Principles of Management

The big picture

On Monday you were the best barista in the store. On Friday your district manager promotes you to shift supervisor, and suddenly the job is unrecognizable. The Saturday schedule has a hole in it, a customer wants a refund your till cannot process, the new hire is steaming milk wrong, the grinder is making a noise grinders should not make, and somewhere in there you are still expected to pull shots. Nothing in your barista skills tells you which of these problems to touch first. You have just discovered, the hard way, that management is not senior barista work. It is a different job.

This course is about that job. Management is getting things done through other people by using an organization's resources effectively and efficiently: effectively meaning the right goals get reached, efficiently meaning without wasting time, money, or effort. The definition sounds bland until you notice the phrase "through other people." Every difficulty in this course flows from it. Machines do what you set them to do. People have their own goals, information, moods, and opinions of you, and yet nearly everything modern life depends on, from vaccines to grocery deliveries, requires coordinating many of them at once.

One honest note before we start. A text course can give you frameworks, the evidence for and against them, and worked practice on realistic cases. It cannot give you the sweaty-palmed experience of your first real feedback conversation. Treat every lesson here as flight simulation: genuinely useful, far cheaper than crashing, and not the same as flying.

This lesson maps managerial work three ways: by function, meaning what management accomplishes; by role, meaning what the day actually looks like; and by level and skill, meaning how the job changes as you rise from shift supervisor to chief executive.

The four functions: a map of the whole course

The oldest and still most useful map of management comes from Henri Fayol, a French mining executive whose 1916 book distilled decades of running a large industrial company. Textbooks have compressed his list into four functions, and this course is organized around them:

• Planning: setting goals and deciding how to reach them. Module 2 covers goals, strategy, and the decisions behind both.
• Organizing: arranging people, tasks, and authority so the plan can actually happen. Module 3 covers structure, culture, and hiring.
• Leading: motivating and influencing people to do the work willingly and well. Module 4 covers motivation, leadership, teams, and communication.
• Controlling: measuring results against the plan and correcting course. Module 5 covers metrics, operations, and change.

Two warnings about this tidy list. First, it describes purposes, not a schedule. No manager spends Monday planning and Tuesday organizing; a single ten-minute conversation about a late shipment can involve all four functions at once. Second, the functions form a cycle, not a line. What controlling reveals feeds the next round of planning, which is why organizations that never look honestly at their results keep writing the same plan every year.

Key idea: Management is the work of planning, organizing, leading, and controlling: four purposes that repeat as a cycle rather than four tasks done in order.

What Mintzberg saw when he actually watched managers

Here is the picture most people carry around: the manager as reflective planner, sitting in a quiet office, studying reports, thinking long strategic thoughts. In the early 1970s a young Canadian researcher named Henry Mintzberg tested that picture in the simplest way imaginable: he followed five chief executives around, week after week, and wrote down everything they did. His 1973 book The Nature of Managerial Work, and a famous 1975 Harvard Business Review article summarizing it, reported what he found, and the findings demolished the folklore.

Real managerial work ran at an unrelenting pace and was chopped into fragments. Half of the executives' activities lasted less than nine minutes, and only about one in ten lasted more than an hour. The managers were interrupted constantly, jumped between unrelated issues, and showed a strong preference for live conversation, phone calls, and drop-ins over long written reports. They were not calm planners working through a system; they responded to a stream of events, and the effective ones had learned to do their thinking inside the fragments rather than waiting for a quiet stretch that never came. If your mental image of managing is a clean desk and a Gantt chart, replace it now with a hallway conversation that gets interrupted twice.

Out of thousands of observed episodes, Mintzberg distilled ten roles that managers play, in three families:

• Interpersonal roles: figurehead (representing the unit at ceremonies, signings, and visits), leader (staffing, motivating, and developing people), and liaison (maintaining a web of contacts outside the unit).
• Informational roles: monitor (scanning for information), disseminator (passing information inward to the team), and spokesperson (representing the unit's information to outsiders).
• Decisional roles: entrepreneur (initiating improvements), disturbance handler (dealing with the grinder that just broke), resource allocator (deciding who gets budget, time, and attention), and negotiator (working out agreements).

Watch a store manager for one morning and you can label every move: greeting the visiting regional vice president (figurehead), coaching the new hire (leader), calling the manager of the neighboring store about a shared parking problem (liaison), reading overnight sales (monitor), briefing the team on a price change (disseminator), answering a local reporter's question (spokesperson), proposing a new floor layout (entrepreneur), handling the broken grinder (disturbance handler), setting the Saturday schedule (resource allocator), and talking the milk supplier down on price (negotiator).

How do the ten roles square with the four functions? They are two views of one job. The functions describe what management accomplishes over weeks and months; the roles describe what the work looks like hour by hour. Planning rarely shows up as a visible role because planning happens inside the roles, in fragments, between interruptions. Understanding both views keeps you from the two classic errors: believing the job is all quiet strategy, or concluding from the chaos that no strategy exists.

Key idea: Observed managerial work is fast, fragmented, and conversational, and Mintzberg's ten roles describe the hour-by-hour activity through which the four slower-moving functions get done.

Levels and skills: how the job changes as you rise

Management jobs differ more by level than by industry. Three levels cover most organizations:

Level	Typical titles	Time horizon	Core question
First-line	Shift supervisor, team lead, crew chief	Days to weeks	Is today's work getting done well?
Middle	Store manager, plant manager, regional director	Months to a year or two	How do I turn strategy into results in my unit?
Top (C-suite)	CEO, CFO, COO, chief officers generally	Years to a decade	What should this organization become?

In a 1955 Harvard Business Review article, reprinted for decades afterward, Robert Katz argued that effective managers need three skills in proportions that shift with level. Technical skill is command of the actual work: pulling shots, closing the books, writing code. Human skill is the ability to work with and through people: reading a room, giving feedback that lands, building trust. Conceptual skill is the ability to see the organization as a whole system and understand how a change in one part ripples through the rest. First-line managers lean hardest on technical skill, because they train and troubleshoot daily work. Top managers lean hardest on conceptual skill, because their decisions are about the whole. Human skill stays large at every level, which is Katz's most durable finding.

This explains one of the most common career accidents in business. The best engineer on the team is promoted to engineering manager because she is the best engineer, which is precisely the skill the new job needs least. Unless she builds human and conceptual skills, the organization has simultaneously lost its best engineer and gained a struggling manager. The remedy is not to stop promoting strong performers. It is to treat a first management job as entry into a new profession, with training, coaching, and honest feedback, rather than as a larger version of the old job.

While we are at this table, a word in defense of its middle row. Mocking middle managers is an old sport, and every few years a company announces it is deleting the layer entirely. The evidence suggests caution. One study of the video game industry by the researcher Ethan Mollick matched hundreds of games to the producers and designers who made them and found that the identity of the middle manager, the producer, explained more of the variation in a game's revenue than the identity of the creative designers. Coordinating specialists, translating strategy into assignments, and unblocking work is real production, even when it leaves no visible fingerprints on the product.

Key idea: As you rise, the skill mix shifts from technical toward conceptual, human skill never stops mattering, and promotion into management is a change of profession, not a reward level.

Does management actually matter?

A fair skeptic could ask: maybe firms succeed on products, prices, and luck, and managers just ride along. This is an empirical question, and it has been tested. Since 2004, the World Management Survey, run by the economists Nicholas Bloom, John Van Reenen, and colleagues, has scored thousands of firms in dozens of countries on eighteen basic practices: does the firm track performance, set targets, fix problems when they appear, promote on merit? The pattern is consistent. Firms that score higher are more productive, more profitable, grow faster, and survive longer, and the spread in scores is enormous, including within the same industry and country.

Correlation invites the obvious objection, so the same team ran an experiment. In a randomized trial published in 2013, large Indian textile plants were randomly assigned to receive intensive consulting on those basic practices or to serve as controls. Treated plants raised productivity by about seventeen percent in the first year, largely through quality and inventory improvements, and the gains persisted for years. Basic management practice, taught and adopted, caused better performance. That is about as close to a clean answer as social science gets.

Even the skeptics' home turf produced the same verdict. Google, founded by engineers who openly doubted the value of managers, ran an internal study starting in 2008 called Project Oxygen to test whether manager quality showed up in team outcomes. It did: teams with better-rated managers performed better and stayed longer, and the behaviors that mattered most were coaching, empowering rather than micromanaging, and caring about people's success and wellbeing. A company that would have been delighted to prove managers unnecessary concluded the opposite and built manager training around the findings. And the labor market prices the work accordingly: the U.S. Bureau of Labor Statistics reports that management occupations carry the highest median annual wage of any major occupational group, well above one hundred thousand dollars.

Key idea: Management quality is measurable, and both large observational datasets and a randomized experiment show that adopting basic management practices raises productivity: the job matters.

Try it

Here are six entries from the Tuesday log of Maya, who manages a fourteen-person bicycle shop. (1) At 8:05 she walks the floor and checks yesterday's sales against target. (2) At 8:40 she meets a supplier rep and presses for faster delivery of a back-ordered cargo bike. (3) At 10:15 she calms an upset customer whose repair went wrong and authorizes a free fix. (4) At 11:00 she interviews a mechanic candidate. (5) At 1:30 she emails the team a summary of the new e-bike rebate rules. (6) At 3:00 she sketches a plan to add a mobile repair van next spring. Label each entry with a Mintzberg role, and say which of the four functions items 1 and 6 serve.

Answer: (1) Monitor, and because it compares results to a target it is the controlling function. (2) Negotiator. (3) Disturbance handler. (4) Leader, since staffing and developing the unit belong to that role. (5) Disseminator. (6) Entrepreneur, and as goal-setting for the future it is the planning function. Notice that one ordinary Tuesday touched six of the ten roles before 3:30 in the afternoon.

Common misconceptions

• "A manager is just a senior worker." Management is a different job with different skills. The promotion that treats it as a reward for technical excellence produces struggling managers.
• "Managers spend their days in quiet planning." Mintzberg's data show the opposite: fragments, interruptions, and constant conversation, with half of activities lasting under nine minutes.
• "Management and leadership are the same thing." Leading is one of the four functions. A manager who only inspires but never plans, organizes, or controls is failing three quarters of the job.
• "Middle management is bureaucratic fat." Evidence from project-based industries shows individual middle managers account for large differences in results.
• "Management is just common sense." If it were, the Indian textile plants in the 2013 experiment would already have been using basic quality and inventory practices. They were not, and adopting them raised productivity seventeen percent.
• "You need the title before you can practice." The monitor, liaison, and entrepreneur roles are available to anyone: track how your team is doing, build contacts, propose improvements.

Recap

• Management is getting things done through people, judged by effectiveness (right goals) and efficiency (little waste).
• The four functions, planning, organizing, leading, and controlling, form a repeating cycle and the map of this course.
• Observed managerial work is fast, fragmented, and conversational; Mintzberg's ten roles in three families describe it.
• First-line, middle, and top managers differ mainly in time horizon, and the skill mix shifts from technical toward conceptual as you rise.
• Human skill matters at every level, and a first management job is a new profession, not a bigger version of the old one.
• Large surveys, a randomized trial in Indian factories, and Google's own Project Oxygen all show that management quality measurably drives performance.

Sources

1. Bright, D. S., & Cortes, A. H. (2019). What do managers do? In Principles of Management. OpenStax, Rice University. openstax.org
2. Mintzberg, H. (1975). The manager's job: Folklore and fact. Harvard Business Review, 53(4), 49-61. hbr.org
3. Katz, R. L. (1974). Skills of an effective administrator. Harvard Business Review, 52(5), 90-102. hbr.org
4. Bloom, N., Eifert, B., Mahajan, A., McKenzie, D., & Roberts, J. (2013). Does management matter? Evidence from India. The Quarterly Journal of Economics, 128(1), 1-51. nber.org
5. U.S. Bureau of Labor Statistics. (2025). Management occupations. In Occupational Outlook Handbook. bls.gov

The big picture

Every few years a book tour announces a revolutionary new way to manage: self-managing teams, radical transparency, algorithmic productivity tracking. Managers who know no history buy the book twice. Managers who know some history recognize the product: productivity tracking is Frederick Taylor with better sensors, and self-managing teams are the human relations movement with better branding. A century of management thought is not a museum. It is a parts catalog from which every current fashion is assembled, and knowing it inoculates you against both blind enthusiasm and blind rejection.

There is a second reason to study the history critically. Management ideas were forged in real workplaces, on real workers, and some of the forging hurt. Taylor's methods provoked a strike and a congressional investigation. Ford's five-dollar day came bundled with home inspections. The Hawthorne studies, the most cited experiments in management, showed something different from what most textbooks say they showed. Taking the history seriously, costs included, is how you learn to ask the right questions about the next big idea that lands on your desk.

This lesson walks the main line: scientific management, administrative and bureaucratic theory, the Hawthorne studies and human relations, and the systems and contingency thinking that frames everything modern.

Scientific management: real gains, real costs

Frederick Winslow Taylor, born to a wealthy Philadelphia family in 1856, skipped Harvard for a machine shop and worked his way up at Midvale Steel from laborer to chief engineer. On the shop floor he saw what he called soldiering: workers deliberately pacing themselves well below capacity, partly because faster work earned nothing but a cut piece rate, partly because nobody had ever measured what a fair day's work was. His answer, published in 1911 as The Principles of Scientific Management, was to make work an object of study. Break each job into elements, time them with a stopwatch, eliminate wasted motion, standardize the one best way, select and train workers scientifically, pay more for hitting the standard, and split responsibility: managers plan the work, workers execute it.

The gains were real. At Bethlehem Steel, Taylor claimed his methods raised a pig iron loader's daily tonnage from about 12.5 to 47 long tons for a 60 percent raise in pay, and his shoveling studies cut the yard workforce dramatically while raising wages for those who remained. Frank and Lillian Gilbreth, the movement's most sympathetic figures, used motion study to cut the motions in bricklaying from eighteen to about five, and Lillian, a psychologist, pushed the movement to consider fatigue and worker welfare. Henry Ford's moving assembly line, launched at Highland Park in 1913, applied the same logic to flow: chassis assembly time fell from over twelve hours to around ninety minutes.

Now the costs, which the record documents just as clearly. Taylor's famous pig iron story was embellished: historians who checked the archives found the character he called Schmidt was reconstructed after the fact and the numbers were tidied. Workers experienced time study as surveillance and speed-up, and the deskilling was not a side effect but the design: planning moved upstairs, and the worker's judgment was explicitly unwanted. In 1911, molders at the Watertown Arsenal walked out when a stopwatch appeared, triggering a House investigation before which Taylor testified in 1912, and Congress for decades attached riders banning stopwatch time studies in Army arsenals and Navy yards. Ford's line worked so hard on workers that in 1913 the company had to hire roughly 52,000 people to keep about 14,000 positions filled. The famous five-dollar day of January 1914, roughly double the going wage, was less generosity than a turnover fix, and half the money was profit sharing conditional on passing home inspections by Ford's Sociological Department: scientific management extended to your living room.

Key idea: Scientific management delivered genuine productivity gains by studying work systematically, and it did so by transferring judgment from workers to managers, a bargain workers often experienced as surveillance and speed-up, and sometimes fought.

Fayol and Weber: managing the whole, taming favoritism

While Taylor studied the shop floor from below, Henri Fayol described the enterprise from the top. Fayol spent his entire career at a French mining and metals firm, thirty years of it as chief executive, and rescued the company from near failure. His 1916 book argued that administration itself, distinct from engineering or finance, was a teachable discipline with five elements, planning, organizing, commanding, coordinating, and controlling, and fourteen principles, including unity of command (one boss per worker), fair remuneration, equity, and esprit de corps. Where Taylor optimized tasks, Fayol optimized the organization. The four functions organizing this course are his list, lightly edited by a century of textbooks.

The German sociologist Max Weber, writing around the same period, gave a name to the organizational form the modern world was converging on: bureaucracy. His ideal type had a clear division of labor, a hierarchy of offices, written rules and records, managers appointed on qualifications rather than connections, and impersonal application of the rules. Before you sneer at the word, remember what bureaucracy replaced: jobs handed to nephews, decisions varying with the official's mood, no records, no appeal. Rules that bind the powerful were a reform. Weber himself supplied the warning label, worrying that rule-following could become an iron cage in which the means devour the ends. Both halves were right, which is why you will spend part of Module 3 learning when to formalize and when formalization strangles.

Key idea: Fayol made managing the whole organization a teachable discipline, and Weber's bureaucracy, rules, hierarchy, and merit, was an anti-favoritism technology whose failure mode, rule-worship, its own inventor predicted.

Hawthorne: what the most famous studies actually showed

From 1924 to 1932, at Western Electric's Hawthorne Works outside Chicago, researchers ran the most cited and most misreported studies in management history. The first round tested lighting: raise illumination for a test group and output rose, but it also rose in the control group, and in one run output held up even as lights dimmed toward moonlight levels. Something other than lighting was moving the numbers. In the relay assembly test room, five women assembling telephone relays were moved to a separate room and observed for five years while researchers varied rest breaks, hours, and pay incentives. Output climbed substantially over the period, seemingly regardless of which condition was in force, and the Harvard researcher Elton Mayo drew the conclusion that made the studies famous: workers are not isolated machines responding to physical conditions; they are members of social groups whose attention, cohesion, and sense of being valued drive output.

That conclusion launched the human relations movement, and a simplified version hardened into the textbook Hawthorne effect: people perform better simply because they are being observed. Here is where you should slow down, because the record is messier and more interesting. In the relay room, a group piece-rate incentive was introduced along with the observation, two of the five workers were replaced partway through for low output and talking, the Great Depression arrived to concentrate minds on keeping one's job, and skills improved with five years of practice. Any of those could move output. When the economists Steven Levitt and John List recovered the original illumination data decades later, they found the evidence for the classic effect weak: much of the pattern traced to mundane artifacts, like output rising on Mondays whether or not lighting changed. Meanwhile the least famous Hawthorne study may be the most useful: in the bank wiring observation room, fourteen men on a group incentive enforced an informal output ceiling, mocking rate-busters who worked too fast and chiselers who worked too slow. The group, not the pay scheme, set the pace.

So what survives? Not a magic observation effect. What survives is the founding insight of organizational behavior: a workplace is a social system, informal groups develop norms that can beat formal incentives, and how people are treated changes what they produce. That is worth keeping, and it does not require the legend.

Key idea: The Hawthorne studies are shakier than their legend, confounded by incentives, replacements, and fear of the Depression, but their durable finding stands: workplaces are social systems in which informal group norms regulate output.

Systems and contingency: the end of the one best way

The Second World War pulled mathematicians into logistics and gave management a quantitative school: operations research, inventory models, statistical quality control, the ancestors of every dashboard you will ever read. But the deeper postwar shift was conceptual. Systems thinking reframed the organization as an open system: it takes inputs from an environment, transforms them, returns outputs, and lives or dies by the exchange. Subsystems interlock, so a fix in one department can break another, and a company that stops listening to its environment, customers, suppliers, regulators, is a closed system running down.

From there it was a short step to the finding that ended a half-century of hunting for the one best way. Joan Woodward, studying about a hundred English manufacturing firms in the 1950s, found that the structure that worked depended on the production technology: successful small-batch and continuous-process firms were organized differently from successful mass producers. Burns and Stalker found mechanistic, rule-heavy structures thrived in stable environments while looser, organic structures thrived in changing ones. This is contingency thinking: the effectiveness of a practice depends on the situation. It sounds like a cop-out and is actually a discipline, because it obliges you to name the contingencies: what technology, what environment, what size, what strategy. Nearly every lesson that follows, on structure, leadership, control, will end in a contingency claim, and now you know why.

One more pattern worth carrying. Scholars who track management fashion, such as Barley and Kunda, note that ideas swing like a pendulum between rational control (Taylor, reengineering, analytics) and normative commitment (human relations, culture, engagement), with each swing sold as brand new. When the next revolution is announced, ask three questions: what problem does it claim to solve, what is the evidence beyond testimonials, and what does it cost if it is wrong here? Those questions are this lesson in portable form.

Key idea: Systems thinking sees organizations as open systems of interlocking parts, and contingency thinking replaces the one best way with a disciplined 'it depends,' the stance this whole course takes.

Try it

Match each modern practice to its closest historical ancestor, and note one caution the history suggests: (a) warehouse software that times each picker's route to the second; (b) a company org chart with one boss per employee and written job descriptions; (c) a firm that invests heavily in team lunches, recognition, and belonging surveys; (d) a consultant who answers every structural question with 'it depends on your environment.'

Answer: (a) Taylor's time study; caution: workers experience second-by-second timing as surveillance, and Watertown showed they may organize against it. (b) Fayol's unity of command and Weber's bureaucracy; caution: rules that outlive their purpose become the iron cage. (c) The human relations movement out of Hawthorne; caution: social goodwill cannot substitute for fair incentives, and informal groups may still cap output. (d) Contingency thinking; caution: 'it depends' is only useful when the consultant names the specific contingencies and the evidence.

