The Practice of Management

1. 1Define each manager's job by the task the enterprise's objectives demand, not by what the boss wants done — make it a real job with visible, measurable contribution to enterprise success.
2. 2Grant the broadest scope and authority: state that everything not expressly excluded falls within the manager's authority, so the job is directed by objectives of performance rather than by the superior.
3. 3Install management by objectives so each manager sets goals derived from the objective needs of the task, converting those objective needs into personal goals the manager owns.
4. 4Give each manager measurements of his own performance so he controls his own work — supply the figures fast enough that he can correct his own course before the superior intervenes.
5. 5Replace 'communications down' (talking, telling) with 'communications up': require the manager's letter so lower managers state what they think their objectives are and where misdirection exists.
6. 6Read those letters to surface misdirection, forced compromises, and unclear priorities, then negotiate objectives openly rather than dictate them.
7. 7Appraise each subordinate yourself on proven performance — his ability to set goals and attain them — not on potential via a psychological specialist; use the appraisal to teach and to place the right person.
8. 8Balance present and future in every objective set: refuse goals that buy immediate results by paying out capital or endangering long-range survival.

Henry Ford tried to run the Ford Motor Company without managers, treating management as something he could do without and reserving direction to himself. The result was the near-collapse of the company — because managers are the basic, scarcest, most expensive, and most perishable resource of a business, and a management team that takes years to build was never built. Rebuilding Ford required managing managers: establishing order, structure, motivation, and leadership through actual managerial jobs rather than delegation from one man. The lesson the passages draw is that 'management not by delegation' fails; a manager must manage because the objective needs of the task demand it, acting as a free man who decides he has to act. Ford's near-collapse is the tell for what happens when responsibility is withheld from managers and motivation cannot form because no one owns a real job.

• If a job cannot be tied to a visible contribution to enterprise objectives, then it is not a real manager's job — redefine or eliminate it.
• If a manager acts because he is told or talked into it rather than because the task's objective needs demand it, then you have domination, not self-control — shift to objectives.
• If you cannot give a manager measurements of his own performance, then you cannot give him self-control or stronger motivation — build the measurement first.
• If immediate economic results require paying out capital or risking long-range survival, then reject them as irresponsible regardless of this year's numbers.
• If appraisal focuses on a man's potential and is done by a specialist, then it is poor management — make it the manager's own job and focus it on proven performance.

• Treating 'control' as domination of one person by another rather than the ability to direct oneself; the tell is objectives being used to police subordinates instead of letting them measure their own work.
• Trying to build mutual understanding through 'communications down' and more talking; the tell is management keeps explaining while lower managers stay misdirected and unheard.
• Doing without real managerial jobs or relying on delegation from the top; the tell is Ford-style near-collapse where a team was never built and depreciates fast under misrule.
• Setting a manager's job by the narrowest scope and by the boss's wishes; the tell is managers wait for permission and act only when told, not because the task demands it.
• Buying immediate profits at the cost of long-range profitability; the tell is great economic results while running the company but a burned-out, sinking hulk left behind — fictitious results paid out of capital.
• Appraising for potential via a psychologist instead of proven performance; the tell is the manager goes by hunch and cannot teach or place people well.

• ▸Manager and job title: name the position defined by the enterprise task it exists to perform
• ▸Objectives of my unit as I see them: the manager states, in his own words, what his goals are
• ▸Contribution to enterprise objectives: the visible, measurable result this job must produce
• ▸Scope and authority: everything not expressly excluded is within this manager's authority — list exclusions only
• ▸Self-control measurements: the figures the manager will use to judge his own performance, and how fast he gets them
• ▸Misdirection and forced compromises: where objectives conflict, priorities are unclear, or behavior must change
• ▸Present-future balance: how these goals protect long-range profitability and survival, not just this year
• ▸Proven-performance appraisal note: how the manager's own boss will assess goal-setting and attainment, not potential

1. 1State the objective before you state the control: define what results the manager's unit must produce, in measurable terms, and derive that from the goals of the business as a whole. Without a clear objective, self-control has no yardstick.
2. 2Set the standard the manager is accountable for reaching, and make it visible to the person doing the work — not held privately above them. The standard is the target the manager aims at, not a whip held by a superior.
3. 3Route performance information directly to the manager who acts on it, early enough to correct course. Information is the instrument the manager uses to guide their own work, so the person needing it must get it first, not their boss.
4. 4Have each manager write down what they think their unit's objectives are and what they believe they should be accountable for, then reconcile against what higher management expects. Surface the gaps in this comparison before assigning measures.
5. 5Build the measurement so the manager can read their own results against the standard without waiting to be told — the process must feed back to the point of action, the way the special-handling clerk pulls the exceptions.
6. 6Push responsibility for shaping results downward: managing is active, not passive adaptation, so let the manager plan, initiate, and carry through the changes needed to hit the objective rather than merely react to instructions.
7. 7Balance present and future in every objective you set. A result that buys immediate performance by burning out the unit's long-range capacity is a failure of self-control, not a success — the equivalent of paying out capital.

Consider the Automation control described in the life-insurance claims office. Instead of a supervisor inspecting every claim from above, what is significant is pre-established: policies that require special handling — documents missing, data absent, beneficiary unclear, title clouded — are simply set aside and handed to a separate clerk for individual handling. The standard for a 'normal' claim is defined in advance, so the process itself becomes a self-activating governor. The clerk handling routine claims can see immediately whether a case fits the pattern or is an exception; no one has to stand over them checking. This is self-control by information as a tool: the significant deviation announces itself and routes to the point of action. The mechanics are extremely simple, yet control is maintained because the standard, not a boss, does the measuring.

