Transcript
Mark:
Welcome to People Tech the podcast of the HCM Technology Report. I’m Mark Feffer.
Mark:
My guest today is Eric Sydell, executive vice president of innovation at Modern Hire. They’re a science based recruiting platform with tools that touch multiple steps in the hiring process. We’re going to talk about data, why it’s important and how to use it on this addition of People Tech.
Mark:
Eric, thanks for joining me today. Obviously there’s a lot of tech solutions at work in the talent acquisition space right now, what makes Modern Hire different?
Eric:
Yeah, that’s a good question. And I think the core of that question gets at our approach to research and the scientific method. There’s a lot of us at Modern Hire that are psychologists, that are industrial, organizational psychologists by background. And that means that we went through a lot of school that got at how to conduct research and how to find things out about the world and about human psychology and how humans behave in the workplace. And with that comes a very rigorous stance on how to find out what works and what doesn’t work. You collect data, you create hypotheses, you test those hypotheses, you find out in an objective way whether a solution works or doesn’t work. We do that with everything that we create and build. We build it, we test it, we retest it, we retest it some more and we tweak it and improve it, always. So there’s that scientific method mentality that is behind everything we do.
Eric:
And what that means, in practice, is that we don’t take something like AI and say, “Okay, let’s just smush some AI into this product and call it a day.” Instead we build it out very carefully and do a lot, a lot of testing to see whether it really works to do what we want it to do. And so I think, fundamentally that is the essence of the company and that’s in our DNA to do that. For example, we recently have rolled out a capability that involves using AI to score the spoken word in an interview. So we can take what a person says in an interview and apply AI to that, to create a competency score on multiple different competencies and then use that to help make the hiring process smoother, more objective, faster, fairer, et cetera.
Eric:
But we didn’t just create that capability and immediately roll it out. We created it in the lab and then we tested it out with a ton of different cases of candidate data going through it and seeing what happened. And to find out whether that score actually predicts results on the job and whether it’s fair to all groups of people. That takes some time and it takes a lot of data and working through that to see how it works and whether it works.
Eric:
We’ve also done things in the lab that we don’t deploy. We do that a lot, actually. We test a lot of things out in a lab setting just to see what happens and see if it works and see what we find and see if it’s fair. Sometimes we find something that can predict performance really well, but that isn’t fair or that we’re worried might not be completely fair so that doesn’t get rolled out. That just goes into the research bin and general knowledge of how things work and how predictive capabilities work.
Eric:
And our ability to predict job success changes all the time, because there’s new capabilities and technology coming out. AI, as much as it has revolutionized so many aspects of our world, it’s still developing. There’s new techniques all the time in the AI world coming out that make what we do easier and better and faster.
Eric:
That’s sort of the essence, I think, of what makes Modern Hire different. That was a very long answer, Mark. I’m sorry.
Mark:
That’s okay.
Eric:
That’s what it is. It goes well beyond just designing a product, to really doing a dissertation on the product, collecting a ton of data and testing it out and then continuing to do so in perpetuity in the future too, because things change over time as well.
Mark:
The last time I interviewed you, you said that Modern Hire relies on hardcore, rigorous … I’m sorry. Hardcore, rigorous science and talent acquisition. Can you give me an idea of what you mean by that? I mean, what’s your approach and how does it work?
Eric:
Fundamentally, what we need to sort of do science in our space is data. We need data on candidates, on employees and on how those employees perform on the job. So we need to be able to match up our measurements of what a person is like, and usually those are candidates, sometimes their employees as well. We need to match that up to how they perform on the job, how long they last on the job, how many widgets they sell, how good the customer service is that they deliver, how generally good their solutions are that they’re coming up with and the problems they’re solving, et cetera. And so that’s the basis of everything. It starts with collecting data. And then once we have the data, we can analyze it to see what predicts what, and if a solution predicts success on the job, it predicts job performance. Great. But that’s not enough off because it also has to be fair to all different protected classes. And so we have to test that as well.
Eric:
And then there’s of course, other litmus tests like efficiency and cost effectiveness and candidate engagement, the experience itself. All of those things have to be aligned for a solution to be finalized and sold by us. There’s a lot of considerations, but there’s no silver bullet on predicting what a human will do in a future role. We don’t have a crystal ball. There is no crystal ball and humans are inordinately complicated creatures. You might know that if you’ve met some of us. I mean, we’re confusing and we are inconsistent and we’re amazing and problematic and all of these things kind of rolled into one.
