How to Use Predictive Analytics for Employee Retention
In this episode of the HR Leaders Podcast, we sit down with Neeraj Tandon, Head of People Insights and Analytics at Syngenta Group, to discuss how predictive analytics, cultural transformation, and employee insights are reshaping HR strategies.
Neeraj shares impactful stories and practical approaches from his experience at Novartis and Philips, highlighting how analytics can dramatically improve employee retention, business outcomes, and overall organizational effectiveness.
🎓 In this episode, Neeraj discusses:
Moving beyond PowerPoints to actionable analytics
How Novartis measured cultural transformation and its ROI
Using predictive models to improve retention at Philips
How analytics revealed critical sales and marketing connections
Productizing HR analytics to drive tangible results
Summary
Deel is the all-in-one payroll and HR platform for global teams.
Deel helps companies simplify every aspect of managing a workforce, from onboarding, compliance and performance management, to global payroll, HRIS and immigration support.
Deel works for full-time employees and independent contractors in more than 150 countries, compliantly.
And getting set up takes just a few minutes.
Neeraj Tandon 0:00
One of the things that I have been advocating recently is that let's move away from PowerPoints, because what happens is we we create beautiful models. We see the insights from those PowerPoints. They get presented to xCO, they present they get presented to the supervisory board at times, but often
then they are dusted.
If you really want an impact coming in from the people analytics team. In this case, we need to productize our insights in this particular case like I'll give you an example in Phillips, what we did was we we did the retention analysis. And this is not something which is new. People have been doing retention analysis for last 20 years. But what the difference that we made in this case was that we took the retention analysis to as a kind of a product. So we we actually created close to about, I would say, 200 models, which was guiding our HR business partners, HR managers in this particular case on how the exits are going to look like in future. We were actually using the time series models to predict how the exits are going to look like in future. By capturing the seasonality in this case, we were also guiding them on what are the key attributes that are contributing to attrition in this particular case.
Chris Rainey 1:21
So I'm just showing the presentation fancy deck. You're basically productizing it and delivering it as a solution. That's our point in the beginning of the call. Absolutely,
Neeraj Tandon 1:30
absolutely.
And the best benefit about it the moment you productize it and you're able to get the attention of the HR business partners. In this case, there is an automatic governance that comes around it.
Chris Rainey 2:03
Welcome to the show. How are you, my friend?
Neeraj Tandon 2:05
I'm doing very well, Chris. And how are you doing? I'm good.
Chris Rainey 2:08
I'm so excited for you. Congratulations on the new thank you so much. Thank you so much. Yeah, and day three, and you're already stuck with me, so I apologize.
Neeraj Tandon 2:19
Not at all. I think three days have been a kind of an information overload for me, time when I can actually give a little bit of download to you.
Chris Rainey 2:28
Well, hopefully I can make this as less stressful and you can relax so we can have a conversation the new role your head of people, insights and analytics. That's agenda. Yes, amazing. For those that aren't aware, tell them a little bit more about your personal background, and then maybe just tell them a little bit about its agenda as well. If they're not familiar with a business,
Neeraj Tandon 2:51
sure, I'll start with myself. I I'm Neeraj, and I have close to about 25 years of experience, which is basically I started as a kind of a financial advisor with a travel company, Thomas Cook and I did corporate real estate sales, and then started doing analytics. When Analytics was just building in India as a kind of a practice. I did CRM analytics, I did procurement analytics, and then finally landed my piece, basically, was when I started doing people analytics where I realized that we could add so much value to the organization by improving employee experiences. So yeah, that was where I think I have been in people analytics space for last 17 years, and I've been a practitioner myself, having done this for companies like HSBC, Ericsson, novertis, Philips and now Syngenta, and I've been a consultant myself as well, where I was consulting companies as a part of Accenture and Willis towers, watts, Syngenta is an amazing company, and I think the reason I decided to take this opportunity was because I think they're at a cusp of a very interesting phase of I think moving from traditional product business to selling solution to farmers, we are basically what one of the world's largest agri business company, and we are into crop protection. We are into seeds. And I think sometimes in the last three days, when I've been actually talking to a few business stakeholders, what I realize is there's so much thought process going on on how they can integrate technology and science to the agri business. So I think the same applies for analytics in the HR space. I think the main mandate that we have is how we can actually look at developing skills within our sales force on. On instead of selling products. How can we actually reskill our workforce to basically sell solutions instead so, so I think amazing transformations the company is going through, and there's a lot that people analytics can actually contribute with. Yeah,
Chris Rainey 5:16
