Delegating without abandoning

Delegation is described as an obvious efficiency and is done badly for reasons that are understandable rather than careless.

Why people do not delegate

It is slower initially. Explaining takes longer than doing, for the first several instances.

The calculation is correct in the short term and wrong across any reasonable horizon, and the short term is what is visible when the pressure is on.

Quality anxiety. The work will be done differently and possibly worse, at least initially.

Identity. For someone promoted out of doing the work, the work was the basis of their competence and value. Giving it away is genuinely uncomfortable.

This is the least discussed and frequently the largest factor.

Reluctance to impose. A concern about workload that is sometimes accurate and frequently a rationalisation.

The other failure

Handing something over with no context and no support, and treating any subsequent problem as the other person's failure.

This is not delegation but disposal, and it produces predictable outcomes: work that misses unstated requirements, and someone who learns that taking things on is risky.

What a proper handover contains

The purpose, not just the task. What is this for, who uses it, what happens if it is wrong.

Without the purpose, the person cannot make the hundred small judgements the task actually requires, and will either guess or return with questions.

The constraints. What must be true, what cannot change, what has been tried and failed.

The decision rights. The part most often left implicit and the cause of most friction.

What can they decide alone? What should they consult on? What must come back?

Stating this explicitly prevents both the frustration of being second-guessed and the surprise of a decision you would not have made.

The standard. What good enough looks like, which is frequently lower than the delegator's private standard and should be stated rather than discovered.

The checkpoints. When you will look at it, agreed in advance rather than appearing unannounced.

Scheduled review is oversight; unscheduled review is checking up, and the two feel entirely different to the person doing the work.

The levels

A useful frame because delegation is not binary.

Do exactly this. Investigate and report back. Investigate and recommend. Decide and tell me. Decide and tell me if it goes wrong. Decide.

Being explicit about which level applies removes most of the ambiguity, and moving someone up the levels over time is what development actually consists of.

The common error is assuming a level rather than stating one, with each party assuming a different answer.

Accepting a different method

The discipline that makes delegation work.

Work delegated will be done differently. If the outcome meets the standard, the method is not your business.

Correcting method rather than outcome is the definition of micromanagement, and it teaches the person that they are executing rather than owning.

The exception is where the method matters for reasons they cannot see — safety, compliance, downstream consequences — in which case the constraint should have been stated at handover.

When it goes wrong

It will, particularly early.

The useful response is to establish what was missing from the handover rather than what was wrong with the execution.

Most delegation failures trace to unstated context: a constraint nobody mentioned, a stakeholder nobody named, a standard nobody articulated.

Which makes the fix a better handover next time rather than a conclusion about capability.

Taking the work back at the first difficulty is the most damaging response available. It confirms the fear, removes the learning, and guarantees that nothing is delegated successfully thereafter.

What to keep

Not everything should be delegated.

Work requiring your specific relationships or authority. Work where the consequences of getting it wrong are severe and the person is not ready. Difficult conversations about performance. And a small amount of the actual work, which keeps you connected to what the job involves.

That last one is contested and defensible: managers entirely detached from the work make worse decisions about it.

The measure

Whether delegation worked is not whether the task was completed, but whether it needs to be delegated again next time or has simply become theirs.

That transfer of ownership, rather than the completion of an instance, is the point.