How to Delegate When Nobody Does It as Well as You

You handed the thing over. It came back wrong in the small way that takes longer to explain than to fix, so you fixed it. Eleven minutes. Explaining it properly would have cost forty, and you had a call at two.

Six weeks on, you are still doing it, and the person you gave it to has quietly stopped asking for it.

The conclusion almost everyone draws here is that nobody does it as well as they do. That conclusion is usually accurate, which is why it is so hard to argue with. It is also completely useless, because it describes a permanent condition and hands you nothing to do on Monday.

The advice on offer at this point is always about letting go. Trust your people. Get over the perfectionism. Accept that done beats perfect. That advice has been in circulation for decades and the problem has not shifted an inch, which usually means the diagnosis is wrong rather than the reader.

Here is the diagnosis I would put in its place. You handed over the task and kept the judgement. The task is the visible half, the half that fits in a list of instructions. The judgement is the thing that tells you when the list is wrong, and it never went anywhere. So the work came back needing you, because you had arranged for it to need you.

The Part That Does Not Travel

Some work is hard to do. Other work is hard to judge. Those are different problems, and they come apart far more often than people expect.

Building a pitch deck is hard to do. Knowing which slide is the one that loses the room is hard to judge. Writing a quote is hard to do. Knowing which customer gets the extra ten percent, and which one will read that discount as an admission that the first price was invented, is hard to judge.

When you delegate, the hard-to-do half travels fine. Anyone reasonably capable can build the deck. The hard-to-judge half does not travel at all, because it has never been stated in words, including to yourself. It lives as a feeling you get when something is off, and a feeling is not a transferable asset.

The stakes here are not in dispute. Gallup assessed 143 chief executives from the Inc. 500 and found that the ones who scored high for delegating posted an average three-year growth rate of 1,751%, some 112 percentage points ahead of those who scored low. Nobody needs persuading that this matters.

What a finding like that cannot tell you is what the high scorers did differently, and the honest answer is that most of them could not tell you either. They had the habit without the method. Which is exactly why the advice that filters down to everyone else is written by people describing a feeling, and why it comes out as let go.

How to Delegate the Judgement, Not the Task

“Handle the refunds” is a task. It transfers nothing at all. The person now owns a decision they have no basis for making, so they will either guess, which produces the wrong outcome, or ask you, which produces you doing it anyway with extra steps in between.

A rule looks different. Anything under sixty pounds, refund it and log it. Over sixty, come to me. If they have been with us more than a year and you genuinely cannot tell, the tie goes to the customer. Three sentences, and they cover most of a year of decisions.

The tie-breaker clause does more work than the rule it sits under. Most of the friction in delegated work happens in the cases nobody anticipated, and a person who knows which way to fall when things are ambiguous will be right more often than a person holding a longer list.

Writing the rule takes twenty minutes, which is why it never gets written. Answering the question in the moment takes ninety seconds. In week one the ninety seconds is obviously the better trade, and it stays the better trade for roughly a month.

By month six you are being interrupted all day by questions whose answers are the same four rules, and because each interruption is small, nobody has noticed that it has quietly become the most expensive thing in your week. The interruption loop is a loan. The repayments are invisible and they do not stop on their own.

Write the rules at the tie-breakers. That is most of the trick. “Do X” fails because reality refuses to stay inside X. “When you see A, do B, and when you cannot tell, the tie goes to the customer” survives contact with a situation you did not predict.

Show Them a Version Being Killed

Almost everyone trains someone by demonstrating the task. You do it while they watch, then they do it while you watch, and everyone agrees that this is what teaching looks like.

It transfers the easy half. Watching a competent person do a thing well tells you what the finished article looks like and nothing whatsoever about the twenty versions considered and dropped on the way there.

The demonstration worth giving is the other one. Put two versions in front of them, choose the one you are going to reject, and say out loud why, in specific terms, including the part that is only a hunch. “This one is technically fine and it will make them feel managed rather than helped, and that costs us more than the hour we saved.” Do that eight or nine times and they start hearing your objection before they hand the work over, which is the entire point of the exercise.