Common misconceptions

• "Taylor was simply a villain." His methods raised real output and real wages, and the Gilbreths used the same tools to reduce fatigue. The record supports critique, not cartoon.
• "Taylor was simply a genius ahead of his time." He embellished his most famous case, treated worker judgment as a defect, and provoked a strike and a congressional investigation. Both halves are the history.
• "Hawthorne proved that watching workers boosts output." The studies were confounded by pay changes, replacements, practice, and the Depression, and reanalysis of the illumination data found weak effects. The durable finding is about social systems, not observation magic.
• "Bureaucracy just means inefficiency." Weber's bureaucracy was a merit-and-rules reform against favoritism. Its pathologies are real, but so is what it fixed.
• "Each new school replaced the previous one." They layer, not replace. Your employer almost certainly runs Taylorist metrics, a Weberian hierarchy, human-relations engagement programs, and contingency rhetoric simultaneously.
• "Management history is trivia." It is pattern recognition. Most new fashions are old ideas with new packaging, and the old failures come included.

Recap

• Taylor's scientific management studied work systematically, raised productivity and pay, and centralized judgment in management, at documented human cost.
• Ford's assembly line scaled the logic; the five-dollar day was largely a fix for catastrophic turnover, with paternalist strings attached.
• Fayol defined administration as a teachable discipline; his five elements became the four functions.
• Weber's bureaucracy, rules, hierarchy, merit, was designed to defeat favoritism, and he foresaw its iron-cage failure mode.
• The Hawthorne studies were confounded and their legend overstated, but they established that informal social groups regulate output.
• Systems thinking treats organizations as open systems; contingency thinking makes 'it depends, on these named factors' the honest core of management knowledge.

Sources

1. Encyclopaedia Britannica. (2024). Frederick W. Taylor. britannica.com
2. Bright, D. S., & Cortes, A. H. (2019). Administrative and bureaucratic management. In Principles of Management. OpenStax, Rice University. openstax.org
3. Bright, D. S., & Cortes, A. H. (2019). Principles of Management, Chapter 3: The history of management. OpenStax, Rice University. openstax.org
4. Wikipedia. (2025). Hawthorne effect. en.wikipedia.org
5. Wikipedia. (2025). Scientific management. en.wikipedia.org

The big picture

Ask ten managers what their team's goal is and you will get answers ranging from "grow the business" to a spreadsheet with forty metrics. Both are failures of the same kind. The first cannot be acted on, because no one knows what to do differently on Tuesday morning. The second cannot be acted on either, because forty priorities is the same as none. Somewhere between the slogan and the spreadsheet sits the work of planning, and it is the function most managers say they never have time for, right up until they lose a quarter to activity that pointed nowhere.

Planning is deciding what you are trying to achieve and how you will get there, before events force the choice on you. It gives direction, forces trade-offs into the open while they are still cheap, coordinates people who would otherwise optimize their own corner, and creates the benchmark without which the controlling function has nothing to measure against. It also has a bad reputation among people who have watched thick annual plan documents go straight into a drawer, and that reputation is partly earned. This lesson teaches planning that survives contact with reality: how goals cascade, what the evidence says about goal setting, where goals cause damage, and how to plan under genuine uncertainty.

The planning hierarchy: from mission to Tuesday

Plans exist at different altitudes, and confusing them is the most common planning error.

Level	Who owns it	Horizon	Example (a regional grocery chain)
Mission and vision	Top management, board	Indefinite	Mission: affordable fresh food for working families. Vision: the region's most trusted grocer by 2032.
Strategic plan	Top management	2-5 years	Grow from 20 to 30 stores, and reach 15 percent of sales from prepared meals.
Tactical plan	Middle management	6-24 months	Open four stores this year in named suburbs; launch prepared meals in the ten largest stores.
Operational plan	First-line management	Days to months	Weekly staffing, delivery schedules, waste targets per department.

Two mechanisms are worth naming. Means-ends chains connect the levels: what is an end at one level is a means at the level above, so the store manager's staffing plan is the means to the district's prepared-meals launch, which is the means to the strategic growth goal. When people cannot describe that chain for their own work, the plan has broken somewhere between the boardroom and the floor. The second mechanism is the difference between standing plans, which handle recurring situations (policies, procedures, rules such as how to process a return), and single-use plans, which handle one-off efforts (a store opening, a software migration). Standing plans exist so that recurring decisions do not consume management attention twice. If your team keeps rediscussing the same question, you are missing a standing plan.

Key idea: Plans operate at four altitudes linked by means-ends chains, and the test of a plan is whether a front-line worker can name what it changes about their week.

SMART goals, and the evidence underneath them

The most durable goal-writing checklist is SMART: specific, measurable, achievable, relevant, and time-bound. Compare two versions of the same intention. "Improve customer service" fails every letter. "Cut the average phone hold time in our support centre from 4 minutes 40 seconds to under 2 minutes by March 31, without raising the abandoned-call rate above 5 percent" passes all five and, more importantly, tells a supervisor what to change on Monday: staffing at peak hours, call routing, or first-contact resolution.

SMART is a mnemonic, not a theory, and it is worth knowing the actual research it borrows from. Starting in the late 1960s, Edwin Locke and Gary Latham ran hundreds of studies on goal setting, and the core findings are among the most robust in organizational psychology. Specific, difficult goals produce higher performance than vague exhortations to do your best, and the effect is large. But the finding comes with four moderators that the mnemonic hides. Goals raise performance only when the person has the ability to do the task, is committed to the goal (which usually requires either participation in setting it or a credible explanation), receives feedback on progress, and is working on a task that is not too complex or novel. On genuinely new, complex work, a hard outcome goal can actually hurt, because attention goes to hitting a number instead of learning how the task works. For that kind of work, learning goals ("test three approaches and document what fails") beat performance goals.

Notice too the quiet tension inside SMART itself. The A stands for achievable, while the evidence says difficult goals drive performance. The resolution most organizations use is that goals should be difficult but not delusional, since a goal seen as impossible destroys commitment, which is one of the four moderators. That is a judgment call, not a formula, and pretending otherwise is how stretch goals become fiction.

Key idea: Specific, difficult goals genuinely raise performance, but only with ability, commitment, feedback, and manageable task complexity, and on novel work learning goals beat outcome goals.

When goals go wrong

Here is the part most courses skip. Goals are a powerful intervention, and powerful interventions have side effects. A well-known critique by Lisa Ordóñez and colleagues catalogued them: goals narrow attention so that unmeasured but important work is dropped, they encourage risk-taking near the threshold, they crowd out intrinsic motivation, they corrode ethics, and they can poison cooperation when set individually in interdependent work.

The examples are not hypothetical. In 1992, California regulators investigated Sears Auto Centers after mechanics were given sales quotas for repairs; investigators found customers were being sold parts and services they did not need, and Sears dropped the commission-based quota system, its chairman conceding the goals had created an environment where mistakes occurred. The mechanics were not unusually dishonest. They were ordinary people given a specific, difficult, measurable, time-bound goal with a paycheck attached, and no counterweight. You will meet the same mechanism again in Module 5 under Goodhart's law and the Wells Fargo account-opening scandal.

Three practical counterweights help. First, pair every output goal with a guardrail metric that must not degrade: hold time down and abandoned calls capped, sales up and returns or complaint rates flat. Second, keep the count of goals small, because a goal set that covers everything is a to-do list, and one that covers only what is countable will quietly instruct people to stop doing the rest. Third, ask before launch: if someone wanted to hit this number without doing the underlying work, how would they do it? Then decide whether you can live with that path being available.

Key idea: Goals reliably change behavior, including behavior you did not want, so every serious goal needs a guardrail metric and a deliberate answer to how it could be gamed.

MBO and OKRs: two ways to cascade goals

Peter Drucker introduced management by objectives in 1954: managers and employees jointly set specific objectives for a period, agree how they will be measured, and review results against them, with rewards tied to achievement. The strengths are participation, which builds commitment, and clarity. The classic failure modes are bureaucratic bloat, the annual cycle being too slow for fast-moving work, and, because rewards ride on the numbers, employees negotiating for easy objectives.

Objectives and key results, developed by Andy Grove at Intel and brought to Google in 1999 by John Doerr, is a descendant with three deliberate changes: quarterly cycles instead of annual, public visibility of everyone's OKRs, and, in most implementations, a firm separation from compensation so that people can set ambitious targets without betting their bonus. The format is one qualitative objective plus two to five quantitative key results.

Here is a correct one for our grocery chain's prepared-meals push. Objective: make prepared meals a real reason to shop with us. Key results: raise prepared-meal sales from 4 percent to 9 percent of store revenue in the ten pilot stores; keep prepared-meal waste under 6 percent of production; reach a repeat-purchase rate of 30 percent among loyalty members who buy a meal. And here is the near-universal beginner error: writing a key result as "launch prepared meals in ten stores." That is a task. Key results describe outcomes, not activities, because a team can complete every task on a list and change nothing about the business. If your key results all begin with a verb like launch, build, or hire, you have written a project plan and labeled it a goal.

Key idea: MBO and OKRs both cascade goals through participation, but OKRs run quarterly, publicly, and usually apart from pay, and their discipline is that key results state outcomes rather than activities.

Planning when the future refuses to cooperate

Every plan rests on assumptions, and the useful question is not whether they are right but which ones would hurt most if wrong. Contingency planning answers that with pre-agreed responses: if a key supplier fails, we switch to the qualified second source; if flu season strips 20 percent of staff, these services pause first. Scenario planning goes further, building two to four internally consistent stories about the future and testing the strategy against each. Royal Dutch Shell famously used the technique in the early 1970s to explore a world of sharply higher oil prices, and when the 1973 oil shock arrived, managers who had already thought through that world reacted faster than rivals who had not. The point of scenarios is not prediction. It is rehearsal, so that the response is not being invented under stress.

Finally, correct for the single most reliable planning error. The planning fallacy, named by Daniel Kahneman and Amos Tversky, is our tendency to underestimate the time, cost, and risk of our own projects while knowing perfectly well that similar projects overrun. The Sydney Opera House was estimated at about seven million Australian dollars with a 1963 completion date and opened in 1973 at roughly 102 million. The correction is unglamorous and effective: take the outside view. Instead of building the estimate from your optimistic step-by-step plan, ask how long comparable projects actually took, and start there. If your last four store openings ran an average of five weeks late, the honest plan for the fifth is not four weeks early because this time you have a better checklist.

Key idea: Plans should carry pre-agreed contingencies and be tested against scenarios, and estimates should be anchored on how similar past efforts actually went rather than on the optimistic internal view.

Try it

A gym chain sets this goal for its twelve locations: "Dramatically improve member retention this year." Current annual member churn is 42 percent on a base of 9,000 members, and each lost member costs about $540 a year in revenue. (a) Rewrite the goal in SMART form, choosing a defensible target. (b) Compute the revenue at stake if churn falls to 34 percent. (c) Name one guardrail metric and explain what gaming it prevents. (d) Turn your goal into an objective with two key results.

Answer: (a) For example: "Cut annual member churn from 42 percent to 34 percent across all twelve locations by December 31, without reducing class capacity." (b) An 8 percentage point reduction on 9,000 members retains 0.08 x 9,000 = 720 members, worth 720 x $540 = $388,800 a year. (c) A guardrail such as cancellations-denied complaints or average class satisfaction. Without it, a location could cut measured churn by making cancellation difficult, which lowers the number while worsening the business. (d) Objective: make membership something people renew without thinking. Key results: churn down from 42 percent to 34 percent; share of members attending at least four times a month up from 38 percent to 50 percent.

Common misconceptions

• "A goal is a plan." A goal is a desired outcome. It becomes a plan only when it has actions, owners, resources, and dates attached.
• "SMART goals are backed by research." The evidence supports specific, difficult goals with feedback and commitment. SMART is a memory aid built loosely on that work, and its A pulls against the evidence for difficulty.
• "Stretch goals always raise performance." Only when people believe the goal is reachable and have the ability and feedback to pursue it; otherwise commitment collapses or corners get cut.
• "Goals are ethically neutral." Sears Auto Centers in 1992 is one of several documented cases where quotas produced customer harm from ordinary employees.
• "OKRs are just MBO with new jargon." The differences are structural: quarterly cadence, public visibility, and separation from compensation, all aimed at MBO's known failure modes.
• "Planning is pointless because the future is uncertain." Uncertainty is the argument for contingency and scenario planning, not against planning. The plan's value is largely in the thinking it forces before the pressure arrives.
• "Our project is different, so past overruns do not apply." That belief is the planning fallacy in one sentence. The outside view, what similar projects actually cost, beats the inside view.

Recap

• Planning gives direction, forces trade-offs, coordinates effort, and creates the standard that controlling measures against.
• Mission and vision, strategic, tactical, and operational plans link through means-ends chains; standing plans handle recurring decisions once.
• SMART is a useful checklist; the underlying evidence is that specific difficult goals work given ability, commitment, feedback, and manageable complexity.
• Goals have documented side effects: narrowed attention, gaming, and ethical corrosion, so pair them with guardrails and few in number.
• MBO cascades goals through joint objective setting; OKRs revise it with quarterly cycles, transparency, outcome-based key results, and distance from pay.
• Contingency and scenario planning prepare responses in advance, and the outside view corrects the planning fallacy.

Sources

1. Bright, D. S., & Cortes, A. H. (2019). Management by objectives: A planning and control technique. In Principles of Management. OpenStax, Rice University. openstax.org
2. Bright, D. S., & Cortes, A. H. (2019). Principles of Management, Chapter 17: Organizational planning and controlling. OpenStax, Rice University. openstax.org
3. Wikipedia. (2025). Goal-setting theory. en.wikipedia.org
4. Wikipedia. (2025). Objectives and key results. en.wikipedia.org
5. Wikipedia. (2025). Planning fallacy. en.wikipedia.org
6. Encyclopaedia Britannica. (2024). Peter Drucker. britannica.com

The big picture

Read enough corporate strategy documents and you will notice most of them contain no strategy. They contain ambition ("become the leading provider"), values ("we put customers first"), and financial targets ("grow revenue 12 percent"), which are respectively a wish, a virtue, and a goal. A strategy is the answer to a harder question: given what we are up against, where will we compete, and what will we do differently that rivals cannot easily copy? Richard Rumelt puts the structure of a real strategy in three parts, a diagnosis of the challenge, a guiding policy for dealing with it, and a coherent set of actions that carry the policy out. Most of what fails that test fails at the first part, because facing the actual challenge is uncomfortable and writing aspirational language is not.

Strategy is also where students of management most often mistake vocabulary for skill. Anyone can name the five forces; the work is applying them to a specific industry with specific evidence and then acting on what the analysis says. This lesson does that on a real industry, and it treats each tool as an instrument with a calibration range rather than an oracle.

The environment: Porter's five forces on airlines

In 1979, Michael Porter proposed that the average profitability of an industry is set by five structural forces, not by how hard its managers work. Learn them by watching them explain something famous: why passenger airlines, an industry of extraordinary engineering and enormous revenue, have historically been so bad at making money that Delta filed for bankruptcy protection in 2005, United in 2002, and American in 2011, and Pan Am and TWA disappeared entirely.

Force	Question it asks	Airlines: the evidence	Strength
Rivalry among existing firms	How intense is competition for the same customers?	High fixed costs and perishable inventory: an empty seat at takeoff is worth zero forever, so carriers discount to fill it. Fares are visible to everyone in seconds.	Very high
Bargaining power of buyers	Can customers force prices down?	Comparison sites made price the dominant choice factor for leisure travel, and switching airlines costs a traveler almost nothing.	High
Bargaining power of suppliers	Can input providers capture the profit?	Aircraft come from essentially two large-jet makers, fuel is a volatile commodity nobody controls, pilots and mechanics are unionized and scarce, and airports control gates and slots.	High
Threat of substitutes	Can buyers meet the need another way?	Driving on short routes, high-speed rail where it exists, and video meetings for a large share of business travel.	Moderate to high
Threat of new entrants	How easily can newcomers arrive?	Capital is heavy but aircraft can be leased, and new low-cost carriers keep appearing; slots at congested airports are the real barrier.	Moderate

Four forces strong and one moderate is a structural verdict: this is a hard industry in which to earn sustained returns, and no amount of managerial effort repeals that. The strategic implications follow directly, and you can see them in real behavior. Weaken buyer power by making switching costly, which is what frequent-flyer programs and corporate contracts do. Escape pure price rivalry by segmenting the cabin so that the same flight sells at four different prices. Reduce supplier leverage by simplifying, which is why Southwest Airlines has flown a single aircraft family, the Boeing 737, cutting spare-parts inventory, maintenance complexity, and pilot training costs, and helping it post 47 consecutive profitable years through 2019 in an industry where that record is otherwise unheard of.

Two cautions before you point the framework at everything. Porter's model was built for established, well-defined industries, and it says little about complements, network effects, or regulation, which is why it is a poor first tool for platform businesses. And industries are not fixed: the forces describe structure at a moment, and a strategy can change the structure, as low-cost carriers did to European short-haul flying.

Key idea: Industry structure, not effort, sets average profitability, and the five forces tell you which specific structural pressure a strategy must escape.

SWOT, used properly

SWOT lists strengths and weaknesses, which are internal and present, and opportunities and threats, which are external and future-facing. It is the most-used and most-abused tool in management, and its abuse has a predictable shape: a workshop generates four columns of vague nouns, everyone feels productive, and the page never affects a decision.

Four disciplines fix most of that. First, tie the analysis to a specific decision, because "SWOT of our company" produces mush while "SWOT of opening a second location in Riverside" produces argument about the right things. Second, make every entry relative and evidenced: not "good staff" but "average tenure of 6.2 years against an industry norm near 2, which is why our repeat-repair rate is 3 percent." Third, be honest that a trait can appear twice, since deep specialization is a strength when demand holds and a weakness when it moves. Fourth, and most important, convert the grid into actions by pairing across it: strength plus opportunity gives an attack, weakness plus threat gives a defense or an exit, strength plus threat gives a shield, weakness plus opportunity tells you what to fix or buy before you can move. A SWOT that ends as four lists is a diagnosis with no guiding policy, which is Rumelt's definition of bad strategy.

Key idea: SWOT earns its keep only when it is scoped to one decision, evidenced, and converted into paired actions rather than left as four lists of adjectives.

Generic strategies: three ways to win, one way to drift

Porter argued that a firm's advantage comes from one of two sources, lower cost or meaningful differentiation, applied either broadly or to a narrow segment, giving cost leadership, differentiation, and focus.

• Cost leadership means being the lowest-cost producer at acceptable quality, and then choosing whether to pass savings on or bank a wider margin. It is built from structural choices such as scale, standardization, and process discipline, not from telling people to spend less.
• Differentiation means offering something buyers value enough to pay a premium for: reliability, design, service, speed, brand, or an ecosystem that makes leaving painful.
• Focus means applying either logic to a narrow slice, such as a regional airline serving routes the majors will not fly, or a barber shop that serves only curly hair and can charge for that expertise.

Porter's warning was that a firm which pursues neither clearly gets stuck in the middle, more expensive than the cost leader and less compelling than the differentiator, which describes a good number of failed mid-market retailers. The warning is useful, but the strict claim has been contested for decades with concrete counterexamples. Toyota built both low cost and high perceived quality; IKEA is cheap and strongly differentiated by design and store experience; Southwest was the cost leader and a brand people liked. Modern practice treats the framework as a discipline about coherence rather than a prohibition: your operating choices must reinforce each other. Southwest's single fleet type, point-to-point routing, and fast turnarounds all pull in the same direction, and adding a first-class cabin and hub-and-spoke connections would have fought every one of them. Being stuck in the middle is really the state of having made a set of choices that cancel out.

Key idea: Advantage comes from lower cost or real differentiation, broad or focused, and the practical test is not which label you claim but whether your operating choices reinforce one another.

Is your advantage real? The resource-based view

The five forces look outward at the industry; the resource-based view looks inward and asks why one firm outperforms rivals facing identical forces. Jay Barney's VRIO test asks four questions about any resource or capability you believe is an advantage. Is it valuable, meaning it lets you exploit an opportunity or blunt a threat? Is it rare, meaning few competitors have it? Is it costly to imitate, whether because of history, ambiguity about how it works, or social complexity? And is the firm organized to capture the value, with the structures and incentives to actually use it?

Run it on a claim you will hear constantly: "our people are our advantage." Valuable, plausibly yes. Rare, only if your hiring or training genuinely differs from the market. Costly to imitate, rarely, since competitors can hire your people, which is why individual talent is a weak moat while a culture that makes talented people effective is a strong one. Organized to capture, only if you keep them. Most claimed advantages die at the third question, and a strategy that survives on a resource anyone can buy is a strategy with a countdown timer. What survives tends to be the awkward, slow-built stuff: a distribution network, a regulatory license, a brand meaning built over decades, switching costs, patents, or a genuinely different way of working that competitors can describe but not reproduce.

Key idea: A durable advantage must be valuable, rare, costly to imitate, and captured by the organization, and most claimed advantages fail the imitation test.

Try it

Run an abbreviated five forces on independent coffee shops in a mid-sized city, then decide a strategy. Facts: espresso machines and a lease are the main capital, so a new shop can open for well under a hundred thousand dollars; there are two large chains and about forty independents; customers can walk to another shop in five minutes; beans are available from many roasters; and home espresso machines have improved sharply.

