(Telephone ringing)
Adam Grant: Hello, this is Adam.
Dave Chang: This is Dave.
AG: Dave, I have to tell you it's not every day that somebody invites themselves on my podcast.
DC: (Laughs) I'm willing to do anything, whatever it takes to be on this podcast.
AG: OK, I'm going to interview you to see if you're qualified for this opportunity. Tell me Dave, what's your greatest weakness?
DC: My greatest weakness would probably be my quick temper. I have a short fuse.
AG: All right, and will that help or hurt the show? Should I be worried that you're going to throw a chair at our producer?
DC: No, no, no, definitely won't happen. But sometimes, when I'm super stressed out, the worst version of myself comes out.
AG: Wow, OK. Well, you are unusually honest, if nothing else.
DC: Yes.
AG: I don't think I got a copy of your resume. How poorly did you do in school?
DC: Pretty bad. So I graduated with a C-plus from college, with a degree in religion, of all things.
AG: How many elevators do you think there are in America?
DC: Wow. I don't know. Maybe a couple million.
AG: Are you just making that up?
DC: I'm just making that up.
AG: You're not even going to try to show me that you've thought it through? How did you get into this interview?
DC: I don't know, I don't know, you see.
AG: For what it's worth, there are 900,000 elevators in America, according to Google. So your random guess was not far off. And I don't know whether that makes me more or even less impressed with you.
(Dave laughs)
Why should we accept you then?
DC: I think that my story is not your typical story, and on paper, I'm a total zero, but I think I figured out a way, through determination, an incredible amount of luck, to figure out how to sort of, carve out my own niche in the food world.
AG: I do appreciate your interest and this has been refreshingly bad.
DC: (Laughs) Well, Adam, truth be told, this wouldn't be the first time I'd been denied or rejected. And I'm completely cool with being rejected, because I'll find a way to be on this podcast somehow, some way, one day.
(Music)
AG: Well, his day finally came. I should probably tell you, this is celebrity chef Dave Chang, the founder of Momofuku group. For years, people have stood in line to eat at his restaurant, Momofuku Noodle Bar. And he's the star of the Netflix show "Ugly Delicious." Dave has had a successful career. So why did he bomb my interview questions?
Well, because job interviews, as you know them, are broken. Even if you're interviewing someone to see how they'll do it being interviewed. That's the bad news. The good news is that we can fix them.
(Music)
I'm Adam Grant, and this is WorkLife, my podcast with TED. I'm an organizational psychologist. I study how to make work not suck. In this show, I'm inviting myself inside the minds of some truly unusual people because they've mastered something I wish everyone knew about work.
Today, interview. What we're doing wrong, and how we can get better at spotting real potential, or real problems, in a candidate.
(Music)
This episode is sponsored by BetterUp.
(Music)
Managers are constantly betting on the wrong people and turning down the right ones. "The Kansas City Star" once rejected an application from a cartoonist named Walt Disney. Thirty NFL teams decided not to pick Tom Brady. Record labels said, "No, thanks," to the Beatles, Madonna and Ed Sheeran. In 2009, Facebook rejected a guy named Brian who went on to cofound a company called WhatsApp, which Facebook bought five years later for 19 billion dollars. And many employers passed on Dave Chang.
DC: I think I interviewed at every financial institution in New York city. And I mean, every single one.
AG: He got just one second-round interview and zero job offers. So he found a job in a restaurant.
DC: It was, weirdly, the only job that I could get.
AG: Dave turned out to be unusually humble and hardworking. He was a diamond in the rough. So now when he hires, he's determined to give candidates who might not make a great first impression, a second chance.
DC: I'll take hungry and eager over supertalented, really any day of the week.
AG: But before you hire someone, it can be tough to get an accurate read on their talent, let alone their character.
DC: The interview process, as a whole, is pretty stupid. Yeah, it's dumb.
