DEEP READING · READ 913
A Silicon Valley coach's operating playbook · Matt Mochary · 2019
The book stakes everything on one rather blunt claim: the bottleneck in an early company is almost always the CEO — not the market, not the competition, not an underpowered team. So it skips vision and strategy frameworks and reads like an assembly manual instead: how one person should structure a day, how two people should get a hard thing said, how a room full of people can hold a meeting that isn't a waste, how a company should let its information out. Its ambition is not to make you understand something but to make you swap out three or four specific behaviours this week — because the author's working assumption is that founders are never short of principles, only of the step where a principle gets installed into a calendar, a template, a process.
Matt Mochary is a Silicon Valley executive coach — a founder and investor earlier on, then a full-time coach to CEOs of fast-scaling technology companies; the method was written up as a public curriculum and as this book, which is deliberately thin, deliberately fragmented, almost entirely bullets and templates, and has long been available free online. Its shelf position is unambiguous: this is not Grove's High Output Management, a theory derived from first principles, but a use-it-today checklist — nearly every chapter traces back to somebody else's book (time management out of the GTD lineage, communication out of Nonviolent Communication, hiring out of Who, feedback out of Radical Candor), and the author never pretends otherwise. The value is in the curation and the sequence, not the originality.
The rule is almost crude in its simplicity: write down your single Top Goal, and then spend the first hour of every workday on it alone — no inbox, no messages, no meetings. The author's experience is that this one hour routinely out-produces the entire rest of the day (paraphrased).
What it treats is a disease every manager knows: important-but-not-urgent always loses to urgent-but-not-important. The mechanism is an asymmetry of push. Urgent things arrive with external force behind them — somebody is waiting, something is buzzing — while important things are pushed by nothing but your own will. And the least contaminated stretch of the day is the beginning: the later it gets, the more your calendar fills with other people's agendas and the more your judgment has been eroded by a string of decisions. So "the first hour" is not a time-management trick, it's a land grab — you claim the resource for your own work before the world wakes up.
What it changes is this: a priority that does not occupy a specific block on your calendar is not a priority, it's a wish. You say a thing matters most — go look at your calendar. If it didn't get a single slot this week, your actual top goal is something else. Likewise, that project everyone agrees is critical but which never moves: the problem was never insufficient agreement, it's that nobody was ever assigned a protected stretch of time for it.
This frame is borrowed from Gay Hendricks' The Big Leap, and it sorts what you do into four layers: incompetence (you're bad at it, plenty of people are far better), competence (you can do it, so can anyone), excellence (you're very good, people praise you for it and pay you for it), and genius (you're good at it, only you do it quite this way, and doing it gives you energy).
The insight sits on the third layer. What traps people for a whole career is not the work they're bad at; it's the work they're extremely good at, are constantly asked for, and feel hollow after. Incompetence is easy to hand off — you're visibly bad, nobody fights you. Excellence is brutally hard to hand off, because three forces push at once: people ask for you by name, quality genuinely dips in the short term when you let go, and "I'm great at this" has become part of who you think you are. Which is how founders end up running companies of several hundred people while still personally doing the thing they were brilliant at five years ago.
One honest caveat: this is a coaching self-inquiry frame, not an evidence-backed psychological model — there is no instrument for "zone of genius" and no way to falsify it. But as a question-generator it earns its place, because it swaps the usual question for a better one: don't ask "what am I good at," ask "what am I good at, that the company needs, and that leaves me with more energy than it took" — that intersection is the small patch to defend, and the rest should go, progressively. The threshold for handing something off is equally blunt: if someone can do it to eighty percent of your standard, give it to them, because that missing twenty is worth far less than your hours back in the genius zone.
Concretely: list every activity on the last week or two of your calendar and mark each one "+" (left you with more energy) or "−" (drained you); then take each "−" through four questions — can it be dropped, delegated, automated, or restructured into a less draining form? Then run it the other way and deliberately increase the "+" items.
Why it deserves to be taken seriously: hours are interchangeable, energy is not. Two sixty-minute meetings look identical in a time ledger, but one leaves you able to work well for three more hours and the other costs you the afternoon. Account only in time and you can never see where the real leakage is. And what drains people is rarely volume; it's a few specific categories — the same unresolved personnel problem revisited for the fourth time, a meeting where you can neither contribute nor leave, an argument you're pulled into with no authority to settle. They occupy one small box in the time ledger and half a day in the energy one.
There's a boundary here that has to be stated plainly: "delete everything you don't enjoy" is neither possible nor responsible. Every company has unpleasant, necessary work, and founders least of all get to push it outward — that isn't energy management, it's dumping. The legitimate use of this tool is to find the losses that can honestly disappear: the repetitive, the automatable, and the work that belonged to someone else but drifted onto your desk.
