How Notion Got Its First 1,000 Users: The Real Story
Notion didn't explode in 2016 — it barely survived. Learn how Ivan Zhao's near-dead startup found traction through permutations, patience, and a radical cost cut.
Most founders who study Notion walk away with the wrong lesson. The popular version of the story — two founders, a trip to Kyoto, a viral Product Hunt launch in 2016, instant growth — is a myth. Ivan Zhao has said it himself: real traction didn't arrive until 2018, with the Notion 2.0 launch. The 2016 launch was oxygen, not ignition.
What actually happened in the three years between near-bankruptcy and product-market fit is both more instructive and more uncomfortable than the myth. There was a full technical rebuild, a radical cost reduction, a design methodology that most teams have never heard of, and a founder who turned a support tool's push notifications into a constant physical reminder that real users existed. This is what that period actually looked like.
Why did Notion almost die before anyone had heard of it?
Notion's first version was built on Web Components and CouchDB, with an offline-first architecture that sounded clever in 2014 but produced constant crashes and data loss. Ivan told Sequoia directly: "Bugs came from everywhere. It was unstable. People were losing their content."
The technical problems were fixable. The strategic problem was harder. The founding pitch — software you assemble like LEGO blocks, build your own tools, no-code before "no-code" was a phrase — was too abstract for the market of 2014. Ivan's own words from Lenny's podcast: "We tried that for a couple of years and learned that actually most people just don't care."
By mid-2015, the money was running out. Team members had already quit. The gap between a genuinely interesting vision and a product people would actually open on a Tuesday morning had not closed.
What did the move to Kyoto actually accomplish?
The Kyoto period is usually told as an inspiring founder story. It was also a calculated survival move. Ivan fired what remained of the team, subleased the San Francisco office and apartments, and moved with co-founder Simon Last to a rented two-story house in Kyoto. Ivan's reasoning, from the Figma blog: "If you looked at the burn rate, we all would've died together. It wasn't much of a choice."
Kyoto wasn't chosen for romantic reasons. Ivan told Nira that Kyoto houses were larger than what was available in Tokyo or Osaka, they could bike everywhere, and the cost structure was radically lower than San Francisco. Neither founder spoke Japanese. Nobody around them spoke English. They coded.
What they built in that period wasn't an iteration of the broken product. It was a complete rebuild. They moved from Web Components to React, abandoned the offline-first architecture, and redesigned the data model around a single core concept: blocks. Every primitive in Notion — a paragraph, an image, a database, a to-do item — is the same kind of object, and any block can contain any other block. That architectural decision is what makes Notion structurally different from every competitor. It came out of Kyoto.
Notion 1.0 launched on Product Hunt on August 9, 2016 — not March 2016, a date that gets repeated incorrectly across most coverage. It received 5,537 upvotes, 170 comments, won Product of the Month for August, and took the Golden Kitty award for 2016. A solid launch. Not a growth moment. Ivan confirmed in Notion's own 2024 post celebrating 100 million users that 2016 wasn't when traction appeared.
What is the "permutations" method and why does it matter for product design?
During the Kyoto rebuild, Ivan spent so much time in Figma that the Figma team noticed him appearing at the top of their most active user list without knowing who he was. When they eventually talked to him and asked what he was doing, his answer introduced a concept worth understanding precisely.
Ivan called it permutations, and he drew a clear distinction from iteration. Iteration is sequential — one version, then a slightly improved version, then another. Permutations are parallel — many versions built simultaneously and compared side by side. His direct quote from the Figma blog in 2019: "Our general philosophy is to go crazy with permutations. If the taste is there, you can churn out versions until you find the best one."
The mechanics matter. Ivan's rule was explicit: "Do not replace that text. Just copy and create a new canvas." Every Notion design file from that era contained seven or eight canvases side by side, tracking how language and layout evolved from one to the next. New hires at Notion were shown those original Kyoto canvases as part of onboarding. The taste was treated as an institutional asset.
Two reasons this works structurally. First, quality only becomes visible through comparison — a single design looks acceptable until something better sits next to it. Second, parallel versions separate taste from ego. When you've made twenty versions, you're not defending any of them. You're judging all of them from the outside.
Applied practically, this means: for any decision that will shape your product for years — onboarding flow, data model, pricing page, first email — make at least five visibly different versions before choosing one. Not five variations on the same idea. Five different answers to the same question. Compare them in parallel, not in sequence, because working memory doesn't preserve what a design felt like ten minutes ago.
