The 5 Decisions You Should Never Delegate in Your Business

What decisions should you never delegate?

The decisions you should never delegate are the ones that define the business rather than run it: who you say yes to, what the work is worth, what counts as good, who is allowed to decide what, and what you stop doing.

Everything attached to those five can be handed over, including the research, the drafting, the follow-up and the admin. The judgement at the centre stays with the owner.

Seniority has little to do with it. What decides is whether the answer sets the boundary that other people then work inside.

Who is this for?

This is for a founder whose offer already sells. You have a small team, or a few freelancers, and the work is being done by people other than you.

And yet the same questions keep coming back up to your desk. Can we take this client. Can we do it for that price. Is this good enough to send.

You delegated the work. The decisions came back anyway, one message at a time, usually on a Friday afternoon.

Key takeaways

• Delegation usually breaks at the judgement buried inside the task, which nobody wrote down.
• Five decisions belong to the owner: who you say yes to, price, the quality bar, who decides what, and what you stop.
• A decision you keep must be written as a rule, or it turns into a question someone has to ask you.
• Gallup measured a correlation between delegator talent and revenue among fast-growing US companies. It is a correlation, on a specific group.
• The fastest test is your last ten interruptions. Sort them into work and judgement.
• Keeping a decision does not mean doing the work under it. Those are two different things.

Table of contents

• Where delegation actually breaks
• The five decisions you should never delegate
• What the evidence actually says about delegating
• The task version and the decision version
• The decision a default made for one of my clients
• A ten-minute check you can run today
• Frequently asked questions
• Recommended tools and resources
• Where to explore next
• Final summary

Where delegation actually breaks

Most handovers describe an action. Send the proposal. Book the call. Reply to the enquiry.

Inside each sits a judgement nobody wrote down. Which proposal. Which call is worth an hour. Which enquiry gets a reply today and which one waits.

So the person does the action, hits the judgement, and stops. They come back to you, because that was the only rule they were given.

From your side it looks like the delegation failed. From their side they did exactly what they were asked, right up to the point where the instructions ran out.

This is why hiring a more experienced person often changes less than expected. Experience helps someone guess your boundary faster. It does not tell them where it is.

The decisions you should never delegate are worth naming for one practical reason. Once you know which five are yours, everything else becomes handover-able, and you can stop deciding case by case whether something is yours or theirs.

The five decisions you should never delegate

1. Who you say yes to

Which clients you take, and which you turn down. This sets the shape of your delivery, your margin and your reputation at the same time.

Someone else can screen, score and summarise. The yes is yours, because a bad yes is paid for over the following three months by everyone.

2. What the work is worth

Pricing carries your positioning. Change the number and you change who arrives, what they expect and what you are able to deliver.

Discounts are the same decision wearing a smaller coat. A team member offering ten percent to close something is repricing your business in a chat window.

3. What counts as good

The quality bar is the one thing a team cannot infer. Two competent people will hold two different bars, and both will believe theirs is the standard.

Write it down as examples rather than adjectives. Three pieces that passed, one that did not, and the reason. That is more useful than a page describing excellence.

4. Who is allowed to decide what

This is the decision that removes the other interruptions. It covers who holds the call when the situation is not standard, which is a different question from who does the work.

Most founders never make it explicitly. Authority gets assumed, then tested, then quietly returned to the founder the first time something goes wrong.

5. What you stop doing

Killing an offer, closing a channel, ending a client relationship, dropping a report nobody reads. Nobody below you will propose it, because stopping things looks like failure from every seat except yours.

A business that only adds gets heavier every quarter. This is the one people forget belongs in the list at all.

Read the five together and the pattern shows up. Each one sets a boundary other people then work inside. That test is more reliable than asking whether something feels important.

What the evidence actually says about delegating

In 2014 Gallup studied 143 Inc. 500 CEOs and scored them on what it called Delegator talent. Those scoring high averaged 8 million dollars of revenue in 2013, against 6 million for those scoring low, a difference of 33 percent.

The same group posted a three-year growth rate 112 percentage points above the lower-scoring CEOs, and created 21 jobs in three years against 17.

Now the limit, which matters more than the numbers. This is a correlation inside a specific population: American companies already on one of the fastest-growth lists in the world. Gallup measured a talent these founders had, not a practice they were taught.

It does not show that delegating more produces revenue. Fast growth forces delegation at least as plausibly as delegation produces fast growth. What it supports is narrower: among founders who grew quickly, the ones who handed work over well were the majority.

The task version and the decision version

Here is the same handover written twice, as an example rather than a template.

The task version. "Can you handle the inbound enquiries this week? Reply to everyone, book the ones who seem serious, and check with me if you're unsure."

Read the last four words. That is the whole job, and the person will use it three times a day, because "seem serious" is a judgement they have no way to make.

The decision version. "Reply to every enquiry within one working day. Book a call if they name a budget over X and a start date inside eight weeks. Decline politely below that, using the saved reply. Anything that mentions a competitor or a partnership comes to me, same day."

The second version hands over more work. It also keeps a decision explicitly with the owner, in one sentence, and everyone knows which sentence it is.

