Why are there no specialist company launchers?

1 - The founding of any startup is the most important moment in the company's life, but it's done by the least skilled and experienced people. Finding the right problem, building the first version of the product, hiring founding employees, finding product market fit and steering towards a big, defensible business are all vitally important activities that require skill, and that give outsized returns. First-time founders have literally never done this job before. Second-time founders have done it once, perhaps for somewhere between 6 and 36 months.

2 - The best way to build this skill is through doing it. Figuring out a complex market, wandering through the idea maze, and building the foundations of a huge business is exactly the kind of tacit skill that practice makes you much better at: the feedback cycles are long, the judgement calls are complex, and almost none of it can be learned from a book. Second-time founders have a much higher success rate. It's very unlikely that the improvement curve levels off after just two reps.

3 - But most founders will do this once or twice in their life. Even Elon Musk has only done it five or six times. This is because once founders successfully get to product market fit their reward is to run a growth stage company: a completely different job.

4 - There are many reasons why founders are almost always the best people to run their companies at the growth stage. But they're also the best people to launch new companies. They have incredibly good performance above replacement in both jobs. The question is where their comparative advantage is greatest.

5 - At the moment they're all pushed towards growing their own companies: the market doesn't have a legible mechanism for allocating great founders to the specialist function of launching new companies but not growing them. Every norm pushes them to continue to grow the company that's reached product market fit. Hiring execs to replace founders at the series A is taboo, vesting punishes founders who leave 'early' and a founder who leaves a post-PMF company to start something new is read as a failure or a flake, not as someone rationally deploying their comparative advantage. There's no commonly used title for someone who specialises in launching companies but not growing them. There's no normalized mechanism for handing off a company from the team that launched it to the team that will grow it. Founders who launch a second successful company are encouraged to grow that too once it reaches product market fit.

6 - For some founders, comparative advantage clearly favours the growth stage: it would almost certainly have been a terrible idea to replace the Collisons at Stripe, or Bezos at Amazon. But it seems unlikely that every single founder's comparative advantage lies in growing rather than launching.

Importantly, the replacement pools are also very different. The market is full of excellent growth-stage operators. It has almost no specialist company launchers. When Mike Speiser stepped away from Snowflake, the company replaced him with Bob Muglia, formerly President of Servers and Tools at Microsoft, and later with Frank Slootman. His replacement at the next company was a first-time founder who'd never done it before

7 - Therefore a company that provided a means for talented people to become specialist company launchers, but not scalers, should be able to create and capture a huge amount of value. If there are indeed returns to specialisation and practice in repeatedly launching companies, then this company could harvest them. Their company launchers could launch a company per year for 40 years: they could have 20x more experience in this specialist function and shots on goal than even the best second-time founders.

8 - This company already exists in at least one form: Sutter Hill Ventures, and specifically its managing partner Mike Speiser. The model is simple. Speiser identifies a promising enterprise market, recruits 2-3 technical co-founders, and then steps in as founding investor and interim CEO, doing the zero-to-one work himself, roughly two days a week. He runs customer development, shapes the initial product, and steers the company toward product-market fit. Once it gets there, he hands off to a permanent CEO and moves on. Then he does it again. One new company per year. Roughly 20% of Speiser's incubations have become multi-billion dollar companies, a hit rate that makes the best traditional VC funds look ordinary. Snowflake, which Speiser incubated in 2012 and ran as founding CEO until 2014, grew until their $63B IPO in 2020. Sutter Hill owned approximately 20% of it. His other incubations include Pure Storage, Lacework, and Clumio. He might have launched more successful tech companies than any other single person. For a longer description see this excellent essay from Kevin Kwok (who later joined Sutter Hill himself).

9 - So where are all the other Sutter Hills? If they're getting such great returns running this model, why isn't the market following suit? You could have imagined early on that no one knew about them, but not any more. It's now super legible. You can read excellent essays on Sutter Hill. Their companies are big and famous. What's the limiting factor that explains why we don't have far more?

A few ideas:

Status lags. It will be high status to be a company launcher once that's a legible, understood role, with many successful examples to point to. Until then it'll be higher status to be a founder running a series B company than to attempt this weird, unproven new role.

Misaligned incentives with current investors. If a founder has comparative advantage in launching companies but also has good performance above replacement in scaling them, and decides to leave their first company to launch a second, the investors in the first company are penalised and gains accrue to the investors in the new company (or to the founder if they can price high enough in the first round for the new company). But this shouldn’t be a huge problem and is another reason why companies like Sutter Hill are valuable: because they align the incentives nicely for everyone by owning the entirety of the equity in each company from the start, meaning they want to deploy their founders, or particularly Mike Speiser, wherever he has comparative advantage, because they're agnostic about which company he makes more valuable: they just care about the total portfolio value.

Maybe it's happening, but slowly. There might be two lagging curves here: one of new incubators being launched, and the other of their companies growing big enough for us to notice. Maybe we're in the early innings of other Sutter Hills being launched and we're waiting for their Snowflakes to emerge.

Others doing this

Boulton & Watt — Sam Gerstenzang and others. Founded 2022: seems to be working! Their first company Moxie is at series C and their second, Meadow, is at series A.

We build one company at a time with focused execution and high success rates.

We start one new company every year or so, and we do the work ourselves. We don't work with entrepreneurs-in-residence or invest in other companies.

We think product-market fit is a hard problem that can't be industrialized or handed off. Most of the work of a scientist isn't generating hypotheses, it's running experiments to get the data. That's what we do.

Once we've created real traction, we partner with best-in-class operators to create a second founding moment to scale the company.