There’s a quiet anxiety running underneath every conversation about self-managing teams, flatter structures, and distributed authority, and it’s worth naming plainly: if teams don’t need someone telling them what to do, what are leaders for?
The anxiety produces two bad answers. The first is denial — leaders who hear “complex adaptive system” and keep right on operating the org like a machine with a cockpit, pulling levers that are no longer connected to anything. The second is abdication — leaders who conclude that since control is impossible, leadership is obsolete, and recede into a kind of well-intentioned absenteeism dressed up as empowerment. Both answers share the same flawed premise: that control was ever the part of leadership that created the value.
It wasn’t. And once you let go of it, you can see what was underneath the whole time. I wrote an early version of this argument years ago as a list of things great leaders do in “contemporary” organizations. Returning to it now, I notice something the list-format hid from me: the three jobs that survive the death of control aren’t three separate things. They’re one thing, pointed in three directions. In a complex system, the leader’s enduring work is tending the organization’s flows of information and attention — making sure context flows down, learning flows across, and interference stays out.
Let me take the premise seriously first, briefly. An organization is a complex adaptive system — a network of people and teams, each acting on local information, producing collective behavior nobody is steering. Such systems cannot be precisely controlled by any single actor; they can only be influenced. Teams in this kind of network don’t merely not need someone directing their every move — they are actively damaged by it, because micromanagement severs them from the local information that was their entire advantage. But “can’t be controlled” has never meant “can’t be tended.” A gardener controls nothing — not the weather, not the genome, not a single leaf — and yet the garden with a gardener is unmistakably different from the lot without one. The question is what, exactly, the tending consists of. Three flows.
First flow: context, downward
Whatever else changes about organizational structure, one asymmetry persists: position grants vantage. A leader sits where certain information converges — the board conversation, the market pattern visible only in aggregate, the strategic constraint, the reason behind the decision. Teams close to the work hold the richest information about how things are actually going; leaders hold information about why the surrounding decisions look the way they do. Neither can substitute for the other.
Here’s the reframe that matters: that vantage is not an asset you hold. It’s a debt you owe. Every piece of context a leader possesses and a team lacks is a decision that team will make slightly blind — and in a complex environment, where you want teams making fast local decisions, the quality of those decisions is capped by the context they’re made within. A leader hoarding context while urging teams to “take more ownership” is asking people to drive faster while painting over the windshield.
The discipline of context-sharing is harder than it sounds, because of one specific temptation: attaching instructions. Sharing what you see, and then prescribing what to do about it, isn’t context — it’s direction with a longer preamble. The practice that actually builds capable teams is sharing your vantage without the attached action: here’s what I’m seeing, here’s why that decision above us went the way it did, here’s the constraint nobody talks about — make of it what you will. It’s worth saying that this has nothing to do with flattening. Hierarchy of vantage is unavoidable and fine; it’s the hierarchy of command that complexity punishes. The leader’s altitude is only a problem when it’s used for steering instead of seeing.
Second flow: learning, across
The second flow runs laterally, and it’s the one leaders most consistently neglect because no org chart makes them responsible for it.
A network’s power doesn’t live in its nodes. It lives in its connections — and specifically, in whether things learned at one node can reach the other nodes that need them. Organizational learning is expensive in a way we rarely account for: a hard-won lesson costs weeks of someone’s attention, a failure endured, a breakthrough that took three frustrating detours. When that lesson stays trapped in the team that paid for it, the organization bought the learning at full price and used it once. When it flows, the same purchase pays out everywhere. Silo’d organizations aren’t just slow — they’re running the world’s worst investment strategy on their own experience, paying full tuition in every classroom for the same lesson.
The flow doesn’t stop on its own; it gets stopped — by structures that give teams no occasion to talk, by incentives that quietly rank internal competition above shared purpose, by cultures where pausing to reflect and share reads as slacking. None of these blockages can be cleared by the people experiencing them, because each one is bigger than any single team. This is precisely where a leader’s position becomes useful again: not to carry the information personally (the leader-as-router is a bottleneck with a title) but to find where the pipes are pinched and use their standing to unpinch them — building the occasions where teams show each other their work, making transparency the default working style, dismantling the incentive that made hoarding rational. The leader doesn’t move the water. The leader fixes the plumbing.
Third flow: interference, outward
The first two jobs are about helping good things flow. The third is about keeping bad things out — and it may be the most valuable thing a leader does in any given week.
Teams trying to do focused work inside a large organization are under constant bombardment: the urgent request from above that isn’t actually urgent, the executive drive-by that spawns a week of unplanned work, the new reporting requirement, the reorg rumor, the meeting that exists because someone senior got nervous. Each interruption seems small from where it originates. From inside the team, they are a weather system — and attention, the one resource the team’s work actually runs on, is precisely what the weather destroys.
A leader is positioned to do something about this that no one inside the team can: absorb pressure instead of transmitting it. When the anxious request comes down, the protecting leader metabolizes it — answers what can be answered, negotiates what can be deferred, takes the heat personally rather than converting it into team thrash. They notice which recurring demands produce nothing and spend their political capital killing them. They secure what the team needs — time, room, money, cover — and they treat an unbroken stretch of their team’s attention as the asset it is. This has only grown more vital as the volume of incoming demands has accelerated; requests and “quick asks” are now generated at machine speed, and the human attention they land on has not gotten any larger. In the organism metaphor I’ve always liked for this: teams are cells, and cells need three things from their environment — nutrients in, waste out, toxins kept away. The leader is the membrane.
The price
I’ve saved the uncomfortable part for the end, because it’s the part my younger self left out, and it explains why this style of leadership remains rare despite being widely praised.
All three flows require the leader to spend power rather than accumulate it. Look at the ledger honestly. Sharing context dissolves information asymmetry — and information asymmetry is one of the oldest power strategies there is. Fixing the lateral plumbing eliminates the leader’s role as gatekeeper — and gatekeeping is what made many leaders feel indispensable. Absorbing pressure means taking hits that could have been passed downward — and passing them downward is the traditional perk of altitude. The controlling leader’s importance is visible every day, in every decision that has to route through them. The tending leader’s importance is visible mainly in the counterfactual: things flow, teams flourish, and it’s genuinely hard to point at why.
Which yields the only honest test I know for this kind of leadership: does the system work better when you’re not in the room? Not the same — better. Teams deciding well because they have the context. Lessons traveling without you carrying them. Focus holding because the interference is being absorbed somewhere above. If the answer is yes, you’ve done the job that was always underneath the control — the one that doesn’t expire when the org chart flattens or the machines start drafting the memos. Great organizations were never great because of what their leaders decided. They were great because of what their leaders kept flowing.
A note on lineage: an earlier version of this argument appeared in 2016 as “Three Things Great Leaders Inside Contemporary Organizations Do," written during my decade at The Ready. The three jobs — sharing context, unblocking information flow, protecting team health — began there; unifying them as flows, the context-as-debt reframe, and the argument about spending power are where the thinking has gone since.