High Output Management

Key Insights: High Output Management

Andy Grove built Intel and wrote this book to treat management as a production process — measurable, improvable, teachable. The manager is not a coordinator of other people's work but a producer whose own output can be counted.

5 Pillars of Success

1. Managerial output is a sum of other people's work Grove's definition is blunt: a manager's output equals the output of their own organization plus the output of the organizations they influence. That turns management into arithmetic. Time spent on work that raises nobody's output is low-leverage or actively negative. Leverage has three levers: perform an activity faster, shift your mix of activities toward higher-leverage ones, or raise the leverage of the activity itself. Training ten supervisors once reaches every employee those supervisors manage — the clearest example Grove gives of work that pays twice. Equally, a decision left to waffle holds up everyone waiting on it, which is leverage working in reverse.

2. Every flow has a limiting step Grove's breakfast factory — a small stand selling boiled eggs, toast and coffee — carries most of the book's production logic. The egg boiler takes the longest, so it sets the plant's capacity no matter how fast the waiter moves. Fix the limiting step first: add a second boiler, buy a faster one, or hold cooked eggs in a warming tray against forecast demand. Batching matters too: making toast in larger runs cuts the per-unit cost. Grove applies quality control to the same flow — inspect at the lowest value-added stage, because defects caught early are cheap to discard, and watch the process, not only the finished product.

3. Match supervision to task-relevant maturity Task-relevant maturity is not a person's seniority or general competence; it is their experience with this specific task. At low task-relevant maturity a manager should give structured, task-oriented direction and check progress closely. At medium maturity the style shifts to two-way communication and support. At high maturity the manager delegates, sets objectives and stays out of the way. Grove is equally hard on the opposite error: meddling in a task a subordinate has already mastered tells that person you do not trust them, and it costs you their confidence.

4. Meetings are the medium of managerial work Grove argues that managerial work happens in meetings, so meetings are a production tool, not an interruption. The one-on-one belongs to the subordinate: they set the agenda, raise problems early, and the supervisor listens and supplies context. Staff meetings spread information across a peer group that would otherwise each reinvent it. Decision-making follows a three-step sequence — free discussion, a clear decision, then full support — and Grove illustrates the cost of avoiding it with Intel's exit from memory chips: he asked Gordon Moore what a new management team would do, heard "get out of memories," and realised they should walk out and come back in as that new team.

5. Measure paired indicators and give honest feedback An indicator drives behaviour — the Hawthorne effect, Grove notes, is the rule rather than the exception — so a single number invites cheating against it. Use paired indicators: count eggs served alongside complaints about cold eggs, or sales alongside customer satisfaction, so improving one cannot quietly destroy the other. The same logic governs people. Performance reviews are the most important form of task-relevant feedback, and Grove insists they be honest, specific and example-based, tied to objectives agreed in advance. Motivation follows competence and achievement, so a manager's job is to create the conditions in which capable people can succeed.

Key Takeaway > "The output of a manager is the output of the organizational units under his supervision or influence."

Practical Task 1. Write down your three current objectives and, next to each, the output it is supposed to increase. 2. Audit last week's calendar and tag every block as high, medium or low leverage. 3. Identify your limiting step — the person, approval or process everything else waits on — and take one concrete action to widen it this week. 4. Pick one subordinate and one specific task. Set their task-relevant maturity for it: does it need direction, support, or delegation? 5. Choose one indicator you report on and pair it with a counter-indicator so the number cannot be gamed, then review both in your next one-on-one.