Answer: Entrants: high threat, low capital and no license barrier. Rivalry: high, many similar shops in a small area. Buyers: high power, near-zero switching cost. Suppliers: low to moderate, roasters are numerous, though a specific single-origin bean or a prime corner lease is scarcer. Substitutes: high, home brewing and convenience-store coffee. Structural verdict: thin margins for an undifferentiated shop. A defensible strategy is focus with differentiation: pick a narrow segment such as remote workers who need four hours of reliable seating, power, and quiet, and reinforce it coherently with table density, a wifi policy, and food designed for long stays. Notice that the choice creates a real cost, since fewer seats turn over per hour, which is what makes it a strategy rather than a wish.

Common misconceptions

• "A revenue target is a strategy." A target says where you want to arrive. Strategy says how you will get there against opposition, which is why Rumelt calls goal-stating a common form of bad strategy.
• "Great management can make any industry profitable." Structure sets the average. Skilled airline managers still fight four strong forces, which is why the industry's history is full of bankruptcies.
• "SWOT is a strategy tool." It is a diagnostic input. Unless it is scoped to a decision and converted into paired actions, it changes nothing.
• "You must be either low cost or differentiated, never both." Toyota, IKEA, and Southwest are standing counterexamples; the durable lesson is coherence among choices, not purity of label.
• "Our advantage is our great people." People are usually valuable but rarely rare and seldom costly to imitate, since competitors can hire them. What is hard to copy is the system that makes people effective.
• "The five forces work everywhere." The model was built for defined industries and handles complements, network effects, and regulation poorly, so platforms need additional tools.

Recap

• A real strategy has a diagnosis, a guiding policy, and coherent actions; ambition and targets are none of those.
• Porter's five forces explain average industry profitability, and applied to airlines they explain a century of thin returns and bankruptcies.
• Strategic moves aim to weaken a specific force: loyalty programs against buyer power, fleet simplification against supplier power, segmentation against rivalry.
• SWOT works only when scoped to a decision, evidenced, and converted into strength-opportunity and weakness-threat actions.
• Cost leadership, differentiation, and focus describe sources of advantage; incoherent choices, not impure labels, are what leave firms stuck in the middle.
• VRIO tests whether a claimed advantage is valuable, rare, costly to imitate, and organized for capture; most fail on imitability.

Sources

1. Bright, D. S., & Cortes, A. H. (2019). Principles of Management, Chapter 9: The strategic management process. OpenStax, Rice University. openstax.org
2. Porter, M. E. (1979). How competitive forces shape strategy. Harvard Business Review, 57(2), 137-145. hbr.org
3. Wikipedia. (2025). Porter's five forces analysis. en.wikipedia.org
4. Wikipedia. (2025). SWOT analysis. en.wikipedia.org
5. Wikipedia. (2025). Southwest Airlines. en.wikipedia.org
6. U.S. Bureau of Transportation Statistics. (2025). Airline data and statistics. bts.gov

The big picture

A manager's output is decisions. Not effort, not hours, not the number of meetings attended: decisions, plus the follow-through that makes them real. That is uncomfortable, because deciding is the part of the job we do worst and understand least. We know how to reason carefully when we have time, information, and no stake in the answer. Managers almost never have all three at once.

This lesson has two halves. The first is the machinery: how a decision should be structured, and the well-documented ways human judgment departs from that ideal. The second is a case, the launch of the Space Shuttle Challenger on January 28, 1986, in which a group of intelligent, technically expert, well-meaning people made a decision that killed seven astronauts, and in which every failure mode we will study is visible in the record. The point of the case is not to feel superior to people under pressure. It is the opposite: to see how ordinary the mechanisms are, and then to build the safeguards that catch them.

The rational model and its limits

The textbook rational decision model runs in six steps: define the problem, identify and weight the criteria that a good answer must satisfy, generate alternatives, evaluate each against the criteria, choose the highest-scoring option, then implement and review. It assumes a clear problem, known alternatives, stable preferences, and enough time.

Herbert Simon, who won the Nobel Memorial Prize in Economics in 1978, pointed out that real managers have none of that. Information is incomplete and costly, the mind's processing is limited, deadlines are real, and options are discovered in sequence rather than laid out for comparison. He called the result bounded rationality: people behave rationally within the limits of what they can know and process. The characteristic behavior is satisficing, searching until an option is found that is good enough on the criteria that matter, and stopping there rather than optimizing over a full set. Hiring is the everyday example. Nobody interviews the population; you interview until someone clears the bar.

Satisficing is not laziness, it is economy, and it is usually correct. The failure is applying it at the wrong altitude. A useful rule: match effort to reversibility and stakes. Choosing a caterer is cheap and reversible, so satisfice and move on; choosing a warehouse management system that will shape operations for a decade deserves the slow model. Two more distinctions help. Programmed decisions are routine and should be settled once with a standing plan, while nonprogrammed decisions are novel and deserve real analysis. And decisions vary by information state: certainty, risk (outcomes with knowable probabilities), uncertainty (outcomes without reliable probabilities), and ambiguity, where even the question is unclear. Most interesting management decisions live in the last two, which is exactly where our biases operate hardest.

Key idea: The rational model describes how to decide, bounded rationality describes how people actually decide, and the skill is matching the effort you spend to the stakes and reversibility of the choice.

The biases that bite managers

These are not exotic failings of unusual people. They are regularities in normal cognition, and they show up in your own decisions most invisibly.

• Anchoring: the first number in the room drags the final one. A supplier who opens at $140,000 has shifted your idea of a good deal before the discussion starts. Countermeasure: write your own number, with reasoning, before you hear theirs.
• Availability: vivid, recent, or personally experienced events feel more likely than they are. One dramatic warehouse theft can trigger a security program while a slow, larger loss to inventory shrinkage goes unexamined. Countermeasure: ask for base rates.
• Confirmation bias: we seek and weight evidence that supports what we already believe. It is the reason pilot projects championed by their sponsors so often succeed. Countermeasure: before looking at data, write down what result would change your mind.
• Overconfidence: people's confidence intervals are far too narrow, which is the engine of the planning fallacy from the previous lesson. Countermeasure: forecast ranges, and check them later against what happened.
• Escalation of commitment: having invested, we invest more to justify the earlier investment, treating sunk costs as reasons. The correct question is always forward-looking: given where we are now, is the remaining spend the best use of this money? Countermeasure: set kill criteria before starting, and have someone other than the champion review them.
• Framing: the same facts described as a 90 percent survival rate or a 10 percent mortality rate produce different choices. Countermeasure: restate every important option in both gain and loss terms.
• Hindsight bias: after an outcome, we believe it was predictable, which corrupts learning and makes post-mortems unfair. Countermeasure: keep a decision journal recording what you expected and why, then grade the reasoning rather than the result.

Key idea: Managerial biases are systematic, not random, so the fix is procedural, deciding in advance what would change your mind, what counts as failure, and when to stop.

Groupthink, and the night before Challenger

Groups can decide better than individuals, since they hold more information and more perspectives, but they can also decide much worse. Irving Janis, studying American foreign policy fiascos, named the failure groupthink: a cohesive group under pressure converges on consensus and suppresses dissent, producing an illusion of invulnerability, collective rationalization of warnings, self-censorship by doubters, an illusion of unanimity because silence reads as assent, direct pressure on anyone who objects, and self-appointed mindguards who shield the group from inconvenient information.

Now the case. Challenger's solid rocket boosters were sealed by rubber O-rings. Engineers at the contractor Morton Thiokol had known for years that these rings could be damaged by hot gases during ignition, and that cold weather made the rubber less resilient and the seal slower to close. The forecast for the morning of January 28, 1986 was unusually cold, near freezing at the pad, far below any previous launch. On the evening of January 27, Thiokol engineers, including Roger Boisjoly, presented their concerns in a teleconference and recommended against launching. NASA managers reacted with evident displeasure at the recommendation. Thiokol went off the line for a private caucus, in which a senior executive told a colleague it was time to take off his engineering hat and put on his management hat, and the company returned with a reversed recommendation to launch. No engineer signed the approval. Challenger broke apart 73 seconds after liftoff, killing all seven crew members, including the teacher Christa McAuliffe.

The Rogers Commission, whose members included the physicist Richard Feynman, concluded that the physical cause was a failed O-ring seal and that the decision-making process was seriously flawed. Feynman demonstrated the physics for the public by dropping a piece of O-ring material into a glass of ice water and showing that it lost resilience. Read the sequence against Janis and every symptom is present: pressure to maintain a launch schedule, dissent expressed and then overridden, a group reframing a safety objection as an engineering burden of proof placed on the objectors rather than the proponents, and silence taken for agreement.

The sociologist Diane Vaughan added the deeper mechanism in her 1996 study. She called it normalization of deviance. O-ring erosion had appeared on earlier flights; each time, the mission succeeded, and the anomaly was reclassified as an acceptable risk. Over years, an alarming signal became routine data. No villain is required for this, and that is the frightening part: every organization that treats a recurring near-miss as evidence of safety rather than as a warning is running the same process. Ask yourself what your own workplace has quietly reclassified as normal.

Key idea: Groupthink and the normalization of deviance turn dissent into noise and warnings into routine, and both were visible in the Challenger decision before the launch, not only after it.

Building decisions that resist these failures

Good decision processes are cheap relative to bad decisions, and they are mostly about structure rather than intelligence.

1. Set criteria before seeing options. Deciding what matters after you have a favorite is how criteria get bent to fit.
2. Run a premortem. Gary Klein's technique: before committing, tell the team to imagine it is eighteen months later and the decision failed badly, then have everyone write down why. Prospective hindsight loosens tongues that stay tied when the question is a polite "any concerns?"
3. Assign structured dissent. Name a devil's advocate or a small red team with the explicit job of arguing against, so objecting is a role rather than an act of disloyalty. Note the limit: a token advocate whose objections are ritually dismissed can make a group more confident, so the role only works when the group must answer the objections on the record.
4. Have the leader speak last. When the boss opens with a preference, later speakers are no longer independent sources of information.
5. Collect views independently first. Written estimates or votes before discussion preserve information that conformity would erase.
6. Hold a second-chance meeting. Janis's own recommendation: after reaching consensus, reconvene expressly to air residual doubts before acting.
7. Write kill criteria and a decision journal. Record what you expect, what would prove you wrong, and at what point you stop. This is the single best defense against escalation of commitment and hindsight bias at once.

Janis pointed at the contrast himself: the same Kennedy administration that produced the Bay of Pigs fiasco in 1961 restructured its process, inviting outside experts, splitting into subgroups, and pressing for dissent, and handled the Cuban Missile Crisis in 1962 far better. Same people, different procedure, different outcome. That is the most encouraging fact in this lesson.

Key idea: Decision quality is a property of process, not of individual brilliance, and premortems, structured dissent, independent views, and pre-set kill criteria measurably improve it.

Try it

A logistics firm has spent $1,800,000 of a $2,400,000 budget on custom routing software that is late and buggy. Finishing it will cost the remaining $600,000 and probably three more months. A commercial package would cost $250,000 plus $70,000 a year and could be running in six weeks. The project sponsor argues that abandoning the build wastes $1,800,000. (a) Name the bias in the sponsor's argument. (b) State the correct comparison. (c) Design one process safeguard that would have surfaced this earlier.

Answer: (a) Escalation of commitment driven by sunk-cost reasoning, likely with a dose of confirmation bias in the sponsor's reading of progress reports. (b) The $1,800,000 is spent regardless and is irrelevant. The comparison is $600,000 plus three months for an uncertain custom system against $250,000 plus $70,000 a year for a known one, so over five years the package costs about $250,000 + 5 x $70,000 = $600,000, roughly equal in money but far lower in risk and much faster. (c) Kill criteria agreed at kickoff, for example a review at 60 percent of budget with a pre-agreed rule that missing two milestones triggers a formal build-or-buy reassessment run by someone other than the sponsor.

Common misconceptions

• "Good managers decide from the gut." Expert intuition is real but requires a regular environment with fast, clear feedback. Hiring and strategy provide neither, which is why structure beats instinct there.
• "Satisficing is lazy." It is efficient for reversible, low-stakes decisions. The error is satisficing on a ten-year commitment.
• "Sunk costs should count." Money already spent cannot be recovered by spending more. Only future costs and benefits are decision-relevant.
• "Challenger failed because someone was reckless or evil." The record shows competent people inside a process that inverted the burden of proof and normalized a known anomaly over years.
• "Cohesive teams decide better." Cohesion plus pressure plus a directive leader is the recipe Janis identified for groupthink; cohesion helps only with structured dissent.
• "A devil's advocate always fixes group decisions." Only if the group must answer the objections. A ritual advocate can increase false confidence.
• "A bad outcome proves a bad decision." Decisions are made under uncertainty. Judge the process and the information available at the time, which is what a decision journal preserves.

Recap

• The rational model prescribes defining, weighting, generating, evaluating, choosing, and reviewing; bounded rationality explains why managers satisfice instead.
• Match decision effort to stakes and reversibility, and settle programmed decisions once with standing plans.
• Anchoring, availability, confirmation, overconfidence, escalation, framing, and hindsight are systematic and require procedural fixes.
• Groupthink suppresses dissent through rationalization, self-censorship, illusory unanimity, and pressure on objectors.
• The Challenger decision shows dissent overridden, the burden of proof inverted, and years of normalized deviance around O-ring erosion.
• Premortems, structured dissent that must be answered, leaders speaking last, independent views, second-chance meetings, and pre-set kill criteria improve decision quality.

Sources

1. Bright, D. S., & Cortes, A. H. (2019). Principles of Management, Chapter 2: Managerial decision-making. OpenStax, Rice University. openstax.org
2. Encyclopaedia Britannica. (2024). Challenger disaster. britannica.com
3. Encyclopaedia Britannica. (2024). Herbert A. Simon. britannica.com
4. Wikipedia. (2025). Groupthink. en.wikipedia.org
5. Wikipedia. (2025). Space Shuttle Challenger disaster. en.wikipedia.org
6. Klein, G. (2007). Performing a project premortem. Harvard Business Review, 85(9), 18-19. hbr.org

The big picture

An organizational chart looks like a picture of who is important. It is actually a picture of two decisions: how the work was cut into pieces, and how those pieces are supposed to be put back together. Every structure ever invented is an answer to that pair. Specialization makes people faster and better at their piece. Specialization also creates the exact problem it solves for, because now the pieces have to be reconnected by someone, and reconnection is expensive.

You can watch this in a restaurant. One person doing everything for four tables needs no coordination and no meetings, and will be slow and mediocre at half the tasks. Split the work into host, server, line cook, and dishwasher, and each becomes fast and skilled, but now the order has to travel accurately from table to kitchen and the food has to come back hot to the right seat. The tickets, the expediter calling out orders, the shift briefing: that is all coordination overhead, and it is the price of specialization. Organizing is the function that sets that price deliberately instead of by accident.

A structure is not neutral furniture. It determines who talks to whom, which problems get noticed, whose career depends on what, and how fast a decision can travel. Change the chart and you change behavior, which is why reorganizations are both a genuine tool and a favorite way of appearing to act without deciding anything.

The four building blocks

Four choices define most structures. Specialization is how finely work is divided. Departmentalization is how specialized jobs are grouped: by function, by product, by geography, or by customer. Chain of command and span of control set who reports to whom and how many people report to each manager. Centralization and formalization set where decisions are made and how much is written into rules.

A word on authority. Authority is the right to make decisions and direct work, responsibility is the obligation to perform, and accountability is answering for the result. The rule that matters in practice is that authority must be delegated with responsibility. Give someone responsibility for on-time delivery without authority over the schedule and you have created a person who can only fail and apologize. It is also worth distinguishing line positions, which are in the chain that produces the product or service, from staff positions such as HR, legal, and finance, which advise and support. Much day-to-day organizational friction is line-staff friction, an operations manager who experiences a compliance requirement as an obstacle and a compliance officer who experiences a shortcut as a risk. Both are doing their jobs.

Key idea: Structure is the deliberate answer to dividing work and rejoining it, and delegated responsibility without matching authority guarantees failure.

The main structural forms

Structure	Grouping logic	Strengths	Weaknesses	Fits
Functional	By specialty: marketing, operations, finance	Deep expertise, economies of scale, clear career ladders	Silos, slow cross-functional work, no one owns the customer end to end	Single product line, stable environment
Divisional	By product, region, or customer segment	Accountability for results, faster local response, easy to add or shed a division	Duplicated functions, higher cost, divisions compete internally	Diverse products or markets
Matrix	Both at once: function plus project or product	Shares scarce expertise, balances two priorities, rich information flow	Two bosses, ambiguity, conflict, slow decisions, heavy meeting load	Project work needing scarce specialists, aerospace, consulting
Team-based or flat	Small cross-functional teams, few layers	Speed, ownership, less bureaucracy	Unclear careers, coordination strain at scale, hidden politics	Small firms, fast-moving product work
Network or virtual	Core firm plus contracted partners	Flexibility, low fixed cost, access to specialists	Weak control over quality, dependency, knowledge leaks out	Fashion, film, some electronics

The historical pattern is worth knowing because it repeats. Alfred Chandler's study of American corporations argued that structure follows strategy: firms that diversified into multiple products, notably DuPont and General Motors in the 1920s, found the functional form could not cope and invented the multidivisional structure, with divisions running their own operations under a corporate center that allocated capital. When a company that grew up making one thing adds a second, unrelated thing, the functional chart starts producing arguments that no one has the authority to settle. That is the signal.

Matrix deserves a specific warning. It violates Fayol's unity of command deliberately, and the price is real: employees report to a functional manager and a project manager whose priorities conflict, and the conflicts arrive on the employee's desk. A matrix works when three conditions hold, namely genuinely scarce specialists who must be shared, projects that demand both technical depth and delivery focus, and senior managers willing to resolve conflicts quickly rather than letting them fester. Adopt a matrix without the third condition and you have invented a machine for producing meetings.

Key idea: Functional structures buy expertise at the cost of silos, divisional structures buy accountability at the cost of duplication, and matrix structures buy both at the cost of clarity.

Span of control: do the arithmetic

Span of control is the number of people reporting to one manager. Narrow spans give close supervision and more layers; wide spans give autonomy and fewer layers. The consequences are arithmetic, and doing the arithmetic once will change how you read a chart forever.

Take an organization with 4,096 front-line workers. With a span of 4, each layer is a quarter the size of the one below: 4,096 workers, then 1,024 supervisors, 256, 64, 16, 4, and 1 at the top. That is six management layers and 1,365 managers. Now widen the span to 8: 4,096 workers, then 512, 64, 8, and 1. That is four layers and 585 managers. Same workforce, 780 fewer managers, and two fewer layers for information to climb. If a manager costs $95,000 fully loaded, the difference is 780 x $95,000 = $74,100,000 a year.

Before you conclude that wide is always better, notice what the wide span assumes: work that is standardized enough that people do not need constant direction, employees experienced enough to solve their own problems, physical or digital proximity, and good information systems. Move the same span onto brand-new employees doing varied, high-risk work and quality will fall, because nobody has time to coach. The right span depends on the work, which is contingency thinking again. And the two lost layers are not purely a gain: layers are also where developing managers learn, so flattening can save money now and hollow out your management bench in five years.

Key idea: Span of control drives layers, cost, and communication speed geometrically, and the right span depends on task standardization and employee experience rather than on fashion.

Centralization, formalization, and the limits of flat

Centralization concentrates decision authority at the top. It gives consistency, faster enterprise-wide moves, and tighter cost control, and it is right for decisions where uniformity matters more than local fit: brand standards, safety policy, capital allocation, pricing architecture. Decentralization pushes authority down. It gives speed, local fit, and motivation, and it is right where local knowledge is decisive and errors are recoverable: staffing a shift, handling a customer complaint, adjusting a display. Most organizations are mixed, and the useful discipline is to decide deliberately, decision type by decision type, rather than to adopt a slogan. Ask three questions of any decision: who has the information, how costly is an inconsistent answer, and how reversible is a mistake.

Formalization is how much is written down as rules and procedures. High formalization gives consistency, trainability, and legal defensibility, and it slowly removes judgment. That is precisely Weber's iron cage from Module 1, and it is why customer service scripts produce reliable mediocrity: the script prevents the worst interaction and forbids the best.

Which brings us to the flat-organization dream, and a real test of it. In 2013 the online retailer Zappos began adopting Holacracy, a system that replaces managers with self-organizing circles and defined roles. In 2015 the chief executive offered severance to anyone who did not want to work that way, and roughly 18 percent of employees took the offer. Zappos moved away from Holacracy by around 2020. The lesson is not that flat is impossible; it is that removing the formal hierarchy does not remove hierarchy. It converts an explicit, accountable structure into an implicit one based on tenure, personal networks, and confidence, which is harder for newcomers to read and harder to appeal against. Amazon's two-pizza team rule, keeping teams small enough to be fed by two pizzas, is the more common compromise: small autonomous units inside a conventional hierarchy, with clear owners and interfaces.

Key idea: Centralize what needs consistency and decentralize what needs local knowledge, and remember that abolishing formal hierarchy usually replaces it with an informal one that is less visible and less accountable.

Try it

A regional bakery has grown from one shop to eleven across three cities, plus a central production kitchen. It is still organized functionally: everyone reports through production, retail, or finance, and the head of retail approves every hire, price change, and promotion. Store managers complain that a request to change a display takes eight days. (a) Which structure would you propose and why? (b) Which decisions would you decentralize and which would you keep central? (c) If the chain has 190 front-line staff and moves from a span of 5 to a span of 9, roughly how many first-line supervisors does it need in each case?