It's a table. Every thing in the company that needs an owner — the server bill, the hiring pipeline, support response times, the office keycards — gets its own row, and each row gets one person's name. Not "the ops team." Not two people. One.
The mechanism behind it is the well-documented diffusion of responsibility: the more people present, the less pressure any individual feels to act — shared ownership doesn't produce redundancy, it produces two people each assuming the other has it. So "two owners" is functionally "no owner." And the real power of the AOR table shows up not day to day but the moment something breaks: a ball drops, and the first move isn't hunting for who slacked, it's opening the table and asking two questions — is this on the list, and whose name is on it? If it isn't listed, that's a hole in the table; patch it. If it is listed and that person didn't do it, now you have a real accountability conversation, and both sides know exactly what it's about.
What this changes is your default attribution for organisational failure. Most teams reach first for attitude and ability ("he doesn't care," "she isn't good enough") — explanations that wound and can't be acted on. AOR offers one that's more often true and far more repairable: the thing was never assigned to a specific human being. As a bonus it doubles as an org health check: finish the table and you'll usually discover two things at once — some names carry a dozen rows (that person is your bottleneck and a resignation-shaped hole waiting to happen), and some important rows are blank.
He gives the weekly meeting a fixed four-part structure: (1) good news to open — one item each, work or personal; (2) review the metrics dashboard together; (3) put every open issue on one list and vote it into priority order; (4) work down from the top vote, discussing each until it resolves — and the end state of every item is "who · does what · by when." All status updates and progress reports are written in advance and read before the meeting.
Each part has a reason worth spelling out. "Good news first" is not team-building filler: people in a defensive or braced state instinctively say less and say safer things, and a meeting whose whole purpose is surfacing problems depends on people volunteering the ugly ones; a round of good news is a very cheap state change. Voting solves the biggest waste in meetings — an agenda hijacked by whoever talks most or spoke first; a vote lets the room's collective judgment set the order instead of volume. And writing status updates in advance follows from a more basic principle: the only genuinely scarce resource in a meeting is that these particular people are in the room at the same time — and one-way information transfer doesn't require that condition at all. It should be asynchronous text. Move reporting out and you return the entire meeting to argument and decision.
So there's exactly one test for whether a meeting was any good: did it produce new "who + what + by when"? A meeting that generated information but no ownership and no deadline produced nothing, however lively it felt.
Feedback has three rules: give it immediately (while the memory is fresh and the cost is still small), give it privately (criticism in public triggers defence instantly, and from then on the person is rehearsing a rebuttal rather than listening), and ask permission before you start — "can I give you some feedback?" The seemingly redundant question does real work: it moves the other person from being subjected to something to consenting to it, and the defensiveness drops a notch. Praise runs the opposite way: the more public the better. And the rule most often skipped and most effective: as the leader, ask for feedback first, and thank the person who says the hard thing in front of everyone — whether a team tells the truth depends entirely on what happened to the first person who did.
Conflict is handled with a mirroring process descended from Marshall Rosenberg's Nonviolent Communication. Three people are present, the two parties and a neutral third: A states their view and feelings; B may not rebut, only reflect A's words back until A says "yes, that's it." Then they swap. Only after both have been mirrored to their own satisfaction may anyone propose a solution — and the rule is that each person must propose the solution the other would be happy with.
What powers this is a counter-intuitive fact: most conflicts are not fuelled by the disagreement but by "I wasn't heard." Once someone is genuinely satisfied that the other side understood them — and accurate reflection is the only way to prove that; nodding and "I hear you" don't count — the emotional level visibly drops, and the disagreement itself usually shrinks into something specific and negotiable. You can use just this one piece today: next argument, don't counter — restate their position and ask "did I get that right?" A surprising share of fights end there.
Hiring largely follows Who (Geoff Smart and Randy Street): before you meet anyone, write a scorecard — the mission of the role, the three to five measurable outcomes it must deliver in the first year, and the competencies required. Why that step can't be skipped: without a written definition of what success looks like, an interview inevitably degrades into "do I like this person," and people reliably prefer people like themselves (similarity bias), so the team keeps cloning itself, the résumés keep looking better, and delivery is never actually secured. The scorecard converts interviewing from impression into comparison: every question tests a stated outcome rather than sampling the vibe.
Firing he puts more bluntly: by the time you are seriously turning over whether to let someone go, you generally already have your answer (paraphrased) — and the cost of stalling doesn't land on that person, it lands on everyone else. The team almost always knows before you do, and every extra month you wait is paid out of your credibility with them: what they conclude is "so underperforming here is fine," and once that conclusion sets, the whole standard drops a level. On execution he pairs "fast" with "generous": real severance, a dignified exit, no humiliation — because everyone who stays is watching how the person who left was treated.
