What If We Engineered Management Systems?

Few companies have a designed management system. They have layers, policies, handbooks, dashboards, culture decks and more. But a system, in the true sense of the word? One that was designed, tested, refined, and maintained with the same care you'd apply to a new product? That's rare.

We inherit most of what we call management. Different members of the founding team designed different parts of it when the company started. As circumstances changed, they and their successors patched it in ad hoc responses to new problems, and everyone has lived with the duct tape. The idea that we could engineer the entire system, on purpose, is barely written about and rarely attempted. I only know of a few examples. Alfred Sloan deliberately designed General Motors’ decentralized units with centralized controls—a brilliant system for its time. The U.S. Department of Defense created a comprehensive management system it required of its contractors, which enabled oversight of a sprawling supply chain. Perhaps the clearest example is the U.S. Constitution—an intentionally designed system enabling effective governance while protecting freedom.

Engineering a management system isn’t about designing a utopia. It’s about removing the hidden friction that slows everything down. It means making the tradeoffs visible. It means asking, out loud: “Do our metrics drive the behavior we actually want?” And, more often than not, realizing the answer is: “Not quite.”

It means noticing when well-meaning managers work around the system because the official process gets in the way. It means realizing that tolerance for those workarounds is sometimes compassion, sometimes pragmatism, and sometimes an unspoken confession that the system no longer fits.

Most of all, it means designing something coherent enough that people can trust it and adaptive enough that it doesn’t need to be burned down every five years.

History warns us about the long-term risks of even the best-engineered systems. Sloan’s decentralized structure eventually ossified and nearly destroyed the company. The DoD’s oversight systems, initially effective, became bloated and dysfunctional. Over time, the American government has evolved through patchwork reforms, informal precedents, and reactive adjustments, straying further and further from its original design, and now controls more of American life than George III would ever have attempted. This contrast underscores a fundamental truth: systems left unexamined will degrade.

A well-engineered management system should permit us to reduce bureaucracy, stripping away unnecessary complexity to create clarity. The rules we set should align with the behaviors we measure, the outcomes we reward, and the actions we allow.

If that sounds hard, it is. It requires education, alignment, and a willingness to touch third rails. Because real systems work across boundaries: incentives, power structures, reporting lines, sacred cows. You can’t fix the metrics without revisiting the assumptions. You can’t fix the culture without changing what you tolerate.

But if you don’t engineer your management system, you’re just hoping that the pieces you inherited will line up by accident. Sooner or later, that will blow up.

Where to Begin

Start by noticing the misfires. Look at where decisions get stuck, where performance drifts, where good people quietly work around the rules to get things done. Those moments are signals—early warnings that the system isn’t doing what it was meant to do.

From there, it helps to get clear on a few essentials: What are we measuring, and why? What are we tolerating, and at what cost? Where are we rewarding individual heroics instead of system-wide clarity?

This isn’t something most teams can figure out alone. Not because they lack intelligence or effort, but because they’re already inside the system they need to examine. Sometimes what looks like common sense from the inside is clearly a trap from the outside. The goal isn’t to replace what works. It’s to design what’s missing, remove what’s in the way, and realign the moving parts so the whole thing actually runs.

You don’t need a blueprint for perfection. But you do need a process, and someone who knows how to guide it.