Cole Napper - People Analytics in the World of AI

[00:00:03] Welcome to the HR Data Labs Podcast, now part of the WRKdefined Podcast Network. Join us as we explore the vital role of compensation, strategy, data, and people analytics in navigating today's complex business world. With the resources of WorkDefined, we're now bringing you deeper insights and actionable ideas from top experts. Now, here is your host, David Turetsky. Hello and welcome to the HR Data Labs Podcast. I am your host, David Turetsky.

[00:00:30] And I have with me someone who I've been dying to get on this podcast for a really long time. He's brilliant and you will see that from the moment he opens his mouth, from HR Bench, Cole Napper. Cole, how are you? Thanks for setting the expectations high. Always. The expectations are always high. Well, thanks for having me, David. This is going to be a lot of fun. This is, this is. And we're going to totally geek out on people analytics.

[00:00:57] And for those people who follow the podcast forever, our roots in people analytics. So we're going to totally geek. Before we do, tell people a little bit about you and about HR Bench. Sure. So I have a background with a PhD in industrial organizational psychology, been in the people analytics space since before that was a term. So over 15 years, I published a book last year called people analytics, talking about how AI is sort of disrupting the profession right now.

[00:01:27] I have something called the data driven HR Academy, have my own podcast and newsletter called directionally correct. Right. And recently joined HR Bench as the chief people intelligence officer. So a lot of stuff going on in the space right now, for sure. And Cole, if it doesn't get intrusive enough. Now we ask what's one fun thing that no one knows about Cole Napper? Oh, goodness. I don't talk about it.

[00:01:52] I mean, not that no one knows, because obviously my siblings know, but I'm a middle child and I suffer from a hard case of middle child itis. So it's pretty uncommon thing that people don't know. But obviously my siblings know I'm a middle child for sure. But but it's painful because you got the itis in there. Yeah, exactly. See, I'm a middle child. I don't really feel middle child itis. Well, David, you're a more well-adjusted human than me. I won't complain about it. No, it's the medication that does that. Cool. Yeah.

[00:02:19] But today we have a very special topic because, as I said, this goes back to the roots of where HR data labs came from. And it's people analytics in the world of A.I. And we're going to get into it in a second.

[00:02:54] The use case podcast is where technology vendors get to talk about themselves. And it's a wonderful place for vendors, investors and practitioners to listen to the story of the solution, the features, benefits, attributes, etc. And we get to know the CEO or founder during the during the call. And we also get to know the tech. So subscribe to the use case podcast.

[00:03:24] So, Cole, let's get back into it. And you've mentioned you've been in it. You've been in people analytics now for 15 years, which is tremendous. And, gosh, I started my people analytics journey at ADP back in, oh, my goodness, 2014, which is what, 12 years ago? Yeah. Maybe a little bit before then. But really, 12 years ago is when, you know, kind of I came into people analytics.

[00:03:55] How far have we come in the world of people analytics? Because for a while there, you had to convince people that it was the thing. Yeah, there was a definitely a period of time where no one understood or. No one's I don't want to say didn't care, but because they did care, because I think a lot of leaders were starting to get kind of pushed and their feet held to the fire that they needed to start making, quote unquote, more data driven decisions. Right.

[00:04:23] But they did not understand how to get there. And so like, all right, we'll tolerate having these nerds around. These HR nerds. Yeah, these HR nerds. They're just trying to, you know, get us to use data. We don't want to. We're going to. I feel like over time, though, the conversation really changed around COVID. And then it really changed with the introduction of AI.

[00:04:48] And so, you know, I think the field was still kind of reeling from the newfound credibility that came post pandemic and then AI slapped it across the face. And so we're kind of going through that adjustment right now, for sure. But we're going to get to that. That's the next question. I don't want to go there yet. Okay, not going there. Not, not yet. But, but there have been a few things that have happened in the world to people analytics that have kind of lent some credibility.

