How strategy leaders can achieve their moonshot goals

Summary

Lydia Wu, an expert in people analytics and AI, discusses the importance of people analytics at Panasonic and how it shaped the company’s HR strategy. She shares examples of using data and people insights to achieve strategic business objectives, such as connecting production line data to people data and understanding informal feedback discrepancies by gender. Lydia also addresses the challenges of data quality and accessibility and provides advice on building a business case and where to start with people analytics. Lydia Wu discusses the importance of financial survival in business and the need to tie people dollars to business dollars. She shares her experience of starting with basic tools and data, such as Microsoft Office Suite, and the value of curiosity in starting conversations. Lydia also discusses the rent vs buy debate in HR technology and the need for HR leaders to try out tools before committing to them. She highlights the power of people analytics in providing real-time insights to line leaders and driving conversations around high performance culture and employee well-being. Finally, she explores the future of HR technology and the shift towards targeting individual practitioners before licensing to enterprises.

Key Takeaways

• People analytics is crucial for understanding the impact of the human element on business outcomes.
• Connecting production line data to people data can provide valuable insights into improving productivity and reducing defects.
• Analyzing informal feedback can uncover gender disparities and inform initiatives to promote equality and inclusion.
• Addressing data quality and accessibility challenges requires a combination of manual cleanup, data architecture design, and building trust with employees.

Full Transcript

Abhinav (00:00)
you ever wondered how Fortune 100 companies with tens of thousands of employees do people analytics? Have you ever struggled with connecting people data to its business impact? Have you ever thought how AI is going to change the world of people insights? Are you just intrigued by what is the future of people analytics?

and how important role will it play in helping companies beat its competitors.

Hi Everyone I’m Abhinav and welcome to the Peoplebox Analytics Talk, where we invite incredible leaders to go deep into the fascinating intersection of data, AI and people.

And today, I’m delighted to welcome Lydia Wu. Lydia started her journey at Accenture and Deloitte, providing people analytics and HR transformation consulting.

to large enterprises. Later, she joined Panasonic to head their talent analytics. She’s an advisor to many startups and large companies in this space. And her knowledge and passion for people analytics and AI is second to none. Welcome to the show, Lydia.

Lydia Wu (01:04)
Thank you for having me. Happy to be here.

Abhinav (01:07)
Thank you. Lydia, let’s start with Panasonic. How important is people analytics to Panasonic and what role did it play in shaping companies overall HR strategy?

Lydia Wu (01:19)
Of course. So for Panasonic, I think like many other organizations who are growing, transforming and reshaping their strategy in the current macroeconomic environment, analytics is incredibly important. When we started on the journey back in 2018, we didn’t quite know exactly what we were getting ourselves into. It was more so the fact that understanding we had a lot of data in the cloud and we had to get all the data out of the cloud and make something of it.

But as we went on this journey, as we morphed and as we grew, what we realized was that analytics was really the key to unlock the value HR held in the company. Especially when you think about the manufacturing environment, which I’m working right now, 24 seven, 365 around the clock. It is incredibly important not to only understand what your production line outputs and inputs are, but also to understand how the human element really impacts the production lines.

Because For any leader out there who thinks that as long as you perfect a process, the people don’t quite matter as much, I will for sure tell you that the engagement of your line managers, the engagement of your QA person on that line is actually gonna impact the defect rates as well as the good throughputs that is gonna happen on the line. That is something that we’ve been able to statistically prove and I think that has really helped us have the conversation around how to better support our people because,

It’s not just a good people decision, but at the end of the day, it’s also a good business decision.

Abhinav (02:48)
What was the trigger for the company to build the best practices for people analytics and say, we need data. So we need Lydia.

Lydia Wu (02:48)
Yes.

Absolutely. Well, we need Lydia came through the interview process. Let’s be honest for a moment. We need data was really coming from the complexity of this organization because I think when you think about Panasonic or when you think about any larger brand names that are out there, not everyone necessarily thinks about the complexity of the legal holding structure, the subsidiary structure. And because of the organic nature of those organizational setups,

What ends up happening is that you also have disparate data sources because if your company grew through M &A or if it grew through acquisition of any sort, the data sources that you acquire, the historical data transfers that you acquire, all of that gets stored somewhere, but no one was really looking at it. And we were at that point in 2018 sitting on about 10 years at least worth of historical data.

that we thought we should make something out of. And if you think about the environment back in 2018, the war on talent was really starting to heat up back then. And what we realized was that across buy, build and rent, we don’t exactly have the top dollars in the industry for the engineers, for the top sales folks. And what we really had to do was get smart about what it meant to be a part of Panasonic and really what it meant to…

the talent that we needed and to grow and develop them internally and come up with the quote unquote business case. I have my sentiments about that word, but really the financial metrics behind why it makes sense to invest in your people and not just let it burn and churn and hire additional from the market.

Abhinav (04:34)
It’s so amazing you use the word be smart and that’s so right because not everybody has the money and the brand that many of these large companies like Google, Facebook, Amazon hold, And I think the only mantra for them, like you said, very rightly is be smart and be data driven. Now going even before the Panasonic, you started your career as a consultant with Accenture and Deloitte and I’m sure you must have worked with.

loads of large enterprises. And later you led the same role at Panasonic. So how different was it to be on the execution side than from the consulting one?

Lydia Wu (05:09)
Very. So I actually went to Panasonic because when I was in consulting, first of all, I got the bug to look into people analytics. But I also realized the challenge with working in consulting, especially when you’re a talent strategy, HR strategy consultant, is that your bill rates can only afford 20 % of the work on any transformation or implementation project. And usually for me, what that used to look like was,

You come up with a business case, maybe you’ll do a work breakdown analysis, maybe you’ll map some processes, maybe you’ll come up with a taxonomy. But when the rubber hits the road or really the 80 % of the work where it really, really matters to an organization executing on a transformation, nobody can afford you at that point anymore because they thought that the strategy consultants were supposed to deliver a deck and somehow magic will just happen and the implementation will happen on its own. So having seen that through one too many times and having.

deliver those strategy deck, which in a way is kind of like my brain children if I think about it. What I decided to do was like, you know what, I’m gonna see one through to fruition and let’s see what it’s actually like. Let’s see what people are actually stuck with. What they’re actually challenged with. What’s keeping the business partners up at night? What’s keeping talent management, talent acquisition up at night?

Abhinav (06:26)
And I agree if you don’t understand what’s the problem that they are facing in a day to day, it’s very hard to even understand the value of that data. And a lot of our audience are HR and business leaders and they always try to understand the impact. So Lydia, could you provide some examples of…

how you are able to use these data and people insights to achieve some strategic business objectives.

Lydia Wu (06:53)
Absolutely. So I think the tails are many because it’s been six years and on average, the way I looked at it was that we would deliver about three to four banner projects or like top high level hitting projects on an annual basis to one balance the business need, but also to balance the internal team need to do the research, get the data clean and really get the homework done.

I think one of the most fascinating projects I’ve undertaken in manufacturing was for the first time ever connecting production line data to people data.

For the first time ever, we were able to look at production leaders in the eye and say, hey, you know that line that you thought you need 55 people on? You only need 48 because by person number 49 in the last 12 months,

your defect rate has gone up. Regardless of why, regardless of how, instead of asking for bodies, let’s start looking at why is 48 the magic number for you. Line two, hey, let’s start looking at why 46 is a magic number for you, because every single line was so different. And that was actually how we were able to backtrack to the fact that engagement did matter, frontline manager effectiveness did matter. And it’s not just because satisfaction or employee sentiments.

It’s actually the outputs. And right now I’m in a battery manufacturing world. So for anyone out there who understand what battery manufacturing is like, you would know that when you scrap a battery at the end of the line, it’s really hard to recycle that raw material back into the process again. So scraps for us aren’t like, oh, just grind it into paper pulp and try it over again and you’re fine. It is actually dollars wasted natural resources that the earth has limited amounts of.

So for us, that was incredibly important. And for us, that was genuinely one of the biggest business objectives that we were able to convey that essentially also led us to be able to implement different HR systems, different ERPs, so we can really be smart about how we do the work at a manufacturing plant. So that’s sort of the manufacturing and throughput side of the house. I think the other part of the work that we did that was really fascinating for me was actually during the pandemic era.

This was back when I was with the regional head offices. And at that time, everyone was always talking about personas, personas in your employee base, design your journey according to personas. But it was like a very happy and fluffy concept and arguably to some organizations it still is today. So I went on this journey saying, okay, does personas actually matter? Should we really pay attention to them?

And what we ended up doing was that we ran a series of longitudinal engagement surveys. It was incredibly insightful because in the pandemic era, what it allowed us to understand was that.

hey, the traditional way we’ve been doing benefits with 401k, healthcare, eye care, dental, so on and so forth, it doesn’t actually work for everyone because if you’re a millennial or Gen Z coming into the workforce back then, having vision and dental didn’t really matter to you. Having somebody being able to tell you how to do taxes and financial plan for you and teach you about a 401k was actually what mattered. It sounds again, really, really intuitive, but.

in the face of limited investment dollars. It was also something that a lot of times when we were in boardrooms where the conversation went something to the effect of like, yeah, it’s nice to have, but we have all these other things, so why bother? So first of all, that piece of research led to us establishing a wellbeing credit across the organization to say, hey, in addition to everything that we think you need, here’s a bucket of money that we’re gonna give you for you to figure out what you need. And here’s a category of things that you can spend it on.

