Startups shouldn’t solve technically hard problems (InvertedPassion)
1/ Startups get funded when they’re expected to be valuable, and they’re valuable when they can generate a continuous stream of profits for its investors.
2/ With this view, the value of a startup comes mostly from its expected moat, i.e. how well can it defend its business from competitors once they take notice of the market.
3/ Startups that solve technically hard problems are often in an economically disadvantaged position because solving technical problems is hard, but once a solution is found, it’s not as hard to understand or replicate it.
4/ And if the solution is the major value-driver of a startup but is easy to replicate, the value of the startup is diminished irrespective of how important a problem it is solving or how hard it was to solve that problem.
5/ The usual tool for business defense for such startups is a patent. But patents (outside of pharma) aren’t worth much. Patents are easy to work around and even if they’re not, a startup cannot afford the multi-year, million-dollar patent litigation fighting others.
No wonder, probably >95% of patents are worthless.
6/ One major exception to all that I’ve said so far is pharma/health/medical. There, solving technically hard problems is a profitable strategy, and hence a booming VC and acquisition culture has evolved around medical startups.
7/ Solving hard problems in the medical domain is worth the effort because of three reasons:
a) Due to patents, competitors can’t copy a molecule exactly so they have to modify it a bit
b) Once the leading molecule is demonstrated to work via Phase-1, 2 and 3 clinical trials, it has a lead time of 5-10 years over competitors who will have to do their own trials because their substitute molecules differ from the original one and hence regulatory agencies require fresh clinical trials
c) Because the medical market is price-inelastic, margins are high and the original innovating startup is able to accumulate enough cashflow in the initial years to fight patent battles
8/ So, the value of medical startups comes from the long regulatory period (and not necessarily from patents).
Plus, it helps that people pay a lot for healthcare (as compared to other things they desire).
9/ Thus, medical/healthcare startups can be seen as high-risk, high-return startups.
In contrast, most deep tech startups are high-risk, low-return startups.
10/ Low-return because, as we’ve seen, once a solution to a hard problem has been demonstrated, it’s often easy to replicate (unless there’s a regulatory agency with stringent criteria which implicitly gives an edge to the first one to solve the problem).
11/ A beautiful case study of high-risk, low-return startups is the cleantech boom of 2000-2010.
MIT has a report where they analyzed funded startups in this industry and found out that 90% of them failed to return even their originally invested capital.
12/ According to the report, cleantech startups had high risk in multiple areas:
- Tinkering science takes time, so they were illiquid assets until then
- Expensive to scale as building factories is capital intensive
- Razor-thin margins as they were competing against established commodity markets (oil and gas)
- Potential acquirers not ready to pay a premium as they figured they can develop technologies in house
13/ High-risk per see is not a problem (as demonstrated by returns in healthcare)
But high-risk along with no or weak moat translates into low returns, which made cleantech a bad investment (in retrospect).
14/ I worry that many upcoming “deep-tech” companies (in synthetic bio, climate, food, etc.) will face a similar fate to cleantech.
They need to answer what moat they’re building and be honest with how strong it really is (patents are not moats).
15/ Technical and scientific breakthroughs are essential to push our society forward but they often don’t turn out to be great investments.
Hence, science/tech breakthroughs should be funded by the public (via govt agencies) and startups should comfort themselves in the roles of commercializing already available technologies (rather than inventing new ones).
16/ To sum up, even though it’s often the case that taking market-risk pays more than taking technical-risk, I don’t think it’s a useful classification.
17/ Rather, think about the strength of a startup’s moat once the scale of its innovations is widely known by many.
18/ If once the profit source is known, it isn’t easily competed away, you have a valuable startup. If not, you have something cool (and potentially useful to society). But it likely won’t become a successful business.
This essay is part of the book on mental models for startup founders.