Part one of three · Thinking in Systems

The Wrong Queue

Why the thing you can see is almost never the thing that's stuck.

Maria runs the front desk. She works through lunch most days. The phone starts before the lights are on and there is a queue at her window from opening until close. When patients write in to complain, and they do, they complain about her desk.

She is, visibly and unarguably, the hardest-working person in the building.

She is also irrelevant to how many people this clinic helps.

It treats twenty patients a day. It has treated twenty patients a day for four years.

One of them is a man in his sixties who has now come twice about a cough he has had since March. He waited three hours on Tuesday and left at four to collect his grandson. He will try again next week.

Maria was never the problem. She was busy absorbing the anger that something else was generating.

The station that sets the number — the one every other station is waiting on, whether it knows it or not — has a name. It is the constraint. Every system has one. Almost no system knows where its own is.

Almost everyone says yes.

But look at what full effort at the front desk actually produces. Sixty people checked in, against twenty the doctor can see. The other forty do not vanish. They sit in the waiting room — and a waiting room is not output. It is forty afternoons, spent.

A station that isn't the constraint, running at maximum, manufactures work the constraint cannot absorb. In a factory that work is inventory, stacked and going stale. Here it is people.

So a well-run system needs most of its parts to have slack — deliberately running below their maximum — so the one part that governs everything can run flat out.

When everybody is at full stretch, the honest reading is not that the place is well run. It is that nobody has found the constraint, so everyone is compensating for it at once.

We know it is the doctor. So hire another doctor.

Not yet. Watch her day first.

Dr Adeyemi has been here nine years. She stays until seven most evenings writing up notes she had no time to write in the room.

Six more people a day. Around fifteen hundred a year. The man with the cough is one of them. The cost of this is nothing.

That is the step almost everyone skips — and it is the same move as the hiring decision, made smaller. You are not adding capacity to the system. You are moving capacity toward the constraint. It just happens that the cheapest place to find some was already inside it.

The next step is harder, because it looks like failure. Everyone else now runs at whatever pace keeps Dr Adeyemi working. Reception deliberately slows down. Every local number you have will show those people getting worse while the clinic gets better. If you cannot tolerate that on a dashboard, none of this is available to you.

Only then do you spend money — and you spend it on the doctor, not on the loudest queue.

Now, a clinic is a tidy example: five stations, one line, one direction. Almost no real work looks like that. There are loops and parallel paths and jobs that come back half-finished, and you cannot draw most organisations as a row of boxes.

The messiness changes the arithmetic. It does not change the property underneath it. In anything where the parts depend on each other, a very small number of places govern what the whole thing produces — and they are almost never where the noise is.

A parent getting older. There are a dozen things to worry about — the stairs, the medication, the driving, whether the money lasts — and you are working on all of them, badly, in the evenings, and getting nowhere. It feels like there is simply too much.

There isn't. There is one thing. The assessment that unlocks the funding that pays for the help, and it needs a referral nobody has requested, and until that happens every other thing you do gets undone by Wednesday.

It is not the biggest problem on the list. It is the stuck one. You can work extremely hard on a system for years without ever once touching the thing that governs it.

It is still a constraint. It just now needs an entirely different set of skills to attack.

Which is why this work never finishes, and why the most expensive sentence anyone in that clinic can say is we solved that one.

All of this was easy. Five boxes, five numbers, find the smallest.

Real organisations do not hand you the numbers. And the constraint is usually not a person or a machine at all. It is a rule.

Breaking one is nearly free. A rule costs nothing to change, next to hiring a doctor. And it usually improves things more than any purchase would.

But you will not find it, because everybody experiences it as normal. A bottleneck machine announces itself: there is a pile of work in front of it. A bottleneck rule has no pile. It hides inside that's just how it's done, and it is defended, sincerely, by good people whose own numbers depend on it.

So: find the constraint. Put nearly everything into it. Ignore, for now, almost everything else.

But the constraint is usually a rule. Rules are invisible. They look normal. They are defended in good faith by people who are not wrong to defend them.

And they connect to the symptoms you can actually see through long chains of cause and effect that cross every boundary on the org chart.

Ask five colleagues what the most limiting thing about your organisation is. You will get five confident, different answers.

On what basis do you decide who is right? Seniority? Confidence? Volume?

Most organisations settle it with some blend of authority, anecdote and whoever tells the best story — which is to say they don't settle it.

Then they spend a year, a budget and a great deal of everyone's goodwill finding out whether the guess was right.

That is not a management problem. It is a thinking problem — which is a more hopeful thing to be, because thinking problems have techniques. This one has had a technique for forty years. Hardly anybody is taught it.

And so, somewhere in that building, someone is working through lunch on something that was never going to matter.