Startup Failure Rate: What the 43% Really Means for Your Idea

Seventy percent of VC-backed startups cite running out of capital as their cause of death. Forty-three percent cite poor product-market fit. Most founders read these as two separate statistics pointing to two different problems. They are not.

Running out of money is how most startups die. Poor PMF is why. A company with genuine product-market fit can raise capital in nearly any environment — because investors bet on fit, not on ideas. A company without PMF burns through capital faster than its metrics justify, fails to make the case for the next round, and eventually closes. Capital failure is the downstream consequence of PMF failure in the vast majority of cases.

CB Insights’ 2024 update to its startup post-mortem database — covering 431 VC-backed companies that shut down after raising institutional capital — is the most cited primary source on this question. This article unpacks what the data actually shows, breaks the failure rate down stage by stage from idea to Series B, explains why the seed-to-Series-A graduation rate has halved since 2018, introduces the four PMF Failure Archetypes™ framework for diagnosing which failure pattern your idea most resembles, and maps a practical diagnostic roadmap by stage.

The Number Everyone Quotes Is Wrong

“90% of startups fail” is probably the most repeated figure in startup culture. It is also the least sourceable. No single authoritative study produces that number for any well-defined population.

The Bureau of Labor Statistics tracks all US employer businesses — restaurants, retailers, law firms — and finds 21.5% fail in year one, 48.4% by year five, and 65.1% by year ten. These are real numbers, but they describe a population that includes entirely different risk profiles from a VC-backed B2B SaaS company.

Harvard Business School researcher Shikhar Ghosh studied venture-capital-backed companies specifically and found approximately 75% fail to return investors’ capital — closer to the informal “most startups fail” narrative but still not 90%, and defined by a different criterion (investor returns rather than business closure). Source: “The Venture Capital Secret: 3 Out of 4 Start-Ups Fail,” Wall Street Journal (2012), reporting Ghosh’s research on 2,000+ VC-backed companies that raised at least $1M between 2004–2010; also covered by Harvard Business School’s own newsroom.

The figure that matters most for a pre-seed or seed founder is not the aggregate failure rate. It is the cause of failure — because not all failure causes are equally preventable, and the cause itself shifts depending on what stage you’re at. This is where the CB Insights data, read alongside Carta’s stage-by-stage graduation figures, becomes genuinely useful.

What the 2024 CB Insights Data Actually Says

CB Insights maintains a post-mortem database drawn from public records and founder post-mortems of VC-backed companies that shut down after raising institutional capital. The 2024 update covers 431 companies — a specific, well-filtered sample (companies that raised institutional rounds, not bootstrapped ventures or sole traders).

The top failure causes by frequency, from founders’ own post-mortems:

These are multi-select: a single company can cite multiple causes. Two additional data points from the 2024 report are less frequently cited but equally important: the median company in the dataset had raised $11M in total before closing, and the median time from last funding round to shutdown was 22 months. And 72% of the companies showed a Mosaic health score decline in the 12 months before officially closing — meaning the end was typically visible well before it arrived.

The $11M median is significant. These were not underfunded seed startups that never got off the ground. They were companies with real institutional backing that ran the full experiment and still failed. The cause was not lack of resources in most cases. It was lack of product-market fit.

Capital Failure Is a Symptom, Not a Cause

The most important reading of the CB Insights data is understanding the relationship between the 70% (capital) and the 43% (PMF). Many founders read these as separate, parallel failure modes. They are not parallel — they are sequential.

Here is the mechanism: a startup without product-market fit cannot demonstrate the retention, referral, and payment metrics that justify continued investment. Investors who see these metrics missing — or present but declining — decline to fund the next round. The company burns through its existing capital with no incoming round. It runs out of money and closes. The founders’ post-mortem notes: “we ran out of cash.”

The cause of death was capital exhaustion. The cause of the capital exhaustion was PMF failure. Treating capital failure as the primary problem — by extending runway through bridge rounds, cost-cutting, or pivoting only in pitch narrative — does not fix the underlying issue.

“A company with strong product-market fit can raise capital in almost any environment. A company without it cannot raise in any environment, regardless of its deck.”

This matters practically: if you are preparing to raise and the primary argument in your pitch is about capital efficiency rather than PMF evidence, you have probably diagnosed the wrong problem.

Failure Rate by Stage: the Full Breakdown

Aggregate failure statistics like “43% cite poor PMF” describe funded startups after the fact. They hide a more useful story: failure at the idea stage has a completely different cause — and a completely different fix — than failure at Series A. The table below breaks the failure rate down by stage, using stage-specific definitions of “failure”:

• Idea stage — percentage of startup ideas that fail a structured validation framework before any product is built
• Pre-seed to seed — percentage of concepts that fail to secure any institutional capital
• Seed to Series A — percentage of seed-funded startups that fail to close a Series A within approximately 24 months of seed close
• Series A to Series B — percentage of Series A companies that fail to close a Series B or achieve profitability within a reasonable operating window

All percentages are approximations. “Did not” figures from Carta represent lack of graduation within the tracked window, not confirmed shutdown.

Idea stage. This is where most startup failure is seeded — but it almost never shows up in founder failure post-mortems, because founders who fail at the idea stage typically build something first. The failure becomes visible at the seed stage when they cannot show traction, or at Series A when they cannot show retention. Most ideas that reach structured validation have at least one claim that doesn’t hold, and a large share earn a KILL verdict: the evidence base does not justify moving to build. The primary driver is a market validation gap — the submitted evidence for customer demand is either assumed (founder inference), weak (friends and family feedback), or missing entirely. This mirrors the 43% PMF figure above — except at the idea stage, the failure hasn’t happened yet.

