Management
in a Hurry

The Whole Thing in One Page

A manager can look busy while making work harder. They can fill calendars, collect dashboards, approve small decisions, demand accountability and still leave a team unsure what matters, who may decide, where problems go and whether telling the truth is safe. Management is easy to imitate because its visible rituals are not the job.

The job begins with interdependence. The moment useful work is split among people, one person's success depends on somebody else's judgement, timing, information or permission. Specialisation creates enormous productive power. It also creates hand-offs, queues, conflicts, blind spots and unequal authority. Management is the practice of making those dependencies produce a result rather than friction.

That makes management more than coordination. It is coordination carried out through power. Managers can allocate work, control information, approve spending, judge performance, influence pay and careers, and sometimes end employment. The subtitle therefore matters. Getting work out of people humanely cannot mean being pleasant while the system remains arbitrary. It means using authority so that the bargain is intelligible: people know the result they owe, the resources and discretion they receive, the constraints they must respect, how they will be judged, and how they can challenge a bad decision without being punished for the challenge.

Seven ideas do most of the work. First, management exists because divided labour creates dependencies. Second, every role needs a workable bargain between outcomes, authority, resources and constraints. Third, decisions should sit close to relevant information unless risk, irreversibility or cross-organisational consequences justify moving them upward. Fourth, measures and incentives change behaviour; they never merely observe it. Fifth, capability must be selected, built and supported before failure is treated as character. Sixth, information must be able to travel upwards, especially bad news. Seventh, as organisations scale, today's choices harden into structure, culture and distributions of power, so humane management requires routes for correction as well as routes for command.

The practical tools follow from those ideas. Goals convert purpose into commitments. Roles divide contribution. Decision rights place authority. Hiring and onboarding create capability. Meetings and written systems move information. Feedback and performance management compare reality with expectation. Incentives shape attention. Organisational design decides which conflicts are local and which travel upward.

Every tool can become its opposite. A target can clarify or invite gaming. A process can preserve knowledge or stop judgement. Delegation can create ownership or transfer blame. Psychological safety can support candour or be reduced to a slogan that protects weak performance. A flat structure can remove layers while leaving informal power untouched. A kind manager can still be unfair if they avoid difficult decisions until colleagues carry the cost.

The recurring test is therefore concrete. Does this arrangement help the person doing the next necessary piece of work understand the result, exercise appropriate judgement, see the consequences, surface trouble and receive a fair response? If not, the management system is consuming the capacity it was meant to coordinate.

Good management does not remove hierarchy, conflict or pressure. It makes them usable. It puts authority where it can improve decisions, makes trade-offs visible, distinguishes system failures from individual failures, and keeps enough voice in the system that people can report when the plan has stopped matching reality.

That is the book.

Why You Should Care

Imagine a competent employee who is failing. Their manager has given them three priorities, each described as urgent. Two senior people can reverse their decisions. The relevant data arrive late. A weekly target rewards speed, while complaints caused by rushed work are counted elsewhere. The employee asks which goal wins and is told to “use judgement”. Three weeks later they are criticised for choosing the wrong one.

Nothing in that story requires a bully. The harm comes from ordinary design choices. Management matters because it controls the conditions under which other people exercise effort and judgement.

The economic evidence is unusually clear that these conditions can affect output. In a randomised field experiment in large Indian textile firms, a package of management practices covering areas such as quality, inventory and production monitoring raised productivity by 17 per cent in the first year. A separate study inside one technology-services company found large differences between supervisors: replacing a boss near the bottom of the estimated quality distribution with one near the top increased team output by more than adding another worker to a nine-person team. Neither result is a universal effect size. Together they establish something more modest and more useful: managers and management systems can materially alter what a given workforce produces.

The leverage cuts both ways. A weak individual contributor can make their own work late. A weak manager can make lateness rational for ten people, then mistake the pattern for ten separate attitude problems. They can create rework by changing priorities, silence by punishing bad news, queues by retaining decisions, and burnout by treating overload as a test of commitment. Because the consequences are distributed, the cause can remain invisible.

The humane part is therefore not decoration. Managers hold asymmetric power. They influence workload, status, access to information, career opportunities and the interpretation of performance. Employees cannot respond to that power as equals while remaining inside the employment relationship. Fairness requires more than good intentions. It requires clear standards, relevant evidence, proportionate consequences, realistic capacity, opportunities to explain, and some protected ability to disagree.

This is also why management is not reducible to motivation. People do not enter work as empty containers waiting to be inspired. They bring motives, skills, obligations, fears, preferences and limits. The manager changes the environment in which those motives operate. Autonomy can support better judgement when the person has competence and context. The same autonomy can feel like abandonment when goals and resources are missing. Tight supervision can be useful for a novice or a dangerous task and destructive when it persists after expertise has grown.

You should care even if you never manage anyone. Management literacy helps you distinguish a genuine constraint from an arbitrary preference, a capability gap from a broken process, a useful target from a gameable proxy, and a real decision from a meeting that exists because nobody will own one. It also tells you when personal effort is subsidising a bad system. The employee who remembers every informal promise, chases every hand-off and works late to absorb every scheduling error may look indispensable while making organisational failure cheaper to ignore.

The same mechanics appear outside companies. Hospitals coordinate clinicians, beds, tests and risk. Schools coordinate teachers, pupils, timetables and standards. Governments coordinate expertise, budgets and political authority. Charities coordinate missions, donors and delivery. The purposes differ. The management problem remains: distributed knowledge must become collective action under constraints.

There is no universal management style because work differs. A nuclear control room should not be managed like a design studio. A mature warehouse process can tolerate less improvisation than a research programme whose method is still being discovered. The manager's craft is choosing the right amount of clarity, discretion, measurement, review and escalation for the work in front of them.

Once you see management this way, the role stops looking like a personality type. It becomes a set of design and judgement problems with moral consequences. The best managers are not those who appear most managerial. They are those whose teams can do difficult work with less confusion, earlier truth and fewer avoidable indignities.

The Core Ideas

1. Divided Work Creates a Coordination Debt

Adam Smith's pin factory remains useful because it contains management in miniature. Divide production into specialised operations and output rises. The same division means no worker now owns the whole object. Wire must arrive before it can be drawn, drawn wire before it can be cut, cut wire before it can be pointed, headed, finished and packed. Each local success becomes dependent on somebody else's timing.

Specialisation therefore creates what might be called a coordination debt. The organisation gains through concentration of skill and attention, then owes the work a mechanism for reconnecting what it separated. The debt appears as schedules, standards, meetings, interfaces, escalation routes, managers and information systems.

The larger the division, the more important the joins become. A software engineer can write correct code that breaks another service. A salesperson can close a profitable-looking contract that operations cannot fulfil. A procurement team can cut unit cost while increasing delays. A hospital department can optimise bed utilisation while making the whole patient flow worse. Local competence does not guarantee collective performance.

There are several basic ways to coordinate. Repetitive work can use standard methods. Professional work can rely on standardised training and shared norms. Units can coordinate through defined outputs. Plans can sequence dependent actions. Colleagues can adjust directly through conversation. Hierarchy can settle conflicts when local agreement fails. Markets and prices can also coordinate between units when internal exchange is sufficiently clear. Most organisations mix these methods.

Trouble begins when the method does not match the dependency. A novel problem is forced through a rigid procedure. A routine decision goes to a committee. Two teams are told to collaborate without knowing who has final authority. A manager centralises a decision because they are accountable for the outcome, then becomes the bottleneck that makes the outcome worse.

This is why the organisation chart is a poor map of the work. Boxes show reporting lines. They do not show where information must cross, where one role can block another, which decisions are reversible, or where a failure propagates. A better manager looks for dependencies first: who needs what from whom, by when, at what quality, with which decision rights if reality departs from plan.

The number of possible relationships also grows quickly as teams expand. Three people create three possible pairwise links. Ten create forty-five. Nobody needs to manage each link explicitly, but the arithmetic explains why adding people can increase communication load faster than capacity. Growth without clearer interfaces can make a larger team slower.

Management is therefore not primarily supervision. It is the deliberate repayment of coordination debt. The manager decides what should be standardised, what should remain discretionary, what information must move, what conflicts deserve escalation and where one person's local optimisation can damage the whole.

That work carries power. Somebody decides whose deadline moves, whose method becomes standard, whose judgement is trusted and whose error becomes a lesson for everyone else. The coordination problem and the moral problem arrive together. A humane system does not pretend power away. It makes the rules of its use more visible and gives people meaningful ways to supply information, exercise judgement and contest mistakes.

