The Confidence Gauge Is Broken, Not You.

The Assumption

Every founder has heard some version of this: self-doubt and overconfidence are two flaws at opposite ends of a spectrum, and the job is to find the calm middle — "grounded realism." Coaches sell frameworks for it. Books promise the fix. The idea underneath all of it: measure yourself accurately enough, and you graduate out of the swing and land somewhere stable.

Unfortuantely, this doesn't survive the contact with what actually happened at three companies and their founders that, on paper, had every reason to know better.

The Inversion Test

Here's the arc most founders actually live, whether or not anyone names it for them. Early on — first prototype, first customer, first hire — self-doubt tends to dominate. Everything is new, you have the least evidence, and the gap between what you're projecting and what you privately know is at its widest. That's the classic imposter feeling. A little further in — seed funding closed, a small team in place, some real signal from the market — is roughly where grounded realism becomes possible, because enough evidence now exists to check belief against reality, if anyone still bothers to. Later — Series A, Series B, a board, press attention — is where overconfidence risks running unchecked. Not because the founder changed, but because the people around him stopped saying things he didn't want to hear.

It doesn't move as one clean dial, either. A founder can feel unshakeable about the product the same week he's quietly unsure whether he can actually sell it. The three states aren't a personality type you settle into. They're stages nearly every founder passes through, in different areas of the business at different times — and what decides the outcome isn't which one you're "prone to." It's whether anything can still reach you when you need it to.

The Reality Check

That early-traction feeling — the first yes, the first working prototype — is the part every founder recognizes, and it's exactly where this piece starts. What we want to trace is what happens to that same confidence after it survives long past that first moment: through a seed round, through hiring, through a Series B, without anyone ever forcing a recalibration in between. Three companies show that escalation clearly — and by the time each story reaches its critical moment, it no longer has anything to do with a prototype or a first customer.

Three Observations from the Survivor Bias Excavation

WeWork. Adam Neumann co-founded the company in 2010 as a straightforward business: lease office space long-term, redesign it with better furniture and free coffee, sublet it short-term to freelancers and startups at a markup. Neumann insisted, from early on, that WeWork was a technology company, not a real estate one, and investors mostly went along with it. By early 2019 that story had produced a $47 billion valuation, one of the highest of any private company in the world. That same year, the IPO filing exposed the underlying numbers, the offering collapsed, and the board removed Neumann as CEO.

The usual version of this story ends there: hubris, crash, lesson learned. The more useful question is what let this particular confidence run for nine years without correction. Behavioral finance research on CEO overconfidence gives one part of the answer: an overconfident executive making big bets isn't unusual by itself — it becomes catastrophic only once paired with enough capital to act on the bias at real scale, which Neumann didn't have until SoftBank's Masayoshi Son committed several billion dollars to the company. Confidence was the constant. What changed was who was willing to fund it without asking hard questions.

Katerra. A less-cited case, and a different mechanism. Katerra, founded in 2015, aimed to manufacture building components off-site the way a factory builds cars. Its CEO, Michael Marks, wasn't a first-time founder chasing a dream — he was the former CEO of Flextronics, a board veteran of companies like SanDisk, exactly the profile investors are trained to trust. After raising roughly $2 billion, mostly from SoftBank, Katerra filed for bankruptcy in 2021. Reporting at the time described a board that had reportedly received misstated financial reports, and investors who kept extending trust specifically because of Marks's (Katerra) track record — one investor's reasoning, on record: a person with this résumé can't be the one missing the problem. Here the miscalibration wasn't Marks's own confidence. It was everyone else's confidence in his confidence, standing in for scrutiny that should have happened anyway.

Peloton. The case that keeps this honest, because it breaks the "SoftBank ruins everything" shortcut — no Vision Fund money here at all. During the pandemic, CEO John Foley (Peloton) scaled manufacturing hard to meet a demand surge, including a $400 million U.S. factory announced in May 2021 — months after Peloton's own 2020 annual report had already flagged that pandemic-driven demand might not hold. It didn't. Peloton later posted a $2.83 billion net loss for the fiscal year, laid off thousands of employees, and Foley stepped down. On the earnings call, he said the company had scaled too rapidly and overinvested, and that he owned it. Same pattern as WeWork and Katerra — a demand curve and a story extrapolated past their shelf life — but Peloton restructured under a new CEO and kept operating. It never filed for bankruptcy.

The Forensic Analysis

Here's what we actually want to know — not who's to blame, but what was happening inside each founder as the confidence kept climbing. All three plausibly started with the normal mix: real confidence, and some private uncertainty about whether they were really right. What none of them got, as the stakes grew, was any signal back for the doubting half. Not because anyone conspired to silence it — because nobody around them needed that doubt to surface for their own position to keep working. Investors, boards, employees holding equity: all better off, in the moment, if the confident story kept being true.