Answer: (a) A divisional structure by city, each with its own retail leadership and shared central production, since the products are similar but local markets and staffing differ. If instead the growth were into different product lines, product divisions would fit better. (b) Decentralize display layout, local promotions within a set budget, shift staffing, and hiring for hourly roles; keep central the recipes and quality standards, pricing architecture, brand identity, supplier contracts, and capital spending. (c) At a span of 5, about 190 / 5 = 38 supervisors; at a span of 9, about 190 / 9 = 21. The 17-supervisor difference is only a saving if the work is standardized enough that a supervisor of nine can still coach effectively.

Common misconceptions

• "Flat organizations have no hierarchy." They have an informal one. Zappos's experience shows that removing titles moves influence to networks and tenure rather than eliminating it.
• "Wider spans are always better." Wide spans require standardized work, experienced staff, and good information systems. Otherwise coaching stops and quality falls.
• "Matrix structures are modern and therefore superior." Matrix trades clarity for shared expertise and needs senior managers who resolve conflicts quickly, or it becomes a meeting factory.
• "Reorganizing fixes performance." Structure shapes behavior but cannot substitute for strategy, competent people, or working processes. Frequent reorganizations often signal avoidance of a harder problem.
• "Decentralization means losing control." It means moving decision rights to where the information is, while keeping consistency where inconsistency is costly.
• "Silos happen because people are territorial." Silos are the predicted output of functional grouping plus function-based incentives. Change the incentives and the shared goals before blaming personalities.

Recap

• Every structure answers two questions: how to divide work and how to coordinate it, and coordination is the price of specialization.
• Authority must accompany responsibility, and line-staff friction is structural rather than personal.
• Functional, divisional, matrix, team-based, and network forms each trade expertise, accountability, speed, and clarity differently.
• Structure follows strategy: diversification drove the invention of the multidivisional form in the 1920s.
• Span of control governs layers, manager headcount, and communication speed, with large arithmetic consequences.
• Centralize for consistency, decentralize for local knowledge, and expect informal hierarchy to fill any vacuum left by removing the formal one.

Sources

1. Bright, D. S., & Cortes, A. H. (2019). Principles of Management, Chapter 10: Organizational structure and change. OpenStax, Rice University. openstax.org
2. Wikipedia. (2025). Organizational structure. en.wikipedia.org
3. Wikipedia. (2025). Span of control. en.wikipedia.org
4. Wikipedia. (2025). Holacracy. en.wikipedia.org
5. Gitman, L. J., McDaniel, C., Shah, A., Reece, M., Koffel, L., Talsma, B., & Hyatt, J. C. (2018). Designing organizational structures. In Introduction to Business. OpenStax, Rice University. openstax.org

The big picture

Start a new job and within two weeks you will know things nobody told you. Whether it is acceptable to disagree with your boss in a meeting. Whether the 9 a.m. start is real. Whether the person who works Sundays is admired or pitied. Whether saying "I do not know" is safe. None of this is in the handbook, all of it is enforced, and it will shape your behavior more reliably than any policy. That is organizational culture: the shared assumptions, values, and norms that determine how things are done here.

Culture is not a poster and it is not a perks budget. It is the pattern of behavior that a group has learned works, transmitted to new members as the correct way to perceive and act. Because it is learned rather than announced, it can be entirely at odds with the official version, and the gap between the two is one of the most diagnostic things you can measure about a company. This lesson gives you a way to read a culture, an honest account of what culture does and does not do for performance, and a realistic account of how it changes, which is slowly, expensively, and only through behavior.

Schein's three levels: reading a culture

Edgar Schein's model is the most useful diagnostic tool in this area because it separates what you can see from what actually governs behavior.

1. Artifacts are the visible surface: office layout, dress, logos, jargon, meeting rituals, stories people tell newcomers, what hangs on the walls. Easy to observe, easy to misread. An open-plan office may express egalitarianism or simply a cheap lease.
2. Espoused values are what the organization says it believes: the values statement, the strategy deck, what a leader says in an all-hands meeting. These are claims, and claims can be sincere, aspirational, or decorative.
3. Basic underlying assumptions are the beliefs so deeply held that nobody argues about them, or even notices them. Whether people can be trusted without monitoring. Whether bad news should travel up. Whether the customer or the engineering team is the real authority. These are invisible, unwritten, and decisive.

The method is to read the artifacts, note the espoused values, and then look for the gap that reveals the assumption. A firm whose values statement says "we take smart risks" while every failed project has ended someone's promotion prospects has an underlying assumption that failure is career-ending, and no amount of encouragement will produce risk-taking. The practical test is simple and brutal: watch what happens to the person who acts on the stated value at real personal cost. That is the culture.

Enron is the standard illustration, and the details matter. The company's stated values were respect, integrity, communication, and excellence, displayed prominently at headquarters. Its actual reward system paid enormously for booking deals and reported earnings, its performance review process was famously brutal, and its accounting became fraudulent, ending in bankruptcy in 2001 and criminal convictions. The espoused values were not lies told by everyone; they were simply irrelevant next to what was measured, paid, and promoted. Employees are excellent readers of that difference.

Key idea: Culture lives in basic assumptions, not in stated values, and the gap between the two is visible in what happens to the person who takes the stated value seriously.

Where culture comes from, and how it reproduces

Cultures are not chosen. They accumulate through five mechanisms, and every one of them is a lever.

• Founders and early history. The first twenty people set defaults that persist for decades, because their solutions to early problems become the way things are done.
• What leaders pay attention to, measure, and reward. Schein considered this the single most powerful mechanism. If a leader asks about safety in every meeting for a year, safety becomes real. If the only recurring question is this quarter's number, everything else becomes decoration.
• How leaders behave in a crisis. Ordinary days teach little. The layoff, the recall, the scandal, the lost customer: those reveal what is actually valued, and people remember for years.
• Selection, socialization, and promotion. Hiring people who fit and promoting people who exemplify the norms is how a culture reproduces itself. It is also how a culture becomes homogeneous and blind, which is the argument for hiring for values while deliberately seeking difference in background and thought.
• Stories, rituals, and language. The story about the engineer who stopped the line, the Friday demo, the shorthand nobody outside understands: these carry the culture more efficiently than any document.

Culture is also not uniform. Most organizations have subcultures, and the sales floor, the engineering team, the night shift, and the finance department can differ sharply in norms while sharing a few core assumptions. Much cross-functional conflict is subculture friction rather than personal difficulty, which is a more useful diagnosis because it points to shared goals and joint work rather than to personality repair.

Key idea: Culture forms from founders, from what leaders attend to and reward, from crisis behavior, from who is hired and promoted, and from the stories a group tells about itself.

Does culture drive performance?

You will often see the line "culture eats strategy for breakfast," usually attributed to Peter Drucker, though the attribution is doubtful and the sentiment is looser than it sounds. Take the underlying claim seriously and it splits into two questions.

The first is whether culture affects results. There is reasonable evidence that it does. Kotter and Heskett's study of over two hundred firms found that cultures which valued adaptability and attention to customers, shareholders, and employees outperformed over more than a decade, while strong cultures that valued internal consistency alone did not. Note the qualifier, because it is where most popular writing goes wrong. Strength is not the same as fitness. A strong culture aligned to a changed environment is a liability, and its very strength makes it harder to correct: everyone agrees, dissent feels disloyal, and the market has already moved.

The second question is what kind of culture. Cameron and Quinn's competing values framework offers a workable map with four types: clan (collaborative, family-like, high commitment), adhocracy (entrepreneurial, innovative, tolerant of failure), market (results-driven, competitive, externally focused), and hierarchy (structured, controlled, efficient). None is best. A nuclear plant with an adhocracy culture is terrifying; a research startup with a hierarchy culture is dead. Fit to the work, and to the strategy, is the criterion.

It is also worth being honest that "culture fit" has a well-documented failure mode in hiring: it often decays into hiring people who feel familiar, which narrows the organization while sounding rigorous. The improvement is to define the specific behaviors you require, such as giving direct feedback or writing decisions down, and to assess for those, rather than asking whether you would enjoy a drink with the candidate.

Key idea: Adaptive cultures outperform over the long run, strength without fit is a liability, and the useful question is not how strong your culture is but what specific behaviors it produces.

Why culture change is hard, and what actually works

Culture change fails for a structural reason: culture is downstream of incentives, and most change programs try to change culture directly with communication. You cannot announce a new assumption. Posters, values workshops, and rebranded conference rooms move the artifacts layer while the underlying assumptions sit untouched, and the visible mismatch usually increases cynicism. If leaders declare that speaking up is now valued and the first person to speak up is punished, the change program has taught the opposite of its message, and taught it more convincingly.

What does work is changing the things assumptions are made of. Alan Mulally's arrival at Ford in 2006 is a well-documented example. Ford's culture punished the reporting of problems, so executives arrived at reviews with everything coded green. Mulally instituted a weekly Business Plan Review in which every executive rated their programs green, yellow, or red against plan. For weeks, everything was green while the company was losing billions, which was itself the diagnosis. When Mark Fields finally showed a red for a launch delay, the room waited for the execution. Mulally applauded. That single response, repeated and reinforced, taught the organization more about the new culture than any statement could, and honest status data became available to management for the first time. Ford went on to restructure without taking the federal bailout that General Motors and Chrysler required in 2009.

Extract the general recipe from the specific story. Change what is measured and reported. Change who gets promoted, since promotions are the loudest statement an organization makes about what it values. Change what leaders visibly do, especially under pressure. Change the structures and processes that make the old behavior rational. Then, and only then, tell the story about what changed, using real examples rather than aspirations. Expect the whole thing to take years, and expect turnover, because some people joined for the old culture and will not want the new one.

Key idea: Culture changes when incentives, promotions, measurement, and visible leader behavior change, and communication only ratifies a change that has already happened in practice.

Try it

A software firm's values page reads: "Fail fast. Bring bad news early. Everyone has a voice." Observations: the last three promotions all went to people who shipped on schedule; a manager who flagged a slipping deadline was moved off the project; meetings run by seniority and juniors are not asked to speak; there is a monthly award for the biggest bug caught in production. (a) Sort these into Schein's levels. (b) State the likely underlying assumption. (c) Name two specific changes with a real chance of shifting behavior.

Answer: (a) Artifacts: the values page, the monthly bug award, the meeting ritual ordered by seniority. Espoused values: fail fast, bring bad news early, everyone has a voice. The promotions and the reassigned manager are enacted values, and they are the evidence for the assumption. (b) The underlying assumption is that schedule adherence is what actually counts and that reporting problems is dangerous to your career, so problems will be hidden until they explode. (c) For example: change the promotion criteria to include documented early escalation and require the promotion committee to cite an instance for each candidate; and change meeting practice so junior team members give their read first, with the leader speaking last, which also imports the decision-quality safeguard from Module 2. Note that a new poster would change nothing.

Common misconceptions

• "Culture is the values statement." That is one artifact. Culture is the assumption set revealed by what is rewarded, punished, and tolerated.
• "Perks create culture." Free lunch is an artifact. It changes convenience, not the assumption about whether bad news can travel upward.
• "A strong culture is always good." Strong plus misfit is worse than weak plus adaptable, because strength suppresses the dissent that would signal the mismatch.
• "Culture change is a communications project." Communication ratifies change; incentives, promotions, and leader behavior produce it.
• "Every organization has one culture." Subcultures are normal, and much cross-functional friction is subculture difference rather than personal failing.
• "Hiring for culture fit is objective rigor." Undefined, it selects for familiarity and narrows the organization. Define required behaviors and assess those instead.
• "Enron's people were uniquely corrupt." Enron's stated values were unremarkable. What differed was a reward system that made the stated values irrelevant, which is a design failure available to any firm.

Recap

• Culture is the set of shared assumptions that determine how things are actually done, learned rather than announced.
• Schein's levels move from visible artifacts through espoused values to the invisible basic assumptions that govern behavior.
• Culture forms through founders, what leaders attend to and reward, crisis behavior, selection and promotion, and stories.
• Adaptive cultures show better long-run performance; strength without fit to the environment is a liability.
• Competing values gives four cultural types, and the right one depends on the work and strategy rather than fashion.
• Culture changes through incentives, measurement, promotions, and visible leader behavior, as Ford's status-reporting turnaround illustrates.

Sources

1. Black, J. S., Bright, D. S., et al. (2019). Organizational Behavior, Chapter on organizational culture. OpenStax, Rice University. openstax.org
2. Wikipedia. (2025). Organizational culture. en.wikipedia.org
3. Wikipedia. (2025). Edgar Schein. en.wikipedia.org
4. Encyclopaedia Britannica. (2024). Enron scandal. britannica.com
5. Society for Human Resource Management. (2025). Organizational and employee development resources. shrm.org

The big picture

Human resources exists as a department, and every manager still does human resources. HR writes the policy, but you decide who to interview and what to ask, you decide whether the new hire's first week is organized or improvised, you decide what to say in the performance conversation, and you are the one whose careless sentence in a termination meeting becomes an exhibit. Delegating these to a department is not possible, only pretending to.

This lesson covers the four things a line manager must be able to do competently: hire, onboard, manage performance, and stay inside the legal and ethical lines. It closes with an evenhanded look at diversity, equity, and inclusion programs, an area where legal requirements, moral commitments, and empirical evidence are three different conversations that get mixed together, usually to everyone's confusion.

One necessary caveat: what follows is general education about U.S. practice, not legal advice. Employment law varies by state and country, changes often, and applies differently by employer size. Any real decision about discipline, accommodation, classification, or termination should involve your HR team or an employment lawyer.

Hiring: what actually predicts performance

Hiring begins before the posting, with a job analysis: what does this person actually do, what results define success, and which knowledge, skills, and abilities produce those results? Vague postings produce vague interviews, and vague interviews get decided on likeability. Write the success criteria first, in observable terms, and the rest of the process becomes an evidence-collection exercise against them.

Then use methods that predict. Decades of personnel-selection research consistently find that structured interviews, in which every candidate gets the same job-related questions scored on a defined rubric, predict later job performance substantially better than unstructured conversations, and that work samples, having the candidate do a representative piece of the job, are among the best predictors available. The precise validity numbers have been revised over the years, and a recent re-analysis corrected several long-quoted estimates downward, so treat any specific coefficient with care. What has not changed is the ordering: structure beats chat, and job-relevant demonstrations beat impressions.

That finding is unpopular because unstructured interviews feel informative. They are mostly a measure of social similarity and interviewer confidence, and they are where bias enters most easily, since an undefined standard is a standard that drifts toward whoever seems familiar. Three practical moves: write the questions and the scoring rubric before you meet anyone; have interviewers record scores independently before discussing, for the same reason the leader speaks last in a decision meeting; and give a realistic preview of the job's genuine downsides, since realistic job previews reduce early turnover by letting people select themselves out before you have paid to train them.

Key idea: Define success first, then use structured, job-relevant methods and independent scoring, because unstructured interviews mostly measure familiarity.

Onboarding and the first ninety days

Onboarding is the cheapest retention investment available and the one most often improvised. The new hire's first week teaches them what your culture actually is, and a week of waiting for a laptop and a login teaches something specific.

A workable structure. Before day one: equipment ready, accounts created, a written plan for the first two weeks, a peer buddy named, and the team told who is arriving and what they will own. Week one: meet the people they will depend on, complete one small real task end to end, and understand how the work reaches a customer. Days 30, 60, and 90: explicit checkpoints with written expectations, since the most common onboarding failure is a new hire who has been busy for two months without ever hearing what good looks like. Ask two questions at each checkpoint that most managers skip: what is still confusing, and what did you expect that turned out to be different? The second one is free consulting on your own organization from the only person who can still see it clearly.

The financial case is straightforward. Replacing an employee typically costs a substantial fraction of annual salary once you count recruiting, lost productivity, and the time of everyone involved in training. On a $60,000 role, even a conservative estimate of half a year's salary implies $30,000 per avoidable departure. A structured onboarding process costs a few hours of manager time per hire.

Key idea: Structured onboarding with named owners and 30-60-90 day expectations is a cheap, high-return retention practice, and the new hire's early confusion is valuable data about your organization.

Performance management that improves performance

The annual performance review has a poor reputation, largely deserved. Feedback arriving eleven months after the event cannot change the event, ratings are compressed by managers avoiding conflict, and tying a rating to pay turns a development conversation into a negotiation. Several large firms responded by rebuilding the process: Adobe replaced annual reviews with frequent check-ins in 2012, and Microsoft dropped its stack-ranking system in 2013.

Forced ranking, popularized as the vitality curve at General Electric under Jack Welch, requires managers to sort employees into fixed proportions, typically rewarding a top group and dismissing a bottom group. The argument for it is that it forces managers to make distinctions instead of rating everyone above average. The arguments against are that it assumes performance is distributed the same way in every team, which is false for a small team of excellent people, and that it makes colleagues into competitors, which damages exactly the cooperation most work depends on. Most large adopters have abandoned it.

What works better is unglamorous. Set expectations in observable terms at the start of the period. Give feedback close to the event, when it is still specific and actionable. Separate the development conversation from the compensation conversation, even if only by scheduling them weeks apart. Keep contemporaneous notes, both because memory is bad and because documentation is what protects a fair decision later. And know the rating errors you will personally commit: halo, letting one strong trait color everything; recency, over-weighting last month; leniency or central tendency, rating everyone high or everyone in the middle; and similarity bias, rating people like yourself more generously. You cannot introspect your way out of these; you can only write things down as they happen and use defined criteria.

When performance genuinely falls short, the sequence is clarity, support, and consequence: state the specific gap and the standard, agree what support and timeline are reasonable, document the conversation, and follow through consistently across people. Inconsistency is both unfair and the fastest route to a legal problem.

Key idea: Frequent, specific, documented feedback against expectations set in advance beats annual ratings, and forced ranking damages cooperation while assuming a distribution that small teams do not have.

The legal floor: at-will employment and EEO

Employment at will is the default rule in most U.S. states: absent a contract or statute to the contrary, either party may end the employment relationship at any time, for any reason or no reason, with or without notice. Montana is the notable exception, having enacted a statute limiting discharge without good cause after a probationary period. But "any reason" has always meant "any lawful reason," and the exceptions matter more than the rule. You may not fire someone for a reason prohibited by statute, in violation of public policy such as refusing to commit an illegal act or serving on a jury, in breach of an implied contract created by your own handbook or promises, or in retaliation for protected activity such as filing a complaint. Union contracts and individual employment agreements displace at-will terms entirely.

The main federal equal employment opportunity statutes, enforced primarily by the U.S. Equal Employment Opportunity Commission, are worth knowing by name and coverage:

Law	Year	Protects against discrimination based on
Equal Pay Act	1963	Sex-based pay differences for substantially equal work
Title VII, Civil Rights Act	1964	Race, color, religion, sex, and national origin; the Supreme Court held in 2020 that sex includes sexual orientation and gender identity
Age Discrimination in Employment Act	1967	Age, for workers 40 and older
Americans with Disabilities Act	1990	Disability, and it requires reasonable accommodation absent undue hardship

Two more statutes shape daily management even though they are not EEO laws. The Fair Labor Standards Act governs minimum wage, overtime, and the exempt or non-exempt classification that determines who must be paid overtime, an area where well-meaning managers create expensive problems by giving someone a manager title and assuming overtime rules no longer apply. The Family and Medical Leave Act provides eligible employees of covered employers up to twelve workweeks of unpaid, job-protected leave for specified family and medical reasons. Coverage thresholds and details matter, which is again why HR exists.

The practical translation for a line manager is short: base employment decisions on documented, job-related criteria, apply rules consistently, keep records, take complaints seriously and route them properly, never retaliate against someone who raises a concern, and ask before acting when you are unsure.

Key idea: At-will means either side may end employment for any lawful reason, and the lawful qualifier plus consistent, documented, job-related criteria is what keeps managerial decisions defensible.

DEI: three separate conversations

Programs grouped under diversity, equity, and inclusion attract strong feelings, and the disagreements become clearer once you separate three questions that are usually run together.

The first is legal, and it is not really contested: discrimination on protected characteristics is unlawful, and organizations must comply. The second is a values question. Some people hold that an organization should actively work toward representation reflecting the society it serves, as a matter of fairness and opportunity; others hold that the fairest approach is to evaluate individuals strictly on job-related merit and that group-conscious programs conflict with that. Both positions are held sincerely by reasonable people, they rest on different underlying values rather than different facts, and a course cannot settle them for you. It is also worth noting that the legal landscape is active: the Supreme Court's 2023 decision on race-conscious university admissions did not directly govern private employment, but it prompted many employers to review program design with counsel.

The third question is empirical, and here the honest answer is that the evidence is more contested than advocacy on either side suggests. The widely quoted consultancy findings correlating executive diversity with higher profitability have been challenged, including by researchers who attempted to replicate the U.S. results and could not. Correlation-based claims of this kind are also hard to interpret causally, since successful firms can afford to change their hiring and their boards. Meanwhile research on interventions finds that mandatory diversity training frequently produces little lasting behavior change and can provoke backlash, while structural measures show better results: targeted recruiting that widens the applicant pool, mentoring and sponsorship, transparent promotion criteria, and assigning specific accountability for outcomes to named managers.