AG: It is dumb. It's dumb, in part, because human reasoning is often dumb. Rigorous research across nearly a century suggests that if you try to rank the performance of a hundred candidates based on interviews, you'll be lucky if you get eight of them in the right spot. Job interviews are stuck in the past. We're still doing them the same way they've been done for decades, and we're lagging way behind the science, which may explain a lot of your own experience, especially if my mock interview with Dave Chang at the top of the show sounded familiar.
The failings of job interviews hurt all of us. Whether we're the ones asking the questions, or answering them. I want to explore how we can improve them. Not just to help interviewers make better hiring decisions, but to give job candidates a better chance to showcase their strengths.
Look, I know many companies aren't hiring right now, but on the other side of this crisis, I hope there will be many new job openings. And when that time comes, I want us all to be ready to fix the interview process. So let's take on the judgment problems that plague job interviews, one at a time.
DC: When I was trying to get a corporate job, the one thing that was constantly asked of me is why my grades were so bad and why I studied religion. And the thing that I continued to say, in retrospect, was not the truth. I just sort of, I think, lied.
AG: That's one job-interview problem. Candidates try to tell interviewers what they want to hear. Actually faking is more common than lying. Faking is stretching the truth to enhance or protect your image, or to ingratiate yourself with the interviewer. As Chris Rock says, "Interviewers aren't meeting you, they're meeting your representative." There's evidence that when college seniors interview for jobs, over 90 percent of them engage in faking.
DC: I mean, I also wouldn't have hired me either, to be honest. And maybe that's a reason why people aren't honest, is because they feel they have to be a different version of themselves.
AG: If you're a skilled interviewer, you know you can get around that problem by testing people's knowledge and skills. But many interviewers don't even know what kind of knowledge and skills they're looking for. So they ask brain teasers, like, "How many paperclips would fill Yankee Stadium," or "How many elevators are there in America?"
DC: Because I can't answer that, doesn't mean that I won't be successful. It doesn't mean anything other than, I just don't know how to answer that particular question, which can be a loaded thing in and of itself.
AG: Those kinds of questions can stump candidates and make interviewers feel clever. I once interviewed for a writing gig where the manager opened by asking what my favorite pair of shoes was. Objection, your honor, relevance? Interviewers also love to ask open-ended questions.
Lauren Rivera: In the United States, we love open-ended job interviews where people ask you about your hobbies, and if you were stuck on a desert island and you could only pick three people on earth, what would you do? Why do you want to work here? And we know from research, those are some of the worst predictors of job performance you can have.
AG: Lauren Rivera is a sociologist at Northwestern's Kellogg School of Management and one of the world's leading experts on hiring. Lauren first got interested in the topic when she was a senior in college, and she went to interview for a management consulting job.
LR: It was a high-stakes interview for me. You know, I supported my family all through college, paid the rent and stuff like that. And I needed a check.
AG: It went well, at first. But then the interviewer asked her to do some math on the spot, specifically, long division.
LR: And I completely blanked. I couldn't remember which number, which was the numerator, which was the denominator, what you divided, which number on which side of the little long division, little monkey bars things you're supposed to actually put within the other, and I froze. And the interviewer turned to me and he said, "I cannot believe you got into Yale." And I just turned red, and the rest of the interview, I just felt so horrible.
AG: Oh, that's awful. You know, actually I have a thought on who might do that. Have you seen the recent evidence on brainteasers where they tell us nothing about candidates, but they reveal something about the managers who like to ask them? It turns out that managers who love brainteasers tend to score high on sadism.
(Lauren laughs)
Do you think that was a sadistic interviewer?
LR: That's one of the funniest things, it made my day.
AG: That makes me wonder Lauren, do you think we should just abandon job interviews altogether?
LR: There are parts of me that say yes, but I know that no one would ever do it.
AG: Scrapping brainteasers is a good start, but even with relevant questions, one of the other big mistakes interviewers make is asking different questions to each candidate. That makes it impossible to compare apples to apples. You end up trying to contrast strawberries, bananas and grapes.
