Analytics: Driving business value matters more than perfect models (podcast episode #133)

Action Box: Claim your scholarship to boost conversions

Want to boost your conversions and collaborate with customer-first marketers? Claim your free 3-month AI Guild scholarship (from MeclabsAI, MarketingSherpa’s parent company).

Want to move up in your organization? Here’s the best advice I’ve gotten in my career – ‘train others to replace you.’

Hey, it’s a bit scary. If others can replace you, it’s easy to fire you, right? However, if you’re truly providing value, once others can do your current tasks, it also makes it possible for you to move up to the next level in your organization.

Now here’s another version of that same wisdom – ‘automate your job.’

That’s a lesson from my most recent guest. To hear the story behind that lesson, along with many more lesson-filled stories, I talked to Prateek Shrivastava, advanced analytics manager, Cummins.

Cummins is a public company traded on the New York Stock Exchange. It reported $34.1 billion in revenue for 2024.

Shrivastava leads a worldwide team of five data scientists, and his team estimated they generate $50 million in annual savings with their models.

Hear the full episode using this embedded player or by clicking through to your preferred audio streaming service using the links below it.

Listen on Apple Podcasts | Listen on Spotify | Listen on Amazon Music

Lessons from the things he made

Embrace failure

As a data scientist, Shrivastava had built some groundbreaking models – but also plenty of garbage ones. Sometimes, in tackling complex problems, we lose sight of the fact that not everything is just a mathematical challenge. One particular model he worked on was technically brilliant, but the business had no immediate need for it. After an entire year of development, it was shelved in a single day.

That experience taught him a crucial lesson: there's a difference between problems that are good to solve and those that need to be solved.

Pivoting is a superpower

When Shrivastava entered the workforce in 2012, the idea of a "forever career" was still prevalent. He initially struggled with the variety of projects and technologies he was thrown into, thinking he needed a single, well-defined path. But one particular project sparked his passion, leading him to shift his career focus entirely.

Shrivastava realized that his true strength wasn’t in choosing a rigid path but in his willingness to keep learning and evolving. That adaptability has been the key to his success.

Driving business value matters more than perfect models

A technically sound model doesn’t automatically translate to a successful business project. Shrivastava once worked on an initiative that delivered millions of dollars in incremental savings, not because the model was perfect, but because it directly addressed a real business need. Contrast that with a previous project where the model was technically superior but ultimately useless because it didn’t align with business priorities.

The key takeaway? The best solutions are the ones that drive impact, not just the most complex ones.

Lessons from the people he made it with

Helping others creates genuine connections

via Shashank Saxena, CEO, Pantomath

Back in 2011, when Shrivastava was just starting out, his mentor, Shashank Saxena, went out of his way to guide Shrivastava – even though he had no obligation to do so. Today, Saxena is a successful venture capitalist, and his advice still sticks with Shrivastava: help selflessly, and you’ll build meaningful relationships.

That philosophy has shaped Shrivastava’s career, proving time and again that generosity fosters authentic professional growth.

Automate your job

via Nick Hamilton, SVP, Kroger Precision Marketing

At 84.51°, his leader, Nick Hamilton, gave him a perspective that changed the way Shrivastava approached technology. Hamilton told him, “Your job isn’t just to build models – it’s to automate your work to the point that you’re no longer needed for it.” That advice reframed his approach: rather than just solving problems, Shrivastava started focusing on creating scalable, long-term solutions that eliminate redundant tasks.

Be your own brand

via Thiag Loganathan, CEO, Cardinality.ai

At DMI, his leader, Thiag Loganathan, once told him to wear shorts to a business meeting – not as a gimmick, but as a statement. His message was clear: You are your own brand. That moment forced him to rethink his professional identity. It wasn’t about what Shrivastava wore but about how Shrivastava represented himself and his expertise.

It made him more intentional about how he presents himself to the world, balancing individuality with the responsibility of representing his organization.

Discussed in this episode

Social Media Marketing: Educate your executives (podcast episode #123)

Value Proposition Power: 3 ways to intensify the force of your value proposition

Get more episodes

Subscribe to the MarketingSherpa email newsletter to get more insights from your fellow marketers. Sign up for free if you’d like to get more episodes like this one.

For more insights, check out...

This podcast is not about marketing – it is about the marketer. It draws its inspiration from the Flint McGlaughlin quote, “The key to transformative marketing is a transformed marketer” from the Become a Marketer-Philosopher: Create and optimize high-converting webpages free digital marketing course.

Transcript

Not ready for a listen yet? Interested in searching the conversation? No problem. Below is a rough transcript of our discussion.

