In the winter of 2004, a twenty-three-year-old German immigrant sat in an Ottawa coffee shop, trying to sell snowboards on the internet, and failing — not because nobody wanted snowboards, but because the internet itself seemed to resist the transaction. Every e-commerce platform he tried was, in his later estimation, an insult: clunky payment gateways that took three months to provision, templating systems that produced storefronts ugly enough to repel customers on sight, backend architectures so rigid they seemed designed to punish anyone foolish enough to attempt online retail without a team of engineers. Tobias Lütke had moved across the Atlantic for love, arrived without a work visa, and — having been told by his girlfriend that now was the time to jump — landed on the idea of selling snowboarding equipment from a site called Snowdevil. The store would prove less important than his rage at the software required to build it. He scrapped the existing tools, opened a text editor, and began writing his own.
What emerged was not a snowboard shop. Or rather, it was a snowboard shop that contained, inside its code, the embryo of something else entirely — a platform that would, two decades later, power more than four million merchants in 175 countries, process hundreds of billions of dollars in gross merchandise volume annually, become the second-largest e-commerce platform in North America behind only Amazon, and make its creator one of the wealthiest people in Canada, with a net worth that Forbes pegged at roughly US$12.5 billion by late 2021. The snowboards were incidental. The anger was the point.
This is the paradox at the center of Tobi Lütke's story: a man who never intended to build a company, who dropped out of school at sixteen, who learned more about leadership from StarCraft and Asheron's Call than from any management text, who draws a $1 annual salary, who publicly questioned whether financial analysts should be held accountable for their track records, and who — in April 2025 — issued an internal memo declaring that the reflexive use of artificial intelligence was now a "baseline expectation" for every employee at Shopify, himself included. He is simultaneously the most consequential CEO in Canadian history and one of the least interested in the ceremonial apparatus of corporate power. He wears a newsboy cap and a hoodie on video calls with . He builds things. He ships them. He plays video games with his kids. He cannot quite believe any of this happened.
Part IIThe Playbook
What follows is a distillation of the operating principles that have governed Tobi Lütke's two-decade run building Shopify — drawn from his public statements, interviews, memos, and the decisions that define the company's trajectory. These are not motivational aphorisms. They are engineering specifications for a particular kind of organization.
Table of Contents
1.Fall in love with the problem, not the solution.
2.Build the tool you wish existed.
3.Use the secondary market as a strategic advantage.
4.Install the trust battery as organizational firmware.
5.Classify every process. Kill the third category.
6.Subtract before you add.
7.Make the hundred-year question your product filter.
Arm the rebels, never become the empire.
In Their Own Words
I think the best trick to pull off in life is to find your calling early and hone it into a craft, and then share it in some way.
The act of creating is sort of a conversion of energy...like passion for a field...or emotions such as anger or dissatisfaction. You take those and you channel them into making something.
Statistically, [entrepreneurship is] a very bad way to make money...you might actually lose money.
It is incredibly powerful if you solve the problem you actually have yourself.
At some point in life, you realize that everything around us is made by people who are no smarter or more talented than us.
Something that potential investors must understand: we do not chase revenue as the primary driver of our business. Shopify has been about empowering merchants since it was founded, and we have always prioritized long-term value over short-term revenue opportunities.
Every product in the world, at the end of the day, is simply a reflection of how much the people who created it gave a sh*t.
You don't even need to predict the future that well — if you just live in everyone else's relative future, you're ahead.
I work under the assumption that we have no idea how to build companies yet, and that 50 years from now people will look back at the companies of today and they will seem like the black-and-white footage of the first hockey games.
Your Audience is Your Lifeblood.
Time is energy that you can invest in things, and money is energy that you can invest. Time has significantly more leverage than money in terms of how much energy you get out of time.
I think focus of audience ends up being, amongst all the things that you could have focused on, one of the most significant predictors of the success of a company.
I really encourage people to find a place that sort of culturally aligns with what they're trying to accomplish, and then just go for it.
When we start new interns in our R&D team, we make sure that within their first week they actually make a change to Shopify that impacts our customers.
He was born on July 16, 1981, in Koblenz, a small city at the confluence of the Rhine and Moselle rivers in what was then West Germany. His parents gave him an Amstrad CPC when he was six years old. He played Pac-Man. He played Space Invaders. The games were scarce — this was the late 1980s, and software distribution in provincial Germany was not what it would become — so he did what a certain kind of child does when the games run out: he opened them up. By twelve, he was rewriting the code for his computer games, not to cheat but to understand, and the understanding produced a sensation he would later describe with the precision of someone who has spent decades trying to name it. The computer was, he said, "more interesting than anything else and it was certainly more interesting than school."
This is not the standard biographical detail it appears to be. Lütke has spoken extensively about the role of video games in shaping his thinking about attention, systems, and leadership — arguing that games are misunderstood in the same way that chess would never be misunderstood if a child played it all day. "Everyone always ends up coming into these conversations very apologetically," he told Shane Parrish on The Knowledge Project. "Almost like, 'I as a parent lost control and now I need to course correct.' I'm always frustrated with this because obviously, you can have too much of a good thing, but there's a weird way that people perceive video games in society." He has said, with evident seriousness, that he learned more about resource management, competitive strategy, and the dynamics of multiplayer coordination from StarCraft and other real-time strategy games than from the business literature he later consumed. The games taught him that attention is the scarcest resource, that systems are more powerful than individual actors, and that the meta-game — the game about the game — is where the real decisions live.
He was also, by his own account, dealing with ADHD and dyslexia, conditions that made the traditional German educational system a poor fit. He left school after the tenth grade, at sixteen, and enrolled in a vocational apprenticeship program at the Koblenzer Carl-Benz-School — a Fachinformatiker program, Germany's dual-education track combining classroom instruction with on-the-job coding. He completed the apprenticeship, worked briefly as a programmer at Siemens, found the work stultifying ("the driest of the dry kind of stuff"), and might have drifted indefinitely through the German corporate software ecosystem had he not, in the early 2000s, met a Canadian woman in a virtual world.
