HRx Radio – Executive Conversations
Guest: Christy Whitehead, Chief Talent Economist, ENGAGE Talent
Air Date: September 6, 2019
Christy Whitehead is the Chief Talent Economist at ENGAGE Talent. She leads the development of the science behind ENGAGE’s predictive algorithms and machine learning models as well as the product as a whole. Christy holds a PhD in Economics from Clemson University and has focused her career on employment and labor economics.
Her previous experiences include key data science roles with PeopleMatter (acquired by Snag A Job) and Equifax. Christy is a frequent speaker in HR Analytics and machine learning conferences. She is also an active member of the Federal Reserve Roundtable
Website | Twitter | Facebook | Linkedin
Audio MP3AudioPlayer.embed(“f-html5audio-0”, soundFile: “https://www.hrexaminer.com/wp-content/uploads/2019/09/2019-09-06-HRExaminer-Radio-Executive-Conversations-Episode-338-Christy-Whitehead-Chief-Talent-Economist-ENGAGE-Talent-Final.mp3”);Audio MP3AudioPlayer.embed(“f-html5audio-0”, soundFile: “https://www.hrexaminer.com/wp-content/uploads/2019/09/2019-09-06-HRExaminer-Radio-Executive-Conversations-Episode-338-Christy-Whitehead-Chief-Talent-Economist-ENGAGE-Talent-Final.mp3”);if (jQuery.browser.mozilla) tempaud=document.getElementsByTagName(“audio”); jQuery(tempaud).remove(); jQuery(“div.audio_wrap div”).show() else jQuery(“div.audio_wrap div *”).remove();
Subscribe to Podcast or Download
Subscribe in iTunes
Subscribe via RSS
Download MP3 File
Join John Sumser at this year’s HRTech conference
Important: Our transcripts at HRExaminer are AI-powered (and fairly accurate) but there are still instances where the robots get confused (or extremely confused) and make errors. Please expect some inaccuracies as you read through the text of this conversation and let us know if you find something wrong and we’ll get it fixed right away. Thank you for your understanding.
Full Transcript with timecode
Today’s Show is brought to you by the Human Resource Executive Magazine’s HR Technology Conference and Exposition, held October 1st through 4th at the Venetian in Las Vegas.
Join me and thousands of your colleagues at the world’s largest exhibition of HR Technology. Act now using the code HREX and you can receive a $300 dollar discount on your ticket.
Thanks. We’ll see you there. And by the way, don’t miss the Women In Technology segment.
Good morning, and welcome to HRExaminer’s Executive Conversations. I’m your host John Sumser and today we’re going to be talking with Christy Whitehead who is the Chief Talent Economist at Engage Talent.[00:01:00] Now that’s a lot of mumbo-jumbo for Christy is the most interesting data scientist in all of HR technology and so she’s got a PhD from Clemson and spends her time thinking about employment and labor economics and it’s going to be a great conversation. Hi Christy.
Christy Whitehead: [00:01:19] Hey John, thanks for having me today.
John Sumser: [00:01:21] Yeah. I’m really looking forward to this conversation. Take some time and introduce yourself if you would.
Christy Whitehead: [00:01:27] Sure, yeah, so, you know as you mentioned I’m the Chief Talent Economist for Engage Talent and what does that mean?
[00:01:33] So I lead up our product our research and our data science team. I’ve been working with Engage for almost four years now. And I’ve been in the HR Tech space for almost 10 years. As you mentioned I have my PhD and I absolutely love to learn I would have stayed in school forever if I could but I found a great balance with Engage continuing to do research and learn and also kind of find a way to contribute value.
John Sumser: [00:01:58] That’s amazing. Tell me a little bit more about the day-to-day. How does have a talent economist work?
Christy Whitehead: [00:02:06] Yeah, that’s great. So, we spend a lot of time reading and finding new data sources and pulling in that data to really exploring everything but other people are talking about making sure that we’re staying kind of on that Cutting Edge of the way.
[00:02:20] People are thinking about Talent the flow of talent through understanding management science understanding the way people make decisions. And you know using everything that we have to kind of explore research and then build models to really leverage that information and they useful way.
