Can You Avoid Overhiring?

Headcount Dynamics for Software Businesses

A friend of mine who runs his own startup recently posed the question: “How do teams grow so big? I can’t imagine my company ever needing more than 20 people.”

This question triggered a rant that in some form or other I’ve had brewing in me for a few years, so here we go. Note that I will focus on the “initial spurt” of overhiring that new companies do but it is trivial to deduce how the process self-perpetuates for larger companies.

---

Within every organization, there is a gravitational pull to HIRE MORE people. It is generated by the existing incentive structures for all of the company employees.

An IC (individual contributor) is always overworked, necessitating a new hire. There is always too much work and not enough time to do it. There are a few reasons for that:

1. The word "work" includes everything from "needs to be done right now or we go bankrupt" to "would be nice to do at some point in the future". The scope is always infinite.

2. Seeming busy is good for your career. It makes you look important. Imagine a colleague who spends 4 hours every day on the golf course - you'd probably subconsciously think that if they have the time to do that, they must not be terribly critical to the company's operations. If you think about it - it might or might not be true! They might be the only person at the company who understands Critical System X. Or they might be slacking. We don't know a priori, but the bias is always there.

3. There's always some kind of work that an IC doesn't like. When you ask an IC whether we should hire a new employee, they will always think something along the lines of "If we hire a new person, they can do the TPS reports, and I can focus on the more creative task of Concepting. And, after all, I AM overworked."

A Manager always needs more people on their team. There are two reasons for that (or rather a reason and a justification):

1. Having more subordinates is always good for the Manager's career. That's true within an organization and doubly true when switching jobs. Companies cannot really evaluate candidates very well during the interview process, so they rely on proxies. For a manager, the number of people they manage is the most reliable proxy because every company can relate to that. The more people you manage, the higher your title/salary when you switch jobs.

2. All of their ICs are overworked! That's all managers see and hear from below. Overworking is bad for morale and retention. More people must be hired so that the most valuable ICs don't burn out.

Similar forces are acting on Founders.

1. A Founder's social status outside of their immediate niche (with some exceptions for recognizable brand names) is determined by their number of employees. Most people don't know whether a Widgetmaker Inc is truly important within their niche of synthetic biology/materials design/SaaS/etc. But they do understand that if one hundred people work for someone, that person is probably more important than if fifteen people work for them. Similarly to Managers, a high headcount is useful for clout when starting future companies, but mostly it's an immediate social status thing.

2. All of the projects in the company are always behind schedule. And the Managers say that's because they are understaffed and if they could only hire one more person, they'd be on track. And all of the ICs do really do look overworked.

3. When raising the latest round of capital, the Founders promised their investors robust revenue growth. To reach that plausible but ambitious rate of growth, the company needs to... Do more things! And doing more things requires more people.

That last point is where things get interesting. For the vast majority of businesses in human history, to do more things you NEEDED more people. To make more cars, you need more factory workers (and more factories). To fly more passengers, you need more pilots and flight attendants. To serve more beer, you need more bartenders.

Sure, there are labor-capital tradeoffs you can make, for example you can buy a beer-pouring machine:

…and make your factories more efficient and fly fuller planes, etc., but at some point, you do actually need more people to do more things.

That relationship is a lot less rigid for software companies. There are examples of companies serving tens or hundreds of millions of users with very few employees - the most famous ones being Instagram, which had 13 employees serving 27M people when it sold itself to Facebook, and WhatsApp, which had 55 employees serving >400M users when it sold itself to Facebook.

Sure, some software companies, especially B2B ones, may need to grow headcount to grow business when it comes to sales, customer support, partner engineering, etc. But the broader and more interesting dynamic is that software can scale with a lot fewer employees than any businesses that had existed before computers.

On top of that, prediction in software (especially in startup software) is hard. When you're building your 5th car factory, you can do the math pretty well - it will employ X people, produce Y cars, and give us Z% return on invested capital. When you're hiring your 5th product manager... Damn, I don't know. They might come up with a product that will double your growth rate. They might ship useless redesigns that will antagonize your users. They might create process and drama and, as a result, worsen the morale of your ICs. I've seen all of those happen, the only comment I have is that when hiring the 5th PM, both the candidates and the executives focus a lot more on the first possibility than on the second and third ones.

To sum up what we have so far:

1. There is always a gravitational pull in every organization to HIRE MORE.

2. Software companies can scale with fewer people than any businesses that have come before them.

3. Predicting the outcomes of hiring more people in your engineering/growth orgs is very hard, on balance tech people are optimistic, and that leads to HIRING EVEN MORE.

Combined, that has led to many a software company overhiring to the point where the last X people they have added have not only not contributed to their revenue growth/product development, but have actually detracted from it. The post-acquisition Twitter cull has demonstrated that you can fire a ton of people from a tech company AND increase product shipping speed at the same time¹.

---

You cannot remove the HIRE MORE gravitational pull. The incentive structures are what they are, calling them out doesn't help. You can have your entire team discuss the part above before the next hiring cycle, and it will not change a thing. There are, however, forces that act in the opposite direction.

The first one is externally imposed by the market - capital efficiency requirements. We haven't had much of that in the tech world lately given the prolonged period of zero-interest rates we've just had, but it is a reasonable request in many times and places for investors in companies to require return on equity math when investing in companies. It doesn't solve the "prediction in software development is hard" problem, but it does make startups consider the costs of hiring more seriously.

