How to Fix Your Hiring Process

AI is transforming the world. And it starts with the right compute. Arm is the AI compute platform trusted by global leaders. Proudly nasdaq listed, built for the future. Visit arm.comdiscover.

Welcome to the hbr ideacast from harvard business review. I'm alison beard. Here's a question. Is your company good at hiring? Are you attracting strong applicants? Are you screening and interviewing them well? And ultimately, are you putting the right people in the right jobs? Since this is a special show for us, we're taping in front of a live audience at the Kimpton nine zero hotel in downtown Boston. I am going to ask our attendees. There are about 60 people in this room from a variety of organizations and industries. So those of you who would answer yes to all the questions I just asked, shout out yes. Ooh, that's a lot of yeses. Okay, now if you would answer no. Let's hear that no. Okay. So evenly split, sort of surprising. Some of you must work for unusually excellent companies. Others may be more the norm because according to our guest today, the vast majority of employers aren't using best practices when it comes to hiring. That means they are not only wasting time and money, but they're also missing out on a lot of high potential talent. Peter Capelli is a professor at the University of Pennsylvania's Wharton School of Business. He's also the director of its center for Human Resources, and he's the author of the HBR article, your Approach to Hiring is All Wrong. Peter, thanks so for joining us today.

My pleasure. Thanks for having me.

You guys can clap. So, Peter, we have a lot of HR leaders and hiring managers in the room. We've talked to a few recruiters before we started. In broad strokes, what are some of the mistakes that we're all making?

Well, the biggest one and the one that just completely stunned me is how few companies bother to check to see whether the people they're hiring are good or not. So there's at least three separate surveys that all found almost exactly the same number, which about 30% of companies checked to see or collected the data that would allow them to check to see whether they were hiring good people or bad people. And if you're not doing that, you know, this is sort of like, you know, driving and you don't know where you're going. You're not looking out the window to see where you're going. It's hard to imagine being good if you're never checking to see if you're doing well or not at this.

So what are some ways to measure whether you're making good hires.

Well, the simplest thing is you want to see whether the candidates you've hired are good ones. And then we have a problem of people saying, well, we don't have good measures of what's good. Okay, but if you don't try, you're not going to get anywhere. So a simple question to ask is just to ask the hiring managers, would you have hired this person? Again, that's a simple yes, no question. It's not hard to capture that data. If you believe your performance appraisal measures tell you anything at all, you could use those. You could look, how long does this person stay with us? You could look at turnover. So it's not like there is a single measure that is necessarily perfect. The more you've got, the better.

So once you collect that data and you know a bit about what makes a candidate ultimately a good hire, how do you then apply it to your hiring process, to your recruiting, to your screening, to how you interview candidates?

Yeah. So the technique for this is a very old one. So employers have been doing this since World War I. Industrial psychologists have been asking exactly the same question and been looking at it more or less the same way since World War I. And that is you look at the performance of the people that you hire and you look at the scores on whatever criteria you were using to hire them in the first place. Right. So if you were giving them various kinds of tests, now let's look at the test scores. If you're doing interviews, let's see if we can quantify what the interviewers think about the person. And then we look at the relationship between those scores, your interview scores, your test scores, your ability, whatever we're looking at, and how well people actually do. If you're not doing that, among other things, one of the problems you've got is you can't defend yourself against bias claims. Right. So if you get sued and somebody claims that your hiring practices have an adverse impact on protected groups, the only defense against that is to be able to show that your practices actually predict who's a good hire.

You talk in the article about how HR functions have been largely outsourced by a lot of companies in the past decade. Is that helping this issue? Are they doing a good job of creating metrics and pushing companies toward more research, or they making things worse?

It could, you know, the outsourced companies, the recruitment process outsourcers are typically big companies. Many of them are quite sophisticated. They could do a better job for you. What they say, though, is their clients Ask for two metrics. And the metrics are time to hire, how long does it take to fill a position, and cost per hire. Right. And that they're not asking about quality of hire. And so if you think about measuring you're hiring by how fast you're filling the job and how cheap it was to fill, imagine if we rated restaurants that way, right? McDonald's would win every year, year after year. If you're not looking at quality of the food, if you're not looking at quality of the hire, and you're just looking at cost per hire and time to fill, you're probably not going to end up with a lot of great hires because you're trying to do it fast, you're trying to do it cheap.