Prateek Shrivastava : But if you keep on moving to different interesting problems, then you will create a personality who knows everything about the business and who can create solutions in different spaces. And I think, with this and what I learned and everything in my career, I realized that that's what that's what I have to do. So every couple of years, you keep switching.

If you don't want to switch companies, switch teams, at least understand different areas of business.

Intro : Welcome to how I made it in marketing. From marketing Sherpa, we scour pitches from hundreds of creative leaders and uncover specific examples, not just trending ideas or buzzword laden schmaltz. Real world examples to help you transform yourself as a marketer. Now here's your host. The senior Director of Content and Marketing at Marketing Sherpa, Daniel Bernstein, to tell you about today's guest and.

Daniel Burstein : Want to move up in your organization? Here's the best advice I've gotten in my career. Train others to replace you. Hey, it's a bit scary if others can't place you. It's also easy to fire you, right? However, if you're truly providing value once others can do your current tasks, it also makes it possible for you to move up to the next level in your organization.

Now here's another version of that same wisdom. Automate your job. That's a lesson for my next guest. Joining me now to share the story behind that lesson, along with many more lesson filled stories, is Prateek Shrivastav, Advanced Analytics Manager at Cummins. Thanks for joining me.

Prateek Shrivastava : Thank you, Dana.

Daniel Burstein : So let's take a look at background real quick so you know who I'm talking to. He's been a senior statistical analyst at DMI, a senior data scientist at 8451 and for the past four years he's been at Cummins. Cummins is a publicly traded company on the New York Stock Exchange. It reported at $34.1 billion in revenue for 2024.

Prateek leads a worldwide team of five data scientists, and his team estimated they generate $50 million in annual savings with their models. So, Prateek, give us a sense. What is your day like as advanced analytics manager?

Prateek Shrivastava : So I usually start the day with looking at supply chain problems. So we have built several dashboards that that track all the metrics metrics. It's sometimes our forecast is good, sometimes it's not that great. I'm just looking for red flags over there. If there is some, then then we we get, into a more, deeper dive into into the problems that we might be having.

Then, I usually talk to my peers in several business groups. There's, there is a project that we are doing for transportation. There is a project that we are doing with the plant managers. We are trying to reduce the inventory for them. So all of these meetings happens throughout the day. And once we have an understanding of what we are going to do, I go usually go deep dive into deep dive into the models that we are going to build, and provide technical leadership as well as, mentorship to all of my team members, helping them out, building better models and, creating better solutions.

Daniel Burstein : Okay. So normally at this point I'll just jump into the lessons in the stories, but I got a question on that number we started with. So as we were talking normally I got like close on, we're talking about a marketing budget that sort of thing. But I thought it'd be really interesting to talk to you because you got an interesting role.

And also in your podcast guest application, I mean, you had really good stories. You took it through and you didn't just see this number as you saw it as like, okay, business wisdom. But. All right. Five data scientist you lead around the world $50 million in annual savings with your models. That is a heck of an ROI.

How are you doing that? How does that.

Prateek Shrivastava : Work? Yeah. So, just to give you an example, like, the revenue is incremental, right? So when I started first year, we my generated like $5 million worth of savings. Next year we generated another $5 million in savings. But what happens is that, every year when you are generating $5 million in savings, third year, you have already have like 10 million in savings that you are generating.

So it's an incremental revenue. And over period of time, it just adds up. For example, last year, one of the projects that we were working on, we were trying to get, the supplier recovery costs, so to come in space, a lot of money for, for all of these suppliers. Sorry, all of these claims. But whenever a claim happens.

But some of that claim cost comes from the suppliers. So we needed to make sure that, we are truly tracking that we are getting to the correct supplier who supplied us the part that has failed, and to build a solution for that. Initially all of the processes were manual, so so we needed to convert that into a better data science model.

That increased that have like the savings by three, $4 million. So that's an another incremental thing that would happen. So so that's how that's how the 50 million number comes into picture.

Daniel Burstein : That's fantastic. Well, let's take a look at another way that $50 million number comes into a picture. Right. Because you didn't just, you know, pop out of the womb knowing how to do this. Yeah. You built up this career to get to where you are today. So let's look at how you did that. Some of the key lessons from there.

So the first lesson you mentioned is embrace failure. Yeah. How did you learn this lesson.

Prateek Shrivastava : Now this is a very, interesting example because I think, one of the models that we created was to predict, when a truck might fail. So, for example, trucks are driving on the road all the time. We have, like, telematics boxes that are sending us signals. And it's a widely, it's a wide use case that happens across the industry.

