The Architecture of Flow

Two mornings

Same creek, same four men, same boat, two Saturdays apart.

The first Saturday the coach gives us a long, easy row. Low rate, light pressure, forty minutes. It is the kind of session that does real physiological good and asks almost nothing of the mind. By minute twelve I am thinking about breakfast. By minute twenty I am thinking about an email I have not answered. The boat is fine. I am somewhere else.

The second Saturday someone has the bright idea of racing a younger crew in a full-pressure piece we have not trained for. Within a minute the rate is too high, the catches are ragged, and my body is busy doing arithmetic it cannot win. I am not thinking about breakfast now. I am thinking about survival, and about how the young men in the other boat are probably not thinking about either.

Neither morning was flow. One was too small for my attention and the other too big for it. Somewhere between those two Saturdays sits a third kind of morning, the one from the first article, when the work is hard enough to hold every part of us and not so hard that it scatters us.

You know these two mornings even if you have never rowed. The piano student playing scales she mastered years ago, and then the same student sight-reading a concerto in front of her teacher. The new nurse given nothing to do on a quiet shift, and then the same nurse alone with three emergencies at once. The skill did not change between the two. The fit between the skill and the moment did.

The first article asked what flow is. This one asks what tends to make room for it. The honest answer is that flow cannot be built to order. But it does seem to arrive in a recognizable architecture — a set of conditions that make it more likely, the way certain weather makes fog more likely without promising it.

The channel between boredom and anxiety

Csikszentmihalyi’s first model was a simple diagonal. Put challenge on one axis and skill on the other. When challenge runs well ahead of skill, we get anxious. When skill runs well ahead of challenge, we get bored. Between them runs a channel where the two stay roughly in step, and the people he interviewed seemed to find flow inside it.

It was a lovely idea with a hole in it. In the 1980s two researchers at the University of Milan, Fausto Massimini and Massimo Carli, pointed out that a perfect balance of almost no challenge and almost no skill — washing a cup, waiting for a bus — sits squarely in the channel and is plainly not flow. Their eight-channel model, and the quadrant model Csikszentmihalyi published with Judith LeFevre in 1989, measured challenge and skill against each person’s own average. Flow, on this view, is not where challenge equals skill. It is where both are high for you.

Read the map loosely. It describes perception, not hard limits, and the center moves with the day. The same piece can sit in the flow zone on Tuesday and in anxiety on Saturday, because sleep, confidence and whatever the week has been doing to us all shift where we stand.

What the evidence says about balance

The quadrant model came out of a clever method. Instead of asking people to remember their best moments, Csikszentmihalyi’s group gave them pagers that went off at random times through the day. Each time it buzzed, people wrote down what they were doing, how challenging it felt, how skilled they felt, and how they were experiencing it. Thousands of these small reports, across ordinary weeks, made it possible to see where in daily life absorption actually happened. One famous finding was how often people reported it at work, while saying they would rather be at leisure.

Since then the challenge–skill idea has been tested many times. A 2015 meta-analysis by Carlton Fong, Diana Zaleski and Jennifer Leach pooled 28 studies. The link between challenge–skill balance and flow was real but moderate, and it varied by culture, age, setting and how balance was measured. Balance was a meaningful correlate of flow, alongside clear goals and a sense of control. It was not the whole story.

The details complicate the tidy picture in useful ways:

• Balance helps; high balance helps more. A 2025 study of 528 online board-game players, mostly chess and Go players, found flow was higher when both challenge and skill were high than when both were low.
• Slightly too easy beats slightly too hard. The same study found flow was higher when skill exceeded challenge than when challenge exceeded skill. The authors caution that this may hold mainly for structured, rule-bound activities like games, so treat it as a finding about that setting, not a universal rule. Earlier work by Stefan Engeser and Falko Rheinberg found something similar: when stakes are high, feeling a little ahead of the task may be what lets people relax into it.
• It is perception, not measurement. Every one of these studies asks people how challenged and how capable they feel. Two people facing the same task can sit in completely different zones of the map.

For anyone hoping to arrange more absorbing work, that last point is the practical one. You cannot fix the challenge–skill fit from the outside alone. A teacher can set the right problem; the student still has to believe she can get somewhere with it. A coach can design the right piece; the crew still has to feel it is within reach.