The distinction matters because organisations contain interests as well as dependencies. A deadline can be good for the customer and miserable for the team. A staffing cut can improve this quarter's cost and transfer risk into next quarter's reliability. Managers do not merely discover one technically correct answer. They arbitrate among claims made by people with unequal power. The quality of management depends partly on whether those claims are made visible before the strongest party silently wins.

2. Every Role Needs a Workable Bargain

A job description lists duties. A functioning role is a bargain.

The organisation wants a result. The employee needs enough authority, information, time, tools and skill to influence it. Constraints define what cannot be traded away. Evaluation determines how the result will be judged. Pay and career consequences make the bargain consequential. If these parts do not line up, the role can become unfair even when every document looks professional.

Start with outcomes. An outcome is a changed state, not motion. A customer receives a working installation. Stock accuracy reaches a defined level. A month-end close finishes by a certain date with specified errors below a threshold. A new analyst can handle the normal queue independently. “Write reports”, “attend calls” and “support the team” may be necessary tasks, but they do not tell anyone what success changes.

Outcome clarity matters because tasks are easy to perform ceremonially. A weekly report can be written for months after it has stopped informing any decision. A meeting can occur on schedule while the underlying conflict remains ownerless. An employee can complete every assigned activity and still be blamed for a result nobody gave them the authority to influence.

A good commitment therefore has at least five parts: the result, the owner, the evidence, the constraints and the review point. It also needs a capacity assumption. When managers assign more commitments than available labour can support, the team does not magically create capacity. It rations. Maintenance waits. Quality slips. Learning stops. People work unpaid extra hours. The choice still occurs, but at the level least able to admit it.

Priorities become real only when they displace something. Calling a new project “top priority” while protecting every existing deadline is not prioritisation. It is concealed overload. The managerial act is to state the trade: this work starts, that work pauses, this customer waits, this risk is accepted. The higher the manager's authority, the more responsibility they have to make those losses explicit.

Authority must match accountability. If a person is answerable for customer response time but cannot change staffing, queue rules or escalation, the system has created responsibility without control. That may sometimes be unavoidable, but it should be visible. Managers often use the language of ownership while keeping all important choices above. The employee then owns the consequences and borrows the decisions.

Goals can improve performance, especially when they are specific, demanding and supported by commitment and feedback. Yet precise targets are not automatically good. On unfamiliar work, a hard outcome target can make people repeat a poor method or hide uncertainty. A learning goal can be better: test three approaches, identify the failure mode, or establish which assumptions hold before promising the final number.

The bargain also includes reciprocity. Employment is not a gift from the organisation to the employee, nor a favour from the employee to the organisation. Each side exchanges something. Managers damage that exchange when they demand flexibility while offering none, celebrate commitment only when it means extra hours, or change the standard after the work is complete. A humane manager does not promise symmetry of power. They make asymmetry less arbitrary.

That includes explaining who bears which risk. An organisation can ask an employee to accept uncertainty, changing priorities or stretch assignments, but it should not pretend those costs are shared equally when pay, job security and decision authority are not. Naming the asymmetry does not remove it. It makes the bargain more honest.

This is the first place where kindness can mislead. A manager can be personally considerate while maintaining an impossible workload, vague expectations and inconsistent consequences. Another can be demanding and still fair because the result, resources, boundaries and evidence are clear. Employees need decency. They also need a bargain they can understand before they are judged by it.

3. Put Decisions Near Information, Risk and Consequence

Hierarchy solves one problem by creating another. It gives an organisation a way to settle disputes and make binding commitments. It also places many decisions in the hands of people furthest from the details.

The person closest to a customer complaint may know within thirty seconds whether a refund is sensible. Their manager sees a category. The director sees a monthly total. By the time the decision reaches the top, the information that made the answer obvious has been compressed into a slide. Centralisation buys consistency and whole-system perspective at the price of delay and lost context.

The answer is not “empower everyone”. Decision rights should be designed around three questions. Where is the relevant information? What is the cost if the decision is wrong? How reversible is the choice?

Low-cost, reversible decisions usually belong close to the work. High-impact decisions that commit scarce capital, expose the organisation legally or create consequences across many teams deserve wider review. Expertise can override rank. A junior specialist may be the correct decision-maker on a technical question while a senior manager remains responsible for the resources and consequences around it.

Toyota's line-stopping practice shows the principle cleanly. Under jidoka, an operator who detects an abnormality can signal it and stop the line. That is not an abolition of hierarchy. The authority is tightly bounded around a condition where local information is immediate and the cost of allowing defects to continue can compound. Standardisation and discretion reinforce one another.

Delegation should be equally explicit. A real delegation specifies the outcome, the decision being transferred, relevant context, constraints, resources, review point and escalation boundary. “Own this” is not enough. Nor is assigning preparation while retaining every meaningful choice.

Managers often fail to delegate because deciding themselves is faster today. That can be rational once. Repeated, it creates a permanent queue. Employees stop forming recommendations because experience teaches them that the answer lives above. The manager's overload then becomes evidence that nobody else is ready. Centralisation reproduces its own justification.

Escalation needs design too. It should mean that a problem has crossed a boundary of authority, risk or coordination, not that the employee wants somebody senior to absorb uncertainty. A useful escalation states the situation, likely consequence, available options, recommendation and time remaining. The senior person should decide, add missing context or deliberately return authority. Merely forwarding the issue upward moves confusion without reducing it.

Written decision records can improve learning when the stakes justify the effort. Record what was decided, the assumptions that mattered and the condition that would trigger review. Outcome alone is a poor judge of decision quality. A sensible decision can fail because an uncertain assumption broke. A reckless one can succeed through luck. Organisations that reward only outcomes become risk-averse after bad luck and overconfident after good luck.

Good decision design therefore keeps authority neither permanently high nor permanently low. It moves it to the level where information, consequence and accountability can be combined at acceptable cost. That is the practical meaning of control: the organisation can act quickly without pretending every choice is equally safe.

4. Measures and Incentives Change the Work

The moment a manager measures something, the measure enters the system it is describing.

Calls per hour may increase calls. On-time delivery may improve because teams promise later dates. Defect reports may fall because reporting becomes painful. Sales can rise while returns, discounts or bad-fit customers rise somewhere else. A metric is not a window. It is an intervention.

This happens because work contains more dimensions than any score can carry. Quality, speed, cost, learning, safety, resilience and customer value compete. Managers usually attach stronger rewards to the dimensions they can count. People respond rationally by protecting what is visible.

The formal version is the multitask problem: strong incentives on measurable tasks can pull effort away from valuable work that is harder to observe. The practical version appears everywhere. A support team rewarded for closing tickets may close them prematurely. A recruiter judged on vacancies filled may accept poor matches. A warehouse judged only on labour efficiency may defer maintenance and training.

The answer is not to stop measuring. It is to treat every measure as partial. Pair speed with quality. Pair acquisition with retention. Pair output with rework or defects. Use leading indicators to inspect the conditions expected to produce a result and lagging indicators to test whether the result arrived. Most importantly, keep managerial judgement available for the information the dashboard drops.

Pay creates the same problem with higher stakes. In Edward Lazear's study of Safelite AutoGlass, moving installers from hourly pay to piece rates increased output per worker by 44 per cent, with the paper separating incentive and selection effects. That is a powerful result in work where individual output could be observed and quality could be monitored. It is not a licence to attach bonuses to every target. In collaborative, creative or safety-critical work, individual incentives can distort cooperation or push effort towards what is easiest to count.

Incentives also include promotion, recognition, workload, access and punishment. If rescuing a crisis earns praise while preventing one earns silence, the organisation will produce heroes and emergencies. If the employee who always says yes receives more work, reliability becomes self-punishing. If raising a risk makes the messenger responsible for fixing it alone, silence becomes rational.

This is why culture often follows incentives more faithfully than speeches. Employees learn what the organisation values by watching what happens after an error, who receives the difficult opportunity, which deadline gets protected, who is believed in a dispute and what behaviour advances careers.

Managers should therefore ask of every target: what would a capable person do if they wanted to maximise this number while caring little about the underlying purpose? The answer exposes the metric's edge. Then ask which important behaviour remains unrewarded or invisible. That exposes the rest of the work.

