创始人是一种态度，CEO 是一门手艺：失败模式、Magic Box 退出与招聘

Jonathan Lowenhar runs Enjoy The Work, a firm whose mission is to help founders become great CEOs. Host Lenny Rachitsky frames it as studying what causes founders pain and failure and turning that into frameworks and training. Lowenhar's own background spans a division of a public company, a private equity CEO role and two startups, and he left the second one, he says, grossly unhealthy and 25 pounds overweight.

While reflecting, he noticed that every one of those companies ran well in the same way, and he came to believe every well-run company has an unmistakable rhythm that you could see in a couple of days of watching. So he began asking investors three questions. They described the ideal founder with words like grit, tenacity, courage and insight, but the ideal startup CEO with skills: building, selling, recruiting, raising capital, organizing people and financial planning.

His third question was how a founder learns those skills while already doing the job. The answer, he says, was blank stares, over and over. That gap is what set him on a ten-year path of teaching founders those skills so investors do not replace them and they can stay in the job as long as they want.

Lowenhar says the founder-mode article "got me hot" because it felt like an excuse, giving founders permission not to learn the job. He describes the ascent of a startup: invent something, get customers to try it, build a repeatable business model, generate demand, build an operation to handle it, and eventually turn revenue into cash flow. The idea that founders need not advance their skills for any of that, or that learning them is negative, is "bananas" to him.

What launching, growing, scaling and exiting a company each require is different, he says, and the best CEOs learn them all along the way. He does grant what is unique about a founder: unlike a professional CEO dropped in from outside, a founder knows what built the company and can drop into a product feature, a customer or partner conversation, or a long-time employee's work, and then rise back up. The point is not that manager mode is bad; the greatest CEOs know which mode to calibrate.

Lowenhar separates failure modes for companies from those for CEOs. For companies he lists choosing the wrong market, failing to build something people want to buy, founder dysfunction (more companies die from suicide than homicide, he jokes) and poor execution. He then moves to a more playful catalog of CEO types, starting with the robot CEO who believes emotion should not exist at a startup. His formula for them is that humans have emotions, startups need humans, so startups are messy.

The pleaser CEO cares more about being liked than running a business, cannot deliver hard news or break ties, and hides, hoping problems go away. The perfectionist CEO has the most beautiful product delivered minutes before bankruptcy; they are slow and build a team that cannot disagree or use gut, when early-stage startups have more questions than answers. The angry CEO wakes up angry and takes it out on the first employee, and Lowenhar notes that no one wants to work for that person, whatever the equity is worth.

The laissez-faire CEO believes great hires can be ignored, when he has never met anyone who did not benefit from good management. The micromanager is the opposite: they treat everyone as a child who needs supervision and do everyone's work until everyone quits. Others are comfortable either with the brake, refusing to spend and getting lapped, or with the accelerator, running out of money fast. Lenny recaps the list, and Lowenhar adds that founders usually have a little of many of these, and that those who refuse to admit it are probably not coachable.

Lowenhar says the most common mode he sees is ready, fire, aim, and that it has grown as founders can no longer paper over weak execution with easy capital. Most CEOs, he notes, have never run a business before and are bad at planning. Planning need not be a months-long bureaucratic exercise; it can be a light bottoms-up look at what you are trying to achieve, how to quantify it, what resources and people are needed, and how to shorten feedback loops.

Companies, he says, are always working backwards from one of four things: an exit, the next fundraise, profitability or winding down. Most choose a fundraise, and he asks founders what must be true, qualitatively and quantitatively, to unlock it. They write it down, work out the actions and cadence, and estimate the resources. With a culture of accountability and communication rituals, the founder no longer has to guess; this only works for a CEO who admits that improvisation got them here but won't get them there.

Lowenhar credits the Magic Box Paradigm to a book of that name by independent banker Ezra Roizen, which his firm has operationalized for teaching. The traditional route is to put up a for-sale sign: build a list of possible buyers, sign NDAs, collect indications of interest, and pray that more than one buyer remains, while terms get re-traded. Magic Box argues the best early-stage outcomes happen differently, with three stages: learn the fantasy, prove the fantasy and quantify the fantasy.

The fantasy is what someone at a larger company imagines happens if they buy you. Large companies want technology and small ones want distribution. Instagram is his example: it reportedly had no revenue and sold for a billion dollars, which no historical math could justify; it was a model of the future. Traditional processes price the past, while Magic Box prices the future.

He tells of a construction-tech company whose video-analytics product caught the interest of a large company sitting on unused video data, where a product leader imagined plugging the platform into their planning. You sell to a person, he stresses, not a company: a champion who wants to find a way to say yes. The cast also includes advocates (pawns who give intel but take no risk), blockers such as legal, procurement or security who can only say no, and the buyer, a committee that cares only about the business case.

Proof exists to arm the champion, who must socialize the deal with the buyer and survive the blockers. Then comes quantification: the CEO tells the champion that the board is asking why so many calories go into this, and builds a rough model of how the buyer's business changes. If the champion responds to the model, you have moved the conversation from history to future. In his story, a sub-$2 million revenue company with a weak Series B outlook sold for generational wealth, after which the CEO spent two-plus years inside the acquirer.

Lowenhar adds two stress points for founders working backwards from an exit. The fantasy includes more than a business impact: the books must be in order, investors must give unanimous consent, key team members must stay and the founder must be a joy to work with, so never puncture it. Corporate development people will appear; they are expert negotiators and can add urgency and momentum to a deal but are far better at negotiating than you.