The Love Story as Immigration Story
Fiona McKean was the daughter of Canadian diplomats who had grown up in Ottawa and several countries around the world. She met Lütke inside Asheron's Call, a massively multiplayer online role-playing game, where their avatars fought virtual monsters on quests separated by six time zones. He stayed up late in Germany to play alongside her. They chatted. The vibes were good. He flew to Canada for a winter holiday and met her in person at the Whistler ski resort in British Columbia. She was about his height, wore horn-rimmed glasses, had her blonde hair cut short. He was smitten.
McKean — who later described herself as "50% Geek, 50% Nerd" — returned to Ottawa to study for her master's degree at the Norman Paterson School of International Affairs, an elite training ground for Canada's foreign service. Lütke followed her in 2002, at twenty-two, intending to support himself as a remote programmer for a German company. The arrangement worked until it didn't: the German contract dried up, and Lütke, who lacked a Canadian work visa, found himself unable to get a job. There was, however, no law against starting a business.
McKean's intervention was decisive. "Look, I'm super busy because I'm finishing this degree," she told him. "You should find something that makes you super busy. We don't have kids; we don't have a lot of costs. Now's the time to jump. So I'll jump with you."
They married. He moved into her parents' house. And at the McKean dinner table, he met Scott Lake — a friend of Fiona's family, twelve years Lütke's senior, athletic, gregarious, a venture capitalist by temperament who made friends easily. Lake would become Lütke's first business partner. Daniel Weinand, a German designer and friend of Lütke's, would become the third co-founder. The three of them decided, in 2004, to sell snowboarding equipment online.
The founding myth of Shopify is inseparable from this domestic geography: a dinner table in Ottawa, a woman finishing her master's degree, a boyfriend without a work visa, a sport they all loved. It is also inseparable from the specific texture of immigrant ambition — the way that a young man who had dropped out of school at sixteen, who spoke English with a noticeable German accent, who had no network in Canadian tech, found himself building infrastructure for the internet from a borrowed laptop in a coffee shop in a city that was not, by any stretch, a primary talent market.
The Rebellion of the Platform
The conventional founding narrative goes like this: Lütke tried to sell snowboards, found the existing e-commerce tools inadequate, and built his own. But this understates the ambition and the anger. What Lütke actually built, using Ruby on Rails — a web framework he would later contribute to as a core team member — was not merely a better storefront. It was a system designed from first principles around the question: What makes it difficult to start a business on the internet?
He coded alone, mostly in coffee shops, taking no salary. An angel investor in Toronto provided early funding for design work. For the first three months, the platform was free while Lütke finished building it. In June 2006, Shopify launched as a service: a hosted, cloud-based platform that would let anyone — not just people with engineering teams — create and run an online store. The positioning sentence Lütke settled on was deceptively simple: "Shopify is the easiest way to create scalable and beautiful online stores." He later reflected on the discipline this kind of sentence imposed. "You have to have one of those sentences which really is going to guide you through these times where you just have to say no to features."
The early years were precarious. Funding was a constant concern. Shopify had just 300 employees as late as 2013. But the underlying architecture was sound, and the timing — though Lütke would have denied there was any strategic timing — was exquisite. E-commerce was growing. Small businesses wanted online presence. And the alternatives were, in 2006, still terrible.
Shopify's big break in funding came in 2010, when Bessemer Venture Partners led a significant investment. By 2013, more than 50,000 online stores ran on the platform. In 2013, Shopify launched Shopify Payments in partnership with Stripe, allowing merchants to accept payments without a third-party gateway — eliminating the kind of friction that had enraged Lütke a decade earlier when it took him three months to provision a payment gateway for Snowdevil. In 2016, the company launched Shopify Capital, providing cash advances to merchants against future sales. In 2017 came Shop Pay, Shopify's own checkout product. Each layer — storefront, payments, lending, logistics — was an answer to the same originating question, extended one level deeper into the stack of commerce.
We work backwards from what makes it difficult to start these businesses. It's hard to run a small business and be responsible for payroll, for staff, et cetera, even at the best of times.
— Tobi Lütke
The metaphor Lütke settled on was martial: Shopify was "arming the rebels." Amazon, in this framing, was the empire — a centralized marketplace where merchants were interchangeable, disposable, legible only as data points in Jeff Bezos's optimization function. "Amazon's worldview is that merchants don't matter; factories and consumers matter," Lütke told Fortune. "Everything in between is Jeff's opportunity." Shopify's worldview was the inverse: the merchant was the unit of value, and the platform's job was to make the merchant more powerful, not more dependent. It was, in Reid Hoffman's taxonomy, the distinction between a business that captures value and a business that creates it — between Archibald T. Whinnymaker III's lemonade monopoly and Loretta Lovegood's open emporium.
Ottawa, Not San Francisco
One of the most consequential decisions Lütke made was also one of the least discussed: he built Shopify in Ottawa, not in San Francisco or New York or even Toronto. This was partly accident — McKean was from Ottawa, and Lütke had landed there — but it became strategy. He has spoken at length about the difference between building a company in a "primary talent market" (the Bay Area, for tech; Los Angeles, for film) and a "secondary talent market," and his analysis is characteristically precise.
"One of the biggest differences in secondary places is, if I hired someone, the chance that we are gonna still work together in a decade from now is super high," he told Parrish. "That doesn't sound like a big change, but it changes absolutely everything for the company. It means that it's a much better idea to hire for future potential rather than for current skill."