Knowing what matters now

The second condition is almost embarrassingly simple. People in flow know what they are trying to do next.

Notice the word next. It is not the same as knowing your life’s purpose, or this season’s target, or even the goal for today’s session. A rower in the middle of a hard piece is not thinking about the regatta in May. She is thinking about this catch. A surgeon is not thinking about the patient’s five-year outcome while she ties the knot. A cook during service is thinking about the ticket in front of him. Big goals give work its meaning. Near goals give attention somewhere to land.

Psychologists have known this for a long time. In a well-known 1981 study, Albert Bandura and Dale Schunk gave children struggling with arithmetic a long series of problems. Children who set themselves small goals — a few pages per session — made more progress and grew more interested in the material than children working toward one distant target. The near goal did not shrink the ambition. It made it reachable from where they stood.

This is also why vague instructions are so tiring. Be better. Work harder. Write something good. None of these tells attention where to go, so attention goes back to the self and asks, anxiously, whether this is it yet. In coaching I learned to swap ten technical points for one task the athlete could feel: move the boat between strokes. In my years using exercise as medicine, with people recovering from illness and injury, it was the same. Get your strength back is a hope. Stand up from this chair five times without using your hands is a place to put your attention.

Near goals do not have to be rigid. Some of the richest flow happens in open-ended work — improvising, sketching, a conversation going somewhere unplanned. Even there, the next move is clear even when the destination is not. The jazz player does not know where the solo will end. She knows exactly what the next phrase has to answer.

Feedback you can read as you go

The third condition is feedback, and the kind that matters most is the kind that arrives during the work rather than after it.

A boat is extraordinary at this. The hull answers every stroke within a fraction of a second. Run means it worked; check means it didn’t. Nobody has to explain it. A climber gets the same honesty from the rock, a potter from the clay turning under her hands, a teacher from thirty faces, a comedian from the silence after a line. The information is immediate, it is about the task rather than about you, and it tells you what to do next. Then attention can stay where it is instead of leaving to go and look for a verdict.

Compare that with work where the feedback comes in a quarterly review, or a grade three weeks later, or an email from a reader months after publication. None of that is useless. But it cannot hold attention in the moment, and in its absence the second rower fills the gap. Was that good? Will they like it? How am I doing? When the work does not tell us how it is going, we go and ask ourselves, and we are rarely kind.

Two distinctions help here:

• Task feedback versus judgment. “The boat checked at the catch” keeps attention on the water. “You were sloppy” sends it straight back to the self. The most useful coaches, teachers and editors I have known speak almost entirely in the first kind.
• Feedback you can feel versus feedback you have to fetch. A number on a screen can help, and I have built a career partly on measuring things. But a number you have to look away from the work to check is a small interruption, and a stream of them can become a different kind of clipboard.

None of this means removing measurement. It means asking, of any piece of work, whether there is a way for the work itself to tell you how it is going while you are still doing it. Writers read a paragraph aloud. Musicians record a run-through. Nurses watch the patient, not just the monitor. The aim is a short loop between action and answer. Speed is not the whole of it. Feedback has to be readable and pointed at the right thing. In some work, especially creative work, the response comes late or misleads: an audience that laughs at the wrong line, an early reader who dislikes a draft for reasons that have nothing to do with it. Fast feedback helps only when it is relevant.

Choosing it

There is a quieter condition that the original model underplayed, and recent researchers have brought it back to the front: whether we chose the activity, and whether we care about it for its own sake.

Remember where flow was first found. Csikszentmihalyi interviewed climbers on walls they picked, dancers in styles they loved, chess players who would have played for nothing. In 2024 Oliver Durcan and colleagues argued that this is exactly what many laboratory studies leave out. They ask people to do unfamiliar tasks they did not choose, then wonder why flow is hard to find. Their proposed framework puts autonomy — the sense that this is your activity, done your way — alongside challenge and skill.