[00:05:17] Like I know my friend Richard Rosenow has kind of put his, you know, arms around people analytics and kind of dragged it kicking and screaming through the streets to get it some notoriety. And I'm sure you have too. And also Martha Kirione and some of the others that, and obviously Mr. Green, of course, as well.

[00:05:40] But there are those luminaries who've been talking about it really, you know, and I consider you to be one of those to kind of make it be another real pillar of HR. I mean, if not of HR, of operational excellence, at least. Yeah, it's interesting. And I appreciate that. And I want to say like all three people you mentioned, you know, David, Richard, Martha, all I consider to be good friends. I consider them to be luminaries in their own right. I'll actually make fun of Richard here for a second.

[00:06:09] You always joke that he, Richard's the only person I know that would go on vacation literally to meet with other people that do people analytics. And that's how we met back in the day. Oh, really? I'm like, I care about people analytics. I don't care that much. Yes. Well, it's his life. I mean, it has been his life. And in fact, at one point, I think I called him the grandfather of people analytics. Yeah. I mean, I think that the team at Google may have invented the term, but I think he probably contributed in a large part to popularizing it for sure. Yeah. Yeah.

[00:06:38] I mean, he's got a lot of thoughts. He used to be such a prolific writer. I was a big fan of both him and David in the past. And frankly, I've modeled a lot of the things that I've done off of, you know, learning from them over the years. But in the space, you know, I feel like I've kind of found my own voice to a certain extent. And in the thing that for me, I mean, I don't know if you you've experienced this yourself, David, but when you kind of do things largely in a digital environment, you don't realize how much of an impact you have on the real world.

[00:07:07] Because I keep, you know, just sitting at my desk, talking to my computer screen, writing articles. But the world changes around me and they start listening at a certain point in time. And that's sort of bizarre in a way. But also it's nice to see that things I used to talk about and I felt like I was an old man screaming at, you know, the kids on the lawn. To now it's like a lot of these are pretty commonly held beliefs. It's like, OK, this is making an impact. And that is exciting to see.

[00:07:37] Well, especially when other people quote you. Right. When you see whether it's your words or at least not just your words, but the things that you you've been advocating for. And when other people start advocating for it, well, as well, you're kind of you kind of I don't want to say pat yourself on the back. That sounds awful. It sounds very narcissistic. But, you know, you say, wow, that's yay. Well, I mean, it's like I learned.

[00:08:06] And again, I put like Richard and David and Martha in this camp. But there were folks like, you know, Keith McNulty and Amit Mohindra that used to just like give so much to the community in terms of just like free resources, training, like things that you could learn from. And I always just took that very much to heart and have tried to kind of carry that baton forward to try to kind of pay it forward.

[00:08:30] So to me, it's gratifying because it's like the people who I learned from, I get to kind of hopefully experience what I think they experienced of paying it forward to other people that are coming behind me. Well, and hopefully all that has kind of wound around the people that are now carrying the torches because there is a new generation. And I actually do consider you to still be in the new generation, but you're also in the old one as well.

[00:08:56] But, you know, the people who are carrying the torch for people analytics now, I mean, it's still David Green. You could still read. He's still there. He's still there. And he's still doing a great job of making compendiums on all this stuff. But also there's Ben Weber. I mean, he also does a lot of really great stuff from MIT. And there's a bunch of people who are still carrying the torch, even when I have to focus on other things. So I'm not focused as much on people analytics anymore. I have to focus on everything, but they are still carrying it.

[00:09:25] And I really appreciate that. Yeah, absolutely. I feel the same way. You know, the world keeps spinning like HR. It will change its focus areas over time. So sometimes things will go to kind of go out of the limelight and then back into the limelight and so on and so forth. And but it's really great that there's sort of the journeyman's mentality out there of people who are still trucking along, even if it's not in the limelight. Right. Well, they have to. Right. Yeah.

[00:09:48] And one of the crazy things about where we are today, before we get into the next question, is that a lot of companies are doing a lot of layoffs. And when I say this is crazy, it's crazy because some of these companies have created outstanding teams, especially around people analytics. God forbid these teams are deemed unnecessary and I will replace them. God forbid. I mean, really.

