Diversity is not a nice-to-have

An interview with Joelle Emerson

Joelle Emerson (@joelle_emerson) is the founder and CEO of Paradigm, which partners with leaders of innovative companies to consult and advise on diversity and inclusion strategies. She has written extensively about diversity, inclusion, and unconscious bias, and her work has been featured in The Wall Street Journal, The New Yorker, Fortune, Fast Company, Business Insider, and several other outlets.

Before founding Paradigm Joelle was a women’s rights employment lawyer. As a Skadden Fellow at Equal Rights Advocates, she represented women in gender discrimination and sexual harassment litigation, and she advocated for local, state, and federal policies to ensure equal pay and other workplace protections for women. Joelle is a graduate of Stanford Law School.

Joelle Emerson is at the heart of the conversation on diversity, advising rapidly growing giants like Slack and Airbnb on how to develop the strategies and mindsets that will equip them to recruit and retain diverse teams. Using a research-heavy, data-driven approach, her firm is demonstrating why varied perspectives make companies stronger and training startups to effectively source them.

I had the opportunity to speak with her about why a culture of inclusivity actually yields better products, and how growing tech startups can leverage the lessons of the companies that went before them.

Elad Gil: Let’s start with the basics: Why is diversity important, and how can companies start thinking about it early?

Joelle Emerson:
There are a lot of reasons why diversity matters, and it’s important for every company that’s going to invest in this—invest time, invest resources, invest energy—to have a reason that’s specific to them.

The first reason is that there is a lot of research showing diverse teams are stronger when it comes to analytical thinking and complex problem-solving— basically the things that contribute to innovation. There’s also research that shows a correlational relationship between companies that have diversity and positive financial outcomes. I tend to find that people aren’t as persuaded by the latter (because it’s hard to know what’s the cause, what’s the effect, is there some third factor), but they are interested in the wealth of research that shows a causal relationship between adding diversity to teams and those teams’ ability to solve hard problems.

When it comes to small companies that are thinking about hard problems, they’re interested in building the team that helps them do that best. And we know that when you bring together people from different backgrounds— and this has been studied in terms of gender, in terms of ethnicity, even, as surprising as it sometimes seems, in terms of political affiliation—when you bring together those different perspectives, you can create a team that’s better able to innovate.

Then there’s a second category of reasons why this matters for some companies, which is that if you’re building for a diverse customer/user base, it’s helpful to have perspectives on your team that can help you design and build products or services for those users.

One example of this I really like is the story of when YouTube launched a mobile product in 2012 that allowed people to upload videos from their cell phones. They saw that about 10% of videos were being uploaded upside down, and they ultimately found out that that’s because left-handed people hold their cell phones differently. There were no left-handed people among the product or design team that worked on that product. I think we’re just better able to design for a broader group of people when we have people with different perspectives designing products.

The third reason is that the talent pool is generally very narrow. Companies are just generally struggling to find good talent to build their companies. People understand that, “Okay, it’s hard to find people, and if I rely only on my own network and I don’t think about building a company that attracts people from different backgrounds that’s limiting me logistically from hiring more people.” I work with a couple of companies that are actually primarily motivated by that, by creating a company that is the most attractive to the widest set of people so they can get the best talent in.

And then the fourth reason, which it’s important not to ignore, is that a lot of companies actually think this is the right thing to do, to include all kinds of communities in the design and development of technology and in creating the products that are going to define what our world looks like over the next 10, 20, however many years. They think it’s problematic that groups are being left out of that creation of technology.

Those are the four reasons I see fundamentally motivating the companies we work with. And I think they’re all good reasons. To start, you should know which one resonates with you so that you can design your strategy to be responsive to what you really care about.

Elad: What do you think are the biggest obstacles to companies actually being able to source and hire a diverse set of people for their teams?

Joelle: It’s helpful to break it down into the two parts you articulated, source and hire.