In Silicon Valley, employees leave. The half-life of a good engineer at a hot startup is measured in months. In Ottawa, people stay. This inverts the standard calculus of human capital: if your investment in developing someone doesn't walk out the door, you can invest far more aggressively in development. You hire for aptitude and train for skill. You build a culture of long tenure and deep institutional knowledge. The secondary market, paradoxically, becomes an advantage.
When BlackBerry — then Canada's dominant tech company — announced in September 2013 that it was cutting 4,500 jobs (having already cut more than 7,000 in the two prior years), Lütke saw opportunity. Shopify, then just 300 employees, expected to absorb many of the castoffs. It was a literal transfer of talent from Canada's fading tech champion to its rising one. Today, BlackBerry makes security software and is a shadow of its former self. Shopify employs thousands globally and has, as of late 2021, a market capitalization that made it the most valuable publicly listed company in Canada.
The IPO Letter and the Hundred-Year Question
On May 21, 2015, Shopify went public on both the New York Stock Exchange and the Toronto Stock Exchange, raising $131 million. The IPO was, by all accounts, a success — the stock surged on its first day of trading. But the more revealing document was Lütke's IPO letter, which he addressed not to investors but to future employees and merchants.
"The first Shopify store was our own," it began. He laid out the founding story, the values, the aspirations — but the letter's real subject was time. Lütke wanted Shopify to be the kind of company that people a hundred years hence would not be embarrassed by. He had articulated this ambition as early as 2011, at the Business of Software conference, where he posed the question to an audience of developers: "How can we build businesses that people won't be embarrassed by in 100 years?" The question sounds naive until you consider its implications: it means resisting every short-term optimization that sacrifices long-term trust. It means thinking of the company as infrastructure, not as an asset to be flipped.
This temporal frame — the hundred-year question — shaped decisions that might otherwise have seemed eccentric. Lütke has spoken about playing the "infinite game," borrowing from James Carse's philosophical framework (later popularized by Simon Sinek): the goal is not to win but to keep playing, to remain a viable participant in a game that has no endpoint. In practice, this meant building the platform to be maximally flexible, resisting the temptation to extract rents from merchants, investing in tools that increased merchant independence rather than merchant dependence.
By 2020, Shopify's revenue had reached US$2.9 billion, up 86 percent from the prior year. By the first nine months of 2021, the company had already surpassed that full-year figure, clocking $3.2 billion. The stock, which had traded at $55 on the TSX in December 2016, hit $2,075 at its 52-week high in 2021 — a 3,327 percent increase over five years. Lütke's personal fortune, which Forbes had estimated at US$1.2 billion when he first appeared on the billionaires list in 2018, ballooned to US$9.8 billion by early 2021, making him Canada's second-richest person behind David Thomson of Thomson Reuters. Catherine and John Phillips, early Shopify investors who each owned 1.5 percent of the company, became billionaires too.
The Pandemic as Stress Test
On March 11, 2020, the World Health Organization declared COVID-19 a global pandemic. On the same day, Shopify announced that its roughly 5,000 employees would work from home. Each was given $1,000 to help set up a home office. Lütke and his wife redirected their private charitable foundation — which had focused on decarbonization — to fight COVID-19. In May, Lütke made the arrangement permanent: Shopify would be "digital by default." "Office centricity is over," he tweeted.
The pandemic was, for Shopify, simultaneously a humanitarian crisis and a commercial accelerant. In the six weeks between the start of lockdowns and April 24, 2020, the number of new stores on Shopify rose 62 percent compared with the prior six weeks. Grocers, farmers, florists — businesses that had never considered e-commerce — were suddenly online, using Shopify's tools to offer curbside pickup and local delivery. "There just wasn't any supply online," Lütke told Fortune. "All these kinds of things have been unknotted now."
Lütke described the company's pandemic response with a word that revealed his engineering temperament: they "deleted" their existing plans. The roadmap — the carefully sequenced product launches, the polish, the QA — was scrapped. What mattered was speed. A curbside pickup feature went out the door as what Lütke candidly called "a duct tape system." The principle was explicit: accept less polish in exchange for getting services to merchants who needed them now. It was the startup ethos applied at scale, under duress, and it worked. Shopify's stock rose 150 percent between mid-March and late May 2020. By June, the company had surpassed Royal Bank of Canada to become the most valuable public company in the country.
We can't help on the front line, we're not doctors and nurses. But there is a second wave coming, which is the economic crisis, which is going to derive from the humanitarian crisis. We can't solve it, but we can hopefully help.
— Tobi Lütke
But the pandemic also created a distortion that would haunt the company. The massive pull-forward of e-commerce demand made Shopify's growth look like a new baseline when it was, in part, a temporary surge. When the world reopened and shoppers returned to physical stores, Shopify's growth decelerated. The stock, which had peaked near $2,140 on the TSX in November 2021, cratered. By May 2022, it had fallen to around $455 — below its pre-pandemic highs — wiping out all pandemic gains. Shopify reported a net loss of US$1.5 billion in Q1 2022, compared with net income of US$1.3 billion a year earlier. Investors who had bid the stock up on the assumption that the pandemic had permanently rewired consumer behavior discovered that reality was more recursive than that.
The Trust Battery and the War on Process
If there is a single concept that captures Lütke's management philosophy, it is the "trust battery" — a mental model he developed and has described at length on podcasts and in interviews. The idea is simple, the implications vast: trust between any two people in an organization is not binary (I trust you / I don't trust you) but a gradient, like a battery charge. When two people start working together in a curated context — they both got hired, they both passed the gauntlet — the trust battery is at roughly 50 percent. Every interaction charges or drains it. A meeting where someone follows through charges it. A missed commitment drains it. Over time, the battery reaches a level that determines how much latitude, autonomy, and benefit of the doubt one person extends to the other.