So. So we are getting all those signals for using those signals. We want to predict when a truck might fail. There could be several, use uses of that. We could, get those parts in the inventory faster if they're, and then then when the truck fails, the downtime is much, much less than what it would be.

The other use cases that, it could get to the dealer faster. We can tell the truck that something is going on with your engine. You need to go to the dealer first. So all of those things, we that's that's what we thought that this could be like practical implications of the work that we are doing. We build a model, to predict that it was it was a good enough model.

I wouldn't say it was the best model of all, but it was it was 78% correct. It was not nothing. It was better than a coin toss. Much better than a coin, as I would say. When we built it, we spent considerable amount of time on it. But once we get to the business, they say that.

Okay, you know, the use cases that you are suggesting, those all require a lot of process change. It's not just the data science thing. It's, process related thing. And, we don't have a control over it, and it would require business to put substantial effort on that. And the value that, that it will generate at the end of it is not that much.

We in the hindsight, when I look at it, I can see that, yes, there are some problems with it, and we shouldn't have spent that much time, the amount of time we spent building it. However, what that taught me is that you need to understand the business value first. You should not just go dive into the problem, spend so much amount and so much money and time into problems which I'm not probably worth solving.

Or maybe the time hasn't come yet. So. So it's not exactly a failure. Failure, I would say. But it's a failure in terms of, understand business context.

Daniel Burstein : So that's an example from you understand that context. Can you give us an example of like what a marketing leader should bring to you to build a successful model? Because that's a great example. You talk about not understanding the process. I remember interviewing a CMO at a major hotel, Las Vegas. I'm forgetting his name right now, but he talked about you wanted to add this, like, kind of loyalty program or do something, you know, check in.

And he's like, hey, this will be a great thing. That was from a marketing perspective. But when he sat down with the actual, you know, customer service team or whoever, it was understanding that this would take five minutes extra at every check in, which doesn't sound like a big deal, but thousands and thousands of guests they have every day or whatever, you know, it was just the cost of it was just ridiculous.

So, you know, a marketing leader, we have these ideas of, oh, you know, math and data models and all these things. It's kind of this magic goo. And it'll just, you know, give us these all these answers and solve all these problems. But if we don't come to you with the correct ideas of what we're trying to solve or process, I imagine, or more likely to lead to failure.

So do you have an example of, okay, here's a marketing year came and they brought the right stuff so we can set it up for success.

Prateek Shrivastava : Yeah. So, that used to happen a lot in my previous role when I was in 84, 51 and 84, 51 is, just to give you a brief background about what 8451 is so 84, 51, 84 and 51 is the longitude of Cincinnati, first of all. And it's this company. It's used to be called enemy USA.

They were doing some work in Tesco in UK and Kroger had similar problems in US. So so Kroger called tandem the UK to come into us and then do this kind of analytics for them. And that's, that's very similar to what you were talking about the loyalty marketing side. So we were doing a lot of marketing analytics for Kroger.

So since the company was set up like that, we had a lot of business interaction, and the people who were in business were directly working with the data scientist to build, build, build tools like that. So, so for marketers, I think the most important thing is to, for us to build better products. To, to better market those products to the audience, marketers needs to have, like, really good business understanding.

And they need to come up with business processes that it probably not tell us, like what to do with the data, but but helps us understand the processes a little bit better. I would say.

Daniel Burstein : Well, I mean, also, is there the sense that data isn't the thing that will solve every problem because so something that we talk about now, I mean, data driven, that word is just used constantly. But sometimes, you know, the data that we find isn't really elucidating the problem we're trying to solve. So, I mean, what do they need to do to give you that actually understanding?

So I'll give you a real quick example. You know, my background and, you know, with my labs and stuff were more ab testing and that really experimentation. And when you talk about embrace Failure, I love it because that's what we're trying to do with experimentation. We're trying to fail sometimes because if we if we don't fail, then we're not really pushing the edge in testing.

But there's different types of failing. There's failing because we have an accurate hypothesis. And we thought this thing would happen, but it didn't. And we learned about the customer versus there's failing because we don't actually have a hypothesis. We just did some stuff. We got some results and we don't really know what it means. So so that's where I see like when they come to like, what are some things that these data models would be good to predict.

And what are some things where that data isn't the answer for everything, or am I wrong? And it is the answer for everything.

Prateek Shrivastava : So in my experience, if, data has been answered for a lot of things, I have not come up with examples where the data has not been. I mean, maybe because the questions that I'm getting, people have already thought of some sort of hypothesis. For example, like one example I came to my mind is that at Kroger we were trying to do some to to my team used to send coupons, to all the customers.

