Richard Rosenow: I think it's right in the name. It's the foundation of everything we get up to in the people data space. So truly, if you try to build people analytics without having that core foundation stability, the confidence in your data, it's like building on sand.
And so I think really starting with, Do I have confidence in my data? Do I understand my data? Do I have the right data?
Before you start asking some of these deeper questions, because otherwise you'll answer those questions. But you might answer them with bad data or the wrong data, or you may not have the data you need by the time you get to the end of the project? So truly, I think in the name of foundations is a starting spot.
In your experience, what are some common mistakes that organizations make when it comes to building their data foundations?
Richard Rosenow: That's a good question. If you've made mistakes, if you found yourself in the kind of trenches of data projects that are not going the right way, give yourself some credit, give yourself some space, because this is a new space for a lot of HR teams. But really, I think I see a couple things about HR specifically, that teams tend to overlook. Because I think a lot of times IT teams will say, Hey, we've got data, other places, we do data projects all the time, we can do them for HR, don't worry.
But the reality is that HR data is different. And it's different for some technical reasons as well it's just like a very big ethical reason, which is, this is data about people end their work.
And the relationship between people and their work is so important that if we get that wrong, if we make mistakes here, it doesn't just like throw off our forecast a little bit or maybe have some downstream impact to the kind of sales but it could change, something like to get paid differently to get a performance review that's not quite right, to have their data misaligned in different ways.
It is really a life changing thing to mess up data. If I think there's that, just that ethical piece right off the bat that this is a human, this is someone who has to livelihood with friends, family, and we've got to treat that with a different level of respect. Then the other side is a lot of times we'll see HR data and time is a massive element. And the relationship between things we do in HR and time is under spoken about, in the sense that it's like if you get promoted after you quit, well, that doesn't quite work. You've got to get those sequences, right.
And you've got to say, okay, when did this person come in? Where did they go? How did they move? How do they transfer? And that sequence of all of those times across all of those systems have to harmonize. That is a massive data problem. And it's one that's really underappreciated by a lot of HR teams, and IT teams. And it really comes down to things like back data changes i.e. hey, that person quit three months ago, we forgot to enter it.
Well, hopefully not that bad but I've seen it happen, and it gets entered in the back. Now, you've got to reinstate the data warehouse, you've got to go back and redo things and rebuild things, and there's some real landmines there.
I think the other one that really comes to mind is around the regulatory.
So HR data, again, being people data is highly regulated, especially in the EU, but also across New York, California, other places around the world, where it's a data set that really has protections in place, which are good, healthy protections, these are good things to be aware of and be thinking about.
But if it's an IT team that maybe isn't as versed in that, they might not have the HR legal background, and when they start to approach those things, they can get a little bit over their heads. Maybe the last one I'll bring up is around maintenance. Everybody likes to build a data foundation. Everybody wants to do a fun project to build from scratch. Let's do it. Let's build all the systems together.
But then suddenly, it's built, and then you change your HRIS, and then you want to change your ATS, and then you want to change the performance system, and oh my goodness, we have to redo everything again, and nobody wants to go back. So I think about maintenance is one that a lot of teams tend to overlook in those early stages.
Felicia Shakiba: Wow. And you I've seen other companies, they've had a restructure and still don't know what their headcount is because they forgot to take people out of the system. And so your data is so dirty. And that type of timeline of making sure and validating that data, it could take nine months to a year to actually figure out.
I understand how those mistakes could have tremendous impact on an organization.
Richard Rosenow (05:16): Ya, and I think a lot of HR leaders as they approach data, it's really the first time they've gone into that, because a lot of times HR professionals, we come up through the ranks very different ways. You get to the top of each organization, you're then faced with this challenge around, how do I handle my data? How do I support my data in these different ways. And without that experience, looking across a lot of companies, you can feel like you're making a lot of mistakes. But the reality is like this space is hard.
This is difficult to dig in here. And as an HR leader, I think sometimes we doubt ourselves in that. To anyone listening just to give them a little bit of confidence here, you're not alone, everybody is struggling with this. HR data is one of the most complex data sets across the business just hands down from someone who has been deep in this industry for a long time, and have some confidence that this is hard.
Felicia Shakiba: I completely agree. And I think that is warranted to give people the confidence to say, Hey, it's okay to mess up. We're all messing up together. If you don't move forward and make progress and fail, you're never going to get to the other side of this story. And data foundations is really the only thing that we're really focused on in this episode. And I could only imagine the amount of conversations that will come from this episode. But with that being said, the importance of the foundation is so critical that has a rippling effect on everything else that we do. So that's why it's important to get it right.
