babel

Business

Lessons from Great Founders

Bring the greatest founders’ hard-earned knowledge to your own business decisions, whenever you need it. Draw on researched choices, alternatives and outcomes to decide what to build, how to price it and how to grow.

Download INKO Early Access

What you asked PachinkoWe are twelve people making one product whose parts all have to fit, and I am about to draw the usual org chart with a head of product and a head of engineering. Is that the right shape for us?

Nvidia is organised like the thing it builds: Jensen Huang arranged the company as a computing stack, with decisions heard by all at once

Your company is growing and you are about to draw the usual chart: divisions, each with a head, with information passed down through the layers. You make one complicated thing whose parts must fit. You are deciding whether the usual chart suits it. Start from the product. List its layers and modules, and put whoever is best at running each one in charge of it, whatever their title. Define the unit of work as a concrete mission, such as shipping a named product, and assemble for it the best people from anywhere in the company. Make decisions in meetings that include the junior people doing the work, so that they hear the reasoning and the decision when the executives do. A manager then gains nothing from knowing things first, and your leaders must earn their positions by reasoning well and helping others succeed. Read other companies' methods for what they mean in your circumstances; do not adopt them whole.

Jensen Huang has run Nvidia since founding it in 1993. On the Acquired podcast in 2023 the hosts put it to him that he has more than 40 direct reports and an unconventional chart, and asked what was special about Nvidia. "I don't think it's something special in Nvidia. I think that we had the courage to build a system like this. Nvidia's not built like a military." The company is "built much more like a computing stack. The lowest layer is our architecture, then there's our chip, then there's our software, and on top of it there are all these different modules. Each one of these layers of modules are people." Title has nothing to do with position in the stack. Whoever is best at running a module "is in-charge. That person is the pilot in command."

He had thought of it that way from the start, he said, "because your organization should be the architecture of the machinery of building the product. That's what a company is. And yet, everybody's company looks exactly the same, but they all build different things. How does that make any sense?" Frying chicken, flipping burgers and making fried rice are different processes; he saw no reason their organisations should match. The standard chart of separate business units made no sense to him.

Work is assigned by mission. "There's a phrase in the company called 'mission is the boss.'" A mission is concrete. When the hosts guessed at something like advancing accelerated computing, he corrected them: "Build Hopper," the name of a chip generation, or build the system for it, or build CUDA for it. The best skills and teams are wired to each mission across the whole organisation, "like a neural network."

He named the cost. "The downside is the pressure on the leaders is fairly high." In a command system a superior has power because he is closer to the source of information. At Nvidia information goes to teams. He had just left a robotics meeting attended by new college graduates, three vice-presidents and two members of his executive staff. "We made a decision, everybody heard it exactly the same time. Nobody has more power than anybody else." Leaders keep their jobs by their ability "to reason through problems and help other people succeed," and privileged information plays no part.

The hosts observed that Nvidia ships flagship products at a pace they could not imagine from other large technology companies. In the fiscal year to January 2024 its revenue was $60.9 billion.

Sources: Jensen Huang, interview, Acquired podcast, 2023, transcript at acquired.fm; Nvidia, Form 10-K for fiscal 2024.

Put the greatest founders on speed dial.

In the spring of 2020 Brian Chesky watched Airbnb lose 80 percent of its business in eight weeks. The company he had to save was no longer the one he had started. It had spread into transportation, travel content, hotels and luxury stays, and by his own account it ran like ten divisions. He cut the initiatives and pulled what was left into one company organized by function. Then came the harder question: how would one company decide things? [1]

  • The operating model
    Steve Jobs, through Ive and Asai

    Bring design, engineering and marketing back together.

    He did not have to guess. Before the pandemic, Jony Ive and Hiroki Asai had sat with him and described how Steve Jobs ran Apple: the decisions behind the products Chesky admired, told by two people who had been in the room. What he took from them was a shape for the company, with design, engineering and marketing pulled back into one whole. [2]

  • The adaptation
    Brian Chesky

    Keep the decisions close to the work.

    Chesky put every project on one roadmap and reviewed the work himself, on a schedule, with the dependencies between projects settled in the room. His executives had to know their disciplines well enough to judge the work in front of them. The cost arrived at once: more decisions came back to him, and the model only holds with leaders who have the expertise and the hours to make them. [2]

  • The shareholder letter
    Jeff Bezos

    Explain what makes the company different.

    Later that year a different problem sent him to a different founder. Airbnb was going public and Chesky had to write the founders’ letter. He read Amazon’s 1997 shareholder letter, and what held him was the way Bezos explained what made the company distinctive and how he intended to run it. From that example Chesky worked out what Airbnb should commit to its own investors. [1]

  • Your next move
    The experience your business needs

    Match the conditions before borrowing the decision.

    Two ways to borrow a founder’s judgment, then: hear how a decision worked from the people who lived it, or read the document in which the founder explained it. Both begin with your problem. A shared roadmap suits teams building one connected product; a business with independent customers and operations may need its authority spread out. Before you borrow the decision, find out what it depended on and whether you have those conditions.

Lessons from Great Founders puts that kind of account on your desk: a library of researched notes filed in 77 sections, each one a founder’s decision with the situation, the alternatives, the choice, the cost and what followed, and the sources named. Ask Pachinko about the problem in front of you and the notes that fit come back, whether you are setting a price, hiring a leader or deciding what to build next.