I love that when you pointed that way, by the way, because it's quite an interesting shift, because people may say, okay, they're not that. That's not that different from selling products versus solution, but it's completely different. I went through a similar, not, obviously nowhere near on the scale you're talking about. We launched, obviously, Atlas co pilot, our AI copilot for HR executives, to upskill and reskill them and their teams. And in the beginning we we, we started, we were selling Atlas as a product, and it was way more challenging selling an AI product. You know, here's the integrations, here's the AI. This is how it works. You have your own agents, etc. And then I was really tough, right? And then as soon as we just shifted nothing to do with the product. We just shifted the messaging and the marketing and the sales towards solution orientated. And here's what problems we're solving. It was like, night and day response, and it was so good because everyone it was tempting, because I was kind of following the market, of like, here's our new AI, right? Here's our new agents. And I kind of, and I kind of got caught up in the hype a little bit, if I'm being honest about look back and reflect, yeah, and then now, like we in the last couple of weeks and months, we've focused on this service and the solution, and it's our sales have almost doubled just by changing the messaging. Our power doing it. So is, this is, and it was actually a guest on my show who recommended to me, who had built and sold many HR software. He was like, Chris, you can't, you cannot lead with the product. You have to leave with the solution. It sounds, it sounded so obvious at the time, but then when I implemented it, I was like, oh my god,
Neeraj Tandon 6:59
it makes all the difference, because then you're talking the customer language, yeah, exactly
Chris Rainey 7:02
not what we want, like, you know, not what we care about. Yeah, exactly right. We think, why do they not love it as much as us and all of our fancy features? People don't buy that, right? They buy, they buy, they buy the value, right as well. So, so what I wanted to do today, obviously, you've just jumped into the role, so I got to give you some time to settle in before I start asking you Syngenta related questions. So let me call to talk about some of the project that you've implemented in your career that you're most proud of get perhaps there's some lessons that our audience can can take away. So let's just jump straight in what would be one of the projects or transformations that you've delivered that you feel really added value to the organizations that you've worked within and the employees and customers, but you're also personally very proud of I
Neeraj Tandon 7:57
think I can talk about multiple projects over that I have implemented over the last 17 years with different organization, but the one that I've been most proud of is where we actually brought in predictive analytics to understand the impact of cultural change on the business performance of an organization. And this is a mandate that we got an avert from our CEO, where he actually took a leaf from the book called as unbossed, yeah, yeah, and brought in a change, saying that we need to empower the 100,000 people with a no voters To perform, instead of just focusing on seven or 8000 managers, and we brought in that entire change, where the entire cost concept was you drive curiosity, you inspire people to innovate, and then you unboss them or unleash them to deliver the results for the business in This particular case. So I think it was a very, very interesting project, very intangible, and people who are in the practice would know culture is a beast which is not very easy to tame. So I think what we did in this particular case was we went around researching a lot of papers. What has been done in this area, and especially around knowledge workers, to say, what are the different components that we can typically capture, especially on the on the culture side, and then incorporated those components as a part of the construct that we did for the culture designed Our engagement survey and our team prospective survey, which was basically the feedback of the team on the managers around those different elements. And then floated a quarterly survey to measure the cultural index, something which we created as a part. Of that entire construct on a quarterly basis. Not only that, we did not stop there, but we actually went ahead and measure the movement and this cultural index with the impact it was actually making on some of the hard variables, like the top line, net sales or the bottom line, that was the profitability. So I think that was a kind of a very, very nice project to work on. Wow.
Chris Rainey 10:25
Let's get into it. I love the fact that you ended the last part by here's the actually D ROI, which is always something that we normally don't get to in many, many of these projects. Sorry, apologies. It's interesting you say that, because in the last couple of years, Novartis has now become known and renowned for the culture of curiosity, absolutely right? And I've heard that through my conversations with Simon Brown, their former chief learning officer. And is it Stefan Bert at the time? Stefan Bart?
Neeraj Tandon 10:58
Exactly. Stefan was the CHRO and I
Chris Rainey 11:02
remember interviewing them both and talking about this culture of curiosity that was probably around a similar time you were maybe doing this work as well, and it was fascinating for me, especially in such a highly regulated environment, how do you unleash the curiosity and innovation, but also in an industry which is renowned for having such strict guardrails, right as well, which is super it almost goes against everything to do that. So talk to me. Where did you start? You mentioned about going and collecting that information. Walk us through. What did that look like? You mentioned, so many employees and managers getting some of the sentiment from them. What did that practically look like?