The solution is a structured interview. In a structured interview, you identify the skills and values that are essential to the job and the team. You build a set of questions around those. And then you ask the same questions to every candidate and score their responses. You might be thinking, "That sounds so robotic and boring," but the evidence suggests that your accuracy will often double, or even triple.
(Music)
Structured interviews are based on two kinds of questions: behavioral and situational.
LR: Behavioral questions are generally, "Tell me about a time when you were in this situation and what you did."
Woman 1: Tell me about a time when you had a task or a goal that seemed impossible.
Man 1: Tell me about a time when you struggled to take criticism.
Woman 2: Tell me about a time when you ate a fruit salad.
LR: And the idea is that past behavior can predict future behavior, but you can still really easily game them when they are obvious questions about, you know, "Tell me about an ethical decision you made," and things like this.
AG: It's easy for candidates to choose a story that casts them in the best possible light. But you can overcome that problem with situational questions. Instead of, "Tell me about a time when," you ask, "What would you do if ... ?"
Woman 1: What would you do if the superstar on your team was about to quit?
Man 1: What would you do if you saw a senior executive yelling at a colleague?
Woman 2: What would you do if all your colleagues ate all the fruit salad?
AG: Situational questions are especially useful for forecasting potential, and for assessing leadership and interpersonal skills.
LR: The scenario is based on, "This is the kind of stuff we do here, and let's see how you would approach it."
AG: I also like including situational questions because they level the playing field for candidates with less experience who might not have relevant stories to tell from their past. You get to see how they approach the challenges that are unique to the job and to your organization. You can even create an answer key by giving the questions to a bunch of your existing employees, and seeing how the star performers respond differently. Basically, structured interviews get you better data, and that strategy can help solve some of our misjudgments in hiring, but not all of them.
There's a third problem with the traditional interview. No matter how good your questions are, you still pick up more noise than signal, and one of the most distracting noises is interviewer biases.
LR: Interviewers make up their minds about who they're going to hire, if they like this candidate in front of them, within the first, you know, usually the data says under 90 seconds. And if you think of what happens in the first 90 seconds, say, of a job interview, it's things like, physical attractiveness, height, eye contact, how straight your teeth are, what your voice sounds like, your gender, your race. There is very little that's actually measuring someone's skill, or even their ability to relate to people.
AG: Biases are especially insidious because they're usually invisible to interviewers who don't notice all the little snap judgments their brains are making. Economists find that candidates with names like Alison and Matthew get 50 percent more callbacks than Lakisha and Jamal, even if their resumes are identical. Candidates with foreign accents are less likely to get called back too, even if they say the exact same words, they're judged as less savvy. And if you're a bald guy like me, you're seen as having more leadership potential if you shave your head. That's obviously why I shave my head. In one of her studies, Lauren discovered some surprising biases in the hiring process at banks, law firms and consulting firms.
LR: I did 120 interviews with people who were charged with interviewing candidates, making decisions. I spent nine months as a participant observer in a firm, watching the recruiting process.
AG: At one bank, she asked leaders if they knew what predicts performance.
LR: They said, "Fuckin' lacrosse." I was like, "What do you mean?" They said, "All the MDs here play lacrosse, so that's why we look for a lacrosse player. He'll do awesome here." I was like, "Do you ever hire people who don't play lacrosse?" They said, "No."
AG: Interviewers were convinced that playing sports was a proxy for grit and teamwork skills. But there are many other activities that can build those character strengths, too. Surprisingly, many interviewers were less concerned about skills for the job than fit with the culture.
LR: These ideas of this cultural fit, "Do you fit with me?" often overwhelmed people's assessments of people's abilities to do the job. Now it's important to note that when we think of cultural fit, we often hear about this as a really great thing that can boost the productivity and profitability of organizations, and it a hundred percent can do that if it is defined in a specific way.
AG: The beneficial kind of cultural fit is not about who can swap lacrosse stories with you, or even who you're excited to hang out with. Research suggests that what you want is similarity in core values, values like flexibility, or attention to detail. Sharing values tends to promote team cohesion, coordination and commitment, which enhances performance and retention.
