This is a tough topic. Few founders want to talk about it. But statistics don't lie: around 80% of startups fail, with 20% failing in their first year and 50% not making it past their fifth year[^1]. According to a Harvard Business School study, 75% of venture-backed startups never return capital to investors[^2]. CB Insights analyzed 101 startup failure post-mortems and found that the top reasons for failure were no market need (42%), ran out of cash (29%), and not the right team (23%)[^3]. And although we do our best to pick the best founding teams and support them as much as we can, this also happens in our portfolio of founders. It has also happened to me personally in the past.
Failure never feels good. But I think we need to talk about it and also need to work on it in a way that a) it can be prevented and b) recognized as early as possible.
In general, I think that failure is an important part of success. Having said this, let me explain: I think that failing granularly can prevent you from failing epically. By building success based on iterations, tests, and pivots, you continuously will create winning and losing scenarios. And the granular winning and losing might lead to epic winning in the end. But this is not what this chapter is about.
This chapter is about the moments where you notice that it is not working out. And that you want to do something else.
Several things are important: transparency, clarity, inclusivity, and a good sense of reality. Sometimes some of the superpowers of excellent founders (a selective perception of reality, an intense conviction in their idea, and an unshakeable belief in their success) can stand in the way of recognizing the inevitable: the company is not working, and it might have to be ended.
One of the most valuable insights from post-mortems is the recognition that warning signs were present long before failure. As founders, we've often found ourselves in post-mortem situations asking "why did this fail?" and we noticed that we knew exactly when the issue started. We were able to notice the weak signal(s), but we did not act on them accordingly.
Research from the University of Michigan supports this observation, showing that 70% of business failures exhibit clear warning signs at least 6-12 months before the actual collapse[^4]. A study in the Journal of Management Studies found that early-stage startups that implemented formal early warning systems were 37% more likely to pivot successfully rather than fail outright[^5].
What I have started to do is conduct pre-mortems: I simulate how plans might go wrong and then look for weak signals that this is happening. This technique, developed by psychologist Gary Klein, has been shown to increase a team's ability to identify potential problems by 30%[^6].
Pre-Mortem Tip: LLMs and reasoning models are exceptionally strong at doing this. Try prompting them with: "I am planning to do [project and explanation], please give me 4 vivid scenarios how this will fail and the according countermeasures to take. Include the weak signals to look out for."
When a crisis is developing, in my experience, everything is much easier if you do multiple things at once:
Research published in the MIT Sloan Management Review indicates that transparent crisis management can reduce recovery time by up to 30% and significantly improve stakeholder trust during difficult transitions[^7]. The Harvard Business Review found that companies that maintained high transparency during crises were 21% more likely to maintain key relationships through the turmoil[^8].