On the sourcing side, one big obstacle is that people’s networks are really homogenous, especially along racial and ethnic lines and along educational- background lines. Our networks are filled with mostly people who are similar to us, and when you’re at a fairly early stage you rely heavily on your networks to build your company. Companies that want to benefit from the value of diversity need to be willing to invest some time in going outside of their own networks.

I acknowledge that doing that is an investment of time. So any company has to decide how big of a priority it is for them to benefit from diversity. And they have to decide how early they want to do that, because we know that when you wait to start bringing in people from different backgrounds, it just gets exponentially harder to get that first, that second person in. Because who wants to join a company where 50 people look one way and you look a different way? It’s much easier when it’s five people or 10 people, 15 people.

A second barrier on the sourcing side is unconscious bias, and maybe sometimes conscious bias—that is, certain beliefs about who’s going to be a good fit for a role. With unconscious bias, our brains naturally take cognitive shortcuts and engage in pattern matching. So people think of who they’ve seen in a role like this in the past, and that can color who they then go look for.

Then there’s what I call the culture of brilliance—which I think is a really unhealthy part of the tech community—this belief in innate talents, innate abilities, the myth of “the natural,” looking for people who just “have it,” rock stars, any of these sort of words or traits. We know that when we think about abilities, talents, and intelligence as fixed, innate traits, stereotypes are much more likely to guide our decision-making than when we have more of a growth mindset and we think about talent and abilities as malleable— as things people can develop. Those are two really big barriers I see come into play on the sourcing side.

On the hiring side, once candidates get in the door, early-stage companies generally have pretty terrible processes for making hiring decisions. Usually they’re really subjective, ad-hoc, and different from candidate to candidate or interviewer to interviewer. We know that the more structure you have in place to guide hiring decisions, the more accurate those decisions will be and the less barriers like unconscious bias are going to creep in and influence outcomes.

For example, when companies hire for “culture fit”—even big companies, but especially companies that are 20 people, 30 people—I hear people define “fit” differently, and articulate different approaches as to how they assess for it. Some people are looking for “Who do I want to get stuck in an airport with for 12 hours?” And these are just really, really poor methods of assessing people’s abilities or assessing who’s going to be a good fit in the role. So I think one critical solution there is just creating some structure in the process. Be clear about what it is that you’re looking for, use job-relevant work samples, actually see how people do work, and then design interviews that are objective and structured to produce relevant information.

Elad: At Color we actually went through an exercise defining what we mean by “culture fit,” specifically to get rid of biases—not just from a diversity perspective, but because we wanted to make sure that everybody was aligned in terms of what we really meant by that. It was about emphasizing things like excellence and pragmatism and those sorts of traits in terms of what we were looking for, although those are particularly hard to grok sometimes.

Joelle: That’s great. And I hope that the Valley in general moves away from the term “culture fit.” The word “fit” almost primes us to think about “who is like people we already have on the team,” beyond the specific traits we’ve articulated.

I really like when companies think about it more in terms of values alignment or work-style alignment, people that are going to do work the way that work gets done in our organization. We’ve also worked with companies that call this “culture add.” “Who’s going to bring what we most need?” rather than “Who fits in with what we already have?” The approach to assessing for this matters more than what it’s called, but words do matter, and the way that we talk about these things can have an impact as well.

Elad: One other obstacle I’ve seen on the sourcing side is that—for my company and a lot of the companies I’m involved with—startups tend to source out of larger, more successful tech companies. And since those larger tech companies themselves seem to have a reasonably homogenous teams, that in and of itself limits who you end up recruiting. Because you start off with the numbers against you. How do you think about overcoming those obstacles? Or where should startups look alternatively? How do you start thinking about broadening the set of networks, to your point earlier?

Joelle: When you are looking at bigger companies within those networks that you’re sourcing from, make diversity intentional. Intentionally look for people that would add diversity. Different companies define diversity differently, it could be in terms of age, educational profile, ethnicity, gender. For example, we know some companies that spend 90–100% of their outbound recruiting, their active sourcing efforts, focused only on candidates from underrepresented backgrounds. They know that their inbound is not diverse, and so the only way to move the needle is in their active efforts.

