Every owner has run the experiment: hand a task to someone, get it back wrong or late, redo it at midnight, and conclude the obvious. It's faster to do it myself. That conclusion is true, for that week. Compounded over years it becomes the most expensive true sentence in small business, because it caps the company at the size of one person's hours and keeps your best people permanently junior. The fix isn't summoning the courage to "let go". Failed delegation is almost never a courage problem or a people problem; it's a missing system, and systems can be built.
Delegating outcomes, not activities
The classic failure is handing over an activity while the standard stays in your head. You say "send the quotes"; you mean "quotes go out fast, follow our template, and anything unusually large or off-standard pricing comes past me first". The person executes the visible task, misses the invisible standard, and gets marked down against criteria they were never shown. Both parties leave the exchange more cautious than they arrived.
The repair is a written definition of done for anything you delegate: the outcome you want, the quality bar it has to meet, the deadline, and the boundaries around it. One short paragraph. It feels pedantic the first time and like infrastructure every time after, because the same paragraph trains the next person too, and the one after that. This is the smallest unit of the broader work of systemising a business: getting standards out of the founder's head and into a form other people can meet.
Making decision rights explicit
Most of the anxiety on both sides of a delegation traces to one unanswered question: what can I decide alone? Answer it out loud, in three buckets:
- Decide and don't tell me. Routine, reversible, inside the standards. The bulk of daily work should live here.
- Decide and tell me after. Judgement calls within agreed limits. A clear threshold, whether a spend level or a category of customer, works wonders because it replaces a feeling with a line.
- Ask before acting. The genuinely irreversible or sensitive: firing a supplier, committing to significant spend, anything touching a key relationship.
People given explicit decision rights act. People left to guess either escalate everything, which trains you to stay the bottleneck, or escalate nothing, which trains you to hover. Then comes the harder half: enforcing the buckets on yourself. When someone brings you a bucket-one decision, the discipline is to hand it back warmly, along the lines of "that one's yours, what do you think?", because every decision you make in their lane re-teaches them to bring you the next one.
Checkpoints instead of hovering
What owners describe as trust issues is usually the absence of any legitimate way to see progress, so they compensate by hovering, which everyone hates, or by abdicating, which fails later and more expensively. The alternative is a checkpoint agreed at handover: for a short task, a quick draft-before-send; for a longer one, a scheduled midpoint look at work in progress. The key is that the checkpoint is part of the delegation, named in advance, rather than an ambush. Inspection you asked permission to do at the start is support; inspection you spring on people is surveillance.
Match checkpoint intensity to track record, not to your mood. Someone new to a task gets closer checkpoints; someone who has nailed it repeatedly gets fewer, and knows why. Progressive trust, made visible, is one of the strongest development tools you own and costs nothing.
When it comes back wrong
It sometimes will, and what you do next determines whether the system compounds or collapses. Resist the midnight redo. Instead, run the returned work against the definition of done, together. If the gap traces to a standard you never wrote down, that's your fix: the paragraph gets a line added. If it traces to a standard that was written and missed, that's a coaching conversation, specific and unheated. Either way the system improves, which is the entire point. Silent rework improves nothing and guarantees a repeat.
What this buys you
Run this way, delegation stops being an act of hope and becomes a machine with parts you can inspect: a standard, a set of decision rights, a checkpoint, a feedback loop. Each handover makes the next one cheaper. Over time the business stops needing you in every decision, which is worth more than any single task you could have done faster yourself; it's the ground-level mechanics of reducing founder dependency. The worry you're handing over was never really about other people's competence. It was about the absence of a system that let competence show.
About the author
Andrew Northcott
Founder & Chairman, Valont
Andrew is the founder and chairman of Valont and the parent group Wattlestone. He has spent two decades building and running Australian SMEs, and writes about the realities of ownership — cash, people, systems, and the decisions that compound.
LinkedIn →