So who should be responsible for turning all of this around? Do companies need to rebuild their HR departments? Should hiring managers get more involved?

Well, I think it's probably important inside a company to ask the question, why are we not doing that now? Why have we stopped looking at what determines who's a good hire and why we stopped checking? Most companies used to do this a generation or two ago. Bigger companies, a lot of them still, still do it. But why are we not doing that? And I think the answer typically has to do with folks at the very top of the organization. They don't see the value. And a lot of times this is because the people who are running the companies have not had a lot of management experience themselves, particularly if it was a startup that's now gotten big. A lot of the people running US Companies, the preferred track to the CEO job is through finance. A lot of people who come through that track don't know very much about management. So they see the cost, but they can't see the benefits. I think it starts with persuading those folks that there really is something here you should take more seriously. And I think some of that is simply spelling out what are the costs of making a bad hire. Right? And here's where we do a lousy job on this. So we typically are looking only at the cost of if somebody quits, what are the replacement costs? Well, those aren't very big. I think the figure is about $4,000 per employee, something like that. By the way, we fill 66 million jobs in the US per year, though, so that's a lot. But the real costs are much bigger than that. Right? The real costs include the amount of time it takes people to figure out how to do the job well and how they're not performing well along the way. Right. The costs in terms of disruption. When you leave a position open, you start totaling those numbers up and you get to some eye popping figures. Right. And then if you start talking to the people at the top about the difference in performance between your best employee and your worst employee. So that's always an interesting question. Think about the worst employee we've got and then think about the best employee and how much more valuable is the best employee than the worst? How much does the best employee get paid as compared to the worst? Right. And you typically see the performance differences are enormously bigger, like five, 10 times bigger, and the pay differences are about 30%. And if you put it that way, you get this sense pretty clearly that we're way under investing in the value of having good employees, including trying to figure out who's good before we hire them.

Right. Let's talk a little bit about external recruiting. You argue in the article that companies do too much of it, but doesn't it behoove you to look both in and outside your organization to make sure that you're getting the best talent?

Well, I think it's useful to recognize that these are not equivalent decisions. That if you had two equal candidates, there's a big cost to going outside as opposed to inside. So my colleague Matthew Bidwell did a nice study of this a few years ago. So he went to some big banks and he looked at say, the director's position and he compared people who were promoted from within to director as opposed to people who were hired from the outside into director's jobs. He found people who were promoted from within. It took them seven years to make as much money as the people hired from the outside. Right. So it's not just that you're paying a premium to go outside. The premium takes a long time to wear down. And the people who are hired from the outside, it took them three years to perform as well as the people hired from within. So if you start totaling those costs up, it's enormously more expensive to go outside. And then if you start thinking about what happens to the turnover of the people below those positions. Right. So if I'm a senior manager and the next job up as director and that position gets filled from the outside, one of the things we know is turnover of those people just below jumps up. Right. Because I get the sense that the way to get ahead is to leave. So you start totaling all those things up and you see that we really ought to be biased toward advancing people from within rather than going outside. Even if you thought you had Equivalent people on the outside. It is so much more expensive to go outside. And then, you know, that that ignores too the motivation effects of being in an organization where you are retained when you think you can grow and develop. And you know, the main reason people seem to be saying they're quitting these days. What's the number one reason historically been pay? It's not pay, it's I hate my boss. And the second one has been I don't fit. Which is something that's almost completely preventable. You go into a company and gee, I'm surprised, I didn't know it was gonna be like that. That's cause somebody lied to you on the recruiting side. Right. But the reason we're hearing now from a lot of people is they're saying, I can't get ahead, so I'm leaving to find a way I can get ahead. Right. So if you start totaling up all those costs associated with going outside, you really should be pretty careful about making that decision to go outside. Right. So here's a stunning statistic from the article. One of the surveys, I think this was a LinkedIn survey that when we talked to, when they talked to talent acquisition executives in their own companies, only 28% of those thought that their internal candidates, that is our current employees, were a good source of talent for advancement. I mean, that's stunning, right? That's really stunning.

