How to Delegate as a Founder Without It Coming Back to You

You gave vendor negotiation to your purchase head in March. In August you spent forty minutes on the phone yourself, arguing over a ₹3 lakh order, because he had forwarded the comparison sheet "just for your confirmation, sir" and you did not like the second line of it.

Nothing dramatic happened between March and August. He never refused the work. You never took it back in a meeting, never said anything that would look like a reversal to anyone watching. The task simply walked home.

What you performed in March was not delegation. It was the handover of work — and work is one of four things that have to move, and the least important of the four. Move the work alone and the task returns to you inside a month, having consumed his hours and yours on the way, and having taught everybody in the room something worse than the cost of the task itself: that delegation does not work here.

That is the expensive part. Every failed handover raises the price of the next one, because your team now has evidence: they watched a serious attempt to give something away, and they watched it come back. The next time you say someone owns something, they will wait before believing you.

This is the delegation half of a larger structure. Ownership needs five conditions present at once, and a handover that skips three of them cannot produce it; the full argument is in employees not taking ownership.

The playbooks are answering the wrong question

Almost everything written on how to delegate as a founder is about selection. Rank your tasks by rupee value per hour. List what only you can do. Sort into the four boxes. Start with something low-risk and build up.

None of that is wrong. All of it is beside the point. You already knew in March that vendor negotiation should not be on your desk — you did not need a matrix to tell you. Selection was never where this fails.

It fails in what does not travel with the task.

The Four Handovers

Four separate things have to move for a task to actually leave you. Most founders move the first, assume the rest are implied, and are genuinely surprised when the work comes back. It comes back because three of the four are still sitting on your desk, and everyone can see that they are.

The work

The task itself. The calls, the sheet, the follow-ups, the vendor, the file.

This is the easy handover and the only one most founders perform. It looks most like delegation from the outside, which is why it survives. You explained the task, answered his questions, told him to come to you if he was stuck. In the room, that felt complete.

But this handover moves your calendar, not your load. The forty minutes on the phone in August was not the work coming back. It was everything you never handed over becoming visible at once.

The standard

The standard is what "done well" means, in a form the person can check himself, without asking you.

That last clause is the whole thing. If he has to bring it to you to find out whether it is good, you have not delegated the task. You have queued it. The work now takes longer than it did when you were doing it, involves two people instead of one, and still ends with your judgement — which is why founders so often conclude, honestly and wrongly, that it is faster to do it themselves.

Gallup's 2023 study of 18,665 US employees found that only 47% strongly agree they know what is expected of them at work, down from 56% before the pandemic. Less than half, in companies with far more written process than yours. In India the same failure shows up on the outcome side: Gallup's Indian performance-management work found that among employees clear about the criteria on which they were evaluated, 66% were engaged, and among those who were not clear, 1% were.

In a promoter-led business the standard has a specific shape. It exists on paper, in the KRA sheet, and the KRA sheet was signed. But everybody knows the operative standard is whatever you say when you see it, and that this is not predictable in advance. So the rational purchase head stops trying to hit the written standard and starts trying to predict you — which requires showing you things, which is exactly the behaviour you are complaining about. He is not being timid. He is doing the job as it actually exists.

A handed-over standard sounds concrete and slightly boring. Three quotes, lowest technically-compliant one wins, deviation only for delivery risk, and if a vendor's price moved more than 8% since last order, note why. He can now score his own work before it reaches you. Until he can, nothing above this handover holds.

If your written standards and your real standards have drifted apart, that gap is the mechanism examined in why your team waits to be told.

The decision

The decision handover is the specific set of calls he can now make alone, with a stated boundary.

Stated. Written, spoken in front of others, and numeric where a number is possible. A rupee figure. A customer tier. A named class of exception. "Use your judgement" is not a boundary. "You have full authority on this" is not a boundary either, and everybody who has worked for a founder knows it — it means full authority until something goes wrong, at which point the boundary turns out to have existed all along, retroactively, at whatever level makes the founder's discomfort correct.

Unstated boundaries are not boundaries. They are a test the person can only fail.

The real version is dull to write and immediately usable: he approves any order up to ₹5 lakh alone; ₹5 to ₹15 lakh he decides and informs you the same week; above ₹15 lakh, or any change of vendor on the top four SKUs, comes to you first. Now he knows the shape of his own authority, and so does everyone who reports to him.

Founders resist the number because a number can be wrong. It can. A wrong number is a far smaller problem than no number, because a wrong number gets corrected once, in the open, while no number gets tested privately by every person in the company forever. Where to write these down, and what changes when authority stops depending on your mood that afternoon, is set out in the decision bottleneck.

