Arram Sabeti Put Lunch on a Spreadsheet - Then Learned to Leave the Table

The first useful thing Arram Sabeti built was not software. It was an arrangement: restaurants, office managers, dietary preferences, delivery times and one founder trying to keep the whole lunch hour from collapsing into cold noodles. The machinery lived in an inbox and a spreadsheet. Eventually the spreadsheet reached 500 columns, which is less a data structure than a cry for help. Still, lunch arrived.

That stubborn, inelegant system became Zerocater, the corporate food-service company Sabeti founded in San Francisco in 2009. By 2015, the company had passed $100 million in cumulative sales. It had begun with roughly $7,000, a list of restaurants and the unsettling realization that an ordinary workplace nuisance could support a serious business. Sabeti did not so much invent office lunch as consent to notice how difficult it was.

His route to that nuisance was wonderfully indirect. At 20, he had become absorbed by technology startups. He sold his car, moved to the Bay Area and treated a stack of Paul Graham essays as an informal curriculum. The timing was poor. He was 21, without a degree or an obvious professional specialty, in a brutal summer for job seekers. He used computers at the Berkeley Public Library to find a cheap studio in El Cerrito and walked into local businesses asking for applications. One said yes: Ben & Jerry’s.

Chapter oneThe apprenticeship hidden inside the chore

The ice-cream era lasted two weeks. A Hacker News listing led Sabeti to interview at Justin.tv, the live-streaming startup that would later produce Twitch. He arrived with the delicate negotiating style of someone who had almost no leverage and a great deal of honesty. He told the founders he planned to quit in a year and start his own company. They declined to give him the advertised job, then invented another position for him.

It was a grab bag of neglected work: testing site releases, handling applicants, doing whatever the young company needed and, critically, ordering lunch. Sabeti worked near Justin Kan and Emmett Shear and watched capable founders operate at close range. He later called the experience an education he could not have received elsewhere. It also showed him that a startup is often a pile of unclaimed tasks with a persuasive story attached.

After a year and a half, the promised company had not appeared. Sabeti gave himself six months. He put a countdown on his desktop and changed his work password to “SixMonthsQuit,” forcing himself to type the deadline repeatedly. A password is a small place to store an existential ultimatum, but the arrangement worked. When another office asked for his carefully assembled restaurant list, Sabeti offered to manage its lunches instead. The office manager was relieved. Relief, in business, is usually worth inspecting.

One customer became two, then three. He asked friends for introductions to office managers. Three months into his six-month countdown, there was enough work to leave Justin.tv. The grand startup he had crossed the country to build turned out to be hiding in the duty nobody else wanted.

Chapter twoAn inbox learns to sell

Sabeti was embarrassed that his technology company contained no technology. The first Zerocater was an inbox, a gigantic spreadsheet and an increasingly practiced sales pitch. Then he got over it. If someone at a party mentioned that a company ordered food, out came his phone. If a salesperson called him, he proposed an exchange: he would hear the pitch if the caller would hear his.

Once, while buying a phone, a Verizon employee asked what he did. Sabeti answered by reflex. Give Zerocater the headcount, the vegetarians, the vegan, the days and the budget; the right food would appear. By the time he got home, San Francisco’s Verizon stores had signed up to feed employees during an iPhone launch. It is difficult to teach that degree of shameless readiness. It is also difficult to grow a company without it.

The difficult parts multiplied. The schedule expanded toward those 500 columns. Billing alone consumed about 20 hours each week. Restaurants varied, customers had preferences, diets differed, deliveries went wrong, and every successful lunch created the obligation to make tomorrow’s feel just as effortless. Zerocater entered Y Combinator’s Winter 2011 batch, but near the end of the program Sabeti’s technical partner left. Investors tend to prefer technical founders and teams of more than one. He was neither. He kept building.

The early constraint became part of the company’s character. Sabeti later argued that money does not convert trivially into success, particularly in a business that looks easy until one tries to run it at scale. In Zerocater’s first 18 months, cash was tight enough that losing one customer could threaten the rent. He also spent more than a year attempting to do everything himself, a decision he later described as a mistake that cost growth.

This is the refreshing complication in his solo-founder story: he survived it without recommending it. When a friend pointed to Sabeti as evidence that a cofounder was unnecessary, Sabeti told him to stop. Much of the difficulty had been needless. Persistence is admirable; arranging to require quite so much of it is not.

Chapter threeWhat the spreadsheet could not measure

By 2015, Zerocater was serving tens of thousands of meals a day and had crossed $100 million in total sales. Those figures give the story scale. A taquería owner gave it consequence. At a party for food vendors, the owner showed Sabeti a video of his family in a hot tub. The presentation was confusing until he explained that Zerocater’s business had enabled the family to take its first vacation in ten years. “My heart grew three sizes,” Sabeti recalled.

The company sat between office appetites and local kitchens, translating aggregated demand into revenue for restaurants that might otherwise have struggled to reach corporate buyers. Its real product was dependable coordination. The food was visible. The confidence that it would arrive, fit the room and avoid repetition was the quieter commodity.

Sabeti’s management preferences followed the same logic. He looked for an “ownership mentality,” people capable of caring about an outcome rather than merely completing an assignment. Years later, writing about hiring, he said he had interviewed roughly 3,000 people during nine years at Zerocater. One exceptionally polished candidate passed every conversation and impressed the entire team, but references exposed a different record. Interviews still mattered, Sabeti concluded, yet a strong account from someone who had actually worked beside a candidate could reveal what a fluent hour in a conference room could not.

In 2018, Sabeti moved from chief executive to chairman of the board. Ali Sabeti became CEO. The founder remained connected to the company while no longer making himself its daily center. For a person who once said his goal was not to be comfortable, the transition suggests a more interesting ambition than permanent control: build something sturdy enough to outgrow your grip.

Chapter fourAfter the half-billion-dollar lunch

On his personal site, Sabeti compresses the corporate chapter into one dryly spectacular line: “Once I made a company that sold half a billion dollars of lunch.” The tense matters. Zerocater continues, and he remains its chairman, but his public self-description now begins elsewhere. He makes strange, beautiful things. He writes. He meditates.

The newer work has a familiar operating style. His newsletter describes itself as “Mysticism without the mist.” Recent essays examine contemplative practice for readers who like rationality, careful definitions and claims that survive contact with skepticism. Where office lunch involved translating taste and logistics into a repeatable system, the writing tries to translate old practices into language a technically minded reader can test. The object has changed from the workplace to the mind. The impatience with fog has not.

He has also invested in startups, including meeting-intelligence company Fathom. That is another kind of remove: helping someone else’s machine begin, without insisting on sitting in the driver’s seat. The young founder who warned Justin.tv that he would soon leave to build his own company has become a chairman, investor and writer whose identity no longer depends on one title.

There is a temptation to make the story symmetrical. First ambition, then contemplation. First the spreadsheet, then the empty page. Life is rarely that tidy, and Sabeti’s own phrase “Assume you can” still connects both halves. The difference is in what follows the assumption. Early on, it meant selling the car, taking the odd job, making the pitch and staying close to the edge. Now it appears alongside making and inquiry, with less need for the company to carry the entire weight of a life.

The best detail remains the spreadsheet. Five hundred columns are absurd. They are also evidence: each column held some fact that the world had insisted mattered. The eventual software was not speculation. It was compressed experience. Arram Sabeti built a company by respecting a tedious problem longer than most people could bear to look at it. Then he performed another difficult trick. He let the company become something he had made, rather than everything he was.