Notion’s Ivan Zhao: The Refounder

Navigating founder mode, delegation, and early growth at Notion

The concept of 'founder mode' versus 'manager mode' has significantly influenced how CEOs lead, with some interpreting it as a license to maintain a hands-on, deeply involved approach. This shift is notable as leaders like Jack Dorsey and Brian Armstrong are now even reshaping their companies to integrate new focus areas like AI.

Ivan Zhao, CEO of Notion, experienced his own variations of these modes. He describes a 'small detour' from a purely founder-driven approach, where he focused on delegating responsibilities to facilitate growth. This was a critical step as Notion scaled its operations.

In its early stages, Notion maintained a lean team, with fewer than ten people in 2019 when it achieved product-market fit by early 2020. This small team size contributed to the company's profitability. Despite its lean structure and eventual profitability, Notion did secure venture capital, enduring a challenging five-year period of C-stage funding.

first five years just C-stage money, drag it out for five years.

Notion's Scaling Journey and Lessons on Management

Notion faced significant challenges in achieving product-market fit, a period described as "despair" before finding its footing. The company maintained a small, profitable team early on, which allowed for easier fundraising and a "default alive" status, fostering resistance to rapid headcount expansion.

As Notion grew and achieved product-market fit, its founder learned the necessity of delegation and introducing professional management. This shift, influenced by conventional wisdom among investors, was initially resisted as the company preferred a lean structure.

Reflecting on this growth phase, Ivan acknowledged a mistake in delaying the establishment of a sales team, having initially believed in creating an entirely new product-led growth motion. Notion did, however, opt for traditional approaches in building product marketing and engineering functions, hiring experienced SaaS executives who shaped the company's early DNA.

Notion's Blurry Roles and Technology-First AI Development

Developing products with language models differs significantly from traditional software engineering. Instead of a predictable process like building a bridge with clear handoffs between roles, working with AI is more akin to brewing beer; the underlying outcome isn't entirely predictable. This means the development approach shifts from being primarily customer-driven to technology-driven, prioritizing experimentation with new capabilities.

This unpredictable nature forces a change in how product teams operate. Traditional role boundaries become blurry, encouraging designers, engineers, and product managers to collaborate closely. They sit together, working with evaluations and experiences to discover new possibilities that the technology offers.

Notion's organizational DNA was inherently suited for this shift. From its inception, the company favored individuals with versatile skills, such as designers who could code. This foundation fostered a culture where engineers possess strong product sense and product managers are hands-on, exemplified by one PM being among the heaviest token consumers in the company, constantly experimenting.

Building classic software is like engineering a bridge... and building with language model... it's like brewing beer. You can't truly predict the thing.

AI Transforms Traditional Organizational Hierarchies

Traditional hierarchical organizational structures, often visualized as large triangles, are widely considered inefficient for the flow of information. Thinkers like Dorsey and Armstrong suggest it's time to fundamentally rethink these age-old structures, particularly with the advent of artificial intelligence.

The core idea involves placing AI at the center of organizations to delegate tasks, automate decision-making, and provide context. This approach aims to streamline processes that historically required humans to act as intermediaries, pushing information or making fuzzy decisions, which is often termed 'paper pushing'.

Language models are uniquely suited for this transformation. They can generate software to design efficient information and decision pipelines within a company. Furthermore, they are capable of performing the 'meaning-decision making' that frequently occurs between various decision points, reducing the need for human intervention in these intermediary roles.

The impact of AI on organizational design is likened to the impact of steel on architecture. Before steel, buildings were limited to about five or six floors. Just as steel allowed for much taller and more complex structures, AI and software act as 'the steel for organizations,' enabling fundamentally new and more efficient organizational structures to emerge.

Language model plus software is that, it's the steel for organizations. What does the organization look like now that we have the steel?

Notion prioritizes agency and curiosity over experience in engineering hires.

Organizational structures are evolving, with some companies like Notion embracing flatter models in the age of AI. While human nature might traditionally lean hierarchical, the current trend sees increased direct reports, with some leaders managing 15-25 people, a shift from the typical 7-9.

Notion has significantly altered its engineering hiring strategy over the past two years. The company no longer prioritizes traditional experience; instead, it optimizes for candidates' agency, energy, optimism, and curiosity, believing these intrinsic qualities are more indicative of future success.

The company's talent framework considers a person's capability, their taste or value system, and their agency or will. With large language models democratizing basic capabilities like writing and programming, a person's fundamental skill set is becoming normalized.

As core capabilities become more accessible and widespread due to AI, an individual's 'taste' and 'value system'—their inherent drive, what they aspire to create, and the direction they want to pursue—emerge as crucial differentiators in assessing talent.