Elad: What are some examples of companies that do that?

Joelle: The one company that I know that’s talked about it publically is Gusto. They committed I think 100% of their outbound sourcing to identifying candidates from underrepresented backgrounds.

Elad: One of the points that you raised that I thought was really key is around being at a state where you as a company can invest in people. Many companies want to do it, but even things like hiring more junior people tend to take a lot of bandwidth from the team when the company is still on the smaller end of 50 or even 100 people. So it seems really hard to invest in people until you really reach critical mass as a company.

Joelle: I agree that doing that is really hard when you’re early. As someone who has a five-person company, I know that is really hard. But even before you do that, you can achieve easier wins, like focusing your outbound sourcing on people who are from underrepresented groups, for example, or posting job descriptions on job boards and with organizations that have demographically diverse representation.

When you’re really early, yes, it’s hard. But the other nice thing is you usually only need one or two key hires to make a really big difference. The numbers needed to dramatically move the needle aren’t that big. So it’s an investment in getting a couple of those hires in the door.

“You usually only need one or two key hires to make a really big difference.”

– Joelle Emerson

I know there are candidates from underrepresented backgrounds that can meet early-stage companies’ qualifications if those companies spend the time finding them. And the return on investment, as companies benefit from those employees’ networks as they grow, will be so worthwhile. And then as you grow, then yes, all the things you said about thinking about growing people and having a wider sense of what you’re looking for and that sort of thing, those things come into play.

Elad: You’re making a great point. Because if you’re really growing off of your network early on, then you should be consciously asking yourself who you want in your network early on, so you can grow off of that. Really thinking about it as a foundation early is important.

How do you think about diversity at the board level, which I think is a whole different ball game in some sense?

Joelle: It is a different ball game, and we actually don’t spend a ton of time working on this, mostly because early-stage companies we work with are like, “We’re not adding anyone to our board, period. We don’t want a big board. We want it to be us—founders—and a couple of key investors. We just don’t want to add anyone, at all.” Okay, well, that’s a conscious decision you can make. If you don’t have any investors who are from underrepresented groups, and your board members are only you and investors, that’s tough.

But as you grow, and you think about adding people with specific areas of expertise, thinking about diversity is incredibly important, and I think the only way to really do it at the board level is just do it. To actually just say, “We believe it’s critical that the people issuing guidance to our company represent different perspectives.”

There’s a lot of research—again this is correlational research—that shows positive financial impact, and less of the types of damaging decisions companies can make financially, when you have women, in particular, on your board. So I think you just have to say, “We’re going to do this, if we’re adding a board seat. This is what our board looks like now, so for this seat we need someone that brings these talents and skills and that would add diversity, and that’s what we’re going to consider.”

Again, it’s usually a matter of adding one person, so you have to take time. The hard thing with companies that are growing quickly is a lot of this stuff just takes a little bit more time. I think the benefit is worth it. But there are more men that have served on boards, so if you’re trying to fill a board seat quickly it’s just going to be much easier to find a man than a woman. But if you want the best person and you want to create the best board, you just have to commit to it and not waver as people send other candidates your way.

Elad: Jack Dorsey actually spent a lot of time on this, both for Square and now for Twitter. So I know that if you can prioritize it then you can accomplish it. To your earlier point, it strikes me that where it starts is on the investor side. Because if you don’t have any investors from underrepresented groups, then effectively you’re not going to build those relationships early on. Often what I’ve found is that the first independent board member will often be an investor or other advisor or friend of the company who’s already somehow involved, either formally or informally.

Joelle: Yeah, I agree. And there’s so much responsibility on the shoulders— or there should be responsibility on the shoulders—of the venture capital industry in general to do better. I think tech companies have taken this much, much more seriously—or have begun to take it much more seriously— this kind of thoughtfulness around diversity and inclusion than VCs have. And that’s really, really hard, because it’s the VCs who help you build your network. They’re sending you candidates, they’re sending you recommendations for board seats, and if they don’t have networks that are diverse it’s going to make it much harder for you. So I’m hoping that as tech companies start to ask for this more, VCs respond. That’s what we’re beginning to see: tech companies are asking their investors to help them with this. And then we’re hearing from their investors saying, “How do we help them with this?” So that’s good, and I hope that that trend continues.

