Failure Criteria for Startups

In Silicon Valley a startup is a vehicle for performing an experiment. The motivation is usually a hypothesis of the form:

If we build and promote this product/service correctly and with the right timing, we will generate a scalable business that will significantly change an existing market or create a new one.

Once a startup has succeeded, it’s possible to construct a narrative showing how the hypothetical assumptions played out as the founders had hoped. For unsuccessful startups, it’s frequently the case that the experiment runs out of funding before the hypothesis has been fully put to the test. The participation of investors adds an implicit constraint to the experiment:

It’s possible to carry this out in stages of funding. The first stage requires $X, and will reach milestone M within T months.

Of course, experiments can and do fail before the money runs out. When this happens, it’s rare that founders acknowledge it and return the remaining funds to investors. Most often, failed startups linger on until they run out of cash. If you’re in Silicon Valley, I bet you already have an example in mind. If you’re not, know that many companies you’ve read about on Techcrunch are ghost ships. Long written off by their investors, they fly on autopilot like Payne Stewart’s jet. They may show signs of life once in a while because employees have to do something (“we just added wings to this boat because we like wings, it probably won’t fly but I’ll write a blog post about how pretty it looks“). Unlike human beings, companies tend to look healthy to outsiders even (and perhaps especially) when they are hopeless.

In my view, this is the worst possible failure mode for the founders. They often spend the remaining funding on keeping the startup alive for as long as possible. Sometimes they downsize, take pay cuts, and stretch the funding to last for years. This behavior is fueled by biases such as the IKEA Effect (people place an irrationally high value on what they have created), or the reluctance to accept that the effort spent thus far (sunk cost) has yielded little value. Some of these startups find metaphorical soft landings, but a startup is less like a physical aircraft and more like an expensive flight simulator.

When I was running my last startup I tried to avoid the above scenario. During the fundraising process I wrote down where I’d need the company to be at certain points in the future in order to want to go on. I saw this as a sort of contract with myself. Lucky for me, we always reached our goals (not always with much room to spare, of course) and were doing well when we were acquired. After that I tried other self-funded experiments by myself, for which I did the same thing. The latest of them was Cointipping, an app for sending small amounts of Dogecoin. I started it in early 2014, and I abandoned after a few months without thinking twice about it.

If I were to start a company today, I’d formalize the process. I’d have a set of minimum goals to accomplish in six months, and a tentative set for six months after that. If when the six months have passed I haven’t reached my goals, I’d fire myself. Otherwise I’d evaluate the next set of goals to see if it still makes sense and adjust it accordingly. Repeat until the company has reached momentum or died trying.

Final thought: this kind of self-imposed accountability is what most politicians abhor: when you’re not motivated to fulfill your promises but you want to keep your job, you do not want to be evaluated based on the results. However, a startup is not about staying in power. It is (to me, at least) about having a realistic chance of building something at scale. There are unlimited ways to do it, and no need to hold on to a particular one because “hey, we still have seven months of runway. Let’s pivot into something that won’t piss off our investors even though we’re not really excited about it.”

I leave you with an episode of the British TV show “Yes, Minister” that inspired this post (the relevant part starts around 5:40).

Leave a Reply

Your email address will not be published. Required fields are marked *