Lessons from Ivan Zhao

Lessons from Ivan Zhao

Ivan Zhao co-founded Notion to let non-technical people build their own software tools. Before the app became a workplace standard, he and his co-founder nearly went broke, fired their team, and moved to Kyoto to rewrite it from scratch. The notes below cover his approach to design, computing history, and running lean companies.

Part 1: The Kyoto Reset

1. On running out of money: "We built a platform that no one wanted and crashed the app. We had to sub-let our San Francisco office and move to Kyoto to survive." — Source: [Lenny's Podcast]
2. On isolation: "We didn't speak Japanese, so there was nothing to do but code. We programmed for eighteen hours a day in our underwear." — Source: [Figma Config]
3. On firing friends: "Letting go of the early team was the hardest thing. It was a reset we had to make to save the company, but it fundamentally changed how conservatively I hire today." — Source: [Lenny's Podcast]
4. On rebuilding from scratch: "We threw away all the old code. We realized we were trying to build the entire house, when we just needed to build the LEGO blocks first." — Source: [Station F]
5. On finding focus: "The constraints of Kyoto stripped away all the Silicon Valley noise. There were no coffee meetings or networking events, just the product." — Source: [Forbes]
6. On surviving the trough: "You have to genuinely love the craft of building the software, because if you are just in it for the success, the dark years will break you." — Source: [First Block Podcast]
7. On rewriting the architecture: "We switched to a block-based architecture. Every paragraph, image, and checkbox became a distinct unit in the database." — Source: [Notion Blog]
8. On returning to San Francisco: "We didn't come back until we had a product that we knew worked. The reception for Notion 1.0 was immediate validation." — Source: [Possible with Reid Hoffman]
9. On early pricing: "We launched on Product Hunt with a very simple subscription model because we desperately needed the cash flow to hire our next engineer." — Source: [Grit Podcast]

Part 2: Computing History

1. On Douglas Engelbart: "Engelbart's 'Mother of All Demos' in 1968 showed us that computers are meant to augment human intellect, not just process data." — Source: [Ness Labs]
2. On Alan Kay: "Alan Kay taught us that the computer is a medium, like paper or clay. It should be malleable enough for anyone to shape it." — Source: [Station F]
3. On lost computing ideals: "The pioneers of the 1970s imagined a world where everyone was a programmer. We lost that when software became rigidly packaged applications." — Source: [First Block Podcast]
4. On toolmaking: "Humans are toolmakers. Software should be a tool that lets people make more tools, rather than a pre-fabricated solution." — Source: [Lenny's Podcast]
5. On Ted Nelson: "Project Xanadu and the idea of deep, bi-directional linking heavily influenced how we structured pages in Notion." — Source: [Notion Blog]
6. On Xerox PARC: "They invented the GUI by studying how people think and work. We try to study cognitive science the same way to inform our interface." — Source: [Grit Podcast]
7. On computing as a bicycle: "Steve Jobs called the computer a bicycle for the mind. We want Notion to be the custom assembly kit for that bicycle." — Source: [Possible with Reid Hoffman]
8. On the command line: "The slash command is a modern take on the terminal. It gives you the speed of a keyboard interface without having to memorize complex syntax." — Source: [Figma Config]
9. On historical continuity: "We are not inventing new paradigms. We are just executing the unfulfilled promises of the personal computing revolution." — Source: [Station F]
10. On medium versus tool: "A tool does one job well. A medium allows you to express your own ideas. We wanted to build a medium." — Source: [Lenny's Podcast]