• If a manager cannot state, unprompted, what results their unit is accountable for, then you have no basis for self-control — fix the objective before any measurement.
• If performance information reaches the superior before it reaches the manager who acts on it, then it is a tool of control-from-above, not self-guidance — reroute it to the point of action.
• If what is significant has not been pre-established as a standard, then deviations cannot govern the process themselves — define the significant threshold first, as with the flagged exception claims.
• If an objective raises present results by endangering long-range capacity, then reject it — unbalanced present-and-future results are fictitious, achieved by paying out capital.
• If the manager is only reacting to instructions rather than initiating changes toward the objective, then self-control has collapsed into passive adaptation — restore responsibility for shaping results.

• Using information to police rather than to guide: the tell is that reports flow upward for judgment before the doer ever sees them, so the manager learns their standing from a superior instead of from their own instrument.
• Setting controls without objectives: the tell is elaborate measurement of activity nobody can tie to a result the business needs.
• Confusing self-control with being left alone: the tell is a manager who adapts rationally to events but never plans or initiates — the 'trader' who shifts resources but shapes nothing.
• Optimizing the present at the expense of the future: the tell is strong immediate economic results that leave behind a burned-out, rapidly sinking hulk.
• Failing to pre-establish what is significant: the tell is that every case is treated as needing individual handling, so no self-activating governor emerges and someone must inspect everything.

• ▸Unit objective: the measurable result this unit must produce
• ▸Link to business goals: how this objective derives from the goals of the whole
• ▸Standard I am accountable for: the specific target I aim at
• ▸Significant deviation threshold: what pre-established condition flags an exception
• ▸Information I need: which measures, and by when, to guide my own action
• ▸Delivery point: confirm this information reaches me before any superior
• ▸Present/future balance check: what long-range capacity this objective must not deplete
• ▸Action I will initiate: the change I plan and carry through to hit the target
• ▸Self-read date: when I will measure my own results against the standard

1. 1Write each manager's objectives in three parts: what his own unit must produce, what his unit contributes to other units' objectives, and what he can expect from other units toward his own. This puts teamwork at the center from the start.
2. 2Derive every objective from the goals of the business enterprise, not from functional pride. Force each manager to state his contribution to every area that affects survival — and where a unit has no direct contribution to an area, say so explicitly.
3. 3Identify which parts of the work are too big, complex, or strenuous for an individual. Organize those as a community working as a team, not as a series of individuals linked mechanically.
4. 4Give each team member a clearly assigned, clearly defined role — teamwork requires more internal organization and definiteness of assignment than individual jobs, not less.
5. 5For new-style mass production or process production, replace the functional chain of command with a task-force pattern: many small centers of information and decision far down the line, cutting across functional lines, with design, production, and marketing working together from the start.
6. 6Control managers by the objective requirements of the job and measure them by results, not by span of control. This removes the levels-upon-levels that impede cooperation and lets span of managerial responsibility (how many you can assist, teach, and help) govern structure.
7. 7Build a right of nullification between integrated units — let each buy or sell outside when quality or price is better. Its existence, even if rarely exercised, keeps both parties efficient and responsible.

In a WWII aircraft engine plant, time pressure meant engineers could not detail every man's job. Each group of four to six men setting cylinder heads and pistons into cylinder blocks worked out its own organization, its own pace and rhythm, and its own group structure. The groups began competing to do the job faster and with fewer rejects. The production standards these self-organized teams set for themselves were consistently higher than what the industrial engineers had thought appropriate. The parallel Soviet workroom case makes the same point: workers rotated who was the daily 'shock brigade' while the others held norm and lent spare time, producing the highest income for workers, the maximum output, and the lowest per-unit cost for the company — all because the work organization served the group's social cohesion instead of cutting across it.

• If a job is too big, complex, or strenuous for one person, then organize it as a team of individuals, never as individuals linked mechanically.
• If the production system contains any element of Automation (automatic materials handling or feed-back controls), then functional centralization is impossible — decentralize into low-level task-force centers of information and decision.
• If a manager is controlled by objectives and measured by results, then abandon span of control and let span of managerial responsibility set how many report to him.
• If integrated units keep producing 'surface harmony' but poor cost and quality, then install a right of nullification so each can go outside — as with the oil company whose tanker division and refineries may deal externally.
• If a top job exceeds one person's scope, skills, and temperament, then organize an explicit executive team with each member holding a defined role, rather than pretending a one-man chief executive exists.

• Treating a team as 'chaos made into a virtue' — no defined roles. The tell: no one can state their own assignment or how it links to others'.
• Setting functional goals as ends in themselves. The tell: a unit reports 'we sold eighteen new personnel programs to line managers' instead of 'we raised employee productivity 5 percent.'
• Building levels upon levels of supervision. The tell: the head of a huge unit (like Chevrolet's general manager) has no direct access to the man who makes final decisions; layers intervene until the job is no longer 'manageable.'
• Keeping functional centralization under new-style mass or process production. The tell: design starts in engineering, then plant tools up, then sales pushes — sequential handoffs instead of a team working from the start.
• Prizing integration over performance and killing the right of nullification. The tell: rate disputes between owned units get bumped up to top management for decision, and 'it all goes into the same corporate pocket' is used to excuse high cost.

• ▸Manager/Unit: name the unit and its head
• ▸Own unit performance: the results this unit alone must produce
• ▸Contribution to others: what this unit gives each named unit toward their objectives
• ▸Expected from others: what this unit needs from named units to hit its own goals
• ▸Link to company goals: which business-survival area(s) this unit contributes to
• ▸No-contribution areas: areas this unit is explicitly not expected to affect
• ▸Team tasks: work here too big/complex for individuals — assign as a defined-role team
• ▸Nullification right: where this unit may buy/sell outside and on what condition
• ▸Measure of results: the business outcome, not the professional-activity count, used to judge performance