I'd love to bring the audience in here just to ask you about this idea of being poached or getting promoted. How many of you have been poached from one company to another? Okay, fair amount, 30%. I told Peter he would have to help me with the estimates.

I told Allison nobody knows if she's making it up either. It's radio. Right.

And how many people feel as if they can rise within their organization? There are lots of opportunities to advance. Okay, A little bit less.

Yeah. So 20%. Well, here's a couple of statistics which are eye opening on that. We hear a lot about kids, young people in particular, hopping from company to company. But if you look at the average 50 year old, 54 year old in the U.S. you know how many different employers they've had according to the census data? 13. Right. So are people moving from company to company? Yeah, and they've been doing it for a long time. If you looked at the people who changed employers last year, the majority of those folks said they were not looking to move. Right. So almost everybody who changes employers is enticed in some way, but most people were not actively looking.

And you Argue that those passive candidates aren't necessarily good employees once they get in.

Well, I think here's the complication is that these terms get used in lots of different ways. Right. So employers seem to like the idea of people who are not interested in applying for your jobs. Right. Because you know the Groucho Marx line, I wouldn't join any club that would have me as a member. Right. I wouldn't want to do that. And that somebody who's applying for a job, maybe there's something wrong with them, they want out of their current job. Right. But the evidence seems to suggest that that's really not true. That people who apply for jobs, the main reason they say that they apply is because they feel they're blocked in their current job. People who are not applying for jobs. Right. These are really passive. It appears that a lot of those people are not interested in moving because they're perfectly happy where they are. They don't want to go anywhere in terms their career either. And if you ask them what would cause you to move, it's money. So for people who are actively looking, they want career advancement. People who are not actively looking, they want money. And the people who are not actively. Who are actively looking, sorry, they're the ones who also report they're doing more to keep their skills up to date and to try to be kind of more marketable. Right. So I think we are spending a lot of attention and a lot of budget chasing people rather than maybe trying to look at people more carefully who are actually want to work for us. Right.

In a tight labor market, though, you have to expect that some poaching will be necessary for companies that want to grow.

Sure. Yeah. No, you got to go advertise, let people know what the jobs are. And, and you do have to persuade people to come. I think though, this is where we. I make big mistakes. So if you think about what could we do differently? I'd say the first thing we want to do differently is to be clear about our job requirements. That's the biggest basic problem. Right. And that is we have a committee, we have five people. We're asking about what are we looking for in this job? What should somebody have? And we end up just adding up everything that everybody said among the five people, and we end up with these joke kind of job descriptions that are impossible to fill. The five years of experience in a programming language that doesn't exist yet kind of thing. And one of the reasons companies have struggled to fill positions is they've been unrealistic about what they're looking for. But the second thing I think, where we make a mistake is that we don't tell people the truth about what it's actually like to work in our business. It's sort of like saying, don't let him see Peter because they're not going to like him, but you're gonna have to work next to him. So what are we trying to do? Trick you into working for us? And then they spring me on you. Right. And then you're just gonna quit and then you gotta repeat the process. Right. So the thing that we want to do, I think, is you do want to reach out and you are interested in getting people to apply for your jobs, but you want them to know what they're getting into. And the idea of trying to fill up the funnel, as they say, you know, that is try to get as many people as possible to apply, is a mistake. And the reason it's a mistake is because we're not good at sorting out candidates. And so what we would like to be able to do is get the applicants to sort themselves out before they apply. And that means if I'm pretty sure I'm not what you're looking for, I'm not going to bother to apply. Right. If this is the kind of place that tells us up front we don't have work life balance, I might as well tell you that up front if we really don't have it. Because you're going to find out. Right. And you really want that don't apply. Right. So we want people to screen themselves out up front so that you don't end up with people who are probably not good fits. And then you're going to try to figure that out with your interviews. And that's really difficult. Right. Because people are. This is a shock. Are not telling you the unvarnished truth about themselves when they're trying to persuade you to hire them.