Elad: What are the best approaches that you’ve seen, for people to either get rid of unconscious bias from the job application and interview process?

Joelle: Structured interviewing is the one thing that I think is completely critical to doing that effectively, to actually hiring the best people and increasing diversity. I think of this as a four-step process.

Step one is articulating the relevant qualifications for every role, technical and non-technical (if there are non-technical qualifications).

Step two is designing specific questions to assess for each qualification. It’s important not just to be clear about what matters, but be clear about how you’re going to determine whether someone meets that qualification, and that everyone is aligned on what questions to ask for each competency.

Step three is limiting the domains that each interviewer assesses. You shouldn’t have to go in and try to decide, “Should we hire this person?” as an interviewer. You should decide, “Does this person meet what we need on these two things?” Because what we know is when cognitive load is heavy, that’s when people are most likely to take shortcuts. And when you’re trying to assess people along five different lines, that’s just really hard to do, for any person, especially in a 30-minute, hour-long interview.

Then step four, and this one’s really critical, is creating rubrics to help interviewers evaluate answers to the questions that they’re asking, or to grade work samples that they’re getting. A rubric can be as simple as: A great answer will hit on these three things, a medium answer will hit on maybe two of those things, a bad answer will not talk about any of these, or maybe it’ll hit on one. These kinds of things really help anchor people in what they’re looking for in interviews, and they limit the common biases that we know affect hiring decisions, like confirmation bias, where you answer a couple of questions really well and I’m primed to look for all the things you do well the rest of the interview. Or I’m really impressed by your resume, so I’m primed to really think you’re going to be great and pay attention to the things you do well. When you’re anchored around the rubric, you’re much better able to focus on the answer and figure out if the answers each person is giving are actually good. So that’s one thing that I think is just critical.

And then if you’re growing, think about how your job descriptions look: How are you attracting people? How are you talking about your company, and where are you posting jobs? Are you writing job descriptions that are likely to attract a narrow slice of people because you’re using, for example, gendered language? Or are you talking in terms of innate abilities like brilliance?

Research shows, in particular, that women are less likely to apply for jobs that are described in terms of a fixed mindset, like brilliance, rather than a growth mindset, like learning and growth. And then where are you posting your jobs? Are you posting them on sites and in places where people from different groups are most likely to look?

Elad: There’s bringing people in the door or hiring people, and then there’s integrating them, mentoring them, creating an environment where a variety of people can thrive. Have you seen companies fail on that front? And if so, how can they avoid that?

Joelle: Most companies really struggle with that, even the ones that are doing hiring decently well. It’s hard. But thinking about how to create a culture where everyone can feel like they belong is really critical, and not just a nice-to-have.

There’s a lot of research that shows that when people question whether they belong, their work performance suffers, their engagement goes down, they’re more likely to leave—all of these bad things happen. What companies typically do is build a culture that works well for the dominant group, but doesn’t work well for people that may not be like the people already there.

There’s a lot of research on a topic called “ambient belonging.” That is, what are the little signals that we’re getting from the world around us that tell us whether we belong or not? This can be really small things like the posters on your wall, or the types of social events you host. Are all your networking events or social team events happy hours after work? Well, if so, that sends a pretty strong message to people who may have obligations outside of work or who may not drink. Or do you also have events and activities that include people who might come from different backgrounds or who might not be able to do the things that you’re doing? Is your office designed in a way that feels like a place anyone could feel comfortable?

