Andy Grove: Management as a Rigorous, Learnable Discipline

Andy Grove: Management as a Rigorous, Learnable Discipline

Andy Grove led Intel through some of the most consequential decisions in the company's history, including the wrenching exit from the memory-chip business to refocus entirely on microprocessors — a decision he later described as something he could only make by imagining he'd just been fired and a new CEO was starting fresh, asking what that person would do differently. That same unsentimental, output-focused clarity defines his enduring contribution to management thinking: the book High Output Management, still assigned reading across the tech industry decades after it was written.

Output, not effort, is what management is judged on

Grove's central argument was that a manager's own output isn't their personal work product — it's the output of their entire team and every neighboring team they influence. This reframes what a manager's time should actually go toward: activities with the highest leverage on other people's output, not tasks a manager could just do themselves faster. It's a direct rebuttal to the instinct many new engineering managers have to stay heads-down in code because it feels more "productive" than an hour spent unblocking someone else.

OKRs started here

The goal-setting framework known today as OKRs (Objectives and Key Results) — now used at Google, and countless startups since — originated at Intel under Grove, built on his belief that objectives should be few, ambitious, and paired with measurable key results that make it obvious whether you actually achieved them or just felt busy. This is the same discipline behind good sprint goals: not "work on the checkout flow," but a specific, falsifiable outcome you'll know you hit or missed.

One-on-ones as a real management tool, not a status meeting

Grove was explicit that a one-on-one's agenda should be set primarily by the report, not the manager — the manager's job is to listen for problems the report might not even fully realize they have yet, since a manager typically has visibility the individual contributor doesn't. That's a meaningfully different posture than a manager running through their own checklist of questions.

The takeaway for engineering leads

Grove's whole framework rests on a single reframe worth sitting with: your output as a lead isn't measured by what you personally shipped this week, it's measured by what your team collectively shipped because of decisions you made about where to spend your attention. If your 1:1s are status updates you drive, or you're regularly the person who "just does it faster" instead of delegating, Grove's model would say you're optimizing for the wrong output entirely.