The model is useful precisely because it makes an invisible social dynamic legible and discussable. If a team is dysfunctional, you can ask: what is the trust battery level between these two people? You can have a conversation about it without it becoming personal. It is, in Lütke's framing, a way of reducing emotional noise in organizational decisions — turning a feeling ("I don't trust her") into a diagnostic ("the trust battery is at 20 percent because of these specific interactions"). It is also, characteristically, a systems-level intervention: rather than telling people to trust each other, you build a shared language for tracking trust as a dynamic variable.
His war on process is equally systematic. Lütke has articulated a taxonomy of processes that is almost comically precise: there are processes that make previously impossible things possible (good), processes that make previously possible things significantly simpler (also good), and then everything else. "I bet you 99.9% of all process that exists in corporate America is the third category," he has said, "which is actually just telling people to behave slightly different from what common sense tells them to do." The implication is that most organizational process is not enabling but constraining — a tax on intelligence, a substitute for trust, a way of managing the lowest common denominator that penalizes everyone else.
This is not libertarian anarchism. Lütke is precise about the distinction: the first two categories of process are essential. Infrastructure is process. Payment systems are process. The platform itself is, in some sense, a very large process for making commerce possible. What he opposes is the third category — the creeping bureaucratization that large organizations accumulate like sediment, each layer a response to some past failure, each layer making the next failure slightly more likely because it has drained the system of the flexibility to respond.
"The best thing founders can do is subtraction," he told Parrish. "It's much, much, much easier to add things than it is to remove things. Adding things is a lot more expensive than removing things. However, it requires some measure of bravery and risk-taking. If you say no to a thing, you say no to one thing. If you say yes to a thing, you actually say no to every other thing you could have done during this period of time."
The Harley Finkelstein Question
Every founder-CEO needs a counterpart, and Lütke's is Harley Finkelstein — Shopify's president since 2020, a ten-year veteran of the company, and in many ways Lütke's temperamental opposite. Where Lütke is quiet, technical, and allergic to self-promotion, Finkelstein is voluble, commercial, and relentlessly public-facing. A lawyer by training who originally joined Shopify as a merchant himself, Finkelstein was named to Fortune's 40 Under 40 in 2021 at age thirty-seven. He has been instrumental in commercializing Shopify's tools, shaping its enterprise strategy, and guiding its expansion into fulfillment and payments.
Finkelstein's role illuminates something important about Lütke's leadership style: the CEO does not do everything. He does not want to do everything. Lütke is, by his own admission, a programmer first and a CEO second — the role he occupies because the company he built required someone to occupy it, not because he aspired to corporate leadership. He took over as CEO in 2008, four years after co-founding the company, and has described the position as a continuous exercise in learning to do things he was not trained to do. The programming, the systems thinking, the product intuition — these came naturally. The people management, the investor relations, the public performance of leadership — these required deliberate acquisition.
Scott Lake, Shopify's co-founder, put it simply in an interview with the Globe and Mail: "He is the brain trust. Tobi is Shopify." The statement is both a compliment and a constraint. If Tobi is Shopify, then Shopify's trajectory is inseparable from Tobi's intellectual evolution — his reading list (Thinking in Systems by Donella Meadows, High Output Management by Andy Grove, A Guide to the Good Life by William B. Irvine, The Courage to Be Disliked), his gaming habits, his shifting views on work and attention, his capacity to absorb new paradigms and embed them in the organization's operating system.
The AI Memo and the Death of Optionality
On April 7, 2025, Lütke released an internal memo — originally leaked, then published on X by Lütke himself — that declared reflexive AI usage a "baseline expectation" for every employee at Shopify, from the most junior ranks to the executive team. The memo was roughly 1,100 words long, direct, and unapologetic. Six expectations were outlined:
AI usage would be a fundamental requirement. "Frankly, I don't think it's feasible to opt out of learning the skill of applying AI in your craft," Lütke wrote. "Stagnation is almost certain, and stagnation is slow-motion failure. If you're not climbing, you're sliding."
AI must be part of every prototype phase. Before building anything, teams would use AI to explore, learn, and create artifacts for others to evaluate.
AI usage questions would be added to performance and peer review questionnaires. The company would measure not just output but AI fluency.
Learning would be self-directed but shared. Slack channels, internal tools, and monthly business reviews would be dedicated to AI integration.
Before requesting headcount, teams must demonstrate why AI cannot do the work. "What would this area look like if autonomous AI agents were already part of the team?" Lütke asked.
Everyone means everyone. Including Lütke and the executive team.
The memo landed in the tech industry like a grenade. Commentators compared it to Brian Armstrong's 2020 "mission-focused company" memo at Coinbase and Mark Zuckerberg's "Year of Efficiency" — documents that reshape corporate culture across the industry by articulating a new operating principle with sufficient clarity and authority to become memetic. Jeff Morris Jr., writing in his New Internet newsletter, noted that "yesterday's memo will shape how founders approach using more AI and hiring less people in 2025."
The most striking line was buried in the middle: "My sense is that a lot of people give up after writing a prompt and not getting the ideal thing back immediately." This is Lütke the programmer speaking — the person who understands that tools require practice, that fluency is a skill, that the gap between a novice prompt and an expert prompt is as wide as the gap between novice code and expert code. It is also Lütke the systems thinker, who recognizes that the organizational adoption of any new technology follows a learning curve, and that the curve must be managed, not merely announced.
But the memo also contained an implicit threat: if you cannot demonstrate that your team needs more humans, you will not get them. In an era of mass tech layoffs and anxious discourse about AI displacement, this was not a comfortable position. Shopify spokesperson Jackie Warren, when asked about guardrails to ensure AI-generated work is accurate and unbiased, did not respond to the inquiry.