But when we think about starting to build a solid foundation, what are those key components of an effective data architecture?
Richard Rosenow (06:43): It's a good question. And I'm going to take this a little bit away from the technical side a little bit more to the outcomes you should be looking for. One of the things I think about when somebody has a really strong data foundation within the workforce is confidence. You should be walking away confident that you know what reality is. That's a funny way to say it, but a lot of times it's a little bit unclear, like what is our headcount? As we said earlier and they're not just random teams, that's fortune 50 companies that are struggling with someone who's kind of just, What is the reality of our company today from a headcount, attrition, diversity, talent, you name it there.
I think about you should have confidence when you go into those conversations with finance leaders, with business partners to say, Hey, I know my stuff is clear, clean, architected appropriately, and I can speak to what's happening on the ground. That immediately removes so many debates.
Because when you get into the room, you say, hey, it's XYZ number, and then Finance says, hey, it's actually ABC, and suddenly, you're in this back and forth with this team, and that small piece of data - it's only off by three or four, and I've seen that. Like headcount or 10s of 1000s are off by three or four, and then suddenly, the questions.
And it's this barrage of questions that just takes you away from the real goal of meeting, because the goal of the meeting is to drive the business forward, it's to drive business outcomes, and use data to get there. But when your data is questionable, suddenly that just starts to crumble. If I think again, coming back to the core word of foundation, like having that confidence, to come into a conversation and say, I know my stuff. This is what's happening across the workforce. Let's move forward from there. That's how you know when you have a strong foundation.
Felicia Shakiba: I think what you brought up was really important, which was headcount.
When you think about headcount, how do organizations get this wrong? Talk to me a little bit more about that.
Richard Rosenow: This is a funny one. And it's one that I've used this question in interviews when I'm talking to people analytic leaders, which is like, Hey, could you give me six calculations for headcount? And it's one of those like, you could start to see the wheels start turning, because there's not one way to do headcount. There's 50, 60, hundreds of ways to do headcount. You could do 30 day rolling averages, you could do a start of period, end of period on an average, you could do point in time at a given day.
Do you count the people who quit that day in that headcount number? It's one of these places, I tend to refer to this as quantum HR. So we talk about data as the atomic level, like you kind of zoom out from like, programs and processes all the way down to the data of HR, but then suddenly, you get to that atomic level, and you go further, and you zoom into this, like quantum world, that's all much bigger than it should be, but the reality is these metrics, these measures, there's so many blogs and people out there that are saying, Here's the 10 things you should do, this is what you need.
And it simplifies it to a point that it's over simplified, because every company is so different. And we see that at One Model looking at across all the different customers we work with, we define headcount uniquely to each customer the way they need. And headcount is such a basic metric that has so many profound effects on all the other metrics and measures you start to measure. But some companies need it a very specific way for very specific reports. And if you need that- need to define it that way, you go after it there. And you've got to have that in your report in that specific way.
Felicia Shakiba: And I hear exactly what you're saying, because in global organizations, you're talking about headcount under a function, headcount in a location, headcount in a country, headcount of full time employees within a function within a location. There's so many different ways to parse it. Yeah, so I hear what you're saying, especially in those larger organization the amount of different outputs that one could really look at and choose to look at for decision making. So it is a very big deal in some of these larger organizations.
Richard Rosenow (10:12): I'll say though, architecture reduces that conversation. So the purpose of having nice architecture in place and having a data foundation in place and saying, Hey, across this globalization, we have 50 different ways to define headcount today, that's got to be one. Or maybe it's two, and there's a very specific use case for that second one, but it's got to be less.
And suddenly all of that noise that takes up so much time, energy and space to validate and correct and change reports and dig in on questions evaporates, and you've got more time to spend on strategy. And I think that's one of the real powers of getting a strong data foundation, and having that data governance in place, is just reducing the amount of noise you spend validating.
Richard, how do we get started? What steps should an organization take to improve their data?
Richard Rosenow (10:55): Oh, there's so many editing services. I hate to say it, but it's one of those like consulting kind of things, like it depends. But anyway, there's a little bit of depends here. I think getting access to your data is step one. So the HR organization, I've seen a lot of organizations that struggle with either tech systems that won't relinquish their data, or maybe they don't have a tech system in place. Or maybe they're keeping track of things in spreadsheets here and Google Sheets there and the data might be everywhere.
Take step one, it's a little bit of maybe a page out of Marie Kondo, like you've got to get it all out, you got to get it all out of the closets, get it all under the bed, and then work through it from there. And to be able to kind of parse through and understand what do I have? Where is it happening? And what reports are going out? I think it's that inventory moment, and then reducing. A lot of times, once you start the inventory, you can say, Okay, we've got a lot of extra work happening in a lot of places, we've had a lot of people touching this and reusable ghost factories of data.
