Ops Knowledge Garden

Build the Sprinkler System

Work in progressLast tended 6 min read

Build the Sprinkler System

I did not invent the phrase.

I heard a version of “fire sprinklers or putting out fires” during an operations conversation years ago. It stayed with me. Later, frustrated by an operation that seemed to survive one crisis only by moving immediately into the next, I wrote:

We are always putting out fires. But I want to build a sprinkler system.

That became the cleanest description I had for the kind of work I wanted to do.

Putting out a fire is visible. It has urgency, stakes, and usually somebody who looks competent when it is over. Building the sprinkler system is slower. It requires finding patterns across fires, spending money before the next one happens, and defending work whose success may look like nothing happened.

Organizations say they want prevention. Their attention systems usually reward recovery.

Heroics can make a weak system look healthy

Most essential operations are full of people who refuse to let the mission fail.

The schedule has a gap, so somebody stays late. A process is unclear, so an experienced employee translates it for everyone. Two departments do not communicate, so one conscientious person carries the information between them. Equipment is unreliable, so the crew builds a workaround. A supervisor answers every question because waiting for the actual system to respond would take too long.

The outcome is saved.

That is good, in the immediate sense. It may also be the reason the underlying problem survives.

The organization sees that the shift was covered, the customer was helped, the call was completed, or the deadline was met. The cost remains inside the person who absorbed the gap. It does not arrive on a decision-maker’s desk as a system failure.

Competence can hide the deficit.

I think of this as success becoming the failure mode: the frontline succeeds so consistently at recovery that the system never receives a strong enough signal to fund prevention.

The fleet-safety project made this concrete for me

The fleet-safety work is the clearest example in my own career.

An ambulance accident happens. The immediate response naturally focuses on the driver, the damaged vehicle, the report, and whether a policy was violated. Those things matter. But if the process ends there, the organization becomes very efficient at processing accidents without becoming better at preventing them.

The work changed when we stopped treating every collision as a self-contained event and looked across the record. Accident types, training, policy, technology, locations, behavior, and the way supervisors coached after an event were connected.

The measurable result during my project period was a 30% reduction in the accident rate, from 1.0 to 0.7 per 1,000 trips in the first year.

The number matters. The more important lesson for me was that the reduction did not come from finding one root cause or writing one stronger rule. It came from building a system that could notice, learn, intervene, and continue measuring.

That is the sprinkler system.

Prevention is difficult to narrate

If the system works, the next event does not happen—or happens less often, or is less severe.

There is no dramatic save to point to. You are making a claim about an absence.

That creates a political problem inside organizations. Recovery has a story. Prevention has a forecast, a trend line, or a near miss that somebody must take seriously before it becomes an injury.

This is why leading indicators matter, but it is also why numbers alone are not enough. A hard-braking event, a reopened case, an employee quietly correcting the same process every week—each can be an early signal. The organization still needs a way to hear it.

Recognition can help make adaptation visible. Structured review can show where the same workaround repeats. Better incident learning can recover the conditions that a formal report was never designed to carry.

The point is not to eliminate human adaptation. Adaptation is how complicated work gets done. The point is to stop requiring the same person to make the same save indefinitely while leadership calls the operation resilient.

Not every fire deserves infrastructure

There is a danger in taking the metaphor too far.

Some problems are genuinely isolated. Some risks are too small to justify a new system. Some preventive controls add more friction than safety. An organization can bury itself in procedures built to prevent events that occurred once under unusual conditions.

Building the sprinkler system does not mean turning every lesson into a policy.

It means developing the ability to distinguish:

  • an isolated event from a recurring condition;
  • a useful adaptation from an unsustainable workaround;
  • a person who needs coaching from a system that keeps producing the same trap;
  • a control that sounds reassuring from one that can actually be observed and tested.

That last distinction is becoming increasingly important in my work on chatIR. A candidate control is not a conclusion. It is something the responsible team can validate: was the barrier available, was it used, did local conditions make it impractical, and what evidence would tell us whether it helped?

The unresolved incentive problem

The cure often takes longer than the person approving it expects to stay.

The workaround produces a benefit today. The preventive system may require budget, cooperation across departments, and months or years before the trend is convincing. Everybody can benefit from kicking the cost forward. The next leader, next budget, next shift, or next employee will deal with it.

I do not think this is always a moral failure. Often it is an incentive structure working exactly as designed.

That means the argument for prevention cannot rely only on asking leaders to care more. The cost of repeated recovery has to become visible enough that building the system competes with continuing the workaround.

I am still working on how to do that honestly. Organizations are already paying for weak systems, but the cost is scattered across overtime, turnover, rework, delays, injuries, frustration, and the attention of their best people. Combining those into one dramatic ROI number can create false precision. Leaving them invisible guarantees that prevention continues to look expensive.

The sprinkler system needs a technical design. It also needs an accounting system capable of seeing the fires that competent people prevented anyone else from noticing.

Connected notes

Graph View

Backlinks