So that’s part one. Part two, as a subsidiary part of that research, we also looked into talent management because again, personas, what the heck does that even mean for talent management? It sounds very happy fluffy. Why can’t you just do bare bones talent management, get an annual performance review done and call it a day? Why are we spending money on this? It was all of these great questions that were coming up. And I think it’s questions a lot of HR departments are still facing today.

And what we did was that we connected the longitudinal satisfaction data to talent management to performance feedback data, hooked it all up to demographics, because the beauty of analytics is you’re allowed to wire data sets together that didn’t historically go together.

And it actually floored us to find that from a formal feedback perspective, so your annual performance reviews, everyone was about the same in terms of satisfaction. Nobody really loved it, but they understood it. And they were like, yeah, things are going well. We get it. Let’s move on. But what was amazing with us realizing the informal feedback, so the casual check -ins and the hey, how am I doings, those had such a discrepancy between gender.

in terms of satisfaction rates, that we were actually paused in the midst of a conversation and meeting to say, hang on, what’s going on here? And when we do these analysis, we actually run regressions against all demographics. So not just gender, but like ethnicity, generation, whatever region, so on and so forth, managerial population. And gender was the only one that stood out enough with a significant score that we’re like, wait, what’s going on? Ran a focus group. And that was actually when we uncovered that,

As women in the organization are going through the day to day experiences, they were actually finding it a lot harder to have these water cooler conversations with their most often are male. And what ended up happening is that because of that level of discomfort in terms of how we’re socialized and just how things work in general, they weren’t getting as much feedback. They weren’t getting as much insights as their male colleagues were in terms of like, hey, how was your weekend? Like the water cooler chats that eventually go into work.

Again, as I’m telling you this, it sounds so intuitive, but as HR, you always wonder, is this anecdotal or is this legit? And we actually statistically proved out the fact that it was legit and it is a genuine statistical problem in the organization. So coming straight out of that, what we did was that we actually instituted a formal mentorship program, even in playing fields across genders, across different parts of the population.

all because we had to wear without to say we’re gonna gather the data and we’re gonna let the data guide our investment and guide our decisions.

Abhinav (13:17)
This is fascinating, Lydia, because actually both the examples, the assembly line and the watercolors, are so insightful. And I’m wondering, it’s not just for the HR, but even for the leaders and the managers, just getting the data could help them so much with achieving their strategic objective, retaining their team, improving the whole employee experience.

In both of the examples, you talk about real large set of data. And whenever I’m talking to the HR leaders about, you know, starting people analytics or just using the whole aspect of people insights, one of the major challenges that they spoke about is the data quality and accessibility, For a large organization like Panasonic, I can imagine, or I can’t even imagine the magnitude of this problem.

How did you overcome that?

Lydia Wu (14:09)
Yes. So first of all, for whoever’s listening to this, we’re all going to virtually hold hands for a moment and just acknowledge the fact that unless you turn off your HCM system, you are never going to have a hundred percent clean data. It’s a pipe dream. And I think most of us who work in analytics have given up at this point. So a couple of different things in terms of how we dealt with this. When I first started in 2018, it was a one woman army, one person shot.

Abhinav (14:16)
Hahaha!

Lydia Wu (14:37)
and on a shoestring budget as well. So it was a lot of cleaning after the fact. It was a lot of dumping everything out into Excel, recognizing that different companies were using different data fields differently, because under the region column, I would have somebody use it to identify full -time, part -time employees. I would have someone identifying the home region. I would have somebody else identifying the office region. And it was a little crazy to see what all the values were within the broader ecosystem.

And on top of that, the way we had gender was like F, capital F for female, and then, or lowercase f female, or like FEM, and just all the variations. So it was a lot of manual cleanup to start, but I actually really appreciated that exercise because what that then allowed me to do was truly understand the power of data architecture and really the power of designing your data input processes and your data storage mechanisms. So, fast forward.

What do we do today? Step one, I do not skip the step of data architecturing when it comes to system implementation, when it comes to release management, when it comes to plugging in a new system. Sometimes when we go through system implementation, it sounds very easy to say, oh yeah, it’s a six week thing. Just plug it in, run a flat file, SFTP integration, and boom, there you go. But the problem is,

Unless you know the ROI that you’re trying to get out of that system and the broader picture of what you’re trying to achieve, plugging in a CRM on top of your ATS is easy. Getting that CRM to measure the funnel, measure the effectiveness of your programmatic advertising, measure if Indeed or JobCase or whatever the posting size that Procutor works better for you, that is a challenge. And unless you design that data in, unless you design that measurement step in a friend,

It’s really hard to build it in later in terms of retrofitting processes and kind of ripping a bandaid off people who just want to enjoy the fruits of their labor. So for me right now, step one is always understanding what am I trying to get to? What is my five to 10 year strategy? What do I need to convince the business to help me with my five to 10 year strategy? And therefore taking a step back, what do I need to capture and measure?

to deliver that message to the business accordingly. So that is the philosophy of the data architecture. The second part of it is then getting technology to really help control the data input. I think HR as an industry love opinions. We love to give people that open life space to say, open comments, other, tell us more. And I think it’s a phenomenal idea and it’s phenomenal value for the information we get. However, when you’re designing a system,

it becomes a little crazy when you let everyone freehand everything that you need to collect in the system. So what we then do is we basically gather all of the, what I call the 90%, the values that we expect 90 % of the time, turn them into multiple choice. So at least we know what our data catalog looks like. And then from there, the 10%, we give the other option. If needed, we’ll have somebody follow up on the other option. But most of the time, the 90 % catches everything.

Abhinav (17:46)
Lydia, when I talk with a lot of HR leaders, right, who are fascinated about data, who really want to be, not that they’re HR team, but the business leaders to be more data driven, they mainly speak about these two challenges. Okay, one is how do we go about building a business case to the leadership to invest more in people, data and insights? And second, and very like…

Quite obvious is where to start, what should be our first step, because even if the CEOs or the CXO approve, they say, okay, what’s going to be our first step? And a lot of times they don’t have clear idea. So a lot of these leaders must be in our audience. What would you advise them? One on building, how to build a business case, and second is where to start.

Lydia Wu (18:31)
Delete the fact that you’re an HR leader. Delete the fact that you’re looking at a people data. Let’s look at it from a home mortgage or a home loan perspective. When you go to a bank and say, I need money, give me money, what’s the first thing they ask you? Okay, well, what are you going to show for it?

And most of the time, it’s the evaluation of your house, the fact that your house is worth more than what they’re giving you. So push comes to shove, they can still make about 20 % in liquidating your home. And let’s hope nothing ever comes down to that. But it’s a very cut and dry mathematical equation. That equation doesn’t change in the corporate world. It doesn’t change just because we’re talking about people data. For some reason, a lot of leaders and a lot of practitioners I talk to, they think that people function and people data is different.

But at the end of the day, when you’re running a business, what is incredibly critical is the financial survival of that business. So you can pay everyone and make sure before you make sure that they’re happy working for you. Because if you can’t pay them, well, engagement isn’t necessarily top priority at that moment in time. So working from that logic backwards then, when you’re creating a business case, it’s not just about like, oh, we’re going to have a cost avoidance. We’re going to…

be able to make people happier, more satisfied, more engaged, all important, all incredibly valid. But at the end of the day, your CFO is gonna look you in the eye and say, what the heck am I getting out of it? Where is that extra penny for every dollar I put into this? And how are you gonna guarantee and prove to me that you’re gonna squeeze that extra penny out of that dollar? And it sounds incredibly crude to some. I’m sure some of our audiences are listening to me say this and going, oh my God, you cannot possibly.

But at the end of the day, when you’re trying to get money, when you’re trying to grow, that is the equation. And that is unfortunately the game rules that have been written and the game rules that as HR we have to play by. So how do I look at it? First of all, I never approached a question of how do we create a business case for people data? People data like technology. It’s a tool. It’s a mechanism. It’s not a be all end all. It’s not the end. So take a step back and figure out what is the business strategy and what are you trying to do with the HR function and people in general?

because then you have a case of, okay, let me tie people dollars to the business dollars. Once you figure that out, take another step back to say, okay, of that people dollar, let’s say I want half a million, but finance can only afford quarter million or 200 ,000. Then how do I efficiently squeeze out that $300 ,000? Because the answer honestly, 95 % of the time lies in technology, automation, data, intelligence, research, so on and so forth.

That is where that business case of HR analytics and data comes in. It’s not like, hey, leader, I need money for more data to build an HR specific data lake because most CEOs will look at you like you were a second head and tell you to go bugger off and go to IT and figure it out with what’s available today. It’s the angle of which you attack that conversation that I think builds the most successful business cases. And the angle should never be data necessity led. It should always wire itself back to the people problem and ultimately back to the business problem.

that the whole organization is trying to solve. I think related to that, one of the most critical questions any HR practitioner can ask their business first day on your job, if not during the interview process, is how do we make money? Because until you understand how your organization makes money, you always are going to feel like you’re running into a wall every time you’re trying to ask for funding and every time you’re trying to ask for money. And that’s really the balance of the equation. So that’s part one. Part two, where do you start?