Yeah. So you're talking about the advent of technology, social media platforms, job boards, algorithms that we use to then screen this wide funnel of people. Wasn't the idea behind all of that to expand the talent pool, sort of overcome some of the biases that we traditionally see in hiring, get beyond our networks?

No.

Okay.

No, that's not what happened. What happened was in the 1990s, during the last really tight labor market, companies were desperate to get people to apply. And so they were trying to make it easier and easier and easier to apply. Then we hit the downturn of 2001 and lots of people started to apply. And that's when companies had to start creating software applicant tracking systems to sort them out because they simply had so many applicants that they couldn't, as recruiters, look at them all. So then they were off on this slippery slope. It was never intentional. And then they couldn't get out of it. Right. So then they're still in this process now of having or during the Great Recession, you know, thousands, I mean, you know, 20, 30,000 applicants. Right. So I wrote a book about this called why Good People Can't Get Jobs during the Great Recession. You know, when companies then were reporting, you know, for atypical engineering job, they'd get 20 or 30,000 applicants. Right. And then the process of trying to screen those out became extremely difficult to. To do. So they didn't do it with the idea of trying to reach outside their network. That's a smart thing to do. You could do it purposefully, but just opening up the applicant saying, anybody apply? We're a wonderful place to work. It's not clear that that's going to give you the kind of diversity that you really want. What it is going to do is fill up your pipeline so that you're going to spend a ton of time trying to figure out how to screen people who you probably know. A lot of those people really don't fit your organization. You know, only 2% of applicants get interviews in typical companies right now. Right. So when we hear that, you know, companies are struggling to hire, it's not because there aren't a lot of applicants still, it's that typically they're the wrong applicants.

So apart from writing an effective job description, how do you find all those candidates who are suited to the job?

Well, I think so go back to the appropriate job description. That's part. And I think telling them accurately what the organization is truly like is another interesting thing. The cleverest thing I've seen on the selection side, which is an old thing that Texas Instruments did more than a decade or two ago, is they would give you on their website their test. That said, here's the test we use to try to see whether people fit or not. You can take the test and we'll give you your score and we'll tell you how people with your score have actually performed in the company before. We're not going to see your score. We're not looking at this at all. We don't keep it. We don't know it. There's the score. Here's what it tells you. Now, you decide whether you want to apply. And the thing that's so smart about that is that every applicant costs you money. Every applicant exposes you to legal liability. You'd really like to cut down the number of people who apply to only those who really kind of fit. Now from there then, you got this problem. You still got a ton of applicants, and how do you start weeding them out? And, you know, there are simple things. Do they actually have the qualifications, right to do jobs? And after that, we have a long history, literally 100 years of things that predict something about job performance. The problem is, this is the realm of personnel psychology. We got any personnel psychologists in the room? Good.

There was one in our audience listening.

Just cover your ears. And that is that we haven't gotten much better over the last hundred years in that line of inquiries in terms of trying to figure out what is going to predict who's going to be a good hire or not. But some of those things do predict. IQ does seem to predict. There are some things about personality, conscientiousness in particular, that does predict. And here's a stunning statistic. I think it's stunning. 40% of US employers do some kind of tests. These are big employers, but 74% do drug tests. Way more employers seem to be concerned about whether you're taking using drugs, including marijuana, even in states where it's legal now. Far more concerned about that than are in terms of giving you kind of objective tests. And here's where we could talk about algorithms for a minute, right? So the data scientists have an enormous hole that they have jumped into. If you look at the best predictors that they use, like if you looked at conscientiousness, how much does that explain of job performance? Well, they talk in correlation. So, you know, 0.3 correlation, but that means they're explaining about 9% of the variation in candidates. And that's assuming you don't know anything about the people. Once you start knowing things already, personnel explains less and less and less. So the data scientist guys go at this quite differently. They're saying we have no priors about what predicts. Right. Maybe personality does. Maybe your prior experiences in a job, maybe your height. I mean, we have no idea. We're just going to throw everything we know about you into a model, and we're going to come up with one answer, and that's a score. Right? And we're going to build this model, talk about training the data based on how well your employees in the past have performed. If you do it right, and then we're going to give you a score. And the score is going to be how close to the prediction are you in terms of mapping onto good performers in your organization? And that has the potential to be way better as a prediction device than what we're doing now, even with testing, is we got an answer about iq, we got an answer about personality, we got an answer about your GPA in college, which predicts nothing, by the way. So the data scientists have got a real opportunity to make progress here. A score, one number for each candidate that tells you how good a fit is this with people who have performed well before. Right. And that's why these guys are now, they don't do it very well. And the reason is because most of the vendors who produce these data science algorithms don't know much about employment. So a lot of the things they do don't make sense in terms of the logic of doing these. And a lot of the things they do risk violating law too. Right. So you got to be really careful with them. But my bet is those guys are coming in, they're pushing out the personnel psychologists, and they're probably going to do a better job because all they care about is doing a better job of prediction. That's it.

