The Innovation Equation

October 29, 2021

A large proportion of the world’s scalable innovation comes from the venture ecosystem — venture-backed startups, plus public companies that used to be venture-backed startups.

If we want to figure out how to improve it, we have to be solving the right problem. How could we model this innovation?


Economics is good at finding ways to model things.

In 2007, as an undergrad econ major, I took a graduate-level seminar in economic history that had deep discussion of innovation and economic growth.

We read perhaps the most clever economics paper ever written, Nordhaus’s lighting paper, which created an alternative metric of economic growth over the last 10,000 years by measuring lighting efficiency (lumens / watt) over time, from Paleolithic fires to Babylonian lamps to candles and finally electricity.

We also read a landmark paper that one of the professors teaching the class, Petra Moser, had published. Petra (as we called her) had digitized and classified records from an exhibition in the Crystal Palace in London circa 1850. She’d painstakingly digitized the nearly 15,000 exhibit records, and carefully analyze what drives innovation — looking at the effects of patent laws and well as concentrations of industries by geography — chemicals in Germany, agricultural implements in the US, and so on.

A common theme in economics models is isolating key “inputs” into a particular output.

A lot of what makes a good model is picking a simple model. For example, here’s the Cobb-Douglas equation:

Y=A\times K^{\alpha }\times L^{\beta }

Where A is “productivity”, K represents capital, and L represents labor. Estimates of alpha are around 0.25 and beta around 0.75.

All the conversations around the Great Stagnation are driven by measurements of A and the great slowdown in growth of A since 1971.

If we were to create some sort of innovation equation, it would probably have a similar structure around five inputs:

The Innovation Equation: I = F(ecosystem engineers, founders, barrels, venture capital, tooling)

where barrels are people who can scale functions and lead initiatives and tooling is anything from CRMs to business analytics to better programming languages. 1

These inputs are complements, so the input in relative shortage tends to change over time.

For example, in B2B SaaS, 2005 it was probably tooling; in 2008 with “Good Times RIP” it was venture capital; by 2012, with Stripe, AWS, and renewed VC investment, it shifted to founders; by 2016 as these startups started scaling it was engineers, and by 2021 post-COVID with remote work and the flood of venture capital and growing startups we’re probably mostly barrel-constrained.

And in the startup ecosystem, at any given time most ambitious people tend to be tr45rwedwsa343remhinking about how to increase at least one of the inputs, if only for their company.2

The ecosystem often gets caught up in odd histrionics when the supply of any one of the factors increases too fast relative to the others — “there are so many startup tooling startups these days!” “what is Tiger helicoptering money going to do to the ecosystem?”

I have a pretty simple point of view here. More innovation is good. That means more venture capital is good. More startup tooling is good. More developer tooling is good. More founders are good. More engineers are good. More barrels are good.

Also, things that significantly expand the pool of one or another resource are good. Migration to NYC/Miami/Austin is good. CA SB9 is good. Venture-funding open source companies is good.

YC is really good. Remote work tooling is really good. Virtual reality and the metaverse will be really good. You get the point.

If you care about about increasing innovation, it’s worth thinking about how to increase innovation inputs at scale.

1 There are other factors, but in my experience they are less likely to be bottlenecks.

Other ICs roles can usually be filled with folks who are {smart, personable, hungry, analytical, highly organized | choose 2}. Glue roles like product manager and engineering manager can often be promoted from within. And in a world where three-person firms are putting $5 billion into the ecosystem in a matter of months the number of VCs is probably not a limiting factor (though perhaps board governance, good advice to founders, and funding on the fringes are somewhat of a constraint).

2 My personal example: from 2011-2013 I was working on “founders” via trying to create a loophole in the immigration system. From 2013-2017 I was mostly thinking about engineers, both because I was an IC engineer, hacking recruiting processes, and interviewing hundreds of engineers.

From 2017 until now building Gatsby, I’ve had to think about all five: we’re first-time founders; as an open-source startup we think a lot about both developer and general company tooling; a lot of the team has never worked in the startup ecosystem before; we’ve needed to hire barrels as we’ve scaled, and of course we need to think raising venture capital and in general building a company that can generate outsized returns.

Profile picture

Written by Sam Bhagwat, cofounder & chief strategy officer at Gatsby; programmer, analyst, writer; follow me on Twitter!