The Manager's New Brief

1.1

Ch. 1 — What Management Is For (and What It Isn’t)

Part I — The mandate

💡 Core idea

Management is a distinct job: creating the conditions for a group of people to do important work well, consistently, and together — not being the best individual contributor in the room.

The promotion often feels like a continuation of the old job, plus meetings. It is not. Your output is now mostly indirect. If you still take the hardest work because you are fastest at it, you have not started the job. You have added a queue.

This is not only a philosophical claim. Gallup estimates, in the State of the American Manager report, that managers account for at least 70% of the variance in employee engagement scores across business units (Beck and Harter, 2015; based on Gallup’s engagement database of 27 million employees and more than 2.5 million work units). That is Gallup’s estimate, not a public replication, but the direction is the one this chapter is written for: the manager is a first-order variable, not a layer of coordination. Google’s Project Oxygen reached a similar managerial conclusion from internal data: teams with highly effective managers were more likely to achieve better results, be happier, and experience lower turnover.

Three jobs inside the job

Everything that belongs to you fits in three buckets.

Direction. What winning looks like this month and this quarter, what you will not do, and who owns which calls.

People. Hiring, coaching, feedback, fairness, and the conditions in which people will tell you the truth early.

Systems. Cadence, writing, decision records, and workflows — including human-plus-tool workflows — that still work when you are not in the room.

If a task does not serve one of these, it is probably residual individual work. Some of that is inevitable. Most of it should shrink.

The identity shift

You will be tempted to prove you still “have it” by doing the work yourself. That impulse is understandable. It is also how you become the bottleneck.

Wes Kao names the usual excuse: it would take too long to explain. Preparing the handoff is what makes delegation pay off. Bjorg adds the other half: good delegation means both sides feel trusted, even when the other person will not do it exactly your way.

Delegation alone is still not the job. Jason Cohen argues that you scale by building a team stronger than yourself, not by emptying your plate. If the team cannot do the work well without you, you have moved tasks, not built capability.

The honest objection, especially if clients still buy your expertise, is that sometimes they really do need you in the room. That is sometimes true. The next chapter is about choosing, for each task, whether you Own it (mode 1), Scope it (mode 2), or Release it (mode 3) — so “I have to be there” is a decision, not a default.

Early mistakes cluster here. Ben Kuhn, writing about his first time as a manager, includes delaying hard questions. Terrible Software catalogs the same family of errors: weak delegation, a lost sense of what “good” feels like when you no longer produce the work yourself, and clinging to the old identity. Read those as diagnostics, not as a personality test.

Myths to retire

Management is not status, surveillance, or being the smartest person in every meeting. It is not being liked.

Kore Nordmann is blunt: the power imbalance is real, and wanting to be friends with your reports is a dishonest frame. You can be warm, fair, and on their side. You cannot pretend you do not affect their livelihood. Honesty about hierarchy is part of the job, not a failure of culture.

A newer myth is that the layer itself is obsolete — that once tools can draft, summarize, and route work, managers are overhead. Rands treats “RIP middle management” as a fashion that will pass, as holacracy did. Colin Breck argues that AI will mostly amplify existing leadership challenges, and that leaders may need to stay closer to the real work, not further from it.

An April 2026 commentary by Silicon Carne made a sharper claim about that month’s tech cuts: they had “un profil précis : le middle manager” — not the engineer or the salesperson, but the person whose job is to forward information between teams, consolidate reports, and serve as a “courroie de transmission” between direction and execution. That is commentary, not a workforce study. The job this chapter describes is not that relay.

Camille Fournier, in a piece the author’s research digest summarizes as follows: visibility is not understanding; a machine that cannot be held responsible cannot decide; spend more time with people than with a model. That is the cleanest statement of why this job remains. You can use tools. You cannot outsource accountability.

The operating model is what remains, and what tools will inherit. Stuart Winter-Tear writes: “Agentic AI is a mirror. It does not invent coordination, clarity, or trust. It reflects whatever is already present.” And: “AI won’t fix your system. It will only show you what’s broken, faster.” Atlassian’s AI Collaboration Index (2025), a vendor survey by its Teamwork Lab of 180 Fortune 1000 executives and 12,000 knowledge workers, reports that workers “say AI makes them an average of 33% more productive,” while “transformation is rare”: “96% of companies have not seen dramatic improvements in organizational efficiency, innovation, or work quality.” Those figures are self-reported perceptions, not audited outcomes. If the job is only a relay, tools will expose it. If the job is direction, people, and systems, that is the model agents will inherit (Chapters 12 and 15).

There is also no “one weird trick.” Bjorg includes “just use AI” in that category. The work is still to understand the stakes and focus the team.

For a concrete comparison of expectations and manager/IC balance, see the 37signals Manager Playbook. Use it as a specimen, not as your policy.

What “good” looks like in year one

A useful first-year test is simple. Can each person name the priority? Can they disagree with you without punishment? Are they blocked by you less often than a quarter ago?

When a team is struggling, do not start with character. Molly Graham’s Waterline Model (2026), drawn from National Outdoor Leadership School practice, asks you to diagnose four levels in order: structure (goals, roles, context), then dynamics (how decisions, conflict, and information actually work), then interpersonal tension, and only then the individual. The rule of thumb is “snorkel before you scuba”: start with the shared system, which affects the most people, before you write a story about a person.

John Shook (2010), who helped introduce the Toyota system at NUMMI, the GM–Toyota joint venture in Fremont, describes “the very same workers” in a plant where “Absenteeism routinely ran over 20%” and quality had been “some of the worst in the GM system.” After the production and management system changed, “The absenteeism that had regularly reached 20% or more? It immediately fell to a steady 2%. The quality that had been GM’s worst? In just one year, it became GM’s best. All with the exact same workers.” Toyota did not lower the bar. Asked whether workers would have “the right to stop the line,” the answer was: “No, we intend to give them the obligation to stop it — whenever they find a problem.” Mike Fisher keeps the limit this book also keeps: “The individual is accountable for effort, behavior, and judgment. The leader is accountable for the conditions.” Do not hide in the system, and do not start with character when the system is still the problem. Chapter 10 returns to the judgment side.

Tools

🤖 AI-era tip: Start with a leverage audit, not a tool stack

Once a week, use an approved assistant to sort last week’s calendar and task list into “only I can do,” “someone else should do,” and “a tool can draft or check.” Then you decide. Do not outsource the judgment of what the team actually needs. Do not add a workflow until you can say what managerial work it is supposed to free — coaching, decisions, presence. There is no “just use AI” shortcut. And do not let a summary tool replace talking to the people who do the work.

Takeaways

Evidence