In 1994, two Stanford professors walked into nearly 200 Silicon Valley startups with clipboards and a question nobody was asking: What happens to companies based on the employment decisions founders make in their first year?
James Baron and Michael Hannan didn’t know it yet, but they were launching one of the longest-running studies of startup culture ever attempted. The Stanford Project on Emerging Companies — SPEC — would track these firms for more than 15 years, following them through the dot-com boom, the bust, and everything after.
The findings would upend a lot of conventional wisdom about what makes startups win.
Five Blueprints, One Clear Winner
Baron and Hannan identified five distinct “organizational blueprints” that founders used to build their companies. These weren’t vague vibes. They were concrete choices about three things: why employees would stay, how work would be coordinated, and what criteria would be used to hire people.
The models ranged from “engineering” cultures (hire the smartest people, let technical excellence drive everything) to “star” models (recruit top talent from elite institutions) to “commitment” cultures (hire for cultural fit first, build deep emotional bonds with the organization).
“The commitment model treats the company as a family,” Baron explained in a Stanford GSB interview. “Founders who adopt this approach are deliberately selecting for values and cultural alignment before they look at technical skills.”
Most founders in the sample didn’t choose commitment. The engineering model dominated, especially in Silicon Valley, where technical prowess was religion. But the data told a different story about what actually worked.
The Numbers Were Not Close
Companies built on the commitment blueprint had a five- to six-fold advantage in their likelihood of reaching an IPO compared to engineering-model firms. That’s not a marginal edge. That’s a different category of outcome.
Even more striking: as of January 2000, not a single commitment-model company in the study had failed. Zero disbanded, zero disappeared, zero acquired in distress. The failure rate across the rest of the sample topped 13%.
Hannan put it bluntly: “The effects of the founder’s original blueprint on the company’s trajectory were remarkably persistent. These early choices created organizational DNA that was very difficult to alter later.”
Diane Burton, who joined Baron and Hannan as the third principal investigator from MIT, found that the pattern held even after controlling for industry, company age, and the amount of venture capital raised. Culture wasn’t a proxy for something else. It was the thing itself.
Why Culture-First Hiring Creates an Unfair Advantage
The mechanism isn’t mysterious once you see it. Commitment-model companies invested heavily in onboarding and socialization. They were slower to hire. They were more deliberate about who got through the door. And because employees felt genuine attachment to the organization — not just to their project or their stock options — turnover was dramatically lower.
Lower turnover in a startup’s early years is a compounding advantage. Every person who leaves takes institutional knowledge with them and creates a recruiting cost. In a company of 30 people, losing five in a year is an existential disruption. In a commitment-model company, it almost never happened.
There’s a connection here to what researchers now understand about intrinsic motivation and incentives. The commitment model works precisely because it doesn’t rely on external carrots. People stay because they belong, not because they’re being bribed.
The Trap of Changing Your Blueprint
One of SPEC’s most important findings was about what happened when founders tried to change their organizational model after the fact. It was a disaster.
Companies that shifted from one blueprint to another experienced significantly higher turnover and dramatically lower long-term stock valuations — nearly six times lower than companies that stuck with their original model. The act of reorganizing the social contract was more destructive than having a suboptimal model in the first place.
“Resistance to blueprint change, far from being pathological, seems to be telling us something important,” Burton noted. “Employees experience organizational model changes as a fundamental betrayal of the deal they signed up for.”
This helps explain why so many companies struggle after bringing in a “professional CEO” to replace the founder. The new leader often imports a different organizational philosophy, and the immune system of the original culture rejects it.
What This Means for Anyone Building a Team Now
The SPEC research is over two decades old now, and if anything, the findings have become more relevant, not less. In an era where manager engagement has hit historic lows, the idea that organizational culture is set in the founder’s first year — and then calcifies — should terrify anyone who’s building casually.
Baron and Hannan didn’t argue that the commitment model was right for every company. They argued something more subtle: that founders make these choices whether they realize it or not, and the ones who make them deliberately have a massive structural advantage.
The startups that hired for culture first weren’t soft or sentimental. They were playing a longer game. They understood something that most of their engineering-model competitors learned too late: you can always teach someone a new programming language, but you can’t teach them to care about the mission.
Fifteen years of data made that pretty hard to argue with.