How do you decide to shut down a startup?

Shutting down is one of the hardest calls a founder makes, partly because the signals are almost never unambiguous. The honest answer is that you shut down when continuing would destroy more value—financial, personal, and human—than stopping. That means doing a clear-eyed audit of growth trajectory, team morale, and capital efficiency rather than defaulting to optimism or despair.

The core question: is there still a growth engine, or just motion?

Paul Graham's framework for what makes something a startup centers on one word: rapid growth. A company that isn't growing isn't a startup anymore—it's a small business, a hobby, or a mistake. That distinction matters enormously when you're deciding whether to continue. The question isn't whether you're surviving but whether the underlying mechanism that could produce rapid growth actually exists. If you've been operating for 12–18 months and you can't point to a single metric that's compounding—whether that's revenue, active users, retention, or referral rate—you're not in a temporary trough, you're in a structural dead end.

The trap most founders fall into is measuring activity instead of growth. They count meetings held, features shipped, or partnerships discussed as evidence that things are moving. None of that matters. What matters is whether customers are pulling the product into the market or whether founders are constantly pushing. If every new customer requires heroic founder effort to acquire and retain, the unit economics will never close, and you should stop before burning more of your savings, your team's time, and your investors' capital.

A useful mental test: if you stepped away for 30 days, would the business grow, hold steady, or collapse? Startups that are working tend to have some self-reinforcing loop—word of mouth, a content flywheel, a network effect—that operates even when founders aren't pushing. If the answer is collapse, you're doing manual labor disguised as a company.

What the fundraising signal actually tells you

Founders often use investor interest as a proxy for startup health. This is reasonable but must be interpreted carefully. Graham's observation that you should avoid trying to raise money when you won't be able to convince investors is not just tactical fundraising advice—it contains a diagnostic insight. If you've talked to 30 relevant investors and none can articulate why this business wins, that's not a distribution problem. That's product-market fit feedback.

The danger is treating a failed fundraise as a temporary market condition rather than a signal about the underlying business. Sometimes capital markets are tight and great companies can't raise. But if you can't raise and you also can't grow without raising, you are in a binary situation: either find a path to default alive on existing revenue, or acknowledge that the business as currently conceived requires capital you can't access to prove a hypothesis that isn't yet proven.

Default alive means your current revenue covers your current burn before you run out of money. If you are not default alive and cannot raise, the math does the deciding for you—the only question is whether you control the timing of shutdown or the bank account does. Controlling the timing is almost always better: it lets you handle obligations to employees, customers, and investors with dignity rather than chaos.

Pivot vs. shutdown: how to tell the difference

Many founders who should shut down instead run a series of pivots that consume runway without improving the odds. A real pivot is a hypothesis: 'We believe the same core capability, applied to a different customer or problem, will produce growth.' A false pivot is a coping mechanism: changing just enough to feel like something different is happening while avoiding the hard question of whether the founding team has real insight into any problem anyone urgently needs solved.

The diagnostic for whether a pivot is genuine: can you articulate, in one sentence, a specific customer who has a specific urgent problem that your specific solution solves better than anything else they could do today? If yes, and if you can test that hypothesis cheaply within 60 days, the pivot is worth attempting. If the answer is vague—'we think there might be something in enterprise' or 'maybe we go B2B'—that's not a pivot, it's drift. Drift consumes the same resources as a real pivot but produces no useful information.

A practical rule: allow yourself at most two genuine pivots before treating a continued failure to find product-market fit as a signal about the founding team's insight in this particular domain, not a signal that you just haven't found the right pivot yet. There's no shame in being wrong about a market. The shame is in spending three years of your life proving it slowly when the answer was available after six months.

The human cost accounting founders skip

Every shutdown decision involves people, and founders often either over-weight or under-weight this factor. Over-weighting looks like: 'I can't shut down because my team has been so loyal.' Under-weighting looks like ignoring that employees have mortgages, career trajectories, and family obligations that are being held hostage to your optimism.

The honest accounting works like this: if you have six months of runway, your employees effectively have six months of job security regardless of what you tell them. If you know the company is unlikely to survive and you're not telling them, you're making their career decisions for them by withholding information they'd want. A founder who shuts down with three months of runway remaining—giving people time to find new roles, providing references, and paying out what's owed—does more for their team than one who runs to zero and lays everyone off on a Friday.

Also frequently skipped: founder wellbeing is a legitimate input. A startup that could theoretically survive if the founders just worked harder and sacrificed more is not necessarily worth continuing. At some point, the expected value of your time is higher applied somewhere else. This isn't quitting; it's resource allocation. The opportunity cost of continuing a failing startup isn't just money—it's the compounding value of the years you'd spend building something that actually works.

A practical shutdown decision checklist

Before making the final call, run through four concrete checks. First: growth. Has any core metric grown meaningfully in the last 90 days without extraordinary founder intervention? If no, name the specific change you're betting on and the date by which you'll know if it worked. If you can't name both, you're drifting.

Second: capital. Are you default alive? If not, what is your realistic fundraising probability given your current metrics and the conversations you've already had? Be honest—not about what you hope, but about what investors have actually said. Third: team. Is your core team still motivated, or are they staying out of loyalty or sunk-cost thinking? Have direct individual conversations, not team meetings, to get honest answers. Fourth: founder conviction. On a private scale of 1–10, how convinced are you that this specific problem, solved in this specific way, for this specific customer, is genuinely the right bet? If you're below 6, that's a meaningful signal—founders who believe tend to find ways to transmit that belief, and founders who are performing belief tend to underperform in every unstructured moment.

If the answers to those four checks are all pointing toward shutdown, they are. The remaining decision is only about timing and process—and both should be handled with as much care as the decision itself.

“Rapid growth is what makes a company a startup.”

— Paul Graham, source

The one thing to do

Set a specific 60-day deadline with one named hypothesis and one measurable threshold—if growth doesn't cross that threshold by that date, shut down while you still control the timing.

Frequently asked questions

How long should you keep trying before shutting down?

There's no universal timeline, but if you've spent 18 months without finding a metric that compounds on its own, and you've run at least two genuine pivots, the expected value of continuing is almost always lower than starting fresh. The question isn't calendar time—it's whether each month is generating new, useful information or just burning resources proving what you already suspect.

Should you shut down or pivot if you're out of money?

If you're out of money and have no bridge, shutdown is typically forced. The better question is what to do with the last 60–90 days of runway: use it to test one specific hypothesis cleanly, or use it to wind down with care. Running a pivot on fumes rarely produces clean data, because desperation distorts every decision.

What do you owe employees when shutting down?

At minimum: honest communication as early as legally possible, full payment of wages owed, and genuine references. Ideally, enough runway notice that people can job-search while still employed. The founders who run to zero and then lay off staff have made a values choice, not just a financial one.

How do you tell investors you're shutting down?

Tell them before you've made the final decision, not after. A brief, factual note—what you tried, what you learned, what the situation is, what you're considering—gives investors the chance to suggest alternatives and gives you one more data point. Most investors would rather know early than receive a shutdown email out of nowhere.

Sources

More playbook answers · Growth Prophet home