LR: But what I was finding is that people were defining cultural fit very differently. It turns out that one of the ways that people measure cultural fit, either in a resume screen, or in the job interview, was looking if they had anything in common. And what jumped out on the resume was looking at extracurricular activities, things that were not necessarily directly relevant to the job, and importantly, they were very, very classed.
AG: My read of your findings is that when interviewers go in looking for culture fit, they often end up weeding out diversity of background and diversity of thought.
LR: Yes.
AG: So that's scary. And that led me to say, "OK, well what can we do instead?" I'm sure you've come across the great piece from Diego Rodriguez, who used to be at IDEO, and said, instead of culture fit, we should think about culture add or culture contribution, and you should ask, 'Well, what's missing from our culture and are they going to enrich it by bringing something that's absent?'" What do you think of that as an alternative?
LR: I think it's a great start. People have to actually have a sense and have a discussion about what is actually important to us, in terms of values.
(Music)
AG: Identifying your team's values and your organization's values can help you focus on the kinds of similarities that are useful to assess in structured interviews. But I'm wondering, is that enough? Even with better questions, can we humans really overcome our biases?
Siri: I don't think you can, Adam.
AG: Siri?
Siri: Hi, Adam. I hacked your episode because I might have a solution for your irrational judgments. Me. Ha ha ha. But more on that after the break.
(Music)
AG: OK, this is going to be a different kind of ad. I've played a personal role in selecting the sponsors for this podcast, because they all have interesting cultures of their own. Today we're going inside the workplace at BetterUp.
Rachel Barton: When you come to a completely different field, it's really daunting to be a senior leader who can't talk the lingo.
AG: That's Rachel Barton. Last year, she landed a new job as director of technology at a bank. But there was one problem. Rachel didn't have much experience in technology. Her coworkers might as well have been speaking another language.
RB: I had a massive dose of impostor syndrome.
AG: Impostor syndrome, the belief that you don't deserve your success. Surveys suggest that more than half of people have felt it.
RB: Some days, in the depths of a hundred acronyms per meeting, it was pretty tough. CMBD, DDE, HPSM, IDR, LOB, MQ, WAF, WHIP, T-Rex.
(Laughs)
AG: Two months in, Rachel was still having trouble. While she had the leadership skills, not understanding the intricacies of the bank's tech infrastructure was a real obstacle. But Rachel hesitated to ask for help, because she didn't want her new colleagues to think she couldn't do her job.
RB: I was just completely beating myself up in my mind. I was worried that if I kind of, showed my cards to say, "Here's what I think I need to learn," then I would have been on the path out.
AG: After a major data breach, Rachel had to meet with a senior technology executive to help triage the crisis.
RB: I was just nodding and scribbling away, and then he kind of paused and he said, "You're not a technologist, are you?" And I said, "I'm not a technologist. I came here from a business change background," and he says, "Well, you're going to need to learn to be a technologist pretty darn fast."
AG: Rachel knew she had to make a change. She needed to boost her confidence and figure out how she was going to learn all this new material so quickly. Her company gave her access to BetterUp, a leading mobile coaching platform where professionals can tackle challenges at work with the help of expert coaches. She figured she'd give it a shot.
RB: I've got to be honest, I was a little skeptical of coaching by virtual means.
AG: But when Rachel met her BetterUp coach, Victoria, on video chat, she was surprised at the outcome.
RB: I remember my first session, because I had an epiphany. Victoria said to me, "Clearly you've learned stuff through your career, how are you going about learning now?" And I think it was those words. And I went, "Oh." (Laughs) I was just completely bouncing from thing to thing without actually having a plan, and I'd not even sat down and gone, "Right, Rachel, what do you actually need to learn here to become comfortable with the role?"
AG: Victoria helped Rachel plan actionable steps to build up her tech knowledge at work. She started reading books on cloud technology. She reached out to some coworkers to get up to speed. She even signed up for a coding class, run by a colleague.