[00:10:16] It just it rankles me that we get companies who invest a lot of time, a lot of energy. And then they say, well, this looks like it's going to be something that gets replaced soon. Yeah. Can I talk about that for a second? Because I have a lot of thoughts. One of the my greatest strengths, but also the things I struggle with the most is I'm usually early. And what that means is a lot of times I'm recognizing things that other people aren't seeing and then I'm reacting to them and people are going, what are you talking about? Yeah, exactly.

[00:10:45] And so in 2022, I gave this big keynote presentation at a few conferences called Is People Analytics a Luxury? And kind of like the subtitle that we cannot afford. And it was it was like a precursor because I saw this thing with the end of the Zerp era. Interest rates were increasing and teams are starting to have to justify their existence. And this was like right before all the layoffs started to occur. And and I was talking to folks and everybody's like, no, everything's great.

[00:11:15] We're coming off the pandemic. Everybody thinks we're super seriously. I'm like, you guys are not you're not recognizing this is a problem. And and then just I went through my phone at one point in like 2024 and I counted. I'm not like one of these people has thousands of people on their phone. I kind of had 48 people that were in my friend group that had been laid off in that period between 2022 and 2024. It was it was a culling. Yeah. It's depressing. It was it was horrendous. The good things I'll say.

[00:11:45] And so in 2023, I'd written an article something about like how people analytics leaders were coming disenchanted. No one liked it. And so I just kind of seeing that there was this malaise that came over the field. I say all that to say I'm actually talking to people right now. And a lot of people are still kind of reeling like, oh, the layoffs, things are so hard. We're being challenged to do with more with less. I'm talking to people right now. They're excited. I really think that we're about to see another boom. I do.

[00:12:12] And I'm crazy about that. But I actually see like some of the early signals that this is.

[00:12:48] And I'll tell you why. We're kind of creating the future that we're going to live in. And so I don't know if that's a hot take, but that is something that I'm seeing right now that I don't hear a lot of other people talking about. So what that says is because those people analytics professionals have their pulse on the data.

[00:13:08] They have their pulse on how to utilize ML and AI and some other pieces that the companies are leaning on them to kind of steward. The HR department into the HR department into the AI realm. It is that.

[00:13:27] And there's one other component, which is, I think, almost more existential, which is like it requires an experimental mindset and almost like the ability to like try new things and be somewhat of a risk taker. So if you were a person who joined people analytics in 2015, you were a risk taker because there was nothing. Yeah. It was like you had to be willing to kind of write your own script because there was no playbook to follow.

[00:13:56] And that type of person is a unique person. And they're actually pretty well equipped for transformational times like the moment that we're in. And so if you're looking around the HR function, HR has typically been considered a risk mitigation function. Right. And risk mitigators aren't really good at dealing with transformative change.

[00:14:16] And so it's actually kind of that mindset plus the skills with data, plus the skills with, you know, machine learning and AI and LLMs and the kind of that product based mindset that many people have as well. Because a lot of what is being built is needing to be productized at the moment. I think all of those things are essential for what the transformation that HR functions are trying to undergo. And so I actually think that makes people analytics pretty invaluable at the moment.

[00:14:45] Until now, let's talk about that. So the second question we just started to talk about was AI is now going to change people analytics forever. Yeah. How and why? And I think we've started talking about it, that it's leading in certain areas, but it also it needs help, right? Yeah. Well, I'll tell you this. Like I've been saying lately, people are becoming a commodity, like the core parts of it. And I think that is such a good thing.

[00:15:13] I know that puts me out of step with a lot of folks. Like if you're a person out there and all you do is build Power BI dashboards, you hate when I say things like this. But the reality is like, man, we don't ever need to build another dashboard, another report. Like the technology exists. It has existed for some time now. Like it just doesn't need to be done. And so the people that are out there that have made a whole career out of doing that, you know, more apologies to you.