The consequence

The outcome now belongs to him. Visibly, in front of other people, in both directions.

This is the handover founders keep, and they keep it because keeping it feels like decency. When the vendor deal goes badly you absorb it — you smooth it over, you do not name him. When it goes well, you say the team did well. Both feel like protection. Both quietly return the ownership to you.

Consequence has to run in both directions to mean anything. Most Indian firms have a downside that is fast and personal and an upside that is slow and shared: a bad quarter is discussed loudly and specifically, a good quarter dissolves into the general prosperity of the business. Against that asymmetry the intelligent employee minimises variance. He never fails badly and never particularly succeeds, which is the behaviour you keep reading as a lack of ambition.

The visible version is not complicated. In the monthly review, the vendor numbers are presented by him, not by you, and questions about them are answered by him while you stay quiet. When the saving comes through, it is named as his in the same meeting where a miss would have been named as his. The mechanics of attaching outcomes to people without turning it into fear are in the accountability gap.

The Return Test

Here is a diagnostic you can run this week, without a workshop.

Take any task you believe you have delegated. Count how many times in the last month it came back to you — as a question, as an approval, as a review, as a forwarded mail marked FYI that clearly wants a reply. Include the corridor conversations. Include WhatsApp.

More than twice in a month and the task is not delegated. Which handover is missing is usually obvious from the shape of the returns:

• Questions about how, or "is this okay" — the standard did not move. He cannot check his own work.
• Approvals and sign-offs — the decision did not move. There is no boundary, or the boundary is too low to be useful.
• Nothing comes back at all, and the results are mediocre — this one is not a success. It usually means the consequence did not move, so nobody is watching the outcome closely, including him.

Run this across the six or seven things you think you have given away. Founders who do it honestly typically find that two are genuinely delegated and the rest are in a queue with their name on it. That count is the real measure of how much of your week belongs to you.

If the count is high across every task and every person, the problem is not this delegation. It is the pattern, and the pattern has its own diagnosis in why founders micromanage.

What you should not delegate

Most advice on this subject is a list of things to offload. The more useful list is short and runs the other way, because founders who finally decide to let go tend to let go of exactly the wrong three things.

• The definition of the standard. You can delegate every piece of the work and none of the definition of what good is. A founder who hands over the standard has not delegated, he has abdicated, and the business will drift to whatever standard the most senior available person happens to hold. Define it, write it, then let other people meet it without you in the room.
• The first hire into any new function. The first finance head, the first person running exports, the first quality manager. That hire sets the standard for the function for years, and the person who hires the second one will hire to match the first. You can use a consultant, a search firm, three rounds of panel interviews. The call is yours.
• The relationships that are actually yours. Some of them genuinely are. A banker who extended a limit on your word in 2016, the customer whose father dealt with your father. These are not transferable by announcement, and pretending otherwise damages both the relationship and the person you hand it to.

Everything else you are holding, you are holding out of habit. Pricing approvals you set at ₹50,000 when the company was twelve people. The final read on every proposal. Junior interviews. Site visits you enjoy. Habit is a human reason to keep something, but it should be named as habit and not defended as necessity — the two require completely different conversations.

The relationship objection

Every promoter reaches the same sentence at this point. The market here is relationship-driven. They only deal with me.

You are usually right about the top ten accounts and wrong about the other ninety. Test it rather than assuming it: list your accounts by revenue and mark the ones where a specific person has explicitly asked for you, in the last year, when someone else was available. That list is almost always shorter than the belief.

The other ninety deal with you because you answer, not because they require you. And the ten that genuinely are yours are not an argument against delegation — they are the argument for handing over the ninety properly, so the ten get the attention they deserve. If the volume of the ninety is now beyond what any handover can absorb, the question has quietly become a structural one, and it is treated in when to hire a COO.

Where to start

Take one task you already believe you delegated. Do not pick a new one — the point is to repair a handover, not to start another.

Run the Return Test on it, find the missing handover, and complete just that one. Write the standard so he can score himself, or state the rupee boundary in front of his team, or let him present the number in the next review while you say nothing. One task, one handover, one month. If the return count drops, you have found the mechanism, and you can apply it to the rest without anybody running a programme.

If you would rather see where every decision in the business currently lands before you start, the Business Pulse diagnostic maps where decisions are actually made against where your org chart says they are. For most promoters, the distance between those two maps is the honest answer to how much has really been delegated.