Eric:
And so understanding who a person is and whether they’ll fit a job, whether they’ll fit a role, whether they’ll fit a company, it’s complicated. We are doing, I think, better and better all the time because of modern analytics being as strong as it is and new AI techniques and the big data world that we live in we’re so much data is captured. There’s more and more we can do to make sure that that fit between an individual and a job and a company is ideal. It’s ideal for the company and that they’re getting an employee who performs well. And it’s ideal for the employee, who is getting a role where they can hopefully thrive and enjoy it and stay a long time and grow in their career, et cetera.
Eric:
There’s just tons of opportunity for that match to be optimized out there in the world. It all starts with data. It all starts with some hypotheses about what we think will make a person successful in a role, and then testing those out because we’re wrong a lot of the time in the hypotheses we come up with. I joke, our puny human brains are not really that good at determining what will make a person successful in a role.
Eric:
I mean, everybody thinks that experience is a good predictor of success in a job and that’s why when you go look at a job posting, it says it must have three to five years of experience or five to seven years of experience and that’s very rarely a useful measure. Really, what matters is the quality of experience and what a person’s learned and skills that they’ve acquired and things of that nature, not so much, just the amount of time they’ve done a job. But in our human minds, that’s a very simple metric that we think must be relevant for success. That’s just an example of the type of hypothesis that’s probably not going to be born out when you look at data. The amount of time is not usually related to success on that same type of job.
Eric:
That’s the process. It’s collecting data. It’s analyzing the data. It’s seeing what works. It’s seeing what doesn’t work. It’s making sure it’s fair, et cetera. And so, again, no silver bullets, no crystal balls. Humans are complicated and the more data we get, the better the AI becomes, the better a job we can ultimately do at creating a successful match between that candidate and the organization.
Mark:
Let’s step back for a minute and take a look at the market. We’re in a unique labor market right now. How would you say Modern Higher fits into it? Are there certain types of roles you are one way or another optimized for … Can you deal with any population?
Eric:
Yeah. Yeah. So short answer is, yeah any population. Longer answer is that we deal with a lot of high volume hiring solutions. And so they are most often deployed, I think, for roles that have a lot of applicants and a relatively large headcount. You would look at things like call center representatives and field service technicians and things like that. But you can also go into professional level jobs, certainly, and managerial roles as well. And you can go across every sector of the economy, any industry.
Eric:
At this point, we’ve been doing this for over two decades now, I think. Yeah. Over two decades. And we’ve studied almost every job under the sun. We may not have a solution for CEO of a biotech firm, specifically. That might be a job that we’ve never done a validation study for and studied specifically, because it’s so rare but we have other tools that can be useful at that level to try to help identify that level of person. But below that, when you look at the range of jobs in the economy, we have solutions developed for many of them and we’ve studied a lot of them over the years.
Eric:
And so in general, though, our place in the HR tech marketplace is, we are a firm that collects data on candidates and uses that data to help enable better decisions in the hiring process. We don’t do sourcing, per se. We don’t do, sort of above the funnel tools very frequently, but once a person becomes a candidate, then they go through various gates that collect data on them. Interviews and assessments are chief among what we do. We collect data from interviews and assessments, and then we process that. We use advanced analytics, AI, et cetera, to process that and help enable a better decision.
Eric:
A recruiter might then get a list of candidates that have applied for their job, and they can sort it based on scores that are validated, scientifically determined to relate to, or to predict performance on the job. And so we’re sort of that middle part of the hiring funnel, where you have all this data that we’re collecting and then using it to enable a better decision.
Eric:
Basically, we’re trying to make the hiring process scientific. It can be very difficult as a recruiter or a hiring manager to figure out who to hire. Let’s say, you post a job, you get a hundred applications. What do you do with that? You’ve got to read through every single one. You’ve got to figure out what their experience means and how it may, or may not, relate to the job you are hiring for. Nobody can do that, adequately. I mean, I’m a industrial psychologist, I’ve been one of my entire career, I’ve studied people my whole life, I can’t do that. I don’t know whether the school a person went to is a meaningful predictor of their success on the job that I’m hiring for. I don’t know how conscientious do they need to be? I don’t know all these things about them. I don’t know this in my head. I can’t know that because nobody can. I mean, it’s really the type of thing that can only be understood when you look at aggregate data, when you analyze that data and you see what the trends in the data are.
