The Delegation Ladder: A 5-Level System to Stop Being the Bottleneck

A COO I coached in Hamburg last year had 17 direct reports and was working 75 hours a week. He was convinced that the problem was his team — they weren't as senior as they needed to be, they couldn't be trusted with the harder decisions, and he was the only one with the context to make the calls that mattered. After six weeks of observation, the diagnosis was the opposite. His team was capable. He was a bottleneck by reflex, not by necessity. He was delegating at a single level — roughly, "do this thing and bring it back to me" — on decisions that should have been entirely theirs, and hoarding the decision authority that would have made them grow.

This is the most common senior-leader failure mode in mid-sized companies, and it's almost always about delegation levels rather than delegation volume. The leaders who scale past 50 people are the ones who understand that delegation isn't binary — it's a ladder with five distinct rungs, and different decisions belong on different rungs. Push the wrong decisions onto the wrong rungs, and you either suffocate your team or expose yourself to risk you can't afford.

The model I'll describe is a modified version of Michael Hyatt's delegation ladder, stress-tested against the operating problems I've actually seen in 15 years of working with senior leaders.

The Five Levels

Level 1: "Research it and bring me options"

The person does the legwork, gathers the information, and presents options. You make the decision. Very little judgement is being exercised by them — they're a researcher for you.

When to use: high-stakes, one-off decisions where you don't have the bandwidth to do the analysis yourself but the call really does need to be yours. Example: evaluating three possible acquisition targets, where the decision will shape the company for the next five years. The analysis is delegable; the decision isn't.

When it's wrong: routine operational decisions, things your team has seen twenty times before. If you're still at Level 1 on recurring decisions two years into a role, you're the bottleneck and you're telling your team they're not trusted.

Level 2: "Recommend a specific option and explain your reasoning"

One step up. The person does the research and picks a recommended path, with a reasoning summary. You still approve, but they've done the cognitive work. This forces them to actually form a view — which is the skill most junior managers lack — and lets you calibrate whether their judgement is ready for the next rung.

When to use: material decisions where you want the person to practise forming a view, but you still need final say because the cost of being wrong is high. Most executive hiring decisions, for example, should be at Level 2 with your direct reports — they recommend, you approve.

When it's wrong: if you find yourself accepting their recommendations 95% of the time, you're adding friction without adding value. The decision should be moved to Level 3 or higher.

Level 3: "Decide, and tell me before you act"

The person makes the decision and commits to it — but tells you before execution, so you can catch a mistake. This is the middle rung of the ladder and probably where about 40% of senior-to-direct-report decisions should sit in a mature organisation.

When to use: decisions the person is qualified to make but where the downside of being wrong is large enough to justify a second set of eyes before action. Hiring a senior individual contributor. Approving a meaningful change in pricing or commission. Signing a non-standard customer contract.

The discipline, on your side: if they tell you before acting and you veto, you need a specific, defensible reason. "I would have decided differently" is not a defensible reason — it's a retreat to Level 1. The veto at Level 3 should be for genuinely catastrophic errors, not stylistic differences.

Level 4: "Decide, act, and tell me after"

The person decides, executes, and reports back after the fact. You have no operational role in the decision. Your role is to absorb the information, learn about what's happening in the business, and adjust your model of their judgement.

When to use: decisions inside their domain of expertise, where a mistake costs money but is recoverable. A marketing manager picking a creative agency. A head of engineering choosing between two technical architectures for an internal tool. Operational hires at their own level or below.

When it's wrong: if you're still being told after the fact about genuinely routine decisions that have no material impact on the broader business, you're still in the loop unnecessarily and you're generating noise.

Level 5: "Decide, act — I don't need to know"

The top rung. Full autonomy. You don't hear about the decision unless something goes catastrophically wrong. This is where genuinely senior leaders should be operating on 70-80% of their day-to-day decisions.

When to use: decisions that are entirely within the person's ownership, low enough stakes that a mistake is absorbable, and — critically — decisions that are recurring enough that you've already calibrated their judgement through observation at lower rungs. You don't skip straight to Level 5 with someone new. You earn your way up.

When it's wrong: never at Level 5 on decisions that could meaningfully harm the company, that break regulatory constraints, or that touch areas where you have specific context they don't.

The Rules That Actually Make the Ladder Work

Rule 1: Be explicit about the level

The single biggest failure mode is assumed levels. You think you've delegated at Level 4; they think you're at Level 2. They bring you every decision; you wonder why they're not autonomous. They were doing what they thought you wanted.

The fix: when delegating, state the level. "On this one, I want you at Level 3 — decide, but tell me before you act." Clumsy the first time, natural by the tenth. Most of my coaching clients put the level into 1:1 agendas explicitly until it becomes habit.

Rule 2: The level should rise over time

If someone is at Level 2 on a category of decision in month three, they should be at Level 4 in month twelve, assuming performance is good. The ladder is a growth vector, not a steady state. Static levels signal either that you're not trusting growth or that they're not growing — both are problems to diagnose.

A useful ritual: every quarter, on each direct report, review what categories of decision they're making at what levels and whether any should move up. Most senior leaders don't do this actively. The ones who do, scale.

Rule 3: Don't mix levels on the same decision

A common pathology: delegating the execution at Level 4 but the approval at Level 2. "You can run the project, but come to me before any key decision." This is a recipe for stalled work and resentful reports. Either it's their project, or it's yours. Mixed ownership is almost always worse than clean ownership at whichever level.

Rule 4: Expect and tolerate variance

At Level 3 and above, your direct reports will sometimes make decisions you wouldn't have made. Some of those will be wrong. Some will be different-from-you-but-also-right. You need to distinguish.

A useful rule from Marcus Buckingham: if their decision is in the top 75% of the outcome distribution for the problem, let it go. Only intervene when it's in the bottom 25%. This sounds arbitrary. It's actually the only way to scale without micromanaging — you accept a ceiling on the average quality of decisions below you, in exchange for actually being able to run a bigger operation.

The Specific Decisions That Should Never Be Delegated

Even with a mature ladder, a specific set of decisions should stay at Level 1 or 2 — the senior leader's, not the team's. Roughly:

• Firing a senior person. The decision itself, not the execution. The judgement about when someone's tenure should end is rarely delegable below the person two steps above them.
• Compensation for senior individuals. The individual comp numbers for anyone reporting directly to you should be your call. Delegated compensation design is fine. Delegated individual approval is how disasters happen.
• Non-recoverable strategic moves. Entering a new market, shutting down a product line, acquisitions, major pivots. These need to stay at the top.
• Legal and regulatory exposure. Decisions where a mistake creates exposure that can't be fixed with money — brand-damaging PR, regulatory violations, material risks to employees or customers.

The explicit carve-out matters. If you tell your team "I delegate everything" and then second-guess them on senior firing decisions, you've broken the contract. Say what's not delegable upfront, and then — for everything else — delegate cleanly at the right level.

The Honest Ceiling

The delegation ladder only works if you actually trust the people you're delegating to. If the underlying calibration is off — you have someone in a senior role who genuinely can't make good decisions at Level 3 — the ladder doesn't rescue them. It exposes them faster.

This is often the real problem the COO I mentioned was facing. Two of his 17 direct reports genuinely weren't ready for Level 3 autonomy, and the fact that he was holding everyone at Level 1 was masking that specific weakness in the team. Once the ladder was applied properly — the capable 15 moved up, the two weaker ones moved into different roles over six months — the organisation became capable of scaling. His hours dropped from 75 to 55. The business grew 30% the following year. The ladder didn't fix the organisation. It revealed what needed fixing.