John Sumser: [00:02:40] So you’ve built an extraordinary model at engage down with them and I had somebody some years back told me what it was like to watch you when you were working artists and they said when she is really doing her best work.[00:02:53] She’s out in a chair in the courtyard with her eyes [00:03:00] closed getting
Christy Whitehead: [00:03:00] some sun. I have a very active in our mind. So I really love to think about problems and really explore various Solutions. And so it definitely one of my favorite pastimes. It’s
John Sumser: [00:03:17] really interesting because the engaged count model. At least I’m not exaggerating when I say there are 40,000.
[00:03:24] And is that right?
Christy Whitehead: [00:03:26] Yeah, so we are, you know directly or indirectly with collecting data from over 40,000 different sources and that equates billions of data points. So, yes, we are collecting and aggregating a large amount of data about people about organizations about jobs and the economy as a whole to really understand the full context of the talent.
John Sumser: [00:03:46] So we didn’t really cover this yet. But why don’t you talk a little bit about what engaged telling actually does that’s it’s a pretty interesting
Christy Whitehead: [00:03:56] idea. Yeah. Yeah, so, you know gate Talent is really kind of a [00:04:00] total pal intelligence platform. So we like have mentioned kind of we’re working to understand that flow of talent in and out of work on phaeacians kind of what creates a misalignment between a person and their job and the company that they’re working for.[00:04:14] To really understand why people change jobs how recruiters can better engage that counts are interested in and really understand the market as a whole other competitors and really help to make better decisions about their talent strategy. Our core foundational for is what we call our engaged for and that’s something that we built on years of academic research. [00:04:37] And we kind of identified a lot of different factors that play into the reasons why people paying job and we still have been able to build a really robust model that really helps to identify which people are most likely the same jobs at any time.
John Sumser: [00:04:55] That’s pretty interesting stuff. So where does it go from there?[00:04:58] You’ve got these [00:05:00] 40,000 inputs billions of data points. And you’re trying to make predictions about about individual choice in staying or leaving a job. What’s the next Horizon?
Christy Whitehead: [00:05:12] Yeah, so I think really being able to understand how the workforce is changing right? It’s not static. He was an open sign in a dynamic world is always constantly evolving and changing the thinking about how people were are people moving to more of a condition.[00:05:30] I like kind of mobile Workforce. How does that impact people’s decisions. You know, what does the labor force look like in the future? And how does that impact the way that we find talent and recruit talent and maintain that strategy that’s going to keep the business productive. And so that’s
John Sumser: [00:05:49] the core right.
[00:05:51] Now. The value proposition is that you can predict when somebody is likely to be receptive. Good job offer. Does [00:06:00] it work?
Christy Whitehead: [00:06:01] Yeah, so we definitely find that people if our customers Target those people that we think are most likely to change jobs, but they kind of get that 2 x is 7x Improvement in response rate from those candidate and that really can help them.
[00:06:15] Not only understand the people because we do this at the People level, but we also do it, you know at the company level of the industry level and really at the macroeconomic level understand not only who they should Target but also how their competitors are doing, which is our. Target for talent, how was the industry as a whole how do they compare to the industry and really get a full picture and really understand that whole context and not just what’s going on with an individual person fascinating.
John Sumser: [00:06:41] So I’m kind of Imagine do it five or six years old in a sandbox somewhere. Deciding that you wanted to become a scientist.
John Sumser: [00:06:54] I don’t think it might have come from Counting grains of sand, but I can’t imagine that that’s [00:07:00] actually the case. So tell me about how you evolved from wherever you started to is somewhat bizarre fascination with complicated puzzle.
