LLenny's Podcast
← All episodes
Ivan Zhao (CEO and co-founder)06 March 2025

Notion’s lost years, its near collapse during Covid, staying small to move fast, the joy and suffering of building horizontal, more

6Frameworks
14Insights

Frameworks in this episode

Insights & moments

The myth-busts, hot takes, explainers, and tools worth keeping.

Hot Take· 2

Hot Take20:00

Building for Your Values vs. Building for the Business

Ivan draws a distinction between building for a business — where you want users and revenue and the market is an arena to optimize a scorecard — and building for your own values, taste, and aesthetic. Too much of yourself and you have no users, just an art project; too much for the business and you build a commodity. He finds building for value more lasting and fulfilling.

  • Building a product/business means optimizing for users and revenue, like a sport with a scorecard
  • Building for value means putting your taste and aesthetic into the world
  • Too much self yields an art project with no users; too much business yields a commodity
  • Building for value is the more durable, fulfilling energy source through dark years

too much of yourself then there's no users you're just just doing art project

Ivan Zhao · 21:30

too much for a business you're building a commodity

Ivan Zhao · 21:30
#product-philosophy#values#founder-mindset#craft
Hot Take1:06:00

Steal Ideas From Outside Tech

Ivan's closing advice: tech is obsessed with the now and what happened six months ago, but reading across other industries and back through history surfaces a massive reservoir of patterns, shapes, and tradeoffs you can borrow. Looking sideways and backwards lets product and business makers solve the problem in front of them with ideas from outside their domain.

  • Tech, Hacker News, and Twitter over-focus on the immediate present
  • Other industries and history hold reusable patterns, shapes, and tradeoffs
  • Borrowing clever patterns from other domains makes the problem in front of you more interesting and solvable
  • Echoes 'good artists copy, great artists steal'

if you go back into history there's a massive amount of patterns and shapes and tradeoffs you can steal from

Ivan Zhao · 1:06:30
#creativity#learning#history#product-makers

Explainer· 4

Explainer30:00

Why Notion Stays Small: The 'Small Bus' and Talent Density

Ivan explains Notion's decade-long commitment to staying lean — profitable, with most raised money still in the bank, no salesperson until $10M ARR, first PM at ~50 people. His view: better systems beat solving problems through more people, since most overhead is internal communication. Internally they track talent density and revenue per employee, and use a 'small bus' metaphor — a smaller bus turns corners and accelerates faster.

  • Notion has been profitable for years with most raised capital untouched; no salesperson until >$10M ARR
  • Most company overhead is internal communication and aligning minds, not headcount
  • They track talent density and revenue per employee rather than number of people
  • The 'small bus' turns corners, accelerates, and maneuvers faster than a big one

we don't try to track number of people but we're try to track how Talent Den Revenue per employee

Ivan Zhao · 32:00

the smaller the bus is easier to turn Corners easier to accelerate easier to maneuver

Ivan Zhao · 33:30
#lean#management#talent-density#efficiency
Explainer52:00

Selling Horizontal Software: Lego Bricks vs. Lego Boxes

Ivan shares the hardest lesson of building horizontal: segmentation. Only hardcore fans want raw Lego bricks; most people want ready-made Lego boxes, and enterprises buy 'solutions' that sit on a P&L. A vertical product's category is its natural wall, but a horizontal product going after the entire software market has no wall — so you must create one yourself through go-to-market and positioning, while still building bricks behind the scenes.

  • Most customers want the ready-made box, not the raw bricks; only hardcore fans want bricks
  • Moving up-market and into B2B requires shifting from a brick mindset to a solution mindset
  • Vertical software's category is its built-in wall; horizontal software has none
  • You must create your own wall via positioning while still building bricks behind the scenes

for the longest time I felt too much in the Lego briak mindset now in the solution Lego box

Ivan Zhao · 53:00

building horizontal there's no wall at least for in our space we notion go after entire software Market but then you need to create a…

Ivan Zhao · 53:30
#horizontal-software#positioning#go-to-market#segmentation
Explainer56:30

Notion's B2C2B Engine: Notes as the Billion-User Top of Funnel

Ivan describes Notion's growth strategy, which he calls B2C2B: consumers adopt Notion for the simplest use case — note-taking and document sharing, something over a billion people do daily — then discover its database power and bring it to work. About half of Notion's B2B customers came from prior personal users. That logic drives new products: Ivan targets other billion-user categories, shipping a calendar last year and building email now.

