Founders are told to delegate, and most try. The attempt frequently reverts within weeks, and the reason has more to do with how the judgement was formed than with any reluctance to share control.
The decision being delegated is compressed experience
A founder deciding which customer to accept or which supplier to trust is drawing on hundreds of small observations accumulated over years.
What gets handed over is the decision, not the observations, so the person receiving it applies general reasoning to a situation that rewards specific memory.
Their answer is defensible and frequently wrong in a way that is hard to explain, which the founder experiences as confirmation that delegation does not work here.
Reversal teaches faster than instruction
Overturning a delegated decision, even once and for good reason, tells the team that the authority was provisional.
Afterwards, people bring the question back rather than deciding, because checking is cheap and being reversed publicly is not.
The founder then observes that the team keeps asking, concludes they are not ready, and the pattern reinforces itself in both directions.
Delegating outcomes is different from delegating tasks
Handing over tasks keeps the founder as the point where everything is assembled and judged, which means the load does not actually reduce.
Handing over an outcome means someone else owns the result and the decisions inside it, including the ones the founder would have made differently.
The second is what creates capacity, and it requires accepting a range of acceptable answers rather than one correct one.
The cost of a mistake determines what can move
Decisions that are reversible and small are where delegation should start, because the organisation learns from the errors without paying much for them.
Decisions that are difficult to unwind, such as major commitments or anything affecting the company's obligations, reasonably stay with the founder for longer.
Separating the two explicitly is more productive than treating delegation as a single decision about how much to trust people.
The bottleneck is visible before it is admitted
Companies where every decision routes through one person develop a recognisable pattern: work waits for approvals, and progress tracks that person's availability.
The symptom is usually attributed to team capability, since the founder sees people who will not decide rather than a structure that trained them not to.
Founders who resolve it generally start by writing down which decisions no longer require them, because the constraint is more easily changed on paper than in habit.