Salary in my Startup: a Thought Experiment

I have written previously about trends, for instance self-driving cars; shared best practices, say on how to work with corporate VCs; and offered principles, for instance how to pitch to a VC. In the spirit of experimentation I have been playing with breadth, depth and style in them and learned much from people’s feedback (thank you). Today I want to continue trying different things and do a thought experiment. I may very well find out all the ways in which my hypothesis won’t work but if it gets people thinking I will have achieved my goal.

One of the issues that has puzzled me the most as an operator in a big company, entrepreneur in my startup, and VC investing in others, is around salary. What I would really like is a correlation algorithm, grounded on enough startup data, which takes in a few of my variables as input (industry, geography, company stage etc) to give me the salary I could expect.

But until we build a good enough AI to get there, how can we make sure to offer and ask for the right amount in a marketplace with imperfect information? There are incredibly rich articles everywhere online discussing this topic: First Round and YourStory offer some fantastic food for thought. I can also imagine economists who have spent many years thinking this through but I wonder if there is a way to distill into a basic formula. Especially because the way most startups still tackle the challenge of salaries is by surveying the market, a bottoms-up approach that is accurate but draining. My central argument is that approach can be complemented with a broadly applicable framework, a tops-down approach that can be always implemented.

So let’s make three generalizations.

One, that as the company moves from seed to A to B and beyond salaries approach asymptotically the corporate numbers.

Two, that a startup will pay as much as a corporate job once it hits series E. Consider now that investors will typically look for 25% ownership of a company given its valuation in the A, 20% in the B, 20% in the C, 10% in the D and E.

Three, let us ignore the many other variables to a salary including signing bonus, annual bonus, salary increases, and severance package. All of these tend to be significantly less common in startups anyways so it’s a reasonable assumption.

I would argue that reverse-engineering this paradigm can be a useful framework for salary negotiations. In other words, if an engineer can command $250K at the series E, comparable to what they would be paid in a corporate, then ~$225K (10% lower) is reasonable in the D, ~$200K (10% lower) in the C, ~$160K (20% lower) in the B and ~$130K in the A (20% lower). If in the future the engineer can command $300K at the series E based on market comps aka a corporate salary, then it would be ~$270K for the D, ~$240K for the C, ~$190K for the B and ~$150K for the A.

Underlying this framework is that there is something inherent about risk and ownership that is captured by VCs in each round of financing. I am sure there have been attempts to build a formula that could work broadly across geographies, sectors and market dynamics but I have never seen one that works. Yes, there are a huge number of catches, including the obvious limitation of reverse-engineering previous to a series A when it’s a non-priced round. The biggest catch is probably equity — the seed stage engineer will hold substantially more equity than an equivalent new hire, here is a previous post discussing how equity decreases exponentially not linearly. But if there is indeed an internally consistent model around salaries then we can all save ourselves much grief around them and focus on the truly exciting part of a startup: building cool things.

What do you think? How well does this framework hold to (messy) reality? Is it even possible to have a mathematical model around salaries given companies are so different from one another and are constantly evolving? Thanks to Ankit Jain for his ever valuable feedback.