Since Paul Graham published “Founder Mode” in September 2024, founders have been handed a tempting new permission slip. Graham, drawing on a talk by Airbnb’s Brian Chesky, argued that the conventional advice to hire good people and give them room to work is often “disastrous” for founders, and that founders can do things managers can’t. The essay struck a nerve, and many founders concluded they’d been right to stay involved in every detail all along.
There’s just one problem with turning this into a leadership philosophy: Chesky himself rejected that framing. “I never called it founder mode,” he said days later. “I just described my experience.” The real question was never whether to be a founder-mode person or a manager-mode person. It’s which decisions deserve which mode, and the evidence on that is far more useful than the vibe.
The evidence cuts both ways, on purpose
Start with the case for staying deeply involved, because it’s strong. Bain’s research found that since 2015, founder-led companies outperformed their peers by 2.1 times in total shareholder returns. A study by Purdue professors found founder-CEO firms generated 31% more citation-weighted patents and made bolder bets. Founders carry conviction, context, and a licence to override consensus that hired executives rarely have. Bain promotes its own framework around founder leadership, so that context is worth keeping in mind. Even so, similar patterns appear in independent research.
Now the case for stepping back, which is equally well-evidenced. A study of more than 2,000 public companies found the founder advantage is real early but the premium dwindles to zero about three years after IPO, after which founder-CEOs actively start detracting from firm value. And Noam Wasserman’s research on startup founders is sobering about how this usually ends: by the third year, around half of founders are no longer CEO, fewer than 25% lead their company through its IPO, and roughly four out of five of those departures are forced, not chosen.
So involvement wins early and can poison later. Both conclusions can be true. A binary about founder personas can’t hold both. A framework about decisions can.
Classify the decision, not the founder
The most useful tool here comes from Jeff Bezos, who in his 2015 letter to shareholders split decisions into two types. Type 1 decisions are “one-way doors”: consequential and effectively irreversible, and they deserve slow, careful, deeply-involved judgment. Type 2 decisions are two-way doors: reversible, changeable, and best made fast by the people closest to them. Bezos’s warning was that organisations default to using the heavyweight Type 1 process on Type 2 decisions, which produces slowness, timidity, and a failure to experiment.
For a founder, that framing dissolves the whole founder-mode debate. Stay deeply involved in one-way-door decisions. Step back and delegate the two-way doors.
The one-way doors are surprisingly few, and they map closely to what the venture investor Fred Wilson argues a CEO actually does: set and communicate the vision, recruit and retain the key people, and make sure the company doesn’t run out of money. Add to that the irreversible, identity-defining product and strategy calls. These are where a founder’s conviction is an asset and delegation is genuinely dangerous, because a hired executive optimizing for consensus may unintentionally smooth out the very decisions that made the company distinctive.
Almost everything else is a two-way door. Which vendor to use, how to run a given campaign, the details of a feature that can ship and iterate, the process a team adopts. In these situations, founder involvement often isn’t diligence. It’s a bottleneck. It teaches your best people that their judgment isn’t trusted, which is how you lose them.
How to tell which door you’re standing in
The practical test takes less than a minute. Ask: if this goes wrong, how expensive is it to reverse? If the answer is “cheap and quick,” it’s a two-way door, and your job is to make sure a capable person owns it, not to own it yourself. If the answer is “we can’t easily undo this,” slow down and get involved, because that’s precisely the kind of call founders exist to make.
If I had to name the single most common founder mistake here, it’s applying one-way-door caution to two-way-door decisions, and calling it high standards. It isn’t high standards. It’s a growth ceiling built around your own calendar. The reverse mistake, delegating an irreversible strategy call to a committee, is rarer but more fatal, which is why the classification matters more than the instinct.
One honest caveat: the doors aren’t always obvious in the moment, and some two-way doors quietly become one-way as a company scales, when a “reversible” decision now affects thousands of customers or a regulated process. Re-classify as you grow. A decision you could safely delegate at ten people may need you again at two hundred, and part of the founder’s job is noticing when a door has changed.
The version that actually works
Graham was right that founders can do things managers can’t. Chesky was right that it isn’t about adopting a founder persona. Bezos offered the missing piece: focus on the door.
Stay ruthlessly involved in the handful of irreversible decisions that define what your company is, delegate the reversible many with genuine trust, and re-check the classification as you scale. The goal isn’t adopting a leadership persona. It’s making the right decision, one choice at a time. That’s the difference between a founder whose involvement compounds the company’s value and one whose involvement slowly becomes the thing holding it back.
Image by magnific