[00:15:39] But probably the skills that took you for you to learn Power BI are the same skills that are going to take you into this new AI generation. And so you're actually pretty adjacent from like a skill standpoint compared to many other people. Like you're going to be able to make the journey, right? And it's just like I am so excited because like I remember sitting around with my team like 2017, 2018 saying, I never want to build another dashboard again. Can we automate all this stuff?

[00:16:05] Like there is no incremental value that is created for an organization because all of the same data every organization needs. Like there's nothing new under the sun, right? And so if you know there's nothing new under the sun, why don't we just commoditize the building of it, which has happened. And then let's focus on, again, they always say like doing the strategic work. But there's so much strategic work to be done. Like let's focus there. And so I really think people analytics becoming a commodity is a good thing.

[00:16:32] And that's why I've been talking more lately about this concept of people intelligence as being kind of the successor to people analytics, which is what is all that value added work? Well, guess what? It fits within people intelligence. Well, before we get to people intelligence, though, let's talk a little bit more about how people analytics has grown with AI. Like, for example, you're mentioning a lot of you never have to build another dashboard again. Well, why?

[00:16:58] There still has to be a place where you get notified as to stuff's going off the rails, dude. And here's what's going off the rails. So there's got to be some kind of interaction, whether it's a chat, whether it's notifications, whether it's something. People have to be notified as to what's going on if there's no dashboard for them to go like a destination. Where are they seeing these things or how are they consuming them? Yeah, I think, well, the cynical take, first of all, is they were never consuming them.

[00:17:27] Because a lot of times, like if you look at the usage of most dashboards, which any team that I've led, I've always created kind of like usage metrics. The most common number of logins is zero or one, right? Yes. And so, and yeah, you can create kind of these push notifications and even like Slack integrations and things like that that will send people based on like algorithms that are showing like this is outside of a certain range.

[00:17:55] Therefore, you should, you know, kind of contextualize the results. And some teams are doing that and some technologies are doing that. But I'd still say it's more rare than it is common. But the reality is there has to be a desire to do something with the data. Otherwise, the push never turns into a pull. Because what I've always, I mean, I kind of was brought up on Lean Six Sigma. And, you know, you're trying to create pull-based methodologies based on the Toyota production system. I used to work at Toyota for a little bit.

[00:18:25] And the goal is to get to pull. And the old dictum I used to say is no push can ever turn you into a pull. And so what that means is there needs to be demand. And you like no amount of automation can create demand. What creates demand is business need and curiosity, right? Business need is good. Curiosity is good and also bad sometimes because curiosity also creates waste. That's something you always have to measure against.

[00:18:53] But I remember this is a long time ago. And I was interviewing for a role with a CHRO at the time. And they told me to describe like a typical day as a people analytics person. And I started talking about like all the HR business partners and business people I would be speaking with. And, you know, trying to understand the problems that they're facing. And then bringing the data to help solve those problems and make the key decisions in a just-in-time manner for them. And they said, and they said, Cole, you don't sound like an analytics guy.

[00:19:21] You know, you sound like, you know, maybe an OD person or a change manager or just like a business person more generally. And I was like, well, first of all, that's like the utmost compliment you could ever give me. But second of all, if you want me to just be the kind of guy that sits by my computer and crunches numbers all day, I'm probably not the guy you want to hire. Right. And the reality is a lot of folks in the field have just been the sit behind your computer, crunch your numbers kind of people. And the field is less than as a consequence.

[00:19:49] But I think that's because they started out as reporting people. They weren't starting out as people analytics people. And now we have that, as we were talking about before, the new generation that are people analytics people because they're people analytics people, not because they got thrown into the world.

[00:20:06] But I want to go back to one of the things you just said, which I think is really important, which is that the efficacy of people analytics in the world of AI isn't that it's going to blow people away who've never seen this stuff before. Because we've had dashboards forever and it's been field of dreams. People have not come. You've built it and they have not come.

