"Founder Mode" shouldn't be a Rorschach Test
Paul Graham coined Founder Mode over the weekend, but he left "what is founder mode" and "how to do it" as an exercise for the reader.
That has left room for the term to become a Rorschach test... random observers have shared their own views on how to be a founder or run a company, co-opting "Founder Mode" to justify it.
This is my attempt to define and contrast Founder Mode and Manager Mode in a concrete, falsifiable way. I’ve tried to limit my own takes, instead distilling source material from Brian Chesky, Paul Graham, and Steve Jobs:
Founder Mode means leveraging a founder's context & moral authority to create urgency, clarity, and conviction across the organization.
Founder Mode isn't "better"; it's different. It depends on what an organization needs: roughly speaking, Founder Mode vs. Manager Mode is a tradeoff between value creation and value preservation. Startups are obviously better off in Founder Mode, while your local power grid needs Manager Mode. Given how single-threaded Founder Mode can be, it is critical to set corresponding expectations with your team which you abide by, in order to avoid a toxic culture and mass revolt.1
Defining Founder Mode based on source material
Brian Chesky defined a good bit of the underlying concept on his podcast episode with Lenny last November. Here's a summary of the critical points:
When a crisis (pandemic) happened, people wanted Brian more involved.
He ruthlessly cut scope – not one team on 3 things, but 3 teams on one thing. This made it possible to centralize the roadmap into a single living document, with 30-40 senior people engaging on it via a single continuous conversation.2
He enforced a rule that managers must operate as player-coaches. Leaders must be in the detail. (If you're not in the detail, how can you measure work or create accountability?)
The CEO must review all the work, which enabled attribution and created accountability
His overall realization is that everyone really just wants clarity, to row in the same direction really quickly, and he set up systems to enable this.
In another interview, he referenced the founder's contextual and moral authority to be more decisive: being the "biological parent," "having permission to make change", and "knowing [how it works and hence] how to rebuild the company."
Jobs (heavily referenced by both Chesky and Graham) references singleton owners at Apple for each major part of the company: iOS, hardware engineering, etc., and that they get together to talk about the whole company, holistically, for 3 hours. "We all work on the same thing, and bring it together as a product."
Graham in his post also describes a few other characteristics that seem to describe Founder Mode:
Avoiding treating parts of the organization as black boxes that are delegated away completely
Lots of discussion outside the reporting line (i.e., skip-levels and more)
The DNA of companies (and their leaders) should dictate how delegation and autonomy works
The concept will get misused by people who want to co-opt it (already happening!); Founder Mode increases upside but also increases potential damage caused by a bad leader
The resulting definition from the common themes:
Increase speed and clarity by centralizing decision-making (with the CEO + small group of other people); by contrast, consensus is slow and sclerotic.3
Maximize context for fewer decision-makers, who are "in the details" (a player-coach), including the CEO. Create a single, simple context where decisions are made by that small group (one meeting, one roadmap, etc.).4
Avoid excessive delegation, because it hurts centralizing context, slows down decision-making, and eliminates accountability.
Tailor to the DNA of the company and its leaders, instead of installing "best practices."
So what’s Manager Mode, and why is it common?
The more mature a company, the more "responsibilities" it has: customers, revenue, employees, shareholder value... "Move fast and break things" doesn't scale. So it's natural that a company should be more deliberate and "safe" with decisions.
The obvious – and lazy – way to enforce safe decisions is by consensus. This is what Managers usually deploy – committees, stakeholders, buy-in, etc. In a word, bureaucracy.
So the characteristics of Manager Mode are:
Decrease the rate of change and prefer maintaining the status quo, by creating more dependencies; this reduces the risk of bad decisions (i.e., false positives); by contrast, Founder Mode causes rapid change that is hard to check.
Decentralize decision-making to reduce the reduce the "bus factor", i.e., dependency on a few linchpins within the company. This makes the organization more resilient than Founder Mode, which relies on a few people sticking around.
Delegate heavily, since larger organizations have more resources to decentralize the work, and delegation involves creating process – which standardizes and reduces risk with the work being done.
Create a more generic management structure, through best practices; this enables anyone to step into a company and pick up where the last person left off.
It’s a game of tradeoffs
Manager Mode exists because bureaucracy has its uses; it creates checks and balances and limits the damage any one person can do.
The tradeoff is that it's slow, and nothing new and amazing ever got designed by committee. Speed is often a critical input into being an elite, high-functioning organization.
