Almost every owner who says "I tried delegating and it didn't work" is describing the same event: a task was handed over with an outcome in mind that only existed in their head, the result came back wrong, and they took it back. The conclusion drawn — nobody can do this like I do — is comforting and usually false. The handover was incomplete, and an incomplete handover produces an unpredictable result no matter who receives it.
The reframe that fixes most of it: delegation is not giving someone a task. It is transferring a defined outcome, a set of constraints, a level of authority, and a review point — all four, explicitly. Leave any one of them implicit and you have not delegated; you have created a guessing game with your name on the result.
Why owners hold on
Worth naming honestly, because the reasons are rational and they are what you have to work against.
It is genuinely faster to do it yourself this once — and it is faster every single time, which is precisely how a bottleneck becomes permanent. The quality risk is real and lands on your reputation. Explaining the task properly costs more than doing it. And for a lot of founders, some of the work is identity: the part of the business you are actually good at is the hardest to hand over, even when it is no longer the highest-value use of your time.
The counterweight is arithmetic. Every hour you keep is an hour of capacity the business cannot buy more of, and the ceiling of the company becomes the ceiling of your calendar.
Separate decisions from tasks
The single most useful distinction. Tasks are things that get done. Decisions are judgements about how they get done.
Most people delegate tasks and unconsciously keep every decision, which produces the worst of both worlds: someone else is doing the work, but they have to come back to you constantly, so you carry all the interruption and none of the relief. That is not delegation, it is remote-controlling.
The alternative is to be explicit about which decisions move with the task. Some genuinely should not — anything that commits significant money, sets pricing, or changes a client relationship may reasonably stay with you. But most do, and saying so out loud is what turns a task into ownership.
The four things every delegation must specify
Write these down. Verbal handovers reliably lose two of the four.
1. The outcome. What "done and good" looks like, in observable terms. Not "handle the supplier onboarding" but "new supplier has signed terms, is set up in the system, and has had a first order placed without escalation." An outcome you cannot check is not an outcome.
2. The constraints. Budget, deadline, tools, tone, legal or safety requirements, what must not happen. Constraints are where "I'd never have done it that way" hides — if you have an unstated preference, it is a constraint and it belongs on the page.
3. The decision rights. Which calls they make alone, which they make and inform you about, and which they bring to you. Ambiguity here is what produces both over-checking and unwelcome surprises.
4. The check-in points. When you will look, and at what. Agreeing this in advance is what lets you stop hovering — you have a scheduled moment, so you do not need an unscheduled one.
Levels of authority
A useful shorthand when handing over decisions is to name the level explicitly, rather than leaving it to be inferred:
- Ask first. Bring me options; I decide.
- Recommend. Tell me what you would do and why; I confirm.
- Decide and inform. Act, then tell me what you did.
- Own it. Act; report on results, not on decisions.
People do not start at "own it" and they should not stay at "ask first" forever. Say which level applies today and what would move it up — competence demonstrated over a few cycles, usually. Naming the level removes the most common friction in delegation, which is two people holding different assumptions about who decides.
Document the standard once
If you find yourself correcting the same thing repeatedly, the problem is not the person, it is that the standard lives in your head. Writing it down once is cheaper than correcting it indefinitely, and it also makes the next handover nearly free.
The lightest version that works: what "good" looks like, the three or four ways it usually goes wrong, and the non-obvious constraints. This does not need to be a manual — a page beats nothing, and it beats an unwritten process that only you can audit. Our guide to standard operating procedures covers turning that into something durable.
The best time to write it is while someone is learning the task: have them write it, and you review. It is more accurate, because they document the parts that were actually unclear.
The handover sequence
Quality collapses when the transition is a single event. Stagger it:
- You do it, they watch. You narrate the judgement calls — that is the part that is invisible in a written process.
- You do it together. They drive on the parts they are ready for.
- They do it, you watch. Resist correcting mid-flight unless something is genuinely about to go wrong; note it and debrief after.
- They own it. You see the output at the agreed check-in, not the process.
The step people skip is the third, and it is the one that surfaces the misunderstandings while they are still cheap.
Reviewing work without taking it back
This is where most delegations quietly die. Work comes back at 80% of your standard, you fix it yourself, and everyone learns that the real reviewer is you. The next submission will be at 70%.
Instead: review against the outcome you specified, not against how you would have done it. If it meets the standard but differs from your approach, it passed — and this is the hardest discipline in the whole exercise. If it misses, hand it back with the specific gap named, and let them close it. Handing back feels harsher than fixing it silently; it is the opposite, because fixing it silently withholds the information they needed.
Watch for reverse delegation, too — the "quick question" that ends with the task back on your desk. The response is a question, not an answer: "what do you think we should do?" You will find you are answering things they already knew.
Signals you got the level wrong
Too little delegated: you are still the bottleneck, decisions queue behind your calendar, your team asks permission for things they clearly understand better than you, and you are working in the business rather than on it — the position our operations guide is largely about escaping.
Too much, too fast: repeated surprises, quality issues reaching customers, someone visibly out of their depth, or decisions being made with information they were never given. That is rarely a reason to take it all back; it usually means dropping one level of authority and adding a check-in for a while.
FAQ
What should I delegate first? The work that is frequent, rule-based, and low-risk to get slightly wrong. It builds the habit on both sides and buys you time to hand over the harder things properly. Save irreversible, high-stakes decisions for last.
How do I delegate when there's nobody to delegate to? Then the first delegation decision is a hiring or outsourcing one. Define the outcome you would hand over before you look for the person — it is also the clearest brief you can give, as covered in our guide to the first key hire.
What if they make a mistake that costs money? Set the authority level so the worst plausible outcome is survivable, then let it be survivable. Mistakes inside agreed limits are the tuition cost of building capacity; mistakes outside them mean the limits were wrong.
How often should I check in? Frequently at first, then less as the level of authority rises. Scheduled and predictable beats frequent and random — random checking reads as distrust and teaches people to wait for you.
Isn't it faster to just do it myself? Once, yes — always. That is exactly why the calculation has to be made across the year rather than across the afternoon.
Take the one task you complain about most, write down its outcome, constraints, decision rights, and check-ins, then hand it over properly. For more practical management and operations guidance, visit Consulting Firm USA.
