When you manage people, you can still see the work. When you manage managers, you mostly see managers. The hardest part of becoming a leader of leaders is that your influence is now second-hand, it travels through other people's judgement, and a lot of what you used to do well no longer applies. The leaders who struggle aren't the ones who lack talent; they're the ones who keep doing their old job while quietly hoping the new one takes care of itself.

The quick version

  • Managing managers is a distinct role, not a bigger version of managing people. Your job shifts from delivering work through your own team to building leaders who deliver through theirs.
  • The work becomes indirect. You influence outcomes through other people's decisions, so coaching, selecting and holding managers accountable matters more than your own hands-on output.
  • The most common failure is staying a great individual contributor or a great first-line manager, solving problems you should be developing someone else to solve.
  • Span and attention are real constraints. You can lead more teams than you can do their work, but not infinitely, and the leadership team you build, not the headcount you carry, is what scales.

The idea in depth: a passage, not a promotion

The clearest map of why this feels so disorienting comes from Ram Charan, Stephen Drotter and James Noel in The Leadership Pipeline (Jossey-Bass, 2000; revised 2011). Drawing on decades of succession-planning work inside large companies, they argue that a leadership career isn't a smooth ramp but a series of six passages, each demanding a different set of skills, a different use of time, and, most awkwardly, different work values (what you believe is valuable to spend your day on). The passage that concerns us here is the second: from managing others to managing managers.

What makes this passage treacherous, in their account, is that it looks like a small step and is actually a large one. A first-line manager still does some of the work and coaches people who do the rest. A manager of managers does almost none of the work directly. Their job is to select and develop first-line managers, hold them accountable for managerial work (not just technical output), deploy resources across teams, and, this is the part people skip, coach managers on how to manage. Charan and colleagues note that this layer is frequently the most neglected, because organisations promote strong first-line managers and then assume the leadership of leaders will come for free.

When you manage people you can still see the work; when you manage managers, you mostly see managers.

So the move is: audit where your hours actually go for two weeks, then sort each block into "doing the work," "managing the work," and "developing the managers." If the third bucket is nearly empty, you haven't made the passage yet, you've just kept your old job and added meetings. The deliberate fix is to schedule the development work first (a recurring coaching one-to-one with each manager, focused on how they lead, not status updates) so it stops being the thing that gets crowded out.

Why you have to let go of the work, and what replaces it

Linda Hill, the Harvard Business School professor, spent a year tracking nineteen newly promoted managers for Becoming a Manager (Harvard Business School Press; revised edition 2003). Her central finding was that the move into management is a genuine identity change, not a skills top-up: people have to stop seeing themselves as the doer and start seeing themselves as the person who gets results through others, and most find the loss of direct control genuinely uncomfortable. That discomfort doesn't disappear at the next level; it compounds. Now you're getting results through people who are themselves getting results through others, and you're two steps removed from the keyboard.

The reason letting go matters isn't modesty. It's arithmetic. Your attention is finite, and every problem you personally solve is a problem one of your managers didn't learn to solve, and will bring you again next month. Hill's work, echoed in practitioner accounts like Julie Zhuo's The Making of a Manager (Portfolio, 2019), points the same way: the leader's output is no longer their own work but the multiplied output and growing capability of the people below them.

flowchart TD
  A(["Individual contributor
output = your own work"]) --> B(["First-line manager
output = your team's work + coaching"])
  B --> C(["Leader of leaders
output = managers you develop,
×their teams"])
  C -.->|"failure mode"| D(["Super-IC / super-manager
solving problems
you should be teaching"])
Each step up trades direct control for leverage, and the trap at every step is reaching back down. Leaders Loop

So the move is: when a manager brings you a problem, resist solving it and instead ask, "What are you thinking of doing?" Then coach the decision rather than make it. You're slower the first few times and faster forever after, because you're building a leader who handles the next ten problems without you. The exception that proves the rule: in a genuine crisis you do step in directly, leadership of leaders isn't passivity, it's choosing the rare moments your hands actually belong on the work.

An honest limitation. The Leadership Pipeline is a practitioner framework built from consulting experience, not a peer-reviewed, controlled study, and Hill's research, though rigorous and influential, follows a relatively small cohort. Both describe a strong, recognisable pattern, but they're best used as a lens for "which passage am I actually in?" rather than as laws. Real careers skip levels, combine roles in flat organisations, and vary by industry. Treat the model as a diagnostic, not a destiny.

