What Is the Job That Only the Person at the Top Can Do?

When people are promoted into responsibility for something larger than their own work, most of them respond the same way. They work harder on the parts. They review more drafts, sit in more meetings, and personally fix the problems that land on their desk. For a while this looks like leadership, and it is often rewarded as leadership.

At the scale of a whole organization, the approach quietly fails. The results no longer come from any single person's effort, however large. They come from other people, other people's capital, and systems that keep running when the leader is somewhere else.

The chief executive's job makes this plain because it is the one role whose work is the whole. The best way to understand that job is not as a bundle of duties but as leverage thinking: a leader's output is everything that happens because of their decisions. Seen that way, the role narrows to a handful of decisions no one else can make, and each of those decisions follows a logic that can be learned and applied well below the top of any organization.

Output is what happens because of a decision

Andy Grove, who led Intel through its most dangerous years, put the principle in one line: a manager's output is the output of the organization they influence. The consequence is uncomfortable for anyone who built a career on personal production. The work shifts from doing the work to deciding what work gets done, by whom, and with what resources.

This is not an argument for doing less. It is an argument for measuring differently. The useful question stops being how much got finished and becomes where one hour of attention lifts the most other hours. A leader who spends the week on tasks others could do has spent the week on low-leverage work, however hard it felt.

The same logic defines which decisions matter. The former Procter & Gamble chief executive A.G. Lafley argued that the chief executive is the only person who connects the outside world of customers, markets, and society to the inside of the organization. The decisions that come from that position cannot be delegated: direction, capital, people, the operating system, and judgment under pressure.

The outside view reveals what familiarity hides

The hardest of those decisions is usually the one closest to home. In 1985 Intel was losing the memory-chip business it had been founded on. Grove and Gordon Moore had debated the problem for months without resolving it.

Grove then asked Moore what a newly hired chief executive would do. Moore answered immediately: get out of memories. Grove's reply was to walk out the door, come back in, and do it themselves. Intel left memories and bet the company on microprocessors.

Nothing new was learned in that exchange. What changed was the vantage point. A newcomer carries none of the history, pride, or sunk effort that makes an obvious move feel impossible from the inside. Imagining a capable successor taking over tomorrow is therefore a repeatable tool, not a one-time flash of insight, and it works on a product line, a project, or a career as well as on a company.

The budget is the strategy made real

If direction is where an organization should go, capital allocation is whether it actually goes there. It is the least discussed part of the top job and arguably the most consequential. William Thorndike's study of chief executives with exceptional long-run records, including Henry Singleton at Teledyne and Katharine Graham at the Washington Post, found that their shared strength was disciplined decisions about where cash goes.

Every surplus dollar has five possible uses: reinvestment, acquisition, dividends, share buybacks, or paying down debt. The strongest allocators compared the likely return on each, every time, instead of repeating last year's pattern. Warren Buffett's discipline sharpens the comparison: growth creates value only when the return on invested capital exceeds its cost, so growth for its own sake can quietly destroy value.

This is why a plan and a budget can tell two different stories, and why the budget is the more honest one. Where the next dollar, the next hour, and the next hire go reveals what an organization actually believes. When Steve Jobs returned to Apple in 1997, he cut a sprawling product line to a grid of four: consumer and professional, desktop and portable. Every good project funded is a great one delayed.

Choosing who decides, and building what runs without you

Allocation applies to people as well as money, and here the leverage is greatest. Jim Collins's principle of first who, then what holds that the right people adapt as strategy changes, while the wrong people struggle even with the right strategy. Selecting who decides shapes every decision that follows.

Delegation then depends on fit, not rank. Grove's idea of task-relevant maturity holds that how closely to manage someone depends on their experience with the specific task in front of them. Clear decision rights, stating who recommends and who decides, turn a team of capable people into speed rather than into a queue waiting for approval.

What keeps those decisions working is less a matter of inspiration than of design. Jeff Bezos put it directly: good intentions do not produce reliable outcomes, mechanisms do. A mechanism is a tool, an owner, and a review that runs every time.

Alan Mulally's weekly Business Plan Review at Ford shows the idea at full scale. The company was forecasting heavy losses while every project on the executives' status charts showed green. When one executive finally presented a red chart, Mulally applauded, and the truth began to travel upward. A fixed operating cadence is only as useful as the honesty that flows through it, and a leader hears only what the organization believes is safe to say.

Decision speed should match reversibility

The last of the non-delegable decisions is how to judge when the stakes are high and the information is incomplete. Bezos divided decisions into one-way doors, which are hard to reverse, and two-way doors, which can be walked back.

Most decisions are two-way doors. They should be made quickly, on roughly seventy percent of the information one would like, because the cost of waiting usually exceeds the cost of an occasional correction. Deliberation belongs to the few irreversible choices, and a leader who treats every decision as a one-way door slows the whole organization to the pace of its most cautious call.

Crises compress this judgment into days. When Tylenol capsules were tampered with in 1982 and several people died, James Burke's Johnson & Johnson pulled the product from shelves nationwide, communicated openly, and put customer safety ahead of short-term cost. The brand recovered, and the response became a model of crisis leadership built on speed, candor, and visible presence.

Running any scope as its chief executive

Taken together, these ideas form one position rather than a list. The top role is defined by a small set of decisions that belong to whoever owns the whole: the outside view on direction, the discipline of comparing returns before committing resources, the leverage of choosing who decides, the mechanisms that make good behavior automatic, and decision speed matched to reversibility.

None of this requires the title. A team is a company, a project is a portfolio of bets, and a career is an enterprise with one employee and many stakeholders. Anyone who owns an outcome can write a single page that states the diagnosis, the few choices, the resource shifts, the people bets, and one mechanism to install.

That page is the job at any scale. The measure of it is not how much the author personally produced, but everything that happened because of the decisions it recorded.