Delegation goes wrong in two opposite directions, and most people are fluent in both. Micromanaging hands over the task but keeps the judgment, so the other person becomes a pair of remote hands. Abdicating hands over the task and vanishes, so the other person guesses at what success means and often guesses wrong. The skill is the narrow path between.
Delegate the outcome, not the keystrokes
The unit you hand over should be a result, not a recipe. "Book the venue" invites a hundred check-in questions; "arrange a space for 40 people on the 12th, under $2K, decision by Friday" hands over a real problem with clear edges. Define the outcome, the constraints, and the deadline — then let the how belong to them. If you specify the how, you've kept the interesting part and delegated only the typing, and you'll be back doing it yourself within a week.
Say the quiet part: what "good" looks like and what can't move
The person taking the work needs three things you probably have in your head and haven't said: what a great result looks like, which constraints are hard versus nice-to-have, and where the landmines are. "Finance is sensitive about anything over $2K — go over and it needs sign-off" saves them a costly wrong turn. Front-load this and you replace ten interrupting questions with one clear brief.
The context that matters most is usually why the task exists. Someone who knows the venue is for a client dinner where the point is quiet conversation will make a hundred small decisions correctly that you never thought to specify. Someone who only knows "40 people, under £2K" will book a sports bar, and they will not have done anything wrong.
Name the decision rights explicitly
Most delegation friction is a disagreement about authority that neither party knew they were having. They think they should check before spending; you think you handed them a budget. Both are being reasonable and both are irritated.
Say it in one sentence. Something like: decide and go ahead, decide but tell me before you commit, or bring me options and I'll pick. Different tasks warrant different levels, and the same person can hold different levels for different work — which is fine, as long as it's stated rather than inferred.
Delegation grows by moving people up that ladder deliberately. Someone who has run three of these on "tell me before you commit" is ready for "decide and go ahead," and saying so out loud is most of what people mean by feeling trusted.
Make the handoff visible and owned
A delegated task that lives only in a hallway conversation will be remembered differently by each of you. Put it somewhere you both can see — a card on a shared action board with their name on it and the deadline attached — so the ownership is unambiguous and neither of you has to nag or wonder. Visible ownership is what lets you stop hovering: you can see it's moving without asking.
Check in on the schedule you agreed, not on your anxiety
Micromanagement is check-ins driven by your nerves; good oversight is check-ins driven by the work's rhythm. Agree up front on when you'll sync — a midpoint review, a check before anything irreversible — and then hold to that instead of pinging every time you feel a flicker of worry. The midpoint check is the crucial one: early enough to correct course, late enough that they've done real thinking you can respond to.
If the anxiety is genuinely unmanageable, that's information about the delegation rather than about them. Either the brief was too vague, the check-in cadence is too sparse for the stakes, or the task was too large a jump for where the person currently is. Fix the structure instead of resolving the feeling by interrupting them.
Let them do it worse than you would
The hardest part of delegating is tolerating a result that's 85% of what you'd have produced, done a way you wouldn't have chosen. If you claw it back every time it's imperfect, you'll train the person to stop trying and you'll never get the time back that delegation was supposed to buy. The 15% gap is the tuition. Coach it afterward, don't seize it midway — that's the difference between growing someone and renting their hands.
There is one legitimate exception, and it's worth naming so it isn't used as cover: intervene when the gap is not stylistic but consequential — a legal exposure, a customer relationship at risk, a decision that can't be undone. Even then, intervene by raising the specific problem, not by taking the work back. "This clause exposes us — can you rework it?" keeps the ownership where you put it.
Delegating badly is faster than delegating well, once
The reason people don't delegate is honest: the first time, explaining takes longer than doing. That's true, and it's also the entire point. You're paying an upfront cost for something that pays back on every subsequent occurrence — and the tasks worth delegating are precisely the recurring ones where that payback compounds.
The corollary is that delegating a genuine one-off, under time pressure, to someone who has never done it is often just a worse way to do it yourself. Delegate the things that will come round again.
Close the loop
The step that gets skipped: tell them how it landed. Not just "thanks" — what the client said, what happened as a result, what you'd do differently next time. Without that, the person has no way to calibrate, and the next attempt will contain the same gap.
That feedback is also what makes the next delegation cheaper. Somebody who has heard, specifically, why the second option would have been better is somebody you can hand more to next time.