Part 3: Product Strategy

1. On sugar-coated broccoli: "You have to give users something immediately useful, like a simple text editor (the sugar), before exposing them to the powerful database features (the broccoli)." — Source: [Lenny's Podcast]
2. On horizontal products: "Building a horizontal product is agonizing because you are competing with everyone. But if you succeed, your total addressable market is anyone who uses a computer." — Source: [Possible with Reid Hoffman]
3. On the LEGO block model: "If you give people pre-built castles, they get bored. If you give them blocks, they will build things you never imagined." — Source: [Ness Labs]
4. On replacing apps: "Our goal wasn't to replace Google Docs or Trello specifically. It was to replace the need to switch between five different apps to do one project." — Source: [Station F]
5. On templates: "Templates are how you educate users on the possibilities of a blank canvas without overwhelming them with an empty screen." — Source: [First Block Podcast]
6. On feature requests: "We read every feature request, but we rarely build exactly what users ask for. We try to build the underlying block that allows them to solve the problem themselves." — Source: [Grit Podcast]
7. On product-led growth: "The product has to be so good for a single user that they organically force their team to adopt it. We never had a sales team in the early days." — Source: [Lenny's Podcast]
8. On speed as a feature: "A note-taking app has to open instantly. If there is a one-second delay, the user's thought is gone, and the product has failed." — Source: [Possible with Reid Hoffman]
9. On constraints: "We deliberately limit font choices and colors. Too much formatting freedom distracts from the actual work." — Source: [Figma Config]
10. On bundling: "The history of software is unbundling and bundling. We are a deliberate attempt to bundle the fragmented modern workspace." — Source: [Station F]

Part 4: Design as a Medium

1. On cognitive science: "My background in cognitive science taught me that memory is spatial. That is why drag-and-drop and nesting pages feel natural to the brain." — Source: [Lenny's Podcast]
2. On typography: "We spent weeks choosing our default sans, serif, and mono fonts. Typography is the voice of the product, and it has to feel calm." — Source: [First Block Podcast]
3. On blank states: "A blinking cursor on a white screen is terrifying. We use placeholders and subtle prompts to invite the user to start typing." — Source: [Grit Podcast]
4. On the aesthetic of quietness: "Enterprise software is usually loud, full of colorful buttons and alerts. We wanted an aesthetic of quietness, like a clean sheet of paper." — Source: [Figma Config]
5. On illustration: "The black-and-white, hand-drawn illustrations we use are meant to feel human and slightly imperfect, contrasting with the rigid software grid." — Source: [Ness Labs]
6. On craftsmanship: "If the pixel alignment is off by one unit, the user might not consciously notice, but they will unconsciously feel that the software is fragile." — Source: [Possible with Reid Hoffman]
7. On physical metaphors: "We design software with physical metaphors in mind, like paper, binders, and blocks. It grounds abstract data in physical reality." — Source: [Station F]
8. On minimalism: "True minimalism isn't removing features. It is hiding complexity until the exact moment the user needs it." — Source: [Lenny's Podcast]
9. On dark mode: "Dark mode shouldn't just be an inversion of colors. It requires a complete re-balancing of contrast to reduce eye strain." — Source: [First Block Podcast]

Part 5: Lean Operations

1. On jazz mode: "I prefer running the company in jazz mode. You don't need a rigid hierarchy if you have a group of highly skilled individuals playing off each other in sync." — Source: [Lenny's Podcast]
2. On hiring slowly: "We operated with a team of less than 20 people while serving millions of users. Adding people usually slows down the product cycle." — Source: [Forbes]
3. On generalists versus specialists: "In the early days, we only hired engineers who had a deep sense of design. You can't separate the code from the user experience." — Source: [Grit Podcast]
4. On preserving culture: "Culture degrades when headcount doubles too fast. We intentionally constrained our hiring to maintain a high talent density." — Source: [First Block Podcast]
5. On office space: "Our first office looked like an art studio. The environment you work in dictates the quality of the product you ship." — Source: [Figma Config]
6. On funding: "We delayed taking venture capital for as long as possible because we wanted the time to get the product right without the pressure of a ticking clock." — Source: [Possible with Reid Hoffman]
7. On decentralization: "A healthy organization pushes decision-making to the edges. If everything requires my approval, we are moving too slowly." — Source: [Station F]
8. On refounding: "A company has to be refounded at different stages of scale. The systems that got you to ten employees will break at a hundred." — Source: [Lenny's Podcast]
9. On meeting culture: "We write everything down. If a proposal isn't documented clearly in a page, we don't have a meeting about it." — Source: [Ness Labs]

