The Fail Fast, Fail Often Myth: Failing Doesn’t Teach You Anything On Its Own

A “failure resume” sounds like self-awareness. The actual data on founders says the resume that predicts your next win is the one with a success already on it.

Startup culture has spent the last decade and a half turning failure into a personality trait worth listing on a slide. Failure conferences, complete with their own circuit and keynote speakers. Failure resumes, circulated proudly at networking events, one line per venture that didn’t make it. Founders opening pitch decks with the venture that didn’t work, framed as evidence of grit rather than evidence of, well, a venture that didn’t work. The underlying claim is rarely stated this bluntly, but it’s there in every “fail fast, fail often” T-shirt: failing quickly and repeatedly is itself the mechanism that produces eventual success.

Here is the myth: that failure is inherently instructive, that failing faster means learning faster, and that a founder’s collection of failed ventures functions as evidence they’re due for a win.

Here is the bust: large-scale data on actual venture-backed founders doesn’t support that comforting story. Prior failure barely moves the needle on your odds next time. Prior success moves it enormously. The two are not remotely the same predictor, and the mantra quietly assumes they are, without ever citing the numbers that would actually back it up.

Where the Mantra Comes From

The modern version traces to Eric Ries’s 2011 book The Lean Startup, which proposed a specific, structured methodology: build a minimum viable product, measure how real customers actually respond to it, learn from that data, and repeat the loop quickly, at low cost, before committing serious capital to an unvalidated idea. The whole point of the cycle was to shrink the cost of being wrong, not to celebrate wrongness for its own sake. “Failure” in Ries’s original framing wasn’t a virtue in itself — it was the acceptable byproduct of testing a hypothesis cheaply, fast enough that a wrong guess cost you weeks instead of years, with a specific next experiment already queued up to test the lesson.

That distinction did not survive contact with startup Twitter. Over the following decade, “build, measure, learn, and pivot quickly” got sanded down into “fail fast, fail often,” a slogan that keeps the speed and drops the methodology. The compressed version is easier to put on a poster and considerably easier to follow without doing any of the actual measuring.

What the Data Actually Shows

In 2010, Harvard Business School researchers Paul Gompers, Anna Kovner, Josh Lerner, and David Scharfstein published a study in the Journal of Financial Economics examining thousands of venture-capital-backed companies founded between 1986 and 2003, tracking what happened to their founders’ subsequent ventures, with success defined as reaching an IPO or filing to go public. The question was direct: does a founder’s history of success or failure predict what happens next, once you control for the obvious confounders?

Founders with a prior successful venture went on to succeed 30% of the time in their next one. First-time founders, with no track record at all, succeeded 21% of the time. Founders whose prior venture had failed succeeded 22% of the time — statistically indistinguishable from someone who had never tried before, once you account for normal variation in the data. Failing once didn’t erase the odds of a first-timer. It barely moved them, despite everything the failure was supposed to have taught along the way.

Not All Failure Is Created Equal

Harvard Business School professor Amy Edmondson, who has spent decades researching psychological safety and organizational learning, argues the mantra’s real flaw is that it treats all failure as the same substance. In her 2023 book Right Kind of Wrong, Edmondson lays out three archetypes: basic failure, a preventable slip from a known process; complex failure, caused by multiple factors converging in a system nobody fully controls; and intelligent failure, the result of a genuine hypothesis, carefully tested in new territory, that simply turns out to be wrong.

Only the third kind reliably produces learning, and it requires a specific structure to do it: a real hypothesis stated in advance, a test designed to actually check it, and a review of what the result implies. Most “failing fast” inside an actual startup looks much more like the first kind — the same avoidable mistake, made with more confidence and at higher volume, mistaken for iteration because it happened quickly. Speed was never the ingredient doing the work. The structure around the speed was, and the slogan dropped it.

Why the Myth Is Attractive to Founders and Investors Alike

None of this means destigmatizing failure was a bad idea. Founders who fear failure too much won’t take the risks that occasionally pay off spectacularly, and a culture of blame makes people hide problems instead of surfacing them early, which is worse for everyone involved, including investors trying to see a portfolio company’s real health. That’s a real, defensible reason “fail fast” caught on, and it’s worth keeping. The trouble is what got bundled in alongside it: the further claim that failure itself, independent of what caused it or what was learned, functions as career progress on its own. That version is much more convenient for everyone involved — it lets founders reframe a shut-down company as a chapter in an inevitable success story, and it lets investors treat portfolio losses as the expected cost of a strategy that’s “supposed” to look like this, rather than asking harder questions about what specifically should change before the next check gets written.

Why This Myth Stuck

None of this is really about startups specifically. A failure resume is a better story than a spreadsheet of base rates, the same way a moment of founding clarity beats years of unglamorous skill-building in the myth next door. Survivorship bias does the rest of the work: we hear from the founders who failed once and then succeeded, and we don’t hear nearly as much from the much larger group who failed once and then failed again, because that story doesn’t get invited to speak at a conference, and it certainly doesn’t make for a compelling opening slide.

What To Actually Do With This

None of this means you shouldn’t take risks, start things that might not work, or talk openly about what went wrong. It means failing quickly is only valuable when it’s structured enough to actually produce information, and a resume of losses isn’t the same asset as a resume of wins, however the slogan makes it feel in the moment. The founders the data actually favors are the ones who treat a setback as a specific, answerable question, not the ones who simply rack up the most attempts.

• Don’t treat “fail fast” as license to skip structured testing — the value comes from the method, not the failure itself.
• Distinguish preventable mistakes (the same error, repeated) from intelligent failures (a genuine hypothesis, tested and disproven).
• Track what specifically changed in your approach after a failure — if nothing did, no learning actually occurred.
• Don’t assume failure alone builds a track record that predicts your next success; a history of wins does that far more reliably.
• Seek mentorship or partnership with people who’ve actually succeeded before, not only people who’ve failed before.
• Treat “fail fast” as a call for faster, cheaper tests — not permission to be careless with the same mistake twice.

Frequently Asked Questions

Related Myths We’ve Busted

Gear Mentioned in This Article

As an Amazon Associate, Sarcastic Motivators earns from qualifying purchases.