Process Becomes the Point

Category: Organisations and systems Systems persist long after their original purpose fades.

A software company introduced a change-approval board after a bad outage. The logic was sound: before any change went to production, a small group would review it, ask a few questions, and catch the obvious risks. In its first year the board genuinely prevented several bad releases. It existed to protect customers from breakage, and it did.

Years later the board still met, twice a week, without fail. But the work had changed. Deployments were now automated, small, and reversible in seconds; a bad change could be rolled back before the board’s minutes were even written up. The questions the board asked had been designed for a world of large, risky, irreversible releases that no longer existed. Engineers prepared the paperwork, sat in the meeting, answered the ritual questions, and got their stamp. The stamp protected nothing. But getting the stamp was now the job — a change without one was a violation, regardless of whether it was safe.

No one could say when the board had stopped catching real risk. The meeting still happened. The form still got filled in. The approval was still granted. The thing it was built to deliver — fewer outages — was no longer something anyone was actually checking. They were running the process beautifully. They had simply forgotten to ask what it was for.

The Principle

A process is built to deliver an outcome; over time the process becomes the thing people serve, and the outcome it was meant to deliver is quietly forgotten — so the system keeps running perfectly while producing nothing of value, and questioning the process feels like the deviance, not the dead outcome.

The move at the heart of this is a substitution of ends. The means becomes the end. Following the steps correctly turns into the definition of doing the job well, fully decoupled from whether the steps still produce what they were invented to produce. Compliance is rewarded; outcome goes unmeasured. The process has been promoted from servant to master, and nobody signed the promotion.

This is the close cousin of the metrics trap, where a measure becomes a target and stops being a good measure — but it is the non-metric form of it. Nothing here is being gamed and there is no number in sight. What gets served is a procedure. A process is, in the end, a frozen answer to a question someone once asked. The danger is that the question can stop being asked while the answer goes on being given, on schedule, forever.

Why It Is Inevitable

Processes are easier to follow than to evaluate. Checking that you did the steps is cheap and objective — they are written down; you either did them or you did not. Checking whether the outcome still matters is expensive and contestable, and it requires someone to stick their neck out. Faced with a cheap, clean test and an expensive, awkward one, an organisation will quietly adopt the cheap one as its definition of doing well.

The people who run a process also have no standing or incentive to question its existence. Their competence is defined by executing it well. Asking whether the whole thing should still exist is, from inside the role, an act against oneself — it is the one question whose answer could make the job disappear. So the people closest to the process are the least likely to challenge it.

The original purpose, meanwhile, was often tacit, and it left with the people who set the thing up. The founders of the change-approval board knew exactly what it was for; the engineers inheriting it ten years later were handed the steps without the why. They learned to run it, not to interrogate it, and there was no one left who remembered it had ever been optional.

And running the process feels like productivity. It produces visible activity — meetings held, forms completed, stamps granted — which is socially safer than stopping to ask whether any of it should be happening. Doing the dance looks like work. Questioning the dance looks like trouble. Of course the process becomes the point: following steps is cheap, evaluating purpose is expensive, and no one’s job is to ask.

How It Shows Up

• The reason a process exists can no longer be stated by the people who run it — only the steps can.
• Doing the process correctly is treated as the job; whether it produced anything is no one’s question.
• “That’s the process” is cited as a complete answer to “why are we doing this?”
• Effort and pride attach to running it smoothly — better forms, tighter meetings — never to whether it should run at all.
• Skipping it is treated as deviance even when it changes no outcome; following it is treated as virtue even when it changes no outcome.

Why It Causes Damage

Real cost and attention go into producing nothing. People’s time, the most expensive thing an organisation has, is spent feeding a ritual that no longer feeds anything back. That alone would be wasteful. The worse harm is that the process looks like control, so it conceals the absence of the thing it was meant to deliver. The board still “protects” releases, on paper, which means nobody is building the protection the new world actually needs — the dead process occupies the space where a live one would go.

It also crowds out the judgement it replaced. A process exists to spare people from thinking the same thing through every time, which is genuinely useful — until the day the situation changes and the steps no longer fit. By then people have stopped thinking, because the steps were thinking for them, and they do not notice that the steps and the situation have come apart. The ritual runs on, smoothly, over a gap it can no longer see.

And the dead process is the hardest thing in the building to remove, because it is self-defending. Questioning it reads as recklessness or laziness; following it reads as diligence. The compliance culture that keeps it alive also punishes anyone who points at it. So when something does go wrong, the instinct is never to ask whether the existing process is still connected to a purpose — it is to add more process. The graveyard fills, and every headstone is defended.

How To Counter It

• For every standing process, write down the outcome it exists to produce — in outcome terms, not activity terms. If no one can state it, that is the finding.
• Periodically test the process against its purpose, not against itself: would removing it change any real result? Run the experiment wherever you safely can.
• Give every process an owner who is accountable for the outcome, not for the running — so that someone has standing to retire it.
• Build in expiry or review by default: a process should have to justify its continued existence, not be presumed permanent.
• Separate “did we follow the steps” from “did we get the result” in how people are judged — and reward the result, not the ritual.

What Good Looks Like

In a healthy organisation, every process can answer “what would break if we stopped?” with a real answer. Retiring a process that has outlived its purpose is treated as good housekeeping, not heresy, and the person who spots the dead one is thanked rather than eyed with suspicion. Following the steps is never confused with doing the job.

Processes still exist there — they are useful, and a place with none would re-solve the same problems forever. The difference is that they remain servants of an outcome someone is still watching. The change-approval board in such a place either evolves to match the new reality or is quietly dissolved, because someone never stopped asking what it was for.

A Reflective Question

Think of a process your team follows without question — can you state, in terms of an outcome, what would actually go wrong if you stopped, or only what step you would be skipping?