So we were looking for like, who were the highly engaged shoppers at Kroger and buying all those product. And then we would target them with the best coupons available to us. So when we were doing that, we used to do a lot of a B testing, we will have like, people who are going to get some coupons, but they'll not get anything.

And then we will test their behavior. How is it change? How is it, working out compared to people who are getting those coupons? And one of the things that people come up with is that they don't it's it's a very silly example. Okay. So they said that, instead of sending a coupon for, body wash like a particular brand, let's say dove, instead of don't send dove coupon, but instead send, send a coupon just for any, any body wash.

So it will give people more freedom to purchase whatever they want. It will still generate some value for Kroger, because they're still going into the shop and buying whatever they want. And then later on, we realize that while it sounds a really great idea, it isn't really a great idea because what it has done well, we are not getting any money from any of the, companies because now you're just targeting anything across Kroger.

So Kroger is paying for that. And at the same time, people really valued their brand loyalty. People really wanted those coupons. So even we saw engagement go down, but with those kind of questions. So it was a while hypothesis, but sometimes it just it just came out to be, incorrect. I would say.

Daniel Burstein : That's a that's a great example. Sometimes we think we're better serving the customer and we're really not.

Prateek Shrivastava : But yeah, I mean, I'm the now I'm trying to think like, if there is any other example where marketers would have come with, with a different hypothesis if it did something that could not be solved with data. And I don't have really solid examples on that.

Daniel Burstein : The okay, that's fine. It makes it. You're the data guy, and if data works for things, that's great. Let's take another look at another, lesson here. You mentioned pivoting is a superpower. How did you learn this lesson?

Prateek Shrivastava : So that is from my career, I would say. So back in 2011, 2012, I graduated from university and I was looking for a job. And, the big hype thing, at that point of time was business intelligence. I mean, it still is, but at that point of time, it was the it was the data science, it was the artificial intelligence, or 2011.

I would say. So, everybody wanted to go into business intelligence. I was like, oh, this is the shiny thing. Let's, let's go behind it and let's, let's try to get a career in that. And that's how I joined DMI. And we were building B AI solutions for a different companies. But later on I realized that, no, this is not really, what I want to do throughout my career.

And the career Q career since then has been pivoted to a lot of different things, data science to data, from business intelligence to data analytics. Then I took some courses. I tried to go into more into data science. From there I actually made a switch to data science, but now it seems like, the data science is again changing.

All the roles that we would all the things that we used to do in data science, like building better models, a lot of things are getting automated. Everything is, more like a click of a button. So now the career is going into more machine learning and artificial intelligence, right? And it's not like I am pivoting.

It's just industry. It's pivoting. But as long as you are aligned with what industry is going towards and then you are constantly learning new things, and, and switching to the things that, that are happening in the industry, I think that would help, your move your career substantially.

Daniel Burstein : Yeah, definitely. I mean, we've heard about pivoting a lot here on how, I mean, marketing. And I wonder if there's an example from you, specifically from an analytics approach, like we're talking about data before, like, are you data driven in your career? Do you an example of using data to figure out that pivot? So I'll give you a quick example.

I interviewed Carlos Gayle, a US market brand evangelist at Get Response on how I made it a marketing, and one of his lessons was about pivoting fast and recognizing opportunities is key. And he told the story. He heard a lot of pivoting stories during Covid, and he told the story during Covid launching Covid, outlaw masks during the pandemic.

But his he's got a social media background, so he's looking at it through the lens of, okay, this is, you know, I can leverage TikTok. It'll be viral. You know, his pivots focused on social media. I'd say for me, I'm a writer. So every decision I make, even in my career, I take through like a writing approach to try to like right through it.

So for you fatigue, I worry like, yes, you talk about the world is changing. Yeah. Our audience is feeling that. I remember feeling and I started my career writing. Printing. Right. Yeah. And then the world was changing and I had to figure out how do I pivot with my writing digitally? Yeah. So in your career, you're working with big companies, you're going to big brands.

You're giving them data to make decisions. How do you use data to pivot in your career to make career decisions?

Prateek Shrivastava : So, I guess the way I have been looking at it is, mostly through a lens of, what is, what are people talking about? So not just social media, right. So I think, like the first major switch that happened in the industry was like 2017. There was, so Google, usually do this dev conferences, I think every year in March, like me, I think every year in May, they have a big dev conference in, San Francisco.