1. 1Define productivity as the balance of ALL wealth-producing resources, not output-per-labor-hour alone. Inventory the factors the passages name: time, product mix, process mix, and organization structure — treat each as a lever, not just headcount or wages.
2. 2Test the mix of processes against the type of production you actually run. Classify your operation as unique-product, mass production 'old style,' mass production 'new style,' or process production, because each demands a different structure to yield productivity.
3. 3Reject the assumption that management 'only adapts to the forces of the market.' Decide what innovations in resource use will produce the greatest output for the smallest effort — as Rosenwald did inventing the mail-order plant and General Wood did with merchandise planning at Sears.
4. 4Push decision and integration centers to the level the production type requires. Under unique-product work, coordinate at the top; under mass-production new style and process production, staff cross-functional teams that tackle design, production, marketing and work organization simultaneously.
5. 5Redefine 'productive labor' before measuring it. Ask what actually adds economic value versus what merely adds cost or motion, then reduce the effort factor rather than only chasing the output factor.
6. 6Build measurement tools even where none yet exist. The passages admit that for productivity 'we hardly know more than what ought to be done' — so run concentrated, imaginative work to define what to measure, exactly as firms did to achieve today's clarity in market standing.
7. 7Guard against the decline pattern: aging capital equipment and productivity that improves fast only in new industries while stagnating in old ones. Continuously improve management performance rather than defending what you already have.

Take Sears as the passages present it. Sears did not respond to 'forces'; people made it a business. Rosenwald's innovations invented the mail-order plant, organizing physical resources so output rose against effort. A generation later General Wood changed its basic nature through merchandise planning and manager development, sustaining growth through the depression and World War II. Each shift was a productivity decision about the mix of resources — process, structure, and the development of managers — not merely a sales push. The passage warns Sears may again have to think through what its business is, where its markets are, and what innovations are needed, because productivity is never settled. The lesson: superior productivity comes from managers who create the conditions rather than adapt to them.

• If you are judging performance by rising sales volume alone, then stop — volume 'tells little about performance' and a firm's sales can climb while it heads for collapse (see the oil-refinery-equipment maker whose spurt came only from replacement parts for dying equipment).
• If your production type is mass production 'new style' or process production, then functional centralization is impossible — build cross-functional teams and push decision centers down.
• If your production type is unique-product, then centralize coordination at the top and bring functions together only where company policy is set.
• If capital equipment is aging and productivity gains appear only in your newest lines, then treat this as the early Britain-of-1880 decline signal and act on process and structure, not just cost.
• If you can only state intentions rather than goals and measurements for productivity, then treat that as a signal to build measurement tools, not as license to skip the target.

• Equating productivity with labor cost or output-per-worker. The tell: your improvement plans touch only wages and headcount, never time, product mix, process mix, or organization structure.
• Believing management merely 'adapts the business to the forces of the market.' The tell: you explain outcomes by external forces and have no Rosenwald-style innovation decisions of your own.
• Reading rising sales as proof of health. The tell: nobody measures share against market potential or against competitors, so a dying field masquerades as success.
• Forcing top-level functional centralization onto mass-production-new-style or process operations. The tell: design, production and marketing collide late instead of being tackled simultaneously by one team.
• Waiting for perfect metrics before acting on productivity. The tell: you keep 'progress reports' and statements of intention because 'the subject is brand new,' instead of doing the hard, concentrated work that produced clarity in market standing.

• ▸Production type: classify as unique-product / mass 'old' / mass 'new' / process
• ▸Time: where does elapsed time inflate effort without adding output?
• ▸Product mix: which products consume disproportionate resources for their return?
• ▸Process mix: which process choices raise or lower output-per-effort?
• ▸Organization structure: at what level do decision and integration centers sit vs. where they should?
• ▸Productive labor test: what work here actually creates economic value?
• ▸Innovation decision: what Rosenwald/Wood-style change would raise output for less effort?
• ▸Measurement gap: what productivity factor is stated as intention with no metric — and what tool will you build?
• ▸Decline signal check: is capital aging and are gains confined to new lines only?

1. 1Start by analyzing your customer and market—identify who represents a separate, distinct, or under-served segment, as Sears saw the isolated American farmer with low individual but large aggregate buying power.
2. 2Define what that customer considers 'value,' not what you think you produce; value is decisive because the customer alone converts resources into wealth by paying.
3. 3Locate every activity where you can provide a better or more economic good, service, or process—innovation spans product development, distribution, pricing, supplier relations, organization, and even store location, not just the product.
4. 4Redesign products for the actual market's purchasing power (Sears redesigned 'class market' refrigerators for a 'mass market' with limited money) rather than porting existing offerings unchanged.
5. 5Build the enabling infrastructure the innovation requires: new distribution channels, supply sources, plants, and organizational structures without which the business is 'physically impossible' to run.
6. 6Innovate in the human organization too—when you change the business, develop the managers the new mode demands (Sears' retail push stalled for 10–15 years on a manager shortage until it invented manager development).
7. 7Set objectives at the desirable first, then decide what concessions to the possible are required—manage as a creative task that creates or changes economic conditions rather than adapting to them.
8. 8Periodically re-ask 'what is our business and what should it be'—as Sears did when the young married working mother became as distinct a market as the isolated farmer once was.

Sears illustrates innovation as a business-wide function in two phases. In its first phase around 1900, reaching the isolated farmer required innovation in five distinct areas: systematic merchandising to find sources of supply at prices the farmer could pay; a factual, reliable mail-order catalogue that became the farmer's 'wish book'; the shift from 'caveat emptor' to 'caveat vendor' via 'your money back and no questions asked'; the mail-order plant to fill large orders cheaply and quickly; and a human organization to run it all. In its second phase under General Wood, after the farmer's isolation ended and cities became one homogeneous market, Sears switched emphasis to retail stores—forcing a new series of innovations: designing products, developing and financing hundreds of small suppliers (largely T. V. Houser's merchandise planning), building manager development, decentralizing organization structure across 700 stores, new compensation policies, and even reinventing store location, architecture, and physical arrangement. Each innovation was 'as distinct a contribution to the American economy' as the catalogue had been. The common thread: innovation was never confined to product—it reached every activity because the purpose was to create a customer.