No longer optimized for people's experience, optimized for people's agency, energy, optimism, curiosity.

Notion's Evolving Hiring Philosophy and Decentralized Marketing Structure

Notion employs a distinctive hiring strategy for engineers, favoring a "barbell shape" model. This means they look for both extremely junior, recent graduates and highly senior individuals. Senior engineers are valued for their "taste" and ability to provide direction, often managing multiple junior engineers or even AI-powered coding agents, while junior hires contribute raw output and speed.

The company's approach to roles like designers and product managers is also evolving. Designers are increasingly expected to function like PMs by driving initiatives and interacting directly with customers, going beyond just visual craft. Notion's hiring criteria has shifted from merely seeking "slope" (raw intelligence or learning speed) to emphasizing traits like "taste," "curiosity," and "agency," indicating a desire for individuals who can offer nuanced judgment and self-direction.

Notion has recently restructured its marketing department by eliminating the traditional Chief Marketing Officer (CMO) organization. Instead, marketing functions, such as storytelling, are now decentralized and integrated more closely with product teams. This change positions marketing activities to be more directly tied to product development and evolution.

We no longer have a CMO organization.

Notion adopts a decentralized 'jazz band' leadership model in a fast-paced AI environment.

Notion has restructured its marketing function, moving away from a centralized CMO model. Marketing now sits directly with product and social teams to keep up with rapid development cycles, and another segment focuses purely on demand and lead generation. This decentralized approach allows teams to solve problems directly without lengthy information round trips.

The company has also revised its hiring and compensation strategies to adapt to the current fast-paced environment. For roles like engineers, designers, and especially sales representatives, the focus has shifted from resumes to evaluating what candidates have actually built. Compensation is becoming more meritocratic, moving away from the 'peanut butter' approach common in the earlier SaaS era.

Ivan Zhao describes the current business landscape as 'wartime' compared to the 'peaceful' and somewhat 'boring' SaaS era. He finds this new environment more engaging and alive, requiring a different organizational approach. He believes you 'can't peanut butter things' in this new era.

Notion's leadership embraces a flexible, 'jazz band' approach, with 50-60 ex-founders among 1,000 employees capable of leading initiatives bottom-up. This allows for a fluid interplay between top-down guidance and bottom-up innovation, where leaders are empowered to take ownership. Zhao himself can 'dance around really smoothly' between these modes, contributing where he can provide unique value without encountering territoriality.

Where a company a thousand people, we have about fifty, sixty founders in the company, ex-founders, it's like they really can lead things, right?

Ivan Zhao on AI's financial impact and Notion's Kyoto re-founding

Ivan Zhao discusses the profound shift in strategic planning due to AI's rapid evolution. While financial planning still provides a useful "treadmill speed" metric, product strategy has become a weekly, rather than quarterly, endeavor due to the fast-changing market and technology. This new pace requires a more adaptive, "jazz band" approach to product development, contrasting with a traditional "marching band" structure.

The emergence of AI also introduces new cost dimensions and impacts gross margins. Companies now have to actively consider infrastructure costs. While frontier models are crucial for areas like coding agents, knowledge work products like Notion can often leverage second-tier or open-weight models, which are less powerful but also less expensive. This distinction suggests different gross margin profiles across various AI product categories.

Zhao recounts Notion's first re-founding, a pivotal moment born from necessity. After struggling for four to five years to find product-market fit, he and his co-founder laid off their small team. This difficult decision was driven by the realization that they would fail if they didn't fundamentally change their approach.

To facilitate this rebirth, the co-founders moved to Kyoto, Japan. They chose Kyoto for its larger, more affordable apartments compared to Tokyo, and because it was a completely new environment that could help shift their morale and mindset. The experience of intensely coding and rebuilding Notion in Kyoto was ultimately liberating, providing a fresh perspective to refine their product.

It's not easy, but sometimes your body just tells you you have to do it. Like, you know you're dead if you don't do it.

Notion's Rebuild Rooted in Craft and 'Tools for Thought' Lineage

The decision to rebuild Notion from the ground up in Kyoto was deeply influenced by the city's rich craft culture. The founders saw parallels between building physical tools like knives or ceramics and developing software tools for humans, fostering a unique inspirational environment for their work.

The commitment to rebuilding stemmed from a profound personal obsession with the "Notion shape tool problem" since college. The founders were part of the "tools for thought" community, tracing their lineage directly back to early computing pioneers and the Grateful Dead era, viewing Notion as a continuation of this historical trajectory rather than just another startup idea.

This approach contrasts sharply with much of modern Silicon Valley's "tinker culture," which often lacks an understanding of computing history or the intersection of humanity with technology. The founders lament that many in tech are unaware of figures like Douglas Engelbart or Alan Kay, indicating a broken lineage and a potential loss of the craft ethos present in earlier tech movements.

