Finding the Blueprint for Thriving Organizations

Q: What questions have guided your career?

The question I began with was, “What drives inequality in society?” Early projects led me to believe that the way organizations structure themselves and manage their employment relationships has a significant impact on why some people get ahead and some people don’t. At the time, organizations were getting essentially no attention from scholars interested in inequality, so it was an area where I thought I might be able to contribute to our understanding.

Later work on emerging firms in Silicon Valley got me curious about how choices around people practices and culture impact not only the employees but the organization itself—its prospects for surviving and thriving.

And over time, my interests have broadened to include almost anything that has to do with either organizations or their employees.

Q: Would you describe some of the early work?

The Department of Labor produced something called the Dictionary of Occupational Titles from the 1930s until the 1990s. To gather the necessary information, federal occupational analysts around the country went into companies and mapped out the requirements of all the jobs, who was in the jobs, and what the promotion paths were. The data was a sociologist’s dream.

I was able to access the records for the California office of the Department of Labor for my dissertation. Those records were paper—hundreds of thousands of pages. Working with the data started with xeroxing every page, then hiring people to key punch each record onto 80-column punch cards.

It was a huge undertaking, but in the process of digitizing the records, we saw that it was rare that women and men were in the same job—rare enough that when it did happen, we would double-check the paper record because it was usually a key-punching error.

It dawned on us that this extreme sex segregation was quite astounding and quite sociologically interesting. Until then, scholars had only focused on broad occupations—for instance, what’s the representation of women and men among lawyers or restaurant servers? At that level, there was a lot of segregation, but there were women who were attorneys and there were men who were attorneys. There were women who were servers and there were men who were servers.

However, when you looked at specific jobs inside organizations, you got a much different picture of the labor market; women and men were almost never in the same job in the same enterprise. That research provided the first comprehensive picture of just how segregated the world of work was.

It also showed that promotion paths were almost entirely segregated, too. Since women and men weren’t in the same jobs, it wasn’t a surprise that their paths for advancement were markedly different. But there were many fewer jobs that women could be promoted into, resulting in fewer and shorter paths for advancement. And when there were such jobs, they were women’s jobs, not jobs that were accessible to men and women.

Q: Why was there such segregation?

The predominant theory in economics at the time was called statistical discrimination, which argued that employers had beliefs about the typical aptitudes and abilities that women had or men had. Based on those beliefs, employers would decide how to staff a job—something like, “This job involves nurturing. We believe women are more nurturing, so this job is for women.”

The theory had gotten entrenched in the social sciences as a justification for why, essentially, discrimination was rational: in the absence of detailed information about specific individuals, employers relied on beliefs about the average qualities of different groups of individuals. But we were able to provide compelling evidence that the theory didn’t really stand up.

Most of the jobs we looked at had a mix of requirements, some that people at that time would have tended to associate generally with women and some that people would associate with men. If stereotypes were driving hiring, and the tasks required by the jobs were a mix, then we would have expected rational employers to hire a mix of women and men. But the jobs were completely segregated, despite displaying some duties that were viewed as more typical of men and others viewed as more typical of women. That data went against the theory and made a little bit of a splash.

It didn’t provide insight on how to reduce the discrimination, so, with a later project, we looked at agencies of the California state government to see whether external pressures affected efforts by organizations to remediate inequality.

Q: What did you find?

First, we found that the conditions in an organization when it was founded had a profound, enduring effect on the organization over time. Inequalities, such as pay inequality by gender or race, were almost never remediated by recalibrating what existing jobs were paid. A key driver for the organizations to become less discriminatory was to transform their division of labor—that is, kill off old jobs and create new jobs.

In addition, external pressures had a significant impact. Budgetary slack, oversight that made agencies feel accountable to the state legislature, and changes in senior leadership had a substantial effect on increasing the rate at which organizations moved toward more equal division of labor and pay structure.

Q: Your project on Silicon Valley startups required a decade-plus of data gathering and years more to analyze and publish the findings. How did such a long-lived undertaking come about?