• If a customer segment is isolated from existing channels or has distinct needs and untapped aggregate buying power, then treat it as a separate market and innovate a distribution channel to reach it.
• If your existing product targets a 'class market' but the opportunity is a 'mass market' with limited purchasing power, then redesign the product and build the supply base to produce it cheaply at volume.
• If a strategic shift creates a bottleneck in people (e.g., store managers), then make manager development an explicit innovation rather than assuming existing staff can adapt.
• If an operation cannot be run from headquarters because it is local in character, then decentralize structure and create measures of performance plus compensation to reward local autonomy.
• If the business is currently successful, then re-ask 'what is our business and what should it be'—that is when the question matters most, because markets like Sears' shift underneath you.

• Treating innovation as only product invention—the tell is that marketing, pricing, supplier policy, organization, and location are excluded from the innovation agenda, as they were not at Sears.
• Believing 'management only adapts the business to the forces of the market'—the tell is passive, adaptive decisions aimed at accommodation to the possible rather than creating conditions.
• Defining the business by what you think you produce instead of what the customer considers value—the tell is products designed around internal capability rather than customer purchasing power.
• Executing a strategic shift without building enabling infrastructure or people—the tell is a multi-year bottleneck like Sears' 10–15 year manager shortage that stalls the whole move.
• Making executives security-conscious and over-emphasizing promotion—the tell is managers who look at work from a viewpoint of security and won't pioneer, innovate, or reach for the new.

• ▸Target market segment: name the separate/distinct/under-served customer group
• ▸Defined 'value': what this customer considers value and will pay for
• ▸Product innovation: how the good/service must be redesigned for this market's purchasing power
• ▸Channel/distribution innovation: the new way to reach and supply the customer
• ▸Supply innovation: sources or suppliers that must be found, built, or financed
• ▸Organization innovation: structure, decentralization, and performance measures required
• ▸People innovation: manager or staff capabilities that must be developed
• ▸Policy innovation: pricing, guarantees, or supplier-relations rules to change
• ▸Re-examination trigger: the market shift that will force asking 'what should our business be?'

1. 1Put economic performance first in every decision. Management is the specifically economic organ of an industrial society, so treat the economic dimension as the first dimension of every act, deliberation, and choice — not one factor among many.
2. 2Define your business by starting outside it: identify who the customer is and what they buy. The enterprise exists to produce economic results outside itself, in the market and for customers, so 'creating a customer' is the primary objective, not internal organization.
3. 3Set objectives in every key result area. Do not measure performance by one dimension alone; name each area where results are required, write objectives for it, and use those objectives to direct, steer, and measure the business.
4. 4Apply the profitability test as the test of validity. Treat profit not as a motive but as the verification that your marketing and innovation actually created something the customer will pay for at a price that maintains and improves wealth-producing resources.
5. 5Manage present and future simultaneously. Keep the enterprise profitable now (or there is no future) while investing in research, plant, and human organization whose pay-off may run ten to twelve years — knowing action on present results directly determines future results.
6. 6Judge by achievement, not knowledge. Achievement remains both proof and aim; systematic analysis and tools are secondary to actual business performance, so audit whether the tools you use are enabling the performer to do the work.

Consider a new plant investment. A half century ago a new plant was expected to pay for itself in two or three years; today, with capital investment per worker ten times that of 1900, the pay-off period in the same industry runs ten or twelve years. So when you approve the plant you commit present, visible resources to a result that will not validate itself for a decade — an Edison-scale innovation might take his successors fifteen years to complete. You cannot justify it by knowledge or projections alone; only actual business performance in the market will prove it valid. Meanwhile you must keep the existing business profitable in the present, because otherwise there is no enterprise left to enjoy the future. This is the time-squeeze unique to management: your research and plant decisions profoundly influence visible present results even as they aim at future ones.

• If a decision is being weighed, then apply the economic dimension first — every act, decision, and deliberation carries an economic dimension as its first test.
• If you want to know whether marketing and innovation worked, then check whether you created a customer and whether profitability validates it — profit is the test of validity, not the motive.
• If a present-results action (like cutting research or plant investment) also degrades future results, then treat it as destroying capital, not saving cost.
• If someone proposes measuring performance by a single dimension, then insist on objectives across every key result area instead.
• If management skill is offered as transferable to government, church, or army, then discount it — only the analytical and administrative skills transfer, and those are secondary to each institution's primary objectives.

• Seeing only the internal dimension — organization, policies, human relations, authority within — while ignoring that the business exists to produce results outside itself. The tell: all your metrics point inward.
• Treating profit as the motive rather than the test of validity. The tell: you optimize profit directly instead of asking whether you created and kept a customer.
• Trying to make management an exact science or licensed profession, eliminating risk, ups and downs, and 'irrational' consumer choices. The tell: you demand predictability and drift toward cartelization (as Gantt and the German Rationalization movement did).
• Living in only one time dimension — sacrificing the present for the future or vice versa. The tell: you either starve current profitability or defer all investment, unlike the manager who must live in both at once.
• Measuring performance in one key result area alone. The tell: strong numbers in one dimension mask neglect elsewhere because no objectives were set for the other areas.

• ▸Customer created: name who the customer is and what they buy
• ▸Market standing: your position/objective in the market for that customer
• ▸Key result areas: list each area where results are required
• ▸Objective per area: one measurable objective for each area listed
• ▸Profitability test: does profit validate the marketing/innovation? yes/no + evidence
• ▸Present-results actions: current decisions and their visible economic effect
• ▸Future-results investments: research/plant/people commitments and pay-off period
• ▸Capital check: does any present-results action destroy future capacity? note it
• ▸Achievement verdict: is this proven by actual performance, not just knowledge?