It wasn't like another idea kicking around, this is your life's work.

Notion Undergoes Second Refounding After Encountering GPT-4

Ivan Zhao discusses the importance of founders trusting their inner intuition and being willing to make drastic changes, even refounding a company, when it's stagnating. He suggests that many founders might benefit from taking a step back and redefining their path, especially given current market dynamics.

Notion's second refounding occurred in Cancun when the company had 500 employees. The catalyst was getting early access to the GPT-4 model. Ivan describes his first interaction with GPT-4 as a "religious experience," instantly recognizing its potential to fundamentally alter everything and necessitate a complete shift for Notion into an AI company.

Despite some internal skepticism from employees who questioned if AI was another passing trend like crypto, Ivan and co-founder Simon had a clear conviction about the necessity of this pivot. They launched their first AI product, a writing tool, two weeks before ChatGPT, which provided a meaningful revenue boost.

However, subsequent efforts to build an agent product throughout 2022 and 2023 proved to be a difficult "slog" for a year and a half. Despite collaborating with Anthropic and OpenAI for fine-tuning, the technology wasn't mature enough, leading to low morale until the solutions began to work and revenue growth inflected later on.

GPT-four is a religious experience for me. It's like, holy shit. It's, it's a full body religious experience, like you have to do something This, this is what change everything. Anything you do, if you don't do this, it will be meaningless.

Founders Must Get Hands-On to Decalcify Companies with AI

Ivan Zhao recounted the challenging period of repeatedly rebuilding Notion's AI foundation, likening the early pre-product-market fit despair to the difficulty of navigating an unknown path with low growth. He advises established companies, especially those with 500+ employees, to approach AI development by focusing directly on the product, treating it like "brewing beer, not building bridges."

A critical aspect for integrating new technologies like language models is for founders to be directly hands-on. Ivan stresses that leaders need to "feel the AI" by building with it themselves, rather than just reading about it or watching videos. Businesses are path-finding entities, and this direct engagement helps identify new opportunities and angles that open up with new ingredients in the market.

To combat "calcification" in older, larger companies, Ivan suggests intentional acquisitions. He describes acquired founders as "decalcified meathead machine nerds" who are driven to break and rebuild things, thus regenerating innovation within the company. This strategy helps maintain a dynamic, product-first mindset.

Motivating acquired founders is crucial; they should be given platform support and leverage to continue working on their original domains within the larger company. For example, the person leading Notion's enterprise search was the founder of an enterprise search product, allowing them to feel appreciated and empowered to build with greater resources.

You can't read about it, you can't watch YouTube, you gotta feel it.

Future Organizations Will Maintain Hierarchy with AI Enhancing Decision-Making

Ivan Zhao predicts that future organizations, even with advanced AI, will likely retain a hierarchical structure. This is attributed to the static nature of human behavior, where observing natural phenomena like bamboo or chimpanzee groups reveals inherent hierarchies. He emphasizes that humans as a species do not change quickly, and the division of labor naturally leads to specialized roles.

He identifies unchanging "invariables" such as human nature and the current legal systems that require human accountability, like CEO and CFO signatures. These factors suggest that a hierarchical system will persist to manage groups of humans with their diverse biases, preferences, and values. The challenge then becomes how to effectively use new technology to facilitate information flow and decision-making within these established structures.

AI's role is envisioned as a tool to enhance these human-led organizations. Initially, AI makes a small percentage of decisions, perhaps one percent. However, as AI systems gain more context and historical data, they will become better at making decisions than humans in certain areas. Humans will then shift to training, providing context, and refining the "taste" of these AI systems, allowing the AI to increasingly handle decision-making.

Zhao also highlights that modern knowledge work is a relatively recent invention, only about 150 years old, unlike ancient concepts such as fire or language. This relatively short history suggests ample room for evolution and new approaches, especially with the integration of AI. The once "shit industry" of knowledge management is now becoming incredibly important, ideally suited for platforms that can manage and leverage organizational knowledge.

If you just observe bamboo or chimpanzees, right? It's like very, if you watch nature program, there's hierarchy. That doesn't change.

Ivan Zhao's Strategies for Introverted Leadership and Energy Management

Ivan Zhao, an acknowledged introvert, describes his journey to comfortable public speaking as a leader. He initially delegated all-hands meetings but realized the necessity of direct communication for team trust. Instead of a teleprompter, which he finds difficult due to English being his second language, he prepares by speaking his thoughts aloud and using speech-to-text models to create a script. His company now holds weekly all-hands or AMA sessions, where he emphasizes the importance of direct answers to avoid missing opportunities to rally the team.