Part 6: Building Notion

1. On community support: "Our community of ambassadors built Notion's early momentum. They made YouTube tutorials and shared templates when we didn't have a marketing budget." — Source: [Station F]
2. On international expansion: "We saw massive adoption in Japan, Russia, and Korea before we even localized the app. Good design translates across languages." — Source: [Grit Podcast]
3. On scaling databases: "During COVID, our database capacity was pushed to the absolute limit. We were rewriting the infrastructure while the plane was flying." — Source: [Lenny's Podcast]
4. On dogfooding: "We run the entire company on Notion. If a feature is annoying for our users, it's annoying for us, and we fix it immediately." — Source: [First Block Podcast]
5. On early feedback: "I used to personally sit in coffee shops and watch people try to use the app for the first time. It is the most humbling exercise a founder can do." — Source: [Figma Config]
6. On building for teams: "A product is only a collaborative workspace if it handles permissions gracefully. That was one of the hardest technical hurdles we solved." — Source: [Possible with Reid Hoffman]
7. On branding: "We wanted the brand to feel like a friendly local bakery, not a massive enterprise software conglomerate." — Source: [Ness Labs]
8. On the first thousand users: "We manually emailed every person who signed up in the first year to ask them what they were using the tool for." — Source: [Station F]
9. On pivots: "Notion 2.0 was the real pivot. Moving from a document-centric model to a database-centric model unlocked the real power of the product." — Source: [Lenny's Podcast]

Part 7: AI and the Future

1. On generative AI: "AI isn't just a new feature, it is a new interface paradigm. Natural language is replacing the mouse as the primary way we manipulate software." — Source: [Possible with Reid Hoffman]
2. On context in AI: "An AI assistant is only as smart as the data it has access to. Because all your company's knowledge is already in Notion, the AI has perfect context." — Source: [No Priors Podcast]
3. On writing assistants: "We integrated AI not to write for you, but to help you mold your thoughts, like a sparring partner for your ideas." — Source: [Lenny's Podcast]
4. On automation: "In the future, you won't need to manually update a status board. AI agents will read your commits and update the documentation automatically." — Source: [First Block Podcast]
5. On the fear of AI: "Technology has always displaced certain tasks, but it elevates human ambition. We will just focus on higher-level problem solving." — Source: [Grit Podcast]
6. On search: "Keyword search is dead. With Q&A AI, you can ask a question and get a synthesized answer based on thousands of internal documents." — Source: [Possible with Reid Hoffman]
7. On AI agents: "We are moving from software as a tool to software as a teammate. Agents will work asynchronously alongside you." — Source: [No Priors Podcast]
8. On interface evolution: "The chat box is just a transitional interface. Eventually, AI will be invisibly embedded into the fabric of the workspace." — Source: [Station F]
9. On retaining human touch: "Even as we automate tasks with AI, the aesthetic output must still feel distinctly human. The interface must remain calm." — Source: [Lenny's Podcast]

Part 8: Personal Philosophy

1. On childhood hobbies: "I learned to code because I wanted to modify video games, and I learned art through watercolor painting. Both disciplines require patience and an eye for structure." — Source: [Substack Profile]
2. On patience: "It took us five years to become an overnight success. You have to be willing to be misunderstood for a very long time." — Source: [Lenny's Podcast]
3. On reading: "I read a lot of history. If you understand how the Industrial Revolution reshaped society, you can better predict how the information age will do the same." — Source: [Grit Podcast]
4. On optimism: "To be a founder, you have to possess a delusional level of optimism. You have to believe you can bend reality to your will." — Source: [Possible with Reid Hoffman]
5. On success: "Success isn't reaching a valuation. It is walking into a cafe and seeing someone using the tool you built to run their life." — Source: [Station F]
6. On failure: "Our near-collapse in Kyoto was the best thing that ever happened to me. It burned away my ego." — Source: [First Block Podcast]
7. On craftsmanship versus scale: "The tension of running a company is balancing the slow, meticulous work of a craftsman with the aggressive growth expectations of venture capital." — Source: [Lenny's Podcast]
8. On art and tech: "Software engineering is the most accessible art form in the modern era. You write lines of logic, and they manifest as an interactive experience." — Source: [Figma Config]
9. On continuous learning: "The moment you think you have figured out how to run the company, the scale breaks your systems, and you are a beginner all over again." — Source: [Possible with Reid Hoffman]
10. On legacy: "I don't care if people remember my name. I want them to remember the feeling of empowerment they had when they built something in Notion." — Source: [Ness Labs]