Handing over an outcome with the authority to decide how. Handing over tasks while keeping every decision is just queueing.
12 min read · Level 3
Handing over tasks while keeping every decision is not delegation, it is queueing with extra communication — and it caps the company at your own throughput. The harder truth is that real delegation means someone doing it differently from you and it working. If only your method is acceptable, you have not delegated an outcome; you have hired an assistant and limited the result to your own ideas.
Delegation is transferring responsibility for an outcome, along with the authority to make the decisions that outcome requires.
It is not assigning tasks. Assigning tasks while retaining every decision leaves you as the bottleneck with extra communication overhead.
Because the founder becoming the bottleneck is the most reliable way a growing company stalls. Every decision routing through one person caps the company at that person's throughput.
And because people who are given tasks but no authority disengage. The good ones leave; the ones who stay stop thinking.
Not delegation: 'Can you write the onboarding emails? Send them to me before they go out. Use the template I made. Check the wording with me first.'
The founder still makes every decision. They have added a person and a review cycle and removed none of their own work.
Delegation: 'You own activation this quarter. It is 8% and I would like it above 25%. Onboarding emails are probably part of it, but that is your call. Budget is $2,000. Check in with me weekly, and tell me if you need a decision only I can make.'
The difference: an outcome (activation above 25%), authority (it is your call), a constraint (budget), and a rhythm (weekly check-in).
Now the founder is not in the loop on wording. And crucially, the person can do something better than the founder would have — which is impossible under the first version.
First-time founders often delegate only the tasks they dislike, keeping everything interesting and everything important. That produces a team of assistants rather than a team of owners.
The second mistake: delegating without a check-in rhythm, then being surprised a month later. Delegation is not abdication — it needs a scheduled conversation, which is precisely what makes letting go safe.
The third: taking it back at the first mistake. If you reclaim ownership the first time something goes imperfectly, nobody will take real ownership again.
'Activation above 25%' rather than 'write these emails'. The outcome is what makes ownership possible.
What can they decide alone? What needs you? Ambiguity here means everything comes back to you by default.
Budget, timeline, anything that must not change. Constraints enable autonomy rather than limiting it.
Weekly or fortnightly. This is what makes delegation safe, and it is the step most often skipped.
A different approach that works is a success. If only your method is acceptable, you have not delegated.
Coach through it. Reclaiming ownership once teaches everyone that ownership is not real.
From your first hire, and increasingly as the company grows. By around twenty people, most of your job is this.
Do not delegate something you have never done yourself and cannot evaluate. And do not delegate the things only a founder can do — direction, key hires, culture.
The difference between "write these onboarding emails and check them with me" and "you own activation this quarter, it is 8% and I want it above 25%, budget two thousand, check in weekly" is not politeness. The second removes the founder from the loop and permits a better answer than they would have produced.
Delegate a measurable outcome with stated authority, a constraint and a check-in date. Missing any of the four and it reverts to task assignment.
Reclaiming ownership the first time something goes imperfectly teaches everyone that ownership is provisional. After that, people stop taking real responsibility — not out of resentment, but because they have correctly understood the arrangement.
Coach through the first error rather than reversing it. What you do the first time something goes wrong sets the rule.
Create a free account to read the rest and apply it to your own startup.