The Hidden Costs of Startup Failure

On April 6, 2020, Jeffrey Katzenberg and Meg Whitman launched Quibi, a streaming service designed for the little gaps in the day. Ten-minute shows for the subway, a coffee line, the moments between things. There was a problem. Much of the world had just been ordered to stay home.

Quibi had raised $1.75 billion. It would spend a reported $63 million on television, digital and print advertising in 2020. Yet its app initially made sharing clips or screenshots difficult, even as social sharing helped people discover competing entertainment. On October 21, barely six and a half months after launch, Quibi announced it was closing. It said it would return $350 million to investors and sell its assets. The company had employed 265 people that April.¹

The pandemic mattered. So did the larger question Quibi never answered convincingly: Would enough people pay for brief videos on a new service when they already had plenty of ways to watch them for free? More advertising could make Quibi familiar. It could not make the underlying habit valuable.

For a quarter-century, startup culture has repeated a tidy instruction: fail fast. It sounds almost cheerful. Nobody puts the second half on a conference banner: then tell the people who believed you that the money is gone, the jobs have ended, and the years will not be coming back.

In 2025, investors put more than $500 billion into venture-backed companies worldwide, according to KPMG. That is an astonishing amount of money. It is not an estimate of what those companies lost, or will lose. Some investments are still operating. Some will return a fraction of their cost. A few may return many times their cost. The financial outcomes unfold over years, not within the calendar in which the checks were written.²

Then there is the famous claim that 90% of startups fail. Fail by when? Fail according to whom? Is a quiet shutdown equivalent to a company that survives but never repays its investors? What about a business sold for less than its founders hoped?

The U.S. Bureau of Labor Statistics found that 34.7% of private-sector business establishments born in 2013 were still operating ten years later. That says something about business survival, but those establishments are not a cohort of venture-backed startups. Harvard Business School professor Tom Eisenmann uses a different measure: more than two-thirds of startups never deliver a positive return to investors. Both figures are useful. They are not interchangeable.³⁴

AngelList studied 1,115 money-losing venture investments. Among those losing deals, the median investor lost the entire principal. The average loss was 81%. That is a sharp picture of what happens when an individual bet goes wrong. It is not evidence that 81% of every year’s venture funding disappears. The study deliberately examined investments that had already lost money.⁵

You cannot multiply a broad startup failure rate by the year’s funding total and call the product an annual loss. Yet that is the kind of arithmetic that keeps turning up in startup commentary. It produces a spectacular number and tells us remarkably little.

A defensible total for the past 25 years cannot be calculated from those headline figures. What we do know is that an immense amount of capital has been placed at risk. And money is the easiest part of the bill to recognize.

A venture fund can expect some companies to fail and build its portfolio around that expectation. The founder doesn’t have a portfolio of lives. There is one career, one household, one set of relationships, and a finite number of years in which to recover.

Christina Wallace left a well-paid job at Boston Consulting Group to start Quincy Apparel with her Harvard Business School classmate Alex Nelson in 2011. They raised $1 million and built a line of women’s workwear with a different approach to fit. Customers responded. The problem, Wallace later explained, was production: small runs, complicated operations, tight turnaround times, and margins that could not carry the model. Quincy shut down in January 2013.⁶

Wallace had gone eight months without pay to start the company. She had no family or partner to fall back on financially. After the shutdown, she talked about the limits of her emotional support system and the work of separating her identity from a company that had failed. None of that appears in a venture fund’s write-off.⁶

The strain does not begin on closing day. In a 2023 Startup Snapshot survey of more than 400 founders, 72% reported that entrepreneurship had affected their mental health. Thirty-seven percent reported anxiety. Eighty-one percent said they were not really open about their stress and fears, and 77% did not seek professional support. These figures describe the pressures of being a founder, not the psychiatric consequences of failure itself. That distinction matters. They show how much can be at stake before anyone turns off the lights.⁷

Employees absorb their own losses: interrupted careers, vanished equity expectations, sudden job searches. Suppliers and contractors can lose expected work. Families have to live with decisions made in pitch meetings they never attended. Some of these costs can be counted. Others are recorded in postponements: a move that never happened, a degree delayed, savings that no longer exist.

There is, sometimes, a way back. Research discussed by Harvard Business School found that entrepreneurs leaving an initial venture-backed failure often landed jobs roughly three years more senior than comparable peers. Employers evidently value the broad operating experience. But the finding concerns a particular population, not every failed founder. A better job later is not proof that the earlier loss was painless.⁸

Quibi and Quincy are worth putting next to each other because they failed in different ways. Quibi struggled to make its proposition compelling enough for paying viewers. Quincy had customers who wanted the product, but the machinery of producing and delivering it did not work economically. Demand and a viable business are related. They are not the same thing.

This is where marketing has a job much larger than buying ads. In too many startups, marketing is hired to create demand for something the company has already spent a fortune building. The founder hands over a finished proposition and asks for reach, leads, awareness, conversions. If the numbers disappoint, the response is often to adjust the copy or spend more.

But the early marketing questions are harsher. Who has this problem badly enough to pay? What are they doing instead? Do they return when the introductory offer ends? Does each sale leave enough money after acquisition, delivery, service and support to make another sale worth pursuing? Does the business still work when the founder stops doing everything for free?

Quibi could buy awareness at scale. Reports at the time describe $63 million in ad spending and an internal fight over whether to sell the brand or individual shows. Its basic commercial problem remained. And Quincy shows the opposite trap: a product can find its customers and still be defeated by the cost of serving them.⁹¹⁰

The point is not that a marketer can prevent every shutdown. Nobody can. It is that a founder can usually make ignorance cheaper. Put the riskiest assumption in front of real customers while the company is still small. Ask for the purchase, not the compliment. Watch what people do after the first week. Calculate margins with the boring expenses included. Decide in advance what evidence would trigger a change of course.

Every startup is a collection of hypotheses. Some will be wrong. The expensive mistake is allowing confidence, fundraising and publicity to turn an untested hypothesis into a permanent obligation.

The question to ask before the next round, the next hire or the next campaign is not simply, ‘How do we grow?’ It is, ‘What result would convince us to stop?’