But this needs the half the author underplays: "fire fast" is, in many cases, a cover for a management failure. When someone can't deliver, the common causes are a vague role definition, expectations never made explicit, feedback never given — debts owed by the manager, not evidence the person is bad. Employment law and workplace norms also differ enormously between countries, and this Silicon Valley tempo is in places neither legal nor appropriate. The correct use is: first confirm you actually said it clearly, gave the feedback, and offered a real chance to correct — and if all three are true and it still isn't working, then stop dragging it out.
His default is to share nearly everything: finances, runway (how long the current burn rate leaves you), board materials, strategy discussions, even compensation structure, unless there's a hard legal or competitive reason not to. The argument isn't moral, it's operational: withholding information doesn't calm people, it makes them guess; and once people guess, they guess the worst case and then act on it — job-hunting, hoarding, forming private alliances. Opening the numbers costs one round of discomfort; sitting on them costs continuous friction. The same logic applied to boards is the no-surprises rule: materials out several days ahead, bad news first, and the meeting itself spent on the genuinely hard questions rather than reading slides aloud.
Here too the missing half needs saying: transparency has real failure modes. Transparency without psychological safety underneath it turns into surveillance — everyone's numbers public means everyone is permanently being ranked. Financial detail delivered to people with no decision rights often supplies anxiety without supplying judgment. And at certain moments — a raise in progress, layoffs under consideration, an acquisition — premature disclosure can produce a self-fulfilling collapse. The sturdier reading is: default to transparent, but ship the interpretation alongside the data — how to read these numbers and what you can do about them. Numbers without explanation just outsources the anxiety to your team.
There is really only one through-line, and the ordering is deliberate: fix yourself, then fix the interfaces between people, then fix the interface between the company and the outside.
Layer one, yourself: get everything out of your head into a trusted system (inbox empty, calendar as the single source of truth), lock the most important work in with a Top Goal plus the first hour, use the zone of genius to decide what leaves your desk, use the energy audit to find the leaks, and underneath it all sleep, exercise and food — on the grounds that the CEO's physical state is a company asset. The premise: an exhausted founder whose schedule belongs to other people cannot make good decisions, and in an early company decision quality is very nearly everything.
Layer two, the interfaces: the AOR table settles who owns what; the meeting structure settles how a group decides together; the feedback rules settle how the hard sentence gets said without detonating the relationship; the mirroring process settles what to do once it has detonated; scorecards and fast-but-generous exits settle who comes in and who leaves. The premise: output is lost mainly in the gaps between people, not at the ceiling of any individual's ability.
Layer three, outward: default transparency, no surprises for the board, and fundraising run as a time-boxed parallel process — everyone in motion at once to create real competition, raising on momentum rather than at the point of need.
And running through all of it, the judgment that best reveals the book's character: a CEO has only three actual jobs — set the direction and keep communicating it, recruit and retain the best people, and make sure there is always money in the bank (a formulation usually traced to the investor Fred Wilson, which Mochary is quoting; paraphrased). Everything else can and should be handed off. Every tool in the book ultimately serves one purpose: moving the CEO's attention out of the daily flood and back onto those three things.
1. The bottleneck is almost always the CEO. So the first repair isn't the company, it's this person's day: wherever the founder's energy goes, that's where the company grows.
2. Write down one Top Goal and spend the first hour of every workday on it alone — no inbox, no meetings. Urgent work comes with external force behind it; important work is pushed by nothing but you, and the start of the day is its least contaminated stretch.
3. A priority with no specific block on your calendar is not a priority, it's a wish. To learn what your real top goal is, don't listen to what you say — look at this week's calendar.
4. What traps people is never the work they're bad at; it's the excellence zone — very good at it, constantly asked for it, hollow afterwards. Change the test: not "what am I good at," but "what am I good at, that the company needs, that leaves me with more energy." If someone can do it to eighty percent, hand it over.
5. Read your calendar as a P&L: mark every activity + or −. Hours are interchangeable, energy isn't — two sixty-minute meetings look identical in a time ledger while one buys you three productive hours and the other costs you the afternoon.
6. One name per line. Diffusion of responsibility is real: two owners is functionally no owner. When a ball drops, ask "is this on the table and whose name is on it" before you ask who wasn't trying.
7. The only scarce resource in a meeting is that these people are present at the same time — so write the status updates in advance and use the room only for what needs interaction: vote the issues into order, then talk each to a decision. One test: did it produce new "who + what + by when"?
8. Feedback immediate and private, opened with "can I give you some feedback?"; praise in public. And whether a team tells the truth depends on what happened to the first person who did — so ask for feedback yourself first, and thank them for it publicly.
9. Most conflict is fuelled not by disagreement but by "I wasn't heard." Make each side mirror the other back until they hear "yes, that's it," and only then discuss solutions — each proposing the one the other would be happy with.
10. Withholding information doesn't calm people, it makes them guess — and they guess the worst case, then act on it. Default to transparent, no surprises for the board; but ship the interpretation with the data, because numbers without explanation just outsource the anxiety.