[00:20:31] So you can't, to your point, you can't just assume that just because I say you have a turnover problem, half your staff have left. Yeah, I know half my staff have left. I don't need that. That's one of those moments where you go, why is HR bothering me? It needs to be able to give them clear understanding about what their business situation is. It's not their HR situation. To your point about being a better business person than you are being an HR person or a people analytics person.

[00:21:01] Yeah, I don't talk about this as much as I used to, but a lot of my early writings talk to like, what does it mean to be a good analyst? And one of the common riffs I used to have is like, the difference between a good analyst and a great analyst is a good analyst will answer your question. A great analyst will answer your question behind the question. Yeah. And so like, I got to this point where I just can't hear what people ask me anymore. I only hear what I think the thing behind their asking me is. Right, right. And it... You're cutting through the BS.

[00:21:31] Exactly. And so the simple way of putting it is, usually if somebody says, you know, can you give me my turnover? What they mean is they're actually asking four different questions. They're asking a descriptive question. What was happening with my turnover? They're asking a diagnostic question. Why was that happening? They're asking a predictive question, which is, what is going to happen with my turnover? And then they're asking a prescriptive question, which is, what should I do about it? Right.

[00:21:57] And then they're asking a financial question, which is, can I afford to do something with that? And is the juice going to be worth the squeeze of doing something about it? Right. And so what you realize is, I just outlined six questions within a very simple question, which is, what is my turnover? Right. And the reality is, every leader, depending on, again, I'm going to probably probe a little bit more. Every leader wants all those six answers all at once and in an immediate fashion. Right.

[00:22:27] And people on these teams largely historically were not equipped to do that. That's like a six-month project. And I've been given one of my common riffs lately is like what used to take six months, then with the advent of some certain like data management technologies became six weeks. Right. Because it came easier because the data at least wasn't siloed. It wasn't in different places. And then it moved from six weeks to six days because maybe we had the good business intelligence tools on top of that data.

[00:22:55] And then it moved from six days to six hours because of some of the advents that we had done with machine learning. And we had some of the questions kind of pre-baked in the oven, ready to go when people are asking. And then it moved from six hours to six minutes because of the advent of things like Cloud Code and Cloud Cowork, which are going to allow you to kind of do this in a self-service way. And the question is, what's going to happen when it moved from six minutes to six seconds? And what is going to be the thing that needs to happen there?

[00:23:22] Because what you're seeing is a logarithmic improvement. And you realize if it's already going from six months to six minutes, the nature of the job of somebody who does people analytics has fundamentally changed. And most people are still living in six-month land, if not six-week land. And so the reality is, it's like have to say, okay, if people can answer all of their questions in six seconds, what is it that you do that adds value?

[00:23:50] I think the bigger problem then becomes context. Yes. Do we have enough data? Do we have the data to tell that person, even in six seconds, that we know that people are leaving because they hate you? No, seriously, though. How many times have you seen turnover? Term reasons. No, but I mean, so many times it's the relationship between the employee and the manager.

[00:24:17] And no one ever captures that in an exit interview because they don't want to put they hit their manager. But that's the fundamental reality of it. And all those models will fail because the term reason says resigned. Yeah. Action, action, reason. The action reason is resigned. Sure. That tells you nothing. So you could do it six months. You could do it six minutes or six seconds. It doesn't matter. It's still going to be garbage.

[00:24:47] Well, there are ways of like what I try to do is, and this is not something I invented. There's plenty of other people out there that do this is you try to get the signals before they get corrupted. Right. Right. And so what you're seeing there is a case of trying to index to lagging indicators rather than leading indicators. Right. Right. And so I want to know in as immediate terms as possible, when is that relationship between the employee and the manager start to sour?

[00:25:18] Right. And then I can start my stopwatch and say, okay, I know three months from now, six months from now, 18 months from now, they're gone. Because the leading indicator is what really started the stopwatch ticking. But what's that signal though, Cole? What? What's that signal though? Is it taking excessive absence? Is it lateness? Is it disengagement if you have that as a metric? You know, there's a variety of ways of looking at it. There's some that are more creepy than others.