That the doubt existed somewhere is not really in question. Foley said as much afterward, in his own words, on the record. The harder question is why it didn't surface before the damage — because the founder himself found it inconvenient to sit with, or because the world around him gave him nothing to check it against: no counter-voice, no data point, no moment where raising it wouldn't have felt like betraying a story he'd already sold to everyone, including himself.

The Counterfactual Scenario Builder

The three scenarios below aren't about who should have stopped the story — a stronger board, a tougher investor. They're about whether the founder's own doubt, in each case, ever had anywhere real to land.

Scenario A — WeWork: Was there a point where Neumann's own private doubt, if it existed, had a genuine chance to surface — or had the story already outrun any way, internal or external, of testing it?

Scenario B — Katerra: Marks's confidence was partly borrowed from his own résumé, not from the evidence in front of him that quarter. Is trusting your own track record instead of the current, specific facts a different failure than plain overconfidence — one that looks like humility from the outside, because it isn't about ego at all?

Scenario C — Peloton: This is the one with a paper trail. Foley's company had written the doubt down — in the 2020 annual report, as a stated risk — before the $400 million factory commitment went ahead anyway. The doubt wasn't missing. It was on the record, in black and white, and it still lost to the confident read. What does it mean that having the doubt in writing wasn't enough?

What the Research Doesn't Cover

Three founders, the same basic mechanism but each time a story that outran its evidence, with the founder as its author and its most consistent performer. We use that role, storymaster, as our marker for what gets called overconfidence (Dunning Krueger Effect). Whether it was performative or genuinely felt, we'll never know from the outside. What we do know is that the three endings diverged sharply — engineered collapse, negligence-enabled collapse, survived collapse — based on factors that had nothing to do with how convincingly the story was told: who was funding it, at what scale, and whether anything in the room had any reason to push back.

Entrepreneurial overconfidence is a well-studied field — but almost all of it works from outside signals: valuations, public statements, outcomes after the fact. The closest adjacent work, a peer-reviewed critique in the Review of Managerial Science, argues that researchers often call founders "overconfident" using only outside information, when the founder may be acting on private information nobody else in the room has. It's a fair point about how outsiders judge founders — but it doesn't quite fit what Peloton shows us. Foley's doubt wasn't private information he was quietly sitting on. It was written down, shared, on record — and it still lost.

What we don't have — and what the field hasn't yet produced — is research into what happens in the room, in real time, before the story hardened: the moment when a founder's own documented doubt met his own confidence, and one of them won. We only have the aftermath. That gap is what this piece is pointing at, and it's why we're asking.

Open Questions

• Right now, which stage am I in on a given decision — the one where I don't yet have enough evidence to trust myself, or the one where I have the evidence and I'm choosing not to look at it?
• Is there something I've already written down somewhere — a risk note, a line in a deck, a comment to myself — that already contains the doubt I'm not acting on?
• Who in my circle is close and competent enough to see a crack forming, but has no stake in whether I hear about it — and when did I last actually ask them?
• The last time someone raised a doubt about my direction, what did I actually do with it in the next five minutes — sit with it, or find the fastest reason it didn't apply to me?
• A year from now, looking back at this decision — am I telling the story in a way that makes me look right regardless of the outcome, or in a way that would have actually warned me before it was too late?

PS. We reconstructed all three cases from the public record — filings, statements, reporting after the fact. What we don't have is what these three men were actually telling themselves in the moment, before the story hardened into what everyone now calls a lesson. If you've sat inside a moment like this yourself — doubt on the table, confidence winning anyway — we'd genuinely like to hear about it. Write us!

Sources & further reading

• Chwolka, Anne, and Matthias G. Raith. "Overconfidence as a Driver of Entrepreneurial Market Entry Decisions: A Critical Appraisal." Review of Managerial Science, vol. 17, no. 3, Apr. 2022, pp. 985–1016. Springer, https://doi.org/10.1007/s11846-022-00552-6.
• Farrell, Maureen, and Eliot Brown. The Cult of We: WeWork, Adam Neumann, and the Great Startup Delusion. Crown Currency, 2021.
• Primack, Dan. "Katerra's Collapse Is a Failure of Pattern-Matching." Axios, 8 June 2021, https://www.axios.com/2021/06/08/katerra-bankruptcy-softbank.
• Thomas, Lauren, and Sara Salinas. "Inside Peloton's Rapid Rise and Bitter Fall — and Its Attempt at a Comeback." CNBC, 19 Feb. 2023, https://www.cnbc.com/2023/02/19/peloton-rise-fall-attempted-comeback.html.

Destruction Desk
We perform autopsies on innovation’s failed assumptions.

This newsletter was edited by Manfred Lueth.

You received this email because you signed up for this newsletter from DestructionDesk.com.
To stop receiving this newsletter, unsubscribe or manage your email preferences