Eric:
And so what our tools do is they put a number, they put a score on different characteristics that we measure based on volumes and volumes of data that we’ve analyzed. And then they are able to help you zero in on who is likely to actually be a success in the job and who isn’t. When we do that, when we look at the results that these assessments and interviews can help companies to achieve, they’re big. Gosh, we could, literally, go on for days talking about the results that our systems achieve. And that’s because we have a team of dedicated analysts that is continually analyzing the results that come out of our systems.
Eric:
For example, if you deploy a system and you hire people who score higher on the assessment that we deploy let’s say, over time performance in that role of new hires is going to go up and turnover is going to go down and then you can calculate the ROI impact to that. And it’s tremendous. It’s always huge, because people have a huge impact on the bottom line. For a large client, a large retailer, that sort of thing, the ROI of these types of things is in many, many millions of dollars in terms of better performance and lower turnover. That’s the type of gain that can be made for a large organization and for any organization. Hiring better people is always a tremendous value.
Mark:
You know, there’s so many systems out there, [inaudible 00:15:25] Modern Hire, Lord knows how many competitors that you’ve got. Do you think that recruiters and talent acquisition functions are getting the most out of these systems?
Eric:
No. I mean, I would doubt that they are, in general. I think that well-developed, scientifically built systems are common in certain areas of the economy, in certain organizations, but a lot of organizations haven’t fully utilized that type of technology across their footprint yet. I’m a big believer that there’s a ton more that can be done with better data and better analytics. And I think we live in this big data world, we all know this, where tons of data is constantly being created, but that doesn’t mean that we’re learning from that data. That doesn’t mean that companies are optimized to actually look at that data and learn what it means. Most companies are not yet optimized for that.
Eric:
It’s very common for a client of ours to say, “Hey, can you look at this data and tell us what it means?”And we can analyze it and we can crunch it and see what it means for them. But sometimes they’ll say, “Hey, can you take a look at this huge sample of resume data that we had and analyze it for us?” And we’ll say, “Yeah, sure.” And then our HR contact will come back a week later and say, “Actually, I don’t know how to extract that data and send it to you for analysis. There’s seems to be no way for me to do that.” There’s situations where companies are storing a lot of data, but have no legitimate or easy a way to pull it out and analyze it.
Eric:
Just because we live in a big data world doesn’t mean that that data is analyzable. I think there’s, there’s a ton that can still be done. And what I recommend companies do is think about the data that they’re collecting and that they may have access to and work toward mapping out where it all lives and understand what the rules are that govern whether you can access and analyze that data, because sometimes there’s restrictions on it, of various sorts. And sometimes you can’t extract it and put it in a format that’s analyzable.
Eric:
But the thing that’s exciting now is, with AI, like I used that example of resumes, right? Analyzing resumes. Five years ago, we couldn’t analyze resumes very well, but now we can because of AI because of natural language processing and deep learning. And these types of things are brand new. In the scheme of things, these are very recent developments and very recent technologies. And the things we’re talking about now and looking at were not possible just a few years ago. That’s exciting. I think there’s tons of potential for companies to do better with data and to collect data and to enable companies like us, or even themselves, internally, to analyze that information and extract meaning from it. And when you do that, you can better predict who’s going to be a success on the job and you can learn to do it fairly, as well, because, ultimately, AI and any sort of modern analytics can help us to create a more just and fair world by looking at criteria that actually predict performance and that do so fairly.
Mark:
We’re basically at the turn of the year right now moving into 2022. What are the things go on in talent acquisition that you’re keeping a particular eye on?
Eric:
Well, of course, with the pandemic that has changed so many things, so drastically. I think one of the things that seems apparent to me now is that there is no consistency in what to expect. It’s difficult to say, “This is a trend that’s going to continue.” What we know is, we don’t know a lot. And the amount of applicants that companies have may go down, may go up. Companies need to be prepared for different eventualities. One of the things that, I think, is important is for companies to have a solid base in their hiring process that can help them source and screen and hire candidates and onboard candidates in times of candidate abundance and scarcity. And to be able to roll with whatever happens and with AI, big data, et cetera, there’s no reason why companies can’t be positioned for that.
