ShopShop

Home Startups Why Startups Fail: What the Post-Mortems and the Data Actually Show

Startups

Why Startups Fail: What the Post-Mortems and the Data Actually Show

Layla LevesqueOctober 7, 20266 Mins read0

Why Startups Fail: What the Post-Mortems and the Data Actually Show

Photo: Empty office meeting room via Rawpixel, CC0

On December 27, 2024, Bench Accounting, a Vancouver company that had raised more than US$100 million, told thousands of small-business customers it was shutting down, effective immediately. GeekWire reported that nearly 500 employees were affected. Within days, a buyer had stepped in and rehired many of them. It was a dramatic end to a company that, three years earlier, had raised a US$60 million Series C.

Stories like Bench’s get attention because they are sudden. Most startup failures are not. They unfold over months, sometimes years, through signals that founders and investors can see if they know where to look. The trick is separating the cause of death, which is almost always “ran out of money,” from the conditions that made the money run out.

We looked at what the data actually says: CB Insights’ latest analysis of venture-backed shutdowns, US government survival statistics, and Statistics Canada’s figures on how long Canadian businesses last. The picture is less bleak than the “90% of startups fail” line suggests, and more useful.

What the survival data shows

Start with the broadest numbers. These cover all new businesses, not just venture-backed tech startups, but they set a baseline.

In the United States

The US Bureau of Labor Statistics tracks every cohort of new establishments through its Business Employment Dynamics program. Its survival table, updated with data to March 2025, shows a remarkably consistent pattern:

Of establishments that opened in the year ending March 2015, 79.6% were still operating a year later.
After five years, 50.2% survived.
After ten years, 34.7% survived.
Of the cohort born in the year to March 2005, 19.5% were still around twenty years later.

The first-year rate barely moves from cohort to cohort. Establishments that opened in the year to March 2024 had a one-year survival rate of 77.9%. Roughly one in five new US businesses does not see its first birthday, in good years and bad.

In Canada

Canadian numbers look rosier at first glance, but they measure something different. Innovation, Science and Economic Development Canada’s Key Small Business Statistics 2025, which draws on Statistics Canada’s National Accounts Longitudinal Microdata File, tracks employer enterprises with 1 to 99 paid employees created between 2001 and 2022. The survival rates:

Years after creation	Share still operating
1 year	94.8%
2 years	86.9%
5 years	68.0%
10 years	48.2%
21 years	24.2%

Two caveats. The Canadian data counts enterprises with employees, so many solo ventures that never hire are excluded. The BLS counts establishments. The figures are not directly comparable, but both point the same way: roughly half of new businesses are gone within about a decade.

One detail from the ISED data stands out. Size at birth matters. Businesses that started with 1 to 4 employees had a five-year survival rate of 63.0%, while those that started with 20 to 99 employees reached 75.0%. Starting bigger is not always possible, but it is a reminder that undercapitalization is a risk from day one.

The current climate is also worth noting. Statistics Canada’s monthly estimates for February 2026 put the business closure rate at 5.0%, 0.4 percentage points above its 2015 to 2019 average, while business insolvency filings rose 10.1% from the month before.

What venture-backed failures look like

Venture-backed startups are a special case. They take on more risk by design, and investors expect most of them to fail or return little. In March 2026, CB Insights published an analysis of 431 VC-backed companies that had shut down since 2023, drawing on public post-mortems, founder interviews and shutdown announcements. The most cited reasons (companies could list several):

Ran out of capital: 70%
Poor product-market fit: 43%
Bad timing or macro conditions: 29%
Unsustainable unit economics: 19%

CB Insights makes a point that we think is the single most useful idea in this whole topic. Running out of money tops the list, but it is “almost always the final cause of death, not the root problem.” Cash is the symptom. Weak demand, bad timing or a business model that loses money on every sale is the disease.