Founder Mode is riskier for Value Preservation, but is more likely to produce Value Creation.
Manager Mode is better at Value Preservation, and bad at Value Creation.
Endorsing Founder Mode is making a choice to trade off stability for potential upside; it is picking a side in the Innovator's Dilemma.
Of course, these are broad strokes; there are examples and counter-examples for each:
Jobs, Zuckerberg, Bezos, and Musk are clear evidence of building multi-trillion dollar companies on Founder Mode. Specifically, they each built multiple groundbreaking products or businesses in rapid sequence because of their speed, conviction, and clarity. That's not a fluke.5
IBM is an example of a company on Manager Mode that continues to sustain – while they're no one's idea of an innovative company any more, they're still generating billions of dollars in new revenue off AI. The famous phrase "nobody got fired for buying IBM" is about as succinct an argument for Manager Mode as I've ever heard.
Boeing has been in Manager Mode for decades, but ended up destroying brand and value. This is because Manager Mode lacks “central planning”; when left to itself for long periods of time creates layers of bureaucracy, cruft, and misaligned incentives.
Adam Neumann and Elizabeth Holmes are prominent examples of Founder Mode leaders who wrote checks that reality couldn't cash. They were, unsurprisingly, notorious for being mercurial and unpredictable, and even idolized Founder Mode people (Musk and Jobs respectively).
For a company that is "too big to fail", or has to prioritize serving existing customers the same way, or providing a public good with minimal risk... Manager Mode is objectively better.
But it's clear why Founder Mode is the most likely path to produce new things: it's inherently a higher variance route, which produces either moonshots or craters. The best practitioners of Founder Mode can add fuel to the moonshot (sometimes literally, like Musk), and minimize the radius of the crater. But they don't shy away from some rate of failure.
It's not an accident that the people most likely to endorse Founder Mode are techno-optimists, and any other way seems like anathema. Techno-optimists believe that change is intrinsically good.
The best practitioners of Founder Mode are incredibly rare; most leaders (founders or otherwise) probably aren't capable of it. It's even harder to find someone who can practice it at 10 people and 10,000 people. This is why most companies default to Manager Mode. But if a company has one of those rare generational leaders – like Brian Chesky – it's a disservice to everyone to "regress to the mean" too soon.
It's all tradeoffs. I recognize the fact that stability and continuity has its value in many organizations. But I do think too many companies switch to Manager Mode before they should, and never rediscover that spark that made them successful in the first place. I wish we had more thoughtfully practiced Founder Mode companies – whether it's a founder at the helm or not.
Thanks to Amar Sood, Noah Itovitch, and Barbara Pascetta for reviewing and providing inputs to drafts of this post.
The worst scenario is using Founder Mode to micromanage or create a toxic culture, which is why setting expectations and working norms with the leadership team and organization at large is critical. Parachuting in unexpectedly is bad; but setting the expectation that the CEO is final authority on X/Y/Z is totally fine. It's also awful and obviously counterproductive to move the goalposts after the fact, be a jerk, or to ignore the smart people around you.
In analyzing Founder Mode, it helps a lot to internalize Price's Law: the square root of the number of contributors to a project will produce ~50% of the total output. This means you can get a read on the pulse of the entire organization with shockingly few people (e.g., at a 10,000-person company, you can get surprisingly high context by just polling or involving the right 100 people rather than everyone).
Coinbase deciding that every hire will have to be centrally approved is a clear example of Founder Mode at work – note that the decision making includes not just Brian but Emilie the COO – who is not technically a founder but has been recognized with the moral authority to make these one-way decisions.
Scott Belsky wrote about "collapsing context" which is directly in line with Founder Mode practices.
Founder Mode doesn't have to require a founder; I'm a big believer in ex post facto cofounders and of course you have to include more people as the company scales. But I think it does require people of a certain archetype: forces of nature with great instincts, with a bias towards action, who have loads of context & moral authority. Non-founder CEOs have a harder time with this because they haven't earned the context or the moral authority to exercise this conviction and control. However, founder-CEOs are treated as ipso facto competent, because they built the company to that stage to begin with, and they naturally have all the context. I don't think it's an accident that Satya Nadella, who is the closest thing to Founder Mode in a non-founder CEO lately, came up through the Microsoft org over decades – he's earned the context and has the credibility of being a "lifer." Costco and some other companies notoriously promote from within, and I think these people inherit Founder Mode the same way.