I wrote a short post about servant leadership back in 2010 and I meant every word of it. I’d read something in Inc. that summed the whole idea up as identifying and meeting the needs of others rather than chasing power and status for yourself, and it landed hard, because I’d started out managing people the other way. I genuinely believed management was about control and direction and strength — the Alan Sugar school, more or less — and it took me a few uncomfortable years to work out that leadership is influence rather than control. Sixteen years later I still think that post was right.

What I’d add now is that the question sitting at the heart of servant leadership, “How can I help?”, is doing an enormous amount of work without anything much drawn around it. It’s a good question. It reverses the traditional relationship, in which everyone below you exists to make your life more manageable, and it tells the person in front of you that their success matters, that part of your job is making their work possible, and that they should bring you problems rather than hide them until they get too big to hide. All of that’s worth having.

The difficulty is what the question can be heard to promise. Asked without a boundary, it sounds like an offer to take over whatever has just been described, and people are very good at hearing offers.

Say one of your people tells you the supplier can’t hit Friday’s deadline and asks what you want to do about it. You say, “Leave it with me, I’ll give them a call.” That’s a kind and sensible thing to say, and it has just moved the job. The project still has their name on it and they’re still accountable for it at the next review, but the thing that has to happen next is now yours, which means they’ve stopped working on it and started waiting for you.

William Oncken and Donald Wass described this in the Harvard Business Review in 1974, in an article called “Management Time: Who’s Got the Monkey?” that has been reprinted ever since. Their monkey is the next move — the specific action required to push the problem forward one step. It starts on the employee’s back, and phrases like “let me look into it” or “I’ll find out what’s happening” are the moment it jumps across. What makes the idea useful half a century on is that it detaches delegation from the org chart — whatever the plan says about ownership, the person expected to act next is the one who really has the job.

The part that bothers me is that I described this failure back in 2010, and I confidently attributed it to the opposite style of management. My complaint about top-down control was that teams stop being able to think for themselves: they get very good at completing tasks and very poor at making decisions, either because they’re frightened of getting it wrong or because they’ve grown comfortable waiting to be told. I was right about the symptom and wrong to assume only one disease produced it. A manager who takes every awkward problem off your hands builds the same team as a manager who never let you near a decision in the first place. One does it through control and the other does it through kindness, and the people end up in the same position — capable of executing, out of practice at judging, and reliant on somebody else for the hard part.

Robert Greenleaf, who coined the term in a 1970 essay, set a test for this that we mostly don’t apply. He asked whether the people being served grow as a result: whether they become healthier, wiser, freer, more autonomous, more likely themselves to serve. Comfort never came into it: the test was whether people were becoming more capable of handling difficulty on their own. By that standard, declining to take something over isn’t a betrayal of servant leadership at all. It’s often the more faithful version of it. Servant leadership increases somebody’s capacity; subservient leadership substitutes for it.

None of which makes the instinct to step in a character flaw, and I want to be careful here, because the managers who do this most are rarely the lazy ones. Taking the work back is usually quicker than explaining it, coaching it and then reviewing it. It provides immediate relief to somebody who is struggling and immediate satisfaction to you for having relieved them. And it lets you avoid a conversation that would otherwise involve telling a decent person their work isn’t good enough yet, which is a conversation most of us will pay a surprising amount to dodge. There’s also a less flattering ingredient, which is that being the person who can sort anything out is a pleasant thing to be, and doing the work yourself can look thoroughly selfless while feeding a need to be indispensable.

The costs turn up slowly enough that nobody connects them to the cause. Your own week fills with other people’s immediate problems, so the strategic and developmental work — the work that actually required somebody at your level — keeps sliding to next month. Your team gets fewer of the difficult experiences that build judgement, because you keep intercepting them, and their confidence erodes in step with their practice. And the business starts to gather itself around whichever managers are most conscientious, until a worrying number of things can’t proceed without one particular person being available. That last one is the dangerous one, because it looks so much like commitment. A manager who solves everything appears indispensable, when what’s really happening is that the organisation is becoming dependent.

I’m not sure where this leaves somebody who is genuinely out of their depth. The gap between a useful struggle and a person quietly drowning is hard to read from the outside, and I’ve misjudged it in both directions.

What I’m confident about is that the question should stay and simply stop being the last thing said. It works far better when something follows it that establishes what kind of help is actually needed — what they’ve tried, what they think ought to happen, what’s genuinely blocking them, which part of this needs your authority rather than your effort, what decision they want from you. None of those withhold support. They define its shape, and most of the time what comes back is a plan that only needed permission or a bit of confidence to proceed. You can own the obstacle without owning the whole problem: if the supplier fix needs money that only you can approve, approve the money, and let the supplier relationship and the delivery plan stay exactly where they were.

The practical version of this is a habit worth building if you suspect your week is filling up with other people’s next moves. Before a problem-solving conversation ends, get four things straight:

  1. What is the next move? The one specific thing that has to happen next.
  2. Who owns it? One name. If you can’t answer this without hedging, it’s probably drifted to you.
  3. When will it happen? A date attached to the move, so the answer isn’t “soon”.
  4. What do they need from you? Authority, information, a decision, an introduction, cover — help that flows in without the job flowing out.

Run the supplier problem through that and the conversation changes shape without becoming any less supportive. Instead of offering to ring the supplier, you ask what alternatives they’ve already put to them. They suggest a split delivery and a backup for the remainder. You tell them they’ve got the authority to agree the split, and that if the backup takes it over budget you want the number before three so you can decide today. You’ve given them a decision, a boundary and your time. They’ve still got the monkey.

Two warnings, because an argument like this is easy to over-apply. There are situations where taking over is the right call and not a failure of nerve — a genuine emergency, anything touching legality or safety, a decision that was always reserved to you, work that was wrongly delegated in the first place, a structural problem the person has no realistic power to solve, or somebody who has been handed responsibility without ever being trained for it. What separates a responsible intervention from a habit is whether it’s deliberate, proportionate, explained, and followed by handing ownership back. Stepping in once is management. Becoming the default owner is the problem.

The second warning matters more, because “don’t reverse-delegate” must never become the house translation of “don’t bring me problems”. People have to stay able to raise risks, mistakes and obstacles early without calculating what it will cost them, and a manager who greets every escalation with a lecture about ownership will simply stop hearing about things until they’re expensive. The aim is to stop the act of coming to you from automatically transferring the job; you want people coming to you just as often as before. I’ve argued before that owners are often right to stay close to the work rather than delegating themselves into irrelevance, and I still think that — but you can stay close to the work without taking possession of everyone else’s next move, and the two get confused constantly.

Sixteen years on, I’d keep the 2010 post more or less as it was. What I’d change is the ending. Enabling and supporting people really does give you more reach than controlling them, but only while the enabling is making them more capable of acting without you. The moment your help starts arriving as relief rather than as capability, you’ve stopped serving your team and started serving their workload, and the question that got you there was a good one asked one step too far.