I just had to get the conversation started. And in doing the incredibly painful dashboards and getting the conversation started, what I was able to do was generate a sense of curiosity in the organization. Because when somebody sees their turnover number, when somebody sees their demographic breakdown, the immediate next question was always like, OK, but how did it happen? How do you know? What do I do now?

The moment you get that hook, you can keep the conversation going. And once you keep the, once you grow from a one person team to HR having that whole conversation holistically together, and that was about a year and a half’s worth of journey, it’s a lot easier to then go to the business and say, hey, you know that data that you’ve been asking us for the last five years on? I actually have it. Let me show you what I’ve got. This is a month by month, incredibly painful process.

So can I get like $100 ,000 from you if you think that’s interesting so I can invest in something to make this a little less painful?

Abhinav (23:01)
Moving now to both of our favorite topic, which is AI. How did Panasonic leverage AI, and especially you there, leverage AI and machine learning to its people analytics and all the insights initiative?

Lydia Wu (23:18)
Yeah, absolutely. So right now I think we are still in the discovery and build phase of AI. So here’s how I look at the HR technology world. If you look at the last three to five years, I think the development of technology phases trends, especially in the world of HR has happened at a quicker pace than we’ve ever seen before in the industry. You start, initially it was like UX UI and then it was like, oh, HCM cloud. And then it was like, oh, employee experience. Those were slow. It was like,

two, three years apart, but in the last year alone, it was more so skill set than it was AI. And then it was sort of, how do you apply all of that to everything that you’re working on? Here’s a problem. I don’t think most of us out here are working on a solid technological foundation to be able to adapt to a quicker pace of technology evolution in our ecosystem.

We’ve been able to duct tape it. We’ve been able to sort of like wire together, hold it together with bubblegum. But my thinking has always been until you have a really solid technical foundation, you are always going to feel like you’re getting hit sideways with all of the innovation, with all of the technology. And your employees are never going to feel that you’re on top of it because you will always tell them, here’s why we cannot do something and not here is why we’re embracing something.

So I am actually in the midst of a HR system implementation right now because I have decided that we’re going to rip out and re -foundation and re -architecture and re -layer the sort of basement and foundation level of how we run HR from an infrastructure perspective. So we’re not really duct taping and trick and wiring everything. So we’re actually having the proper support beams and the concrete pours and things along those lines. So we’re not standing on stilts. And in doing that,

It’s really done with the future in mind. So how we architected the data, how we designed roles and responsibility, how we even designed a field of employee ID, which I’m happy to get to in a bit, was really the thinking of how are we going to now use all of this to propel us into the world of AI, into the future of HR technology, regardless of AI, regardless of the skill sets, regardless if it’s something else that’s going to hit us sideways in three months time at the current development cycle.

Abhinav (25:33)
which now brings us to this, I think, ever going debate of build versus buy. I’m curious to know, and I’m sure most of our audience would be curious to know, what side are you on?

Lydia Wu (25:48)
I am on the rent side of things, to be honest. So here’s my fundamental problem with build versus buy, and I’ll be very candid about it. Until somebody experiences the pain or joy of build, they’re never going to truly understand what build actually means. Because very similar to how solution consultants demo the most perfect version of a tech product,

Abhinav (25:51)
You

Lydia Wu (26:15)
When you are exploring the build phase internally, it’s always layered with assumptions of like, oh yeah, it’ll be easy because it’s easy because I’m assuming you only have five data fields. It’s easy because I’m assuming you only have so many historical data. It’s easy because I’m assuming there’s only one single organization. You’ve cleaned the data, you validated before you fed the data over to the data lake. So yeah, absolutely. We can build it for you because we are assuming humans are like widgets and things never change. That’s never the case in our world. Now, I also understand a

organizations with incredibly robust IT organizations and possibly HR leaders who just don’t want to touch data. That’s totally fine. It’s not for everyone who want to take that build approach. And I think definitely go for it. Try it out, but isolate yourself so that you’re not fully all in on it and have to peel back. Always build in sort of what I call the gate check periods in your build journey for you to say like, is this working? What are the indicators of it’s working? And if it’s not working, let’s just peel back and pivot another way.

So that’s my opinion on build. In terms of my opinion on buy, my God, they are expensive. I feel like I just share the sentiment of most HR buyers out there, right? Because if you look at a solution, it’s a beautiful solution, and somebody tells you like, oh, it’s $50 per employee per year, you pause. Because depending on the size of your organization, depending on the belief of your leaders that HR analytics is actually going to work, you pause because that’s a

hefty chunk of cash you’re about to shell out.

So here’s the reason why Lydia advocates for rent. Because in rent is what I call the pilot projects. It’s the easily accessible tool sets that you don’t need a lot of technological sophistication to be able to do.

It’s not a solution where you’re either all in or all out. It’s the dip your toe in the water, see how you feel. If you like it, let’s keep going. If you don’t like it, that’s fine. Let’s back out of it. Not enough people in the environment are doing this and not enough HR folks in the ecosystem are asking to say, hey, can I try it out before I buy it?

Abhinav (28:15)
Lydia, you have built so amazing systems, you have rented it, you have bought some of the systems at Panasonic as well. I’m personally very curious about what’s the most bad-ass thing that you have been…

able to see people analytics doing for you.

Lydia Wu (28:29)
It is genuinely being able to tell our line leaders in near real time how the people they have assigned to different parts of their production line are impacting the outputs of their production line. Because if you think about the sort of the office environment, it’s almost like everyday managerial training. I’m like, yeah, pay attention to your people. But if you think about the production line environment, especially when you’re running a giant facility that’s 24 seven around the clock,

Most line leaders think about output. They don’t think about the people and how that impacts output. Everything impacts output, but most importantly, you think about the output. And that’s fine, because that’s what we hired them to do. We need them to obsess themselves over the output so we can maximize our overall productivity. But the ability to tell them, hey, by the way, here’s how your people actually impacts your output. It’s not just your machine operating uptime, downtime, and maintenance time.

it’s the bodies that you’re assigning to those pieces of work as well. It sort of gave them the aha moment to say, oh, let me check in on how that person is doing. Let me be a little more human that if somebody needs to step out for 15 minutes and take a call or whatever, let’s do that.

Abhinav (29:42)
That’s so true. And Lydia, at the end of the day, when it comes to people analytics, the owner and the implementer is always going to be HR. You have been in this industry for 12 years now. Do you see that HR is becoming more and more data -driven with time

Lydia Wu (29:59)
It’s interesting. So I think the owners and implementers will always be HR, but I almost want to tell everyone, don’t ignore your IT department. It’s not HR or IT. It has to be HR with IT because whether you like it or not, your HR system, your data system, and everything else ultimately has to plug into the broader ecosystem. So find your best friend in IT, take them along for the ride because it’s going to serve you in the long run. That would be one part of it.

I think in terms of it has HR become more data driven. Yes, absolutely. HR has become more data driven. Is data literacy and data maturity still a challenge? Yes, absolutely. And this is where I will look at all of the educational institutions for future HR resources and ask them, what are you doing to teach people about data literacy, to teach the future generation of HR practitioners about data and the utilization of data?

Abhinav (30:50)
And the reason I asked this, and very rightly, you also said this is always this debate about balancing data -driven approaches to the human element of HR. I mean, you’ve obviously been to both sides. How do you advise HR to not just rely on one thing, but like use the power of both and balance it correctly?

Lydia Wu (31:09)
Yeah, absolutely. I think Data is actually what makes the human side things of HR more human. And the reason for that is a lot of times we talk about human side of things, we talk about the anecdotes, we talk about the qualitative things. But the problem with the anecdotes and the qualitative things is that you don’t always get the full picture. It’s the saying the squeakiest wheel always gets the oil essentially. And what data allows you to do is being able to look at the

qualitative side of HR, aka the human side of HR, with a lot more objectivity, with a much broader coverage, to be able to definitively and logically say, is this a problem in our organization? Do we need to act on it? How big of a priority is it for us? Because you’d be surprised at the number of organizations I talked to when I asked the question, like, so what did you implement last quarter? It’s like, well, what that one guy who put up his hand during our last quarterly said, that’s what we did.

Okay, well, is that representative of the whole few hundred that you have in the organization? Or is that just one person who was courageous enough to speak up and is actually the minority whom now you have forced to become the majority? So a lot of times I think in being people, people and being more people focused, HR is actually doing the organization and their employees a disservice without looking at the data side and without looking at the broader picture of what it is that they’re trying to do.

Abhinav (32:34)
That is so powerful, Lydia. I am certainly taking that note. And that brings us to the end of this talk. Lydia, thank you so much. So, so much for talking with me. I really enjoyed our conversation. And the work that you have done is inspiring for so many HR and business leaders. We definitely need more people like you in every company. I’ll just say keep up the great work, keep inspiring, and have a great day. Thank you so much again.

Lydia Wu (33:00)
Awesome. Thank you for having me.