A humane incentive system recognises that employees are adaptive, not mechanical. It does not assume every response to a target is moral failure. If the system makes gaming rational, management helped create the game. Accountability still matters, especially for deception or misconduct, but the manager must examine the signal they designed before condemning the person who responded to it.

5. Capability Is Built Before It Is Judged

Organisations frequently promote the person who excels at one job into another job and call the move meritocratic. Then they are surprised when the new manager struggles.

Evidence from sales organisations illustrates the mismatch. Alan Benson, Danielle Li and Kelly Shue found that firms placed substantial weight on current sales performance when promoting managers even though other observable characteristics better predicted the later performance of subordinates. Promotion can still motivate current workers, so the pattern reflects a trade-off rather than simple irrationality. The important lesson is that performance in role A is not a complete prediction of performance in role B.

Management requires a different bundle of capabilities from many individual-contributor roles: allocating attention, judging evidence, coaching, making trade-offs, handling conflict, giving clear feedback and taking satisfaction from other people's output. Technical credibility can help. It does not replace the new work.

Hiring should start from that work rather than from a vague picture of a good person. Define the outcomes, recurring decisions, difficult situations and capabilities needed on entry. Separate what must be present now from what can be learned after joining. Replace soft criteria such as “executive presence” or “culture fit” with observable behaviours relevant to the job.

Structured interviews improve selection by standardising the main questions and scoring process around job-relevant competencies. Work samples can go further when the work can be represented fairly. Ask a candidate to analyse a realistic problem, write a short plan, conduct a mock conversation or prioritise competing requests. The aim is not to make hiring mechanical. It is to reduce the amount of unexamined discretion available to charm, similarity and first impressions.

Selection is only the beginning. Onboarding should create useful capability, not calendar exposure. A new employee needs context, tools, relationships, examples, practice and feedback. Give them an early real outcome small enough to be safe and complete enough to reveal the full loop. Explain where questions go, which errors are cheap enough to learn from and what abnormal conditions require escalation.

Development continues through work design. People learn when tasks are difficult enough to stretch judgement and supported enough to avoid repeated failure. A manager who hoards consequential work prevents others from building the evidence required for trust. A manager who delegates without support transfers risk and calls the resulting failure development.

Workload belongs here too. Performance cannot be judged independently of capacity forever. High demands can sometimes be energising, especially when people have autonomy, support and resources. Sustained demands without recovery and control contribute to strain. Before labelling someone weak, a manager should ask whether the role is possible under the hours, tools and competing priorities supplied.

Capability and commitment are separate diagnoses. Someone may be willing and unable, able and unwilling, both, or neither. The intervention differs: training, clearer context, different work, stronger consequences, or exit. Treating every problem as motivation insults the willing person who lacks a method and protects the capable person who has chosen not to meet the standard.

A humane system therefore earns the right to judge. It sets a valid standard, supplies reasonable conditions, observes relevant evidence, gives timely correction and distinguishes learnable gaps from persistent mismatch. Fair performance management is demanding because it must examine both the person and the system that claims to measure them.

6. Bad News Is a Management Input

Every hierarchy has a built-in reason to hear good news first. People lower down depend on people higher up for approval, reputation, opportunities and sometimes continued employment. Telling a senior person that their plan is failing can therefore carry personal risk even when the organisation needs the information.

This creates one of management's most expensive failures: silence that looks like agreement.

Research on employee voice shows that people weigh both the likely usefulness and the social cost of speaking. They are more likely to stay silent when they expect retaliation, futility or damage to relationships. Psychological safety describes a related team-level condition in which people believe interpersonal risk-taking, such as asking for help, admitting an error or challenging an idea, is safe enough to attempt.

The concept is often softened in corporate language until it means comfort. That is a mistake. A psychologically safe team can have exacting standards, sharp disagreement and consequences for poor performance. The relevant question is whether a person can reveal information the group needs without paying an arbitrary interpersonal price for the revelation.

Managers create the answer through repeated reactions. Ask for concerns, then punish the first person who raises one, and the system has learned. Publicly interrogate whoever reports a problem while leaving the original decision-maker untouched, and future problems will arrive later. Thank people for surfacing risk, investigate it seriously and separate the messenger from the cause, and candour becomes more rational.

Voice also needs a destination. “My door is always open” is weak design if the employee must decide when an issue is serious enough, whether the manager is available and what evidence to bring. Teams need explicit channels: one-to-ones for private concerns, operating reviews for exceptions, escalation routes for urgent risk, retrospectives for learning, and protected routes outside the immediate manager where the manager is part of the problem.

Feedback travels in both directions. Good downward feedback stays close to the task. State what was observed, why it matters, the standard and the next change required. Ask for the employee's account because missing context can alter the diagnosis. Avoid turning one event into an identity. A large meta-analysis by Kluger and DeNisi found that feedback interventions improved performance on average but that more than one-third reduced it. Feedback is not beneficial merely because it is labelled feedback.

Conflict also becomes easier when its object is named. People can disagree about facts, goals, methods, resources, roles or relationships. Facts need evidence. Goals need a decision about priorities. Methods may need expertise or an experiment. Resource conflicts need a trade-off. Role conflicts need redesign. Relationship damage needs direct acknowledgement and behavioural repair. “Communicate better” is useless when the real conflict is that two people have been given incompatible authority.

The humane standard becomes clearest when performance is poor. Tell the person what category of conversation they are in. Developmental coaching is different from a formal warning. Serious concern disguised as friendly mentoring protects the manager from discomfort and denies the employee a fair chance to respond. Clarity is kinder than prolonged ambiguity.

Bad news is therefore not an interruption to management. It is one of management's primary inputs. A system that cannot carry error, dissent and uncertainty upward will eventually make senior authority less informed than the people it governs.

7. Scale Turns Choices Into Institutions

A small team can coordinate through memory. Everyone knows who decides, why the odd customer gets an exception and which informal promise matters. Growth breaks that arrangement. New people do not share the history. Exceptions multiply. Managers become queues. The organisation responds by writing rules, adding roles, introducing systems and creating layers.

This is necessary. It is also where management choices harden into institutions.

A repeated judgement becomes a policy. A recurring approval becomes a workflow. A successful manager's preference becomes “how we do things”. A temporary reporting line becomes a career ladder. The organisation begins storing old decisions in structure.

Structure solves coordination by deciding which conflicts happen where. Functional organisations group expertise, strengthening professional standards while forcing end-to-end results across departmental boundaries. Divisional structures group around products, markets or regions, strengthening ownership while duplicating expertise. Matrices preserve two dimensions and therefore preserve their conflict. None is universally best. Each makes some coordination easier and some more expensive.

Hierarchy also distributes status and opportunity. Managers control access to information, visibility, projects and promotion. Informal networks can matter as much as formal charts. Two employees with identical job titles may experience different organisations because one receives sponsorship, context and forgiveness while the other receives scrutiny. A purely technical account of structure misses this political reality.

Culture develops from the same accumulation. Edgar Schein's model distinguishes visible artefacts, espoused beliefs and deeper assumptions. The practical point is that employees learn culture from repeated consequences. If the company says quality matters but promotes people who ship fastest, speed is the stronger lesson. If it says speak up but treats dissenters as disloyal, loyalty means silence.

Scale also increases the temptation to manage by abstraction. Dashboards replace conversations with the work. Headcount becomes capacity. “Performance distribution” can make individual circumstances disappear. Remote systems and algorithmic tools can extend managerial visibility while narrowing what counts as evidence. Standardisation can reduce bias in one place and institutionalise it in another if the standard itself is poor.

The manager's job therefore changes with scale. Early on, they solve cases. Later, they should redesign the condition producing the cases. A repeated question becomes training. A recurring exception becomes a rule or a clarified boundary. A chronic hand-off becomes a new interface. A meeting that exists to reconcile two teams may reveal that the structure has placed one result across two incompatible owners.

Pruning matters as much as adding. Organisations remember the failure that created a rule and forget the cost the rule imposes on every future case. Every approval has a story behind it. Mature management asks whether the risk still exists, whether the control still reduces it and whether a cheaper control now works.

This is where the humane argument returns to the beginning. Division of labour creates the need for management. Scale makes the coordinating machinery powerful enough to become a world people must live inside for much of their waking week. The structure can enlarge human capability by connecting knowledge, resources and collective effort. It can also shrink people into compliance units whose local knowledge is ignored and whose cost is absent from the metric.

