Your Output Is Your Team's Output

The Core Insight

Andy Grove opens a management book with a breakfast order. A three-minute soft-boiled egg, toast, and coffee, served hot and at the same time, about five to ten minutes after the customer arrives.

The charter is three demands at once: a scheduled delivery time, an acceptable quality level, and the lowest possible cost. Grove ran Intel manufacturing before he ran Intel, and the book is that kitchen applied to managerial work.

Ask managers what they produce and the answers come back as tasks: judgments made, direction given, resources allocated, mistakes caught. Grove files every one of those under activity. A manager's output is the output of his organization plus the output of the neighboring organizations under his influence.

An hour of a manager's time produces nothing by itself, and its worth appears only in what other people ship. A coach alone scores no touchdowns.

Grove wrote for middle managers, and he counts know-how contributors who shape work without commanding it. He calls them micro CEOs, and tells them to stop waiting for good practice from the top.

The Framework

The vocabulary comes off the factory floor and never leaves. Every flow has a limiting step, the longest or hardest or most expensive one, and you build the schedule backward from it. Grove calls the backward stagger an offset. The egg takes three minutes, so the toast starts about a minute before the egg is done.

The limiting step moves when the constraints move. Queue for a single toaster and the three-minute egg becomes a six-minute egg. In college recruiting the limiting step is the plant visit, so phone screening exists to raise offers per visit.

Value accumulates as work moves, so fix a problem at the lowest-value stage available. Reject the rotten egg at receiving, cut the candidate on campus, catch the fault at unit test.

Grove runs the rule against criminal justice, a production line ending in jail. Loading the whole cost onto the people who reach jail puts one conviction well over a million dollars. A cell costs some 80,000 dollars to build and 10,000 to 20,000 dollars a year to hold a person. Cell supply limits the flow now, the wrong step doing the limiting.

Five moves carry the rest of the book.

• Find the limiting step in any flow, then schedule every other step backward from it.
• Fix faults at the lowest-value stage available, because every later stage costs more.
• Pair each quantity indicator with the quality it can damage.
• Count a manager's output as the output of the group, then spend hours by what they produce.
• Match the direction you give to the maturity the person has in that specific task.

Key Ideas

You Steer Where You Look

The breakfast factory runs on five daily indicators: the sales forecast with yesterday's variance, raw material inventory, equipment condition, manpower, and a complaint log. Each aims at a specific operational goal, or it earns nothing.

The mechanism is a bicycle: you steer where you are looking. So each indicator gets a partner that measures the counter-effect. Inventory level pairs with shortages, and a software completion date pairs with the capability delivered.

Two criteria filter the rest. Measure output rather than activity, so you judge a salesman by the orders he gets and not by the calls he makes. Measure a physical, countable thing: vouchers processed, square feet cleaned, people hired. Any measurement beats none.

Leading indicators cut windows into the black box. Adopt one only if you will act on it, because monitoring an indicator you will ignore produces anxiety and nothing else. Grove calls the drift between successive monthly forecasts the sharpest business signal he ever found.

A Few Activities Carry the Week

Grove logged one of his own days and counted about twenty-five separate activities. They sort into five types: gathering information, giving information, deciding, nudging, and acting as a role model. Two thirds of the day went to meetings.

The activities are not equal, and the difference is what each produces downstream. One manager reaches many people at once, as when he defines a planning process for about 200 participants. A brief, well-focused set of words changes a person's behavior for a long stretch.

The same arithmetic runs negative. A manager who arrives unprepared wastes the room, and a depressed manager infects the group. Waffling costs the most, because no green light reads as a red light. Meddling narrows what the subordinate believes he is expected to do.

Delegation is monitoring at the lowest-value stage. Both people need a shared information base, or the work needs specific instructions and the return collapses. So delegate the work you know best, because that is the work you can monitor.

Monitoring means checking that an activity runs in line with expectations. Review rough drafts of the reports you delegated, and vary the sampling rate by the person's experience with that task. Delegation without follow-through is abdication.

Ninety Minutes Buys Eighty Hours

The one-on-one is a meeting between supervisor and subordinate, for mutual teaching and exchange of information. Grove sets an hour as the minimum and holds it in or near the subordinate's work area.

Frequency comes from task-relevant maturity. Meet once a week with someone inexperienced in a specific situation, and once every few weeks with an experienced veteran.

It is the subordinate's meeting. He sets the agenda and the tone, brings an outline, and walks the supervisor through it. A supervisor with eight subordinates otherwise prepares eight times while each subordinate prepares once.

The content is indicators showing trouble, events since the last meeting, people problems, and potential problems. Hunches count, because a hunch triggers a look into the organizational black box. Grove's rule for the supervisor fits four words: ask one more question.

Grove prices the practice. Ninety minutes of a supervisor's time improves a subordinate's work for two weeks, or some eighty-plus hours.

Grove puts a manager's hour at about 100 dollars, so ten managers for two hours costs 2,000 dollars. Scheduled meetings absorb maybe 80 percent of the issues. Grove treats more than 25 percent of a manager's time in ad hoc meetings as a broken organization.