Christy Whitehead: [00:07:11] Yeah, absolutely. So, you know, it’s a great question and you know, the job title data scientist was not around when I was 5 years old. So it’s for the last 10 years or so. So I definitely didn’t grow up dreaming of being a data scientist, but I always did love math and I always loved art and being creative.[00:07:35] And those are two things that you know, I can I continue to see in my life and I can remember coming out of undergraduate trying to think about how can I use data to help businesses and people can make decisions and create value through the data and I can remember having these conversations with my dad and him helping me brainstorm different areas of operational research and this and that and the other [00:08:00] and I wonder how does that I’ve really fallen in love with economic. [00:08:04] Really? What economics is about it gives you all these underlying theories and Frameworks for thinking about the way that people and businesses make decisions and kind of gives you those tools to analyze those decisions and build these models. So I guess it kind of just developed out of the fact that you know, I love math and I love being creative and I feel like the data science. [00:08:25] This mole is really about research and. Creative problem-solving as much as it is about math and numbers, you know, there is a very technical side of it that it that is very technical but there’s also a creative piece where you have to think about, you know, how to measure different things that might not be easily measured and how do you model those in a mathematical way? [00:08:48] So in a lot of ways is very creative process in addition to being very technical.
John Sumser: [00:08:53] Amazing show seems like you like solving puzzles tell me about the kinds of puzzles do like [00:09:00] the salt and the ones that currently have your attention.
Christy Whitehead: [00:09:04] Yeah, so, you know, I think I really I love all kinds of puzzles everything from literal puzzles Sudoku puzzles to thinking about wife as a puzzle.[00:09:15] And and how do you kind of you know, Make those things happen if you want to happen. So to me almost everything is puzzle. And I really like challenging and difficult problems. I kind of have a love-hate relationship with those really hard problems because it’s usually at that brink of frustration where you’re just kind of beating your head against the wall trying to figure out how to solve it that you get this light bulb and you’re like, oh wait, you know, this is what I need to do. [00:09:40] This is how I need to solve it and there’s this huge gratification and satisfaction when you’re actually successful any kind of huh? That precipice after looking at a problem for a long time and exploring our options. There’s a huge satisfaction in being successful when you’re solving hard problems.
John Sumser: [00:09:57] Well tell you what, that’s what I do for a living [00:10:00] in a very different way of stuck this check which about medical and Technical but I do like getting into something that I don’t understand that finding my way out of it. What has your attention now? What are the things that you’re curious about?
Christy Whitehead: [00:10:14] This can sound a little bit kind of idealistic probably but I like to think about what if we all lived in a world where we love to work with everybody loves to go to work. What would that look like? How awesome would that be? You know, we all have to find a way to make a living and provide for ourselves and our families.[00:10:31] Our work is a really big part of our Lives whether it’s inside the home or outside the phone and if we love what I what we do if you have a huge impact on our lives. But can also have a huge impact on the businesses that we work for. We know that people that love their work are more productive and more effective in their job. [00:10:49] And so not only does it help. As individuals and businesses and society as a whole because we make it bigger strides in Innovation and change and can [00:11:00] really help improve the way Society works together. So like I said, this is very ideal with the world. But the problem that I like to think about when I’m trying to kind of think about that highest level problem is how can we best connect the people in the organization’s with the job in a way that they’re going to be happy and excited about its kind of a lofty goal no doubt, but I figure if you can try to. [00:11:20] All the world’s information I can strive to help connect people with what they love. And so what we’re trying to do is help identify those people that may not be enamored right with a job that they’re doing right now and might be up for considering a new role and helping them by hopefully connecting them with something that might be new and interesting and challenging for them.
[00:11:42] Okay. So let’s drift back
John Sumser: [00:11:44] a little bit to engage Talent you’ve got this this Mom. 40,000 flows of data that are input and you predict the likelihood that [00:12:00] person actually will be interested in moving on to the next job help me understand and kind of a simple way what that model looks like because that’s just organizing the pile of data strikes me is worse than having to organize a closet.[00:12:18] The question is how do you organize and give me some sense of how you make sense out of all of that? And then how do you remember what’s
Christy Whitehead: [00:12:29] there? Yeah, great question. So, you know, I think a lot of people don’t think that data science is very glamorous. Definitely heard of it reference is one of the sexiest jobs of the century, but it is really a lot of grunt work. [00:12:41] So 80% of what data scientists do is all about organizing understanding cleaning and dissecting the data before you can figure out how to leverage it and build a model on it and in fact because there’s so much what we do is gathering data from disparate data sources. A [00:13:00] lot of the work that we do is actually entity resolution. [00:13:04] Which is taking data point and mapping them together and understanding that they’re referring to the same entity whether that’s a person or company and mapping people to companies and understanding that this person that is a data scientist is a data scientist with engaged. And I’m kind of understanding what the landscape of and a gage Talent looks like and so for pulling together all of the kind of disparate data and you know, one of the great tools that we Leverage is kind of academic research and there’s really provocative theory that was built in the 90s that we’ve been able to kind of validate with some of our research and has been well validated by the academic Community as well that really helps us break down the problem of when and why people change jobs in a way. [00:13:49] You can have a lower level of unfolding model of voluntary turnover.
