Business thinker

Andy Grove's OKRs, one-on-ones and managerial leverage

Updated 4 November 2026

Not affiliated with or endorsed by Andy Grove or their companies.

Andy Grove ran Intel and wrote High Output Management, a book on management that opens with the basics of production. Three of his ideas are covered here: OKRs as he ran them at Intel, the one-on-one as a core management tool, and the equation that counts a manager's output as the output of their team plus the teams they influence. The book is not freely available online, so this page works from Ben Horowitz's foreword to its 2015 edition and from John Doerr's What Matters site.

A manager's output is the team's output

Horowitz's foreword quotes the equation that opens the book's account of management: "A manager's output = the output of his organization + the output of the neighboring organizations under his influence." Horowitz reads it as the difference between a manager and an individual contributor: what a manager knows is worth something only when it raises the output of others.

The foreword also quotes Grove on why someone isn't doing their job: either they can't or they won't, so the manager's ways to raise that person's output are training and motivation. The book's last chapter is titled "Why Training Is the Boss's Job".

Take a fictional support lead who spends 14 hours a week answering tickets. Applied this way, those hours count only as the lead's own work. An hour spent teaching a newer agent how to escalate, or raising support's needs in the product team's meeting so releases stop surprising customers, counts toward the output of the team and of a neighbouring team.

Task-relevant maturity: how hands-on to be

Horowitz writes that the book's answer to "hands-on or hands-off?" is that it depends on the person's maturity in the specific task. Someone new to the task needs hands-on training; someone experienced in it can be delegated to. He quotes Grove's objection to letting a struggling subordinate learn from their own mistakes: the tuition is paid by the customers.

In the support example, an agent who has handled billing tickets for two years can take them over from the lead outright; an agent who joined last month needs pairing on escalations first.

One-on-ones

The foreword notes that the book's section on meetings starts from first principles with how to conduct a one-on-one. Horowitz gives his own view that the one-on-one may be the best source of organizational knowledge a manager has, and that managers who skip them understand little of what happens in their organizations. The book's detailed advice on how to run one is not in the sources we could open, so this page doesn't restate it; the library's one-on-one prep skill covers the meeting itself.

OKRs as Grove ran them at Intel

John Doerr's What Matters site says Grove built the method at Intel on Peter Drucker's management by objectives, first calling it iMBOs (Intel Management by Objectives), and rarely mentioned objectives without tying them to key results. It lists the differences from MBOs: Grove's were set quarterly, not annually, and kept separate from compensation. His grading was yes or no: did you meet the key result or not.

The site also says that when Grove first introduced OKRs at Intel, only one type existed, aspirational; the split into committed and aspirational OKRs came later at Google. It describes Grove as favouring output key results, and quotes him: "Stressing output is the key to increasing productivity." Doerr, who learned the method in Grove's seminars, named it OKRs and introduced it to Google in 1999.

The foreword adds a planning point that fits OKRs: it quotes Grove that "today's gap represents a failure of planning sometime in the past." The time to work on next quarter's output is now.

How to apply it to your own team

1. List a typical week's activities with hours.
2. Label each one: does it raise your team's output, a neighbouring team's, or is it your own hands-on work?
3. For each piece of your own work, name who on the team could own it and how experienced they are in it: delegate to the experienced, train the new.
4. For each performance gap, write whether it looks like can't or won't, and the evidence; plan training for one, a conversation for the other.
5. Check who hasn't had a one-on-one recently.
6. Check that next quarter's key results measure output that can be graded met or not met, and that none is tied to anyone's pay.

What the sources say, and what they don't

The equation, can't or won't, task-relevant maturity and the planning line are Grove's words or ideas as Horowitz's foreword quotes and summarises them; the view of one-on-ones as the best source of knowledge is Horowitz's. The OKR history and practice come from What Matters, John Doerr's site, not from Grove directly. We left out the book's detailed rules for one-on-ones and its list of high-leverage activities because we could not open a source that states them. The support team example is ours.