That fits a much larger body of work. Edward Deci and Richard Ryan’s self-determination theory, built over four decades, holds that people engage most deeply when three needs are met: autonomy, competence, and relatedness, the feeling of being connected to others in what we do. Flow is not the same thing as intrinsic motivation, but they often reinforce one another, though neither is a guarantee of the other.

This is not an argument that we only flow when we do exactly what we like. I have rowed thousands of sessions I did not design, at hours I would never have picked. What made them mine was something else. I chose the sport, I chose the crew, and I cared about not letting the other men in the boat down. Autonomy does not always mean control over the details. Sometimes it means having said yes to the whole. That is not quite the same as enjoying every hour for its own sake. Valuing a commitment and loving the activity often travel together, but a person can fully endorse a choice and still not enjoy all of what it asks.

The same is true of work most people would never call flow-friendly. A nurse on a night shift did not choose the patient who arrives at 3 a.m. But she chose the profession, and if she still cares about the person in the bed, there is room for the hours when the shift stops being something she endures and becomes something she is fully inside. That is not the same as every demand feeling self-directed. Much meaningful engagement happens under heavy outside constraint — schedules, protocols, other people’s emergencies — and some of it never feels chosen at all. Saying yes to the whole can make room for absorption. It does not guarantee it.

The thieves of attention

A boat at dawn has one great advantage over most modern work: there is no phone in it. Nobody checks their messages between strokes. For an hour, the only thing asking for attention is the water.

Most of us now work in conditions designed to do the opposite. The organizational psychologist Sophie Leroy has spent two decades studying what happens when we switch between tasks. In a 2009 paper she gave the problem a name: attention residue. When we move from one task to another before finishing the first, part of our mind stays behind with it, and performance on the next task suffers. Time pressure made it worse. Simply finishing one thing before starting the next was not always enough to clear it.

That matters for flow because flow can take time to develop, and it often does not arrive in the first few minutes. An interruption can send us back toward the start of the climb, sometimes carrying a little of the last task along for the ride.

Leroy and her colleague Theresa Glomb later found a simple counter-measure. When you are interrupted, take a moment to write down exactly where you were and what you planned to do next. That small “ready to resume” note lets the mind set the unfinished work down instead of carrying it. It may be the cheapest useful advice in this whole series.

The broader point is not that we must live like monks, though several of the monks I have read about were very productive. It is that many kinds of flow benefit from a protected stretch of time, and we usually have to build the wall ourselves:

• One task, one stretch. Decide in advance what this block of time is for.
• Close the loops on the way in. Write the ready-to-resume note for whatever you are leaving.
• Make interruption cost something. Phone in another room, notifications off, a door that closes.
• Expect a slow start. The first ten minutes are often the least absorbing. They are the toll, not the verdict.

Two roads to excellent work

Here is a finding that takes some of the romance out of flow and replaces it with something more useful. Flow is not the only state in which people do their best work.

In 2016 the sport psychologist Christian Swann and colleagues interviewed professional golfers soon after tournaments where they had played exceptionally well. The golfers described two different experiences. One they called letting it happen: easy, absorbed, confident, the classic flow account. The other they called making it happen: intense, effortful, deliberately focused, usually arriving late in a round when something was on the line. A follow-up study across many sports confirmed the pattern, and Swann’s team proposed an integrated model with two states. Flow, and what they called clutch.

Treat the table as a description of two different experiences of performing well, not two sealed boxes. Swann’s own team found athletes moving between the states within a single performance, and nobody has shown that they are distinct states in the brain.

I recognize both from the boat. There are pieces where the crew drops into rhythm and the work seems to row itself. And there are last 500 meters where nothing about it is effortless, where everyone is deciding, stroke by stroke, to keep going, and where we still row as well as we ever have. The second is not a failure to reach the first. It is a different tool for a different moment.

The same is true well beyond sport. A surgeon may be in flow for an hour of routine work and then switch into clutch when something bleeds. A parent at the end of a long day can be all deliberate effort and still be doing it beautifully. A writer can draft in flow and edit in clutch.

This matters for the series, because it takes away a pressure the flow literature can accidentally create. If you think excellence requires effortlessness, every effortful moment feels like evidence that you are doing it wrong. It isn’t. Sometimes the right state is the one where you are trying very hard, on purpose, with your whole attention.