Neeraj Tandon 11:46
I think we started with a few external research papers that were done on what drives curiosity, what inspires employees in a knowledge culture, kind of an environment for the knowledge workers in this particular case. So we researched a few papers in this particular case, and then we went around talking to a few companies that were experts or that were consulting on the culture side. In this case, there are
Chris Rainey 12:14
multiples. Then it came from the research.
Neeraj Tandon 12:17
I think the biggest insight that came in was the variables that that were basically mentioned in those research papers, which said that this is what would drive curiosity among the knowledge workers. This is what would drive the the inspirator, or will inspire the knowledge workers to innovate in this particular case. So the elements were something which we we captured as a part of those research papers, and then we sat with the behavior scientists within the company to see how we can basically create a questionnaire where we can capture the sentiment of our employees on these elements in this case, so we created a kind of a construct for our our own construct for the culture, which reflected the the three elements that we were talking about, the curiosity, the the inspiration and the unbossed culture that
Chris Rainey 13:14
came out of that. Do you remember some of the questions you was asking? I just always grade. I'm always fascinated by those questions. Yeah, so you did that. You did that global survey, like to everyone, absolutely.
Neeraj Tandon 13:32
And it was a quarterly survey, so it was a kind of a kind of a good frequency for us to see how the needle was moving in this particular case,
Chris Rainey 13:40
what were some of the key insights from the survey results that
Neeraj Tandon 13:44
you I think what was important was, during the entire course of the entire implementation in this case, what we learned was that the survey fees back was very, very relative, and they Were variances that could you could actually see coming in from different countries, like Japan would always have low scores
Chris Rainey 14:10
Mexico, low scores on what on
Neeraj Tandon 14:13
the feedback that the employees are providing on a survey. Okay,
Chris Rainey 14:19
because culturally they're less likely, culturally they're more conservative. Yeah, exactly okay.
Neeraj Tandon 14:26
Like Mexico would always be in high 90s,
Brazil would always be in high
so I think the first thing that you needed to do in this particular case is to treat data to make sure that what you're measuring is is uniform across the world, and you're not being
Chris Rainey 14:44
you account in Fauci Yeah, you're accounting experience exactly. Because otherwise, if you only looked at the data without the cultural nuance, you wouldn't see that.
Neeraj Tandon 14:53
You wouldn't see that absolutely. And you would tend to think that the culture index is very, very high in. Mexico, and it's very low in Japan. So that's not the right inference from the entire thing what you need to do. So what we did in this particular case was we actually brought in the country benchmarks,
Chris Rainey 15:09
okay, in what sense are in the platform
Neeraj Tandon 15:14
that was actually helping us do this entire exercise, which was Clint in this particular case, helped us with the average scores by country.
Chris Rainey 15:24
Oh, I see the number of correct.
Neeraj Tandon 15:29
So instead of taking an absolute number, what we did was we took the variance to the benchmarks, love it, as a kind of input to our predictive models.
And that's where I think our results started
Chris Rainey 15:43
to come in. What was the sorry? Carry on. Apologies. No worries.
Neeraj Tandon 15:47
I think you also mentioned what was where you're finding. In this particular case, I could one of the things that I can summarize as a part of our finding was that one of the things that we found out when we were specifically trying to find the relation of some of the cultural variables to net sales or profitability, was that the HR variables had very small contribution to the actual impact on net sales and the ROI so one of the lessons that we learned in this case was that it's not very prudent to develop your HR models when you're trying to predict the business results. Your HR variables can contribute to the business models in this particular case, not create an independent model in this case, interesting.
That's really interesting. It makes sense
Absolutely. So we tried it in every possible manner, trust me, and the differentiated impact of HR variables was not more than 6% in this case really.
Chris Rainey 16:53
What was the areas that moved in need or the most?