The two versions differ in one place. The first delegates an action and keeps the judgement unwritten. The second writes the judgement down as a rule, so the action can genuinely leave.

The decision a default made for one of my clients

One of my clients, an adviser running a local practice, was sending her prospects into a generic industry simulator. It looked like a sensible piece of software to use, and the enquiries were being captured.

They were landing in a shared dashboard. A competitor in the same region could see them.

She had never decided that. She had picked a piece of software, and the software had decided who owned her leads. That is decision number four, who is allowed to decide what, being answered by a vendor's default settings.

We rebuilt the simulator under her own brand, capturing directly into her own dashboard. The competitor lost access the same week.

The point is not the software. An unmade decision does not stay unmade. Something fills it, and that something is usually whatever was easiest to install.

The same boundary shows up in the other direction. Another client, a production company, was sending its team into a meeting to evaluate a service. I pushed back on the frame before anyone booked a room: go in looking for introductions to the larger contractors rather than as a buyer.

Four minutes, and the meeting was for something else. Nobody could have delegated that, because it was not a task. It was decision number one wearing a calendar invite.

A ten-minute check you can run today

Open your messages right now, whichever app your team actually uses, and scroll back through the last ten things someone brought you.

Put each one in one of two columns. Work, meaning they needed a pair of hands. Judgement, meaning they needed your answer before they could move.

Count the judgement column. Then ask a second question of each line: is this one of the five, or is it a decision that could have been a written rule?

Most founders find seven or eight in the judgement column, and most of those are not the five. They are boundary questions nobody wrote down, so they get re-asked every time the situation comes round.

Each one you turn into a sentence disappears from next month's list. It is rarely done because answering the question takes ninety seconds and writing the rule takes ten.

Frequently asked questions

What is the difference between delegating a task and delegating a decision?

A task is an action with a known outcome. A decision sets the boundary that determines which action is correct. You can hand over the first safely once the second is written down somewhere the person can read.

Should I delegate hiring decisions?

You can delegate almost all of it: sourcing, screening, scheduling, reference checks and first interviews. The final yes belongs with you while the team is small, because each person changes the standard for everyone who arrives after them.

Can I delegate pricing if I give someone a price list?

A price list is the rule, so yes. What stays with you is the authority to move off it. Agree in advance what discretion the person has, in a number, and everything inside that number is theirs.

How do I know if I am keeping too many decisions?

Count the interruptions rather than the hours. If people are waiting on you more than once a day for something that is not one of the five, you are holding decisions that should have been written as rules.

What if my team makes a decision I would have made differently?

That is the cost of the boundary being real, and usually cheaper than the interruption it replaces. Check whether the rule was wrong or the judgement was, then resist rewriting the rule after a single bad outcome.

Does this change when the business grows?

The five stay yours for longer than most advice suggests, and eventually move to named roles rather than disappearing. What changes first is the work attached to each one, which can leave well before the decision does.

Can AI take any of these decisions?

No, and it is genuinely useful around them. It can prepare the comparison, surface what happened last time, and draft the rule you are about to write. The judgement stays with a person, because the consequences do.

What should I write down first?

The decision that interrupted you most recently. Recency beats importance here, because it means the situation is live and you still remember what you actually decided and why.

How detailed should a decision rule be?

One or two sentences, with a number or a named condition in it. Rules that read like policy documents do not get used. Rules that fit in a message get quoted back to you, which is the sign they are working.

Is keeping these five decisions the same as micromanaging?

They are close to opposites. Micromanaging is holding the actions while the boundaries stay vague. Keeping the five means writing the boundaries clearly and then letting the actions go.

Recommended tools and resources

• Your last ten interruptions, sorted into work and judgement. The cheapest diagnostic available, and you already have the data.
• A one-page decision sheet per area, naming the owner, the rule, and the number or condition that marks the edge of their discretion.
• A written discretion limit for anyone client-facing, so discounts and exceptions stop being invented in chat windows.
• Gallup's study of 143 Inc. 500 CEOs, if you want the delegation research in its original framing, correlations and all.
• The Studio, if you would rather map the five against your real business with someone: one working session, €750, credited in full against L'Atelier if you go on.

Where to explore next

• How to Delegate Without Losing Control
• How to Document a Process Without Writing a 40-Page Manual
• 7 Signs Your Business Has Outgrown Its Systems
• How to Know What to Fix First in Your Business
• How to organise a business that still depends on you
• Hire vs Automate: How to Tell Which One Your Business Actually Needs

Final summary

The decisions you should never delegate are the five that set boundaries rather than fill them: who you say yes to, what the work is worth, what counts as good, who is allowed to decide what, and what you stop doing. Everything attached to them can leave your desk.

Keeping a decision is not the same as doing the work under it. The work leaves when the judgement gets written as a rule, in a sentence short enough that someone quotes it back to you.

An unmade decision never stays empty for long. A default, a vendor setting or whoever is closest fills it, and you find out months later, usually from the outside.

Which decisions are still coming back to your desk?

If the same three questions reach you every week and you cannot tell whether they are yours to answer or yours to write down, that is the conversation worth having before the next hire. Book a call.