Summary

In this episode of Peoplebox Analytics Talk, Abhinav interviews Krishna Raghavan, former Chief People Officer of Flipkart, about the intersection of data, technology, and people in HR. Krishna shares his unconventional journey from software engineer to HR leader and discusses the importance of data-driven decision-making in HR. He highlights the evolution of people analytics and the role it plays in solving business problems. Krishna provides real examples of how people analytics can be used to predict attrition, improve candidate experience, and drive employee engagement. He also addresses common myths about people analytics and offers advice for HR leaders looking to build a data-driven culture.

Key Takeaways

• Data-driven decision-making is crucial in HR and can lead to better business outcomes.
• People analytics should be a strategic consulting arm of the company, not just an HR function.
• Investing in data education and producing noticeable wins can help gain credibility for people analytics.
• Data democratization is important, allowing everyone in the company to access and use people analytics.
• Data privacy and access control are essential to ensure employee data is not used for surveillance.
• Starting small and focusing on concrete use cases can help build a business case for investing in people analytics.
• People analytics can enhance fairness, transparency, and accountability in the company.
• Overcoming data fragmentation and ensuring data comprehensiveness and completeness are ongoing challenges in people analytics.
• Resistance to data-driven HR can be overcome by demonstrating the efficacy and value of data in decision-making.
• Data analytics can make HR leaders more effective and help them drive better people outcomes.

Full Transcript

Abhinav (00:00)
When was the last time you met with a Chief People Officer of a $40 billion company only to find out that before taking on this coveted role, he has never spent a single day in HR. When was the last time you witnessed the ascent of a senior VP of engineering to the position of chief people officer

at the largest e -commerce company in Asia. Sounds intriguing? Buckle up. Hi, everyone. I’m Abhinav, co -founder of Peoplebox, and I’m super excited to kickstart the first episode of Peoplebox Analytics Talk, where we invite trailblazing leaders to delve into the fascinating intersection of data, technology, and people. And to make it even more special, I’m delighted to welcome our first guest, Krishna Raghavan. Krishna, in his last role, donned the hat of

Chief People Officer of Walmart owned Flipkart, And unlike most HR heads, his journey has been nothing but a barrier breaking one. Welcome to the show, Krishna.

Krishna Raghavan (00:54)
Thank you so much. It’s a pleasure to be here.

Abhinav (00:56)
Krishna, you started your career as a software engineer, worked in tech giants like Yahoo, Oracle, became CTO of ClearTrip, then joined Flipkart as the SVP engineering, and then became Chief People Officer, a path barely taken. One question that probably would be top of everyone’s mind in our audience is that when you were young, did you ever imagine that one day you would be at Peoplebox Analytics Talk talking to us?

Krishna Raghavan (01:21)
Definitely not Abhinav.

Abhinav (01:24)
but jokes apart, it’s truly our honor, Krishna, to have you here. So let’s dive right into this. Talk to our audience about how did you end up snagging the top HR spot at Asia’s largest startup after a super impressive engineering career.

Krishna Raghavan (01:38)
Yeah, the story is an interesting long one, but I’ll try to keep it short so that I don’t bore our listeners. But I think the journey started way back before I can even realize that it was happening to me, which is I always, I think gravitated towards problems that involved people, culture, teams and building them to scale. Right.

And even my prowess as an engineering leader was always within that space. I didn’t realize that as much until probably much later in my career. And that’s when I went through a sort of a life transformation where I did a program. You can call it midlife crisis or whatever, but I did a program and I sort of discovered where my superpowers lie and where I should spend most of my energies in the coming days, weeks and months. Right. And the answer was.

extremely evident to me in front of me which was go and try to take on a role where I could do talent, culture, building at scale not just within my function within engineering. So that’s when you know lot of things came in place and the opportunity opened up at Flipkart applied for it and little did I know in about a month’s time I got approved for the post and I got there and

and I couldn’t have imagined in my wildest of dreams that the COVID pandemic would follow in a couple of months, but rest is history. But that’s really been my journey to get on to this particular role.

Abhinav (03:09)
That is amazing.

getting into the position of Chief People Officer of the largest startup and then COVID hit, which probably nobody prepared anyone for. And so before we go to the COVID one, I’m really interested to know what was everyone’s reaction? You know, your peers, the HR team, you know, getting somebody as heading the people function who has never got a single paycheck.

which writes title HR.

Krishna Raghavan (03:38)
there are lots of people that tell you, are you freaking crazy? That’s the question I got many a times. Some of my peers in engineering in particular, they said that you’re actually taking a disastrous move and you shouldn’t be doing this. You should be staying in technology. That’s where I think most of your promise lies. And,

Some of my well -wishers were obviously backing me, but I think there were lots of naysayers along the way. And I would say that even within the HR team, there was a lot of, I would say, suspicion of what to expect an engineering guy coming into the post of heading HR.

Abhinav (04:16)
let’s talk about engineering, you know, Krishna, you are an engineer at heart. I started my career as a software engineer. there’s one thing that I’m very sure about that. They love data.

And our audience must be very curious to know How did you leverage the engineering mindset and that data -driven culture in making the right people decisions?

Krishna Raghavan (04:35)
No, it’s a very important question Abhinav. I think one of the first things that I sort of brought into the role and many times I’ve been asked this question. What are the things that you actually brought into the role and what are the things that you actually jettisoned by coming into the role because these are completely different roles. But one of the things that I brought in was this data orientation or the mindset of looking at data for everything, right? And I’d like to start with basics. I think as I entered the function,

The most important thing became the question became what to even measure Abhinav because you know, often times there’s a lot of data out there and companies pride themselves on putting together 40 metrics on a spreadsheet and everybody’s pouring over those metrics. But actually do you need to look at 40 metrics to make decisions? That’s the first question to ask. So it became my, my sort of initial focus became.

You know what, what are we here to do from a people function perspective? How is that in alignment with our overall business strategy? And then go to define the metrics and the metrics, which were important. We had to actually come up with them. In some cases, the instrumentation was also not even there for the metric. And you had to put the instrumentation together. In some cases, the metric was already there. And in the other cases, wherever there is noise, where there were extraneous metrics, we actually just kind of, you know,

remove them from all the dashboards. So that became my like initial focus as I came in.

Abhinav (06:06)
But defining the most important metrics or OKR is one thing, but then getting the data, especially for such a large company, how hard was that?

Krishna Raghavan (06:14)
Very hard. Like some of the metrics were obviously there, instrumented like I said earlier Abhinav. But in some cases, there was a fair bit of data fragmentation and I’m sure we’ll speak to it at some point later in our conversation. But you know, disparate systems, systems used for different use cases. But when you actually look at a metric, the metric is actually a blended metric. It’s an output metric of many input metrics. But these input metrics are actually present in different systems.

So how do you actually get them together? In some cases, the systems themselves were not even talking to one another. So it became plumbing, like what I call data plumbing, which is like, okay, you know what? I want to instrument this metric, but the heck I can’t even figure out how to do it. Let me actually now put my head around this problem itself. Let’s instrument first. Let’s build a data pipeline. That became the first order problem. And I started to solve some of the instrumentation problems.

Abhinav (07:08)
I love the phrase data plumbing. I’m probably going to use it in one of the pitches we use. But coming back to the whole uses of data or say the whole function of people analytics, you know, for most of the companies that I speak with, it starts with hiring a reporting or a people analyst, you know, who would help creating reports primarily for the leadership or the HRBP. Is people analytics more than just creating reports? And if it is, what all does it entail?

Krishna Raghavan (07:36)
Yeah, there’s actually a pretty good paper on this Abhinav and I would urge our listeners to look it up. There’s actually a Deloitte study on a people analytics maturity model. And, you know, there are stages of evolution of how they look at people analytics. Obviously, many of these consulting firms do this, you know, as their primary gig, right? But I think just to summarize, initially, when you build a people analytics function, it tends to be

a data provider function like you aptly described it, which is, you know what data is not my problem as an HR functionary. It’s the people analytics is problem. Whenever I want some data, I send a request. I get data back. Neither does people analytics as a function know why I requested to this data in the first place, but they become more data service providers. Right? That is like, I would say.

you know, point zero on the scale of zero to 10 in terms of people analytics maturity. And I won’t elaborate on the entire evolution path evolutionary path Abhinav, but at stage 10, it’s almost like people analytics is like a strategic consulting arm of the company, not the HR function. You know, like the CEO says, you know what? I need to figure out.

Where do I need to invest in terms of my best talent in the company and what are the skill sets I need to build in though in that talent. Now that’s a very fuzzy question when you ask that at a scale of an organization that could be Flipkart size. It’s almost like people analytics has to anticipate that problem and like a strategic consultant going in tell the CEO this is what I think this is the decision or set of decisions or recommendations I can actually give you.

And in some cases, even short circuit and say out of these set of recommendations, by the way, this is the one I would pick. And the CEO has, you know, the ultimate veto choice to make that decision. But it’s almost like moving from data provider to decisioning for the company, not just the HR function. That’s what I see the entire evolutionary curve to be for people analytics.