Notice that the interventions with the best evidence are the same practices this lesson recommended for hiring quality: defined criteria, structured evaluation, independent scoring, wider sourcing, and documented decisions. Whatever your view on the values question, process discipline improves both the fairness and the accuracy of your decisions, which is a rare case of a genuinely non-partisan recommendation.

Key idea: Separate the legal requirement, the values debate, and the contested empirical claims; the structural practices with the best evidence, defined criteria and structured evaluation, also improve hiring accuracy regardless of where you stand.

Try it

You are hiring a customer support lead for a 12-person team. Your draft process is: a resume screen by gut feel, one 45-minute chat with you about background and interests, and a reference call. Turnover in this role has been high. (a) Name three specific improvements and what each one fixes. (b) Draft one structured interview question with a scoring rubric. (c) Estimate the annual saving if better selection and onboarding cut turnover in the role from two departures a year to one, at a replacement cost of half of a $68,000 salary.

Answer: (a) Write success criteria from a job analysis, which fixes drifting standards; add a work sample such as handling three realistic escalated tickets and writing the customer replies, which tests the actual job rather than self-description; and use two independent interviewers scoring the same structured questions before discussing, which limits halo and similarity bias. Add a realistic preview of the hardest part of the role, which reduces early regret quitting. (b) Question: "Tell me about a time you had to deny a customer request that they felt was reasonable. What did you say, and what happened?" Rubric: 1 = vague or blames the customer; 2 = describes the outcome but not their own words; 3 = gives specific language, acknowledges the customer's position, states the constraint clearly; 4 = all of level 3 plus a follow-up action and a change they made to prevent recurrence. (c) Replacement cost per departure is 0.5 x $68,000 = $34,000, so one fewer departure a year saves about $34,000.

Common misconceptions

• "HR handles the people stuff." HR sets policy and supports; the manager still hires, onboards, coaches, documents, and decides.
• "I can tell in the first five minutes." That confidence is well documented and poorly calibrated. Unstructured impressions predict performance weakly compared with structured methods and work samples.
• "At-will means I can fire anyone for any reason." It means any lawful reason. Statutory protections, public policy, implied contracts, and anti-retaliation rules all constrain it, and Montana limits at-will discharge by statute.
• "Giving someone a manager title makes them exempt from overtime." Classification depends on duties and pay tests under wage and hour law, not on the title, and getting it wrong is expensive.
• "Annual reviews are how you improve performance." Feedback works when it is close to the event and specific; annual ratings mostly allocate rewards and record impressions.
• "Forced ranking guarantees high standards." It assumes an identical performance distribution in every team and converts colleagues into competitors, which is why most large adopters dropped it.
• "The business case for diversity is settled." The widely cited profitability correlations have failed replication attempts and are hard to read causally; the values case and the legal duty are separate questions from that empirical one.

Recap

• Start hiring with a job analysis and observable success criteria, then use structured interviews, work samples, and independent scoring.
• Realistic job previews and structured onboarding with 30-60-90 day expectations reduce costly early turnover.
• Frequent specific feedback beats annual ratings; separate development from compensation conversations and keep contemporaneous notes.
• Forced ranking has largely been abandoned because it misassumes distributions and damages cooperation.
• At-will employment permits ending employment for any lawful reason, constrained by EEO statutes, public policy, contracts, and anti-retaliation rules.
• DEI raises three distinct questions, legal, moral, and empirical, and the interventions with the strongest evidence are structural practices that also improve selection accuracy.

Sources

1. U.S. Equal Employment Opportunity Commission. (2025). Laws enforced by the EEOC. eeoc.gov
2. U.S. Department of Labor, Wage and Hour Division. (2025). Family and Medical Leave Act and wage and hour compliance. dol.gov
3. U.S. Bureau of Labor Statistics. (2025). Human resources managers. In Occupational Outlook Handbook. bls.gov
4. Society for Human Resource Management. (2025). Talent acquisition and performance management resources. shrm.org
5. Bright, D. S., & Cortes, A. H. (2019). Principles of Management, Chapter 11: Human resource management. OpenStax, Rice University. openstax.org
6. Wikipedia. (2025). At-will employment. en.wikipedia.org

The big picture

Two people do the same job on the same pay. One takes real care over it and the other does the minimum that avoids trouble. Managers reach for personality as the explanation, which is comfortable and usually wrong, because the same person is often diligent in one job and indifferent in the next. Motivation is not a fixed trait people carry around; it is largely a property of the situation, and situations are things managers design.

This lesson is a tour of the major motivation theories, sorted by how well they hold up. That sorting matters. Most management training presents Maslow, Herzberg, expectancy, and self-determination theory as an undifferentiated buffet, which leaves you unable to tell that one of them is a fifty-year-old speculation with weak support and another has hundreds of experiments behind it. You will learn each theory, learn what it is good for, and learn what its evidence actually looks like.

Content theories: what people want

Content theories ask what needs drive behavior. Maslow's hierarchy, from a 1943 paper, arranges needs from physiological through safety, belonging, and esteem to self-actualization, and claims people pursue higher needs only once lower ones are satisfied. It is the most famous psychological model in business, and here is the honest position: the strict hierarchical ordering has never held up well empirically. A well-known 1976 review by Wahba and Bridwell found little support for either the five-level structure or the progression claim, and later work has not rescued it. People pursue meaning while broke and chase status while secure. What survives is weaker and still useful: different people are driven by different unmet needs at any moment, so a single incentive will not move everyone equally. Keep that; drop the pyramid's determinism.

Herzberg's two-factor theory separates hygiene factors, meaning pay, working conditions, policies, supervision, and job security, from motivators, meaning achievement, recognition, the work itself, responsibility, and advancement. Bad hygiene causes dissatisfaction; good hygiene removes the complaint without creating enthusiasm. Only motivators produce genuine engagement. The methodological criticism is real, since Herzberg's critical-incident interviews invited people to credit themselves for good times and blame conditions for bad ones. But the practical distinction survives repeated use: fixing the broken payroll system removes a grievance, and it does not make the twelfth identical call of the morning interesting. Those are two separate projects with two separate budgets.

McClelland's acquired needs, achievement, affiliation, and power, add a useful individual-differences lens. A high-achievement person wants challenging goals and clear feedback; a high-affiliation person wants belonging and does badly with pure competition; a high-power person wants influence, which is constructive when directed at organizational rather than personal ends. The point is managerial rather than theoretical: the reward that motivates you may bore the person across the desk.

Key idea: Maslow's ordering is poorly supported and should be taught with that caveat; the durable content insight is that people differ in which unmet needs currently drive them.

Process theories: how motivation gets produced

Expectancy theory, developed by Victor Vroom, is the best diagnostic tool in this lesson because it breaks motivation into three multiplied links, meaning that if any one is zero, motivation is zero.

• Expectancy: if I work hard, will performance actually improve? Broken by missing skills, bad tools, unclear standards, or an impossible target.
• Instrumentality: if I perform, will the promised outcome actually arrive? Broken by unkept promises, opaque reward decisions, or rewards distributed by favoritism.
• Valence: do I want that outcome? Broken by rewards nobody values, such as a plaque for someone who wanted a schedule change.

Use it as a diagnostic, not a slogan. A team is not "unmotivated"; one of the three links is broken, and each break has a different fix. If experienced staff have stopped trying for a bonus because last year's went to the manager's favorite, that is an instrumentality problem, and no amount of encouragement or extra bonus money repairs it. If the target requires software the team does not have, that is expectancy, and it is your problem, not theirs.

Equity theory, from J. Stacy Adams, says people compare the ratio of their outcomes (pay, recognition, opportunity) to their inputs (effort, skill, experience) against a referent other, usually a comparable colleague. Perceived under-reward produces predictable responses: reduce effort, seek a raise, reframe the comparison, undermine the comparison person, or leave. The managerial lessons are three. Perception drives behavior, so a fair decision that looks arbitrary functions as an unfair one. People compare locally and specifically, most often to the colleague two desks away. And pay secrecy does not remove comparison; it removes accuracy, leaving people to compare against guesses, which are usually flattering to the guesser.

Key idea: Expectancy theory locates a motivation failure in effort-to-performance, performance-to-reward, or the value of the reward, and equity theory explains why perceived unfairness reliably reduces effort.

Autonomy, competence, relatedness, and the design of work

Self-determination theory, developed by Edward Deci and Richard Ryan over decades of experiments, is the best-evidenced framework here. It holds that people have three basic psychological needs whose satisfaction produces high-quality, self-sustaining motivation: autonomy, meaning volition and choice over how work is done; competence, meaning growing mastery; and relatedness, meaning connection to others who matter. It also distinguishes intrinsic motivation, doing something because the activity itself is satisfying, from extrinsic motivation, which ranges from purely external compliance to values a person has genuinely internalized.

Its most famous finding is the overjustification effect: paying people for an activity they already find interesting can reduce their motivation to do it once the payment stops. In the classic study, children who enjoyed drawing and were given an expected reward for it drew less in free time afterwards than children given no reward. Meta-analysis found the same pattern for tangible, expected, performance-contingent rewards on interesting tasks. This does not mean pay demotivates. It means that for work someone already finds meaningful, controlling incentives can crowd out the internal reason for doing it. Unexpected recognition after the fact, and rewards that read as information about competence rather than as control, do not show the same effect.

The most practical application is the job characteristics model of Hackman and Oldham, which identifies five features of work that predict motivation and satisfaction: skill variety, task identity (completing a whole, identifiable piece of work), task significance (the work matters to other people), autonomy, and feedback from the job itself. The first three produce experienced meaningfulness, autonomy produces experienced responsibility, and feedback produces knowledge of results. The model also predicts a moderator that many managers ignore: enriched jobs motivate most strongly among people with high growth needs, so redesign is not equally welcome to everyone.

Made concrete: a hospital billing clerk who processes one field of every claim has low identity, low variety, and no feedback. Restructure the role so each clerk owns all claims for a set of clinics, sees the resolution, and can call the clinic directly to fix errors, and you have added identity, variety, autonomy, and feedback without spending a dollar on pay. That is job enrichment, adding depth and control, as distinct from job enlargement, which merely adds more tasks of the same kind and often reads to employees as more work for the same money.

Key idea: Autonomy, competence, and relatedness produce durable motivation, and the job characteristics model turns that into concrete redesign: whole tasks, real discretion, visible significance, and feedback from the work itself.

Does money motivate? Yes, conditionally

The evidence supports a nuanced answer rather than either slogan. Money reliably affects who applies and who stays, and pay perceived as unfair reliably damages effort. For simple, measurable, individually attributable output, performance pay works and the effects can be large: when the auto-glass firm Safelite moved windshield installers from hourly wages to piece rates in the 1990s, the economist Edward Lazear found productivity rose roughly 44 percent, about half from people working faster and about half from the change attracting and keeping more productive workers.

Now the conditions that make that success non-transferable. Piece rates work when output is countable, quality is verifiable, the work is individual, and gaming is hard. Move any one of those and the same scheme produces damage: sales commissions on unverifiable quality gave us the Sears repair quotas from Module 2, and incentives on complex interdependent work reliably produce metric gaming and reduced cooperation. That is why most professional work uses base pay plus modest variable components rather than pure piece rates.

A defensible pay philosophy for a manager looks like this. Get base pay right and roughly transparent in structure, because that is hygiene and equity. Do not attempt to buy engagement with variable pay on complex work, because that is where crowding-out and gaming live. Spend your design energy on the motivators and the job characteristics instead, since those are the levers that are both cheaper and more durable.

Key idea: Financial incentives work well for simple, countable, individual output and poorly for complex interdependent work, where fair base pay plus enriched job design outperforms clever bonus schemes.

Try it

A 22-person claims team has rising errors and falling output. Facts: the bonus depends on a department-wide target that the team has missed for three straight quarters; two people do all the interesting escalations while the rest process one field of each claim; the last two promotions went to people from another department with no explanation; and a survey shows staff like their colleagues and find the work pointless. Diagnose using each theory, then propose three changes.

Answer: Expectancy: the effort-to-performance link is intact, but a department target missed three times running has destroyed expectancy that effort changes the outcome, and unexplained outside promotions have destroyed instrumentality. Equity: the promotions created a perceived unfair ratio with an external referent, predicting reduced effort, which is what is happening. Job characteristics: task identity, variety, autonomy, and feedback are all low for twenty of the twenty-two, explaining pointlessness; relatedness is the one need being met, which is why they have not all left. Herzberg: the bonus is hygiene and cannot supply the missing motivators. Three changes: give each clerk end-to-end ownership of claims for named clinics with authority to resolve errors directly, which adds identity, autonomy, and feedback; replace the unreachable department bonus with a team-level goal the team can actually influence, restoring expectancy; and publish promotion criteria with a written rationale for each decision, restoring instrumentality and perceived equity.

Common misconceptions

• "Maslow's hierarchy is established science." The strict ordering has weak empirical support. Use it as a reminder that people differ, not as a sequence.
• "Money is the only real motivator." Pay is powerful for attraction, retention, and simple countable output, and weak at producing engagement in complex work.
• "Money never motivates." Equally wrong, as the Safelite piece-rate result shows. The question is always: which work, measured how?
• "Some people just are not motivated." People are motivated toward something. An apparently unmotivated employee is usually facing a broken expectancy link or a job with no identity, autonomy, or feedback.
• "Pay secrecy prevents comparison problems." It prevents accuracy, not comparison. People compare against guesses, and guesses tend to favor the guesser.
• "Job enlargement is job enrichment." Adding more similar tasks adds volume; enrichment adds depth, discretion, and feedback.
• "Rewarding good work always increases it." For work people already find interesting, expected controlling rewards can crowd out intrinsic motivation once withdrawn.

Recap

• Motivation is largely designed into situations rather than carried around as a trait.
• Maslow's ordering is weakly supported; Herzberg's hygiene-motivator split remains practically useful despite methodological criticism.
• Expectancy theory diagnoses which link is broken: effort to performance, performance to reward, or the value of the reward.
• Equity theory explains reduced effort from perceived unfairness, judged against local referents and driven by perception.
• Self-determination theory, the best-evidenced framework, identifies autonomy, competence, and relatedness, and warns about crowding out intrinsic motivation.
• The job characteristics model gives concrete redesign levers, and financial incentives work best on simple, countable, individual output.

Sources

1. Bright, D. S., & Cortes, A. H. (2019). Content theories of motivation. In Principles of Management. OpenStax, Rice University. openstax.org
2. Black, J. S., Bright, D. S., et al. (2019). Organizational Behavior, motivation chapters. OpenStax, Rice University. openstax.org
3. Wikipedia. (2025). Maslow's hierarchy of needs. en.wikipedia.org
4. Wikipedia. (2025). Expectancy theory. en.wikipedia.org
5. Wikipedia. (2025). Self-determination theory. en.wikipedia.org
6. Wikipedia. (2025). Job characteristics theory. en.wikipedia.org

The big picture

Leadership is the most written-about and least honestly reported subject in management. Airport bookshops sell the memoirs of executives who succeeded, and the stories are constructed backwards from the outcome: the trait that looks like vision in a winner looks like stubbornness in the same person had the market moved differently. The result is a field where confident storytelling has outrun evidence, and where you can be handed contradictory advice with equal conviction on the same afternoon.

So this lesson does two things at once. It teaches the models you are expected to know, because they structure how organizations talk about leadership and you need the vocabulary. And it tells you which ones have real support, which are useful heuristics with thin evidence, and which are mostly folklore. Leadership itself we can define plainly: influencing people to work willingly toward a shared goal. Note what is missing from that definition. No title, no charisma, no authority. Those help sometimes, and none of them is the thing itself.

Traits: what a century of looking actually found

The oldest theory is that leaders are born with distinguishing qualities, the so-called great man view. Researchers have looked hard for a century, and the honest summary is that traits matter modestly and no profile is decisive. Meta-analyses relating personality to leadership find the clearest signals for extraversion, conscientiousness, and openness, with correlations in a range that predicts something but far from everything. Integrity and general cognitive ability also help. But the effects are small enough that a room of effective leaders will not look alike, and any confident claim that leaders are or are not born is overstating the data in one direction or the other.

There is a specific trap worth naming: the gap between leader emergence and leader effectiveness. The traits that get someone chosen as leader, such as talking early, talking often, and projecting confidence, are only weakly related to the traits that make a team perform well once they are in charge. Groups reliably mistake confidence for competence. This should change your behavior in two ways: be suspicious of your own instinct about who "seems like a leader" in a meeting, and if you are quiet, understand that emergence is a skill you can practice separately from effectiveness rather than evidence that you cannot lead.

Key idea: Trait research finds modest, real associations and no decisive profile, and the traits that make people emerge as leaders are not the ones that make them effective.

Behavior and situation: two useful dimensions and one caution

When traits disappointed, researchers turned to behavior. Studies at Ohio State and Michigan in the 1940s and 1950s converged on two independent dimensions that remain the most durable finding in the field. Initiating structure, sometimes called task orientation, covers defining roles, setting standards, scheduling, and clarifying expectations. Consideration, sometimes called people orientation, covers respect, trust, warmth, and concern for members' welfare. Because the dimensions are independent, a leader can be high on both, and meta-analytic work supports what the grid version of this idea assumed: both dimensions relate positively to outcomes, consideration somewhat more strongly to satisfaction and initiating structure somewhat more strongly to performance. Blake and Mouton's managerial grid plots the two and recommends being high on each, which is sound if unsurprising advice.

Contingency models then argued that the right behavior depends on the situation. Fiedler's model classified situations by leader-member relations, task structure, and position power, and argued that task-oriented leaders do better in very favorable or very unfavorable situations while relationship-oriented leaders do better in the middle. House's path-goal theory holds that the leader's job is to clear obstacles and clarify the route to rewards, choosing directive, supportive, participative, or achievement-oriented behavior according to what the task and the follower lack. Path-goal has held up reasonably well as a way of thinking.

Now the caution, because it is the most-taught model in corporate training. Hersey and Blanchard's situational leadership prescribes matching four leader styles, telling, selling, participating, and delegating, to follower readiness. It is intuitive, memorable, and widely licensed, and its empirical support is weak: attempts to validate the specific prescriptions have generally failed. Keep the underlying insight, which is that new people need direction and experienced people need room, because that is sensible and consistent with better-supported work. Do not treat the four-box diagram as a validated instrument, and be suspicious when a consultant charges by the certification.

Key idea: Task and people orientation are independent and both matter, the situation genuinely conditions what works, and the popularity of a situational model is not evidence for it.

Transformational, transactional, and servant leadership

James MacGregor Burns distinguished transactional leadership, an exchange in which performance is rewarded and deviations corrected, from transformational leadership, which raises followers' motivation and connects the work to a larger purpose. Bernard Bass developed the transformational side into four components, easy to remember as the four I's: idealized influence, meaning modeling behavior that earns trust; inspirational motivation, meaning articulating a vision people want to pursue; intellectual stimulation, meaning challenging assumptions and inviting new approaches; and individualized consideration, meaning coaching each person according to their own needs.

Transformational leadership scores correlate with satisfaction, commitment, and performance across many studies, and it is the dominant framework in leadership education. It also attracts a serious methodological critique worth understanding. The construct mixes behaviors with their supposed effects, so items that ask whether a leader inspires you are partly measuring the outcome the theory is meant to predict; ratings usually come from the same followers who rate the outcomes, which inflates correlations; and the components tend to be so highly intercorrelated that the four-part structure is hard to defend. The practical translation: the behaviors are good advice, and the effect sizes in the literature are probably optimistic.

Transactional leadership is often dismissed by comparison, which is a mistake. Contingent reward, meaning clear expectations and reliable follow-through on what was promised, is consistently one of the better predictors of performance in the same research. Most of what a competent manager does is transactional in this sense, and doing it well is a precondition for anything more ambitious. A leader who inspires but does not deliver the promised raise has broken the instrumentality link from the previous lesson, and no amount of vision repairs it.

Servant leadership, articulated by Robert Greenleaf in 1970, inverts the usual framing: the leader's primary role is to serve the people doing the work, removing obstacles, developing people, and treating authority as a responsibility rather than a privilege. Evidence is more limited than for transformational models but generally positive, particularly for trust, retention, and citizenship behavior. Its main risk in practice is misapplication as conflict avoidance. Serving your team includes telling someone their work is not good enough, which is a service to them and to everyone whose work depends on theirs.

Key idea: Transformational behaviors are worth practicing while the evidence for them is inflated by measurement problems, reliable transactional competence underpins everything, and servant leadership is service, not avoidance.

Power and influence: what actually moves people

Leadership without power is preaching. French and Raven's classic taxonomy identifies five bases, and the distinction between them predicts what happens when you are not in the room.

• Legitimate power comes from your formal position. It produces compliance and is the weakest lever for discretionary effort.
• Reward power comes from controlling desirable outcomes; it works while the rewards last and while people believe they will arrive.
• Coercive power comes from the ability to punish. It produces minimum compliance, resentment, and information hiding, and it is the most reliably overused base in management.
• Expert power comes from knowledge others need. It generates genuine commitment and survives changes in title.
• Referent power comes from being respected and trusted enough that people want to align with you. It is the strongest and the slowest to build.

The two personal bases, expert and referent, produce commitment; the three positional bases mainly produce compliance. That is the whole practical lesson, and it explains why a technically respected supervisor can move a team that an equally senior but distrusted one cannot.