And it relies on the fact that people don’t just become unsatisfied in their job and quit
[00:14:00] actually people will stay in a job unhappy for a long time. This mall was send that over 50% of voluntary turnover is related to a soccer an event that happens if something happens that triggers that decision.
John Sumser: [00:15:09] Interesting. So one of the things I’ve noticed over the years, is that how an organization defines what were actually is is highly variable thing and so in some organizations are consulting firms from the mine when you walk into the office.[00:15:29] What you see are a lot of meeting rooms because in those organizations partly because their billable by the our meetings are actually sort of the work. And people who work in those places are really good going to meetings and then you go over to a deeply technical company and you see almost no meeting rooms because in those environments meetings are the opposite of work. [00:15:55] So I think that what work is what [00:16:00] employees do inside of an organization and how they earn the currency that the company gives them for the time they spend there is. Somewhat unique to the organization. And so when I think about why people move and where they go, it seems to me that a lot of the time the issue is I think the work one way in my organization think of work another way. [00:16:23] And so there’s this apparent friction in the relationship. Do you have ways of seeing that kind of thing in your date?
Christy Whitehead: [00:16:30] Yeah, and I think you know, there’s a lot of interesting things that you can learn about, you know and organizations based on the type of people that they can hire right? What type of how long did they say it?[00:16:46] A given person, you know what companies they work for in the past. How long did they say at those companies in those jobs the types of jobs that they have what type of skills do they have? All of those things can give you a little bit of insight about how an [00:17:00] organization work and the types of people that are likely to stay at that company versus the type of people that are likely to leave that organization Jim of the things
John Sumser: [00:17:09] that I’ve been to the show boxing the last year show is the idea that recruiting.
[00:17:16] Has a fifty percent failure rate and that then that that is a dismal embarrassing truth about how recruiting work. Yeah, when you look at your data, do you see that in the data?
Christy Whitehead: [00:17:30] You’re not gonna question. I don’t know that we’ve really framed the question that way what I can say is that, you know, one of the things that we look at is how long people tend to stay in various organizations Etc.[00:17:43] And there’s definitely an overwhelming friend that over the between. The first I would say six months is this is particularly for lower-level jobs, you get into higher level jobs more like non-management lower-level jobs what you’ll see is that between about [00:18:00] three to six months and two years. You see kind of like that peak in the likelihood that they’re going to change jobs to the people that you know, as far as how long people on average stay in a job that first two years. [00:18:12] One of the most pivotal and after that you’ll see that if they say they tend to stay for quite a while. It’s almost to me and I do this all the time is that compared job-hunting to dating so the fifty percent success rate with a job in the 50% divorce rate. Maybe there’s some parallel there, but. I think there’s a lot of unknowns right when you’re going into an organization about what they tell you in the interview. [00:18:39] And then what it’s like when you actually get there and vice versa people will tell you what you want to hear in an interview and they come into work and then behavior is a little bit different. And so I do think that some extent the first couple of years of really make or break when you’re kind of evaluating is this a long-term relationship or this away? [00:18:58] So I definitely see those [00:19:00] Trends and it is an interesting problem to try to solve.
John Sumser: [00:19:04] So we’re going to ask you about I want to start by as I listen to you talk started to imagine that you probably can already. Make some gross estimate of a person’s likelihood of success in the job based on their background, right? And so so the first ethics question is if you can make an estimate of the potential won’t be a hundred percent guarantee. It’ll be some probabilistic formulation of the likelihood that they will succeed. How do
Christy Whitehead: [00:19:37] you like
John Sumser: [00:19:38] how do you deliver that kind of information without tarnishing people with a label?