1. 1Set profitability as a minimum, not a maximum. Compute the profit needed to cover the three costs of staying in business—replacement, obsolescence, and market risk and uncertainty—plus the capital you must attract for innovation and expansion.
2. 2Derive the profitability objective from the capital-market rate for your desired financing. Ask: what return, on money already in the business, produces the additional capital we need at a cost we are willing to pay? Express it as a rate (the passages cite '25 per cent return on capital before taxes').
3. 3Add the time factor to every profitability number. State profitability as anticipated total profit over the number of years the profit can be expected, not as a snapshot rate, because a rate without a time-span is meaningless.
4. 4Set multiple objectives across every key area where performance directly and vitally affects survival and prosperity. Do not hunt for one 'magic formula'; multiple objectives narrow judgment's range, focus it, and ground it in facts and reliable measurements.
5. 5Balance present and long-range future in each decision. Reject any move that buys immediate profit by endangering long-range profitability or survival, and treat 'economic results' achieved by paying out capital as fictitious.
6. 6Define 'what is our business' from the customer's satisfaction outward, not from the product inward—AT&T's 'our business is service' governed research policy, employee indoctrination, and financing.
7. 7Ask what unsatisfied wants the customer still has. The ability to ask and answer this correctly separates a growth company from one that merely rises (and later falls) with the tide of its economy.
8. 8Fund innovation as a survival factor. Treat 'advancement of the arts'—applying increased knowledge to do better—as a task of the enterprise and a major factor in survival and prosperity.

The manufacturer of fuse boxes and switch boxes faced the survival question in 1943: what to do after World War II. He asked what unsatisfied want the customer had and answered correctly—homeowners needed a switch-and-fuse panel carrying much higher electricity loads and more circuits than existing equipment, which had been designed before household appliances became general. But the answer was constrained by prosperity economics: the new panel had to carry almost twice the load while costing, fully installed, much less than two old panels and not much more than one. That way a homeowner needing more circuits would find it easier and not much more expensive to have an electrician rip out the old panel and install the high-load one. This joins innovation (new capability) to profitability (a price the market bears) to growth (capturing the replacement demand)—the three legs of prosperity in one product decision.

• If a decision buys immediate profit by risking long-range profitability or survival, then reject it—the immediate results are fictitious and paid out of capital.
• If a profitability figure is stated without a time-span, then treat it as meaningless; convert it to anticipated total profit over the years the profit can be expected.
• If you are searching for a single objective, then stop—you are seeking a magic formula to avoid judgment; set objectives in every key area affecting survival and prosperity instead.
• If your development depends on 'the rising tide' of your economy or industry, then assume you will fall with that tide, and go find the customer's unsatisfied wants.
• If earned profit does not cover replacement, obsolescence, and market risk plus the capital-market rate, then you are under-earning the minimum needed to stay in business.

• Chasing maximum profit instead of the required minimum. Tell: the objective is framed as 'as much as possible' rather than a rate derived from the cost of capital you must attract.
• Producing great results while running the company but leaving 'a burned-out and rapidly sinking hulk.' Tell: capital equipment is aging and reinvestment is deferred while reported earnings look strong.
• Under-earning without knowing it. Tell: the business treats profit as a residual surplus rather than as a cost of staying in business, so no one asks whether it covers replacement and obsolescence.
• Defending what you have rather than advancing—the Great Britain of 1880 pattern. Tell: productivity improves only in the newest lines and stagnates elsewhere; smugness replaces vision and effort.
• Answering 'what is our business' from the product, not the customer. Tell: policy is defended by attacking critics rather than built on creating customer satisfaction, as AT&T learned it could not be.

• ▸Cost of replacement: annual capital needed to renew wearing/aging assets
• ▸Cost of obsolescence: reserve for equipment made obsolete by new methods
• ▸Cost of market risk and uncertainty: buffer for demand and competitive shocks
• ▸Capital-market rate: the return required for your chosen type of financing
• ▸Minimum profit target: sum of above, expressed as a % return on capital
• ▸Time-span: number of years over which this profit can be expected
• ▸Innovation/expansion capital: additional funds needed for tomorrow's business
• ▸Unsatisfied customer want: the specific need not met by today's offering
• ▸Present/future balance check: does this target endanger long-range survival? (yes/no)

1. 1Start by defining what your business is and what it should be, then derive the overall goals of the enterprise from that answer — objectives for any manager's job flow down from these business goals, never up from personal ambition.
2. 2Set objectives in the key result areas, not everywhere; identify the few areas where performance and results directly determine the survival and prosperity of the business, and state each manager's targets there.
3. 3Have each manager draft the objectives of his own unit before the superior dictates them; the objectives are then fixed through mutual understanding between superior and subordinate, so the manager commits to targets he helped set.
4. 4Anchor every objective to the overall goals: a manager's targets must specify what contribution his unit owes the larger unit it belongs to, so that reaching them advances the business as a whole rather than one function's private drive.
5. 5Attach measurements to each objective at the point of setting it — decide in advance what evidence will show performance, so the objective becomes the standard against which you later direct and measure.
6. 6Route reports and procedures back to the manager doing the work, not upward as control tools; self-control means the person accountable gets the measurements first and steers his own performance.
7. 7Review objectives against results on a defined cycle, and revise when the business goals shift or the key result areas change — objectives are the steering instrument, not a one-time contract.

Take the profitability objective, one of Drucker's key result areas. Instead of chasing the economist's maximum profit, a company sets the minimum profit it must produce to survive and obtain new capital: "We shoot for a return on capital of 25 per cent before taxes." Decoded, this states that a 25 per cent pre-tax return is the minimum needed to get the kind of capital wanted, in the amounts needed, at the cost the firm is willing to pay — a rational, derived objective rather than a slogan. Following Joel Dean, it is refined by adding the time factor: state anticipated total profits over the life of the investment, discounted to present cash value, instead of an annual rate. This frees the target from "the unnecessary tyranny of the accounting year," which wrongly assumes the calendar year has economic meaning. Each manager whose unit affects that return then carries a derived objective specifying his contribution to it, measured on the same discounted basis.