Neeraj Tandon 16:59
I think it's the overall business environment that makes a lot of difference. Geopolitical environment, other factors like the market size, your market share. How quick are you to capture the market share? Specifically, when you're talking about a new product, interesting and in pharma, it's the there is a, kind of a huge first advantage. It's
Chris Rainey 17:25
the whole way the industry works, really, like whoever gets their person gets the approval for through the FDA, etc, that you directly capture the 60% of the market immediately, just from that. But then, to your point, no, you then needed the ability to have an organization that can quickly upskill those employees to then capitalize on that market. Because now you then have to train all of those sales reps to sell that new product, and upskill and reskilling fast with that product knowledge to then go and capitalize on that absolutely,
Neeraj Tandon 18:00
absolutely. So I think it's also around a lot of networking, how the marketing teams were, especially around the launches, yeah, where again, I think I have a very good business case study to talk about, where we use the Microsoft vivo insights data to understand how the organization networks were getting formed between the marketing and the sales teams in this particular case to see where our product launches were meeting their business ambitions, in countries where they were meeting the business ambitions and the places where they were not meeting the business ambitions. In this particular case, how did the sales teams network with the other teams in this case to make the launches successful in this case, and we had some very interesting so
Chris Rainey 18:50
you could see through the organizational network analysis that teams that meet more often are more networked, more not more networked, have a increased sales. Basically, wherever
Neeraj Tandon 19:03
we saw kind of a close combination, or a close networking between the sales and the marketing team, there were certain roles that were being pivotal. In this case, wherever you had a stronger person and in in those markets there, the launches were successful than the other markets. It was not successful. So you could identify the critical roles for the for the launches. In this particular case,
Chris Rainey 19:30
that's super interesting, and it's like, again, good hazard data, but also not that surprising. Anyone who's worked in many companies knows there's unless you have that, synergy between sales and marketing, if you because in some companies, they compete against each other. Yeah, right. It's kind of and it whereas when, which is how my previous companies operated, and it was always like, this tension. It's like, wait a minute, we're trying to achieve the same results, but we're, we're, our targets aren't aligned for us. What you know together, so I can sure you can see that straight away. That's super interesting. And then that allowed for you and the team to to adjust and make the relevant investments. We could
Neeraj Tandon 20:10
always provide those invest insights to the teams and turn for them to to obviously adjust accordingly.
Chris Rainey 20:16
Yeah, once you've what would your advice in terms of how to best present these insights back to the team? I
Neeraj Tandon 20:29
think also depends on the kind of analysis that you're doing, Chris, because I think one of the things that I have been advocating recently is that let's move away from PowerPoints, because what happens is we we create beautiful models. We see the insights from those PowerPoints. They get presented to xCO, they present, they get presented to the supervisory board at times, but often,
then they are dusted.
If you really want an impact coming in from the people analytics scene. In this case, we need to productize our insights in this particular case like I'll give you an example in Phillips, what we did was we we did the retention analysis. And this is not something which is new. People have been doing retention analysis for last 20 years. But what the difference that we made in this case was that we took the retention analysis to as a kind of a product. So we we actually created close to about, I would say, 200 models, which was guiding our HR business partners, HR managers, in this particular case, on how the exits are going to look like in future. We were actually using the time series models to predict how the exits are going to look like in future. By capturing the seasonality in this case, we were also guiding them on what are the key attributes that are contributing to attrition in this particular case.
Chris Rainey 21:55
So I'm just showing the presentation at the fancy deck. You're basically productizing it and delivering it as a solution. Back to our point in the beginning of the
Neeraj Tandon 22:03
call, absolutely, absolutely.
And the best benefit on that it the moment you productize it and you're able to get the attention of the Azure business partners. In this case, there's an automatic governance that comes around it. I see, but I think governance is very, very important, Chris, because otherwise you're just generating some insights, but unless somebody is doing something with it, how? How do you either? How would you measure the impact that those insights have really created?
Chris Rainey 22:37
Do people normally do it their way around? They try and do the governance and vend launch, and then that slows you down.
Neeraj Tandon 22:47
I think it also depends on the I would say governance doesn't add a lot of value people would, if people would, and it doesn't come in with a stick. Governance would always come in when people see value in the data or the insights that you're really creating. Okay, it's
Chris Rainey 23:10
a different lens to look for it, because I think when most people think governance, they don't see it from that lens. They see it from a sort of a command and control risk.
Neeraj Tandon 23:21
I also have an example where we were doing this kind of a study in India, and the CEO actually wanted us to refresh the data every month, and what he would sit down and having the benefit of not being restricted by GDPR for an Indian company, we could actually go down to an individual and identify who are your high probables of leaving the company. And we could also give the top three reason on why X or Y person is leaving your company. The CEO, in that particular case, used to sit down every month look at those high probables and call in the CEO minus one leaders to say, what are they doing to retain this person? Wow, because this was a large PPO, KPO, and every person lost worth actually loss in revenue,
Chris Rainey 24:16
yeah, because they're, they are, they are the products, right? Exactly.
