Part two of a series on project leadership — read part one first.
In part one, the argument was that divide and conquer — despite its track record and its logic — has a structural flaw: it optimises for individual visibility, not project success. People deliver their piece. The integration struggles. The last 20% lives permanently in the gaps between the parts.
So what's the alternative?
Not less structure. Not self-organisation without a centre. Not the kind of distributed ownership that sounds good in theory and produces excellent meeting attendance with unclear outcomes. Something more specific — and more demanding.
Direction isn't a goal statement or a project charter. It's a shared understanding of what success actually looks like — at the level that no single deliverable captures. Not "the portal is live" but "the organisation can operate without the systems we're replacing, and here's what that means for each team, each process, each dependency."
When people have that picture, they make different decisions. They flag problems at the boundaries, not just in their own area. They think about handoffs, not just outputs. They ask whether their piece is actually serving the whole — because they can see the whole clearly enough to check.
This is what changes when you share the full picture instead of dividing it: people stop optimising for their deliverable and start optimising for the outcome. Not because they're better people. Because they have enough information to aim at the right thing.
Sharing the picture isn't enough on its own. The next step is genuine ownership — not of a task, but of an outcome. The difference matters more than it sounds.
A delegated task has defined inputs, defined outputs, defined scope. The person doing it is accountable for execution, not for the result. If the task is the wrong task, that's someone else's problem.
Outcome ownership is different. The person understands what needs to be true at the end, and has the authority to decide how to get there. They can change the approach if the approach isn't working. They can raise dependencies before they become blockers. They can say "this won't achieve what we need" without waiting for someone above them to notice.
That's not the same as unlimited autonomy. It's autonomy within a direction — which is both more trusting and more demanding than dividing up the work and checking it off.
Here's where Direction over Division differs from pure distributed models like holacracy: strategic ownership doesn't get distributed. It stays at the core — not to control the parts, but to watch the connections between them.
This is the part that divide and conquer consistently gets wrong. The connections between areas — the dependencies, the overlaps, the places where one team's output becomes another team's input — are where programmes actually succeed or fail. They're also the places that nobody in a divided structure naturally owns. They fall between the defined areas. They become nobody's problem until they're everyone's problem.
A strategic core that holds the full picture can see these connections before they break. It can prioritise across areas when two things conflict. It can absorb the complexity that doesn't fit neatly into any individual ownership area — and make decisions about it without fragmenting the ownership of the pieces.
This is why the last 20% gets delivered. Not because the centre controls more — because it's watching the things that would otherwise fall through the gaps.
The pattern repeats across programmes and organisations. A project divided by area, by system, by team — each part delivered, integration handled separately, result: 80%, with the remainder living quietly in workarounds and systems that never quite got switched off.
The shift that changes the outcome is always the same: ownership structured around what the organisation actually needs, not around what's easy to divide. Full picture shared across everyone involved — not just their area, but the processes, the dependencies, the definition of done at the level that matters. And strategic ownership staying at the centre to hold the connections, set the priorities, and make the calls that cut across areas before they become the problems nobody owns.
The uncomfortable decisions — the ones that require seeing the full picture — only get made when someone is actually holding it. In a divided structure, those decisions either don't get made or arrive too late.
Direction over Division is harder than divide and conquer. Not technically — structurally. It requires someone willing to hold the full complexity without simplifying it prematurely. It requires trusting people with the full picture before you know what they'll do with it. It requires keeping strategic ownership even when distributing it would be easier to explain.
And it requires accepting that success in this model is harder to attribute individually — which makes it less appealing to people whose primary interest is in claiming it.
That's not a bug. That's the point. A framework where individual success and project success are the same thing is a framework worth building. Direction over Division is one way to get there.
Part two of a series on project leadership. Read part one: Divide and conquer optimises for the wrong thing.