• If an objective cannot be traced back to the overall goals of the business, then it is a personal drive, not a management objective — discard or redirect it.
• If the superior alone sets the target and the subordinate merely receives it, then you have misdirection by the boss, not mutual understanding — re-set it jointly.
• If an area does not directly determine survival and prosperity, then it is not a key result area — do not burden it with formal objectives.
• If a report flows upward as a control before it reaches the manager doing the work, then convert it to self-control by delivering the measurement to him first.
• If a profitability target names a maximum rather than a minimum, then reframe it as the minimum needed to obtain capital at the cost you will pay.

• Treating management as "the boss" and objectives as orders handed down — the tell is that subordinates cannot explain how their targets connect to the business's goals.
• Managing by drives: a burst of attention to one area (inventory, then quality, then costs) while others slide — the tell is that this month's crisis, not the key result areas, sets the agenda.
• Misdirection by the boss, where the superior's stated priorities pull managers away from the business's actual goals — the tell is aligned individuals producing a misaligned whole.
• Measuring maximum instead of minimum profit — the tell is a target justified by "how much we can make" rather than "the least we need to obtain capital."
• Letting the accounting year dictate objectives, ignoring the time factor — the tell is annual rate-of-return targets with no discounting over the investment's life.

• ▸Business goal this job serves: state the overall business goal the objectives derive from.
• ▸Key result area: name the area where this unit's performance affects survival and prosperity.
• ▸Contribution owed: state what this unit must deliver to the larger unit it belongs to.
• ▸Objective (target): write the specific result to be achieved, minimum not maximum.
• ▸Measurement: define the evidence that will show whether the objective is met.
• ▸Time dimension: state the period and, for financial targets, the discounted present value.
• ▸Set jointly by: name superior and subordinate who reached mutual understanding on this.
• ▸Self-control report: name the measurement that goes to the responsible manager first.
• ▸Review date: set when objective and result are compared and revised.

1. 1Decide the unit type before drawing the boxes: split into 'federal' product businesses when each unit can own a distinct market and carry its own profit-and-loss (Ford's fifteen autonomous divisions, Sears's seven hundred stores), or into 'functional' units when each owns one distinct stage of a single process.
2. 2Give each federal division a complete management with full authority to make all decisions needed to attain its results — not delegated tasks, but genuine authority matched to accountability.
3. 3Break out separate cost and result figures for every unit. The old Ford failure was one set of figures for the whole; decentralization requires the steel-mill superintendent to know the price of his own coal.
4. 4Set each unit's objectives by the contribution it must make to the parent — the division general manager's goals come from what the division owes the company, not from what he personally wants.
5. 5Require each manager to develop and set his own unit's objectives, then have higher management approve or disapprove — reserve the veto, but make objective-setting the manager's first responsibility.
6. 6Build measurement of performance for each unit head (store managers, division managers) against those objectives, so autonomy is paired with accountability rather than arbitrary orders.
7. 7Attack the manager-development bottleneck in parallel: decentralization multiplies the number of managers who must decide locally, so build a systematic pipeline to create them — Sears's greatest bottleneck for its first ten-to-fifteen retail years was manager shortage.
8. 8Redesign compensation so unit heads are rewarded for their measured performance and results, not for compliance.

Ford in 1946 was rigidly centralized: all power sat with old Henry Ford, and there was only one set of figures for the whole complex operation. The company owned a 1.5-million-ton steel mill, yet its cost figures disappeared into the overall company totals — the mill superintendent did not even know what price Ford paid for his coal, and purchase contracts were 'top secret.' Henry Ford II decentralized into fifteen autonomous divisions, each with its own complete management fully responsible for performance and results and with full authority to make the decisions needed to reach them. The steel mill became one such division, alongside Ford and Mercury-Lincoln, parts divisions, and an international/export division. He took the concepts intact from General Motors rather than developing them piecemeal. Within ten years a moribund, leaderless company had joined battle with Chevrolet for first place and become a major growth company.

• If a unit can own a distinct market with its own profit-and-loss, then structure it as a federal product business with complete management; if it only owns one stage of a shared process, then structure it as a functional unit.
• If the work is old-style mass production, then functional centralization can still hold (at high cost in efficiency); if it is new-style mass production or process production, then decision centers must move far down and cross-functional teams must decide together.
• If a unit's costs vanish into the company total, then decentralization has failed — insist on separate figures before granting authority.
• If you would give a manager objectives without letting him develop them himself, then stop: define his goals by his contribution upward and require his positive act of assent.
• If the operation must be run locally and cannot be directed from a distant headquarters (retail stores two thousand miles from the mail-order center), then decentralize authority to the local manager.

• Granting responsibility without authority or separate figures — the tell is a unit head who cannot name his own costs, like the Ford superintendent ignorant of his coal price.
• Running a locally-managed operation from central headquarters — the tell is decisions bottlenecking upstream because retail stores 'cannot be run from headquarters two thousand miles away.'
• Handing managers objectives from above instead of having them set their own — the tell is arbitrary orders and managers who don't know what they'll be measured against or why.
• Decentralizing structure while neglecting manager development — the tell is a manager shortage that becomes the operation's greatest bottleneck, as it was at Sears for its first ten-to-fifteen retail years.
• Substituting a 'sense of participation' for genuine responsibility — the tell is human-relations language masking managers who never actually commit to the objectives with a positive act of assent.