The answer is not perpetual informality. It is contestable structure: clear authority, bounded discretion, visible standards, routes for voice, reviewable rules and evidence strong enough to justify consequences. Scale should make coordination more reliable without making power harder to question.

How It Actually Works

Start with the result and the bargain

Begin with a piece of real work, not with the team chart. What changed state must exist when the work is done? Who benefits? What would count as failure? Which constraints cannot be traded away?

Write that result in language a competent outsider could test. Then name one owner for integration. Other people may contribute heavily, but somebody must carry the unresolved trade-offs and return when the commitment can no longer hold.

Now make the bargain visible. What authority does the owner have? Which resources are available? What decisions remain elsewhere? Which assumptions about capacity are embedded in the deadline? How will performance be judged? This step prevents a common managerial trick in which an employee is given a result, denied the decisions needed to influence it, and later told they failed to take ownership.

Prioritisation belongs here. If the new result conflicts with existing commitments, decide what moves. A priority that displaces nothing is additional load.

Translate broad aims into a short chain of commitments rather than a spray of targets. A company may want profitable growth, but a product team can act only on a contribution it can influence: release a defined capability, improve retention among a named customer group, reduce a recurring source of service cost. Each local commitment should preserve enough of the wider reasoning that people can resolve exceptions without appealing upward for permission.

Then test the commitment against capacity. Estimate ordinary workload, known peaks, training time, maintenance, holidays and the interruptions that history says will occur. Managers often plan against theoretical availability and treat predictable variation as employee failure. A credible plan leaves some room for the world to be less obedient than the spreadsheet.

Map the dependencies

Trace what the result needs from other roles. Inputs, approvals, specialist advice, customer decisions, systems, suppliers and downstream acceptance all count. For each interface, specify what must cross, in what form and by when.

Many communication problems turn out to be definition problems. Sales calls a lead qualified; marketing uses a different definition. Product says a feature is ready; support discovers documentation is missing. Finance approves spending after a team has already made the commercial promise. More messages do not repair incompatible meanings.

Choose the cheapest coordination method that fits the dependency. Repetitive work may need a standard. Novel work may need direct adjustment. Cross-unit trade-offs may need a named decision owner. High-risk exceptions may need escalation. Do not create a meeting because a dependency exists. Create a mechanism that changes the state of the work.

Watch the hand-off itself. A useful interface has a definition of ready, a definition of done and a response when either side disagrees. If engineering sends work to testing, what evidence makes it testable? If sales passes an account to implementation, what promise has already been made and what remains negotiable? If a hospital ward requests a scan, what information must accompany the request? These details look administrative until their absence creates rework.

Where possible, shorten dependencies rather than managing them forever. Give one team end-to-end ownership, colocate complementary skills, automate a stable hand-off or remove an approval that adds no information. The best coordination mechanism is sometimes fewer things to coordinate.

Place decisions before designing meetings

List the recurring decisions that make the result possible. Who chooses the technical method? Who may vary price? Who can change scope, deadline or staffing? Who can stop the work when a safety or quality threshold is breached?

Assign final authority explicitly. Give input rights to people with material knowledge. Define what must be communicated after the decision and what boundary triggers escalation. Plain language beats fashionable frameworks if people can recall it under pressure.

Match review intensity to risk and capability. A novice making an irreversible decision needs a closer checkpoint than an expert making a reversible one. Review reasoning and assumptions, not every keystroke. As judgement becomes reliable, widen the boundary.

Distinguish consultation from consent. Some decisions need input from many people but final authority from one. Others genuinely require joint agreement because each party controls a necessary resource or bears a non-transferable risk. Confusing the two produces fake consensus. People sit through consultation believing they possess a veto, or are told they were consulted after the decision was already made. State which kind of participation is being offered.

Hire for evidence, then finish the job after hiring

Translate the role into selection evidence. Which capabilities must exist on day one? Which can be learned? What difficult situations reveal them? Use consistent job-related questions and scoring criteria. Add a work sample when practical.

Do not confuse confidence with competence. Interviews reward social fluency unless the process resists it. Independent scoring before panel discussion can also reduce the tendency for the first strong opinion to become the group's memory.

Once hired, create an entry path towards independent contribution. Provide tools, context, relationships, examples and an early real outcome. Explain how normal work differs from abnormal work and where abnormality goes. A newcomer has finished onboarding when they can handle ordinary cases, recognise exceptions, find help and explain how their decisions affect the wider result.

Use the first weeks to test the role as well as the person. Newcomers see friction that veterans have learned to route around. Ask which instruction contradicted another, which permission was unexpectedly hard to obtain and which piece of context was missing. If every new hire trips over the same obstacle, stop treating the obstacle as an initiation ritual.

Managers should also establish the social map. Formal authority rarely tells a newcomer who carries specialist credibility, who owns institutional memory or which stakeholder must be involved early. Hiding this information turns influence into an insider advantage. Making it explicit reduces the amount of political learning that must occur through mistakes.

Create an operating rhythm

Recurring work benefits from recurring moments for information and decision. Without a rhythm, every issue becomes an interruption. With too much rhythm, the organisation spends its time reporting to itself.

Separate meeting purposes. A one-to-one exists for the employee's work, judgement, development and concerns. An operating review examines commitments, exceptions and dependencies. A decision meeting settles a defined choice. A planning session allocates future capacity. A retrospective studies completed work.

Every recurring meeting should begin with an input and end with a changed state. The input may be a short pre-read, metrics or named decisions. The changed state may be a revised plan, approved choice, owner or escalated risk. If nothing changes, ask whether the information should simply be published.

Protect one-to-ones from becoming private status meetings. Status can often be written. The scarce value of a private conversation is information that status systems suppress: uncertainty, disagreement, workload, development and observations about the organisation itself.

Define what interrupts the cadence. Safety, legal exposure, customer harm, security incidents or irreversible commitments may justify immediate escalation. Everything else should have a known channel. Good operations distinguish urgency from anxiety.

Control the total cost of coordination. Ten people in an hour-long meeting consume ten working hours before preparation and recovery. That cost may be justified, but it should be visible. Invite only people whose information or authority can change the result, and let others receive the record. Seniority is not a reason to attend.

Written communication needs equal discipline. A copied recipient is not informed merely because their address appears in the thread. State the question, owner and deadline. Separate material that requires action from material kept for context. Good written systems reduce memory load; bad ones create a searchable archive of unresolved ambiguity.

The manager should also decide what deserves memory. A temporary workaround can disappear after the incident. A decision that changes policy, customer commitment, safety or architecture deserves a durable record. Organisations become unreliable when important decisions live only in the inboxes of people who happened to attend.

Make the work visible without turning it into a game

Choose a small set of measures that help people see whether the system is moving towards the result. Include quality and flow, not just volume. Track queues and rework where they reveal hidden cost. Pair leading indicators with lagging outcomes.

Then look for gaming before accusing anyone of it. Ask how the number could improve while the real outcome worsened. If the answer is obvious, build a countermeasure or use judgement. Targets should make trade-offs easier to discuss, not hide them behind precision.

Use measures at the level where someone can respond. A weekly quality number may help a team change a process; an annual engagement score may reveal a broad pattern but cannot diagnose the manager, policy or event that caused it. Do not demand local accountability for a metric whose causal structure is mostly outside local control.

Use visual systems and dashboards to shorten the distance between abnormality and response. A red metric that nobody has authority to act on is decoration. A metric that changes every week without a clear owner creates theatre rather than control.

Where feasible, put the measure beside the work. A visible queue, defect board or service backlog can create shared reality faster than a monthly slide. The purpose is not surveillance. It is to let the people capable of correcting the process see abnormality while correction is still cheap.

When several measures move together, resist immediate causal stories. A fall in defects after training may reflect the training, easier work, changed staffing or different reporting. Managers need enough humility to treat operational data as evidence rather than proof. The closer the measure is to a controlled process, the stronger the inference can be.

Review also needs ownership. If a metric repeatedly misses and every meeting ends with a different person promising to look into it, the system has identified a problem without assigning a decision. A useful review ends with one of three states: continue, change the plan, or escalate because the current level lacks authority.

Close the loop on previous decisions before opening new ones. Teams lose trust in reviews when last week's actions vanish into the next agenda. A short record of owner, decision and due condition is enough. The purpose is organisational memory, not minutes for their own sake.

