The instrument that actually develops people is a decision they own, with a consequence they will feel, at a scale the organization can absorb if it goes wrong.
Everything else — the course, the reading, the shadowing, the conference — supplies context. None of it produces judgment, for reasons covered in Chapter 1.
Sizing it
Two conditions, and both have to hold.
Large enough that being wrong costs something the person will feel. If failure produces no consequence, the exercise teaches nothing, because the consequence is what writes the lesson into memory.
Small enough that failure does not damage the organization seriously or end the person’s standing. A development assignment that can destroy somebody’s career is not a development assignment; it is a test they were not told they were taking.
Most organizations fail the first condition rather than the second. People spend years on assignments too small to instruct, then receive one that is far too large, and the resulting failure gets read as a revelation about the individual rather than about the sequence.
Decisions, not work
The characteristic failure is handing over a task while retaining every decision it requires.
Somebody is given a project to run and finds that pricing goes to the owner, scope changes go to the owner, and staffing goes to the owner. They have been assigned work, not delegated responsibility, and they cannot learn anything from executing decisions somebody else made.
Worse, when it goes badly, the visible fact is that a person was given something and did not deliver. The approval queue appears in no account of what happened, because approvals feel like the leader helping rather than the leader constraining.
Before concluding somebody cannot handle responsibility, list the decisions the assignment required and mark which were genuinely theirs. If the marks are thin, nothing has yet been learned about the person.
What has to travel with it
A named boundary. Not general latitude. A figure or a category: you settle anything under this amount, above it we talk. Anything vaguer will be interpreted conservatively, because the cost of overstepping falls entirely on them.
The context you hold and they do not. Why this matters, which constraints are real and which are habit, what happened last time. Leaders under-transfer this constantly because it is unstructured and feels like background. It is the largest single input into whether their decisions resemble yours.
Visibility to everybody else. Tell the people they will need to work with that this is theirs. Delegation communicated only to the delegate is not delegation, because everybody else will keep routing to you and you will keep answering.
A stated intervention threshold. The conditions under which you will step in, named in advance. This protects them from arbitrary intervention and protects you from your own instinct to take it back when they do it differently.
The difficulty is watching
They will do it differently, and differently will look worse, because you can see the failure modes of their approach and cannot see the ones you have navigated around for years without noticing.
The test that separates a real concern from discomfort: if they were doing it this differently and it were going well, would you still want to step in? If yes, the discomfort is about the difference rather than the risk.
Some of these have to be allowed to go wrong. A senior person watching an avoidable loss occur, deliberately, having calculated that the loss is smaller than the cost of the person not learning, is doing the hardest and most valuable thing in this chapter. Very few can.
The debrief is where the learning attaches
Permitted failure without an honest conversation afterward is just loss. The conversation attaches the correct lesson to the case, and the wrong conversation attaches a wrong one that will be carried for twenty years.
The comforting version, in which the account is softened until the person concludes it was mostly bad luck, teaches that outcomes are not connected to decisions. The punitive version teaches that mistakes are dangerous and should be concealed, which removes the raw material permanently.
The useful version separates the decision from the result and examines the reasoning against what could have been known at the time. Sometimes the finding is that the decision was sound and the world did not cooperate, and that finding has to be genuinely available or the exercise becomes a search for fault.
Where to find them
Organizations frequently conclude they have no suitable assignments. They usually have several and have not recognised them.
A negotiation with a small supplier. A decision about a piece of equipment. Running one site or one shift entirely, including the calls that come with it. Preparing and presenting a recommendation to the board and answering the questions personally.
The last one is worth particular attention, because it is available in almost every organization and is routinely handled by the leader out of habit. Sending somebody else to answer the board’s questions, and staying quiet in the room, is among the most efficient development acts available and it costs nothing but discomfort.
Edited by Patrick J. Wolf, PhD