Every piece of advice founders receive points one direction: persist. The stories we celebrate are all about someone who kept going when it looked hopeless.
Survivorship bias makes this advice nearly useless. We do not hear from the founders who kept going for four more years and lost everything, because that story does not get told. So the cultural default is to hold on, and the actual skill — knowing which situation you are in — goes untaught.
I have kept things alive too long. Here is what I try to use now.
The question is not "is it working"
New ventures do not work for a while. That is the normal condition and it tells you nothing.
The useful question is: is it working better than it was, for reasons I understand?
Both halves matter. Improvement alone is not enough if you cannot explain it — improvement you cannot attribute is improvement you cannot repeat. And understanding without improvement means you have a good theory and no evidence.
If the answer is no on both, and has been for a while, you are not being persistent. You are waiting.
The signals I take seriously
The reason keeps changing. It was the market. Then the pricing. Then the onboarding. Then the market again. Each explanation is plausible; the pattern is not. When the diagnosis keeps moving, it usually means nobody knows, and effort is being spent on symptoms.
Customers use it but do not care. The most dangerous position is mild success — real usage, real revenue, nobody upset, nobody enthusiastic. It generates just enough evidence to justify another year and never enough to justify real investment. Businesses can sit here for a decade.
The best case stopped being exciting. Sit down and describe the outcome if everything goes right for three years. If that outcome does not justify three more years of attention, you have your answer, and it did not require anything to go wrong.
You are relieved when you have an excuse not to work on it. Founders are honest with themselves about this less than about any other signal. Notice it.
The only argument left is what you have already spent. Money and years already spent are gone regardless of what you decide next. When they become the main reason to continue, the decision has already been made and is being avoided.
The signals that argue for continuing
A small group loves it. Not likes — loves, would be genuinely upset if it disappeared. A small intense group is a much better sign than a large indifferent one, and it is frequently the beginning of something.
The metric that matters is improving, even slowly. Pick the one number that actually indicates health for this business. If it is moving the right way for understood reasons, keep going, even if everything else looks unimpressive.
The thesis has not been tested yet. Sometimes a venture is failing for a reason unrelated to the actual bet — a bad hire, a technical problem, a launch into a bad quarter. If the core question remains genuinely untested, you have not learned the thing you set out to learn.
Cost of continuing is genuinely low. Some things can idle cheaply. A business that costs little to keep alive and might matter later does not need to be killed; it needs to stop consuming attention, which is a different decision and often the right one.
How to actually kill it
Decide the criteria in advance, when you are calm. Write down what you would need to see by a specific date, and what result means stop. This is the single most useful thing I have learned, because the version of you that is deep in it will not make this call fairly.
Then when the date arrives, hold the line — or consciously extend it once, with a new written criterion and a reason you would defend to someone else.
And when you do stop: do it cleanly. Tell customers early and honestly. Pay people properly. Do not let it die by neglect over eighteen months, which is how most ventures actually end and the way that damages relationships and reputation most.
What it frees
The real cost of a dying venture is not money. It is that it occupies the slot where something better would go, and it occupies your attention, which is the only genuinely scarce resource in a portfolio.
Every venture I have ended freed up capacity that went into something that worked. I have never regretted stopping something. I have regretted stopping late — which is a very different thing, and much more common.