  • B2C2B: consumers start with simple notes/doc sharing, then bring Notion to work
  • Roughly half of Notion's B2B customers came from prior personal users
  • Document notes is a billion-plus-user use case that funnels into enterprise adoption
  • Ivan targets other billion-user categories — document notes, calendar, email — hence the calendar and upcoming email products

half our B2B customers uh coming from prior personal users

Ivan Zhao · 57:00

which other category of software has 1 billion plus user there's document notes there's calendar there's email

Ivan Zhao · 57:30
#growth#b2c2b#product-strategy#notion
Explainer1:01:00

Bundling and Unbundling: The Market Waves of Software

Ivan frames software history as recurring waves of bundling and unbundling, quoting the opening of Romance of the Three Kingdoms. Paper was a fully democratized medium; PCs unbundled into standalone apps in the 80s, Microsoft rebundled into a suite in the 90s, SaaS unbundled again — and now, with the average company using nearly 100 tools, the market is swinging back toward bundling, accelerated by AI and macro pressure.

  • Software oscillates between bundling and unbundling, like empires uniting and dividing
  • PCs unbundled into standalone apps (e.g. dBASE) in the 80s; Microsoft rebundled in the 90s; SaaS unbundled again
  • The average company now uses nearly 100 tools — a sign the pendulum is swinging back to bundling
  • AI and macro conditions are pushing more value toward bundling for now, though the market could shift again

Empires long United must divide long divided must unite as has always been bundling un bundling

Ivan Zhao · 1:01:00

average company to use like almost 100 tools it's not it's kind of Madness

Ivan Zhao · 1:02:00
#bundling#software-history#market-cycles#strategy

Story· 6

Story09:00

How Steve Jobs Saw the GUI at Xerox PARC but Missed the Real Power

Ivan recounts the origin of personal computing at Xerox PARC, where Alan Kay's Smalltalk/Alto system had no separation between users and applications and everything was malleable. When Steve Jobs visited, he only registered the graphical user interface, not the underlying object-oriented environment where users could reshape their own tools. That miss, Ivan argues, locked the industry into the rigid application framework we still live with today.

  • Xerox Alto ran Smalltalk, where there was no such thing as an 'application' and everything could be changed by the user
  • In the demo, engineers instantly reversed the scroll bar direction for Jobs to show the system's malleability
  • Jobs took away the graphic user interface but not the object-oriented power beneath it
  • That framing led to the app-based, SaaS world Notion is now pushing back against

you know what Steve Jobs is actually at fault of this in some strange ways

Ivan Zhao · 09:00

he only saw the graphic user interface he did not see the underlying object orent the environment power

Ivan Zhao · 10:00
#computing-history#xerox-parc#steve-jobs#product-philosophy
Story10:30

Notion's Lost Years and Learning to Hide the Vision

Notion spent its first three to four years — the 'Lost Years' — building an easy developer tool so everyone could make their own software, and discovered most people simply didn't care. The team realized they had to hide that vision inside a form factor people actually use every day: productivity software. Ivan frames it as feeding people sugar with the broccoli hidden inside.

  • The first version was a developer tool; after a couple of years they learned most people don't want to create software
  • The fix was to hide the 'everyone can create software' vision inside a productivity suite people already understand
  • It took two-plus years just to realize they needed to build a productivity tool
  • The maturation was seeing the world from outside their own perspective, not just their own

learned that actually most people just don't care

Ivan Zhao · 12:00

people don't want to eat the broccoli but people like sugar so give them the sugar

Ivan Zhao · 13:30
#product-market-fit#founder-lessons#notion#vision
Story13:30

Why Notion Threw Away Its Codebase Over a Bad Tech Bet

Ivan explains that Notion rebuilt itself repeatedly, and one reset was purely technical: they had bet on Google's Web Components before React won, and the immaturity made bugs impossible to trace. Rather than run out of time on a broken foundation, they reset the codebase and effectively the company. It took four years total to get somewhere.

  • First two years were spent over-building a developer product nobody wanted
  • They bet on Web Components, an unstable pre-React technology where bug sources were untraceable
  • The choice was to rebuild the whole thing or run out of time
  • They reset the codebase and the company to build on the right technical foundation

before react wins there's a competing technology called Web component from Google

Ivan Zhao · 14:30

then we have to restart a company otherwise rebuild the whole thing otherwise we're going to run out of time

Ivan Zhao · 14:30
#engineering#tech-bets#startups#resets
Story15:00

How Ivan's Mom Kept Notion Alive

Staying solvent through the Lost Years partly came down to family. Ivan's mother helped him register a company (required for a Canadian to move to the US) and raise the initial money, which he later returned. When Notion ran out of cash again, he asked her to lend it back to bridge the gap.