Abhinav (09:48)
That’s super interesting. And you mentioned about the report by Deloitte. I actually absolutely love that report and I highly recommend everybody to, you know, all HR leaders to go through that report. Actually it’s authored by a very good friend who is the partner at Deloitte named Nitin Razdan So it’s a fascinating report. Coming back to the usage of people analytics, you know, I think what was every report and talk about that, how useful it is, but

Krishna can you give some real examples of how you use people analytics to achieve some real business objectives?

Krishna Raghavan (10:23)
Yeah, absolutely. I think there’s the holy grail that most companies want to get to, which is the churn prediction model that we called it in Flipkart or the attrition prediction model. We built a model with all the data that we had. This was after I think at least year three of the people analytics journey. You know, we are, we’ve kind of moved along in terms of the maturity curve and we have the data instrumentation in place and all of that.

Abhinav (10:31)
Yes.

Krishna Raghavan (10:51)
It actually turned out to be from a precision perspective. It actually was pretty accurate. Okay. In, in terms of percentages, I think we were able to get to 80 -85 % precision. we were able to employ this in particular teams in the company. Right. and we were able to give this data not just to HRBPs, but actually the line managers, and empower them to actually have these.

conversations with some of their employees that could be on the high risk prediction. So that’s one very, very strong use case. Second Abhinav is I think a lot of companies out there pride themselves on being a top destination for talent. But do you measure candidate experience through the funnel of hiring? And in Flipkart, one of the things that we realized when I was working with the team is that

You know, candidates that got accepted their candidate NPS scores was very good, but the candidates who got dropped at some point in the process because there was probably not a fitment their NPS scores was very less. Now you could argue and say, you know what? I don’t care about the candidates that got dropped, but that does not define a great company because end of the day, your promoters are the ones who also interviewed with you and won’t say, you know what?

I didn’t get through, but I had a great experience through the process. Now the people analytics team was able to give me this data that led to say, if I look at rejected versus accepted, my NPS obviously differs. And these, this led to a series of interventions to improve the experience for rejected candidates as well. Right. That’s my second use case. And the third is we moved away from this annual survey business, which is

once a year, I’ll check employee voice and I will take a set of actions. We moved away and we said we are going to do continuous listening and we are going to actually have a mood score and we’re going to ask you how you feel at a particular day at multiple points in time through a particular week through a particular month. As this data matured, what we found out, this is probably probably common -sensical is that there is a very strong data correlation between mood score.

as the leading factor for attrition. So how do you take action early on from an employee life cycle perspective? Because often companies talk about employee retention. I actually kind of hate that phrase. It’s almost very negative. It’s like you want to go, but I’m somehow trying to hold you back. But is there a reason to stay in the first place? And can I actually engage with you when you are starting to disengage and you’re showing signals of disengagement?

So flip the problem on its head and I think this was one of the biggest shifts that people analytics actually helped us make in the company. So these are two, three use cases that I wanted to talk about.

Abhinav (13:44)
Krishna, the way I look at Flipkart and pardon me if I use the wrong phrase, but.

I look at Flipkart as like the Amitabh Bachchan of business world, you know, the pioneer in setting innovation that everybody look upto And the reason I use the word business and not startup is because even the larger publicly listed companies want to learn from Flipkart. And I believe that the use of data and people analytics must not be an exception here. So give our audience some important learnings from Flipkart, people analytics culture that they can today go and leverage in their business.

Krishna Raghavan (13:52)
Ha!

Yeah, I think it’s a very important question and I wouldn’t say, you know, it’s just people analytics. I think data as a common theme across the company, the data oriented mindset is extremely deep across Flipkart. I think that is something which goes across not just HR, but every function out there. I would probably put across certain big learnings that we’ve had through the, through the journey, right? And in particular, people analytics.

The moment people analytics stays within the domain of HR, you’ve lost the plot. It is not just a HR function or a department that you need to set up in HR. The way you need to think about it is in the business realm, you often have an analytics organization, right? And this analytic organization typically is a horizontal that goes across the company. Initially,

is a data provider then becomes insight provider then starts to actually even recommend and make decisions on behalf of the top management in the company for everything that’s people anybody in the company should be able to access people analytics. So if it remains within the domain of HR, you have lost more than 80 % of the vision of where people analytics can get to. That’s probably the biggest takeaway.

The second I would say is invest a lot of time in data education and I cannot overemphasize this enough even within HR at least I find that the level of data proficiency is probably not where it should be because in today’s day and age there is an explosion of data.

and even within the people realm there will be an explosion of data. The skill actually lies in asking the right questions, connecting the business problem to what we are solving for and asking those pertinent relevant questions and then using the power of data to reveal answers out to you. So if you can’t ask the right questions, you will be barking down the wrong tree many a times. So

Data education is my second biggest takeaway and this has to happen across the board including within HR. Right? And third is produce some very noticeable wins in the company to gain credibility of the function. It should not be something like it’s a pipe dream in year two, year three, by the way, this is the roadmap we have on people analytics and this is what we’re going to deliver. It can’t be that.

Most companies in today’s day and age need answers yesterday, not today. And how do you actually have that business acumen, that urgency and agility in your operation to be able to land some very strong, successful outcomes early on will really establish the credibility of what people analytics can deliver for the company. These are three big takeaways for me.

Abhinav (17:22)
Krishna, I’m so happy that you spoke about the first one, which is data democratization. Because whenever I speak with HR One of the top wishlist for them is the, the ability to provide everybody in the company, even the employees, you know, the power of data. However, the major roadblock

is fragmented employees data in different tools and sheets. And like you also mentioned, right? Some data is in ATS and others on HRIS. ESOP data is sitting on another tool, performance and engagement somewhere else. And there is then tons of data on spreadsheets. How did you overcome that for a company of the size of Flipkart, which has tens of thousands of employees and I’m sure no dearth of tools and sheets.

Krishna Raghavan (18:03)
Yeah, I think if you ask anybody out there on a joking note, Everybody says that this software is the best for this use case and nobody obviously wants to adopt one common platform for all use cases. So what you land up doing is obviously buying lots of different products and services leading to fragmentation and everybody proposes promises the moon when you buy them. But, after that, you realize the actual truth, right? The harshness of.

data fragmentation. So I would say that this was a journey for us, Abhinav, and it’s actually still ongoing. I would say to that extent until very recently, you know. So the way we actually did it is, like I said earlier, we spent a lot of time analyzing what data we want, where does this data reside, and spent time in actually putting together the entire instrumentation pipeline for it.

So we had to build data pipelines across all our systems and all of them flowing into one common data warehouse. Then once we build, you know, the people domain model, the people domain model as in how should we represent an employee, right? The entire data model for us, if you think about it, there are multiple relationships between an employee, a manager, an employee and a skip manager, right?

How do you define the persona of a director and so forth? You define the entire people domain model and the instrumentation pipeline for it. And then what you do is you actually land up building adequate visualization for it because the power of data cannot be revealed, so to speak, or cannot be shown in its all glory unless you have great visualization. So in our case, we also had to pick a platform or a product to visualize.

And then the education thereafter followed. So that’s the journey we’ve kind of been on Abhinav to make sure it’s not been easy at all for sure.

Abhinav (20:01)
I can absolutely imagine it’s not easy. It must not be easy because one thing that we didn’t talk about, is the data sanity. a lot of time data is not in a consumable format. I was talking to somebody and say, you might wonder that it’s so easy to find out the last CTC of an employee, like from a previous company.

And you’d be surprised it’s not because it’s sitting, it’s sitting in notes. Uh, and those are like tens of hundreds of notes. So was that a problem that you also encountered about, you know, cleaning up the data and make it in a consumable or probably a quantifiable way. And then of course, you know, put it into your data pipeline.

Krishna Raghavan (20:41)
Absolutely Abhinav, I would say the two big things that you always think about when you deal with data is you think about data comprehensiveness. Do you have all the data in the first place? Then the degree of data completeness. They are very different by the way, like comprehensive means do you have all the data? Completeness is more the aspect of accuracy. So it’s not enough by the way that even if you…

Take your example, even if you got the CTC of the last employment into a system where you can consume this data, unless you actually refresh this data for future joiners that come in, the data becomes incomplete. So you also need to ensure completeness, not just comprehensiveness, right? So it’s a big problem. And in some cases, frankly, the data is so offline.

it takes a lot of effort to just bring it online. Many of the teams don’t even record this data.

Abhinav (21:38)
Now, Krishna, a bit of a controversial question. You are this round peg in a square hole, bringing this engineering and data driven mindset. Did you get any resistance? Generally, not a lot of companies rely on data when it comes to making people or HR initiatives. How was that going through the journey and was there any resistance?

Krishna Raghavan (22:04)
No, no, I think I would be lying if I said there’s no resistance. There was definitely resistance. Different forms of resistance, right? Sometimes you face resistance when you make a large change, passively or actively, correct? So the active pieces in places like learning and development where you need to define efficacy of your interventions. Sometimes learning and development will say, you know what, we have great participation rates, we have good satisfaction scores. Isn’t that enough?

Why do we need to measure quantifiable business outcomes of our learning intervention? That’s more like active resistance because the question is why? Why do we even need to do that? The passive resistance comes into places where people are not data aware enough and they believe in somebody else’s job to collect the data and take those decisions or they say these decisions are very intuitive. We actually take them based on intuitive thinking. Right? Why do we need to actually bring data into the equation for everything?