Settle the Six Questions Before Anyone Meets

Every decision runs free discussion, then a clear decision, then full support. Support does not require agreement, only a commitment to back the decision. Decisions belong at the lowest competent level, where technical knowledge meets judgment earned from a person's own errors.

Peers alone stall. At Intel's first management training session a group of organizational equals went around in circles for some fifteen minutes without noticing. The returning chairman slapped the table and it resolved at once. Peers wait for a consensus they can state as a group opinion, and they fear sounding dumb.

The fix is peer-plus-one, a senior person in the room to shape the meeting. Grove grounds the rest in one observation: nobody ever died from making a wrong business decision. The senior person uses position power only after the discussion stage closes with no consensus.

Six questions settle the structure before anyone walks in.

• Name the decision that has to be made.
• Name the date by which it has to be made.
• Name who decides.
• Name who gets consulted before the decision.
• Name who ratifies or vetoes it.
• Name who gets informed afterward.

Grove runs them on a Philippine plant expansion that roughly doubles capacity. The plant is needed in two to two and a half years, so the offsets put the decision inside a month. The deciding body is four managers from two organizations, with about equal representation. Grove ratifies, Gordon Moore is informed, and the outcome is a building next door capped at four stories.

Maturity Picks the Style

Grove rotated middle managers among similar groups, and neither the manager nor the group kept its performance level. High output came from particular combinations of manager and group.

Task-relevant maturity is his name for the variable: achievement orientation, readiness to take responsibility, education, training, and experience, read against one specific job. A star field sales manager moved into a factory job was overwhelmed, because Intel confused general competence with maturity on that task.

The style follows from the reading.

• Low maturity gets a structured, task-oriented style that tells what, when, and how.
• Medium maturity gets two-way communication, support, and mutual reasoning.
• High maturity gets objectives and monitoring, and little else from the manager.

Monitoring never stops at any level, because its presence or absence separates delegating from abdicating.

Of the supervisors surveyed, 90 percent saw their own style as more communicating or delegating than their subordinates did. One associate let a junior fail in order to teach him. Grove refuses the trade, because the subordinate's tuition is paid by his customers.

Culture Is the Third Mode of Control

One drive covers all three modes. Buying tires runs on self-interest. Stopping at a red light runs on a contract. Helping at an accident runs on concern for another life.

Free-market forces need an exact price, which most work lacks. The daily work cannot be specified, so a contract also gives the supervisor the right to monitor and correct. Cultural values take over when the environment changes faster than the rules do, and they get built only through shared experience.

The mode follows from motivation crossed with the CUA factor, his shorthand for complexity, uncertainty, and ambiguity. Self-interest with a low CUA factor gets the market. Group interest with a high CUA factor gets cultural values. Self-interest with a high CUA factor gets chaos, because no mode of control works there.

So a new employee gets a structured job with a low CUA factor, and the CUA rises as shared experience accumulates. An outside senior hire lands high self-interest in a high-CUA organization with no shared base. Grove says all you can do there is cross your fingers.

Practical Applications

Write the output equation for your own role first. List the groups whose output you count, including the ones you influence without commanding.

Pick two indicators for each output you own, one for quantity and one for the quality that quantity can damage. Make both physical and countable, review them weekly, and drop any indicator you will not act on.

Put a one-on-one on the calendar for every direct report. An hour minimum, in their area, on their agenda, from their written outline. Set the cadence by task-relevant maturity: weekly for someone new to the task, every few weeks for a veteran.

Answer the six questions on one page before the next decision meeting, and circulate it. Keep the room under eight people, because decision-making is not a spectator sport.

Teach the courses yourself. At Intel an ion implanter drifted out of tune, and the operator did not know the signs. Material worth more than one million dollars went through the machine and got scrapped. The other side is four lectures and twelve hours of your own work. Ten students will work about twenty thousand hours next year, and a 1 percent improvement returns two hundred hours.

Who This Is For

Managers of managers get the most from this book, and so do the know-how people Grove writes to directly.

The book dates to 1983, written out of a semiconductor company Grove ran, with an introduction revised in 1995. The evidence is Intel plus his own logged days, and no outside study checks any of it.

Some of it assumes a large manufacturing organization. The hybrid form and dual reporting need functional groups big enough to matter, and about two thirds of Intel worked in them. The span of six to eight reports and the meeting cost of about 100 dollars a manager hour belong to that period.

The rest transfers to a team of six engineers without changing shape. That covers the output equation, the limiting step, paired indicators, the one-on-one, and task-relevant maturity. Grove does treat the meeting as the only medium for managerial work, which he asserts rather than measures. He wrote before distributed teams existed.

Skip it if you have no direct reports and no group whose output you affect. Skip the compensation and appraisal chapters if you do not set pay.

The Decision

Open next week's calendar and run three counts on it. Grove calls the calendar the medium of a manager's forecast, so it already shows what next week produces.

Count the one-on-ones, one per direct report, an hour each. Count the hours in meetings called at short notice, and compare that total against a quarter of your week. Find the limiting step, the one commitment everything else works around, and check whether the rest of the week offsets backward from it.

Most calendars fail on the second count. Delete the blocks that produce nothing downstream, and put the one-on-ones in the holes they leave.

Change the calendar, or accept the output it already forecasts.