Pre-seed and seeking a first check. Precise data here is hard to source because most companies at this stage are not yet tracked by longitudinal databases. The general industry estimate is that roughly 75–90% of startups seeking a first institutional check do not close one within 12 months. The causes differ from the idea stage: investors are not yet asking whether the market is large — they are asking whether the team is credible and whether there is any signal at all of demand. A structured market validation report does not fix a team-credibility rejection, but it does fix the “how do you know people want this?” question, which blocks roughly 30–40% of pre-seed conversations before the team question is even reached.

Seed to Series A. The Carta Q4 2024 report is the cleanest longitudinal data here. Two findings matter most: graduation rates have nearly halved in four years (30.6% for the 2018 cohort down to ~15.4% for the 2022 cohort), and the 24-month window matters more than founders think — seed rounds that don’t convert within it face a materially harder path. Failure at this stage — not graduating to Series A — is most commonly a retention problem (trial, not habit), a unit economics problem (CAC too high relative to LTV), or a market-signal problem (the seed round’s data suggests the TAM assumption was wrong). All three are visible, and partially addressable, before the seed round is spent.

Series A to Series B. Data here is less clean because “failure” broadens — some fold, some acquire, some raise a bridge and delay, some grow to profitability without a B. The aggregate picture suggests roughly 40–50% of Series A companies do not raise a Series B within a standard timeframe. The failure mode shifts from market evidence to operational scaling: the question moves from “does demand exist?” to “can you build a repeatable sales motion at a cost structure that makes the business viable?” This is a different problem, and one structured idea validation upstream cannot solve.

The table illustrates a key principle: earlier-stage failures are most preventable with evidence-gathering interventions. Later-stage failures are operational and require different disciplines — hiring, GTM architecture, finance — to address.

The Graduation Cliff: Your Series A Odds Just Halved

Carta’s longitudinal cohort data adds a crucial dimension to the CB Insights snapshot. Rather than analysing failed companies, Carta tracked surviving companies and measured what percentage of seed-stage companies graduated to Series A within two years of their seed close.

The Carta data is not primarily a story of more companies failing. It is a story of investors applying higher proof standards before writing the next check. A startup that raised a Series A on a compelling narrative and early traction metrics in 2018 now faces demands for demonstrated retention rates, named customer cohorts, and defensible unit economics before the first institutional meeting. The bar for “we believe you have PMF” has risen sharply.

For a seed-stage founder preparing to raise in 2026: roughly 85% of your cohort will not graduate to Series A. The primary differentiator between those who do and those who don’t is verifiable PMF evidence — not vision, not team, not market size. Evidence.

The 4 PMF Failure Archetypes™

When CB Insights codes a company as a PMF failure, it means the company built something the market did not want at the price required to sustain the business. But that top-level description contains at least four meaningfully different failure subtypes — each with a distinct detection signature, a different point of failure in the founder journey, and a different corrective path.

Understanding which archetype you are dealing with is the difference between a useful pivot and an expensive one.

ThriveFinity’s Idea Validation methodology is built around exactly this pattern: the Demand Mirage and Imagined Market archetypes are among the most common reasons a validated-sounding idea still earns a KILL verdict — problems that founders present as validated demand are most often demand for the category, not for their specific solution. The Unit Economics Trap is the next most common finding in ideas that pass initial demand validation but fail deeper Idea Validation scrutiny. We publish the real aggregate verdict distribution — anonymised and dated — once the sample is statistically meaningful, not before. See how that works →

7 Early Warning Signals Before Your Startup Fails

The CB Insights finding that 72% of shutdowns showed a health score decline in the 12 months before closing suggests that failure is rarely sudden. The trajectory is typically visible in advance. Here are the seven most consistent early signals across the failure archetypes above:

The Walking Dead: 50,000 Startups Already Failing

The CB Insights database captures companies that have officially shut down. It does not capture what the firm calls the “walking dead”: VC-backed startups that have not raised new capital since the start of 2023 and are slowly consuming their remaining runway with no credible path to a next round.

The 2024 CB Insights report estimates approximately 50,000 VC-backed startups globally are in this state. They are technically alive — some may still have 12–24 months of runway — but they are not growing, not raising, and not building toward a viable exit. Their failure has already occurred in the economic sense, even if the shutdown announcement has not yet been written.

This matters for aggregate failure rate statistics: the figures frequently cited today are backward-looking, measuring companies that have already officially closed. The real current failure rate, when walking-dead companies are included, is substantially higher than any figure derived from post-mortem databases alone.

It also matters for competitive dynamics. If you are operating in a space where several competitors have not announced raises in two years, you may be competing against companies that are no longer investing in product or marketing. Understanding which companies in your space are still actively competing and which are consuming runway quietly is a meaningful strategic input.

What to Do with This Data: Your Diagnostic Roadmap

The 43% is not a baseline risk that applies equally to all startups. It is concentrated in ideas that were not rigorously tested before they scaled — and, as the stage breakdown above shows, the specific risk you carry depends entirely on where you are right now. The founders who avoid it are not luckier. They are more precise about what they are testing and what counts as evidence at their current stage.

Here is a practical decision tree based on stage:

The companies that avoid the 43% are not doing something fundamentally different from founders who fail. They are testing the right questions at the right stage, before the wrong assumption becomes a $500K product nobody uses. The 43% is not an argument against building. It is an argument for knowing which archetype — and which stage-specific risk — you are most exposed to before you start.

Pranav Unni

Founder · ThriveFinity Connect on LinkedIn →

Pranav founded ThriveFinity to bring accountable, evidence-based verification to early-stage startups. He runs Idea Validation verdicts and signs every verdict personally.