Eric:
The other big thing that I think we’ve seen a lot of is, the huge emphasis that’s on diversity, fairness, equity, inclusion. That feels to me like a lot of that focus has come out of AI, because AI initially led to some problems of bias when it was applied early on, and I’m sure still is being applied in that manner today, by some folks in some organizations around the world. But AI also gives us a great opportunity to equalize things and to do a better job of analyzing data and making sure that things aren’t biased for different groups of people. In some ways I think AI is the problem that led to our current focus on diversity and also the solution to that problem. That’s an exciting thing. And I think there’s just a lot of power and, and potential in what we can do with AI in the future to help equalize things, equalize the workforce and learn more about how to create a more just and fair world.
Eric:
One of the tools that I’m really excited about that we’ve worked on this past year, Mark, is sort of an above the funnel interest type inventory. And we gave that to a big sample of people. And what we found is that when people go through that tool, they’re recommended jobs in a very gender blind kind of way. And so if we look at a role like field service technician, that role tends to be overwhelmingly male, but in the recommendations that our interest tool provided the candidates, it was recommending that at a very high rate for female applicants. Then that drives female candidates to that field service technician role, because objectively they’re good fits for it. They can be good fits for it, just as good as men.
Eric:
And so all of a sudden you’re able to start equalizing the applicants for that position, which can then help to ultimately equalize the distribution of men and women in that role. There’s no reason women can’t do that role just as well as men. It doesn’t require some kind of outlandish strength or anything like that. It doesn’t need that. As many women might not think of themselves as a fit for that role, as actually are fit for that role, so we can help them to understand that and to drive them into those historically unequal types of, of roles. That’s an exciting thing, and that comes out of data. That comes out of data and being able to analyze it at scale and understand how and where people fit in the world.
Mark:
My last question, we’ve been talking about the market, we’ve been talking about the technology, we’ve been talking about human behavior essentially. With all that said, where does Modern Hire go from here?
Eric:
Well, certainly we have a lot of great existing tools; interviews, assessments, scheduling capabilities, workflow enablement features designed to make the hiring process smoother and more efficient, fair, effective, et cetera. And we continue to iterate on those tools and introduce new features and applications for those tools. One of the big ones this past year was automatic interview scoring, which takes what a person says and applies a score to it. A validated, predictive, fair score, which is many times fairer than what human humans would do if they’re rating an interview response. We have data that we’ve tested, and we found that in, which is very exciting. We’ll continue to expand and roll that type of tool out.
Eric:
Will it continue to use modern analytics to help make the hiring process fairer and more effective? And I think, `beyond that, I look in my role in labs and in innovation to sort of expand the data that we’re able to make sense of. Expanding from just traditional assessments to more unstructured information, written responses, typed responses, spoken responses, et cetera, or other types of data. Could be resumes, could be social profiles. I mean, those are things that we might study in a lab, but that we probably wouldn’t use in the actual hiring process, but there’s a lot to be understood out there. And we view it as our role to help clients understand what all that data means and how to most effectively use it in the hiring process.
Eric:
We’re in a golden age, I think, of research and understanding of what makes people successful on the job because of AI, because of deep learning and natural language processing. For decades and decades, the field of industrial psychology was basically, “Hey, here’s a test and we’ll give you a score for extroversion and conscientiousness and some other things,” and that’s about it. Now, we can score almost anything using modern analytics. And that’s awesome, because in being able to do that, it really opens the door to making the hiring process vastly better for candidates, for organizations, for recruiters, et cetera.
Mark:
Eric, thanks for coming back today.
Eric:
You bet. Always great to talk with you, Mark.
Mark:
My guest today has been Eric Sydell, executive vice president of innovation at Modern Hire. And this has been People Tech, the podcast of the HCM Technology Report. We’re a publication of Recruiting Daily. We’re also a part of Evergreen Podcasts. To see all of their programs visit www.evergreenpodcast.com. And to keep up with HR technology, visit the HCM Technology Report every day. We’re the most trusted source of news and the HR tech industry. Find us at www.hcmtechnologyreport.com. I’m Mark Feffer.
Image: iStock