Use exceptions to choose the depth of discussion. A stable process does not need a weekly forensic examination simply because a meeting exists. Spend attention where the result, assumption or dependency has moved outside its expected range. This keeps the operating rhythm from rewarding whoever can produce the longest update and directs managerial time towards variance that may require a decision.

Review variance and diagnose the level of failure

At the chosen interval, compare expectation with reality. What did we expect? What happened? Where did the difference enter the system?

Separate recurring patterns from isolated events. If a failure appears across several people, shifts or locations, investigate the process, tools, workload or incentives. If one person repeatedly differs under comparable conditions, examine capability, judgement or commitment. Managers create unfairness when they turn system variation into individual blame. They create helplessness when they turn every individual choice into a system excuse.

Feedback should be timely enough to affect the next attempt. Describe the observed behaviour or output, its consequence, the standard and the next change required. Ask for the employee's account. Good praise works the same way: identify what should be repeated rather than supplying warmth without information.

Retrospectives should produce a changed mechanism. “Be more careful” is rarely a changed mechanism. A new checklist, clearer owner, altered sequence, training need, decision boundary or removed incentive may be.

Separate learning from punishment when possible. If every admission of error immediately becomes evidence in a disciplinary case, people will rationally protect themselves. Serious misconduct still requires consequences, but ordinary operational learning needs room to distinguish a bad outcome, a reasonable mistake, a negligent choice and a deliberate breach. Treating them alike destroys useful information.

Handle poor performance without theatre

When performance falls below standard, say so early. Diagnose clarity, capability, capacity, authority, tools, incentives and commitment. Agree what must change, what support will be supplied, what evidence will count and when the position will be reviewed.

Be explicit about the status of the conversation. Coaching, corrective feedback and formal performance action are different. Do not keep an employee in a fog because clarity feels harsh. Jurisdiction-specific law and company procedure still govern formal action, but the management principle is universal: the person should understand the concern and the standard being applied.

Conflict should be decomposed before it is personalised. If two managers want the same employee's time, the conflict is a resource decision. If two functions use different definitions of “ready”, the conflict is an interface. If one person repeatedly humiliates another, it is behavioural. Each problem requires a different remedy.

Sometimes the correct outcome is a role change or exit. The humane version is evidence-based, prompt, private and proportionate. It protects dignity without pretending the mismatch does not exist. Delaying a necessary decision can shift the cost to colleagues who absorb the work and to the person who continues under an expectation they are unlikely to meet.

Do not make the team guess what happened. Privacy limits what can be shared, but silence can create its own story. Reaffirm the standard, the allocation of work and what colleagues need to know without exposing personal detail. Humane treatment of one employee should not require everyone else to operate inside unexplained uncertainty.

Redesign when exceptions become patterns

As the organisation grows, watch for signals that the original design is failing: overloaded managers, repeated approval queues, duplicated work, chronic cross-team conflict, informal coordinators holding everything together, or employees spending large amounts of time translating between systems.

Solve patterns by changing conditions. A recurring approval may become a rule. A repeated question may become training. A persistent cross-team dispute may need a clearer owner. A metric that drives bad behaviour may need a countermeasure or deletion. A manager whose span has become impossible may need a new layer, fewer responsibilities or a different operating model.

Do not add structure automatically. Every layer can shorten one coordination path and lengthen another. Test whether it improves the quality and speed of decisions below, not whether it gives senior leaders more reports.

A useful redesign starts with failure paths. Follow one delayed customer order, one disputed technical decision or one repeated staffing problem across the current structure. Count the transfers, approvals and reinterpretations. Then change the smallest structural feature that removes a recurring loss. Reorganisation by boxes before diagnosis often moves the same problem under new headings.

After a redesign, specify what should improve. Decision time, customer hand-offs, manager load, quality or role clarity should move in a direction that can be observed. Otherwise the organisation can mistake the emotional relief of announcing a new structure for evidence that the structure works. Reorganisations need feedback loops too.

Then prune. Retire meetings whose changed state has disappeared, rules whose original risk has faded and reports no decision uses. Management systems accumulate because adding a control is easier to justify than removing one. The cost of that accumulation is paid in attention.

Growth also changes what managers should spend time on. The first-line manager is close enough to coach work and detect abnormalities. Higher levels should spend more time resolving cross-unit trade-offs, allocating scarce resources, setting decision rules and changing the architecture that lower levels cannot change. When every layer spends its week asking for status, hierarchy adds transmission loss without adding judgement.

How we know

Management evidence is harder to interpret than laboratory evidence because firms select managers, practices arrive in bundles and successful organisations may adopt formal systems after success rather than before it. Single-company estimates therefore show mechanisms and magnitudes in a setting, not universal laws.

The strongest case comes from convergence. Randomised interventions, including the Indian textile experiment, show that management practices can cause productivity gains in specific conditions. Detailed personnel data show large supervisor differences and promotion mismatches inside firms. Organisational psychology provides experimental, longitudinal and meta-analytic evidence on goals, feedback, selection, voice, motivation and work design. Historical and operational cases show how mechanisms function, while also generating myths cleaner than the evidence, as the Hawthorne story demonstrates.

No serious evidence supports one best management style. Task novelty, measurability, danger, interdependence, labour institutions and culture all matter. The reliable conclusion is conditional: management changes performance, the mechanisms can be studied, and any prescription that ignores the work, the power relationship or the surrounding system should be treated with suspicion.

What People Get Wrong

“The best worker will make the best manager”

The rule feels fair because current performance is visible and promotion is a reward. It confuses recognition with prediction. The excellent salesperson, engineer or clinician produces through personal skill. The manager produces through other people's conditions, decisions and development.

Evidence from sales organisations finds that firms strongly reward current sales performance in promotion even when other observable characteristics better predict later managerial performance. That may partly motivate current workers, so the system is not irrational by definition. It is still a trade-off.

The fix is to create progression that does not require management and to test managerial evidence before permanent promotion. Coaching, team leadership, resource allocation and difficult feedback reveal more about the new role than another quarter of individual output. Management should be a different job, not the trophy at the top of every craft.

The organisational consequence is larger than one bad promotion. If the only route to status and pay is management, talented specialists are pushed away from the work they do best and people who do not want to manage learn to perform interest in it. A dual career ladder is not an HR nicety. It protects the organisation from using authority as compensation.

When promotion is still the right move, support the identity change. New managers often keep doing the technical work because it is where competence feels safest. The cost is that coordination, coaching and decision work are postponed until they become emergencies. The role needs time allocated for the work that only the manager can do.

“Good managers keep tight control”

Personal control is reassuring because the manager sees every decision. It can be useful in emergencies, with novices or where errors are irreversible. As a default, it creates a queue and trains passivity.

A strong control system is different. It defines outcomes, limits, abnormalities and escalation, then gives people authority within them. Toyota's line-stopping practice is a useful example because discipline and local discretion coexist.

The hidden cost of centralisation is that delays above are less visible than mistakes below. Managers may therefore believe control is safer than it is. Use staged delegation: transfer a defined decision, review it at a risk-appropriate point and widen authority as capability becomes reliable.

Control also needs a stopping condition. Managers who once needed close oversight can continue it after the reason has disappeared because the habit feels safe. State what evidence will justify loosening review. Otherwise supervision becomes permanent, and the employee can never generate the independent track record that would supposedly earn trust.

“People need to be motivated”

People need conditions in which useful motives can operate. The phrase “motivation problem” often arrives before anyone has checked clarity, workload, tools, authority, fairness or whether the work itself makes sense.

Research on work motivation supports the importance of autonomy, competence and relatedness, but this does not mean every job must be intrinsically enjoyable. Pay, status, security and consequences matter. So do task design and progress.

Diagnose before inspiring. A capable person may be exhausted, blocked or responding rationally to an incentive. An unwilling person may need a consequence rather than a speech. Motivation is an outcome of a person interacting with a system, not a liquid managers pour in.

This also changes the question after a strong performer goes flat. Instead of asking how to restore enthusiasm, ask what changed in the exchange: the task, manager, workload, autonomy, fairness, prospects or belief that effort will matter. Sometimes the answer is personal and outside work. Management still benefits from diagnosing rather than moralising.

“Pay for performance improves performance”

Sometimes it does. Piece rates at Safelite were associated with a large rise in installer output, in work where individual contribution was observable and quality could be monitored. That setting matters.

Performance pay becomes dangerous when the rewarded metric covers only part of the job, when people depend heavily on colleagues or when quality is hard to observe. Strong rewards make proxy optimisation more attractive.