So there you face some degree of passive resistance as well. And what you need to keep constantly doing Abhinav is obviously one as a leader, you role model to you actually make sure that you keep communicating the efficacy of data and how it could lead to better decisioning across the HR function and the company itself. So that’s why I said earlier, producing some wins early is going to be important because talking in theory,

is one thing but actually in practice producing some wins and real examples is much more powerful.

Abhinav (23:36)
I love what you said that you need to be the role model to use anything, or I think to drive any change. But Krishna, tell me honestly, you’ve been an engineer, you’ve been an HR head, and obviously in both of the roles you’ve used data extensively. Do you genuinely believe that the usage of data makes someone a better HR leader or a better HR business partner?

Krishna Raghavan (23:59)
I mean, there’s, you know, it’s an emphatic. Yes, I’ve enough. It definitely makes you there’s no doubt in my mind about this. I mean, everybody would obviously say the answer. Yes to this one. But the degree to which it really helps you become a better leader. I think across the board and actually why only focus the point on HR, HR obviously yes. But the way you lead teams in companies today.

Gone are the days where you can just be very intuitive only as a leader and say, you know what, I think this person’s good. This person’s probably not scaling up enough. You have to now move to an era where you can use data to actually really power your decisions. And let me actually talk a little bit about just one small example. I think what we saw as the biggest transformational change in Flipkart.

is when we started to actually bring the power of data to business leaders around people, we defined what we call is a people dashboard. And we said to a business leader, you know what, you look at business, you look at top line, you look at bottom line, you look at all of this. What if I gave you a people dashboard in conjunction with your business dashboard?

And you look at it also to make decisions for your function. How would that look to you? And tomorrow, you know what? Both the CEO and me are going to hold you accountable. Like we hold you accountable for business outcomes, we are going to hold you accountable to those people outcomes as well. So if your attrition spikes, if your diversity doesn’t get to the target that we want you to get to, we are going to hold you accountable. It changed the game It helped leaders become.

better people leaders. At the same time, it drove the people agenda is not just an HR agenda, but it now becomes a company enterprise agenda. And that’s a very, very powerful shift. And this could not have happened without people analytics.

Abhinav (26:07)
Coming to the more challenges, Krishna, and I know you must, of course, be much more aware about them than I do, is the sad reality of HR world that they don’t get high budgets like a tech or sales or marketing would get. So how would you suggest to HR leaders building a business case for their CEOs or business leaders to invest more in people data or people analytics?

Krishna Raghavan (26:33)
Yeah, I think it’s a very, very, very important question Abhinav. I think that’s where probably most of the companies do not have adequate resources to invest in this particular area. I would actually start with focusing on two or three very concrete use cases for the business. Like it’s almost like when we think about building products for a set of consumers Abhinav.

We always think about product market fit, right? I would actually kind of think about it internally as a CHRO or an HR functionary in similar vein, which is we know it’s important, but how do we actually bring these stakeholders onto the table? The CEO to sponsor the investment in this area. Let’s take a big hairy problem that’s facing the company right now. And actually evidence.

how it can actually be solved elegantly through people analytics leading to direct better business outcomes. It could be around staffing. It could be around talent development. It could be around attrition, right? And you could actually quantify a before and after as well saying that this is what I implemented in a particular function. And so therefore you could have a control group established as well. So I would take this approach.

established two, three concrete use cases. And that I think will be a lot more powerful to sell to your most important stakeholders within the company.

Abhinav (28:04)
Do you see HR and people function in general

is now becoming more and more data driven and did the whole pandemic and remote and gig worker had anything to do with that? Was there a trigger? Was that a driver or is it still the same how it was 10 years ago?

Krishna Raghavan (28:20)
No, certainly not the same. Abhinav I think it is definitely improving as we speak. I think people have gotten a lot more data aware HR functions across the board. I think you have to realize that earlier data was more seen as, you know what I need to get data for a particular use case more as a service provider as I described it earlier. now looking at

How can data power decisioning is something which is dawning upon, I think, HR functions across the board. But they are grappling with data fragmentation as a real problem. So the awareness, the intent is there But now the problem is when rubber hits the road, how do I actually get there?

Abhinav (29:07)
Yeah, absolutely. So I want to now move to our last sort of a section or a round, which is about breaking the myths and as you speak to a lot of, you know, business people, HR people, HRBPs about the people analytics I see a lot of myths and I want to make quick answers from you to our audience about, you know, how do you, how do you react to these myths? So one of the first one that we hear very often is about you need to be

have deep data analytics expertise to use people analytics in the company.

Krishna Raghavan (29:40)
The answer is absolutely not. It all depends on the power of the tooling and the product that you actually employ to solve this. And products have evolved to such an extent where it’s about just a set of clicks. And like I said earlier in my conversation, it’s about asking those right questions. You can get the data that you actually want on your fingertips. So you don’t need to be a data scientist, a data analyst to actually use data.

Abhinav (30:06)
And the second one we hear is that People analytics reduces people just to their data and take the human element out of it.

Krishna Raghavan (30:14)
Actually, it’s more the opposite, right? Which is, I think of it as sometimes, particularly in the people realm, we use our own hidden unconscious biases to drive people decisions. And actually data bust those biases, right? Like, you know, one of the most common things is,

people who are not seen in the pandemic, they are probably not doing as much work and they don’t deserve to get promoted. You know, this could be a huge bias that could actually play out both consciously and unconsciously. But if data was there to our rescue, actually it would even make the company a fairer place to work in where these biases actually do not rule. So if you think about it, data can actually be your

I would say your biggest lieutenant, so to speak, as a leader, your biggest supporter to ensure fairness and transparency in a company.

Abhinav (31:18)
Very well said, Krishna. Another one. People analytics facilitate employee surveillance.

Krishna Raghavan (31:23)
Absolutely not. I think as long as you have standard, very, very strong data privacy rules around employee data, and you have very strong access control, determining who actually can view pieces of data, I think you’re in very, very safe hands. It’s not a surveillance tool at all.

Abhinav (31:44)
I couldn’t agree more. And the last one is to start people analytics, data must be perfect.

Krishna Raghavan (31:49)
Not at all. I think it’s a journey. The data completeness, comprehensiveness journey that I talked about is a journey. You don’t need to be perfect on day one. Start somewhere, start small, establish those wins and continue on your journey. Because what will propel you on your journey is progress, not stagnation.

Abhinav (32:08)
This is really, really helpful, Krishna. Krishna, you have moved on from Flipkart now. 31st December was your last date. And our audience will be very curious to know that after such a path -breaking career, what’s next for Krishna?

Krishna Raghavan (32:23)
Yeah, I’ve been still, you know, sort of thinking through, you know, and scouting for opportunities that I think will really interest me. My heart has always been in the realm of startups and I want to see how I can contribute to one or many. And that’s sort of where I’m sort of focusing my energies on in the coming days, weeks and months. I’ll definitely keep everyone posted, you know, what I’m up to next.

Abhinav (32:52)
I can’t thank you enough, Krishna, for this super insightful session. I enjoyed every bit of it, learned a lot, and I’m sure our audience will have so many insights and so many learnings to take from this talk. Just thank you so much and wish you all the very, very best for the next step in your career.

Krishna Raghavan (33:10)
Thank you so much Abhinav. It was a pleasure talking to you. Very, very insightful questions and thank you for a wonderful conversation.

https://podcasters.spotify.com/pod/show/peoplebox/episodes/Episode-1–The-Impact-of-AI-in-Talent-Acquisition-and-Management-e2n9lg8/a-abgqak1

Summary

Recruitment today is broken. Businesses are grappling with challenges like misalignment between hiring and business goals, overburdened recruiters, and poor candidate communication.

These inefficiencies are costly. According to HBR, the wrong hire can cost a company 5-7 times the employee’s annual salary when considering hiring, training, and lost productivity. Additionally, a study by Glassdoor found that a single job opening costs companies $4,129 on average, with costs increasing the longer a position remains unfilled.

In the first episode of The Peoplebox AI Talk, Abhinav Chugh, CEO and Co-founder at Peoplebox, sits down with Suzanne Salzberg, a veteran talent leader with over three decades of experience at leading tech companies like Highspot and TextNow.

Suzanne discusses how AI can streamline processes, from creating unbiased job descriptions to improving candidate communication and making data-driven hiring decisions.

Full Transcript

Abhinav (00:00)
If there’s one thing everyone is speaking about today, it’s AI. It’s going to disrupt every aspect of our professional lives. And some may say it has already started. So what will be its impact on the future of work, especially how we hire, develop, and retain our most valuable asset, People

and which part of the employee life cycle from hiring to retirement is going to see a biggest change What does it mean for the companies and HR? Hi everyone. I’m Abhinav, and welcome to the new season of our podcast, Peoplebox AI Talk, where we invite incredible leaders to go deep into the fascinating intersection of talent and AI. Today, I’m delighted to have Suzanne Salzberg on the show

Suzanne comes with over three decades of experience in the talent space. She has been the head of talent for big tech companies like Highspot and TextNow, and she’s now also a visiting lecturer at the University of

Welcome to the show, Suzanne.