Let's get that down to one organization. There's places where decentralization makes sense. But I think that's centralization first, to understand your data, govern your data, work with your data, and then relinquishing that back out. There's a time and a place for that. But I think step one is just understand what you have and get access to it.
Which, unfortunately, because of the variety of HR tech tools we have, in technologies that are given, I think it's like 40, or 50 technology some companies have it was written in Stacey Harris's HR tech report, just a tremendous amount of organizational diversity within the HR function, our data sits in so many places. And that's different than other functions. But I think the ability to get it all out, get it organized, get centralized and then see what you have.
So centralize and then create an inventory check, and then create a governance, perhaps of who can access this data and who has access to this data. Just those simple steps could probably take months.
Richard Rosenow: It's an been incredible one. And even just hearing them back and like that's such like a highbrow advice to give people because we all know that's not the way things actually work. Which is the reality is you've got to build the plane while you're flying it, you have to still deliver, you have to still be doing these reports while you're also doing this kind of piece. It is that sort of juggling.
But one of the things we talked about when I was on people analytics teams was the power of inefficiency is sometimes you have to do things badly. And sometimes you do them badly for a little while, and then you approve incrementally over time. And as much as it would be really beautiful to kind of wipe things clean and start fresh, get going, most of the time in HR, we have to just make do with what we have and keep trudging forward. So a little bit of this is like take the steps where you can where you have them and move onward.
I think maybe another easy step that comes to mind, though is a lot of times we think of HR operations teams as very different than data or people analytics teams. But HR operations is really where data starts. A lot of times that's where data entry happens that where processes are defined, that's where the data actually flows from the organization into your tech systems. And so that's actually one of the earliest stocks to start when you're trying to think about, how do I get good data out of my systems? You've got to think about how that data comes into place, too.
Felicia Shakiba: Going back to your comment about time, and that we have to build the plane while we're ready to fly it while it's going down the cliff, right? I think people maybe not understand sometimes that while we're collecting and centralizing this data and finding who has access and who has the capability to input and so forth, those data changes are happening daily, if not by the hour.
And so making sure that the data is valid in real time, or at least on a daily or weekly basis is so critical while we're trying to clean the data and gather the data and centralize it like you mentioned. So it's a heavy lift.
Richard Rosenow: Yeah. Ya, you know, I'll put out there what comes to mind is there's a lot of vendors, I won't name names here, but there's vendors that say, wait until you install our software and our software will finally fix your problem with data. So like when you're switching an HRIS or you're switching your ATS the new one always says, We will fix it, you can stop worrying about your previous one the tech will fix it.
And again is that I've seen a lot of teams just wait to do data governance then. They say hey, when this goes live in two years, then we'll have to governance in place, so we're okay until then. But the reality is it does, it starts with your processes, it starts with what you're doing today. And that incremental change is the only way we make any change happen within a company.
And I would say it's never too late to start, and it's never too early to start. So get involved when you can.
Felicia Shakiba: Now, in many cases, the reality is that people operations teams don't have that green light to prioritize organizing their data, because let's face it, it doesn't have short term value.
So how can Chief People Officers and their teams make the case for prioritizing data, foundations?
Richard Rosenow (15:35): This is really tough. The unfortunate reality a lot of HR teams find themselves in is the HR team understands the architecture a little bit, and the IT team doesn't really care. I wish the IT team can, and God bless your IT team, if they care about people dating, if they're involved with you, you've got a great partner, that's amazing.
A lot of them just don't quite understand what we're trying to do within the people team, because we're in different functions, and that's okay, too. But this architecture space sits really between IT and HR and requires expertise and knowledge across those. I've got kind of two answers here, I think one is that the vendor landscape is getting better. And when you have these sorts of hidden pockets of pain that if you've solved them, no one gives you credit for them, try it outsource those.
Find someone that can come in and help you with that, find a consultant, find an organization, find a tech vendor that can support you with that architecture piece. I think the other part of me is getting increasingly to a point where I'm, I would like more HR teams to just riot a little bit.
I think we're at a point in 2023, that you should not be managing any part of your business without data. It is a shameful thing that HR leaders have to go beg, borrow, steal, when the reality is most of the cost of the business sits within our workforce, we know that our business is made or broken by our people, and the HR leader is asking for clean data. I think that's a terrible spot to be.
And so I think part of me is like, hey, we can accommodate we can figure out ways to get around it, we can give these tricky kind of methods to speak to your peers, but the end of the day, it's like, we've got to start getting loud. And I think we can get a little bit angry at this point, because everybody else has this.