Use variable pay where the measure tracks genuine contribution well enough and the unwanted responses can be detected. Elsewhere, broader pay bands, team rewards, professional standards and managerial judgement may work better. Compensation is a design choice, not a universal lever.

The distribution of risk matters as well. Variable pay shifts some uncertainty from the organisation to the employee. That may be reasonable where individuals control the result and understand the formula. It is harder to defend when bonuses depend heavily on market movements, executive choices or opaque ratings that the employee cannot influence.

The humane test is therefore twofold: does the incentive improve the behaviour the organisation needs, and is the risk being allocated to someone who can reasonably influence it? A scheme can pass the first test and fail the second.

That distinction is why incentive debates framed as “does bonus pay work?” are too crude. The answer depends on what is measured, who controls it, how quality is protected and what cooperation the reward may displace.

“More communication fixes coordination”

Communication has a cost. A team can drown in updates while the underlying decision remains unowned.

Coordination failures often come from unclear interfaces: two teams use different definitions, authority overlaps, or nobody knows which constraint wins. More messages then circulate the same ambiguity faster.

Ask what information must cross, who needs it, when it changes a decision and in which medium. Written updates are good for durable detail. Conversation is good for ambiguity and rapid adjustment. Meetings are good when simultaneous interaction changes the result. Communication volume is not evidence of coordination quality. The goal is enough shared information for dependent work to fit together, delivered with the least avoidable interruption and ambiguity.

A useful test is subtraction. Remove one recurring update or meeting and ask what decision becomes worse. If nobody can name it, the communication may exist mainly to reassure managers that work is visible. Visibility can be valuable, but it should be bought deliberately rather than through permanent interruption of everyone doing the work.

“Culture is what the company says it values”

Values statements describe intent. Culture is what repeated experience teaches people to expect.

Employees watch who gets promoted, what happens after an error, whether deadlines outrank quality, whether dissent damages careers and whether senior people obey the rules imposed on everyone else. Those patterns become the operating assumptions of the organisation.

Change culture through recurring decisions, incentives, manager behaviour and systems. Rewrite the values page afterwards if necessary. When stated values conflict with consequences, consequences win.

This is why culture change is slow when leaders change language but preserve promotion, workload and decision rules. People do not need a workshop to notice the contradiction. They update from behaviour. The fastest credible cultural signal is often a costly decision that proves a stated value still applies when keeping it hurts.

“Humane management means being nice and flat”

Courtesy matters. Flatness can remove unnecessary layers. Neither guarantees humane work.

A pleasant manager can leave expectations vague, avoid performance decisions and let reliable employees carry weak colleagues indefinitely. A flat organisation can concentrate informal power in founders, experts or social networks without naming it. Hidden hierarchy is harder to challenge because it denies that authority exists.

Humane management requires legible power: clear standards, bounded authority, realistic workload, voice, fair evidence and proportionate consequences. Sometimes the fair decision is demanding. Sometimes it is to stop work, move a deadline, reject a request or end a role. Niceness without clarity can become a way of shifting discomfort downwards.

Flatness has a second danger. When formal hierarchy is reduced, expertise, tenure, access to founders and social confidence can become the real authority system. Because those forms of power are unofficial, employees may have fewer routes to challenge them. Removing a title does not remove dependence. Sometimes a named manager is more accountable than an invisible centre of influence.

Use It

Find the dependency before the personality

When work repeatedly fails between two people or teams, resist the first explanation that someone is difficult, lazy or bad at communication. Map the dependency. What must cross? Who decides? Which definition differs? What queue or constraint sits between them?

Personality can matter. Start with the design because recurring friction often survives staff changes when the interface stays broken.

Try drawing the work as a chain. Mark every point where it waits for information, permission or another person's output. The longest delay is often not inside a task but between tasks. That is where management attention can remove more friction than another round of coaching the people already compensating for it.

Ask what bargain the person is being held to

For any role, write four columns: outcome, authority, resources and constraints. Then add the evaluation rule. If accountability is high while authority or resources are low, the role may be structurally unfair.

Use the same test on yourself. If a manager wants ownership, ask which decisions come with it. If the answer is none, the request may mean responsibility for consequences rather than control over causes.

Then inspect whether the constraints are legitimate. Some exist for safety, law, cash or interoperability. Others are inherited preferences. Ask who can change each one. A role becomes easier to manage when people know which walls are structural and which are merely waiting for a decision.

Make the trade-off visible

Whenever a new priority appears, ask what moves. Time, attention and recovery are finite. If nothing is displaced, the organisation has chosen overload while refusing to name it.

This question is especially useful with senior requests. It converts “can you also do this?” from a test of loyalty into a resource decision: yes, and which existing commitment should move?

Use the same discipline for quality. If a deadline cannot move, name what will. Scope may shrink, testing may shorten or cost may rise. Pretending all four constraints can remain fixed does not preserve quality. It merely ensures that the trade-off is made silently by the person with least authority to disclose it.

Make bad news cheap and useful

Notice how the system treats the first person who reports a problem. If they are blamed, interrogated publicly or made sole owner of the fix, future information will arrive later.

Create explicit routes for bad news and ask for recommendations where appropriate. Reward early visibility without rewarding carelessness. The goal is not comfort. It is to reduce the personal cost of supplying information the organisation needs.

Managers can test this by asking for the last piece of bad news that reached them early. If none comes to mind, either the team is extraordinarily lucky or the channel is weak. Ask what people knew before you did and why. The answer is often more useful than another anonymous survey.

Review the metric from the other side

For every important target, ask how a smart person could improve the number while damaging the real outcome. Then look for that behaviour before the target becomes expensive.

Also ask what important work the metric ignores. Maintenance, mentoring, risk prevention and quality often disappear because success is the absence of a visible event. If invisible work matters, create evidence for it or protect managerial judgement around it.

Do not answer proxy problems by multiplying proxies indefinitely. A dashboard with thirty countermeasures can become less interpretable than the work. The aim is enough measurement to reveal direction and abnormality, followed by informed investigation. Metrics should start questions more often than they end them.

Inspect what has hardened

Once a quarter, choose one recurring approval, meeting, report or rule and ask why it exists. Find the failure that created it. Does the failure still matter? Does this control reduce it? Is there now a cheaper way?

Organisations rarely remove management machinery at the speed they add it. Pruning returns attention to the work.

Make removal reversible where possible. Pause a report for a month, shorten a meeting, raise an approval threshold or run a controlled exception. Managers often preserve bureaucracy because deletion feels reckless. Small experiments let the organisation learn whether the control was carrying hidden value or only remembered fear.

Keep a short list of controls on probation. Every new recurring meeting, report or approval should have an owner and a date when its usefulness will be reviewed. Temporary management machinery has a habit of becoming permanent unless expiry is designed in.

The exercise also exposes ownership. If nobody can say who has authority to remove a rule, the organisation has created a control without a governor. That is a small example of how bureaucracy becomes self-preserving.

The limits

Management cannot remove scarcity, conflict or unequal interests. Sometimes the organisation wants more output than employees reasonably want to give. Sometimes pay is low, jobs are insecure, work is unpleasant or bargaining power is one-sided. Better meetings do not resolve those conflicts.

Nor can every decision move towards the front line. Safety, regulation, capital allocation and cross-organisational risk can require central authority. Psychological safety cannot make all speech consequence-free. Performance management cannot guarantee a fair judgement when evidence is weak. Hiring remains prediction under uncertainty.

Management techniques are also context-sensitive. Practices that help a high-volume plant can damage exploratory research. A performance measure useful for one job can be corrosive in another. A manager should therefore ask what mechanism a practice is meant to change and whether the surrounding conditions match the claim.

Most importantly, humane management cannot be reduced to a productivity strategy. Treating dignity only as a route to output turns the moral claim into another incentive. People deserve intelligible standards, fair treatment and proportionate use of power even when the immediate productivity effect is uncertain.

There is also a boundary between management and employment politics. Collective bargaining, labour law, discrimination, ownership and the distribution of profits shape what managers can do and whose interests carry weight. This book cannot settle those questions without becoming Work in a Hurry, Class in a Hurry or a labour-law manual. It can insist on one point: techniques do not make conflicting interests disappear.

The one thing to keep

When someone is failing at work, do not begin with the person. Begin with the arrangement between the person and the work.