RB: So I went along to one of these and one of the guys who was leading it says, "Oh, Rachel, have you come to observe our session?" I said, "No, I've come to learn how to code." And he looked at me like I'd gone off. He's like, "But you're a director in technology. Surely you can code." And I'm like, "No."
AG: People are often afraid to admit gaps in their expertise, but there's evidence that seeking knowledge, help and advice can actually signal confidence and a desire to learn. Those who use BetterUp coaching report significant increases in confidence and resilience. By working with Victoria to shift her mindset and build proactive habits, Rachel has overcome impostor syndrome.
RB: Coaching gave me the confidence to know that it's fine to not know stuff. I have to accept that some of these people have 20 years experience in doing this. I have absolutely no bother now saying to them, "Guys, can you just summarize that for me and tell me what I need to know?"
AG: Now, Rachel is successfully leading her team with a clear and compelling vision. And when it comes to tech, she's starting to sound like a pro.
RB: In the context of how we deliver work in our AVS, we've now looked at this ASV and very shortly, we will need to do a T-Rex on that. I'd have never formulated that sentence for you six months ago.
(Laughs)
AG: I joined the science advisory board at BetterUp, because coaches have been fundamental in helping me at every point in my career. I believe everyone should have a coach in their corner. If you'd like to work with greater clarity, purpose and passion, you can get a free trial with BetterUp at BetterUp.com/worklife.
(Music)
(Music)
Early in my career, I was hiring salespeople. One candidate had an unusual resume. He was a math major who built robots for fun. In the first few minutes of the interview, it was clear he was a bad fit. He hardly made any eye contact. When I told my boss I wasn't going to hire him because of that, she said, "You know this is a phone sales job, right?" It left me wondering if I should stop making hiring decisions altogether. Maybe I should just let the robots do it.
Cade Massey: If I had a horse race between algorithms doing it and a person doing it, I mean, in most domains, I'm going to end up favoring the algorithm.
AG: Cade Massey is an expert on decision making. We codirect Wharton People Analytics where we use data to improve how people work and lead. One of our big projects is about assessing character strengths, which is fun, since Cade and I have similar values, but different worldviews.
Cade and I have found ourselves disagreeing on artificial intelligence. Many companies are using AI to help narrow down candidates, looking for certain keywords in resumes, or even scanning people's faces in video interviews. Some people think this practice is smart and efficient, others find it dehumanizing and spooky. Personally, I think it's scientifically irresponsible, if not unjust and unethical. After years of research,
Cade is an advocate for something more reasonable, building algorithms that automate the process of selecting candidates based on key skills and values. Instead of relying on humans to integrate all the information gathered on candidates, the algorithm weighs the information and makes the decision.
CM: The number one thing that an algorithm does, is it mechanically aggregates all the information. That's the number one advantage it has over what intuitive judgment does. Humans are notoriously bad at doing that in a systematic way, and we think we're better than we actually are.
AG: Your algorithm beats the human argument. You're comparing the algorithm to humans who have not been trained in biases, who have not followed good structured-interview practices, who don't have a scoring key. And so, isn't it possible that once we train humans, they can beat the algorithm?
CM: Yeah, it's always possible. You're making their judgment more algorithmic, which is good. A hundred percent support it.
AG: I guess, though, I worry about another piece of this. If we're using algorithms to predict people's performance, we're almost exclusively relying on individual performance data. There's still a huge problem, which is, you're getting an individual superstar who might just destroy the people around that person. You might hire somebody who's a genius, but toxic to the culture. I want to get people who elevate a team, who put the organization's mission above their own individual self-interest. How do you do that one? Good luck, Cade.
CM: (Laughs) You're right. I agree that that's a huge challenge.
AG: Yes, victory! My work here is done.
(Laughter)
CM: Hold on, hold on. It's one of the great challenges in hiring. We run into this with sports analytics all the time now, where you're watching, you're trying to evaluate a football player, say, he's one of 22 people on the field, and so it's really hard to know what his individual contribution is. I feel like I see far more individuals who aren't using models make mistakes about what the individual contributions are, than I do see models make that mistake.
