How much moving fast and breaking stuff until we're great?

Every organization contains some degree of waste, fraud, and abuse. The answer isn’t in eliminating it entirely—it’s in deciding how much dysfunction we’re willing to tolerate before the fix becomes worse than the flaw.

There’s a moment in every well-meaning project, every corporate initiative, every government overhaul where someone says, “We’ve got to root out the inefficiencies.” This sounds reasonable. Noble, even. But eliminating dysfunction often means creating a new more dysfunctional machine.

Welcome to the paradox of organizational reform. 

The academic roots of this go all the way back to sociologist Robert K. Merton, who in 1940 wrote a paper called The Bureaucratic Structure and Personality. I read things like this because I am fun and great to talk to at parties. 

Dense and boring though the book may be, it contains brilliance: bureaucracies, by their very nature, create problems they cannot solve. 

Merton coined the term “bureaucratic pathology” to describe this. Essentially, organizations need goals in order to function—but because many real-world goals are abstract (like justice, safety, or innovation), bureaucracies create proxies to measure them.

“Justice” is hard to quantify. So we use conviction rates as a benchmark for prosecutorial success. The problem is that conviction rates are only loosely correlated with actual justice. By this metric North Korea's staggering 100% conviction rate would enshrine them as a paragon of righteousness.

This is Merton’s pathology: the benchmark replaces the goal. When a system rewards the proxy and not the principle it optimizes for chasing a phantom.

Fast-forward to now. Organizations everywhere—from your local DMV to transnational organizations are trying to do more with less with a dollop of innovation. Silicon Valley launched this efficiency Renaissance with its move fast, break things ethos. But iterating code and iterating organizations are two different things.

Enter: DOGE.

Elon Musk’s so-called "Department of Government Efficiency" (DOGE) in theory was a satirical but pointed response to the bloated structures he believed were dragging the US government down. DOGE is tasked with streamlining decisions, axing departments, and killing the very bureaucracy it emerged from.

Here's Merton's paradox again. Because if DOGE works too well, it becomes a bureaucracy. Just one with a different hat. The second you build a system to destroy inefficiency, you have created a new system with new proxies. 

Waste becomes cutting jobs and budgets.

Fraud becomes hunting for dead people receiving social security (they aren’t).

Abuse becomes strong arming allies for tariff parity in spite of them having smaller markets.

The snake eats its tail. The Ouroboros of Nonsense spins on.

This isn’t just tech-bro irony. It’s a structural truth of organizations: the fight against inefficiency eventually becomes inefficient or too costly to maintain.

So here’s my heretical take: there is an ideal amount of waste, fraud, and abuse. You’ll know you’ve reached it when the cost of finding and eliminating the inefficiency is greater than the inefficiency itself.

No perfect system exists and one which tears itself apart in pursuit of purity cannot function. 

Look no further than the Roman Empire for a lesson in the carrying capacity of inefficiency. They had a concept called administrative slack—a recognition that too much central control crushed initiative. A certain amount of local leeway, wasteful as it may have been, kept the empire running. The Byzantine Empire? Same deal. In fact, some historians argue that bureaucratic tolerances were what allowed them to outlast their western cousins by a millennium.

So the next time you hear about a new task force, committee, or innovation lab “reimagining systems from the ground up,” take a breath. Inefficiency may be obvious. But trying to fix it might be worse if goals and actions don't align.

We live in a world obsessed with metrics, dashboards, and key performance indicators but proxies have to closely reflect the goals the system hopes to achieve.

The US wants to do more with less overhead. This is an admirable goal, but at what point are we trading one bloated system for another in the pursuit of impossible purity?