Felicia Shakiba: Exactly. I mean, I know that from speaking with HR vendors, they realize that HR is the least tooled function in the business. And I think there is this stigma, this lack of respect for HR teams, because a lot of the data that we gather is qualitative. It sits in our mind, it's it's in conversations from the day to day. And we don't really have any quantitative evidence to share, oftentimes, but I think it is that new year of 2024, that we should be saying, we need to prioritize clean data, in order to have data, in order to show and prove results, that we are valuable as a function. And not just that, but valuable to the business in order to figure out how to make great decisions.
In your experience, I'm curious, what are some examples of how organizations have successfully unlocked new discoveries by improving their data foundations? And how did that impact the business?
Richard Rosenow (18:05): It's a little tricky that data foundation sits early in the value process, because you've got to get your data foundation in place, then you've got to do some analysis, then you've got to tell some stories. And then you've got to make decisions happen. And so by the time the decision happens, the credit kind of sits usually one step before the decision, usually. And we very rarely see that credit go all the way back to this work that happens at this data foundation piece.
It's one of these areas in business, they talk about these as like hygiene areas where it's like, if you have good hygiene, nobody really notices. If you have bad hygiene, people call it out. Data foundations a little bit like that, if it's bad, people notice. But as soon as it gets good, it turns invisible, which is a little tricky. I would say most of the greatest people, analytics teams are built on top of phenomenal data organizations.
I saw just tremendous work when I was at Facebook of teams that were doing some just world class work with diversity and promotions and transfers and just trying to understand how we made sense of a thousand person recruiting organization sourcing talent at sale. That all happened because Facebook had invested in an 800 plus person data engineering team. It was an incredible resource like truly separate from the people analytics team, but an incredible resource led by just brilliant people working day in day out to build the world's best data repositories.
And when we've seen things like that...
Felicia Shakiba: I'm jealous.
Richard Rosenow: Oh, I've been jealous ever since. It's a phenomenal organization over there. And they were able to get to incredible heights because they had this foundation they stood on. And I think it's tough when we hear these stories that a lot of conferences of these brilliant case studies that people do you hear about that final moment where it's like, we change the CEO's mind, but you don't hear about a lot those kind of like it took us a year to get organized on headcount.
On the back end, there's a lot of hidden heroes in the workspace that are doing this enterprise data architecture or HR data architecture.
Felicia Shakiba: Makes sense. Building a foundation is critical, as you've already mentioned, clearly.
What does the maintenance look like for ensuring the ongoing governance of their now clean data?
Richard Rosenow (19:59): Yeah, maintenance is a great, tricky question. I think like we were saying earlier, people love to build this stuff. But then suddenly, you've got to hand off a Workday pipeline. And there is not a single date engineer that went to school saying, someday I'm gonna grow up, and I'm gonna maintain a Workday pipeline, that's gonna be my job. But Workday changes their pipeline, they change their API, they change their data sets, you maybe configure a new thing within workday, it is a living thing.
And I think it's what makes HR data a little bit different than other IT projects or tech projects that are out there is you can't do it once. The HR data is a living, breathing, tech amalgamation of all these different technologies we have across HR. And so as we start to bring in more datasets, as we bring in the next one, the next one, the next one, they start to interplay and interact in different ways too. And the skill starts to get in different ways.
So I think maintenance is a nonstop job. And it's why a lot of people end up looking to vendors like One Model that are out there that say, Hey, we want to build this, but we might not be on to maintain this indefinitely. That's not our core part of our business and be able to outsource that to consulting teams and vendors is the right move for those companies. It's when we hear from a lot of teams that are fed up with maintenance that weren't getting the maintenance, they expected that weren't able to get the new data sets they were looking for, because it's just so big, to try to keep this thing running once you have it.
Felicia Shakiba: And I think HR Business Partners have a really key role into the maintenance, right? They are essentially the point person for a specific client group within the business. And they're responsible for the input data, or when changes get made, and so forth. And so I remember, organizations that I have worked with, have really dedicated an HR Business Partner to kind of have ownership of a client groups data, like an engineering groups data, or, and so forth. And I think that's a lot of responsibility.
The other option is to have the responsibility on the Center of Excellence, the COE, and having a specific person, one person in charge within the COE that has responsibility over the validity of the data for the entire organization. What have you seen that might work best?
Richard Rosenow (21:53): I was reflecting when you said the HRBP. If businesses didn't change maintenance would be easy. But we all know, there's just constant change from - I was a former HRBP way back when it just like constant change, re-orgs, businesses changing business units, and things that are happening, the business takes a left turn suddenly. And it's that, that keeping an eye on that, to understand that and be able to translate that back to these data teams, I think that's a tremendous part of this, because it doesn't happen in isolation.