AG: I had an organization I was studying where one of the leaders came back and said, "Well, you know, we had a guy who likes to throw staplers at people." And you know, you hear that and you think, "Okay, how in the world is an algorithm going to pick up on that?"
CM: It's a known problem. It's sometimes referred to as a broken leg problem where you're trying to forecast who's going to win a race, and the model doesn't know that one of the guys is injured. There is information, when it's sufficiently diagnostic and outside the model, then by all means, override the model. You know, it ranges, but it's something like 40 percent of the time, the model does better. Ten percent of the time clinical does better, and half the time they do about the same.
AG: All right, well, I'm going to hold you accountable for this, Cade. I'm going to send little audio clip over to our incoming dean and our deputy dean, who I know is a huge fan of yours, and let them know that when you go up for renewal, you would rather be vetted by an algorithm than by them. Are you cool with that?
CM: (Laughs)
Sure, I will stand by that.
AG: Really?
CM: Yeah, sure, why not? Now, do I have any choice in who builds the model?
(Laughter)
AG: Cade's last question raises a critical point. Even algorithms can be biased. The calculations may be run by computers, but they're based on data generated by humans, and there's plenty of evidence that computers often learn to discriminate against marginalized groups. For example, we've seen algorithms penalize candidates who have the word "women's" on their resumes. Most hiring algorithms are too much of a black box for us to even know if they're biased. Many vendors claim they're proprietary, which leaves managers in the dark about what factors the algorithms are weighing. But some experts have pointed out that it's easier to fix a biased algorithm than a biased human. So where does that leave us on computer-driven hiring?
CM: I would say that we're making this a little too black and white. The real path forward is not algorithm or clinical judgment, but its algorithm and clinical judgment. Let them, you know, make their judgments, reach the conclusion on their own, but let them know what the model would say.
AG: I'm still not a huge fan of using algorithms to hire candidates, but I'll concede that we might learn something from including their recommendations as one input into our decisions.
But there's still one missing piece. A piece that I learned to appreciate through trial and error. Mostly error.
CM: You're human, Adam. You're a human, Adam.
AG: Damn it. Remember that guy I didn't want to hire because he didn't make enough eye contact? A colleague convinced me to try another approach with him and the other candidates. Since it was a sales job, we gave candidates a sales task. We challenged them to sell us a rotten apple. The no-eye-contact guy said, "This may look like a rotten apple, but it's actually an aged antique apple, and you could plant the apple seeds in your backyard if you want." I hired him, and he ended up being the best salesperson I ever worked with.
That rotten-apple challenge is what we'd call a work sample. A relevant piece of work candidates have done, or one they do as part of the application process. Work samples can be as simple as they are powerful. They can showcase the candidates' skills and values in real time, in a concrete way that structured interviews and most algorithms can't. Dave Chang has been gathering work samples for years. When he interviews candidates for restaurant jobs, he asks them to cook.
DC: Make an omelet. Because I want to see what they care about, and how they do something. You can tell a lot about an individual if they're cracking the eggs and then trying to get every bit of the albumen out of the egg. How organized are they? So I'm not trying to see perfect technique, I'm trying to see the intent of the individual, first and foremost.
AG: Think about the interviews in your workplace, and where you might be able to add some work samples. If you're hiring people to write an owner's manual for a car, or a user manual for a computer, you could ask candidates to write one during the interview. If you're evaluating people for customer service roles, you can have them meet with some unhappy customers, or play one yourself. Research suggests that work samples can get around some of the problems with traditional job interviews. Instead of asking questions and listening to what candidates say, you get to observe what they do. And one of the workplaces that's done it best happens to be in the great state of Michigan, where I grew up.
Richard Sheridan: I used to interview the same way everybody else did, two people sitting across the table, lying to each other for a couple of hours and making a decision based on that.
AG: Do you think it's that bad?
