The proximity trap: when a GCC leader can’t let go of the work

Watch a strong GCC leader 6 months after their reports doubled and you can usually see the strain before they can name it. They’re in more rooms than ever and adding less to each one. They still answer the technical question faster than the person who owns it. They still know the state of the tricky project better than the manager running it. On paper this looks like engagement. In practice the leader has kept the same working distance to the work while the work got 3 times wider, and something has to give. What gives is the part of the role only they can do.

This is a different problem from how the centre is shaped. The number of layers, whether the org is a pyramid or a diamond, how wide any one manager’s span should be, that is the org-design side of span of control, and it matters. But you can have the structure right and still lose the leader, because the trap here lives in the leader’s own head, in the gap between how much they now control and how close they’re still trying to stay to all of it. We call it the proximity trap, and it takes down people who are, by every other measure, very good at their jobs.

The tell is that nothing looks wrong. There’s no missed launch, no resignation, no complaint that reaches you cleanly. There’s a capable leader who is slightly slower on everything, whose calendar is full of work one level below where they should be spending it, and a centre that has quietly stopped getting better without anyone able to say when that began.

Every promotion widens two things at different speeds. Span of control, how many people and outcomes a leader is accountable for, jumps the day the title changes. Span of attention, how much a person can actually hold in their head with enough care to add value, doesn’t move at all. It’s roughly fixed. A leader who could stay genuinely close to 8 people’s work can’t stay equally close to 30’s, no matter how hard they try or how many hours they add. The arithmetic is unforgiving, and it doesn’t care how good the leader is.

So the leader does the only thing that feels responsible: they spread the same finite attention thinner across a bigger surface. They skim where they used to study. They keep a hand in every project, unwilling to admit there are now projects they won’t personally understand in detail. This feels like staying on top of things. A leader who is a little bit close to everything is far from the 2 or 3 decisions that genuinely require them, because those decisions never get the deep, undivided attention that used to be their whole contribution.

The move that actually scales is subtraction. The leader has to decide what to stop paying attention to, on purpose, and trust someone else to hold it. That’s the part good builders find almost physically difficult, because the things they most want to keep watching are precisely the things they used to be best at.

There’s a cruel logic to who this catches. The leaders most exposed are usually the ones who were exceptional at the earlier stage, the ones whose value was being close, in the code or the deal or the model, reachable by anyone with a hard problem. That closeness built the centre. It’s also the identity they earned their promotion on, and identity doesn’t update as fast as a title does.

So when the job asks them to step back, it reads as a demand to be worse at the thing they’re proudest of. Delegating the hard problem feels like abandoning it. Not knowing the detail of a project feels like negligence. A leader who solves the tricky problem themselves gets an immediate, legible reward, the problem is solved, and they solved it, while the leader who coaches someone else through it waits longer, watches a rougher first attempt, and gets no visible credit. Under pressure, almost everyone takes the fast, familiar dopamine of doing the work. It’s the predictable behaviour of a high performer whose sense of their own worth is wired to being the one who fixes it.

Which is why the proximity trap isn’t solved by telling someone to delegate more. They know they should. The instinct fighting them is the same instinct that made them worth promoting. Coaching out of it means giving them a new definition of a good day, one where their fingerprints are on their team’s decisions rather than on the deliverables, and that’s a slower, more personal shift than any process change. It’s closely related to the builder who outgrew the seat, and it responds to the same treatment: change what the leader believes their job is before you change what they do.

Because the trap is behavioural, you diagnose it by watching how decisions move, not by counting boxes. The clearest signal is that the leader has become the bottleneck. Things wait for them. Approvals sit for a day longer than the work needs, because there’s one of them and the queue is now long. When you map where a centre’s decisions actually pause, they pause on the person with the widest span and the least willingness to hand attention off.

The second signal lives a level down. Ask the managers who report to the reports whether they can get a decision without routing it up 2 layers. In a healthy centre they can, and they do, and they’re trusted with the consequences. In a centre caught in the proximity trap, skip-levels have quietly learned that any call they make will be re-examined by the leader, so they stop making calls. They escalate to save themselves the rework. The leader reads the flood of escalations as proof they’re needed, when it’s actually evidence they have trained the layer below them not to think.

The third is subtler and shows up in the leader’s own account of their week. Ask what they personally did, and the answer is full of work that belongs a level or two beneath them, reviewing a design they could have delegated, sitting in a working session they attended out of habit, unblocking something a manager should own. None of it is wrong on its own. All of it together is a leader spending a scarce resource on the wrong altitude, and the tell is that they can’t name the 2 or 3 things only they could have done that week, because those didn’t get done.

The fix people reach for first is structural: add a layer, split the span, hire a deputy. Sometimes that’s genuinely needed, and it’s where the org-design work earns its keep. But a leader in the proximity trap will re-create the trap inside the new structure, because they’ll manage the new deputy exactly as closely as they managed the ICs before. The behaviour follows the person. So the coaching has to work on the person.

The first move is to make the leader’s calendar honest. Have them tag a week of their actual time by altitude: which of this could only I do, which did I do out of habit or comfort. The gap is usually large enough to be its own intervention; leaders rarely realise how much of their week sits a level below them until they see it laid out. The second move is to force a small, deliberate handoff of something the leader is genuinely good at, not something they were happy to lose. The point is to practise the discomfort of watching a rougher version succeed, because that discomfort is the actual skill they need to build.

The third move is to fix the skip-level flow directly. Give the layer below explicit authority over a class of decisions, in writing, and then hold the leader to not re-opening them. This is harder on the leader than on the team, and it’s where a coach or a manager above earns their money, because the leader will reach back in reflexively and needs someone to notice and name it. Do this well and the change compounds: the layer below starts deciding, the leader gets attention back, and they can finally spend it on the few things that were waiting. Do it late, once the leader is fully saturated, and you’re usually reorganising in a panic and getting it wrong, which is its own quiet cost. The proximity trap is coachable, but it’s coachable most cheaply before the leader is drowning in it.