So, okay, we've whittled down our pool of candidates. We have people from inside, we have people from outside. We've put some appropriate screens on with our data scientists. And now we need to evaluate the five candidates we have. But you say that we, most of us do a pretty bad job of this too.

Yeah.

So what are we doing wrong?

Well, interviews are typically a waste of time. Right. So when I was department chair in the management department at Wharton, here's how we used to hire people, they would come in, they would do a serious thing, which is a job talk, which is like a work sample, if you're an industrial psychologist, that is. We want to see how you present. And some places actually have you teach a class, which is also smart. But then you'd go from office to office to office and you'd meet with people and they'd fill out a form about how they thought about you. Right. Whether they thought you were a good candidate or not. So each individual meeting is a different conversation with a different person. And then they'd sort of put them together and see what you thought. Well, we had a colleague who sort of studied this stuff and I asked her to speak to the faculty and she explained why this was completely worthless. Right. Because you're having completely unrelated conversations with, with different scales. Then you put them together and you get complete Noise. So this is what I explained to our colleagues. And how did we hire afterwards? Same way. And the reason is because even though everybody accepted the science of this, everybody thought, okay, I get that, but I can tell they said, right, this enormous overconfidence that I can tell. So the problem is we really think we're good at this, even when we're not. If we wanted to get good at this, here's a simple thing that works well, and that is you'd like everybody to at least hear the same questions, everybody to hear the same answers. You could do this as structured interviews. That's a complete bore, right? If you've ever been a candidate doing that, you meet five different people, they're asking you exactly the same questions, right? So the way to do it is just take the person to lunch, right? The five of you are there listening to the candidate. The candidate doesn't get to eat. But that's okay, you know, it's not for their benefit. And you're all hearing the same questions, you're all hearing the same answers. Your ability to evaluate then gets much better. But even then it only works well if you're asking questions that are really about their prior experiences, you know, asking about their career aspirations and you know, asking what they would take on a desert island. And you know, every top executive thinks they have the magic question that predicts, but they never look to see whether it does. So they persuaded that they're right. The best predictors are have they done something like this before? And how well have they done it? Those are the best predictors. But we all think we're geniuses at selecting people. And frankly, one of the problems in hiring is we've gotten rid of recruiters in a lot of organizations. We've really trimmed them out. So we let hiring managers do everything by themselves. Hiring managers don't hire enough people to ever get good at this. We don't train them and we don't give them feedback on how they did. So of course they're not going to get better. I mean, this is like trying to learn to play golf. If you're practicing at night where you can't see where the ball goes and you don't have a lesson.

So we've conducted excellent interviews. We have a group of candidates. How do we compare them against each other? Who should weigh in?