The Beautiful Unsolvable Problem
In a 2025 conversation with John Collison — co-founder of Stripe, Shopify's longtime payments partner, a man whose own immigrant biography (Irish, arrived in the United States as a teenager, built a payments infrastructure company worth tens of billions) rhymes interestingly with Lütke's — the Shopify CEO offered what might be his most distilled articulation of what drives him. "The best gift in life is finding a beautiful problem that you can never solve," he said. "And even if you accidentally solve it, if you're so unfortunate to solve it, hopefully it has plenty of enlightened problem children."
The problem, for Lütke, is internet commerce. Twenty years after Snowdevil, he remains captivated by it. The Shopify codebase is now 20 million lines of Ruby, plus another 8-to-10 million lines of TypeScript for the admin interface — one of the largest TypeScript applications in existence, and definitively the largest Ruby on Rails application ever built. Shopify processes peak transaction volumes that would qualify it as critical infrastructure in any national economy. The company has expanded from small merchants to large enterprises — Mattel, Steve Madden, Skims — while maintaining the small-merchant ethos that animated its founding. Evercore ISI analyst Mark Mahaney, in upgrading the stock in June 2024, pegged Shopify's total addressable market at $850 billion, encompassing not just online commerce but physical retail point-of-sale and enterprise e-commerce.
And still, Lütke describes the work in terms of fascination, not dominion. "Some people fall in love with solutions," he told Collison. "Some people fall in love with problems." He identifies himself, unambiguously, as the latter. The distinction matters because it determines organizational behavior at scale: a company in love with its solutions becomes defensive, protective, resistant to change. A company in love with its problem remains curious, adaptive, willing to discard solutions that no longer serve.
This is the posture that led Shopify to integrate Amazon's "Buy with Prime" feature into its merchants' stores — an alliance with the company Lütke had spent years positioning as the enemy. It is the posture that led him to embrace AI not as a threat to his workforce but as an extension of the programmer's toolkit, the same way Ruby on Rails had been an extension of the web developer's toolkit two decades earlier. It is the posture that allows him to describe consumerism — a word typically deployed as an indictment — as a problem of product quality: "They throw away things because they hate the things they have. The thing that solves consumerism is quality product."
This is either profound or glib, depending on your politics. But it is consistent. Lütke has always believed that the primary failure mode of commerce is friction — the gap between a person who wants to buy something and a person who wants to sell it. Every layer of Shopify's expanding stack is an attack on some species of friction. The payments. The fulfillment. The capital. The checkout. The AI that can write your product descriptions and answer your customer's questions and maybe, eventually, negotiate on your behalf with an AI representing the buyer. The game has no endpoint. The problem is beautiful. It cannot be solved.
The Evolving Hours
There is one more tension worth examining, because it reveals the distance between ideology and lived experience in the life of a founder-CEO. In 2019, Lütke posted on Twitter that he rarely worked more than 40 hours a week, slept eight hours a night, and had never pulled an all-nighter while running a company then worth $125 billion. The tweet was widely celebrated — a corrective to the hustle-porn mythology of Silicon Valley, proof that you could build something enormous without destroying yourself.
By March 2025, Lütke had deleted the tweet. In its place, a clarification: "Yea, but this is commonly misunderstood. I'm at home for dinner, but I work at least 10 or so hours a day and a lot of the weekend." The shift was not a contradiction, exactly — he is still home for dinner, still sleeping — but it was a revision, and a telling one. The man who had positioned himself as an emblem of work-life balance in tech was now, in a changed industry, working significantly longer hours and saying so publicly.
What happened between 2019 and 2025? The pandemic. The stock collapse. The AI revolution. The intensifying competition with Amazon. The move from a company that mostly served small merchants to one that also courts Mattel and Steve Madden. The weight of being, as one reporter put it, "Canada's most prominent tech company" during a period in which Canada's tech sector was struggling to compete globally. The recognition, perhaps, that the hundred-year game requires more hours than the forty-hour narrative allowed.
Lütke did not frame it as a betrayal. He framed it as evolution. "I don't know if I did a very good job of it, but I certainly did it," he told Parrish about leading through the pandemic's volatility. "I certainly had to deal with volatility, that's true." The honesty is characteristic — the willingness to say I don't know if I was good at this in a public forum, without false modesty or false confidence. It is the voice of a programmer debugging his own performance, noting the failures alongside the successes, resisting the narrative smoothing that most CEOs perform instinctively.
He received the Meritorious Service Cross from the Governor General of Canada in 2018. He serves on the advisory board of Canada Learning Code. He challenged the CRA when they demanded six years of records from 121,000 Canadian merchants, calling it "low-key overreach." He called out the Canadian government's practice of subsidizing foreign tech companies to create domestic jobs, calling the subsidies "toxic" to the domestic tech ecosystem. He competes in amateur car racing — LMP2-class endurance events — which is either a rich man's hobby or a metaphor for his approach to leadership: sustained performance under extreme conditions, the management of systems at high speed, the acceptance of risk as the price of being in the race.
His wife, Fiona McKean — the woman whose Asheron's Call avatar fought virtual monsters alongside his, whose dinner table produced his first business partner, whose injunction to "find something that makes you super busy" launched Shopify — remains largely out of public view. They have three children. They relocated from Ottawa to Toronto. The headquarters, such as it is, followed: Shopify is now a fully remote company, "digital by default," a phrase that sounds corporate until you remember that it was coined by a man who met his wife inside a video game and built his first company from a coffee shop laptop because he didn't have a work visa.
In his conversation with Collison, Lütke described companies as "a form of technology" — systems that evolve, that encode decisions in their structure, that shape the behavior of the people within them as surely as any software architecture shapes the behavior of its users. If that is true, then Shopify is Lütke's most complex program: millions of lines of code, thousands of employees, millions of merchants, hundreds of billions in commerce flowing through a system designed by a man who started by trying to sell snowboards and discovered that the real problem was the system itself.