[00:25:46] My favorite one to rail on, which I give as a bad example, is, and I've seen this presented so many times and it's frustrating because people are like, we figured it out. The second that someone starts to pull their pay stubs. Right. We know that this is it. And the bank balances. Yeah, they're about to leave. And I was like, you know what people also pull their pay stubs for? When they're refinancing their mortgage. And when they're, you know, they're doing this all. And I was like, that is such a faulty signal.

[00:26:13] And so just imagine you're the manager of the HR business partner and you said, hey, I just wanted to check in. How are things going? We noticed that he's been in some odd behaviors lately. Creepy, looking over your shoulder, not helpful. Like, why are they doing this? Well, remember, there was also a Qualtrics, right? That used to read your emails and say, you know, well, they didn't, supposedly weren't reading your emails. They were just kind of figuring out how long did it take for you to get back to someone.

[00:26:41] And they figured that out by the, what do they call it? There was some kind of index. It's like probably the metadata and the digital exhaust type of things. They were talking about something. Yeah. But you know what I mean, right? They were trying to figure out, was that a leading indicator? Did a leading or lagging indicator of your level?

[00:26:59] And were we able to find other signals in the data to show that someone was going to leave because of the time period expanding or contracting and how they actually were acting and getting back to people? Yeah, I think. That's creepy. I've got, like, a few different directions I want to go with this.

[00:27:20] One is, you know, in your career in analytics, I think most people are trying to look for that, like, one silver bullet of, like, I found this insight. No one else has ever found it before. It's brilliant. Guess what? It's the email response time. We figured it out. We figured out turnover. And the reality is the chances are if you figured it out and no one else has, it's probably not true. Right? Or it's probably illegal. Or it's probably illegal. So then there's the second part.

[00:27:49] That's why I wanted to go a couple of different directions. Second of all is this whole concept of, like, informed consent. Do your employees know that you're reading all their emails or know that you're tracking their response time? And what is that going to do to the trust and the culture that you have at your organization? And so I've always thought of trying to predict turnover is not a technical issue. If you hoover up all the data and create the panopticon of an organization, which you can do. It's, like, not impossible.

[00:28:17] You can almost perfectly predict turnover, yes. But at what cost to your organization? So the reality is predicting turnover is not the tough part. It's actually doing something that maintains a high level of trust in your employee base. Right. That is where the juice is worth the squeeze that actually reduces the turnover that has been predicted. Predicting it and then lowering it are two different equations.

[00:28:42] But it goes back to the other point you made before, which was, is it a business problem? Correct. And how are we solving the business problem, not how are we solving the HR problem? The HR problem is we have to backfill because someone left. And we don't want to. That's a cost. That's a big cost. The business problem is who's walking out the door? What intellectual property is walking out the door? And do we care about it? Yep. Couldn't agree more. Well, and that's the problem.

[00:29:09] And that's where when we start talking about predicting turnover, I'm like, yeah, sometimes we don't care. And sometimes we actually want people to leave, especially if they're disengaged. So let's now go to the third question, which is what in the world should we be doing now to kind of not even prepare? Because we're living in a world with AI as part of people analytics. What should we be doing now? Where does this all take us? Yeah.

[00:29:36] I think I want to create like a two by two, which is what where should we be deploying humans? Where should we be deploying AI? And then also, what should we be expecting of our employees to build ourselves versus what should we still go to the market to get those assurances that like, you know, we didn't vibe code our way into destruction. Right. And I don't think.

[00:30:04] And also, can I just make a PSA here for a second? Of course. Like, can we quit treating AI agents as being personified humans? Yeah. This is so annoying. Giving them names, you mean? Giving them ID numbers? All of the things. Right. Because like the reality is the best way to be thinking about AI, like that box, that quadrant, is thinking of it in terms of what some people are calling tokenomics. So what is the cost per token to get the output that you want to see?