The catalyst actually was teaching courses on human capital management while I was at Stanford. When I would argue that managing employment relationships was critically important in determining how successful organizations were, some students would ask, “What’s the evidence for that?” The honest answer was that it was for the most part an article of faith at that time.

The evidence that did exist came primarily from organizations that had been around for a very long time, looking backward and telling stories about how having treated employees well explained their success. That was pretty unsatisfying because it suffers from survivor bias. Many organizations may have done exactly the same things and failed. Without a complete sample, there’s no basis for knowing which factors mattered.

It's the same as asking elderly people how their upbringing impacted their careers. Their stories might be compelling, but there’s no way to know if they’re pointing to the things that actually mattered. If you instead start with a bunch of babies, characterize their early environments, and then follow them forward, you can develop more compelling evidence about what drives success.

This was the early ’90s. Looking around Silicon Valley for a sector where there were lots of relatively similar companies being founded, there was a clear answer—Chinese restaurants. But my co-author had already done a study of Chinese restaurants.

So the next option was tech companies. We ended up studying close to 200 emerging firms, characterizing the intentions and actions of the early leaders. We then traced those companies forward, looking at how the early history affected both economic and non-economic outcomes. They included revenue growth, whether firms survived, whether they went public, stock performance, how much workforce turnover they had, and how much administrative overhead and hierarchy the companies developed.

Q: What were the key results?

The first thing I learned is, you don’t do a study like that unless you have tenure because an untenured faculty member can’t afford to wait around for years and years for things to happen to companies before they publish anything.

In terms of findings, first, founders seemed to be choosing from a relatively small number of cultural models for their organizations. We identified five blueprints that we called engineering, star, commitment, bureaucratic, and autocratic. The predominant blueprints in Silicon Valley were engineering and star.

With the engineering blueprint, you hire a bunch of engineers with specific skill sets and make sure they have a lot of caffeine, sugar, and a ping pong table. Then you give them “cool” work to do and more or less leave them alone. If you launch a tech startup without giving a lot of thought to organizational considerations, the engineering blueprint is where you wind up. You also likely end up using it if you’re building the company to flip rather than endure.

The star blueprint is a variation on engineering. You poach the best and the brightest, then rely on their expertise to figure out what kinds of projects they should be working on. Venture capitalists tend to gravitate toward this model. Most companies that use it fail rapidly, but it’s often fabulously successful when it works. And if you’re a venture capitalist, you don’t care about your average return; you care about your best return.

A rarely used blueprint, commitment, was also among the most successful. It was based on creating a long-lived relationship between employees and the organization, cultivating a sense of family by hiring people with shared values. It wasn’t a very fashionable approach, however. Most people at this time were coming to believe that startups needed to be harder-edged and more flexible and that employees needed to “own their own careers.” The returns of the companies that adopted this blueprint were not extraordinary, but they did moderately well financially, they virtually never failed, they had much lower turnover, and they were able to remain leaner with respect to management overhead as they grew larger and older

So the star model seemed high risk and high return, whereas the commitment model seemed much lower risk and more moderate return. We wondered if we had a tortoise versus hare situation, so we modeled what happened to the portfolio of commitment firms versus star firms in our sample. For the timeframe of our study, even though many of the star firms had gone out of business, the ones that survived had done so well that by the end of our study period they slightly outpaced the commitment portfolio. Still, it stood out that founders who focused on building something enduring—based on the deep commitment of the firm to its employees and vice versa—were thriving, even during a time when everyone had come to believe that commitment was an outmoded concept

The second major finding from the research was that founders who adopted a blueprint early on that they could stick with did much better. Many founders had given more thought to how they were going to scale their telephone system than they had to how they were going to scale the organization. But organizations that could scale with their original blueprint flourished much more than organizations that had to go through a wrenching cultural transformation in order to scale. They were more likely to avoid failure, to go public, and to see their stock price grow after the IPO.

Q: Did that project look at inequality?

It wasn’t something we were able to include in the way we would have liked. In part, it was that these small startups didn't have human resource information systems that could produce that data but also, asking about it was very contentious and raised concerns from founders about what we were doing.

