The Challenge
Notion's first version was a disaster. The product was overly complex, targeting too many use cases at once. The technology was unreliable. The team of approximately 80 people was burning cash with no product-market fit.
Zhao had to make an agonizing decision: shut down completely, or reduce to a skeleton crew and try again. He chose to reduce to just 4 people and relocate to a cheaper location. But before rebuilding, he needed to understand exactly why V1 failed — otherwise he'd repeat the same mistakes.
The Approach — Tools in Action
Zhao applied structured problem-solving to dissect the failure:
Issue Trees broke the failure into MECE (Mutually Exclusive, Collectively Exhaustive) sub-problems:- Product complexity: Tried to be everything — wiki, project management, notes, database — before being good at anything
- Wrong target market: Initially targeted enterprises, but the product wasn't stable enough for enterprise requirements
- Technical architecture: Built on unstable foundations that couldn't scale
- Team alignment: Too many people building in different directions without a shared vision
- Building too many features before nailing the core → focus on blocks-based editing first
- Targeting enterprises before individuals → start with individual users and grow bottoms-up
- Using unproven technology → rebuild on proven, reliable infrastructure
- Having too many people too early → stay small until product-market fit
The Productive Thinking Model guided the rebuild through six structured steps: defining the problem, identifying success criteria, generating solutions, selecting the best approach, building an action plan, and aligning the (now tiny) team.
The Outcome
The comeback was remarkable:
- Notion relaunched with a small team and a focused product — the blocks-based editor that let users build custom workspaces
- Grew to 30M+ users through word-of-mouth and bottoms-up adoption
- Reached a $10B valuation
- Became the defining product in the "all-in-one workspace" category
The lesson: Notion's V1 failed because they built before understanding. V2 succeeded because they diagnosed the failure systematically before writing a single line of new code.
Key Takeaway
Failure isn't the end — but only if you diagnose it rigorously. Use Issue Trees to break down what went wrong, Inversion to identify what NOT to repeat, and structured thinking to rebuild with clarity.
Tools Used in This Story
Issue Trees
Problem SolvingStructure and solve problems in a systematic way
Inversion
Problem SolvingApproach a problem from a completely different angle
Productive Thinking Model
Problem SolvingSolve problems creatively and efficiently