[00:30:34] And then from a workforce planning perspective, looking at the cost of labor of a human that it would have gotten that would have taken to produce the same output and looking at that through a strictly financial and workforce planning lens. Yeah. Free is never free when it comes to AI. It is not free. It's not. And those bills are coming due. Oh, my goodness. They're coming due and they're coming due quicker. I don't know if you use any of these tools, but like. Oh, yeah. All of them. I hit my limit every day.

[00:31:04] Dude, I literally spent $4,000 on perplexity the other day buying it for my entire team. Yeah. And that's the cheap version. And what I mean by that, not that you got the cheaper version. What I mean is you're only paying probably 10% of how much it costs perplexity from a data perspective. Well, yeah. We don't know how profitable perplexity is on this stuff yet. But the point that you're making is the good one, which is it ain't freaking free. It's not free.

[00:31:31] And the quicker because, again, that's why I said earlier it's going to slap teams across the face because that bill is going to start coming due in the next year or two. And organizations that are prepared with the right frameworks in place and aren't giving their AI agent Randy the personified view of like, why are we doing this? It's just it's like a collection of tokens that it takes to do a particular task at a particular cost structure. That's all it is. Right.

[00:32:00] And you can, you know, you can be friends with Randy as much as you want to be friends with Randy, but like you don't need to be. And then there's humans who actually can be friends with. But then there's the part of it, which is, again, HR is different than the business. HR deals with data that is more sensitive than most business data unless it's proprietary business data, which means, again, to use the should we be doing this ourselves or should we be buying off of the shelf? Right.

[00:32:26] Chances are if you're dealing with something that requires high levels of sensitivity and security, you should probably be buying something to deal with that because you're probably not good enough to build it yourself. And make it on prem. Correct. Don't host it in the freaking cloud where it can get hacked. Don't do lots of things, right? Like there's so many things you can do wrong.

[00:32:46] But the flip side of it is if you want to experiment, there's like HR has so many opportunities for kind of low like, you know, risk but high reward type of time savings type of activities. And I think that there's still like so many fruits right for the plucking out there to use that analogy.

[00:33:09] And so, I mean, I think teams sometimes they get a little ahead of their skis where they're either too afraid to do anything. And so they buy everything and do nothing with AI or they're so gung-ho on AI that they're like, we can build everything from the ground up ourselves. I'm like, maybe it should be somewhere in the middle. But before you even get that, I'm going to channel my inner Martha Cirione here. You've got to clean this shit up. Yeah, that too. It's bad.

[00:33:38] I mean, my very first podcast, 270 or some odd ago, was HR data is bad. HR data is just totally unclean. And I've never been proven wrong on that. It's just because we're human. We make mistakes. We don't put in the right date on things. So our effective dates are wrong or we miss a zero or whatever. We're just terrible at keeping this stuff clean.

[00:34:08] Yeah. And that's the stuff you're making decisions on and you're building models on. It's, do you want me to, I have an example of this that I always mention to people, which when they say, well, I mean, how bad could it be? So there's like data that is bad because like a process is bad or somebody, you know, like they incorrectly used the calculation or the data was ported from one location to another, but it didn't have the right ID.

[00:34:38] And therefore it didn't match. But it's the problem. But it's the problem I call the zip code problem. So employee is new to your organization at some part, some form somewhere. They need to put their, their home address. And let's say they just put the wrong zip code. Yeah. They don't recognize that they put it. It's wrong. No one knows it. They don't know that they fat fingered it. And then they never receive whatever mail or important documents to their house because it's just the wrong zip code.

[00:35:06] And the question, and people say, well, you just create an auditing process. And like, what is your auditing process? Right. How do you know it was wrong? How are you going to do that? Right. As long as it was five digits. Is the city state right? Yeah. But the zip code wrong? And so my point to that is, is like, there's like existentially wrong things that unless you just say to everybody, hey, did you type that right? Hey, on the second try, did you also type that right? Hey, on the third try, did you type that right?

