laws of physics are like terrorists: you shouldn’t attempt to negotiate with them.) However, many managers are exactly the opposite: they are entropy accelerators; they speed up the growth of disorder in the organization.

You can do better than that.1 Here are a few suggestions you can apply when dealing with your management peers, so you can act as your organization’s entropy decelerator.

  • Somebody wants to replace multiple existing things with one new thing (the original xkcd 927 scenario): the only circumstance under which you should agree to this is when you already know for certain that the existing things must go away, within a manageable timeframe. For example, the software solutions that your company has been buying from one vendor have had such a massive price hike that they now break the budget, or the legal ramifications of continuing to use them have become untenable. That’s when you have an option of possibly replacing two (or three) things with one. Under all other circumstances, you can hope to replace one thing with one other, at best.

  • Somebody wants to solve a communications issue by adding more channels to your company chat, more categories to your issue tracker, more whatever? That’s your cue to stop that dead in its tracks. Opening more lanes of communication never simplifies anything; it always makes things more complicated. Those new chat channels? Tit for tat. They want three new ones, so they must retire three. No, not two. Three.

  • Somebody wants to “open up team communications”, or “flatten the organization”, so that everyone’s complete graph has way more edges? That’s when you educate them about ${n(n-1)}\over 2$, and what quadratic growth means.

Do you notice how a lot of these involve saying “no” to someone, and that that may place you at odds with the the well-meaning proponent? Congratulations on your realization that leadership is not a popularity contest among your management colleagues.

One word of caution though: even if you fight this good fight — and trust me, it is a good fight — you will still occasionally look back at when you started working in your organization, and realize that despite all your efforts it’s a messier place than when you started. Not just the whole organization, but maybe even your own team or whatever your little corner of the corporate world is. The part where you are responsible for your part of the mess.

This is especially true if you are just in the middle of leaving an organization (or a role therein), and are reflecting on the impact of your tenure: you might fall for the thought of “I tried really hard, but things still are messier than when I got here.” They always will be. The point is not to compare today’s degree of disorder to that when you started. The point is to compare how disorderly it is now, to how disorderly it would have been if you hadn’t been there.


  1. Making a positive contribution to a group in a leadership role is frequently — and somewhat counter-intuitively — achieved by simply focusing on not making things worse for everyone. Canadian astronaut and former ISS commander Chris Hadfield calls this approach “aiming to be a zero” and dedicates a whole chapter in his excellent Astronaut’s Guide to Life on Earth to this idea.