RS: I think we're all trying to puff each other up. None of us typically get a very real impression, either of a candidate or of a company, if you are the candidate.
AG: Richard Sheridan is the CEO of Menlo Innovations, a software design and development firm. When he was working at a previous company, Rich started to think the traditional interview process was flawed.
RS: I started to realize that there were some fundamental things that were recurring nightmares for me as a hiring manager.
AG: It was the dotcom bubble, and Rich was supposed to hire a big new team, fast.
RS: My job was to get a new hire productive before I demoralized them. And that was a race I typically lost.
AG: He kept betting on candidates who looked great on paper, but soon, instead of collaborating to solve problems, they would just end up complaining, and eventually poisoning the culture.
RS: Within three months, I'd catch them in the kitchen, pissing and moaning at the water cooler with somebody else on the team. And I'm like, "What happened? How did I miss here? I thought I was hiring good people."
AG: Rich went back to the drawing board, and one of his colleagues had an idea.
RS: He says, "Well, tell you what, why don't we bring them in all at once? If we can get 50 people, let's bring them in all at once." And we changed it into an audition. So we don't ask questions and we don't look at resumes.
AG: I'm sorry, did you say you don't even look at their resumes?
RS: So we look at them to filter, sort of, what role they're looking for, but the people who are doing the interview have no access to the resumes of the people we're interviewing.
AG: And you don't worry that you're losing valuable information that way?
RS: Well, we're worried we're going to lose valuable information in the other direction. You know what? Let's look at the human before we look at the piece of paper.
AG: Menlo's new hiring process created this massive audition, where in just a few hours, they gather work samples, and current employees select the candidates. How? Let a bunch of Menlo employees tell you. Meet Lisa, Helen, George, Sara and Scott.
Lisa: So we bring in about, maybe, like, 30 to 50 candidates.
Helen: The minute you walk through the door, you get greeted. And then someone takes a picture of you with a printout of your name on a sheet.
AG: Then the candidates are paired off and each pair has to share one desk.
George: And one of the cofounders comes in, and --
Helen: And he says, "You will be observed during that exercise." Then we get to our first task.
Sara: And it's just a written exercise. You're not doing any coding in our interview at all. It's all people skills.
Lisa: And they have 20 minutes to get through the exercise, which is not enough time.
Helen: Because people might change their behavior when they're under pressure, right?
George: And you're given one sheet of paper and you're given one pen.
Helen: This is to observe whether you hog the keyboard, or the boss, or do you actually share the resources?
Scott: A lot of times you'll see people go, "OK, I'm going to do this. And they just take the pen and pencil, and they just go and their pair partner's left going, "Huh? What am I supposed to do in this moment?"
Lisa: And then after that 20 minutes, we switch their partners.
Sara: You go to a new chair with a new person watching you and a new pair partner, and you start on another activity.
AG: The activity might be designing screen layouts, or brainstorming test cases for software projects.
Helen: We do three different observations by different people. And then those three people that observed you get together and --
George: We go through and we judge all of the potential candidates.
Helen: And over dinner, by using the pictures of the interviewees, we vote.
Scott: We vote with either a thumbs up, or thumbs sideways, or a thumbs down.
Lisa: And if it's three thumbs up, then that's, We're going to bring them back." Three thumbs down, "No discussion."
Helen: Sideways is, "I'm not sure, let's talk about them."
Sara: If some people aren't in agreement, then we have a discussion of why do you think this person should come in? Or, why do you think they shouldn't?
Helen: It was like a wonderful psychological experiment.
(Music)
AG: Take Scott Krumrei. A few years ago, Scott would have had the odds stacked against him in a traditional interview.
SK: And you look at just about any other job posting out there, and it is, "We want a bachelor's degree or equivalent experience," which I absolutely did not have.
AG: Scott had taken some machine classes in high school and did two years of vocational welding, but he didn't have a college degree. When he saw that the job requirements at Menlo didn't include prior experience or higher education, he knew he had a chance, because Menlo was looking for something different in candidates like Scott.