The report also found that the failed companies had raised a median of US$11 million before shutting down. These were not tiny side projects. Some were very large: Convoy and Olive each raised around US$1 billion before closing in October 2023.

The patterns underneath

When you read enough post-mortems, the same handful of stories repeat. Here are the ones that show up most often, in our reading.

Building something few people need

Poor product-market fit is the classic root cause. Founders fall in love with a solution and keep building, mistaking polite interest or a few pilot customers for real demand. The symptom is a sales cycle that never gets easier and customers who do not come back.

Growing costs ahead of revenue

Paul Graham, the Y Combinator co-founder, wrote in his 2015 essay “Default Alive or Default Dead?” that “hiring too fast is by far the biggest killer of startups that raise money.” A fresh round makes it easy to staff up for growth that has not yet arrived. If that growth stalls, the burn rate becomes the clock.

Raising on the assumption of the next raise

Many shutdowns since 2023 involve companies that raised in 2021, when money was cheap, and planned on raising again within eighteen months. When interest rates rose and investors pulled back, the next round never came. Bench is an example. BetaKit reported that its last major financing was the 2021 Series C, that it had since closed two bridge rounds, and that, according to The Information, a lender called in its venture debt. An attempt to sell the company earlier that year had not produced a deal.

Debt with strings attached

Venture debt can extend a runway, but it comes with covenants. If the business misses targets, a lender may have the right to demand repayment at the worst possible moment. Founders sometimes treat debt as cheaper equity. It is not.

Timing and markets outside your control

Nearly three in ten failed companies in the CB Insights sample cited bad timing or macro conditions. Some ideas are simply too early, or depend on a funding climate or regulatory regime that changes.

How to spot trouble early

The useful thing about the CB Insights report is that it found measurable signs of decline before shutdown. In the year before they closed, 72% of the companies saw their Mosaic health score fall, by an average of 15%. Partnership activity dropped 44% compared with the prior year, and two-thirds were shrinking headcount in their final six months. Outsiders can see those signals. Insiders can see them much earlier. Here is a short checklist we would suggest founders run every quarter:

Default alive or default dead? Graham’s test: assuming expenses stay flat and revenue keeps growing at its recent rate, does the company reach profitability on the cash it has? If not, you need a plan that does not depend on a new round.
Is retention improving? New sign-ups can hide churn. Track how many customers from each cohort are still paying after six and twelve months.
Does each sale make money? Calculate contribution margin per customer after acquisition costs. If it is negative and not trending positive, growth makes things worse.
How many months of runway, really? Include debt covenants and committed spending, not just bank balance divided by burn.
Are partners and customers leaning in or out? Fewer renewals, slower pilots and quieter partners are early signs.

If two or three of those answers are bad at once, the time to act is now, while there is still money to cut, pivot or sell from a position of some strength.

Failure, in context

None of this means founders should avoid risk. Venture-backed businesses exist precisely because most will not make it and a few will make up for the rest. A founder who shuts down cleanly, pays staff and treats customers well has done something hard and honourable. Bench’s abrupt closure, with customers given little notice, is the opposite case, and that is part of why it drew so much attention.

There is also a Canadian dimension. Canada’s venture market is smaller and leans heavily on US capital (US investors supplied 56% of Canadian venture funding in Q2 2026, according to CPE Analytics data reported by Wealth Professional), which means the funding cliff can arrive faster here when American investors retreat. Founders who build a plan to reach profitability without the next round have more options. If you are weighing a restructuring, wind-down or sale, involve a lawyer and an accountant early; the obligations to employees, creditors and the CRA are real.

The takeaway

Startups rarely die of one thing. They die when a weak fundamental, usually demand or unit economics, meets a funding gap. Government data shows roughly half of all new businesses in both Canada and the US are gone within about a decade. The good news is that the warning signs are measurable. Know whether you are default alive, watch retention and margins more closely than top-line growth, and do not spend money on the assumption that more will arrive. That alone would have changed the ending for a lot of the companies in these post-mortems.