Christy Whitehead: [00:19:43] All right. Yeah, absolutely. You know, we do try to measure on a sale. Right? So we definitely want to recreate these are out from the vehicle. And this is one thing that I think about all the time is the dated me a very useful tool the data is [00:20:00] best to leverage when it’s paired with our own judgment.[00:20:02] So we kind of have a red yellow green scale that we use kind of portray How likely we think somebody is. But we we always say the person is most likely the same job. We would never say this person is going to change we can’t do that. There’s too many factors that come into play and those decisions, right? [00:20:22] So we need to think for ourselves and taking this information and say okay engage those of the person is most likely the same job. What else do I know about them? I use my own judgment and kind of balance that between. Thank you for help and using the information that we have. There’s a famous quote from George box. [00:20:41] But I really love he’s a renowned statistician. This is all models are wrong. But some are useful models are just that their model. There aren’t perfect representation of the world and I’m never going to be a hundred percent right all of the time but given that well built models can and should be used a kind of Osman. That information that we have and you know for ourselves
John Sumser: [00:21:03] so that puts a lot of pain in the end user of the output of the model is that face Justified? I mean, it seems to be that if I get something that says 80 percent likelihood I go. Okay. I’m thinking
Christy Whitehead: [00:21:18] that instead of
John Sumser: [00:21:20] I think what you suggest is that they ought to be more thoughtful than that.
Christy Whitehead: [00:21:25] Sure. Yeah. No, I think I definitely understand where you’re coming from and I think that a lot of people probably do rely a bit to my aunt. I have those predictions. Yeah, I think for the most part we are kind of find out guide you and say, you know, we we do think that our recruiters and the people that leverage the information that we provide they have confidence in themselves.[00:21:46] And for the most part they will absolutely take what we say and Robert with what they know is well. But yeah, you can become overly overly reliant on on these models and that is arrest
John Sumser: [00:21:59] charges. We could spend a long time on that issue. Are there other ethical issues that that grab your attention in your work?
Christy Whitehead: [00:22:06] Yeah. So one of the things that we are thinking about on a regular basis is bias in the workplace and bias in hiring. And how can we help to mitigate that by us and reduce any bias that might be in our all the algorithms and help our users are going to track and understand the bias if there is any monitor that and reduce it by tracking the types of people that are looking at in our tools and also helping them to kind of monitor over time with that looks like oh that’s
John Sumser: [00:22:37] interesting. So you gives them feedback about the kinds of people that were looking at.
Christy Whitehead: [00:22:45] We have
John Sumser: [00:22:47] about anybody doing that before sure so you can you can sort of measure the implicit bias in the sourcing function itself.
Christy Whitehead: [00:22:59] We [00:23:00] also reporting that our users can use understand forgiving role for a given job function for a given company with the break down there look like and leveraging our Floyd kind of compare that to the broader.
John Sumser: [00:23:15] This is a light bulb for me. This is this is this is an actually great thing that you just disclosure have to talk about that after the show, but we’re running out of time. And so I want to be sure that you get to tell the people who are listening to this with you them to know so what they take away
Christy Whitehead: [00:23:35] forever and that, you know, there’s a lot of data and information out there in the world that can help us to make better decisions.
[00:23:41] And I think it’s valuable for all of us to use that information in a responsible way and help us to you know, be more efficient effective and yeah. Use date and use it responsibly. That would be my biggest takeaway.
John Sumser: [00:23:57] Okay, and reintroduce yourself please. Tell people how to bombard you with email.
Christy Whitehead:: Yeah, so I’m Christy Whitehead, I’m the Chief Talent Economist at Engage Talent and you can definitely reach out to me. Christy dot Whitehead, that’s C-H-R-I-S-T-Y Whitehead at EngageTalent.com?
John Sumser: [00:24:19] Thanks Christy I really appreciate you taking the time to do this. It was a great conversation and I enjoyed myself immensely.
And thanks everybody. Okay talk to you again soon. We’ll do this again sooner rather than later Christy.
You’ve been listening to HRExaminer’s Executive Conversations, and we’ve been talking with Christy Whitehead who is the Chief Talent Economist at ENGAGE Talent.[00:24:43] Thanks for joining in.
See you next time [00:25:00], bye.
Join John Sumser at this year’s HRTech conference