This is also where formal training tends to run out of road. A curriculum can teach the task and struggles badly with the judgement, which is part of why so many business graduates end up in work that never required the degree. A credential certifies that someone can do the thing. It stays silent on whether they can tell a good version from a slightly worse one, and that second capability is the one you are actually trying to hire or build.

Decide How Much Worse You Will Accept

Here is the part nobody says from a stage. When you delegate properly, quality goes down for a while. The drop is real, and pretending otherwise is how the whole thing collapses in week three.

If you have never decided in advance how much worse is acceptable, you will take the work back at the first sign of the drop and you will call it standards. Everybody does. The decision gets made at the worst possible moment, by the version of you that is already annoyed.

So make it early and make it specific enough to argue with. For the next eight weeks I will accept work at roughly eighty percent of what I would produce. I will keep my hands off it unless it crosses one of these three lines. Here are the three lines. Written down, before anything has gone wrong.

Delegation without a tolerated quality drop is not delegation, it is supervision. Supervision is a respectable thing to do and there are weeks when it is the right call. It also costs you the same hours you were trying to recover, so at minimum be clear about which of the two you have chosen.

Give Away Outcomes, Because Steps Break

A person handed a list of steps is fine until reality stops matching the list, which happens in the first week. Then they have two options and both are bad. Stop and ask you, or improvise with nothing to improvise against.

It is the same reason most written procedures go unread. When we looked at which standard operating procedures actually get followed, the ones that survived contact with a real week named the outcome and the tie-breaker, not every step.

A person handed an outcome has a third option. “The customer comes off that call knowing exactly what happens next and by when” is something you can steer by once the script has run out. It says what to protect when a choice has to be made.

Steps are also how you end up owning the result while somebody else does the labour. If you specified every move, the plan is yours, and when it goes wrong it went wrong because your plan met a situation your plan did not cover. There is no useful correction to give the person who followed it.

Outcomes carry a quieter benefit that people miss. They make it possible to be surprised upwards. Steps cap the result at your own imagination of the result, and you will notice that this is a ceiling you built.

Start With Something You Are Merely Good At

The instinct is to delegate the work you hate and are bad at. The bookkeeping, the scheduling, the task that has sat at the bottom of the list for three weeks. It feels like the obvious first win.

It is close to the worst available starting point, because you cannot judge it. If you do not understand the work, you cannot tell a good job from a plausible-looking bad one, you cannot explain what went wrong, and you cannot write a rule, because you have never had one. What you have done there is outsource. That is a perfectly good transaction and it is a different transaction, one where you buy a result and take it on faith.

Start instead with something you are competent at and not precious about. You can still see the gap between the version they produced and the version you would have, which means you can name the gap, which means the correction arrives as a sentence rather than a mood. That is the only mechanism by which anybody gets better at anything.

The Reverse Test

There is a test that settles almost every case of this, and it takes about ten seconds.

The work comes back wrong. Before you open your mouth, name the rule they broke.

If you can name it, say it, then check whether it was written down anywhere they could have found it. If it was, you are having a genuine performance conversation and you should have it properly. If it existed only in your head, write it down now and treat the incident as the price of discovering a missing rule.

If you cannot name a rule at all, and what you have instead is a strong sense that this is simply not how you would have done it, then the rule did not exist. The gap is yours. They executed faithfully against an instruction that was incomplete, and the incompleteness is the only defect in the room.

Most people skip this test because it keeps finding them guilty. That is also the reason it works.

Run it for a month and something slightly uncomfortable happens. You start accumulating a written list of the rules you have been carrying around silently for years, and the list is shorter than you expected. Four or five sentences usually covers a job that felt impossible to describe.

“Nobody does it as well as I do” is phrased as a fact about other people, which is precisely what makes it feel like a dead end. Turn it around and it becomes a fact about you. Almost everything that makes you good at this has never left your head in a form another person could pick up and use.

Getting it out is slow, it is mostly writing, and it looks almost nothing like letting go.