Well, the partners in consulting or in partnership firms, or the managing directors in investment banking, they do run the place and they do want to have a say in these things. And if they get to hire somebody that they really care about, they're maybe more likely to look after them afterwards. So it is difficult not to give these folks a say in the process. Right. So you know who should have a say? Well, the person who's going to supervise these folks probably should have some kind of say because you need their buy in. The problem is if they're not trained, they don't know what they're doing. You're just inserting bias into the process. Right. So you might say that you could give them a say when you've got candidates who are so close together that we can't easily differentiate them. My colleagues, Sunny Tambi and Valeria Yakubovich and I have written a paper about algorithms and these algorithmic hiring processes. Right. And one of the things we probably have to recognize is that in many cases, we really can't tell candidates apart. What we ought to do, rather than pretend we can maybe is just admit that, particularly for internal candidates. So we got two people who've applied for this job here, and it's a promotion, let's say, and we can only promote one of them, and honestly, we can't tell which one is better than the other. Right. And rather than kind of force ourselves to make a decision, let's just flip a coin and tell them that we flipped a coin.

That sounds horrible.

It doesn't, actually. There's a long literature on this. It shows people are much more able to accept that, you know, to say, well, they flipped a coin and I lost. Much more willing to accept that than somebody trying to make up a reason why you just weren't quite good enough. Because you probably know that actually you look remarkably similar. Right. So it's easier to accept. People find it. I mean, there's a lot of evidence on this, experimental evidence in particular, that people find it easier to accept these negative outcomes if they're told it was the luck of the draw, rather than there's something weird about you that we can't quite identify and tell you.

But, you know, we talk in HBR frequently about realistic job previews tryouts. Is that something that more companies should be doing?

Well, I think that would be terrific, right, to be able to let people actually experience the job. You know, the reason this became popular, internships in particular, we probably think companies were doing that as a way to look at the candidates, but they were doing it at least as much in some cases explicitly to let the candidates look at them. And the reason was they started to get concerned about retention. So in the 90s, this became a Big deal. We want you to know what you're getting into so that if we offer you a job, you know it's going to look like this. And so what doesn't happen is you come to workforce and then you say, oh, my gosh, I really don't fit here. Right. So if you're looking at exit interviews and people said, I didn't realize it was going to be like this, I don't fit here, that is something that, that's a mistake in the hiring process that you should correct and it shouldn't be that hard to correct it. By the way, exit interviews are really smart things, not a new idea. Right. Henry Ford was doing exit interviews in the 1920s with even hourly workers. But you shouldn't do them yourself, you shouldn't do them for your own employees because they're going to lie to you. One of the reasons why everybody thinks money is the big driver of turnover is because that's what people tell the boss on the way out the door, because they don't want to tell the boss, really, I hate you. That's the reason I'm out of here. I might need a reference or something later on. So they say money, because money is perfectly acceptable. It's the same reason that politicians say they've decided to spend more time with their family. Right. It's the same story.

Right, Terrific. So I think we have 10 minutes left. Thank you. And so Alex is going to walk around and we're going to open it up to you all for a Q.

And A. I had a question. You made the argument of hiring from within. What about if you're in an industry that's going through a large change or transformation or change. Can you sort of give some advice and thoughts about that? Hiring within versus hiring from outside. Yeah. So just in terms of the scale of this. Right. So back in the days of the great corporation, so before 1981 or so, you know, the, the GES, the General Mills, the General Motors, the great corporations, IBMs, they used to fill 90% of their vacancies from within. And at the moment, a typical company in the U.S. you know what the number is? It's about a third. So went from about 90% to about a third. So the bias is strongly. So we're filling 2/3 of vacancies from the outside. Now if you're changing your organization, I mean, I think one of the reasons we moved in that direction is because so many companies, from the CEO down, decide they want to be different and they think it's easier to dump the current workforce and go outside and hire another one. The problem with that is that you're never really becoming a different company. Completely different. Right. There are some things that you'd like to do differently, but it's still important to know how to do stuff in the old business. And by dumping the current employees and bringing in a whole new crop of new people who, who don't know the company, they come in with a different value set, they change the culture of the place. It's a pretty big cost that you're probably imposing on the place. There's a great story, and I think you guys wrote about this at IBM that Diane Gerson did, who's the head of human resources there, when they had consultants who they thought that they were going to have to drop because they were IT consultants, the technology wasn't being used anymore. So the idea was the typical model then was dump them, go hire new people. And she said, let's retrain them. And here's how we're going to do it. We're going to tell them you can work for four days a week, but we're only going to pay you for four days. On the fifth day, we're going to offer you training to develop the skills for the new jobs. You're going to do that on your own dime if you want to do it. And if you don't want to do that, that's fine. There's an exit package. Off you go. And most everybody took up the training and what they discovered, not surprising, is they got terrific consultants after this because they already knew how to be consultants. They already knew IBM. They were already people who had demonstrated that they were good employees and now they had a different set of skills. Right. So I think we are so biased in the other direction toward just dumping people and going out and trying to find new ones that we don't even consider the possibility of reskilling some of these people. It may not be that hard a thing to do. And there are ways to make it work financially as the IBM folks did.

