How to Delegate Without Ending Up Doing It Again

Delegation rarely fails during the work. It fails in the ninety seconds where the task changes hands.

Someone gets a vague instruction, a missing constraint, and no stated standard, then produces something that doesn't match the picture in your head. You fix it yourself, conclude that delegating didn't save time, and take the task back.

The work was never the problem.

Delegate outcomes, not keystrokes

There are two things you can hand over, and they produce very different people.

Hand over the steps and you get someone who can execute this task, under supervision, until anything unexpected happens. Then they come back to you.

Hand over the outcome and the constraints and you get someone who can handle this task and adapt when reality doesn't match the brief.

| Delegating steps | Delegating the outcome | |---|---| | "Pull the numbers from the dashboard, put them in the template, send it Friday." | "Leadership needs to see whether we're on track by Friday morning. Here's last month's version. Budget: two hours." | | Produces a report | Produces someone who owns the report |

The second version is harder to write and it's the whole game. It takes longer up front and stops costing you anything after the second cycle.

Name the decision level

Most delegation friction is a mismatch about authority that neither side ever stated out loud. You thought you'd handed over a decision; they thought they'd been handed a task.

State the level explicitly:

Say the number. "This is a level 4" takes two seconds and removes the most common source of both micromanagement and unwelcome surprises.

Most people delegate at level 1 and expect level 4 behavior. They hand over precise instructions, then feel let down when the person doesn't take initiative — having been given no authority to do so.

The handoff

Six things. Miss any of them and the work comes back wrong in a predictable way.

1. The outcome. What does done look like, described as a result rather than an activity.

2. Why it matters. Who uses this, and what happens next. Without this, someone can't make a single sensible judgment call — and every judgment call becomes a question for you.

3. The constraints. Budget, deadline, tools, what must not change, who must be consulted. Constraints are what let someone act without checking.

4. The decision level. Say the number.

5. A reference. The last good version, an example, or the SOP. "Like this, but for September" is worth a page of description.

6. The checkpoint. Agreed now, before work starts.

That last one matters more than it looks. A checkpoint you schedule at handoff is support. The identical question asked unannounced on Thursday is surveillance. Same words, opposite effect — the difference is entirely whether it was agreed in advance.

Accept the cost curve

The economics of delegation are counterintuitive, and misreading them is why people quit early.

"It's faster if I just do it" is accurate for the first three cycles and wrong forever after. People abandon delegation somewhere in attempt two, having paid the entire cost and collected none of the return.

This is also why you delegate recurring work first. A task you'll do twice more never repays the setup. A task you do weekly repays it within a month.

When it comes back wrong

Check the brief before you assess the person.

Work that comes back wrong almost always traces to one of these:

• The outcome was described as an activity, so they did the activity
• The standard was in your head, not the brief
• The decision level was never stated, so they guessed conservatively
• A constraint was missing, and they solved for the wrong thing
• The reference wasn't provided, so they invented a format

Each of those is a handoff defect. Fix the brief and re-hand it rather than taking it back — taking it back teaches both of you that delegation doesn't work here.

Then give the same task to the same person again. The second attempt on the same task with a corrected brief is where the actual learning happens, and skipping it wastes everything the first attempt cost.

Feedback that transfers

"This isn't quite right, let me fix it" transfers nothing. Two sentences do the work:

The analysis is right. For next time: leadership skims, so lead with the number and put the method underneath.

Specific, forward-looking, tied to why. Then let them redo it.

What not to delegate

• Anything you haven't done yourself and can't judge
• Decisions that are genuinely yours — direction, hiring, firing, pricing strategy
• Work where the method is the point, without stating that it is
• A process still changing weekly, unless changing it is the task
• Something urgent and unfamiliar to a person you haven't delegated to before

What to do next

List the tasks you did more than three times last month. Pick the one with the lowest consequence if it's imperfect. Write the six-part handoff, say the decision level out loud, agree the checkpoint — and then don't touch it until the checkpoint arrives.