The Lego Inside Notion: Ivan Zhao on Lost Years, Small Teams, and Horizontal Bets

Ivan Zhao describes Notion’s first three to four years as the lost years because the team kept trying to turn a big vision into a product. The first version was a developer tool meant to let more people create software, but after a couple of years the team learned that most people did not care about creating software to optimize their work. The realization was to hide the vision—everybody can create their own software—inside a form factor people already used every day: productivity software. That pivot took roughly two years to reach.

Ivan calls this strategy sugar-coated broccoli: people do not want to eat the broccoli, but they like sugar, so you give them the sugar and hide the broccoli inside. He frames the same tension as building a product or business—users and revenue—versus building something you want the world to have, which draws on taste, aesthetics, and values. Too much of yourself and there are no users; too much for business and you build a commodity. Later he adds that tools are extensions of us, and once we shape them and bring them into the world, they can come back to shape us.

Ivan grew up in Urumqi, which he notes is a “small town” of about 4 million people in northwest China, and moved to Beijing with his mother. To get into a good school in the capital he needed to win a competition, so he chose the programming Olympiad because he was already into computer games and could spend all day on computers. He placed second in Beijing, which helped him get into a strong school and into programming.

He later moved to Canada at 16, did not study computer science in college because he already knew how to code, and spent time on video games, art, and science. By graduation, most of his friends were artists who needed websites and portfolios, and as the only nerd in that circle he built three or four sites for them. That experience showed him that most people did not know how to create with computing media, which became the seed for a product like Notion.

Reading Douglas Engelbart’s “Augmenting Human Intellect” was a turning point: Ivan saw that giving people computing power to augment their problem-solving was the highest-leverage thing a software maker could do. That led him to the first generation of computing pioneers—Engelbart, Alan Kay, Ted Nelson—and to Xerox PARC’s Smalltalk system, where there was no separation between users and applications and everything was malleable. He tells the story of Steve Jobs seeing the Alto’s graphical interface but not the underlying object environment, which he argues set personal computing on an application-centric path rather than a malleable-tool path.

Lenny notes that Notion started in 2013 and that Ivan calls the first three to four years the lost years, a contrast to companies that reach large revenue quickly. Ivan explains that the vision was clear but the product shape was not, and the first developer-tool version did not resonate with friends, investors, or most users. People wake up with reports due and jobs to finish; they do not care about creating software to optimize their work. The team therefore hid the vision inside productivity software, but that took another two-plus years to understand.

Notion also threw away its code and rebuilt repeatedly. After deciding to build productivity software, Ivan realized the product sat on the wrong technical foundation: the team had bet on Google’s Web Components, which was new and unstable, and they could not tell whether bugs came from their own code or the underlying libraries. They restarted the company and rebuilt on a more orthodox foundation, which Ivan describes as necessary to avoid running out of time. He calls the result sugar-coated broccoli: a productivity suite that hides no-code developer power so users can create almost any productivity software inside Notion.

Ivan kept the company alive with help from his mother, who helped him register the company to move to the US and later loaned him money to bridge a cash gap. The deeper motivation was that the thing they wanted—Lego for software—did not quite exist at mass-market usability, and both Ivan and his co-founder Simon had grown up with Legos as the toy they wanted. During one rebuild they laid off the five-person team down to Ivan and Simon, subleased their apartment and office, and went to Japan to change the scenery while coding every day. Ivan remembers it as some of the happiest moments, with the pair coding for long stretches, eating, and returning to work.

For founders stuck on an idea, Ivan emphasizes that you need to believe the thing must exist and that you should not be afraid to reset. He distinguishes building a product or business—where the market is the arena and you optimize the scorecard—from building something you want the world to have, which draws on your taste, aesthetic, and values. The two require different energy, and Ivan finds the value-driven version more lasting and fulfilling. The practical balance is to avoid building only for yourself, which leaves no users, and avoid building only for business, which produces a commodity.

Ivan also gives a more technical reason not to fear a reset: better abstractions can recover ground faster than people expect. Humans tend to think linearly, but a smaller, cleaner kernel—like Smalltalk’s tiny Lisp-inspired code base—can compound and unlock complex behavior. If you find the right abstraction, you can catch up to what you built before much more quickly than you think. That is why resetting is not merely self-help; it can be a way to create progress through a better system.

Product-market fit did not arrive as a binary milestone for Notion. Ivan says the high was never too high and the low never too low, and the ramp was gradual: people started caring, paying, and reaching out. A memorable signal came when dog treats arrived at the office from an investor who had found the address and followed the company on Twitter. Ivan later did many investor meetings until Shana Fisher told him he clearly did not need money and was seeking external validation; he went back to hardcore building.

Lenny notes that Notion is profitable, has kept most of the money it raised in the bank, did not hire a salesperson until it passed $10 million ARR, and hired its first PM at around 50 people. Ivan explains that staying small was not a recent fashion choice; it came from the founders’ ability to cover design, marketing, storytelling, and sales themselves. When a few people can do a lot, the company naturally stays smaller.

The bigger constraint is internal communication: getting many minds aligned is hard, and Ivan prefers to solve coordination with systems and tools where possible. Notion itself is a meta-tool for building tools, and the company runs much of its operation on Notion. That helps keep headcount low and profitability high, which puts Notion on a positive treadmill where it can focus on building rather than fundraising every 18 to 24 months. Internally the company tracks talent density and revenue per employee rather than headcount.