Influence tactics have been studied directly, and the results are usefully blunt. Rational persuasion with real evidence, inspirational appeals connected to values people already hold, and consultation that gives people genuine input into how something is done are consistently the most effective, especially upward and laterally. Exchange, ingratiation, and personal appeals are moderately effective and wear out with repetition. Pressure, coalition, and legitimating tactics are the least effective and most likely to produce resistance, though pressure retains a role in genuine emergencies. Notice that the effective tactics all treat the other person as someone who can be given a reason.

Key idea: Expert and referent power produce commitment while positional power produces compliance, and evidence-based persuasion, values-based appeals, and real consultation outperform pressure and coalition-building.

Toxic leadership, named plainly

Most leadership courses stop at models of good leadership, which leaves you unequipped for the more common problem. Abusive supervision, the sustained display of hostile verbal and nonverbal behavior short of physical contact, has been studied for over two decades and is associated with lower performance, higher turnover, reduced wellbeing, and increased deviance among those exposed. It is not rare, and it is frequently tolerated in people who deliver results.

Two ideas help you see it clearly. First, destructive leadership is usually not a matter of one bad individual but of a triangle: a destructive leader, susceptible followers who either conform out of fear or collude for advantage, and a conducive environment with weak checks, high uncertainty, and results-only accountability. Change any corner and the pattern weakens, which is why the fix is structural, meaning upward feedback that is actually read, promotion criteria that include how results were obtained, and exit interviews someone senior sees. Second, be careful with the label. Demanding is not abusive, and giving hard feedback is not mistreatment. The distinguishing features are whether the behavior is directed at the work or at the person, whether it is sustained, and whether it is worse when the target has less power.

If you are on the receiving end, document specifics with dates, use whatever formal channel exists, and take seriously that the most common outcome of a tolerated abusive manager is that good people leave quietly. If you are a manager, the useful discipline is to ask periodically what your team is not telling you, and to notice whether the answer is nothing because everything is fine or nothing because it is not safe.

Key idea: Abusive supervision is common, costly, and enabled by susceptible followers and weak organizational checks, so the countermeasures are structural rather than personal.

Try it

A new engineering manager, promoted from within, has a team of six. He announces a vision for rebuilding the platform, holds inspiring monthly talks, and is well liked. But he has not held a one-on-one in two months, two promised promotions have not been processed because he did not submit the paperwork, and he avoids telling a struggling engineer that her work is behind. Retention is fine, output is falling. (a) Where does he sit on the two behavioral dimensions? (b) Which leadership model is he over-relying on and what is he missing? (c) Which power bases is he using and which is unavailable to him? (d) Name three specific actions for the next month.

Answer: (a) High consideration, low initiating structure: warmth without clarified expectations, standards, or follow-through. (b) He is performing the inspirational side of transformational leadership while neglecting contingent reward, the transactional core. Unprocessed promotions break instrumentality directly. (c) He has referent power and probably expert power from his engineering background; he is not using legitimate authority to set standards, and avoiding the difficult conversation forfeits the corrective side entirely. (d) Restart weekly one-on-ones with a written agenda; complete the promotion paperwork this week and tell both people what happened and when it will land; and hold a specific, documented conversation with the struggling engineer stating the gap, the standard, the support available, and the review date.

Common misconceptions

• "Leaders are born." Traits contribute modestly. Most of what predicts effectiveness is behavior, and behavior can be learned and practiced.
• "The person who takes charge in a meeting is the natural leader." That is emergence, which tracks confidence and talkativeness far more than effectiveness.
• "Situational leadership is proven." It is popular and intuitive; attempts to validate its specific prescriptions have largely failed. The general insight about matching direction to experience is sound.
• "Transformational leadership is settled science." The behaviors are useful, and the reported effect sizes are inflated by common-source ratings and constructs that mix behavior with outcomes.
• "Transactional leadership is the low form." Contingent reward, meaning clear expectations and kept promises, is among the better predictors of performance and underpins everything else.
• "Servant leadership means never saying no." Serving people includes telling them the truth about their work and protecting the team from a colleague who is not delivering.
• "Tough managers get results." Sustained hostility predicts turnover, deviance, and lower performance; demanding standards delivered respectfully are a different thing entirely.

Recap

• Leadership is influence toward a shared goal, and it does not require a title.
• Trait effects are modest, and leader emergence is distinct from leader effectiveness.
• Initiating structure and consideration are independent dimensions that both predict outcomes.
• Path-goal and contingency thinking hold up better than the widely marketed situational leadership model.
• Transformational behaviors are useful but their evidence is inflated; contingent reward remains a strong predictor; servant leadership requires candor.
• Expert and referent power produce commitment, and rational persuasion, inspirational appeals, and consultation are the most effective influence tactics.
• Abusive supervision is common and costly, sustained by susceptible followers and weak checks rather than by one villain.

Sources

1. Bright, D. S., & Cortes, A. H. (2019). Situational (contingency) approaches to leadership. In Principles of Management. OpenStax, Rice University. openstax.org
2. Black, J. S., Bright, D. S., et al. (2019). Organizational Behavior, leadership chapters. OpenStax, Rice University. openstax.org
3. Wikipedia. (2025). Transformational leadership. en.wikipedia.org
4. Wikipedia. (2025). French and Raven's bases of power. en.wikipedia.org
5. Wikipedia. (2025). Servant leadership. en.wikipedia.org
6. Wikipedia. (2025). Toxic leadership. en.wikipedia.org

The big picture

Put six capable people in a room and you do not get six people's worth of output. You might get nine, or you might get three. The difference is not the sum of the individuals, which is the thing organizations spend nearly all their hiring energy on, but the interaction between them, which most organizations leave to chance and personality.

A team differs from a group by interdependence: members need each other's work to produce a shared result for which they are collectively accountable. A group of salespeople with individual territories and individual quotas is not a team no matter how many offsites it attends, and treating it like one produces resentment. Interdependence is what creates both the possible gain and the coordination cost.

This lesson covers what makes teams effective, how conflict works, and the communication skills that carry all of it. Treat the communication section as the practical core of the whole module: most of what you have learned about motivating and leading arrives at other people through a channel you chose and words you picked.

What actually makes teams effective

Google, a company with unusual data and unusual willingness to test its own beliefs, ran a multi-year study of its own teams known as Project Aristotle. The researchers expected to find that the best teams were assembled from the best individuals, or shared friendships, or had a particular mix of personality types. That is not what they found. Who was on the team mattered far less than how the team worked together, and the strongest single factor was psychological safety, followed by dependability, structure and clarity, meaning, and impact.

Psychological safety, defined by Amy Edmondson, is the shared belief that the team is safe for interpersonal risk-taking: that you can ask a question, admit an error, or disagree without being humiliated or punished. Her original finding is a small classic in how easily data can be read backwards. Studying hospital nursing units, she found that better-performing teams reported more errors, which looked wrong until closer investigation showed the better teams were not making more mistakes, they were willing to report them. In the weaker units, errors happened and vanished silently, taking with them any chance of learning.

Two clarifications keep this from turning into mush. Psychological safety is not niceness, comfort, or lowered standards; Edmondson pairs it explicitly with accountability, and the combination of high safety and high standards is where learning happens, while high safety with low standards produces a pleasant team that achieves nothing. And it is built by specific behaviors, not declared: a manager who says "I do not know" in public, who thanks the person who raises the problem, who responds to a mistake by asking what the system allowed rather than who is at fault, and who invites the quietest person's view explicitly. It is destroyed much faster than it is built, typically by one visible punishment of honesty.

Key idea: How a team interacts matters more than who is on it, and psychological safety, paired with real accountability, is the strongest known ingredient.

Size, stages, and loafing

Team size has an arithmetic consequence people persistently underestimate. The number of possible communication links in a team of n people is n(n-1)/2. A team of 5 has 10 links, a team of 10 has 45, and a team of 15 has 105. Coordination cost grows roughly with the links while productive capacity grows with the people, which is why adding members to a struggling team so often slows it down, and why Amazon's two-pizza rule, keeping teams small enough to be fed by two pizzas, has spread so widely. As a working range, five to nine people suits most collaborative work; beyond that, split the work into subteams with a defined interface rather than expanding the meeting.

Larger teams also invite social loafing, the tendency to reduce individual effort when contribution is pooled and unidentifiable. It was first observed in the 1880s by Maximilien Ringelmann, who found that men pulling on a rope together exerted less force each than they did alone. The countermeasures are structural: keep teams small, make individual contributions visible, define who owns what, and set goals people can influence. Note that free riding is rarely laziness in isolation. It is usually a reasonable response to an environment where effort is invisible and outcome is shared, which is a design you chose.

Bruce Tuckman's stages, forming, storming, norming, performing, and later adjourning, are worth knowing as vocabulary and worth holding loosely. As a strict sequence they have never been well validated; real teams skip stages, revisit them when membership changes, and sometimes storm and perform simultaneously. What the model gets right, and it is useful, is that early conflict about roles and standards is normal rather than a sign of a broken team, and that a team which never storms has often just buried the disagreement. Use it as a reassurance and a prompt, not a diagnostic instrument.

Key idea: Coordination cost rises with the square of team size, invisible individual contribution invites loafing, and Tuckman's stages are a loose description rather than a validated sequence.

Conflict: which kind, and what to do with it

Researchers distinguish task conflict, disagreement about the work itself, from relationship conflict, interpersonal friction and animosity, and from process conflict, disagreement about who does what and how. The popular claim that task conflict is good for performance is more optimistic than the evidence: meta-analytic work finds both task and relationship conflict tend to relate negatively to performance and satisfaction, with relationship conflict clearly worse. Task conflict appears to help mainly in specific conditions, namely non-routine decision-making work, in teams with high psychological safety, when the disagreement stays about the problem. That last condition is the fragile one, because task conflict reliably turns into relationship conflict when people are stressed, tired, or unsafe.

The practical implication is to make disagreement about the work structurally easy and personally cheap: argue against the proposal rather than the proposer, ask people to state the strongest version of the view they oppose, and separate the decision meeting from the exploration meeting so people are not defending a position they adopted an hour ago.

For handling a specific conflict, the Thomas-Kilmann framework maps five styles on two axes, concern for your own outcome and concern for the other party's.

Style	Assertive?	Cooperative?	Use when
Competing	High	Low	Emergencies, safety, unpopular decisions that must be made now
Collaborating	High	High	Both sets of concerns matter and there is time to find a better third answer
Compromising	Medium	Medium	Equal power, moderate stakes, a workable answer needed quickly
Avoiding	Low	Low	Trivial issues, or when tempers need time before the conversation can be useful
Accommodating	Low	High	You are wrong, the issue matters far more to them, or preserving the relationship is the priority

No style is best, and the common failure is having only one. Chronic avoiders accumulate unspoken grievances that surface as an explosion over something small; chronic competers win arguments and lose information, because people stop bringing them problems.

Key idea: Relationship conflict damages performance, task conflict helps only under safe and non-routine conditions, and the skill is choosing among the five conflict styles rather than defaulting to one.

Virtual and hybrid teams

Distributed work removes the informal channel that used to carry most coordination: the overheard remark, the two-minute question at someone's desk, the read of a colleague's face. Nothing replaces it automatically, so what was implicit has to become explicit, which is more work and better documentation.

The evidence is genuinely mixed and worth stating carefully rather than tribally. In a randomized experiment at the Chinese travel firm Ctrip, call-centre employees who volunteered were randomly assigned to work from home for nine months; performance rose about 13 percent, roughly two thirds from working more minutes per shift with fewer breaks and sick days, and the rest from more calls per minute in a quieter environment. Attrition fell by about half and satisfaction rose. The same study found a serious counterweight: home workers were promoted at a lower rate conditional on performance, and when the firm let people choose afterwards, many returned to the office. Distance has career costs that do not show up in output metrics, and a manager running a hybrid team owns that problem.

Practical rules that survive the debate: write things down, since documentation is the substitute for proximity; define response-time expectations so that asynchronous does not mean unpredictable; make meetings either all-remote or genuinely equal, since a hybrid meeting with a dominant in-room group systematically silences remote members; invest deliberately in the relationship-building that used to happen for free; and watch promotion and assignment patterns for proximity bias in yourself.

Key idea: Remote work can raise measured productivity and retention while imposing career and coordination costs, so distributed teams need explicit documentation, response norms, and deliberate attention to who is being overlooked.

Communication: channels, listening, and feedback

The basic model is sender, encoding, channel, decoding, receiver, plus noise that distorts the message and feedback that reveals what was actually received. Nearly every managerial communication failure is a failure to close that last loop: the message was sent, and nobody checked what arrived.

Channel choice is a real decision, and media richness is the way to make it. Rich channels, face-to-face first, then video, then voice, carry tone, expression, and immediate clarification, and they suit ambiguous, emotional, or consequential messages. Lean channels, written documents, chat, email, carry precision, permanence, and searchability, and they suit complex detail, records, and anything people must refer back to. The pattern to avoid is the one everyone has experienced: layoffs by email, which uses a lean channel for the most emotionally consequential message, and a forty-message chat thread trying to settle a design decision, which uses a lean channel for something ambiguous. A useful rule: if the message might make someone anxious or angry, pick up the phone; if it must be remembered exactly, write it down; if it is both, do the call and then send the written summary.

Two skills carry disproportionate weight. The first is listening, which in a managerial context mostly means not solving. A person describing a problem is usually still working out what the problem is, and the manager who supplies an answer in the first thirty seconds gets a version of the story shaped to fit that answer. Ask what they have already tried, reflect back what you heard in your own words, tolerate the silence, and ask whether they want help thinking or a decision. That last question, asked directly, prevents a large share of frustrating conversations.

The second is feedback, and it improves enormously with structure. The situation-behavior-impact pattern gives you a script: name the specific situation, describe the observable behavior without interpretation, state the impact you can speak to. "In yesterday's client call, when the client asked about the timeline, you said we would have it Friday. I had not committed to Friday, so I had to correct you in front of them, which was awkward for both of us." Compare that with "you need to be more careful," which contains no information and invites defensiveness. Note that the pattern works identically for positive feedback, which is where most managers underinvest, and that praise gains its value from specificity: "good job" is noise, while "you caught the pricing error before it reached the customer, which saved us a refund and an apology" is data. Deliver corrective feedback privately, promptly, and about behavior rather than character, and be prepared to hear that you were wrong about the facts.

Finally, watch the curse of knowledge: once you know something, you cannot easily reconstruct not knowing it, so you skip the context newcomers need and then interpret their confusion as inattention. The fix is to check what arrived, by asking the other person to say back what they will do next, not by asking whether that made sense.

Key idea: Match channel richness to the message's ambiguity and emotional weight, listen without immediately solving, and give feedback as situation, behavior, and impact rather than as character judgment.

Try it

You lead a 14-person team split between two offices and three home workers. Symptoms: the weekly meeting is 14 people and 55 minutes with two speakers; a remote engineer's design objection was ignored twice and turned out to be right; two members privately complain that a third does almost nothing; and last month's schedule change was announced in a chat thread and half the team missed it. Diagnose and prescribe.

Answer: Size: 14 people is 91 possible communication links, well past the useful range, so split into subteams with defined interfaces and keep the full meeting short and informational. Psychological safety and proximity bias: an objection ignored twice, from the person furthest away, is a signal, so change meeting practice to ask remote members first and record decisions in writing with the objection noted. Social loafing: the complaint is about invisible contribution, so define individual ownership of specific deliverables and make status visible rather than confronting the person as a character problem. Channel mismatch: a consequential schedule change went out on a lean, ephemeral channel, so use a written announcement people can find later plus a short synchronous session for questions, and confirm receipt by asking each subteam lead to state the change back.

Common misconceptions

• "The best teams are made of the best individuals." Google's own research found that how a team interacts predicts effectiveness better than who is on it.
• "Psychological safety means being nice." It means safety for interpersonal risk, and it works only when paired with high standards and accountability.
• "A team reporting more errors is performing worse." Edmondson's hospital study found the opposite; better units reported more because reporting was safe.
• "Adding people speeds up a late project." Coordination links grow with the square of team size, so additional members often slow delivery.
• "Tuckman's stages are a validated process." They are a loose, useful description. Teams skip stages and revisit them whenever membership changes.
• "Task conflict is always healthy." It helps mainly in non-routine work in psychologically safe teams; otherwise it degrades into relationship conflict.
• "Remote work is simply better or simply worse." The randomized Ctrip evidence shows productivity and retention gains alongside a real promotion penalty, so the design of the arrangement decides the outcome.
• "Feedback should be sandwiched between compliments." The pattern trains people to discount praise and hunt for the criticism. Specific situation-behavior-impact statements work better.

Recap

• Teams differ from groups by interdependence and shared accountability, which creates both the gain and the coordination cost.
• Psychological safety is the strongest identified factor in team effectiveness and must be paired with accountability.
• Communication links grow as n(n-1)/2, so five to nine members suits most collaborative work.
• Social loafing follows invisible contribution, and Tuckman's stages describe rather than predict team development.
• Relationship conflict harms performance; task conflict helps only under safe, non-routine conditions; five conflict styles each have their moment.
• Distributed teams need explicit documentation, response norms, equal meeting practice, and vigilance about proximity bias.
• Match channel richness to the message, listen without rushing to solve, and give feedback as situation, behavior, and impact.

Sources

1. Wikipedia. (2025). Psychological safety. en.wikipedia.org
2. Wikipedia. (2025). Tuckman's stages of group development. en.wikipedia.org
3. Bloom, N., Liang, J., Roberts, J., & Ying, Z. J. (2013). Does working from home work? Evidence from a Chinese experiment. National Bureau of Economic Research. nber.org
4. Black, J. S., Bright, D. S., et al. (2019). Organizational Behavior, group and team chapters. OpenStax, Rice University. openstax.org
5. Wikipedia. (2025). Media richness theory. en.wikipedia.org
6. Bright, D. S., & Cortes, A. H. (2019). Principles of Management, chapters on teams and communication. OpenStax, Rice University. openstax.org

The big picture

Controlling is the least glamorous of the four functions and the one that decides whether the other three were real. Plans that are never checked are wishes, structures that are never audited drift, and leadership that never measures anything cannot tell inspiration from noise. Controlling is the process of measuring performance against standards and correcting course, and its purpose is future decisions, not historical record-keeping. A control system that produces beautiful reports nobody acts on is an expensive way to describe the past.

It is also the function with the most dangerous side effects, because measuring something changes it. Every metric you publish is an instruction, and people will follow the instruction you actually gave rather than the one you meant. This lesson covers how to build controls that work, the predictable ways metrics corrupt behavior with a documented case at scale, and the operational core, processes, quality, and supply chains, that most control systems are pointed at.

The control cycle and its three timings

The cycle has four steps: set standards derived from the plan, measure actual performance, compare against the standard, then take corrective action or revise the standard. The fourth step is where most systems fail, because comparing is easy and correcting is political.

Controls differ by when they act. Feedforward controls act before the work: screening suppliers, training staff, checking materials on arrival. Concurrent controls act during the work: a supervisor on the floor, real-time dashboards, a machine that stops itself on an out-of-tolerance reading. Feedback controls act after: monthly financials, customer surveys, quality inspection of finished goods. Feedback control is the most common and the weakest, because by the time it fires, the defective output exists and the money is spent. Mature operations shift effort earlier, which is why a hospital spends more on sterile technique than on infection audits.

Financial control deserves specific mention because every manager meets it as a budget. Variance analysis compares actual against budget and asks why. Suppose a department budgeted $50,000 for the quarter and spent $56,500, an unfavorable variance of $6,500, or 13 percent. The number is not the analysis. The useful questions are whether the overspend bought more output, whether the driver is price or volume, and whether the budget assumption was wrong in the first place. A variance report that generates explanations rather than decisions has become a ritual.

Key idea: Control exists to change future decisions, and shifting effort from after-the-fact feedback toward feedforward and concurrent controls prevents defects rather than counting them.

Balanced scorecards and Goodhart's law

Financial measures alone are lagging indicators: by the time the quarter's revenue is bad, the causes are months old. Kaplan and Norton's balanced scorecard, introduced in 1992, adds three forward-looking perspectives to the financial one.

Perspective	Question	Example measures for a hospital
Financial	How do we look to funders?	Operating margin, cost per case
Customer	How do patients see us?	Satisfaction, wait times, readmission rate
Internal process	What must we excel at?	Infection rate, time from admission to treatment
Learning and growth	Can we keep improving?	Staff turnover, training hours, near-miss reporting

The scorecard's real contribution is the causal chain implied by reading it upward: better learning and growth should improve internal processes, which should improve customer outcomes, which should improve financial results. If you cannot articulate that chain for your own measures, you have a dashboard, not a scorecard.

Now the hazard. Goodhart's law, in the formulation popularized by the anthropologist Marilyn Strathern, says that when a measure becomes a target, it ceases to be a good measure. The mechanism is not fraud but optimization: any proxy for a goal has slack between itself and the goal, and pressure finds the slack. Hospitals measured on emergency wait times can hold patients in ambulances. Call centres measured on call duration can hang up. Schools measured on test scores can teach the test. Police measured on cleared cases can reclassify crimes. In every case the number improves and the goal does not.