[00:35:32] Even the Google API to correct addresses gets it wrong sometimes. Because it can't read your freaking mind. And so these problems of like wrong data, and that's just one example, are just like they're actually very deep and existential at times of like, what does it mean for something to be right? But it even creates a potential violation of law, though, Cole. Because if someone's in the wrong zip code, it might mean that you're that you're getting there, which, you know, HC or whatever it is.

[00:36:02] It might mean that your taxation's wrong and you've been under withholding or over withholding for them. Yeah. There's so many bad things that can happen. And that's just simply because one damn digit got misplaced. I mean, there's people out there that have their wrong name on their birth certificate. Yeah. Right? I mean, talk about unfortunate. Right? You can change that, but it's going to take a while.

[00:36:26] And the reality is, it's like, if I'm that person, all of my proof of documentation that says who I am is wrong. From the beginning. From the beginning. And so, like, there are just really key, deep issues that come to, like, getting correct data that most people just don't appreciate. And so, that, ironically, is the reason why my podcast is named Directionally Correct.

[00:36:51] Because, like I say, like, most analytics people, they get to this fork in the road and they realize that there's only two steps to the fork. One is nihilism, which is nothing can ever be right and everything's always going to be wrong forever and so I might as well give up. Or the other is, well, at least it's directionally correct. And I can live with that. And so, directionally correct is the pathway I've chosen to take. Yeah. But that's, to be honest, that's going to keep you sane. Yeah.

[00:37:18] Because trying to make, and I have this conversation all the time with people when we talk about compensation data and market analysis. And they go, but this is the statistic. And I said, yeah, it's a statistic. It's measuring a sample of the population. That's all it is. It's an art form, not a freaking science. Yes, that's the median. Yes, that's the 75th percentile. They're just numbers. Did we miss somebody in it? Yeah. It's a sample. Yeah.

[00:37:44] I mean, that's my riff on that is like the greatest sin in compensation is the sin that they try to project as if there's like science and there's accuracy to it. I'm like, come on. How many times do you talk to a comp person? They say, this is $20,000 below market. And they come back to you. It's like, oh, I just looked at the data again. It does look like it's $20,000 more. And it's like, no, that's not how this works. Wait, wait. I'm sticking my finger up in the air, checking which way the wind is blowing.

[00:38:12] I got a call from my boss, and now the wind is blowing the other direction. That's the reason why it's the same. But Cole, that's the reason why as a comp person and a reformed people analytics expert, one of the things I keep trying to tell people is, what's the business problem you're trying to solve here? Going back to the comment you made before. It's exactly the right conversation. What's the business problem and how do I help? How do I help the business?

[00:38:38] Because if you're asking me to justify a $20,000 increase on someone who's making $60,000, I'll tell you, I ain't going to do that. But if you're telling me that we've significantly underpaid this person for many years, we'll work on a plan to be able to get them up to the right level of pay. But we can't do that all in one go. So. Dude, I can talk about this all day. This has been such a pleasure. I haven't had this kind of laugh in a long time. So, Cole, thank you so much.

[00:39:07] Thanks for having me, David. I'm glad I could give you a laugh. If you need a hug, I got more hugs. Yes, we need a hug. All the things. Thanks, virtual. But we're going to have to have you back to carry on the conversation. And maybe we'll do another podcast like we were talking about before. We'll just talk about podcasting as well. I think that would be so much fun. I think zero people would tune in and it would be great. No, I think people would vicariously like to live through the lens of two podcasters who've done a ton of podcasts and now want to vent about it. There's a lot of stories, if nothing else.

[00:39:37] That's for sure. Oh, yeah. Oh, absolutely. Sure. Well, again, thank you, Cole. You're awesome. Appreciate you being here. Yeah. Thanks for having me, David. And thank you all for listening. Take care and stay safe. Thank you for listening to the HR Data Labs podcast. Don't forget to hit subscribe and share it with your network. You can also check out the recordings on Spotify or the HR channel now on Roku and Fire TV. Thank you. Take care and stay safe.