To manage his energy as an introvert, Zhao prioritizes deep work in the mornings, waking at 7 AM for thinking and writing at home before heading to the office. He's experimenting with daytime gym visits to boost his energy mid-day. His work schedule involves many 25-minute meeting blocks, with his team helping to bundle meetings to create 1-2 hour uninterrupted thinking chunks, predominantly in the morning when his energy is highest.

Weekends are dedicated to "happy time," allowing him to pursue curiosity-driven interests and reading rather than responsibility-driven tasks. This distinction helps him maintain a balanced energy level and prevent burnout. This approach allows him to sustain his demanding role while honoring his introverted nature.

you can be led by your curiosity rather than by your responsibility

Navigating the CEO Game by Aligning Values and Leaning into Strengths

Being a CEO involves a complex "social game" encompassing elements like status and power, but it also presents an opportunity to pursue personal values. Founders must find a sustainable equilibrium between the demands of market competition and their intrinsic beliefs. This balance is crucial for long-term energy and motivation.

Ivan Zhao emphasized the importance of aligning company goals with personal values, such as the craft of building good tools and focusing on the human side of things, rather than solely engaging in competition for its own sake. He notes that solely pursuing craft without building a business might be better suited for an artist, highlighting the need for a balanced approach.

He strongly advises founders to lean into their strengths rather than expending energy on fixing weaknesses. While certain communication skills become necessary as a company scales, within those boundaries, it is more effective to amplify what one is inherently good at.

In an environment where many capabilities are becoming commoditized by machines, a founder's unique point of view and core strengths are what truly last and differentiate their business.

I need to be in equilibrium with my own value of what this company I want to build with that gives me more gen- more sustainable energy source. I can't just chase something or competition for competition's sake.

Notion's Shift to a Hybrid Enterprise Sales Strategy

Notion initially made the mistake of trying to reinvent their go-to-market motion for enterprise sales, believing they could avoid traditional methods. Ivan Zhao reflects that companies should preserve innovation for a few key areas, rather than attempting to redesign well-established functional playbooks like enterprise sales, which have been refined over decades.

Their initial approach, heavily influenced by product-led growth (PLG), proved insufficient for enterprise clients. This 'system thinker' sales model primarily involved taking orders, making it hard for sellers to pursue more complex or challenging deals when easier, inbound opportunities were available. It became clear that simply filling demand wasn't enough to scale enterprise business.

To address this, Notion successfully revamped its strategy by building a robust enterprise sales structure. They paired Erica, a system-thinking CRO who previously worked at GitHub, with Praveesh, a "meat-eater" head of sales known for his aggressive, results-driven approach. This combination allowed for both strategic system development and energetic sales execution.

This dual-leader model ensured sellers operated within a respected, organized framework while being motivated by a dynamic leader. Notion's culture also played a crucial role, fostering receptiveness to sales within the engineering team, avoiding the common internal friction seen in many organizations.

You shouldn't reinvent new things unless it's really absolutely necessary. I was trying to reinvent a new go-to-market motion. That's stupid.

Companies as a Source of Meaning and the Importance of Refounding

Ivan Zhao views companies as modern-day religions, providing employees with a sense of meaning and purpose, much like faith or sports offer camaraderie. He notes that the root of

culture

is

cult

The first four words in culture are cult.

Developing a new CEO playbook for AI-native companies

A new CEO playbook is emerging for companies operating in the AI-native era, moving away from traditional organizational structures. Historically, companies have relied on a triangular org chart, a model that has existed since the Roman Empire. While functional, this structure becomes increasingly inefficient as a company grows, leading to distorted information flow and slower decision-making across its various levels.

The new approach, observed in companies like Jack Dorsey's and Ivan Zhao's, proposes a circular organizational chart. At the core of this circular model is a centralized AI system that holds all relevant context. Initially, human employees retain decision-making authority.

However, as the AI system is fed more context and becomes smarter over time, the responsibility for decisions gradually shifts. This allows the AI to take on an increasing number of decision-making tasks. This transition is not instantaneous but a phased evolution, requiring a legible set of systems to manage the process effectively.

Companies that embrace this AI-centric circular organization appear to move at a significantly faster pace than those that do not. This accelerated operational speed means they can potentially operate with fewer personnel, execute much shorter planning cycles, and transform what were once considered 'one-way door' decisions into more flexible 'two-way door' decisions, enhancing agility and responsiveness.

Is it's a circle, and in the middle is all of the context in the AI system, and initially humans are making all the decisions, but over time as you give that, that kind of centralized AI system more context, it gets smarter, and you're feeding it context, and you turn over more and more of the decisions to them.