I will say this: I go into tech company offices each week, and about 30% of the time I think, “Wow, I would not fit in here,” before I speak to anyone. Just by the way the office is designed, the things I see on the wall, alcohol all over the place, ping-pong tables. Sometimes I look at tech company websites, and I see that the images they use to show their benefits are pictures of men. They say that they have gym benefits, and the “gym” icon is a man’s arm lifting weights. Those things send a message to people. And it’s easy to miss if you’re from the group that is represented, but it’s pretty powerful. So starting to think about how to build a culture where people can do their best work is crucial.

The other thing we see companies really struggle with is new managers, or inexperienced managers. A lot of people in early-stage companies haven’t managed other people before, or don’t have a lot of experience doing that. So they’re just not great at doing basic manager things, like giving feedback. And giving good feedback, feedback about the process—what went wrong, what went well, how can you do better.

Early-stage companies say, “Well, we can’t really develop manager training. We don’t have the capacity for that. How do we do this?” One thing that you can do that’s really easy is use a vendor and provide coaching. Require managers to have a certain amount of one-on-one coaching per month, or whatever might make sense. Just so that you feel like every manager on your team, even though you can’t provide it yet, is getting some guidance and some direction on how to give feedback.

You can also create sample tools or templates that show people what good feedback looks like. Sometimes little things like that are so overlooked, and they’re really helpful.

You can show what bad feedback looks like. For example: “You’re a good communicator. You did a good job on that project.” In contrast, here’s good feedback: “You communicated well by keeping people up-to-date on the status of your project and being helpful to your coworkers as you worked through challenges.”

Show people what it actually looks like to give process-oriented feedback that helps people grow and develop. I think coaching and using external vendors, external coaches, can be super helpful, and you can do that at any stage of your company.

Elad: Outside of people outright quitting, are there any warning signs or ways to gauge that you’re creating an environment that is going to drive out certain people over time, or not help people meet their full potential at your company?

Joelle: When you’re bigger—I’d say once you’re 50 people—you can start using surveys. And I think surveys are really helpful, and they actually can be very predictive. There’s research that shows—some things you don’t really need research for—but there’s research that shows that one of the best ways to predict whether an employee is going to leave is if they say they’re going to leave when you ask them in an engagement survey whether they plan to be there in a year. So you can actually start to gauge how people feel about your organization by asking them.

As you get bigger, you can survey along demographic lines too, and then cut results by different groups. You need a certain n-count to do this and keep it anonymous, but once you have that you can start to see, “Oh, do women consistently feel differently about things like the quality of their managers than men?” “Do African-American engineers in our organization feel differently about advancement opportunities than white engineers?”

You have to be a certain size to be able to do that. But before you’re that size, you can still create regular opportunities for employees to communicate with senior leadership about how they feel about the company. You can have regular brown-bag lunches that are specifically designed to talk about culture and how people feel. You can have monthly, or even quarterly, events where the leaders of the org get together with people to just ask about how they’re feeling. You can also hold office hours, where once a month you set aside two hours and you say, “This is specifically designed with a focus on building an inclusive culture, so please come talk to me if you think there are things we could be doing better.” These are things that are pretty easy to do even at a really small stage.

One other thing companies need to think about—even if they’re not going to do this early—is how they are going to start measuring and rewarding performance. I think companies typically wait too long to put processes in place for this. Usually it’s because they’ve had a terrible experience with performance review processes in the past, because most of them suck. But what we know is that if you don’t have any process in place, it’s really hard to make outcomes fair, because people just defer to biases. People’s brains are really bad at relying on data to make decisions.

“One other thing companies need to think about is how they are going to start measuring and rewarding performance.”

– Joelle Emerson

So actually thinking, “Okay, we’re going to start that process at 50 people. And before we have that process, we’re going to have some rigor in our leveling decisions or promotion decisions or comp decisions in this way. We’re going to ask these three questions.” Having even a light amount of structure, and then plans for when you’re going to roll out more, is really helpful. It’s important to outcomes not only that organizations be fair, but that people perceive them to be fair. When there’s no process that guides outcomes, people question that a lot. So just having some clarity around how decisions are being made, and how they’re going to be made as you grow, is really important.


This interview has been edited and condensed for clarity.