The snowboards are long gone. The problem remains. He is still writing code.
8.
9.Treat AI adoption as a skill curve, not a switch.
10.Delete the plan when the context changes.
11.Learn leadership from systems, not from management theory.
12.Let your work hours evolve honestly.
Principle 1
Fall in love with the problem, not the solution.
Lütke has articulated this distinction with rare precision: "Some people fall in love with solutions. Some people fall in love with problems." The difference is organizational, not just philosophical. Solution-lovers become defensive — they protect what they've built, resist pivots, confuse their product with their purpose. Problem-lovers remain adaptive — they discard tools that no longer serve, embrace competitors' innovations when those innovations help their merchants, and maintain the curiosity that keeps a twenty-year-old company from ossifying.
Shopify's willingness to integrate Amazon's Buy with Prime feature is the proof case. A solution-lover would have seen this as capitulation to the enemy. A problem-lover sees it as another way to reduce friction for the merchant, which is the actual problem. Similarly, Lütke's embrace of AI — despite its potential to reduce headcount at his own company — reflects the same logic. The problem is making commerce easier. AI makes commerce easier. Therefore, AI is not a threat but a tool.
The practical implication is that your product roadmap should be organized around problems, not features. Features are solutions. Problems persist. The company that orients its strategy around persistent problems — and measures itself by how much friction it has removed — has a navigational instrument that works regardless of which specific technologies rise or fall.
Tactic: Define the single persistent problem your company exists to solve, and evaluate every product decision, partnership, and technology adoption against whether it reduces friction for the person experiencing that problem.
Principle 2
Build the tool you wish existed.
Shopify was born not from market research but from personal frustration. Lütke tried to sell snowboards, found every existing e-commerce tool inadequate, and built his own out of spite. ("I decided to do something about it just to spite them," he later said.) This is the classic founder-as-first-user origin story, and it matters because it produces a specific kind of product intuition: you know what good looks like because you have felt what bad feels like.
The danger, of course, is that your needs are not representative. Lütke's advantage was that his frustrations — clunky payment gateways, ugly templates, rigid backends — were universal among small merchants. He was not an edge case. He was an extreme articulation of a common pain. The lesson is not merely "scratch your own itch" but "scratch your own itch and verify that the itch is widespread."
The deeper lesson is about the emotional fuel of founding. Lütke was not motivated by a market opportunity but by an aesthetic and functional objection to the existing world. He found it unacceptable that it was so hard to sell things on the internet, and that objection — sustained over two decades — has proven more durable than any financial incentive.
Tactic: When evaluating whether to build a product, ask not "Is there a market?" but "Am I personally angry that this doesn't exist?" — and then verify that your anger is shared by a large population of potential users.
Principle 3
Use the secondary market as a strategic advantage.
The conventional wisdom in tech is that you must be in San Francisco (or, increasingly, New York or Miami) to build a world-class company. Lütke built Shopify in Ottawa — a government town, not a tech hub — and turned its supposed disadvantage into an asymmetric strength.
In secondary markets, employees stay. If your talent retention is measured in decades rather than months, the calculus of human capital investment inverts: it becomes rational to hire for potential rather than current skill, to invest heavily in training, to build deep institutional knowledge that would be uneconomical in a market where engineers leave every eighteen months. This is the insight that allowed Shopify to absorb BlackBerry's laid-off engineers in 2013 — not as stop-gaps but as long-term investments.
📍
Primary vs. Secondary Talent Markets
How location changes the human capital equation
Dimension
Primary market (SF, NYC)
Secondary market (Ottawa)
Avg. tenure
18–24 months
5–10+ years
Hiring strategy
Hire for current skill
Hire for future potential
Training ROI
Low (talent leaves)
High (talent stays)
Culture
Transactional, market-rate
Deep, institutional
Risk
Shallower institutional memory
Smaller talent pool
Tactic: If you are building in a secondary market, reframe your location as a talent-retention advantage and restructure your hiring process to evaluate learning velocity and long-term potential over current-skill match.
Principle 4
Install the trust battery as organizational firmware.
Lütke's "trust battery" model is not a metaphor — it is an operational tool. By framing interpersonal trust as a measurable gradient (0–100%) rather than a binary state, he gives organizations a shared language for diagnosing dysfunction without making it personal. A team is struggling not because someone is "untrustworthy" but because the trust battery between two specific people has been drained by specific interactions. This is debuggable. This is fixable.
The model also has a predictive function: it explains why new hires need ramp time (the battery starts at 50%), why small failures compound (each drains the battery), and why some teams perform far above the sum of their parts (mutual trust batteries near 100% enable risk-taking and speed). It explains, too, why organizational restructuring is so disruptive: it resets trust batteries across the company.
For Lütke, the trust battery is also a filter for organizational decision-making. He has described the fork that recurs in every important decision — "do we do what's right or do we do what's easy?" — as a function of collective trust. High-trust organizations can do what's right because the personal cost of hard decisions is buffered by the confidence that others will extend benefit of the doubt. Low-trust organizations default to what's easy because nobody wants to spend political capital.
Tactic: Make trust visible by introducing a shared vocabulary (e.g., the trust battery) that allows team members to discuss relationship quality quantitatively, without blame, as a diagnostic input to organizational design.
Principle 5
Classify every process. Kill the third category.
Lütke's process taxonomy is one of his most actionable ideas. Category one: processes that make impossible things possible. Category two: processes that make possible things significantly simpler. Category three: everything else — which he estimates accounts for 99.9 percent of all process in corporate America.
The third category is insidious because it masquerades as the first two. A weekly status meeting feels like it's making something possible (visibility into team progress) but is often just telling people to report information that could be gathered asynchronously. An approval chain feels like it's simplifying decision-making but is often just distributing accountability so that no individual bears risk. The test is simple: does this process enable something that couldn't happen without it, or does it merely constrain behavior?