What result were they asked to produce? What authority did they have? What information and resources reached them? What incentives shaped their choices? What did the system teach them about speaking up? What part of the failure recurs across people, and what part follows this individual?

Then judge. Good management does not dissolve accountability into systems thinking. It earns accountability by making the system legible first.

Apply the same standard upward. If a team repeatedly misses because priorities change, approvals wait or staffing remains below the planned workload, management performance belongs in the diagnosis. Authority without reciprocal accountability is merely power with better stationery.

This lens is useful in post-mortems. Ask what the worker could have done differently, then ask what the manager could have designed differently. If only the first question is permitted, the review is discipline, not learning. If only the second is permitted, individual agency disappears. Good management can hold both. The discipline is to make the level of explanation match the level at which a change can be made.

It is equally useful before work begins. Ask the person to restate the outcome, constraints and escalation boundary in their own words. Misunderstanding discovered before execution is cheap. Misunderstanding discovered during performance review is often expensive and unfair.

That is the humane edge of the subject. Management has power because divided work needs coordination. The manager's responsibility is to use that power so competent people can contribute judgement rather than guess at intention, tell the truth before it becomes expensive, and know the terms on which their work will be judged.

Terms

Management

The practice of turning collective purpose into coordinated action through goals, roles, decisions, information, resources, incentives and correction. It includes the design of the system around work as well as direct supervision.

Manager

A person with formal responsibility for a unit, team, process or set of resources, including authority that affects other people's work. A manager's output is partly the conditions they create for others.

Coordination

The alignment of dependent activities so separate contributions produce one result. Coordination can occur through standards, plans, shared training, direct adjustment, prices or hierarchy.

Division of labour

The separation of work into specialised tasks or roles. It can raise productivity while creating dependencies, hand-offs and a need to reconnect partial knowledge.

Interdependence

A condition in which one person's or unit's result depends materially on another's action, information or decision. Interdependence is the basic reason management exists.

Hierarchy

A ranked arrangement of authority in which some roles can make decisions binding on others. It can settle conflicts and concentrate accountability while creating delay, filtering and power asymmetry.

Span of control

The number of people or units reporting to one manager. A workable span depends on task similarity, capability, distance, risk, systems and the amount of coaching or coordination required.

Accountability

The obligation to answer for a result, decision or standard. Fair accountability requires a reasonable connection between what a person can influence and what they are judged for.

Authority

The legitimate power attached to a role to decide, allocate resources or bind the organisation. Authority is most useful when explicit, bounded and connected to relevant information.

Responsibility

The duty to perform work or care for an obligation. Several people can share responsibility, while final authority or integration usually needs clearer ownership.

Decision rights

The allocation of final authority, required input, notification and escalation boundaries for a decision. They turn vague collaboration into an operating rule.

Delegation

The transfer of a defined result and meaningful decision authority within stated constraints. The manager remains responsible for context, resources, review and the quality of the boundary.

Centralisation

The concentration of decisions in senior or common roles. It can protect consistency and whole-system trade-offs while increasing queues and distance from local knowledge.

Decentralisation

The distribution of authority towards local teams or specialists. It can improve speed and responsiveness when capability, information and boundaries are strong.

Line manager

A manager with continuing direct authority over employees doing operational or professional work, commonly responsible for priorities, development and performance.

Functional structure

An organisation grouped by expertise, such as engineering, finance or operations. It concentrates professional capability but forces end-to-end results across departmental boundaries.

Divisional structure

An organisation grouped around products, regions, markets or customer groups. It strengthens end-result ownership while often duplicating specialist resources.

Matrix

A structure in which people answer to two organising dimensions, often function and product. It preserves competing needs by preserving their conflict, so decision rules matter greatly.

Cross-functional team

A group containing several specialties and organised around a shared result. Its effectiveness depends on common priorities, authority and protection from conflicting home-team demands.

Standard operating procedure

A documented method for recurring work. A good procedure preserves knowledge and supports training while allowing defined exceptions and revision.

Key performance indicator

A measure selected to signal performance on an important outcome or process. A KPI is always a proxy and can distort behaviour if treated as the whole result.

Leading and lagging indicators

Leading indicators track conditions expected to influence a later result. Lagging indicators record the result after it arrives. Used together, they support earlier correction and later verification.

Objective

A defined result or changed state towards which work is directed. A useful objective includes evidence, constraints, an owner and a review point.

Incentive

Anything that changes the expected reward, cost, status or convenience attached to behaviour. Incentives include pay, promotion, workload, attention, peer norms and punishment.

Psychological safety

A shared belief that interpersonal risk-taking, such as asking for help, admitting an error or challenging an idea, is safe enough within a team. It does not remove performance standards.

Employee voice

Voluntary communication of ideas, concerns, questions or challenges intended to improve work or prevent harm. Voice depends on both the perceived value and personal cost of speaking. It differs from general participation because the information may be unwelcome to someone with more authority. Voice is therefore a test of whether hierarchy can receive corrective information from below.

One-to-one

A recurring private conversation between manager and employee about work, judgement, obstacles, feedback, development and concerns. It should not collapse into a status recital. The recurring privacy of the format makes it particularly valuable for disagreement, workload concerns and information that people may not raise in a group.

Performance management

The continuing process of setting expectations, observing work, giving feedback, developing capability, recognising contribution and addressing shortfalls. Formal ratings are only one possible part.

Culture

The learned assumptions and expected behaviours that tell people how work is done here. Culture is inferred most reliably from recurring decisions and consequences.

Escalation

The deliberate transfer of a problem to a role with greater authority, wider context or relevant expertise because a defined boundary has been crossed. Good escalation carries a question, recommendation and time horizon rather than forwarding confusion.

Go Deeper

The practical manual

Andrew S. Grove, High Output Management (Vintage, 1995). Grove treats management as leverage applied through processes, training, decisions and teams. His treatment of one-to-ones, output and meeting purpose remains unusually usable. Some examples and assumptions belong to late twentieth-century Silicon Valley, and his enthusiasm for measurement needs the safeguards this book has stressed. Read it first if you have recently become a manager and need to turn principles into a working week. The book's strength is operational specificity: Grove asks what meetings are for, how leverage works and why training is a managerial responsibility. Its weakness is the same specificity when readers treat Intel's setting as a universal template.

The classic on effectiveness

Peter F. Drucker, The Effective Executive (Harper Business). First published in 1967, it is less a handbook for supervising staff than an argument about how people with organisational responsibility should use time, contribution, strengths and priorities. The evidence style predates modern organisational research, but the questions remain durable. Read it for the discipline of deciding what deserves attention and what does not. Drucker's language about knowledge work also helps explain why management cannot remain a system in which the boss owns thought and employees supply execution. The examples are dated; the demand for contribution is not. It remains one of the clearest short books on managerial attention.

The reality of managerial work

Henry Mintzberg, Managing (Berrett-Koehler, 2009). Mintzberg resists the tidy picture in which managers plan, organise and control in sequence. He shows managing as a practice across information, people and action, full of interruptions and tensions. It is denser and less immediately prescriptive than Grove. Read it after you have managed long enough to discover that neat frameworks keep colliding with the day.

Voice, learning and authority

Amy C. Edmondson, The Fearless Organization: Creating Psychological Safety in the Workplace for Learning, Innovation, and Growth (Wiley, 2018). Edmondson explains why relevant information stays unspoken and how teams can make candour safer without lowering standards. The phrase psychological safety is now used loosely; the original argument is more demanding. Read it for the management problem that hierarchy is structurally tempted to ignore: getting doubt, error and challenge to travel upward. It is especially useful in healthcare, engineering, technology and other settings where people often discover danger before the person with formal authority does. Pair it with this book's warning that voice also needs decision rights, workload capacity and consequences that are fair when the news is unwelcome.

Notes and Sources

The Whole Thing in One Page and Why You Should Care

Management as coordination and power. The definition used here is a synthesis. Adam Smith supplies the classic division-of-labour problem. Henry Mintzberg and Jay Galbraith are major sources for coordination and organisational design. The power dimension draws on organisational sociology, employee voice research and work on organisational justice rather than treating management as a neutral engineering problem.

Management intervention in India. Nicholas Bloom, Benn Eifert, Aprajit Mahajan, David McKenzie and John Roberts conducted a randomised field experiment with large Indian textile firms. The published Quarterly Journal of Economics paper reports that adopting the management practices raised productivity by 17 per cent in the first year through improved quality and efficiency and reduced inventory. The result belongs to a specific setting and bundled intervention.