The case that shows the full arc is Wells Fargo. The bank set aggressive cross-selling targets, aiming at roughly eight products per household, and pushed them hard down the retail hierarchy with daily quotas and pressure. Employees who could not hit the numbers honestly hit them dishonestly: millions of deposit and credit card accounts were opened without customer authorization. In September 2016 the bank was fined $185 million in a combined action by the Consumer Financial Protection Bureau, the Office of the Comptroller of the Currency, and the Los Angeles City Attorney; about 5,300 employees had been terminated over the conduct; the chief executive resigned; and in 2020 the bank agreed to a $3 billion settlement with the Department of Justice and the Securities and Exchange Commission over the sales practices. Notice how many lessons from this course converge: a specific difficult goal without guardrails from Module 2, a culture whose enacted values differed from its stated ones from Module 3, expectancy pressure with an unreachable target from Module 4, and a control system that measured accounts opened rather than customers served.

Practical defenses: pair every target with a guardrail metric that must not degrade, use a small basket of measures rather than one, audit the mechanism as well as the number (are these accounts being used?), keep some measures qualitative and some sampled rather than universal, and ask the gaming question before launch: if someone wanted this number without doing the work, how would they get it?

Key idea: Balance leading and lagging measures across four perspectives, and assume that any single metric under pressure will be optimized at the expense of the goal it stands for.

Operations: processes, bottlenecks, and the arithmetic of flow

Operations management is the design and control of the process that turns inputs into outputs, and its core insight is that a process is only as fast as its slowest stage. Consider a sandwich shop with three stages: order taking can handle 60 customers an hour, assembly 40, and payment 55. Throughput is 40 an hour. Speeding up order taking to 80 changes nothing except the length of the queue in front of assembly. Only the bottleneck matters, which is the entire content of the theory of constraints and the most commonly ignored fact in operational improvement. Note also that fixing a bottleneck moves it: raise assembly to 60 and payment at 55 becomes the constraint.

Little's law relates the three quantities of any stable process: work in progress equals throughput multiplied by cycle time. If a clinic has 12 patients in the building and treats 4 per hour, average time in the system is 12 / 4 = 3 hours. The law is useful because it makes a common managerial instinct visibly wrong. Told that waits are too long, managers often admit more patients, raising work in progress; with throughput fixed by the bottleneck, that arithmetic can only lengthen the wait. To cut the wait you must raise throughput or reduce work in progress, and admitting more people does the opposite of the second.

Quality practice grew from the same numeric mindset. W. Edwards Deming, an American statistician whose ideas were adopted enthusiastically in postwar Japan, argued that most defects come from the system rather than the worker, that inspecting quality in at the end is wasteful compared with building it in, and that management must act on the process. The plan-do-check-act cycle, statistical process control, and later total quality management and Six Sigma, developed at Motorola in the 1980s and spread by General Electric, all descend from this: measure variation, find causes, reduce variation.

Lean came from the Toyota Production System, and it is worth learning from the source rather than the consultancy version. Its pillars are just-in-time, producing what is needed when it is needed rather than holding large inventories, and jidoka, roughly automation with a human touch, meaning machines and people stop production when something is wrong rather than passing a defect along. The andon cord, which any worker may pull to halt the line, is the physical expression of jidoka, and it only functions inside a culture where pulling it is welcomed rather than punished, which connects directly to the psychological safety of the previous lesson. Kaizen, continuous incremental improvement driven by the people doing the work, completes it. Lean implementations fail most often when firms import the inventory reductions, which save money immediately, without the stop-the-line authority and improvement culture, which cost management comfort.

Key idea: Throughput is set by the bottleneck, Little's law ties work in progress to waiting time, and lean works only when the inventory discipline arrives together with the authority to stop the line.

Supply chains: the efficiency and resilience trade-off

Just-in-time is a genuine advance, and it removes the buffers that used to absorb shocks. That trade-off was theoretical to many managers until a series of documented disruptions made it concrete. The 2011 earthquake and tsunami in Japan halted specialized component plants and idled vehicle production worldwide, and floods in Thailand later that year did the same for hard-drive manufacturing. In March 2021 the container ship Ever Given wedged across the Suez Canal for six days, holding up a waterway that carries a large share of global seaborne trade. Through 2021 and 2022 a semiconductor shortage forced automakers to cut production of finished vehicles for want of chips worth a few dollars each.

The lesson is not that just-in-time was a mistake. It is that inventory is insurance, and firms had priced the premium as pure waste. Sensible responses are specific rather than ideological: hold buffer stock for components that are cheap to store and catastrophic to lack; qualify a second source for single-source parts, even at higher unit cost; map your supply chain beyond tier one, since most firms discovered in 2011 and 2021 that they did not know who their suppliers' suppliers were; and shorten or diversify geography where the risk concentration is severe. Each of these costs money in normal times, which is exactly why they get cut, and why the decision belongs to senior management with an explicit statement of the risk being accepted.

Key idea: Efficiency and resilience trade off directly, buffers are insurance rather than waste, and the practical work is mapping dependencies beyond tier one and deciding consciously which risks to carry.

Try it

A regional bakery has three stages: mixing (capacity 900 loaves per shift), baking (600), and packing (750). It runs one shift a day. Management proposes spending $40,000 on a faster mixer. Meanwhile the store measures its delivery drivers solely on deliveries completed per hour, and customer complaints about damaged boxes have doubled. (a) What is current throughput, and what will the new mixer achieve? (b) Where should the $40,000 go? (c) If 150 loaves are in progress and throughput is 600 per 10-hour shift, what is average cycle time? (d) Diagnose the driver metric and fix it.

Answer: (a) Throughput is 600 loaves per shift, set by the baking bottleneck. The faster mixer raises capacity at a non-bottleneck stage and changes throughput by zero. (b) Into baking capacity, for example a second oven or an additional bake cycle, until baking exceeds 750, at which point packing becomes the new constraint. (c) Throughput is 600 / 10 = 60 loaves per hour, so cycle time is 150 / 60 = 2.5 hours. (d) The single-metric driver target is a Goodhart problem: speed is rewarded and care is not, so damage rises. Pair deliveries per hour with a guardrail on damage claims or customer complaints per hundred deliveries, and audit the mechanism by sampling deliveries rather than trusting the count alone.

Common misconceptions

• "Controlling means controlling people." It means comparing results with standards and correcting course; the object is the process, not the person.
• "More metrics equal better control." A dashboard without a causal chain from learning to process to customer to financial results is decoration, and unbalanced single metrics invite gaming.
• "Goodhart's law is about dishonest people." Wells Fargo shows scale, but the mechanism is ordinary optimization toward whatever proxy carries the pressure.
• "Improving any stage improves output." Only the bottleneck sets throughput, and improving it merely relocates the constraint.
• "Shorter waits come from admitting more people." Little's law says raising work in progress with fixed throughput lengthens waits.
• "Lean is about cutting inventory." Inventory reduction without stop-the-line authority and continuous improvement is cost-cutting wearing lean vocabulary.
• "Just-in-time was proven wrong in 2021." It was proven costly at the tail. The correct conclusion is to price buffers as insurance for specific critical inputs, not to abandon flow discipline.

Recap

• The control cycle sets standards, measures, compares, and corrects, and its value lies entirely in the correcting.
• Feedforward and concurrent controls prevent defects; feedback controls count them after the money is spent.
• The balanced scorecard adds customer, process, and learning perspectives to financial measures with an implied causal chain.
• Goodhart's law predicts that a pressured metric will be optimized against its own purpose, as Wells Fargo's cross-selling targets showed at scale.
• Throughput is set by the bottleneck, and Little's law links work in progress, throughput, and cycle time.
• Deming's system view, TQM, Six Sigma, and Toyota's just-in-time, jidoka, and kaizen form the quality and lean tradition.
• Efficiency and resilience trade off, and documented disruptions justify buffers, second sources, and mapping beyond tier one suppliers.

Sources

1. Bright, D. S., & Cortes, A. H. (2019). Principles of Management, Chapter 17: Organizational planning and controlling. OpenStax, Rice University. openstax.org
2. Consumer Financial Protection Bureau. (2016). Enforcement action regarding Wells Fargo unauthorized accounts. consumerfinance.gov
3. Wikipedia. (2025). Wells Fargo cross-selling scandal. en.wikipedia.org
4. Wikipedia. (2025). Goodhart's law. en.wikipedia.org
5. Wikipedia. (2025). Toyota Production System. en.wikipedia.org
6. Wikipedia. (2025). Balanced scorecard. en.wikipedia.org

The big picture

Ask a room of managers how many change initiatives they have lived through, then ask how many actually stuck. The gap between those two numbers is the subject of this lesson. You will often hear that seventy percent of change efforts fail. Treat the figure with suspicion, because it is repeated far more often than it is measured, and its origins are more anecdotal than empirical. What is not in doubt is that changing an organization is much harder than deciding to change it, and that most of the difficulty is predictable.

Predictable is the operative word. Change fails for a small number of recurring reasons: nobody felt the need, the destination was vague, the early steps produced nothing visible, the incentives kept rewarding the old behavior, or the people asked to change had good reasons not to that nobody bothered to hear. Every one of those is a managerial choice, which means it is fixable.

Why organizations resist

Resistance is usually described as an obstacle, and describing it that way is the first mistake. Split it into individual and organizational sources.

Individuals resist because change threatens something concrete: loss of status, skill relevance, or relationships; habit, since practised routines are cheap and new ones are exhausting; uncertainty about whether they will be good at the new thing; self-interest, which is entirely rational when the change genuinely makes them worse off; and distrust, particularly where previous initiatives were announced with fanfare and quietly abandoned. Notice that only some of these are irrational, and none of them are stupid.

Organizations resist for structural reasons that persist even when everyone individually agrees. Structural inertia means hiring, training, promotion, and reporting systems were built to reproduce current behavior and will keep doing so. Sunk costs in equipment, software, and expertise make the old way cheaper on paper. Group norms punish the person who adopts the new behavior first. And culture, from Module 3, sits underneath all of it: if the underlying assumption is that initiatives come and go, waiting is the rational strategy and everyone knows it.

Here is the reframe that separates competent change leaders from the rest. Resistance is information. When an experienced nurse objects to a new scheduling system, the objection frequently contains a fact the designers did not have. Organizations that treat resistance as an attitude problem to be overcome with communication lose that information and implement the flawed design anyway, at which point the resisters are proved right and trust falls further. Ask the objectors what breaks. Some of the time they are protecting themselves, and some of the time they are protecting you.

Key idea: Resistance combines rational self-protection, habit, distrust, and structural inertia, and the useful managerial move is to mine it for information rather than to overpower it.

Lewin and Kotter: two frameworks, used carefully

Kurt Lewin's model is the ancestor of nearly everything in this field: unfreeze, meaning create a felt need and loosen the current equilibrium; change, meaning move to the new way with support and training; refreeze, meaning lock it in with structures, incentives, and norms so it does not spring back. Its most practical component is force field analysis. Draw the driving forces pushing toward change and the restraining forces holding the current state, then note Lewin's counterintuitive advice: it is usually more effective to reduce restraining forces than to add driving force, because pushing harder against a restrained system raises tension without moving it. If nurses resist a new system because it adds twenty minutes of documentation per shift, no amount of executive urgency substitutes for removing the twenty minutes.

Two honest caveats. The tidy three-stage formulation owes as much to later interpreters as to Lewin himself, and refreezing sits awkwardly in environments that change continuously. Use it as a checklist rather than a physics.

John Kotter's eight-step model, published in 1996, remains the most widely used practitioner framework: establish urgency, build a guiding coalition, form a vision and strategy, communicate the vision, empower action by removing obstacles, generate short-term wins, consolidate gains, and anchor the change in the culture. Three steps carry most of the weight in practice. Urgency is where most efforts start too weak, because a change that feels optional will be treated as optional. Short-term wins matter more than they sound, since a change program with no visible result for nine months has already lost the middle of the organization. And anchoring is the step that gets skipped, because it means changing promotion criteria and incentives, which is slow and contested, and without it the old behavior returns within a year.

Criticisms are fair and worth carrying: the model is linear where real change loops, it is heavily top-down, and its evidence base is practitioner observation rather than controlled study. Use it as a set of prompts for what you have forgotten, which is what it is genuinely good for.

Key idea: Reduce restraining forces rather than pushing harder, and treat Kotter's steps as prompts, paying particular attention to genuine urgency, visible early wins, and anchoring in incentives.

Exploration, exploitation, and ambidexterity

James March drew the distinction that organizes thinking about innovation. Exploitation is getting better at what you already do: refining, improving efficiency, extending the existing product to existing customers. Returns are near-term, reliable, and measurable. Exploration is search: new markets, new technologies, new business models. Returns are distant, uncertain, and mostly negative for any individual attempt.

Every organization needs both, and the two compete for the same resources under decision rules that systematically favor exploitation. Exploitation projects have a business case, a return on investment calculation, and a champion who can promise a number; exploration projects have a story. When budgets tighten, the exploration budget goes first, and the firm becomes progressively better at doing something that is progressively less valuable. This is the success trap, and it looks like prudence right up to the end.

Organizational ambidexterity is the practice of doing both. Structural separation, giving the exploratory unit its own metrics, funding, and freedom from the core business's processes, is the common answer, and its known failure mode is equally common: the new unit is starved, absorbed, or judged on the core business's short-term measures and dies. The management requirement is unglamorous, meaning protected funding, different success criteria, senior sponsorship that survives a bad quarter, and clear rules about when an exploratory bet graduates or stops.

Key idea: Exploitation reliably outcompetes exploration inside normal budgeting, so exploration survives only with protected funding, different metrics, and senior cover.

Disruption theory, used critically

Clayton Christensen's disruptive innovation is the most influential and most misused idea in this area, so learn it precisely. His claim was narrow: a new entrant takes root either at the low end of a market or in a new market with an offering that established firms rationally ignore because it is worse on the dimensions their best customers care about and worse for their margins. The entrant improves along its own trajectory until it satisfies mainstream needs, at which point the incumbent's customers leave. The word rationally is the heart of it. The incumbent does not fail through stupidity; it fails by listening to its best customers and protecting its most profitable business, which is precisely what good management textbooks tell it to do.

Two documented histories illustrate the pattern and its nuances. Blockbuster, at its height a dominant video rental chain, was reportedly offered the chance to acquire the young Netflix around 2000 for a modest sum and declined; Netflix's mail-and-then-streaming model grew, and Blockbuster filed for bankruptcy in 2010. Kodak is the more instructive case because the popular version is wrong: Kodak did not miss digital photography, it built the first digital camera prototype in 1975 and invested heavily in digital technology. What it could not solve was that digital photography destroyed the film and processing economics that produced its profits, and no amount of foresight repeals that. Kodak filed for Chapter 11 in 2012. The real lesson is that seeing the change is not the hard part; restructuring a profitable business to compete with a less profitable future is.

Now the criticism, which you should know because disruption has become a word people use to mean any change they like. Scholars have questioned both the theory's evidence and its predictive power: one study examining the cases most often cited as disruption found that only a small minority fit the theory's own criteria. Christensen's own predictions went wrong publicly, including his expectation that the iPhone would not succeed against established handset makers, which is a fair illustration that the theory explains better retrospectively than it forecasts. And the term is now applied to firms that simply grew fast with a better product, which is competition rather than disruption. Use the framework for what it is good at: explaining why competent incumbents ignore cheap, low-margin entrants until it is late. Do not use it as a prediction machine or as an excuse for any change you happen to favor.

Key idea: Disruption describes a specific mechanism by which rational incumbents lose to low-end or new-market entrants, and it explains history better than it predicts the future.

Try it

A 400-person insurance firm is replacing paper claims handling with a digital platform. Six months in: adoption is 31 percent, the two most experienced adjusters refuse to use it, the executive sponsor announced it at a town hall and has not mentioned it since, no manager's targets changed, and adjusters say the new system takes 40 minutes per claim against 25 on paper. (a) Draw the force field. (b) Which Kotter steps were skipped? (c) What single action would you take first, and why not more urgency?

Answer: (a) Driving forces: executive mandate, long-term cost savings, regulatory reporting benefits. Restraining forces: the system is genuinely slower per claim, experienced staff lose expertise advantage, no incentive change, no visible sponsorship, and the reasonable belief that this too shall pass. (b) Urgency was announced rather than created, there is no guiding coalition of respected adjusters, there are no short-term wins, obstacles were not removed, and nothing was anchored in targets or promotion criteria. (c) First action: fix the 40 minutes. It is the largest restraining force and it is a fact, not an attitude, so more urgency would only raise tension against a real obstacle. Involve the two resisting senior adjusters in the redesign, since their objection contains the information you need and their endorsement is worth more than any announcement.

Common misconceptions

• "Seventy percent of change efforts fail." Widely repeated and poorly evidenced. Change is hard for identifiable reasons; the statistic adds drama rather than knowledge.
• "Resistance is an attitude problem." It is often a rational response, and it frequently contains information about a design flaw.
• "Communicate more and resistance will fade." Communication cannot remove a real obstacle. Reduce restraining forces first.
• "Kotter's steps are validated stages." They are practitioner prompts, linear where real change loops, and most useful as a checklist for what you skipped.
• "Innovation means exploration only." Exploitation funds the firm; the failure mode is letting normal budgeting starve exploration into extinction.
• "Kodak failed because it missed digital." It invented an early digital camera and invested substantially. Its problem was that digital destroyed the economics of film, not that it did not see it.
• "Every fast-growing competitor is a disruptor." Disruption has a specific meaning involving a low-end or new-market foothold that incumbents rationally ignore.

Recap

• Individual resistance comes from loss, habit, uncertainty, self-interest, and distrust; organizational resistance from inertia, sunk costs, norms, and culture.
• Resistance carries information about design flaws and should be mined rather than overpowered.
• Lewin's unfreeze-change-refreeze and force field analysis direct effort toward reducing restraining forces.
• Kotter's eight steps are useful prompts; genuine urgency, short-term wins, and anchoring in incentives carry most of the weight.
• Exploration and exploitation compete for resources under rules that favor exploitation, which produces the success trap.
• Ambidexterity requires structural separation with protected funding and different success measures.
• Disruption explains why rational incumbents lose to low-end entrants, and it is a better retrospective explanation than a forecasting tool.

Sources

1. Bright, D. S., & Cortes, A. H. (2019). Principles of Management, Chapter 10: Organizational structure and change. OpenStax, Rice University. openstax.org
2. Wikipedia. (2025). Kurt Lewin. en.wikipedia.org
3. Kotter, J. P. (1995). Leading change: Why transformation efforts fail. Harvard Business Review, 73(2), 59-67. hbr.org
4. Wikipedia. (2025). Disruptive innovation. en.wikipedia.org
5. King, A. A., & Baatartogtokh, B. (2015). How useful is the theory of disruptive innovation? MIT Sloan Management Review, 57(1). sloanreview.mit.edu
6. Wikipedia. (2025). Netflix. en.wikipedia.org

The big picture

Most business ethics failures are not committed by villains. They are committed by ordinary people who are busy, who face a target, who assume someone above them has thought about it, and who take one small step that makes the next step easier. You have already met several in this course: the Sears mechanics with repair quotas, the Wells Fargo branch staff with cross-selling targets, the NASA and contractor managers who inverted the burden of proof the night before Challenger. None of those required an unusual person. They required an ordinary person inside a particular system.

That is why ethics belongs in a management course rather than only in a philosophy one. You will design the systems. The practical question is not whether you personally are a good person, which you cannot verify and which will not save you, but whether the goals, metrics, and pressures you create leave ordinary people a path to do the right thing without heroism.

This lesson gives you four reasoning frameworks and works them on a documented case, covers whistleblowing honestly including its personal cost, presents the corporate responsibility debate with its best arguments on more than one side, and closes with managing across cultures using frameworks that are useful only when handled carefully.

Four ways to reason about a dilemma

Ethical frameworks do not produce answers automatically. What they do is force different questions, which is exactly what you need when your first instinct is to rationalize.

• Utilitarian: which action produces the greatest overall good across everyone affected? Strength: it forces you to count people outside the room. Weakness: it can justify serious harm to a minority, and it invites convenient arithmetic.
• Rights and duties: what does each person have a right to, and what duties do I have regardless of consequences? Strength: some things stay off the table, including deceiving people for their own good. Weakness: rights conflict, and the framework gives limited guidance for ranking them.
• Justice and fairness: are burdens and benefits distributed fairly, and was the process itself fair? Distributive justice asks about outcomes and procedural justice about how they were decided, and research consistently shows that people accept unfavorable outcomes far better when the process was transparent and consistent.
• Virtue: what would a person of good character do here, and what does this choice make me into? Strength: it captures the cumulative nature of small compromises. Weakness: character talk can float free of specifics.

Work them on a real case. General Motors used an ignition switch that could slip out of the run position, cutting engine power and disabling airbags. Internal knowledge of the defect existed years before the company issued recalls in 2014. In 2015 GM entered a deferred prosecution agreement and paid a $900 million penalty, and a compensation fund it established accepted claims for at least 124 deaths. Run the frameworks on the decision faced by an engineer or middle manager who saw the data in, say, 2006. Utilitarian reasoning appears to favor recall as soon as expected deaths and injuries are weighed against recall cost, and note how easily it can be corrupted by putting a low number on a statistical life. Rights reasoning is more decisive: drivers have a right to material safety information about a product they bought, and withholding it treats them as means to a cost target. Justice asks who bears the burden of delay, namely people with no information and no choice, while the benefit goes to the firm's short-term costs. Virtue asks what kind of engineer signs off, and what the twentieth such sign-off makes of a career. Three of four frameworks converge, which is common in serious cases and is itself useful information: when frameworks converge and you still hesitate, the obstacle is usually career risk rather than moral complexity.