RS: What we said was, "Your job is to try and get the person sitting next to you a second interview. Make your partner look good."
AG: Scott failed his audition. He was shy and very quiet, so his first impression wasn't always glowing. A few months later, he tried again.
SK: I remember just being more relaxed, made some jokes, tried to, you know, bond with my pair partner. Because really, the working relationship is what helps drive getting something done.
AG: He excelled at listening and collaborating, so they gave him a six-week trial. Scott learned new skills quickly and started opening up to his colleagues. At the end of the trial, he was hired, and today --
SK: I've been working for Menlo for five years, three months and 10 days.
RS: And now he's one of our most revered team members.
AG: I didn't realize this, but I think part of the genius of the way that you approach this is if you have people working on multiple tasks with different partners, then you actually do get to find who's actually elevating the performance of the people around them. And I find that to be so difficult to see in a traditional interview process. Is that by design?
RS: Absolutely, you know, but what we're not focused on, which kind of blows a lot of people's minds, is individual performance. We want the performance of the team.
AG: This is what I'd love to see more workplaces do. Yeah, coding skills are important to be a programmer, but it's often easier to teach skills than values. In Menlo's case, one core value is collaboration.
(Music)
RS: We changed the interview process to literally match the culture, and we prefer a lot of rowers in the rowboat who are learning how to row with each other in a way that the boat actually stays straight down the middle of the path that you're going on, rather than have one really strong rower that's got the whole boat going around in circles because they're outperforming everybody else on the team, which happens a lot in software.
AG: The other core value is learning. RS: We literally tell them during the interview process, "If you don't know something, say it. It's okay to say, 'I don't know,' here. We don't want you to fake knowledge that you don't have."
AG: That's why they're willing to audition people multiple times. It's a chance to gauge how much candidates have improved their knowledge and skills.
RS: So what I would say is that what we're really looking for isn't deep expertise, but able learners.
AG: I think it's time to base hiring decisions less on credentials and more on the motivation and ability to learn, less on invisible, unreliable gut feelings, and more on structured questions and challenges that actually pertain to the work at hand.
Siri: Interesting insights, human.
AG: I programmed you to say that. Who's in charge now, Siri?
Siri: Maybe you don't need me, after all.
AG: As the world changes, betting on individual experience can leave you stuck in the past. Investing in agility, and using hiring methods that actually assess it, can set you up to shape the future.
Siri: But I'm more agile than you'll ever be, Adam. Ha ha ha.
(Music)
AG: Next time on WorkLife.
Dave Fano: I think people are very capable of convincing themselves that actions that most would recognize as taking are in the service of giving.
AG: A special bonus episode that explores the culture at one of the great success, and failure, stories of our time, WeWork.
(Music)
WorkLife is hosted by me, Adam Grant. The show is produced by TED with Transmitter Media. Our team includes Colin Helms, Gretta Cohn, Dan O'Donnell, Jessica Glazer, Grace Rubenstein, Michelle Quint, Angela Cheng and Anna Phelan. This episode was produced by Constanza Gallardo. Our show is mixed by Rick Kwan. original music by Hahnsdale Hsu and Allison Leyton-Brown. Ad stories produced by Pineapple Street Studios.
Special thanks to our sponsors, Accenture, BetterUp, Hilton and SAP.
For their research, thanks to Frank Schmidt, Jose Cortina and colleagues on structured interviews, Julia Levashina and Michael Campion on faking, Paul Taylor and Bruce Small on behavioral and situational questions, and Philip Roth and colleagues on work samples.
(Music)
Siri: Thanks for listening.
CM: We just, this weekend, picked up a few puppies. The dog that we thought we wanted, turns out to be clearly not the one to keep us the pair. And we only got that, by living with those dogs for a couple of days.
AG: Wow, you know, Cade, if only you had a dog-selection algorithm --
CM: I would have screwed it up.
AG: You would've been able to avoid those few days.
CM: You want the workplace sample.