I don't know whether we're headed to a world of non-zero interest rates for a long time or not. Right now we're in a weird dual state where there's a 2021-like capital flood in AI and a capital drought everywhere else. The best arguments for moving in opposite directions are *roughly* as follows:

1. We've borrowed as much from the future as possible (on the federal, state, pension system, and corporate levels) and we've run into an inflation-imposed limit on doing more of that. We will now be forced into a world of paying back the debts we've accumulated that will require higher interest rates to avoid falling into an inflationary spiral.

2. We are in a structurally-deflationary world caused by an aging population, offshoring industrial production, and technologically-caused cost decreases. Moreover, the new businesses being built (software mostly) require very little capital, so there is a lack of productive investment opportunities compared to the amount of capital we have. The inflation was a blip caused by the pandemic and the related monetary response and supply chain crisis, but now that it's over, we're going back to the zero-interest-rate world come the next recession.

Both arguments have kernels of truth in them and I am not nearly smart enough to figure out which one will win out in determining the future investment environment. Incredibly smart people have compelling arguments on both sides, if you're interested in learning more about this, I'd look into Stan Druckenmiller and Mike Green, but it's beyond the scope of this post.

The relevant bit for this post is that we're not sure what the externally imposed constraints on hiring are going to be like for startups in the future (and that the uncertainty is greater than at most points in living memory).

The external factors can not only constrain your hiring but also, weirdly, incentivize it. I remember a conversation I had in 2020 with a very successful founder, whom I was arguing, should aim for raising less money and hiring less to de-risk his company. His response was "I could choose to not raise money right now, I don't need the extra $50M but then Softbank will give my competitor $200M and my life will get very difficult".

This is a valid concern in many industries. The way to think about whether it applies to you is whether you operate in an industry where having more capital can buy market share. A great example of an industry where that's possible is ride-sharing - the companies most willing to lose the most money were able to price out their competitors. A good example of an industry where that doesn't apply is single-player video games - if you make the best fantasy RPGs in the world, it is exceedingly unlikely that a random new company can make a game that will drive you out of the market if Softbank gives them 10X the amount of money you have.

---

It is important to recognize that in a bubbly investment environment, it may be a perfectly rational strategy to:

1. Raise as much money as you can in a hot space.

2. Hire as much talent as you can in that space.

3. Sell your company quickly (or SPAC it/go public if you're into that YOLO lifestyle), while the space is still hot².

4. But this strategy is more of a get-rich-quick-gamble strategy than build-sustainable-business-while-minimizing-risks strategy.

This leads us to the next source of hiring discipline. Risk aversion in founders. If the founders seek to maximize the chances of their company surviving in the long term over maximizing the chances of getting paid quickly, they are likely to hire less and get to profitability (default alive) more quickly. Josh Wolfe says that it's a myth that the greatest founders are risk takers, while in reality, they are risk killers.

The risk-killing strategy makes more sense if your goal is not to get rich quickly but to build something that lasts. Founders who prioritize legacy have a stronger internal incentive for hiring discipline.

This creates a bit of a conflict. Legacy is even more unevenly distributed than capital. Which is to say, founders get a larger chunk of legacy than employees when compared to money. That makes for a misalignment - employees (and especially managers) motivated by their career incentives will push for a higher headcount, while founders will push back. It's not the end of the world but it will cause some conflicts.

There is, however, a way to solve this issue for a narrow slice of companies. Here is the key: some of the best ICs you can hire heavily prioritize working conditions only possible at small companies due to a bit of an Ayn Randian mechanic:

The most potentially productive employees want to work in the most productive environments. And the more people you hire, the less productive your most potentially productive employees will be due to process, politics, meetings, etc.

Because of that, some of the most productive people in software development thrive in small teams (the exact size is debatable but it's definitely less than ~100 people with 2 layers of middle management). If you've worked in the industry, you've probably met them - star engineers who only seem to thrive at early-stage companies.

Which means that there's a local optimum point for software companies where the companies don't hire a lot of people and consciously choose to limit their headcount at some small number of people. The founders are happy because they maximize their chances for legacy. The employees are happy because of good working conditions and no bullshit process. The investors are happy because of capital efficiency. As the amount of existing software infrastructure (cloud etc) grows, such companies are able to compete in more and more industries. If I were starting a VC fund today, I'd focus it on such companies that aim for a perpetually-low-headcount from the start.

To succeed as one of those companies, a group needs to fulfill a certain number of conditions:

1. Operate in consumer software or a similar industry where one can grow revenues without growing headcount.

2. Operate in an industry where a competitor getting an ungodly amount of money cannot force you out of the market.

3. Have founders that prioritize legacy over a quick monetary exit and the immediate status gained from a high headcount.

4. Have employees who prioritize productivity over corporate title/prestige/other career benefits.

5. Actively agree from the start on a low headcount, encode it in the founding principles, inform potential investors and hires, and treat every new hire as a very heavy incurred cost.

6. Not be in the empire-building business.

The last point is a bit of an aside, as I haven't touched on it in the article, it's just worth mentioning that if you're in the empire-building business, completely different dynamics apply. For them, the best advice is this quote from Titan:

Taking for granted the growth of his empire, he (Rockefeller) hired talented people as found, not needed.

Further reading: Immoral Mazes, Default Alive or Default Dead

1. You can make a counterpoint about reliability here but we don't know to what extent the tradeoff is headcount/reliability or shipping speed/reliability, my money is that the latter is more accurate.

2. If x.ai went for a SPAC right now after their announcement but before doing anything, I’d be a lot less surprised if they ended up with a market cap of >$10B than a market cap of <$1B.