Suzanne Salzberg (01:00)
Thank you, I’m happy to be here.

Abhinav (01:02)
Suzanne, you have been a talent leader for some of the most high tech companies. Let’s start with speaking about the disruption of AI in the talent space. Where do you see its biggest impact in the employee life cycle?

Suzanne Salzberg (01:16)
Well, I can, I can focus on the candidate’s life cycle in the talent acquisition you know context. I really think the, the biggest impact is going to become, honestly from the minute they read the job description because I think I see AI really helping to create complete, unbiased, easy to understand job descriptions. and it’ll also help hiring managers because one of the biggest pain points for talent teams is waiting for that job description.

The next big impact would be how fast a candidate hears back in the process, whether it means they applied and they get a response or if they’re in first, second, third, fourth interviews getting a response because I talk to a lot of candidates every day. And one of the biggest grievances from candidates is that they’re being ghosted.

And then I think the other major impact will be is how recruiting softwares are actually using AI on their backend and taking some of that load off of talent teams.

Abhinav (02:20)
That’s actually so well said. You’ve been in this industry for over three decades. Where do you believe that recruitment is most broken today and craving for AI to fix?

Suzanne Salzberg (02:24)
Right. Well, I feel like it’s changed, but I feel like right now where it is most broken is because recruiting teams have become so much leaner. The hiring is slowed. So full teams got laid off or there’s one recruiter left. And so those small teams have so, you know time management is a huge problem because now they’re doing four jobs. They’re the recruiting coordinator. They’re the travel coordinator. They’re looking at resumes. They’re doing the interviews. And so it’s almost impossible for them to go through a thousand resumes that they’re getting in a day because they’re also doing, I was talking to a previous recruiter that I worked with and she said that she was, you know, a CMO was coming in, took her eight hours to schedule their onsite, you know, with everybody’s schedules and I’m , and so that’s eight hours that she can’t be recruiting.

And so because AI can help with a lot of those administrative tasks, it’s great at doing travel planning, scheduling, you know, screening those initial resumes, like we said before it can really help with JDs. And so I think that that is a huge place where AI can help. I think the other place that it’s broken, is when you’re hiring for technical roles and engineers and they have to take a coding test, right? And that coding test is only as good as the person writing it. And so many times from a talent perspectiveyou know, someone will get a question wrong and it’s not even something that they’re measuring for that this person is going to do. And so I think if AI can write some amazing coding tests or tests that engineers or QA employees can do, like that will be a huge piece that is a big pain point for recruiting.

Abhinav (04:19)
Do you think that recruiters will be replaced because of AI?

Suzanne Salzberg (04:22)
Well, I think if you talk to anybody in talent acquisition, they’ll laugh thinking like, good luck. There’s no, because I mean, honestly, so many times recruitment and talent acquisition is siloed over here. And we just live in our own world and nobody really knows what we do, right? And so like they don’t, and it’s probably bad on us for saying all the work that we do and what we do, but I just don’t ever see, and anybody that does it, knows that they should not be replaced, right? But there is a fear because the people that don’t know really what we do say, oh we can replace the talent team with AI, right? So, you know, it’s that.

Abhinav (05:03)
So you really have had thousands of resumes in a day?

Suzanne Salzberg (05:07)
Oh, 100%. If you just go on LinkedIn and look at some jobs, it literally will say, a thousand people have applied to this job, for one job, right? So if you think that every job pretty much gets 500 to a thousand resumes, so it’s, and if you know, you have 20 recs open, 20 job openings, yeah, you’re getting a thousand a day for sure.

Abhinav (05:16)
Oh my God, and how do, how do you manage that? I don’t think it’s humanly possible to go through them one by one. And you know, it’s destined to have a

Suzanne Salzberg (05:38)
What I do is I help people navigate the job market and the hiring system. And, and so, and when the first thing I help them with is their resume. And so I teach them that a recruiter probably is going to spend… 10 to 20 seconds on your resume. And that is no lie. So that first half of your resume better be good. because we, there’s you know, fortunately The applicant tracking systems have this thing, it’s called quick review. And you can go into quick review and you’re literally just tapping. And, and that’s why people you know are frustrated that the recruiters aren’t getting back to them. But I said, if you knew what a recruiter was doing right now, all of these things, like that’s, like don’t blame the recruiter. Like it’s the situation that’s happening.

Suzanne Salzberg (06:26)
That’s, it’s real. It is a real thing.

Abhinav (06:29)
I had a very interesting talk, I was at, I was at a conference where there are a lot of job boards and they’re talking about what are the things that they are doing to attract candidates. And they say, we are giving them the ability to not only create their you know resume based on AI, we are giving them the capability to actually apply for say 35 different jobs with 35 different resume, all altered on the basis of the job description.

And we were just laughing that it’s not that AI is going to fight humans or, you know candidates are going to fight AI. It’s AI fighting AI. You know AI from the recruiter side is actually fighting AI on the candidate. And good luck to the Boolean searches and the keyword searches because now everyone’s resume will be built based on the job description. How do you see that world? What are the skills that both the recruiters as well as the candidate will now need to do a better matching.

Suzanne Salzberg (07:34)
I always teach candidates that the job description is the final before the exam.

So I will look at their resume and then I’ll say, look at this job description. And when a applicant tracking systems or LinkedIn or any of those companies are matching you to the Boolean searches, right? they’re literally taking keywords from that job description that recruiter put in. And if that’s not on your resume, guess what? You’re not gonna be one of the top 30 people that show. And so

You can do it without AI. I mean, it’s harder, but like I encourage people to have a couple of different resumes. I mean, sure, if you want to make them, you can tell an AI resume. It’s still at the point where it’s a little bit of a turnoff because like, is it real? Have you searched like what this company is looking for? So there’s a trust factor. There’s a trust factor because AI can make this beautiful resume that literally matches the job description. Then you have to question, do I trust this person? Are these, is this data real?

We’re at the point now where we can tell. We can tell if it’s an AI resume. And I’ve been to, when I’ve been to conferences and you have people that, you know AI still misspells things. You can always tell there’s certain little things. It’s like when you get like a scam email, there’s certain little things you can tell that make it a scam. And so I would encourage people still to do it the old fashioned way, but use that job description as your final, like, and you know, focus on what you’ve done to match the job description.

Abhinav (09:07)
And, you know, when I’m speaking with a lot of, you know, talent heads or, you know, experts in the talent acquisition space, one of the big fear that I hear is about the AI bias.

What do you think both the companies as well as the you know AI tools can do to mitigate this bias in the AI -driven, you know world.

Suzanne Salzberg (09:28)
Yeah, I totally agree that bias can cause AI to make decisions that are, you know, systematically unfair to particular groups of people. it can discriminate based on race, biological sex, national, you know everything.

And so because humans are choosing the data that the algorithms use, and even if like these humans are making a conscious effort to eliminate bias, it can still be baked into the data that they select, right? And so you can do extensive testing and diverse teams can act as effective, you know, safeguards. But even with these measures in place, I mean, you’ve heard the old saying garbage in, garbage out, right?

Suzanne Salzberg (10:11)
And so bias can still enter that machine learning process and AI systems can then automate and perpetuate bias models going forward and then you’re in big trouble, right? And so one of the ways that you can help to mitigate the problem, and I think this is where companies really need to focus on, is that businesses should look to engage their data science teams, all the other functions, like very early on in their organization as early as possible. And so then they can assure that the models accurately reflecting the decision -making process and that you know, the data is just weighed accurately. But engaging those people after the fact is not a smart decision. So just getting everybody in early and preventing that bias early is what’s gonna help because if we don’t then it’s just gonna keep perpetuating that bad bias going forward

Abhinav (11:04)
Yeah, and just about those checks, you know, there are a lot of laws coming up about using AI in recruitment. There are recent law in Europe. New York recently passed a law to build more checks when it comes to recruitment. What are your thoughts? Do you think these laws and compliances, will they help the technology or will they become a blocker in the technological advancement?

Suzanne Salzberg (11:28)
It’s probably a mix of both. I mean, it’s going to be a learning process. think it’s good that there, there are laws. I think there definitely needs to be regulation in AI. And it’s, it’s going to be a lot of trial and error, honestly. I mean, things are going to happen. Things are going to break. And I hope they just fail fast and fix them fast. But I definitely think there should be some regulations. just like when GDPR became a huge thing.

Other countries besides ours were like, you know you have to delete our data within three months and you have to put on there, like, do you want us to delete your data? So yes, it was a blocker, but it was also a good thing, right? And so, anytime something new is, that’s what happens.

Abhinav (12:11)
Absolutely. And I want to talk also about the data. You know data has been one of the most important things when it comes to you know making recruitment more effective, making it better. You know the more the enrichment you can do of the candidate profile, the more you know diverse data you can do that. You know in this world of AI, How, What role do you play that data will take in making better talent decisions? And how can AI further put a fuel to that.