Key idea: Frameworks are question generators, and when utilitarian, rights, justice, and virtue reasoning converge, the remaining hesitation is usually about personal cost rather than ethical difficulty.

Speaking up, and what it costs

Most ethical failures are visible to someone before they become public. The question is what happens to that person. Two well-known cases came from the accounting scandals of the early 2000s: Sherron Watkins wrote to Enron's chairman warning that the company might implode in a wave of accounting scandals, and Cynthia Cooper's internal audit team at WorldCom uncovered massive improper accounting entries. Both events fed directly into the Sarbanes-Oxley Act of 2002, which added corporate governance requirements and protections for employees of public companies who report suspected fraud. The Dodd-Frank Act of 2010 went further by creating a Securities and Exchange Commission whistleblower program that can award a percentage of sanctions collected in successful enforcement actions, and various federal statutes protecting workers who report safety and other violations are administered through the Department of Labor.

Legal protection is real and it is not the whole story. Whistleblowers frequently face isolation, stalled careers, and long legal processes even when they are ultimately vindicated. Any honest treatment says so. The practical guidance for a manager is therefore in two parts. If you are considering raising a concern: document specifics with dates, use internal channels first where it is safe and where the wrongdoing is not being directed by those channels, understand which statute if any covers your situation, and get advice before acting rather than after. If you are the manager receiving concerns: how you respond to the first small complaint determines whether you ever hear a large one. Thank people for raising issues, investigate rather than explain, close the loop with the person who raised it, and never let a reporter suffer for reporting, since retaliation is both unlawful in many circumstances and the single most efficient way to guarantee that the next problem stays hidden until it is a crisis.

Key idea: Legal protections for whistleblowers exist and are imperfect, so a manager's real contribution is making internal reporting safe enough that external reporting is unnecessary.

Corporate responsibility and the ESG argument

Whose interests should a company serve? Milton Friedman's 1970 essay gave the sharpest version of one answer: the social responsibility of business is to increase its profits, within the rules of the game, because executives are agents spending other people's money and social priorities belong to legislatures and individuals. The strongest counterposition is stakeholder theory, articulated by R. Edward Freeman in 1984, which holds that a firm's obligations run to everyone who affects or is affected by it, employees, customers, suppliers, communities, and shareholders, and that serving those groups well is both right and, over long horizons, the way value is actually created.

The triple bottom line, coined by John Elkington in 1994, proposed measuring performance on people, planet, and profit. Worth knowing: in 2018 Elkington himself published a public reconsideration, arguing that the concept had been widely adopted as an accounting exercise rather than the systemic change he intended. It is rare and instructive for the inventor of a management concept to say so.

That brings us to ESG, the practice of assessing companies on environmental, social, and governance factors. Present the debate properly, because you will meet it in the wild and both sides have real arguments.

The case for: several ESG factors are straightforwardly financially material, since climate exposure, safety records, and governance quality all predict future liabilities; investors managing long-horizon money have a legitimate interest in risks that materialize over decades; and disclosure of comparable information lets markets price what was previously invisible.

The case against, or for caution: ratings are inconsistent, and research examining major providers found their ESG scores correlate with each other far more weakly than credit ratings do, which means a company can be simultaneously excellent and poor depending on who is scoring; aggregating three unlike dimensions into one letter grade obscures more than it reveals; the framework invites greenwashing, where disclosure substitutes for change; and critics argue that fund managers should not use other people's retirement savings to pursue social objectives those savers did not choose. There is also a live political and legal dispute in the United States over the use of ESG criteria by fiduciaries, which has changed direction with successive administrations.

A defensible position for a manager, which does not require winning the political argument: treat material risks as risks and manage them because they are risks, disclose honestly rather than promotionally, and be specific about what your firm actually does rather than adopting a rating as a substitute for a decision.

Key idea: The responsibility debate is genuinely contested, ESG ratings from different providers agree far less than people assume, and the durable managerial practice is to manage material risks and disclose honestly rather than to buy a label.

Managing across cultures without stereotyping

Geert Hofstede's dimensions are the most cited framework in cross-cultural management: power distance, the degree to which unequal power is accepted; individualism versus collectivism; a dimension originally labeled masculinity versus femininity, describing emphasis on competitive achievement versus quality of life; uncertainty avoidance, the discomfort with ambiguity; long-term orientation; and indulgence versus restraint. Used well, the framework alerts you that your assumptions about meetings, disagreement, deadlines, and hierarchy are local rather than universal. In a high power distance setting, an engineer may not contradict a senior manager in public, which means the Challenger safeguard of asking the most junior person first needs redesign rather than repetition.

Now the cautions, which matter more than the dimensions. The original data came from surveys of IBM employees around 1967 to 1973, a single company in a single era. Nations are not cultures, and treating a country as culturally uniform ignores region, generation, class, and organization. Most importantly, the scores are country averages and say nothing reliable about the individual in front of you; concluding things about a person from a national mean is the ecological fallacy, and it is where the framework does real damage. Scholars including Brendan McSweeney have criticized both the method and the inferences, and the later GLOBE project offers an alternative set of dimensions with its own debates.

The practical stance is to use the frameworks as hypotheses about what might differ, then verify with the actual people. Ask how decisions normally get made here. Notice whether silence means assent or dissent. Expect that written and spoken directness varies, that time commitments vary in strictness, and that your own norms feel like common sense to you for exactly the same reason theirs do to them. One legal note for managers of international operations: the U.S. Foreign Corrupt Practices Act of 1977 prohibits bribing foreign officials to obtain or retain business, with accounting provisions attached, and the fact that a payment is customary somewhere does not make it lawful for a covered firm.

Key idea: Cultural frameworks are hypotheses about country averages, not descriptions of individuals, and applying a national score to the person in front of you is the ecological fallacy.

Try it

You manage a 40-person team at a firm that must cut costs by 15 percent. Options: lay off six people; cut everyone's hours and pay by 15 percent; or cut the training budget, the bonus pool, and two open positions while leaving current staff intact but with a heavier workload. (a) Analyze with each of the four frameworks. (b) Which procedural fairness steps would you take regardless of the option chosen?

Answer: (a) Utilitarian: hour cuts spread a smaller harm widely and preserve capability, while layoffs concentrate severe harm on six people and their households; but if hour cuts trigger resignations among your strongest staff, total harm may end up higher, so the arithmetic depends on facts you should gather rather than assume. Rights and duties: whichever route you take, people have a right to honest, timely information and to not be misled about their prospects; a duty of care argues against string-along pay cuts if a layoff is inevitable later. Justice: distributive justice asks whether cuts fall on those least able to bear them, and procedural justice governs how the decision is made and explained. Virtue: what would an honest and courageous manager do, and would you be willing to describe your reasoning publicly to those affected? (b) Regardless of option: explain the actual financial situation with real numbers, state the criteria before applying them, apply them consistently and document the application, tell people individually and in person rather than by mass email, give as much notice as is lawful and practical, provide references and support for anyone leaving, and tell the remaining staff the truth about whether more cuts are expected.

Common misconceptions

• "Ethics failures come from bad people." Most come from ordinary people inside systems with pressure, targets, and diffuse responsibility.
• "Legal equals ethical." Law is a floor negotiated in the past. Ethical judgment is required precisely where law is silent or lagging.
• "Ethical frameworks give you the answer." They generate the questions your rationalizations are trying to skip.
• "Whistleblower laws make speaking up safe." They provide real but incomplete protection, and the personal cost is often significant even for vindicated reporters.
• "ESG ratings measure the same thing." Major providers' scores correlate weakly with one another, so a single rating is a weak basis for any decision.
• "The triple bottom line is beyond criticism." Its own originator publicly reconsidered it in 2018 as too often reduced to accounting.
• "Hofstede tells you how to treat this colleague." Country averages say nothing reliable about an individual; assuming otherwise is the ecological fallacy.

Recap

• Systems, not just character, produce ethical failure, so managers design the conditions in which ordinary people choose.
• Utilitarian, rights, justice, and virtue reasoning each ask a different question, and convergence is a strong signal.
• The GM ignition switch case shows the cost of delay: a $900 million penalty in 2015 and a fund accepting claims for at least 124 deaths.
• Sarbanes-Oxley and Dodd-Frank created whistleblower protections and awards, but speaking up remains personally costly.
• Friedman's shareholder view and Freeman's stakeholder view frame the responsibility debate, with the triple bottom line reconsidered by its own author.
• ESG has real materiality arguments and real measurement problems, including weak agreement among rating providers.
• Hofstede's dimensions are useful hypotheses about country averages and dangerous when applied to individuals.

Sources

1. Bright, D. S., & Cortes, A. H. (2019). Principles of Management, ethics and social responsibility chapters. OpenStax, Rice University. openstax.org
2. U.S. Securities and Exchange Commission. (2025). Office of the Whistleblower. sec.gov
3. National Highway Traffic Safety Administration. (2025). Vehicle recalls and safety issues. nhtsa.gov
4. Encyclopaedia Britannica. (2024). Business ethics. britannica.com
5. Wikipedia. (2025). Hofstede's cultural dimensions theory. en.wikipedia.org
6. Wikipedia. (2025). Environmental, social, and governance. en.wikipedia.org

The big picture

Everything in this course so far has been about how organizations work. This lesson is about you, and specifically about the first ninety days of the job where all of it stops being theory. It also covers the setting where a manager wears every hat at once, which is the small business, because that is where a large share of people actually manage.

The transition into management is genuinely difficult, and it is difficult for a reason that is easy to state and hard to absorb: you are no longer paid for your own output. You are paid for your team's output, which means the activities that used to signal a good day, tasks finished, code shipped, calls closed, are now mostly signs that you are doing someone else's job. New managers commonly feel unproductive for months. That feeling is not evidence of failure; it is evidence that the job changed and you noticed.

Managing a small business or a startup

Small firms dominate the count of American employers, and their management problem is the opposite of a corporation's. There is no specialization to lean on. The owner is the head of sales, the head of HR, the bookkeeper, and the person who unclogs the sink, and the constraint is never analysis, it is attention.

Two facts should shape expectations. First, survival is uneven: U.S. Bureau of Labor Statistics data on business employment dynamics show that roughly one in five new establishments closes within the first year, about half survive five years, and roughly a third reach ten. Second, the proximate killer is usually cash rather than profit. A growing small business can be profitable on paper and insolvent in practice, because inventory and receivables consume cash before customers pay. If you take one operational habit from this section, make it a thirteen-week rolling cash forecast, updated weekly, showing money in and money out. It is the single most useful management artifact a small business can maintain.

On planning approach, the honest summary is that the evidence supports neither extreme. Elaborate business plans written before contact with customers are often exercises in confirming what you already believe, which is the confirmation bias from Module 2 with a spreadsheet attached. The lean startup response, meaning customer discovery interviews, a minimum viable product, and a willingness to pivot, is a sensible correction, and it too can become theater if a firm keeps testing and never commits. Do enough planning to establish the unit economics, then get real evidence from paying customers as quickly as possible, and treat both as inputs rather than rituals.

The predictable crisis point arrives at roughly ten to fifteen people. Below that, the founder can hold everything in their head and coordinate informally. Above it, informal coordination fails: nobody knows who decides, quality varies, and the founder becomes the bottleneck for every decision. The fix is exactly the material from Module 3, meaning explicit ownership, a few standing plans for recurring decisions, and delegated authority to match delegated responsibility. Founders who refuse this transition cap their business at their own personal capacity, which is a legitimate choice as long as it is a choice.

Key idea: Small-business management is constrained by attention and cash rather than analysis, and the coordination methods that work at eight people fail at fifteen.

Delegation, with the arithmetic

Delegation is the skill new managers most often fail at, and the reason is always the same sentence: it is faster to do it myself. In the moment that sentence is true, which is why it is so persuasive. Over any horizon longer than a week it is usually false, and the arithmetic settles it.

Suppose a recurring task takes you 30 minutes. Teaching someone costs 20 minutes, and their first attempt takes 60 minutes, of which 30 is the extra time you effectively absorb through review and correction. Your total investment is 20 + 30 = 50 minutes. From the second occurrence onward, that task costs you nothing and them 30 to 40 minutes. Break-even arrives after about two repetitions. If the task happens weekly, you have bought back roughly 25 hours a year, and you have also created a second person who can do it when you are away, which is worth more than the hours.

Delegation fails when it is ambiguous about authority, so state the level explicitly. A simple ladder: (1) do exactly this and tell me when it is done; (2) investigate and bring me options with a recommendation; (3) decide, then tell me before you act; (4) decide and act, tell me afterwards; (5) decide and act, no need to report routinely. Most conflict between managers and reports about delegation is a mismatch of level, where the manager assumed 2 and the employee heard 4. Say the number.

What to delegate: recurring tasks, work that develops someone, anything where your involvement adds no judgment, and, most importantly, whole pieces rather than fragments, since task identity from Module 4 is what makes delegated work motivating. What not to delegate: performance conversations, hiring decisions for your own team, and anything where you alone hold the necessary context or accountability. And delegate the outcome with its constraints, not the method, unless the method matters for safety or compliance. If you specify every step, you have not delegated, you have dictated slowly.

Key idea: Delegation pays back after about two repetitions of a recurring task, and it works only when you name the level of authority explicitly and hand over whole outcomes rather than fragments.

The one-on-one, and scripts for hard conversations

The single highest-return habit for a new manager is a regular one-on-one with each person: 30 minutes, weekly or fortnightly, in the calendar, rarely cancelled. Cancelling it is the loudest thing you can say about priorities, and rescheduling twice teaches people not to prepare.

Structure that works. Their agenda first, because a meeting that opens with your status questions is a status meeting wearing a disguise. Then three standing questions: what is blocking you that I can remove, what did you learn or find frustrating this week, and is there anything you have been hesitating to tell me? The third question is doing psychological safety work from Module 4 and will produce nothing for the first month and something important in the second. Keep a running document with commitments in both directions and check it. Once a quarter, spend the whole session on career: where they want to be in two years, what evidence would show progress, and what you will do about it. And ask directly when you are unsure what someone wants from a conversation: do you want help thinking, or do you want a decision?

For the hard conversations, having actual sentences ready matters more than theory, because the failure mode is vagueness under discomfort. Use situation, behavior, impact from Module 4, and prepare the opening line.

• Missed deadline. "The migration doc was due Friday and I got it Tuesday. I had already told the client Friday, so I had to walk that back. What happened, and what would make the next one land on time?" Then stop talking. The silence is the useful part.
• Interrupting colleagues. "In this morning's review you spoke over Priya twice while she was explaining the schema. She stopped contributing after the second time, and we probably lost something. Were you aware of it?"
• Sustained underperformance. "I want to be direct because you deserve clarity. The standard for this role is closing 20 tickets a week at under 5 percent reopen rate. You are at 11 and 12 percent, and that has been the pattern for two months. I want to fix it with you. Here is the support I can offer, here is the timeline, and I will write down what we agree today."
• Praise, which most managers underuse. "You caught the pricing error before it reached the customer. That saved a refund and an apology, and it is the second time this month you have found something in the checks. I noticed."
• You were wrong. "I pushed the Thursday date without asking about your other commitments, and that was my mistake. I am not going to do that again. What do you need from me now?" Saying this in public costs you nothing you actually had and buys credibility you cannot purchase any other way.

Two more notes for the first ninety days. Managing former peers is uncomfortable and gets better fastest through directness: acknowledge the change once, explicitly, ask what they need from you now, and then behave consistently rather than apologetically. And resist the urge to make large changes in the first month. Learn the system first, since much of what looks stupid from outside turns out to be scar tissue from a problem you have not met yet. Fix the small, obvious irritations early, because that builds credibility, and save the structural changes until you understand why the current arrangement exists.

Key idea: Protected one-on-ones with the employee's agenda first, plus prepared, specific sentences for hard conversations, do more for a new manager's effectiveness than any framework in this course.

Careers in management

Management is not one career, it is a layer that exists in every field, which is why the most reliable route in is depth in something plus demonstrated ability to organize other people. The U.S. Bureau of Labor Statistics reports management occupations as the highest-paid major occupational group, with a median annual wage well above $100,000, and it projects continued growth in the group. Read that with the corresponding cost in mind: longer hours, responsibility for outcomes you do not personally control, and the permanent low-grade discomfort of being evaluated on other people's work.

Practical paths. The most common is promotion from a specialist role, which is why building human and conceptual skills early, from Module 1, matters before you need them. Project and program management is a second entry point, since it grants coordination responsibility without formal authority, which is the harder and more transferable version of the skill. General management programs and MBAs are a third, most valuable for career switchers and for network access rather than for the content, much of which you have just read. Small business and franchise ownership is a fourth, and running a team of six in your own business teaches more per week than any course.

How to get experience before you have the title: run a project end to end and be the person who writes down decisions, volunteer to onboard new hires, take the coordination role nobody wants in a cross-functional effort, and practice giving specific feedback to peers who invite it. Every one of those is real managerial work, and every one is visible to the people who decide about promotions.

Finally, a closing thought for the course. You now know the four functions, the major frameworks, and, importantly, which of them the evidence supports and which are folklore with good branding. That last distinction is the most valuable thing here, because you will spend a career being handed confident advice. Ask what problem it solves, what evidence supports it beyond testimonials, and what it costs if it is wrong in your situation. Then watch what happens to the people around you when they act on it, which is the only data you will ever have that is genuinely about your own organization.

Key idea: The reliable route into management is depth plus demonstrated coordination of others, and the durable skill from this course is judging management claims by their evidence and cost of being wrong.

Try it

You have just been promoted to lead your former team of seven. You inherit: a weekly two-hour team meeting nobody likes, a report you personally used to produce every Monday that takes 90 minutes, one former peer who applied for your job, and a backlog you believe is badly prioritized. Plan your first month. (a) What do you do in week one? (b) What do you delegate and at what authority level? (c) Compute the delegation payback for the Monday report if teaching costs 45 minutes and the first two runs take an extra 45 minutes of your review time in total. (d) What do you not change yet?

Answer: (a) Schedule 30-minute one-on-ones with all seven and ask three questions: what should I keep doing, what is getting in your way, and what would you change if you were me. Have the conversation with the former peer who applied for your job explicitly and early, acknowledging it once and asking what they want from you now. (b) Delegate the Monday report at level 4, decide and act then tell me afterwards, since it is recurring, teachable, and a whole deliverable. Keep hiring decisions and performance conversations. (c) Investment is 45 + 45 = 90 minutes; the report saves you 90 minutes each week thereafter, so break-even arrives after one week and the annual saving is roughly 90 minutes x 50 weeks, about 75 hours. (d) Do not restructure the backlog in month one. Ask why it is prioritized as it is, because the answer often contains a constraint you cannot see yet. The two-hour meeting is fair game to shorten immediately, since it is a visible irritation with low structural risk, and fixing it early buys credibility for the harder changes later.

Common misconceptions

• "A new manager should prove themselves by producing more." Your output is now your team's output. Doing their work is the most common first-year failure.
• "It is faster to do it myself." True this once, false over any horizon. Break-even on a recurring task typically arrives after about two repetitions.
• "Delegating means telling people exactly how." Hand over the outcome and its constraints, and name the authority level so nobody guesses.
• "One-on-ones are a status update." Status can be written. The meeting is for obstacles, development, and things people will not say in a group.
• "Make big changes fast to establish authority." Learn the system first; much of what looks irrational is scar tissue. Fix small irritations early and save structural changes.
• "Most businesses fail in the first year." Roughly one in five new establishments closes in year one, about half reach five years, and around a third reach ten.
• "A profitable small business cannot run out of money." Profit and cash are different. Growth consumes cash before customers pay, which is why the rolling cash forecast exists.

Recap

• Management means being evaluated on your team's output, which makes the transition disorienting and the discomfort normal.
• Small-business management is limited by attention and cash, with a coordination crisis around ten to fifteen people.
• Delegation pays back after roughly two repetitions and requires an explicit authority level and whole deliverables.
• Regular one-on-ones with the employee's agenda first are the highest-return habit available to a new manager.
• Prepared, specific scripts using situation, behavior, and impact make hard conversations possible under discomfort.
• Management occupations are the highest-paid major group in U.S. data, with real costs attached in hours and responsibility.
• The durable skill from this course is evaluating management claims by evidence and by the cost of being wrong.

Sources

1. U.S. Bureau of Labor Statistics. (2025). Management occupations. In Occupational Outlook Handbook. bls.gov
2. U.S. Bureau of Labor Statistics. (2025). Business employment dynamics: Entrepreneurship and business survival. bls.gov
3. U.S. Small Business Administration. (2025). Manage your business: guides and resources. sba.gov
4. Society for Human Resource Management. (2025). Resources for new people managers. shrm.org
5. Bright, D. S., & Cortes, A. H. (2019). Principles of Management, entrepreneurship and career chapters. OpenStax, Rice University. openstax.org