Lütke's insight is that the third category is not just wasteful — it is actively harmful. It replaces judgment with compliance, trust with surveillance, and agility with bureaucratic inertia. For a technology company, where speed of iteration is a competitive advantage, the third category is a tax on innovation.
Tactic: Audit your organization's processes quarterly by applying Lütke's three-part test: does this make something impossible possible, does it make something significantly simpler, or is it merely constraining behavior? Eliminate the third category ruthlessly.
Principle 6
Subtract before you add.
"If you say no to a thing, you say no to one thing. If you say yes to a thing, you actually say no to every other thing you could have done during this period of time." This asymmetry — the hidden cost of yes — is the core of Lütke's approach to product development, organizational design, and personal time management.
Subtraction is harder than addition because it requires bravery: saying no means accepting the possibility that you are wrong, that the thing you are declining is actually the thing that would have changed everything. Addition feels safe because you haven't foreclosed any option. But the aggregate cost of addition is organizational bloat, feature creep, and the slow erosion of focus. A company that says yes to everything ends up building nothing well.
Lütke practices this at the product level (Shopify's guiding sentence — "the easiest way to create scalable and beautiful online stores" — is a subtraction machine, a filter for saying no) and at the organizational level (his war on third-category process is a form of continuous subtraction). The discipline is the same: identify what should not exist and remove it.
Tactic: Before adding any new feature, hire, process, or commitment, explicitly name the things you will be saying no to as a result — and assess whether the aggregate cost of those forgone options exceeds the value of the addition.
Principle 7
Make the hundred-year question your product filter.
Most companies optimize for the quarter. Some optimize for the year. Lütke, since at least 2011, has asked: would people in a hundred years be embarrassed by this decision? The question is deliberately absurd in its time horizon — no one can know what will embarrass people in 2125 — but its absurdity is the point. It forces you out of local optimization and into a mode of reasoning that prioritizes trust, durability, and alignment with human interests over extraction, lock-in, and short-term revenue maximization.
Practically, this means building a platform that increases merchant independence rather than merchant dependence. It means resisting the temptation to become a marketplace (which would put Shopify in competition with its own merchants). It means investing in infrastructure that the company may never monetize directly but that makes the ecosystem healthier.
The hundred-year question is also a recruiting filter: it attracts people who are motivated by craft and mission rather than by compensation and prestige. These people tend to stay longer, contribute more, and care more about the quality of what they build — which, in a secondary talent market like Ottawa, compounds into an enormous advantage over time.
Tactic: Before any strategic decision, ask: "Would someone examining this decision in 100 years find it admirable or embarrassing?" — and use the answer as a tiebreaker when short-term incentives conflict with long-term trust.
Principle 8
Arm the rebels, never become the empire.
Shopify's positioning as "the rebel army to Amazon's web empire" is not just marketing — it is a structural constraint on strategy. If you are arming the rebels, you cannot also be the empire. You cannot compete with your merchants. You cannot extract monopoly rents. You cannot use your data to build competing products. The moment you do, you become Amazon, and your merchants — who chose you precisely because you were not Amazon — will leave.
This constraint has real costs. Shopify does not take a cut of merchant revenue beyond its subscription and transaction fees. It does not operate a marketplace that ranks merchants against each other. It does not use merchant data to launch competing products. Every one of these decisions forecloses a lucrative business model that Amazon exploits ruthlessly. The trade-off is trust: millions of merchants trust Shopify because Shopify has never betrayed them, and that trust is the foundation of the company's $120 billion market cap.
The strategic implication is that platform businesses face a fundamental choice: be the marketplace or be the infrastructure. The marketplace captures more value per transaction. The infrastructure captures more trust per merchant. Over a long enough time horizon — say, a hundred years — trust compounds.
Tactic: Define clearly whether your company is the marketplace or the infrastructure. If infrastructure, codify the boundaries you will never cross (competing with customers, exploiting their data, forcing lock-in) and communicate those boundaries as commitments.
Principle 9
Treat AI adoption as a skill curve, not a switch.
Lütke's AI memo is remarkable not for its ambition but for its granularity. He does not simply declare that Shopify will use AI. He specifies how AI will be adopted: through prototype phases, peer review, shared learning, and demonstrated necessity before hiring. He identifies the specific failure mode — "a lot of people give up after writing a prompt and not getting the ideal thing back immediately" — and prescribes the specific remedy: practice, peer feedback, and organizational patience.
This reflects the programmer's understanding that tools require fluency. The gap between a novice user of AI and an expert user is vast, and most organizations underinvest in traversing it. They buy licenses, announce mandates, and then wonder why adoption stalls. Lütke's insight is that AI fluency is a skill — like coding, or writing, or managing — that must be developed through practice, feedback, and iteration.
The most radical element of the memo is the headcount constraint: teams must demonstrate why AI cannot do the work before requesting new hires. This reverses the default assumption (we need more people) and replaces it with a burden of proof (we need more people because AI isn't sufficient for this specific task). It is subtraction applied to hiring.
Tactic: Treat AI adoption as a training program, not a mandate — with dedicated practice time, shared prompt libraries, peer review of AI fluency, and a demonstrated-necessity standard for new hires.
Principle 10
Delete the plan when the context changes.
When the pandemic hit, Lütke's response was not to adjust the roadmap but to "delete" it. The word is telling. Not "revise." Not "reprioritize." Delete. The plan no longer described reality, so the plan no longer existed. What replaced it was a set of principles (serve merchants, reduce friction, move fast) applied to a new context (lockdowns, surging online demand, desperate small businesses).