How did three product instincts keep Notion alive when growth wasn't coming?
Between the 2016 launch and 2018 product-market fit, Notion survived on three things that had nothing to do with marketing.
They were solving a problem people were already solving badly. By 2013, knowledge workers were juggling Evernote, Google Docs, Trello, and a growing list of single-purpose tools. Notion didn't invent a new behavior — it unified existing demand. That's a different and more durable starting point than creating demand from scratch.
The first five minutes of the product felt better than anything else available. Notion 1.0 wasn't feature-complete. But opening a page, typing, creating a block, and dragging it somewhere felt qualitatively different from every productivity tool that existed. That feeling kept early users around long enough for the product to develop.
Ivan and Simon answered every comment on the Product Hunt launch. When Akshay Kothari joined later as the third co-founder, Ivan made him enable push notifications for every Intercom support ticket. Akshay has described his phone buzzing every three seconds. That level of user contact, sustained through years without visible traction, is what kept the product evolving in the right direction.
None of these are tactics. They're instincts about what a product owes its users at the earliest stage.
What can founders building right now take from Notion's slow years?
Three things, none of them inspirational.
Stop confusing your launch with your product-market fit. Most products that eventually succeed have a quiet two to three year period between launch and real traction. Notion's 2016 launch was respectable. It wasn't the moment. If your launch didn't go viral, that's not evidence the product is wrong — it may just mean you're in the slow years. What matters is whether week-four retention is better than week-two retention.
Use permutations for high-stakes decisions. Not just for UI design. For pricing. For onboarding. For the data model. For the first message a new user receives. Any decision you'll have to live with for years deserves five versions compared in parallel, not one version improved in sequence.
Reducing what you need to survive is a valid strategic move. Ivan's insight in 2015 wasn't just "rebuild the product." It was "rebuild the product at a cost structure we can sustain." The move to Kyoto bought a year of runway at a fraction of the San Francisco burn rate. For founders approaching the end of a runway, the question isn't always how to raise more capital. Sometimes it's how to reduce the monthly cost of staying alive long enough to find the answer.
Practical takeaway
Notion's real lesson isn't about viral launches or founder mythology. It's that three product instincts — solving existing demand, delivering a first experience worth returning for, and maintaining obsessive user contact — can sustain a product through years without growth. Pair that with a design method that generates many parallel options before committing to one, and a willingness to radically cut costs when survival requires it. The combination isn't glamorous, but it's what actually happened.
At FlutterLab, the permutations approach is one we apply directly across product work — not for visual design alone, but for any architectural or UX decision that will constrain the product for years.
Frequently asked questions
When did Notion actually achieve product-market fit?
Ivan Zhao stated in Notion's own 2024 blog post that real signs of traction appeared with the Notion 2.0 Product Hunt launch in 2018, not with the 1.0 launch in 2016. The 2016 launch provided enough oxygen to continue, but it wasn't the growth inflection point most accounts describe it as.
Why did Ivan Zhao and Simon Last move to Kyoto specifically?
According to Ivan's account to Nira, Kyoto houses were larger and more affordable than equivalent spaces in Tokyo or Osaka, and the city was bikeable. The primary driver was practical: they needed to cut their burn rate dramatically while rebuilding the product, and Kyoto gave them more space at lower cost than their San Francisco setup.
What is the "blocks" architecture and why does it matter?
The blocks model — where every element in a Notion document is the same type of object and any block can contain any other block — came out of the Kyoto rebuild. It's the structural reason Notion can function simultaneously as a notes app, a database, a wiki, and a project management tool without those modes feeling bolted together. It's the architectural decision that differentiates Notion from competitors who added features on top of a document model.
What does the permutations design method look like in practice?
Rather than designing one version and iterating on it, you create many visibly different versions simultaneously and compare them side by side. Ivan's rule was to never replace a design — always duplicate the canvas and evolve the copy. This lets you compare the full arc of decisions at once, which is more reliable than trying to remember what an earlier version felt like.
Is there a difference between a product's launch moment and its product-market fit?
Yes, and Notion is the clearest example of why conflating them is dangerous. A launch can generate attention, positive coverage, and a spike in signups without any of that translating into sustained retention or growth. Product-market fit shows up in retention curves — specifically in whether users are still active weeks or months after signing up. Notion had a respectable launch in 2016 and product-market fit two years later. Treating the launch as confirmation of fit would have produced a false sense of security.