Span, attention, and the limits of one head

Leading multiple teams runs into a physical constraint: there are only so many people you can know, coach and hold accountable well. The research on span of control is messier than the tidy "seven plus or minus two" rule of thumb suggests. Gallup's 2025 analysis reports the average number of direct reports per manager has climbed to 12.1, up from 10.9 the year before and nearly 50% higher than in 2013, managers are being stretched. But Gallup's meta-analysis across 92,252 teams found there is no single optimal team size. What predicts whether a larger span works is the quality of management and how much of the manager's time is eaten by their own individual-contributor tasks (a median of around 40%); push past that and engagement erodes as the team grows.

For a leader of leaders this lands as a clear instruction: the constraint isn't headcount, it's attention, and attention spent doing the work is attention not spent developing managers. The number you can lead well goes up precisely as you stop doing their jobs for them.

So the move is: if your span feels unmanageable, don't just ask for fewer reports, first reclaim the hours by handing back the work you've quietly taken on, then invest the recovered attention in the managers themselves. A well-developed leadership team is what lets you carry a wide span without dropping people; a weak one makes even a narrow span feel like drowning.

A worked example

Take a support organisation, call it the company Meridian. (Illustrative throughout; this is a teaching example, not a real org.) Priya has just been promoted from leading one 8-person support team to leading four teams through their four team leads, roughly 35 people in total. For her first quarter she does what made her successful: she jumps into the hardest tickets, joins the tense customer escalations, and personally rewrites the messy handover process for all four teams.

The numbers look fine, briefly. Then they don't. Her four leads stop bringing her decisions and start bringing her the symptom, they've learned Priya will take the hard thing off their plate, so they've stopped building the muscle to handle it. Two are visibly stalling; one is quietly job-hunting because the role feels like a holding pen. Priya is working twelve-hour days and her teams are getting less capable, not more. She has become a super-manager, the failure mode in the diagram above.

flowchart TD
  A(["Priya inherits 4 teams
via 4 team leads"]) --> B{"Where does her
time go?"}
  B -->|"Into the work:
tickets, escalations, process"| C(["Leads de-skill,
one starts job-hunting"])
  B -->|"Into the leads:
weekly coaching 1:1s"| D(["Leads own decisions,
capability compounds"])
  C -.->|"reset"| D
  D --> E(["Wider span carried calmly;
Priya works ON the system"])
The same week, two destinations, set by what the leader chooses to spend attention on. Leaders Loop

The reset is unglamorous. Priya stops taking the escalations and instead runs a weekly 30-minute coaching one-to-one with each lead, built around one question: "What decision are you wrestling with, and how are you thinking about it?" She makes the messy handover process their project to fix, not hers. The first month is slower and a little nerve-wracking. By the second quarter, three of her four leads are resolving the things she used to absorb, the flight risk has re-engaged because the job finally develops her, and Priya has time to do the work only she can do, spotting that two of the four teams are structurally overloaded and rebalancing them. She didn't get better at support. She got better at building people who are.

Frequently asked questions

How is managing managers different from managing people?

When you manage people, you can see the work and coach it directly. When you manage managers, your influence is indirect, it runs through their judgement and their teams. The core of the job moves from delivering and coaching work to selecting, developing and holding managers accountable for how they lead. Charan, Drotter and Noel call this a distinct "passage" with different skills, time use and values, not just a wider version of the same role.

I'm still the best person to solve the hard problems. Why shouldn't I?

Because every problem you personally solve is one your managers didn't learn to solve, and they'll bring it back. Your value at this level is the multiplied capability of the leaders below you, not your own throughput. The exception is a genuine crisis, stepping in then is judgement, not a habit. If you're stepping in weekly, that's the failure mode, not the exception.

How many teams or managers can one leader realistically handle?

There's no universal number. Gallup's research found no single optimal span; what matters is management quality and how much of your time is consumed by individual-contributor work. Average spans have risen (Gallup reports 12.1 direct reports in 2025), but a wide span only works if you've stopped doing the teams' work and invested that attention in developing the managers. Reclaim the hours first, then judge the span.

What if I was promoted but still have to do hands-on work too?

That's common in flat or smaller organisations, a "player-coach" role. The pipeline model still applies; you just hold two passages at once. The risk is that the hands-on work, which is concrete and urgent, crowds out the development work, which is abstract and never urgent. Protect the coaching time on the calendar deliberately, or it will quietly disappear.

How do I hold a manager accountable for managing, not just for results?

Look past the team's output to the leadership underneath it. Are they developing their people, giving feedback, building a healthy team, making decisions instead of escalating them? A manager who hits the number while burning out their team or making themselves indispensable has a managerial problem you should coach now, before it costs you the team later.

Related in the Toolkit

Leading leaders draws on the same foundations as leading people, raised a level: the way you flex your approach (leadership styles & models) and how you give your managers a direction worth leading toward (articulating & cascading vision) both become indirect, you're now setting the conditions in which other leaders lead.

Where to go next