There's not any data foundation org that sits quietly by itself and just waters away it kind of their data world. It's really an interplay of the HR system that says, we'll keep eyes and ears on what's happening, what do we need? What problems are we trying to solve? Do we have what we need to solve those problems? And that interplay back and forth? I think we're gonna start with it, and where have I seen it work well?
I think it's a heavy investment. It's a much heavier investment than people expect, right? I've seen it work well, I've been a customer One Model before I came over here, I hate to keep kind of bringing it that way. But at the end of the day, it's the vendors are going to take this space, eventually. I think about this in the sense that you wouldn't build your own payroll system today. That would be a wild thing to try to do. If they were going to build a payroll system, we're going to maintain it ourselves.
For all these international locations, the payroll vendors have been around really the longest in the HR tech space, they've won that category. You start to see that over the ATS's and HRIS's, you wouldn't try to build that in house. Eventually, these fringe tech vendors come into kind of take over and say, Hey, we got this, because this is massive, and maintaining this by yourself is an undertaking.
That's a different strategy than the business you're trying to run. And so we're starting to see the vendor landscape really evolve in that space. We have multiple people in this space of people, analytics vendors, we've been here for 10 years, I think they're starting to finally realize that started from scratch and trying to maintain this over time, it's just untenable.
Felicia Shakiba: Yeah, it is a heavy lift. If you think about how organizations or the function of HR might change in the future is that once they become more analytic heavy, a significant part of their job would be maintaining the data for their client group, right, that would be a bigger piece of their responsibility, as the function kind of takes shape over the next year or two.
Richard Rosenow: You know, it's funny, I think we've actually hit the peak of how technical HR will need to be. And this is somewhat related to the kind of vendors in the space because they didn't catch up to a great example, they made in 25 minutes. And without saying, I've got to say, and now ChatGPT coming into the picture. But really, what we're seeing is we're starting to see the need for SQL go away. That's not data scientists, data scientists will still need SQL, but there were calls for the past couple years, like HR people should be learning SQL, they should be learning how to code, they should get involved with these different technical pieces.
And some of that still true. And if you're passionate about it, go for it. But the reality is like CheckGPT knows SQL better than I do. I will not learn SQL to the point that I had known in the past ever again in my life because ChatGPT handles it for me and can handle it for me for the work I need to do. I think HR 20 years from now will look more like HR did 20 years ago, than it does today. The goal is to have the technology start to fade into the background, and really the stuff that only humans can do.
That's where HR should be focusing efforts, because we're starting to see more and more automations and technologies come up that can take on a lot of administrative work, that can take on some of these routine things and utilize these data challenges and take those off of the HR person's plate. But what computers are not good at is the stuff where HR has always been strong. And that's listening to people that's working with people that really like cultural ethnography of the workplace, that really only the human mind can do.
I would love to see HR move more that direction, and let the vendors take over more of the tech spot and let that start to fade. I think going forward, we'll see... It could be an interesting time for HR as we move a little bit back towards our qualitative roots. But I see more of the vendors kind of take on more of that tech stuff over time.
And what advice would you give to Chief People Officers who might feel overwhelmed or ill equipped to tackle workforce data and analytics today?
Richard Rosenow (25:49): I think the first piece of advice is it's okay to feel overwhelmed. There was a Stanford study I saw about the people analytics space, and it was academics that we're looking at and tried to understand what do we call this place?
That was the whole study and there's like 60 different names for it. It's a truly remarkably complex thing right now, because we've seen a new function basically come into existence as a function in the past kind of 10-15 years. We've been doing workforce data stuff really since World War II, and psychoanalysis all those good things have been happening for a long time and a lot of places, but as a formal function, like recruiting or total rewards or people operations, people analytics is pretty new.
And I think the first piece of advice is reach out, talk to people and get comfortable with feeling confused, because it's going to be confusing. It's confusing for those of us that are experts.
But I will say the people analytics community bar none is tremendous. People analytics people can't stop talking about it. They love it. They love talking to the people about it. We love get people up to speed. It's a group that really there's a sort of a venn diagram with people and data people. And people analytics, people kind of sit in the middle and they can't get enough of it, and reach out to the community. Find someone to talk to get connected, reach out to me, I'd love to talk to you about people analytics just in general. And if you have any questions about it, please reach out anytime.
Felicia Shakiba: Thank you so much, Richard, this has been a wonderful chat, and we appreciate you being here.
Richard Rosenow: It's been a ton of fun. Thank you, Felicia.