Neeraj Tandon 24:19
They are the product Absolutely. So I think you I've seen those kind of governance models as well. But in Europe, when you're doing these kind of studies, you can't go down to an individual. So it has to be more level, so interesting
Chris Rainey 24:37
as well. What some of the other you meant that was one that's super interesting. What are some of the other projects that really stand out to you?
Neeraj Tandon 24:45
I think the other two I would love to mention in this case is, I think I've already talked about the attrition case study where I think, as a consultant, we were doing this project for a BPO KPO company. They were really. Reeling under. They were a part of the large conglomerate earlier, and they were spun off. They were so long, and they were basically reeling under large attrition of like 57% for their call center. So what would happen is they will bring in employees, train them for a period of like 45 days, and within couple of months, they would leave large leakage in terms of revenues, large leakage in terms of the training investments that they were doing. So we were asked to basically come in to understand the reasons and why the employees there was the turnover was so hard. And we were able to guide them on certain areas where I think the annual incentive that the these guys were paying was a 12 monthly incentive. But I and in India at that point of time, Chris, people would shift for like, 150 years, a hike of 150 years.
Chris Rainey 25:57
That does impact their life, huge, hugely, because
Neeraj Tandon 26:01
these are all like fresh graduates coming out of the college, and they would, this would be a kind of large money for them. So, so what I what, one of the suggestions we gave them was that, instead of paying the bonus on an annual basis or a semi annual basis, can you pay the incentives to these call center agents on a monthly basis, and that changed the entire game. We were able to bring down the attrition by 27% which was a kind of a big win for us. The CEO of that company actually came back to thank us for the insights that they're provided, and he made a big statement saying that you've actually reinstated my belief in consultants.
Chris Rainey 26:48
That's a nice way of saying it. Yeah. I mean, as someone who had monthly sales commission for 10 years, I can't imagine having yearly I would never have stayed like, I can just say right now. Like, as for me, like, the chase of closing the deals to wait a year is crazy. It's crazy. It's crazy. It's a crazy idea to be able to I have some friends, actually, who still have annual bonuses. And I'm always like, it's obviously this. It depends on the type of role you're in. Of, like, if you're talking about sales. It has to be monthly, you know, like, whereas I understand if it's a different type of role yearly, maybe will work, but depends on the type of role. If you just at a lower level with that, meet that, that level of pay, you need a monthly incentive to keep people going. So how did you I mean, that almost seems like a no brainer, right? But what was some of the data and research you did behind it that comes to the conclusion?
Neeraj Tandon 27:46
So we actually pulled in a lot of data sets. We included all the demographic data sets the colleges that they were hiring from. So we were able to identify the colleges where they could retain people. You were looking at the performance of these agents during the training period. We looked at other data sets, which included other factors, like we went to the extent of seeing how far they were from the office. So what was their travel time to the to the office in this particular case as well. So that helped us create a kind of a model profile of an individual who would stay longer. And the biggest thing what we did in this case was we actually looked at these profiles in the lens of people who were staying longer versus people who are leaving very, very quickly. And that helped us basically change the hiring patterns with the company. That's
Chris Rainey 28:39
so interesting. You looked at it from that frame of mind, rather from the negative frame. Let's look at why these people are staying so long, rather than why these people are leaving so quickly, right? And you can see, and I'm like you said, I'm sure that one of the first things you saw was these people will live colloquially. They don't have a lot of them. Yes, they don't have a huge I'm just guessing, I mean, I can imagine that that would be one of those, right? This
Neeraj Tandon 29:04
was a Hyderabad based company, and obviously Hyderabad, being a bigger city, attracts a lot of talent from some of the satellite funds, yeah. So I think if you could prioritize where you're hiring from,
that made a big impact as well. Was
Chris Rainey 29:23
any other key predictors? So you got salary, you got obviously, distance. What was one of the other major factors?
Neeraj Tandon 29:28
I think educational background was one of the major factors. Earlier, these guys were actually hiring a lot of engineers into the call center jobs. And trust me, in India, this is not surprising. A lot of engineers would be starting their careers with
Chris Rainey 29:43
a COVID. Interesting,
Neeraj Tandon 29:46
so, but we very quickly realized that this is a kind of a parking spot that most of the engineers are basically looking for, and they would not look for a
Chris Rainey 29:58
meantime, get a sales job while. So I look for my engineering job, correct. Correct makes sense for them.
More from the HR Leaders Podcast
Anisha Thomas, Head of People at Inscribe.