• ▸Unit name & type: mark federal (own P&L) or functional (single process stage)
• ▸Market or process stage owned: state the one distinct thing this unit is responsible for
• ▸Authority granted: list decisions the unit head may make without approval
• ▸Objectives: define by the contribution this unit owes the parent above it
• ▸Own figures: name the separate cost/result numbers this unit will track
• ▸Measurement: state what performance is measured against and how
• ▸Manager development plan: how the unit head and successors are being built
• ▸Compensation link: how rewards tie to measured performance and results

1. 1Start by mapping the whole process before you divide the work: establish what the planned result is and what control mechanism maintains it, exactly as Automation requires 'that what is significant be pre-established.' Only after the concept is thought through do you assign tasks to people or machines.
2. 2Separate the routine flow from the exception stream. Route the ~98% that is standard through the smooth continuous process, and hand off the handful of cases needing judgment to a person equipped to plan and decide — do not force judgment work into the routine channel.
3. 3Design each job to contain more than a single repetitive motion: build in challenge, skill, judgment, and a planning element. Refuse the assumption that work is made productive by stripping it down; the trend since mass production runs the other way, toward higher skill.
4. 4Marry planning to doing inside the job. Give the person full responsibility for a result and the authority to plan how to reach it, the way Rosenwald 'gave to management people the maximum of authority and full responsibility for results.'
5. 5Extend management-type work downward. Treat the semi-skilled operator and technician as people who must 'understand what' is happening, not just execute; enlarge the management area so many now called rank-and-file do genuine planning work.
6. 6Build the human resource deliberately as a bottleneck to solve: forecast the skilled and trained people the design will demand and develop them systematically before scaling, since 'the major obstacle... will almost certainly be the lack... of enough trained men.'
7. 7Feed results back into the design. Use feedback so the output of the work adjusts an earlier stage, and revise the job design as the market and the whole enterprise (business, human organization, social institution) shift.

Take the life-insurance claims office in the passages. Rather than build one uniform job that grinds every policy through the same steps, the design first fixes the significant variable: whether a policy is complete or defective. Policies missing documents, missing data, unclear beneficiaries, or clouded titles are simply set aside and routed to a separate clerk for 'special, individual handling.' That exception job is where planning marries doing — the clerk exercises judgment on genuinely variable cases. Meanwhile the routine channel processes the roughly 98% smoothly and continuously, even across thousands of payment and beneficiary variations, because 'simple rejection is adequate control to maintain the process.' The design decision came before any machine or gadget: it defined the control concept, then allocated judgment to people and flow to the routine stream.

• If a case requires judgment about incomplete or ambiguous inputs, then route it to a person with planning authority — never bury judgment inside a repetitive station.
• If you can pre-establish what is significant and pre-set a self-activating governor, then let routine flow automatically and reserve human effort for the exceptions.
• If a job contains only doing and no planning, then it is under-designed — add challenge, skill, and a decision element to utilize the human resource's unique qualities.
• If scaling a new design, then treat the supply of skilled and trained people as the binding constraint and develop them first, as Sears did with manager development.
• If the market or enterprise shifts (farmer to motorized, class market to mass market), then redesign the jobs and functions rather than preserving the old division of work.

• Believing new technology deskills work — designing jobs for cheap unskilled labor. The tell: your fastest-growing labor category is unskilled, contradicting the reality that skilled and trained people grow most.
• Dividing work before the process concept and control are thought through. The tell: you bought machines and assigned tasks, then discovered no built-in control maintains the planned result.
• Cramming exceptions into the routine channel. The tell: the smooth 98% flow stalls repeatedly because odd cases clog it instead of being set aside for individual handling.
• Treating 'manager' as a rank rather than as work, function, and responsibility. The tell: people hold titles but do no planning or judgment, and management work is not extended to technicians and operators who need to 'understand what.'
• Scaling the design without building the human organization. The tell: a persistent manager or skilled-worker shortage becomes the bottleneck — as it was for Sears retail for its first ten to fifteen years.

• ▸Planned result: state the specific output this job exists to produce
• ▸Significant variable (pre-established): what condition determines routine vs. exception
• ▸Control mechanism: how the process is maintained — rejection, feedback, or adjustment
• ▸Routine flow (the ~standard %): what runs continuously with minimal judgment
• ▸Exception stream: which cases route to a person for individual, judgment-based handling
• ▸Planning-in-the-job: the specific planning/decision element this doer owns
• ▸Authority + responsibility: what result the person is accountable for and can decide
• ▸Skill/training required: the trained capability this design demands before scaling
• ▸Development plan: how you build that human resource ahead of the bottleneck

1. 1Set objectives in every key area where performance and results directly and vitally affect the survival and prosperity of the business — not one master objective. Appraise against these, because emphasizing only profit misdirects managers into short-changing research, promotion, and capital replacement.
2. 2Direct each individual manager's objectives toward the goals of the business, then let self-control govern execution: management by objectives and self-control is the first requirement. Appraisal measures the manager against targets he helped set, not against a supervisor's mood.
3. 3Structure the manager's job so it allows maximum performance before you appraise the person in it. If the job denies the person the ability to make the needed efforts and produce required results, no appraisal is fair.
4. 4Build measurement into the appraisal using economically meaningful data. Reform accounting so cost allocation reflects the manager's actual contribution (e.g., Contributed Value) rather than across-the-board overhead percentages that make realistic performance analysis impossible.
5. 5Appraise strengths and results, not weaknesses in the abstract: place and promote people where their demonstrated performance builds the market of tomorrow, not merely where they hit the easiest short-run number.
6. 6Administer with justice by making the spirit of the organization recognize and reward excellence. A mean spirit produces mean managers; the appraisal system decides whether the organization rewards excellence or scuttles into placid mediocrity.
7. 7Tie promotion and compensation to the multi-objective scorecard, then use judgment. The system narrows the range of judgment and gives it a foundation in facts and reliable measurements — it does not replace judgment with a formula.

Take the electric-condenser business from the passages. Physically identical condensers serve three distinct lines: new radios, radio repair/replacement, and telephones — and repair condensers split again by Southern customers judging termite resistance versus Northwestern customers judging humidity resistance. A performance-oriented appraisal of the sales manager cannot use one blended sales figure; it must measure standing (in dollars and market-share percentage against direct and indirect competition) line by line. If you appraise only on total profit this quarter, the manager pushes the easiest-selling line and slights the line that is the market of tomorrow, undermining the future while looking good on paper. So you set the seven marketing goals per line as the objectives, let the manager set his own targets and self-control the execution, and appraise against those specific results. Promotion and pay then follow demonstrated performance against the full scorecard, and judgment fills the gaps the numbers can't reach.

