Before There Was a Playbook: What Distributed Delivery Taught Us About the Hidden Work of Leadership
Proximity had been doing half the work of leadership — invisibly, and for free — and we only learned its price by removing it.
The call at the edge of the day
There is a particular hour, familiar now to a generation of programme leaders it was not familiar to five years ago, when the working day of one team overlaps with the working day of another several time zones away. It is early morning at one end of the call and early evening at the other, and into that narrow window — two hours, perhaps three — must be compressed every question, clarification, decision and correction that a co-located team would once have spread unremarkably across its whole day. You can hear the strain of it on the line: the slightly-too-long agenda, the anxious efficiency, the sense of a great deal riding on getting everything said before the window closes and half the programme goes home.
A few years ago this hour did not exist, because the whole team sat within walking distance of one another, and a question was not an event but a turn of the head. The offshoring wave that followed the downturn at the start of the decade changed that almost overnight. Cost pressure was acute, the arbitrage was real, and the delivery centres in India and elsewhere had the scale and the discipline to take on the work. What no one handed the leaders asked to run these newly-stretched teams was any account of what, precisely, they had lost by stretching them — or why the savings the business case promised kept arriving at only half their expected size.
The saving that half-arrived
The business cases were not wrong about the rates. A role that cost a certain amount onshore genuinely could be filled offshore for a fraction of it, and the headline arithmetic — forty per cent off the cost of delivery, sometimes more — was real on the day it was signed. What the arithmetic did not carry was a line for the thing that quietly consumed a large share of the gain: the sheer overhead of coordinating work across distance and time. By the time you had accounted for the extended clarification cycles, the rework born of misunderstanding, the duplicated documentation, the meetings that now had to be scheduled where once they simply happened, a substantial portion of the promised saving had been spent before it was banked.
The pattern that recurs across these programmes is not that offshoring failed — in the main it did not — but that it delivered perhaps half of what the model predicted, and that the missing half was almost never where anyone had looked for it. It was not in the rates. It was in a category the business case had no name for, because it had never before had to be paid for explicitly.
The invisible scaffolding
To understand where the missing half went, you have to notice something we had all been relying on for our entire careers without once being aware of it. Co-located teams are held together by a dense, continuous, almost entirely invisible flow of information that no one plans and no one budgets. You overhear the conversation at the next desk and correct your own assumption before it becomes an error. You read the room in a meeting and adjust. You catch a colleague’s frown and stop to explain. You resolve an ambiguity in the ten seconds it takes to lean over and ask, and neither of you ever records that it happened. This ambient awareness — the constant, unmanaged, peripheral sensing of what everyone else is doing and thinking — is the scaffolding on which co-located delivery stands.
Proximity supplies this scaffolding for nothing. It is so cheap and so automatic that we mistake it for the natural state of things rather than a service we are being given for free. And because it is invisible, it appears nowhere in any model of how a team works, which means that when we move the team apart and the scaffolding vanishes, the loss is felt everywhere and attributed nowhere.
The hardest costs to manage are the ones that were previously free. Proximity was a subsidy we did not know we were receiving, and offshoring was the invoice for it, arriving under a category none of our models had a column for.
A design decision that took a day
Let me make this concrete, because the abstraction understates how ordinary the failure is. On one distributed programme — a composite, but true to a dozen I have watched — a developer offshore reached a point in the build where the design was ambiguous. It was a small ambiguity, the kind that arises constantly: the specification could be read two ways, and only one was right. In a co-located team this is a non-event. You turn to the analyst two desks away, you ask, you get an answer in under a minute, and the whole thing is forgotten before lunch.
Across the distance, that same trivial ambiguity became a full cycle. The developer, reaching the question late in their day with the overlap window already closed, made a reasonable assumption and pressed on rather than lose the day waiting. The assumption was wrong. It surfaced onshore the following morning, by which point a day’s work had been built on it, and unwinding that work and rebuilding it correctly cost the better part of two more days. One ten-second question, unaskable at the moment it arose, had become a multi-day loss — and multiplied across a programme, this single mechanism accounted for a remarkable share of the vanished savings. The tragedy is that everyone behaved sensibly. The developer was right to keep working; the analyst was unreachable through no fault of theirs; the window genuinely was closed. The fault lay not in anyone’s judgement but in the removal of the scaffolding that had always made such judgements unnecessary.
The comfortable explanation, and why it misleads
There is a reassuring way to read all of this, and it is worth stating in its strongest form because a great many people believe it. The comfortable explanation says that these are teething troubles. The tools are immature — give it a few years and the conferencing, the messaging, the shared repositories will close the distance until it hardly matters. The processes are immature — tighten the specifications, raise the delivery maturity, document more rigorously, and the ambiguities that cause the rework will largely disappear. On this view the missing half of the saving is a transitional cost, an artefact of doing something new badly, and it will melt away as the discipline catches up.
There is real truth in it, and only a fool would claim that better tools and tighter process make no difference; they plainly do. But the comfortable explanation mistakes a permanent feature for a temporary one. Better tools can widen the channel through which information flows, but they cannot restore the ambient, unbidden quality of co-located awareness — the correction you receive without having asked for it, the misunderstanding caught before anyone knew it existed. And tighter specifications cannot eliminate ambiguity, because ambiguity is not a defect of specifications but a property of any work complex enough to be worth doing; you can reduce it, but the residue is irreducible, and across a distance the residue is expensive. The overhead is not the cost of doing distributed delivery badly. It is, in large part, the cost of doing it at all. Treat it as a phase to be outgrown and you will chronically under-resource the very leadership work that the distance makes essential.
What leadership turned out to be
Here is where the deeper force finally shows itself, and it reaches well beyond offshoring. What distributed delivery revealed is that a great deal of what we had been calling leadership was, in fact, proximity doing the work on our behalf. The co-located leader looked skilled at keeping a team aligned, catching problems early, maintaining trust and shared understanding — and some of that skill was real. But much of it was supplied, invisibly, by the simple fact of everyone being in the same place. Remove proximity and you find out how much of your leadership was yours and how much was the room’s.
The leaders who adapted well to the distance were the ones who understood this and responded by making deliberate what proximity had made automatic. They engineered the ambient awareness back into existence by explicit means: rituals that surfaced what people were assuming, not merely what they had done; a discipline of writing down the small clarifications that used to evaporate; a deliberate investment in the personal trust that co-located teams build for free over coffee and now had to be built on purpose, across a phone line, by leaders who understood that a team which had never met would not extend one another the benefit of the doubt unless someone worked to earn it. This is harder, slower and more conscious than co-located leadership ever had to be. It is also, I have come to think, a truer form of the craft, because it cannot lean on the scaffolding. It has to supply the structure itself.
“Remove proximity and you discover how much of your leadership was ever really yours — and how much of it was the room’s.”
The playbook we are still writing
We are, all of us running these stretched programmes, writing the playbook as we go, and it would be dishonest to pretend the pages are finished. But the deepest lesson is already legible, and it is not a lesson about time zones or tooling or offshore ratios. It is that the effectiveness we took for granted in co-located work was partly a gift of circumstance, and that when the circumstance is withdrawn, the gift has to be replaced by intention. The organisations chasing the arbitrage would do well to understand that they are not merely buying cheaper hours. They are taking on the obligation to build, by conscious effort and at real cost, the scaffolding that proximity used to supply for nothing — and that the half of the saving they keep losing is simply the portion of that obligation they have not yet learned to see, let alone to fund. The distance did not create the need for that scaffolding. It only, at last, made it visible.