How to delegate decisions, not tasks
Most managers know how to delegate work, and fewer know how to delegate decisions.
This is the case when the difference matters.
When I delegate a task, I can still decide how it should be done. I can define the approach, review the steps and approve the result.
When I delegate a decision, I transfer something more important: a right to decide or choose. That requires more trust, better boundaries and a different understanding of control.
Start with context, not instructions
A person cannot make good decisions if they only know what to do, where the result is - they need to understand why the work exists.
Before delegating a decision, I try to provide four things:
Context. What is happening around this problem? Why does it matter?
Goal. What outcome are we trying to achieve?
Constraints. What boundaries must not be crossed?
Authority and responsibility. What can the person decide independently, and what result do they own?
If I explain only the task, I should not be surprised when the person follows the task literally.
However, if I explain the context and the goal, they can adapt when circumstances change. That is where the real ownership begins.
Boundaries matter more than instructions
Delegation does not mean removing all boundaries, but it means making the boundaries clear enough that the person can operate independently inside them.
The manager still has to define the system, which may include deadlines, security requirements, regulatory constraints, architectural principles or decisions that require escalation. The important part is that these rules should be known before the decision is made.
If a manager repeatedly changes the boundaries after seeing the result, people quickly understand that their authority was not real.
Trust should grow with the cost of mistakes
The manager does not need complete trust before you start delegating - it can be built progressively.
With someone new, I would start with decisions where the cost of being wrong is limited, when I would see how the person thinks, because as their confidence grows, the size and complexity of the decisions can grow as well. My point is not to eliminate mistakes, it is to keep the cost of learning (learning curve) acceptable.
Let people make decisions differently
This is one of the hardest parts.
When a person makes a decision, I may look at it and immediately think that I would not do it this way.
The temptation is to intervene. Sometimes intervention is necessary, but "I would choose differently" is not enough.
Once the decision has been delegated, the employee has the right to use their own judgment, otherwise I have delegated only the execution.
A right test questionnaire is:
Does the decision satisfy the goal?
Does it stay inside the agreed constraints (which were already sounded)?
Is the implied risk acceptable?
If the answer is yes to all of the above, the fact that I would have chosen another path is irrelevant. And if there is “something else” which is not on the table, that means I was not honest to myself when I delegated the task and did not provide all the constraints or concerns.
The difference between my preferred solution and the employee's solution is part of the cost of building an independent team. Without paying that cost, the manager stays the hidden decision-maker.
Intervene because of risk, not preference
When should a manager step in?
Risk is the main boundary. If the employee moves outside the agreed constraints, the situation changes, as that sort of decision may create unacceptable financial exposure or difficult-to-reverse consequences - this is a reason to intervene.
But when? This distinction is essential.
I see two consequences of intervention behaviour:
If managers intervene whenever they disagree with the method, employees learn how to optimise for approval.
If managers intervene only when boundaries or unacceptable risks are involved, employees learn how to exercise judgment.
Control and participation are different things
Delegating a decision does not mean losing control, it means changing the mechanism of control.
There are at least three ways to control work:
1. Control through actions: "Show me what you are going to do before you do it" - which usually works with junior team members.
2. Control through information: "I can see what is happening and what result we are getting”.
Control through the system: "I know who makes the decisions, what principles they use, what the boundaries are, and how we will detect a problem."
As the organisation grows, effective managers move increasingly toward the third model, however the chosen approach still depends on who is in front of them, but, paradoxically, to retain control of a growing system, the manager has to give up control of every individual action.
The leader's job moves upward
If the team can operate without constant involvement from its leader, the leader does not become unnecessary in an absolute sense.
They can spend less time making decisions the team is already capable of making and more time on problems the team cannot yet solve.
This creates a useful way to think about leadership growth:
The growth of a leader can be measured by the complexity of the problems the team has learned to solve without them.
A leader who remains essential to every decision may feel important, but the system cannot scale.
A leader who continuously transfers decision-making capacity to others increases the capacity of the entire organisation.
That is the real purpose of delegation.