Hi, thanks so much for a fascinating discussion tonight.

I've really enjoyed it so far.

So tonight's discussion, I think you've sort of been saying that when looking at positions, we're evaluating internal candidates and external candidates. These are all people who are applying for the job. What strategies are companies using to tap into current employees skill sets and sort of map them to other positions that might be available within the company?

Yeah, so that's a good question because we got rid of a lot of traditional job ladders. You know, where jobs were stacked in a particular way and you knew where the next job for you was because there was a clear hierarchy. So how do you let people know what jobs kind of make sense for them? We moved toward, in the 1990s, the last really tight labor market, these kind of really open internal posting systems. So at one point, I think 95% of US employers, large ones, said they had bidding and posting systems for almost all their jobs. The problem with that was first a good thing that people could apply for any position. And I had a PhD student, JR Keller, who's now at Cornell, did a nice study showing that when people hiring managers posted the positions, they made better hires and they made more diverse hires than when they knew themselves or they thought they knew who the person was who should fill this job. But how do you let people know what jobs make sense for them? This is where some machine learning algorithms appear to be popping up. There's a company started by a couple of Wharton guys called Quine. I got no stake in this company, by the way, where they're trying to do that. Basically what they're trying to do is to say people like you have moved to jobs like these before and succeeded. IBM does this for its own employees internally. Right. So they have looked at this and you're basically. The algorithms are basically just saying, what have people like this done before? Fidelity Investments here years ago did something really simple, really cheap, really smart. I don't know why they still do it. You could pick any job, at least in the management ranks, and it would show you the career paths that the last three incumbents in that job had that took them there. So there's simple things we could do that matter a lot. It doesn't have to be rocket science to give people advice. Right.

Let's do one more question.

Hi, good evening. My name's Everett from Philly. Kind of followed you up. Yay. But question about how big of a problem you think managers poor conception of fit is, as well as a shared understanding of what that even means within an organization and how that also even affects. Or is that tension with companies trying to be more diverse? That's a great question, and I'm glad we got it in because I think it's a real problem and I think employers have to think about this really carefully. Right? Because it sounds like such a neat, cool thing. We should be hiring people for culture fit. That's the phrase, right? Culture fit. And the problem, though is that companies that say that almost never articulate what their culture is, nor do they tell people how you would identify whether somebody is a culture fit or not. So it's just a great big truck full of bias. Right? Since I'm a hiring manager, I get to decide what our culture is. I get to decide who fits. And this is really important in the business. Right. So I think any company that really cares about diversity in particular, really better rethink that exercise. If you really think culture fit matters to so much, then you might want to think about do you really care about having people who are that different in your organization? It's kind of hard to hold both of those thoughts in your head at the same time. But if you're not articulating here's our culture and here's how you would know who fits and who doesn't, then it's just an open door for bias. And I think that's what we're getting.

Terrific. Well, thank you all so much for those terrific questions. I'm very impressed. And Peter, thank you so much. It's been such a fabulous discussion.

Thank you.

We appreciate it. That was Peter Capelli. He's a professor at the University of Pennsylvania's Wharton School of Business and the author of the HBR article, your approach to hiring is all wrong. We spoke to him live at the Kimpton Nine zero Hotel in Boston earlier this. This episode was produced by Mary Dew. We get technical help from Rob Eckart. Adam Buchholz is our audio product manager. Thanks for listening to the HBR IdeaCast. I'm Alison Beard.

Sa.