Suzanne Salzberg (12:39)
Data plays a huge role in making, you know, better talent decisions. I mean, I talked to you about, you know, when CEOs want data, right? So where does this data come from? It used to be like, just my information, right? And so data can play a huge role in making better hiring decisions. We talked about the analytics in the process, huge, right? So just to give you an example, if you build a great dashboard, you can tell in every part of the process how long it’s taking how many diverse underrepresented groups are in the pipeline.

It’s a lot of pipeline analytics that are really important because if you want to do great diversity hiring, then it has to be intentional. And it starts at the beginning of the pipeline. It doesn’t just happen at the end, right? And so let’s say you have a hiring manager that you notice underrepresented groups are just never getting through.

Like, you know, then that’s when you talk with HR and you say, hey, this hiring manager’s never letting through a person from an underrepresented group, right? And so those, it helps you to make better decisions. It also really helps with, one of the biggest mistakes I think people are making is they’re thinking, oh, we need 30 QA engineers.

But AI can help you build these models based on your revenue goals and all these things. And then you can reverse engineer based on these models, how many people should we actually be hiring? And, and what skills should they actually be? Here’s our problem of company X. What skills should the people we hire actually possess.

Suzanne Salzberg (14:22)
The other piece that’s huge is keeping up with compensation. So companies like Radford, I don’t know if you’re familiar with Radford, but companies like Radford that most people use to like look at what we should be paying people, you know, they have 35 ,000 companies that feed input into their, their software.

They do it once every six months, right? And so like say in 2022, what was happening was because salaries were going so high, so fast in the US companies started going outside of the US to hire, right? Canada, South America, that was happening a lot. And the salaries were half, which isn’t great, but, but companies were like, let’s go to Canada cuz you know, but so what was happening, the salaries in Canada and other countries were going up so fast

Because every company got this great idea. And I would come back and say, hey, the senior engineer is now making this much. And they’re like, what? Like there’s no way, right? There’s people just didn’t believe you. And so if AI could, could, you know, take that data and analyze it and quickly and give me like actual data that I can say, no, look here, that would be amazing. And then because you wouldn’t lose candidates, you know, based on the fan companies were just offering 50 grand more per person just to get them, you know? So the smaller companies were like, what do we do? like you know, so, and then the employees that had brought on, before that from those countries were now like making way less. It was, it was crazy. I’ve never seen anything like it in all my years. So, so I think the data on compensation can really help create better models.

Abhinav (16:06)
I think it’s so interesting, they said that how AI or data can actually help align companies business objectives to their hiring goals. And this is one of the big problem that we see that when majority of the companies, this whole talent acquisition and talent managements are staying silos and completely disconnected. You know, I would love to hear your thoughts on how can companies better align both the talent acquisition and talent management to just build a more cohesive ecosystem.

Suzanne Salzberg (16:35)
This is a big problem that’s happening right now in the talent ecosystem, and it’s looming very large right now. And because of the crazy market that’s been, right now employees and candidates are feeling that, you know, going back three to five years, that was, we’re people first. We care about people. We’re people before profit

Like you heard all these things of like people are a number one goal and which is great but now it’s like it’s a bottom line and all we care about and which is you know as a CEO it’s important right? It’s important that you consider the bottom line, but now so what’s happening is Employees are not feeling like companies are loyal to them so guess what we’re not gonna be loyal to you so when Suzanne reaches out as a recruiter I’m getting people way easier to jump ship from a company because they think well we could have a layoff next month because we’ve already had three so I’m just gonna leave right.

The other thing that’s happening is candidates are seeing this or they’re just been burned because they’ve been laid off three times in the last two years so they’re interviewing, they’re getting job offers and then they’re continuing to interview.

And so I’ve had in the past probably six months where candidates literally called the day before and said, oh, I decided to take another job because they gave me 20 ,000 more dollars. so nobody’s loyal. So everybody’s pretty much just like, sorry, not sorry. You know like you haven’t been, you know, helping me at all.

Companies need to be, especially HR, needs to do a better job in hiring better leaders that are understanding hiring the right people and budget forecasting. So we’re trying to mitigate all of this and we just really need to have better succession planning and it’s just instead of everything always being on fire and oh the first, the first solution is to do a layoff, right? And so companies have to really get back to valuing the people, not just staying there, but actually valuing the people. I feel like that’s, that’s really bad right now.

Abhinav (18:43)
We have spoken about hiring Suzzane, but what about post -hiring? You know do you think that AI will make a difference in providing overall employee experience as well, especially say from the day one right from their onboarding

Suzanne Salzberg (18:58)
Yeah, yes, for sure. think it will. I also think that it really helps people. I mean I talk to people that companies every day that’s like, Oh I had to make this PowerPoint presentation and AI helped me make these beautiful slides like that used to take someone a long time to do. Right. So it helps again now that companies are leaner. It’s helping take a lot of those administrative tasks, even off of onboarding and all of those things.

As in recruiting, I don’t think AI should replace onboarding. I think you should have real people doing people’s onboarding because again, it’s the first impression from these companies. I think, this is a really interesting aspect that I think can help a ton, is

If you ask any employee at a tech company, especially right now, is they’re always having to fill out these engagement surveys? How are we doing? What can we do better? Like on and people are very honest because a lot of times they’re anonymous, right? So what happens is the company fills it out, HR is saying, oh make sure you get your engagement survey filled out. And then they have an all hands meeting and they report the findings and people can do you know anonymous questions.

And the biggest pet peeve of employees is they take the time fill these out, they give you the feedback, and then they never hear anything about what’s going to be done to fix what we just told you was wrong. And so I feel like, because it takes HR teams a long time to compile this data, look at where the problems are, and I think AI can do that really fast, and then also maybe recommend some solutions to helping, right? And so then HR is going to be the one that executes them, but they can get that stuff done faster because HR so many things on their plate, again, it just takes that administrative stuff off. So I think it helps everybody, everybody’s employee experience.

Abhinav (20:53)
Absolutely, being an employee engagement platform. I can 100 % you know relate to what you’re saying. It’s not about the effort to take a survey. It is not even about an effort to go and collect the feedback. It’s about what to do with that. And like they say that you know a feedback taken and not done anything is actually worse than no feedback taken, 100%

Because HR has so many different things to do and so many fire to douse, if you just tell them that this is what your number one thing to increase your ENPS or your retention, they would absolutely grab that. And I think that’s a great opportunity for AI. And my last question, you know, we’re speaking about it today, but I want, you know, I would love to hear your thoughts on that one thing, that biggest impact that you believe that AI will make in the talent life cycle in the next 10 years from now

Suzanne Salzberg (21:27)
For sure. You’re gonna see a lot of trial and error which I said well I think you’re gonna see a lot of trial and error which works, but I think the analytics and seeing where the process is working or not working. I think it’s really gonna help make data driven decisions as well as taking the time off the recruiting team on parsing those resumes that we get so many resumes I would say a lot of it depends on the industry for an example

I have a friend who’s a radiologist, and he literally said to me, there will not be radiologists. AI is going to completely take over, just the radiologists that are reading the scans, not the ones that are doing you know surgery. He said they can do it faster, they don’t make mistakes, and, and people can get that information way faster, right? So like you’re sitting there, you just got a scan, and you have to wait for the results. Painful, right? You can literally get it in 10 seconds.

Suzanne Salzberg (22:44)
And so he said, he literally told me, goes, my job is not going to exist. So a lot of it depends on the industry, right? So that is one job that yes, AI can replace that job.

Abhinav (22:44)
Wow.

Suzanne Salzberg (22:56)
Right? And so, yeah. And I think that what’s happening right now is we’re between generations. Right? And so even Gen Z, like they’ve had human interaction and AI right now, but like say a Gen A, like after Gen Z, they’re fine with no human interaction. Right? They’re like, I don’t want to talk to people because they can’t have a conversation. Right? I don’t want to talk to people. I’m fine if an avatar does my interview. So I’m not saying 10 years from now what’s gonna happen

Abhinav (22:56)
my god.

Suzanne Salzberg (23:27)
But I think right now there’s too many people that were in this flux between the generations where that’s why it’s not gonna happen but hey, I am not naive enough to say that like if you have future generations that have no problem never talking to a human it might not, you know replace it. That’s my biggest fear I would be very sad if that happened, but you know, some of it’s probably inevitable

Abhinav (23:49)
I want to say one thing to your radiologist friend that I am talking to, everybody. Like, I’m hearing this from a programmer. I’m hearing from an SDR. I’m hearing from a copywriter, from a marketer that our job will be replaced by AI. And to be honest with you, I honestly, I don’t think so. I think we humans have an incredible capability to adapt and learn and build new skills that an AI can’t do. So I’m

I’m sure there’s both of us to see that what the next 10 years we do, but I’m very optimistic that it’s going to be something good and not very negative. Well Suzanne, thank you so much for taking the time to speak with me. I just loved our conversation and this industry is moving with such a high speed. And I’m sure when we speak again, we’ll have a different you know AI landscape and new challenges.

Suzanne Salzberg (24:24)
Exactly.

Abhinav (24:42)
I’m sure we both and you know all our audience will be fully prepared for that. Have a wonderful day and thank you so much for your time.

Suzanne Salzberg (24:49)
Thank you. I enjoyed it. Have a great day. Bye.

Abhinav (24:52)
Thank you.