  • As a Canadian, Ivan needed a registered US company to move, and his mom helped kickstart it
  • He returned that initial money, then had to ask to borrow it again when they ran out
  • His mother's bridge loan kept the company going during the cashless years

then we run out the money so hey Mom can I borrow that just to bridge us

Ivan Zhao · 15:30
#founder-story#funding#notion#survival
Story17:00

Laying Off Everyone and Coding in Japan

During one of the rebuilds, Notion laid off its entire team, shrinking from five people back to just Ivan and Simon. To change the scenery and reset morale, the two moved to Japan — and by subletting their SF apartment and office, they actually made money while living there. Ivan describes non-stop coding and design as some of the happiest days.

  • Notion cut from five people back to the two founders during a rebuild
  • They moved to Japan to change the scenario after a low-morale layoff
  • Subletting their apartment and office meant they made money living abroad
  • The two coded and designed non-stop, moving through the problem space fast because they knew each other's thinking

at that time notion was five people we lay off everybody back to me and samon two people

Ivan Zhao · 17:30

we sub leas our apartment and office we're actually making money living in Japan

Ivan Zhao · 18:00
#founder-story#layoffs#resilience#remote
Story49:30

The Covid Doomsday Clock: When One Postgres Instance Almost Killed Notion

Ivan tells the story of Notion nearly dying during Covid. For years the whole product ran on a single Postgres instance, scaled by moving to ever-bigger machines — until they hit the largest instance available and started running out of space to store everything. A 'Doomsday Clock' began counting down to a full shutdown; every engineer dropped feature work to shard the database, and they cleared it with maybe three-to-six months to spare.

  • Notion ran on one Postgres instance, scaling only by finding a bigger machine
  • They hit the largest available instance and faced running out of storage entirely
  • As they neared the limit, behavior became sporadic and unpredictable
  • All hands stopped building features to shard the database; the runway was three-to-six months

for the longest time we notion runs on one instance of postris database

Ivan Zhao · 49:30

a Doomsday Clock that when we're going to trly run out this space to store everything in notion and notion got completely shut down

Ivan Zhao · 50:00
#infrastructure#scaling#postgres#crisis

Takeaway· 2

Takeaway22:30

Don't Fear the Reset — Better Abstractions Compound

Ivan reframes sunk cost: resetting isn't just a self-help mantra, it's that progress made through better abstractions compounds and can recover everything you threw away faster than you'd expect. Humans think linearly, not exponentially, so we underestimate how quickly a better approach catches up. He points to Smalltalk's tiny Lisp-inspired kernel — roughly a hundred lines — that produced complex, valuable behavior.

  • Better abstractions create progress that compounds, so sunk cost recovers quickly
  • Humans think linearly and underestimate exponential/abstraction gains
  • Smalltalk and Lisp show a tiny kernel can compound into complex, valuable behavior
  • The same pattern shows up in LLM advances and Chain-of-Thought models that rethink a wrong turn

you can create progress through better abstractions and that thing compounds faster

Ivan Zhao · 23:00

your humans are not thinking not good at thinking in terms of abstraction

Ivan Zhao · 23:00
#abstraction#sunk-cost#systems-thinking#ai
Takeaway34:30

Why Ivan Obsesses Over the Office (No Shoes, Warm White, No Top Light)

Ivan explains the deliberate thought behind Notion's office: no shoes, one shared table, thirty shades of warm white tested on the walls. Part is pragmatic — most offices feel like hospitals with pale walls and dark floors, so cream walls, friendly floors and no top light make people feel creative and at ease. Part is a values/aesthetic stance: ugly things visibly hurt, the same way a bad chair hurts your back.

  • The office should feel like an artist's studio or a home, using home furniture, not corporate fittings
  • Avoid pale white walls, dark floors and top light — they make offices feel like hospitals
  • Cozy space makes people spend more time and feel more creative
  • Ugly visual input hurts, so aesthetics is an ergonomics-for-the-eyes value, not just decoration

most office the top life feels like hospital

Ivan Zhao · 35:00

don't use white use some kind of cream make floor is more friendly colors and don't use top light top light is evil

Ivan Zhao · 35:00
#office#aesthetics#culture#craft