The curbside pickup feature that Shopify shipped during the pandemic was, by Lütke's own admission, "a duct tape system." It was not beautiful. It was not polished. It worked. And it worked because the company was willing to accept a lower quality bar in exchange for a faster response time — a trade-off that is obvious in theory but requires enormous organizational courage in practice, because shipping imperfect work means accepting the possibility of public failure.
The deeper principle is that plans are models, and models are only as good as their assumptions. When the assumptions change — and in a pandemic, they changed overnight — the model is not merely outdated but actively harmful, because it directs resources toward objectives that no longer matter. The ability to delete the plan and re-derive from first principles is, for Lütke, the core competency of a founder-led organization.
Tactic: Build a culture where plans are understood as hypotheses, not commitments — and where the willingness to abandon a plan when context changes is rewarded, not punished.
Principle 11
Learn leadership from systems, not from management theory.
Lütke has been public about the fact that he learned more about leadership from video games — specifically, real-time strategy games like StarCraft — than from business books. This sounds eccentric until you consider what real-time strategy games actually teach: resource allocation under uncertainty, competitive positioning, multi-threaded attention management, the importance of scouting (gathering information before committing resources), and the distinction between tactical execution and strategic positioning.
He has also described Thinking in Systems by Donella Meadows and High Output Management by Andy Grove as formative texts — neither of which is a management book in the conventional sense. Meadows is a systems theorist; Grove's book reads more like an engineering manual for organizational design than a leadership guide. The common thread is that Lütke understands organizations as systems — with feedback loops, leverage points, bottlenecks, and emergent properties — rather than as hierarchies to be managed through authority.
This systems orientation explains his process taxonomy, his trust battery model, his preference for subtraction over addition, and his willingness to redesign organizational structures when they stop producing the desired outputs. It also explains his discomfort with conventional CEO behavior: the speeches, the investor calls, the public performance of leadership. These are management rituals, not systems interventions.
Tactic: Study systems theory and game design alongside management literature — the dynamics of feedback loops, resource allocation, and competitive strategy in games may teach you more about organizational leadership than any Harvard Business Review article.
Principle 12
Let your work hours evolve honestly.
The arc from Lütke's 2019 tweet ("I've never worked through a night… I need 8ish hours of sleep a night") to his 2025 clarification ("I work at least 10 or so hours a day and a lot of the weekend") is not hypocrisy. It is honesty about the fact that the demands of leadership change, and that intellectual honesty requires updating your public statements when your private reality shifts.
The temptation, for a public figure, is to maintain the narrative. The 40-hour-week tweet was enormously popular. It positioned Lütke as the anti-hustle-culture CEO, a corrective to the performative workaholism of Silicon Valley. Deleting it and replacing it with a higher number was a kind of vulnerability — an admission that the earlier framing was, if not false, at least incomplete.
The lesson is not that founders should work 70-hour weeks. It is that founders should be honest about what the work actually requires at any given phase, and that the requirements will change. A company in hypergrowth during a pandemic requires more from its CEO than a company in steady state. A company navigating an AI revolution requires more from everyone than a company iterating on established products. The honesty is the thing.
Tactic: Periodically audit and publicly update your actual working patterns rather than performing a fixed narrative about work-life balance — your team will trust your honesty more than your consistency.
Part IIIQuotes / Maxims
In their words
I got so fascinated with my computer; it was more interesting than anything else and it was certainly more interesting than school.
— Tobi Lütke
Amazon's worldview is that merchants don't matter; factories and consumers matter. Everything in between is Jeff's opportunity.
— Tobi Lütke
The best gift in life is finding a beautiful problem that you can never solve. And even if you accidentally solve it, if you're so unfortunate to solve it, hopefully it has plenty of enlightened problem children.
— Tobi Lütke
Frankly, I don't think it's feasible to opt out of learning the skill of applying AI in your craft; you are welcome to try, but I want to be honest I cannot see this working out today, and definitely not tomorrow. Stagnation is almost certain, and stagnation is slow-motion failure. If you're not climbing, you're sliding.
— Tobi Lütke, AI memo, April 2025
It's much, much, much easier to add things than it is to remove things. Adding things is a lot more expensive than removing things. However, it requires some measure of bravery and risk-taking. If you say no to a thing, you say no to one thing. If you say yes to a thing, you actually say no to every other thing you could have done during this period of time.
— Tobi Lütke
Maxims
Spite is a valid founding emotion. Lütke built Shopify because the existing tools offended him. Moral outrage at bad systems can sustain a company longer than any business plan.
The guiding sentence is a subtraction machine. A single sentence that describes your product's purpose ("the easiest way to create scalable and beautiful online stores") is not marketing copy — it is a decision framework for saying no.
Hire for the decade, not the quarter. In markets where employees stay, invest in potential over current skill. The compounding returns on human development are enormous if the person doesn't leave.
Trust is a battery, not a switch. Make interpersonal trust legible, quantifiable, and discussable — and you transform organizational dysfunction from an emotional crisis into a debugging exercise.
99.9 percent of corporate process is the third category. Most process does not enable the impossible or simplify the possible — it merely constrains behavior. Classify ruthlessly and eliminate the rest.
The platform must never compete with its merchants. If you arm the rebels, you cannot also be the empire. The moment you extract value from the people who depend on you, trust collapses and cannot be rebuilt.
Delete the plan, keep the principles. Plans are models. When assumptions change, models become liabilities. The ability to re-derive from first principles in real time is the core competency of a founder-led organization.
AI fluency is a skill, not a toggle. Most people abandon AI tools after one bad prompt. Treat adoption as a training curve — with practice, feedback loops, and demonstrated necessity as the standard for hiring.
Play the infinite game. The goal is not to win but to keep playing — to remain a viable participant in a game that has no endpoint. Optimize for durability over dominance.
Let the work hours be honest. Leadership demands change. The CEO who pretends otherwise is performing, not leading. Update your narrative when your reality shifts.