From our data, we could say that startups that built in formal workforce procedures and processes were more conducive to providing viable careers for women—blueprints that relied on merit worked for women; those that required fitting into a bro culture didn’t. We didn't have enough data about race or other demographic variables to draw conclusions.

Even today, companies don’t want to talk about what they do with respect to management of their workforce. It’s probably the aspect of organizations that’s the hardest to get them to share data on. Maybe they think it’s their secret sauce or they fear they’re going to get sued; either way, their reluctance tells you it’s important.

Q: To what degree has the tech sector changed since then?

The short answer is: less than many people would think. I think people in Silicon Valley today would still be able to characterize firms using our blueprints.

There has been more explicit attention to the picture for women and people of color than there was 20-plus years ago. But I think women and people of color still tend to feel that their opportunities are somewhat better in organizations that are either highly meritocratic or formalized in the employment relationship.

The commitment model is still rare and difficult for organizations to offer, which is part of what makes it so powerful. In a world where most organizations adopt a transactional approach to their workforce, it’s a significant distinction. And survey after survey of millennials and Gen Z highlight their desire to be affiliated with an organization that allows them to develop and thrive. That says to me there may be more of an appetite for commitment-based organizations than is often recognized.

Q: What led you to Yale SOM?

Yale SOM’s mission was a big part of it. It’s great if you can make organizations and their employees better off. If you can also extend that to positively impacting society, it’s so much more compelling. I find it motivating to encounter students and alumni who are not just thinking about the narrow impact of their actions; they’re grappling with the broader implications too.

Beyond that, most organizational behavior groups focus on either macro or micro issues. Macro, broadly, leverages insights from sociology to look at organizations and how they relate to their external environment. And micro, roughly, leverages psychology to examine individual and group processes. Yale SOM’s organizational behavior group made a very conscious decision to meld both, because we think that the most compelling and important issues, both from a research standpoint and from a practice standpoint, require understanding both the micro and macro aspects of a phenomenon.

For instance, if you want to understand the impediments to organizations creating a truly diverse, equitable, and inclusive culture, you’re not going to get there by focusing on the organizational structure to the exclusion of individuals and their responses, nor are you going to get there by focusing on individual and small-group processes without thinking about the larger organizational context in which those processes are happening.

That approach has meant I’ve published work in sociological, economics, and psychology journals. I also bring that cross-disciplinary approach to teaching, developing courses that call attention to the importance of human capital strategy for how firms perform that draw on sociology, economics, psychology, social psychology, and political science.

Q: How do you think about impact when you look at your work?

When I started, a focus on organizations as important contributors to patterns of inequality in the labor market was unusual—indeed, it was rejected by most scholars. Today it’s very much accepted as something that’s important. I’d like to think maybe I helped to broaden the study of inequality in ways that other people have found to be compelling.

I also think, unless you are someone like LeBron James, most of us positively affect organizations through the way in which we enable others rather than by our own individual contribution. And certainly, I have felt that if I can help make Yale SOM a place where my colleagues can be better and the students can be better, that’s going to be a much bigger impact, in the long run, than any specific piece of research that I do.

People who graduate from Yale SOM are going to go become leaders. They’re going to lead people who lead people who lead people. Helping them understand how to manage better will produce a larger cumulative effect on how organizations perform than helping them become better individual contributors. If they can just make a tiny increment in how effectively the people they lead perform, and each of those people produces a tiny increment in the effect on all their subordinates, the leverage from that is massive.

With a class on cost accounting, students don’t arrive believing, “I’ll be able to just go with my gut and figure out the right accounting system.” But in classes on leadership and managing the workforce, often students do think, “I kind of intuitively know what the right answer is,” even though, oftentimes, those gut instincts are just as wrong as they would be with accounting.

So, to me, there’s great satisfaction in helping students understand, A, that organizational behavior is a lot more complicated than they may recognize, and B, there are frameworks and tools that help to navigate that complexity even if they’re not going to produce answers as precise as what comes out of the capital asset pricing model.

They come away sensitized to issues they otherwise wouldn’t have paid attention to. It’s gratifying when, at the end of a course, a student comes up and says, “You know, I didn’t want to take this course. I didn’t really think these issues were important, but now I understand that they matter a great deal.”