I used to think startups worked like classical mechanics. In classical mechanics, objects are of macroscopic (normal) scale, and the vectors of these bodies – energy, mass, volume, momentum – exist on curves with each other. Some of these curves are linear, some logarithmic, quadratic, etc. In my working with startups over the last dozen or so years, I have operated under the assumption that team formation, product development, and early traction held characteristics of classical mechanics: they were on some sort of curve in relationship with risk over time.
In this model, risk gradually decreases as each of the milestones approaches, and is finally reached. Once the co-founder is selected, a company’s risk profile changed; once the first version of a prototype is built, it is de-risked further; once the product is launched, further still. And finally, when the company reaches product-market fit, or an inflection point where they had discovered a key sticky user feature or scalable distribution mechanism, the risk further lowers.
The graph might be shaped like this:
As investment capital has flooded into the pre- product/market fit ecosystem, each of these stages of a company’s lifecycle should have an efficient capital source for it – and the ratio of price to risk should be mostly consistent in time at the earliest stages. So, perhaps it looks roughly like this:
But what if that’s not true? What if risk in company formation does not look like macroscopic, classical mechanics, but rather like microscopic, quantum mechanics? After all, the earliest stages of a company deal with atomic and subatomic units – a single individual, a single feature, a single line of code. Quantum mechanics are different – at the atomic and subatomic level, objects are often part of bound systems, and have discrete, non-overlapping values at which they can absorb and emit energy.
What does that mean for company formation? What if risk, as a function of milestones, looked less like classical mechanics, where there can be continuous, curve-like de-risking in time? What if all of the activity that comes along with building a company does not in fact de-risk it? What if product market-fit is in fact the only discrete state in the early stage that de-risks a company?
After all, if you have hired an amazing team, but have not made something people want – you still do not have a business. If you have built a product, but not made something people want – you still do not have a business. Once you have made something people want, however, everything changes. What if that was the only functional change that mattered? What if, until that state change, you are effectively as high risk as when you first started? Perhaps the graph would then look more like this:
You’ll notice there is no delineation of time in the above chart – that’s because, of course, it’s very hard to know.* Heisenberg noticed something very unique about quantum object properties – that there was an inverse relationship between the location and the momentum of a particle: the more you knew about one, the less you knew about the other. There is inherent uncertainty in measuring time and energy (or in our case, risk). Similarly, in startups, product-market fit is something of a mystery. Slack started as Glitch, an MMORPG computer game, and raised three rounds of funding over a few years before they abandoned that concept in favor of the collaboration software that currently has millions of users and revenues. Facebook, meanwhile, had product/market fit within days, if not hours. The uncertainty over a given timescale is a core principle of startup formation. I now believe that quantum mechanics is the better way to understand the risk trajectory of a company.
Here’s where it gets interesting: if the market assumption is that startups de-risk according to classical mechanics, but they actually de-risk according to quantum mechanics, there are points in a company’s lifecycle where the market is inefficient in its pricing the risk. Most notably: once there are ‘milestones’ for a company in the early days, prior to product-market fit, the classical mechanics model would suggest the risk is lower than it actually is. And so the prices for investing in those companies is higher than it should be. And the more milestones the company has reached, incidentally, the less rational the pricing is. Therefore: the most rational seed investment is the price that you would pay at the earliest moment in the company’s trajectory – perhaps the moment the founder commits to the idea, or forms the company. Later in the company’s lifecycle, while still prior to product-market fit, the price, at least as the market currently works, is based on a misunderstanding of risk**.
It feels like a trade secret to share this perspective, since it is predicated on everyone else seeing the risk differently. But I suspect it’s counterintuitive enough that I can think out loud. To that point, I am so excited to be spending more and more of my time with people earlier and earlier in their startup trajectory. Extremely talented, passionate, creative people who are at t=0, and in many cases who haven’t yet taken the leap, are my favorite people to work with, and invest in – it is the most rational time to invest in a seed stage startup, after all.
*This is where the analogy is weak, because the states at which a particle can absorb and emit energy are not exactly where the time-energy uncertainty principle comes to bear. But that’s why this is a blog post, not scientific paper.
**There are enough counterexamples to make a rule, of course. For example, sometimes a team is of such high quality that it becomes a valuable company on that basis alone. Sometimes a product has introduced a technical innovation that is very hard to replicate, which also creates unquestionable value. But in high technology and internet go-to-markets, these are exceptions.
Building Mattermark in my attempt to measure the mechanical nature of startup risk and growth leads me to agree that there is something to this quantum analogy for early stage. Mattermark was never helpful enough for early stage folks to evaluate pre PMF companies (and therefore we couldn’t turn those types of investors into paying customers). I wonder what we could do at scale to detect the emergent phenomenon of going from not PMF to post PMF. Also, PMF isn’t static so you might also want to observe when it is waning or strengthening. Thanks for the thought provoking post!