Supervisor effects. Edward Lazear, Kathryn Shaw and Christopher Stanton studied technology-based service workers in one company. They estimated that replacing a boss in the bottom 10 per cent of their quality distribution with one in the top 10 per cent increased team output by more than adding another worker to a nine-person team; workers under better bosses were also less likely to leave.

Sources for the seven ideas

Division of labour and coordination. Smith's pin factory appears in Book I, chapter 1 of The Wealth of Nations. Mintzberg's structural work distinguishes direct supervision, mutual adjustment and forms of standardisation. Galbraith's information-processing view explains why uncertainty changes the amount and type of coordination organisations need.

Goals and work bargains. Edwin Locke and Gary Latham summarise evidence that specific, challenging goals can improve performance under conditions including commitment, capability and feedback. Their work also distinguishes learning from performance goals. Research on psychological contracts and organisational support supplies a broader basis for the reciprocal character of the employment relationship.

Decision rights and Toyota. Toyota's current official description of the Toyota Production System identifies Just-in-Time and jidoka as its two pillars. Under jidoka, machines can stop automatically when abnormalities occur and operators can stop the line. The text uses this narrowly as an example of bounded local authority around information that loses value if delayed.

Metrics, multitasking and incentives. Steven Kerr's essay on rewarding one behaviour while hoping for another remains a classic statement of incentive mismatch. Bengt Holmstrom and Paul Milgrom formalised the multitask problem. Edward Lazear's Safelite study reports a 44 per cent increase in output per worker after a move from hourly to piece-rate pay, while separating incentive and selection effects.

Promotion and hiring. Alan Benson, Danielle Li and Kelly Shue analyse promotions in sales organisations and find firms place substantial weight on current sales performance even when other observable traits better predict managerial performance. The United States Office of Personnel Management's current guidance defines structured interviews through standardised, job-related questions and common scoring processes grounded in job analysis.

Motivation and workload. Edward Deci, Anja Olafsen and Richard Ryan review self-determination theory at work, focusing on autonomy, competence and relatedness. Job Demands-Resources research links sustained high demands and insufficient resources to strain while identifying resources such as autonomy and support as important conditions for effective work and well-being.

Voice and psychological safety. Elizabeth Morrison's reviews of employee voice and silence examine why employees speak up or remain silent. Amy Edmondson's work defines team psychological safety around interpersonal risk-taking and learning. The concept is not equivalent to comfort or freedom from accountability.

Feedback. Avraham Kluger and Angelo DeNisi's meta-analysis covered 607 effect sizes and 23,663 observations. Feedback interventions improved performance on average, but more than one-third reduced it. Later reviews continue to find a more conditional evidence base than the simple rule that more feedback is better.

Culture and scale. Edgar Schein's model of culture distinguishes artefacts, espoused beliefs and underlying assumptions. Mintzberg, Galbraith and classic organisational theory supply the trade-offs among functional, divisional and matrix structures. Research on workplace stratification and organisational inequality supports the warning that formal structure also distributes status and opportunity.

Sources for the operating sequence

Meetings and operating rhythm. Andrew Grove's High Output Management, Mintzberg's empirical work on managers and the workplace-meetings research literature inform the distinction among one-to-ones, operating reviews, decision meetings, planning sessions and retrospectives. The meeting categories are a practical synthesis rather than a single published taxonomy.

Selection and onboarding. Current OPM material was rechecked on 9 August 2026. Its structured-interview guidance requires standardised questioning and scoring and recommends job analysis to select competencies. Work-sample use is consistent with the broader personnel-selection literature.

Performance diagnosis. The book's sequence of checking clarity, capability, capacity, authority, tools, incentives and commitment is a synthesis designed to prevent premature character judgement. It should not be read as a validated diagnostic instrument.

Employment action. The manuscript deliberately avoids jurisdiction-specific dismissal law and HR procedure. Formal performance and termination decisions must follow the law and policy applicable to the organisation.

What People Get Wrong and Use It

Hawthorne. The final manuscript removes the Stage 11 treatment from the narrative because the myth was not essential to the title promise. The literature remains useful as a warning against simple stories in management research, but space is better spent on power, workload and voice.

Humane management. The claim is normative as well as empirical. Organisational justice, employee voice, job demands and psychological-contract research support several mechanisms through which fair treatment, support, workload and voice matter. The book does not claim that dignity is justified only by its effect on productivity.

Bibliography

Foundational and interpretive works

Drucker, Peter F. The Effective Executive: The Definitive Guide to Getting the Right Things Done. Revised ed. New York: Harper Business, 2006.

Galbraith, Jay R. Designing Complex Organizations. Reading, MA: Addison-Wesley, 1973.

Grove, Andrew S. High Output Management. New York: Vintage, 1995.

Mintzberg, Henry. Managing. San Francisco: Berrett-Koehler, 2009.

Mintzberg, Henry. The Structuring of Organizations. Englewood Cliffs, NJ: Prentice-Hall, 1979.

Schein, Edgar H., and Peter A. Schein. Organizational Culture and Leadership. 5th ed. Hoboken, NJ: Wiley, 2017.

Smith, Adam. An Inquiry into the Nature and Causes of the Wealth of Nations. Edited by Edwin Cannan. Indianapolis: Liberty Fund, 1981. Originally published 1776.

Research and original evidence

Bakker, Arnold B., Evangelia Demerouti, and Ana Sanz-Vergel. “Job Demands-Resources Theory: Ten Years Later.” Annual Review of Organizational Psychology and Organizational Behavior 10 (2023): 25-53. https://doi.org/10.1146/annurev-orgpsych-120920-053933.

Benson, Alan, Danielle Li, and Kelly Shue. “Promotions and the Peter Principle.” Quarterly Journal of Economics 134, no. 4 (2019): 2085-2134. https://doi.org/10.1093/qje/qjz022.

Bloom, Nicholas, Benn Eifert, Aprajit Mahajan, David McKenzie, and John Roberts. “Does Management Matter? Evidence from India.” Quarterly Journal of Economics 128, no. 1 (2013): 1-51. https://doi.org/10.1093/qje/qjs044.

Deci, Edward L., Anja H. Olafsen, and Richard M. Ryan. “Self-Determination Theory in Work Organizations: The State of a Science.” Annual Review of Organizational Psychology and Organizational Behavior 4 (2017): 19-43. https://doi.org/10.1146/annurev-orgpsych-032516-113108.

Edmondson, Amy. “Psychological Safety and Learning Behavior in Work Teams.” Administrative Science Quarterly 44, no. 2 (1999): 350-383. https://doi.org/10.2307/2666999.

Holmstrom, Bengt, and Paul Milgrom. “Multitask Principal-Agent Analyses: Incentive Contracts, Asset Ownership, and Job Design.” Journal of Law, Economics, and Organization 7 (1991): 24-52.

Kerr, Steven. “On the Folly of Rewarding A, While Hoping for B.” Academy of Management Journal 18, no. 4 (1975): 769-783.

Kluger, Avraham N., and Angelo DeNisi. “The Effects of Feedback Interventions on Performance: A Historical Review, a Meta-Analysis, and a Preliminary Feedback Intervention Theory.” Psychological Bulletin 119, no. 2 (1996): 254-284.

Latham, Gary P., and Edwin A. Locke. “New Developments in and Directions for Goal-Setting Research.” European Psychologist 12, no. 4 (2007): 290-300.

Lazear, Edward P. “Performance Pay and Productivity.” American Economic Review 90, no. 5 (2000): 1346-1361.

Lazear, Edward P., Kathryn L. Shaw, and Christopher T. Stanton. “The Value of Bosses.” Journal of Labor Economics 33, no. 4 (2015): 823-861.

Morrison, Elizabeth W. “Employee Voice and Silence.” Annual Review of Organizational Psychology and Organizational Behavior 1 (2014): 173-197. https://doi.org/10.1146/annurev-orgpsych-031413-091328.

Morrison, Elizabeth W. “Employee Voice and Silence: Taking Stock a Decade Later.” Annual Review of Organizational Psychology and Organizational Behavior 10 (2023): 79-107. https://doi.org/10.1146/annurev-orgpsych-120920-054654.

Current institutional sources

Toyota Motor Corporation. “Toyota Production System.” Official Global Website. Checked 9 August 2026.

United States Office of Personnel Management. “Structured Interviews.” Assessment and Selection guidance. Checked 9 August 2026.