His rule against a superior negotiator is to move the negotiation asynchronous: say you are not alone in the decision and want to see terms in writing to socialize with advisors, lawyers and co-founders. "Don't negotiate live. You will lose." Compared to enterprise sales, he says the key difference is that you do not push for a compelling event or hint at competing suitors, which tends to be a negative signal until the end. Instead you entice, for example by noting you plan to raise a round in Q1, which would probably make the deal too expensive.

To find the champion without soliciting, he suggests starting with categories of likely buyers, then acquisitive companies with corp dev teams and existing relationships. One tactic from the book is hosting a panel of luminaries where your CEO sits as a peer. Another is a simple, informal LinkedIn note from one founder CEO to another proposing a fun 30-minute chat, and he says roughly one in four to six respond. In the call, ask what they care about in the next year and what keeps them up at night, listening for a fantasy you can squint to fit.

Lowenhar says the CEO has three jobs: make sure everyone knows where the company is going, pick the right people, and give them the tools to win. Most founders hire badly and fall to bias or gamble on a famous employer. His firm has operationalized the book *Who: The A Method for Hiring* by Geoff Smart and Randy Street and updated parts of it, and he names three common mistakes that are easy to fix.

First, hire people who have already done what you need next: do not start from a job description but from twelve months later, imagine clinking champagne, and write what changed in the business. Second, look for a history of raving fans, meaning people who were pulled from job to job by those who loved their work. A quick screen for recruiters is to ask how many of a candidate's former bosses would call and say they are amazing. Third, evaluate core values, since culture at scale is the codification of what matters and the ritual of living it; he splits interviews into culture, functional and technical. Lenny adds that Airbnb used a specially chosen core-values interview team.

For junior roles, he looks for any evidence of the needed behavior, such as selling Girl Scout cookies or tickets for a junior account executive. For executives, he sees three types: the architect who writes the first playbook, the optimizer who tunes it as a team grows, and the scaler who finds ten-times leverage. He would hire an architect only if the candidate had been one before. He ties this to being good at the job, liking it, and the market valuing it, with all three needed to last.

Lowenhar warns that founders often say "I have to hire this person", which invites confirmation bias, and that hiring is never the goal. He returns to the milestones of fundraise, exit or profitability: define what has to change and quantify it, then see which resources are missing, and only then decide whether to hire or rent. He says founders are so deep in the work they rarely step above it, so much of his firm's job is helping them work on the business rather than in it.

For go-to-market, he distills the work into four pieces. The ideal customer profile covers who you want, their qualifications, discovery questions and kill criteria. Marketing covers positioning against enemies (real companies or the status quo) and how to represent that. Demand generation asks how to find the people you want, drawing on the book *Traction* and its nineteen channels, with a two-by-two of impact against effort to pick experiments. Sales is codifying a playbook for discovery, objections, demo and close.

Asked about "brain writing," Lowenhar says he first heard of it from Adam Grant. A founder's voice "has a megaphone attached," so instead of live brainstorming, people write ideas and comments in a doc or survey that others cannot see until combined, ideally without authorship. It dampens the founder effect and evens the field between fast and slow processors and between introverts and extroverts, and Lenny says it matches how he works.

Lowenhar says first-time founders are like twenty-somethings in a bar who say yes to anyone who makes eye contact. With veterans he asks them to imagine building the perfect customer in a petri dish, or, if they already have customers, which one would rave about them on a call. In one case of an enterprise "whale hunting" company with four large customers, one was clearly the favorite, and that reveals the ideal profile.

At the start, he says, you are looking for a white-hot center of opportunity: a small population that is a huge fan and gets huge impact. Adjacencies come later, and Amazon just sold books. The mistake is moving to sales without an ICP, qualifications, discovery questions and kill criteria, meaning conditions under which you do not sign a customer even if they want to. Lenny agrees and Lowenhar calls it the most expensive mistake of the four buckets.

He also cautions against plugging funnel math to hit a revenue target needed for a multiple, which he calls death. Starting from "we must reach 3 million or we're dead" means you are already dead. Instead use the last three to twelve months of history, make reasonable assumptions about how you will improve top-of-funnel, conversion, deal size or length, and then talk honestly about the gap to the financial need.

A founder Lenny emailed said the biggest lesson from Lowenhar was to trust his intuition more. Lowenhar says every startup CEO faces existential moments, such as whether to sell, break up with a co-founder or fire a key employee, and that fear and the lizard brain are poor decision-makers. He asks founders about the quiet voice they hear when still, and says that voice is who you are, while the brain is a pattern-recognition tool.

His firm does not tell founders to go left or right. It trusts the founder's intuition and tries to be the one person fully on their team, since it is not on the cap table, a fiduciary or a board member. He recalls a founder facing a co-founder breakup whose quiet voice had said long before that it was the wrong fit, but it took another year to accept. He distinguishes intuition, which comes from self-knowledge, sleep, food and love, from a fear reaction such as wanting to do a team's job and fire three people.

His parting message is that being a founder is a state of being, an attitude of courage and instinct, while being a CEO is a craft, and both matter. He urges founders to be honest about what they are bad at, like reading a financial statement or running meetings that go nowhere, and to work on the craft alongside the business. In the lightning round he recommends *The Five Dysfunctions of a Team* and *The Untethered Soul*, and lives by the motto "it's one life."