• If appraisal rests on a single objective (usually profit), then reject it — multiple objectives in every key result area are required, or you direct managers into the worst practices of management.
• If a manager underperforms, then first test whether the job is structured to allow maximum performance before judging the person.
• If the objectives were set for the manager rather than by the manager under self-control, then the appraisal lacks integrity — the vision must be the individual's, directed toward business goals.
• If a proposed measure cannot allocate costs in an economically meaningful way, then don't appraise on it — across-the-board overhead percentages make realistic cost analysis impossible.
• If the organization's spirit tolerates mediocrity, then no appraisal system will hold — reward excellence explicitly or the system rewards nothing.

• Chasing the one right objective — the tell is managers short-changing research, promotion, and capital expenditure to protect this quarter's profit and dressing it up as discipline.
• Measuring blended totals instead of lines — the tell is a single sales or profit number covering products that satisfy entirely different customer wants (radio-new vs. repair vs. telephone).
• Appraising the person before fixing the job — the tell is repeated 'underperformers' in a role no one can succeed in.
• Substituting formula for judgment — the tell is a rigid magic-formula scorecard that eliminates the manager's judgment rather than narrowing and focusing it.
• Letting a mean organizational spirit run the appraisal — the tell is talented people becoming stunted and the system quietly rewarding placid mediocrity while claiming to be objective.

• ▸Manager & role — name and the specific job being appraised.
• ▸Key result areas — list every area where results vitally affect survival/prosperity.
• ▸Objectives per area — targets the manager set under self-control (dollars + % where measurable).
• ▸Line-level breakdown — split blended figures by customer want/market, not by physical product.
• ▸Demonstrated result vs. target — actuals against each objective, against direct and indirect competition.
• ▸Measurement basis — the economically meaningful metric used (e.g., Contributed Value), noting any overhead-allocation caveats.
• ▸Strengths leveraged — where this manager's demonstrated strength builds tomorrow's business.
• ▸Justice check — evidence the job was structured to allow performance and the spirit rewards excellence.
• ▸Promotion/compensation decision — action tied to results, with the judgment call recorded.

1. 1State the business's purpose in customer terms first: answer 'What is our business?' by defining who the customer is and what they value, since the enterprise exists to produce economic results outside itself. This becomes the reference point every job connects back to.
2. 2Set objectives, not just tasks. Define key result areas and translate the overall purpose into concrete goals for management people at each level — this is how you 'direct and steer a business and measure its performance.'
3. 3Convert rank-and-file jobs into managerial ones wherever possible. Look for jobs that 'are in effect managerial, or would be more productive if made so,' and give those people responsibility for and a share in decisions about their work.
4. 4Teach people to 'see and think managerially.' Require that technicians and non-managers understand what management is so they connect their function to overall performance rather than to their specialty alone.
5. 5Balance the two time dimensions explicitly in every decision. Show each person how their present work affects long-range profitability so no one purchases immediate results by damaging future capital.
6. 6Build the whole three-dimensionally: the business as economic-results institution, as a human/social organization, and as the innovating engine of tomorrow — and make each person's line of sight run to all three, not just to internal organization.
7. 7Locate decision-making close to the work. Decentralize authority and flexibility so people act on their part of the whole rather than waiting on a small centralized nervous system.

Take the salesman who sells a refrigerator to the Eskimo to keep food from freezing. A parochial view files this under 'sales quota met.' The managerial-vision view sees it as innovation: economically a new product has been created, a new want met, even though technologically it is the same old refrigerator. The salesman understands the enterprise's purpose — to provide better and more economic goods and services and to create markets — and so connects his individual act to the second function of the business, innovation. His work is no longer a functional task in isolation; it advances the whole. That is the difference between a broker collecting a fee in a static economy and a member of a growing enterprise.

• If a person can describe only their own task but not how it serves the customer and overall results, then vision is missing — restate objectives in terms of the whole.
• If a decision buys immediate economic results while endangering long-range profitability, then treat those results as fictitious (capital paid out) and reject the trade.
• If a job requires judgment about the work's contribution but is defined as 'do as directed,' then redesign it as managerial to raise productivity.
• If technicians and specialists cannot 'see and think managerially,' then training in what management is precedes any further automation or expansion.
• If decision-making concentrates at the top while the environment changes fast, then decentralize — the centralized nervous system cannot adapt, like the Saurian reptiles.

• Treating 'management' as merely 'the boss' or 'the fourteenth floor' — the tell is employees, even managers, describing the front office as an unknowable land of ogres and unicorns, with no line of sight to what it does or why.
• Defining rank-and-file workers as mere material resources under 'the laws of mechanics' — the tell is jobs stripped of responsibility that could be, and would be more productive as, managerial.
• Optimizing one internal dimension (organization, human relations, authority) while ignoring the enterprise as an economic and innovating whole — the tell is management books and plans that only ever see the internal side.
• Defending what you have rather than advancing — the tell is aging capital equipment, productivity stagnant outside the newest industries, and a smug 'stay up' posture (the Great Britain of 1880 pattern).
• Measuring by brawn substitution alone (capital per worker) while neglecting 'brain formation' — the tell is capital investment equal to a rival's yet productivity two-thirds lower, from too few managers/technicians and poor structure.

• ▸Role / level: name the position being audited
• ▸Business purpose: state 'what is our business' in customer terms
• ▸This role's key result areas: the 1-3 outcomes it owns
• ▸Connection to whole: one sentence linking daily work to overall performance
• ▸Present-vs-future balance: does this role trade long-range results for immediate ones? note where
• ▸Managerial content: is this job directed-only or does it carry decision responsibility? can it be raised?
• ▸Thinks managerially?: can the person explain how their function serves results (yes/partly/no)
• ▸Decentralization check: is authority for this work held where the work happens?
• ▸Gap + action: the single change that best restores line of sight