Ivan’s advice for leaders who want to stay lean is to understand that abstractions and systems can beat brute force. He uses the metaphor of Notion as a small bus: the smaller the bus, the easier it is to turn, accelerate, and maneuver. As a leader, you choose who sits around you, and those seatmates determine how fast the bus moves and what the work experience feels like. That metaphor helps the company optimize for keeping the bus lean.

Ivan is deliberate about the office because he wants it to feel like an artist’s studio or a home rather than a hospital. Pragmatically, that means avoiding harsh top lighting, using warmer wall colors and friendlier floors, and choosing home furniture so people spend more time there and feel more creative. Aesthetically, he says ugly chairs and walls hurt his eyes the way bad chairs hurt backs, so avoiding visual pain is part of the value system.

Notion names conference rooms after timeless tools—the iPhone, the original Macintosh, Lamy 2000 pens, Toshiba rice cookers, Sony transistor radios—because those objects changed how many people lived and lasted for decades. Ivan wants the team to think about creating software with that kind of longevity, not something that turns over every six or 12 months. He quotes Steve Jobs on the tremendous craftsmanship between a great idea and a great product, and defines craft as applying taste and values to technical know-how to make clever trade-offs. His wife calls him a wood cabinet builder, and he likes that framing: can the cabinet be more beautiful, more useful, and nicer in your hand?

Trade-offs are central because there is no free lunch: you have to give up something, and the craft is giving up the right thing for the market and user at that moment. Ivan sees AI language models as a new material, like aluminum enabling mass air travel or semiconductors enabling computers. New technology unlocks new trade-offs, but product makers must balance those with human behavior, which changes slowly after people reach about 16. The job is to find the sweet spot across multiple dimensions and create something that has a durable right to exist.

Ivan says he is not the soft-spoken archetype in every setting; growing up in China made him more direct, and he describes Notion’s ethos as closer to the East Coast than the West Coast. As the company grew, he had to move from one-to-one storytelling to one-to-many storytelling, even though podcasts and public communication drain him differently than building product. He treats that as a craft and a skill to unlock, much like a video game ability. Every 12 to 18 months Notion becomes a new company that demands a new skill set from him, which he finds fun because it is an infinite game.

Ivan admits that for a period Notion shipped things that were not great, and 2024 was the year he felt the company shipped good things at good velocity, quality, and alignment with its values. The concrete example is project management: rather than building sprints as clusters of tasks—new Lego pieces—the team hardcoded sprint and milestone concepts into the product. That violated Notion’s Lego philosophy, and Ivan says the product felt underpowered, did not work well with the rest of Notion, and felt weird to engineers who had to touch the code base. He describes the result as organ rejection from employees and customers, and it took about a year to a year and a half to realize they should return to the Lego way.

The bleakest operational crisis came during COVID, when Notion could not scale its infrastructure. For a long time the company ran on one Postgres instance, kept finding beefier machines, and eventually ran out of the largest instance available. That created a doomsday clock toward a complete Notion shutdown, so the company stopped building features and put almost every engineer on the problem. Ivan recalls having weeks, or somewhere between three and six months, before running out of time; the fix was sharding, and he now advises planning ahead rather than doing premature optimization too late.

Ivan would not want to build anything other than a horizontal product because Lego for software is inherently horizontal, though he also would not recommend it to most founders. The learning is segmentation: only hardcore Lego fans care about bricks, while most people want a ready-made Lego box. As Notion moved upmarket, Ivan had to learn the term “solutions,” meaning the product must sit somewhere on an enterprise customer’s P&L and optimize their business. For too long he stayed in the Lego-brick mindset; the shift is to sell the box while still building bricks behind the scenes. He describes bundling and unbundling as a long cycle, quoting Romance of the Three Kingdoms: empires long united must divide and long divided must unite.

Lenny shares survey data showing Notion ranking highly as a project management tool, docs tool, and CRM, even though Notion did not intend to build CRM. Ivan explains that CRM is a relational database, so Notion gives users that brick and they build CRM themselves; the advantage is that startups and mid-market companies can address several use cases in one place without each vertical need being too complex. That bundling also helps AI because more information sits together. Notion’s internal strategy is B2C2B: personal users start with notes and documents, discover relational databases and tasks, and bring Notion to work, with about half of B2B customers coming from prior personal use.

Ivan sees language models as a gift for tool builders and a particularly good fit for horizontal, bundled products. Notion first sliced AI writing into the surface where people already write and manage tasks; then it built AI Q&A and connectors so AI can reason over all the information in Notion plus external tools like Jira and Zendesk. The third and most fascinating phase is using AI coding agents to assemble the Lego blocks Notion has spent years building for knowledge work, so customers can create knowledge-work software or agents for their own vertical use cases. He says this was not expected at all: thank you, AI.

Ivan frames tools as extensions that shape us back, quoting Marshall McLuhan: we shape our tools, and afterward our tools shape us. He prefers to amplify creativity and beauty in software rather than zero-sum human tendencies, and he finds it heartwarming when people use Notion in coffee shops or when non-programmers make a living selling Notion templates and workflows. His advice to technologists is to steal ideas from outside tech and from history, because there are many patterns and trade-offs to borrow. In the lightning round he points to complex-systems literature and old BBC documentaries such as Connections, which trace how ideas from different domains inspire one another. His life motto is to treat things as craft: make it better, make it for yourself, and if it is unique enough for you and useful for others, things will follow.