Why most startups die

The largest single category. The product works, the team is capable, and not enough people care enough to change what they currently do. Founders usually know this before they admit it, because the evidence arrives as a pattern of polite enthusiasm that never converts into usage.

The early signal is a gap between interest and behaviour: high sign-up rates with low activation, warm meetings that never produce a purchase order, users who praise the product and do not return. Interest is cheap. Repeated use is the only reliable evidence of need.

Second by frequency and usually the proximate cause even when something else was the underlying one. Running out of money is rarely a surprise; it is the end of a sequence of decisions in which spending was set against a fundraise that had not happened yet.

The signal is visible months ahead in the runway calculation, which is why the founders who survive are the ones who look at it weekly. Below nine months, every hiring decision is also a fundraising decision, and treating them separately is how companies arrive at three months of cash with a full payroll.

Co-founder conflict destroys a striking number of otherwise viable companies. The usual root causes are equity splits that were never renegotiated as contributions diverged, unclear decision rights, and an unaddressed difference in ambition — one founder wants a durable business, the other wants a large exit.

The signal is decisions that stop being made. Meetings that end without an owner, strategy questions that resurface every month, and a slow shift from arguing about the company to arguing about each other. Vesting and a written decision framework prevent more of this than most founders expect.

Being early is functionally identical to being wrong: the market is not ready, the education cost is paid by you, and a better-funded competitor arrives once the category is proven. Being late means competing on execution against someone with a distribution advantage you cannot match.

The signal for a timing problem is a long, expensive sales cycle spent explaining why the category exists. The signal for a competition problem is a steady erosion of win rate against one specific rival that no amount of feature work reverses.

This one kills companies that were doing well. Growth capital arrives, headcount triples, and the acquisition machine that worked at small scale stops working at large scale — acquisition cost climbs, onboarding quality drops, churn rises quietly, and the burn multiple deteriorates before anyone notices.

The early signal is cohort data, not aggregate data. Total revenue keeps rising while each new cohort performs slightly worse than the last. Companies that watch only the top-line number find out roughly a year after the problem started.

The hardest judgement in the job. A company with two years of runway and no evidence of need is spending its remaining option value slowly. A candid conversation — is there a version of this that works, and what specifically would prove it within one quarter — beats another year of iteration on a premise no one has tested.

Stopping well matters too. Returning remaining capital, helping the team find roles, and closing cleanly preserves the relationships that fund the next company. Investors back second-time founders who wound down honestly far more readily than founders who spent the last dollar hoping.

Garage to IPO can end a run through cash exhaustion, a collapse in board confidence, or a product that churns faster than it grows. Random events add pressure — a key hire poached, a competitor's launch, a downturn — but they rarely end a healthy company on their own.

The pattern most players recognise from their first few runs is that the fatal decision was made months before the failure appeared, usually a hire or a marketing budget approved against runway that was already too short.