And every year they come up with like, new things. And I specifically remember 2017, conference because in 2017, what happened was, in one of their keynotes, they, they, they had like, the first version of ChatGPT that you would see. Right? So they, they created an assistant, and the assistant was talking to the restaurant owners and setting up a reservation for that.

So it was like really natural language. And that was the first time the whole world saw that computer can do all of this kind of work. And that made me realize that, this is this is going to happen. This is this is coming in. And then that's when I started learning more into machine learning side that all this natural language processes was happening.

And, so since then I've been, like, keenly following all the trends. So the way I would do it is, is that, I'm always aware of what is happening in the outside world. So always connected with, what are the major industry trends that are happening? And that allows me to learn as well as pivot whenever necessary.

Daniel Burstein : Okay, great. Let's talk about another, brand company corporate example. You said driving business value matters more than perfect models. Yeah, I had to learn this lesson.

Prateek Shrivastava : So, they said. Right. So the first project I was doing at governments, we were we we created a really good model. And then and it was a sexy model, I would say. I mean, it was it was predicting something that it was supposed to predict. And not all data science models are like that. You know, sometimes you you put your effort and then you, you figure out that, no, it did.

It cannot be done. You know, some some of those things are like that, like stock market prediction. People have been dying trying to do that forever, but it just doesn't work. So so that kind of problems also occur in your business when when you are researching something, you spend a lot of time and then you realize that it cannot be done.

However, the other project that I was, saying that, we, we were trying to point all the, the claims that were happening in governments to the actual suppliers, so that we can get some money out of that. That particular project was not very sexy. It's not like, it's not, I wouldn't say that.

It is a data science project. More of it is is moving data and, and just working with data like a data analytics project, I would so, while it is not sexy, it works, and it does create several million dollars in savings every year. So, so that helps you a lot, right? It drives more business value as compared to what?

Traditionally a data scientist would do. But then the, the way I think of it is that my job role is fluid, and, it needs to be. What it needs to do is to drive business value and not, my personal value or my personal value is also important, though.

Daniel Burstein : It is. And that's a good lesson because like I said, when I've worked with data scientists in the past, sometimes they create this a great model. And it's like, okay, well, what is this really doing for us? Yes, there's a bunch of numbers, but how does this help us make a decision or make a better decision? So let me ask you, do you have an example of how you work to understand the customer experience, to build models that ultimately impact the business?

And for example, while you're thinking of that, in value proposition power three Ways to Intensify the Force of Your Value Proposition, our CEO, Philip McLaughlin, teaches separate the means from the end, focus on the customer's experience or the value. And sometimes I think holistically as a business, this is where we're challenge, right? Because you said this has to work for me too.

That's how businesses think. That's how individuals in a business think. We have to hit certain numbers this quarter. We have to. So we're focused on the business goals, right. Which is really a lagging indicator versus the actual customer experience, which is a leading indicator. So do you think of any examples where the yes, we're building business value, but we're doing it by putting the customer first and building a better customer experience.

Prateek Shrivastava : Yeah, yeah, there are definitely so that that there is a lot of what we did at 84, 55 to 8450. And as you said, the motto was customer first. That's what we wanted to do. And then we realized that if we put customer first, the business value, I mean, we will derive business value out of that. So, the coupons that we were sending, they needed to be very specific with what you're buying.

So we were looking at all the transactions that you are making and developing insights from that understanding who you are. Exactly. Without without looking at your data at all. We were just trying to look at that number and from that number figuring out like who this person is and based on that, we were we were sending out the coupons to you.

So the way you measure it, the way you measure it's a product is success or not is to if to to check, how your shopping behavior has changed since you got that coupon or has it changed a lot. So if it has changed and if it has changed or positive, if you are going to Kroger more often, you had by using those coupons.

Using those coupons was like highly indicative factor of if it is actually serving you any value or not, otherwise you wouldn't go and you wouldn't use those coupons if they were not relevant to you. And in the long run, are you generating more business revenue for Kroger or not? These are the proxies that that would help us measure how satisfied the customer is.

And, the other point is like, people will send out these emails to our, CEO, like they didn't send out, does he, Kroger to you as well as it should be? And you telling them how how much they loved those coupons that we were sending out to them. So both both of those things were happening.

Daniel Burstein : Wow. That's some great. Yeah. Lots of people are, you know, see, it's in there. Good coupon. That's pretty good. Yeah. Have, All right. Well, those are some lessons we just talked about from the things batik has made in his career. In just a moment, we're going to talk about some of the lessons from some of the people he made them with, some of the people he collaborated with.

But first I should mention that the how I Made It marketing podcast is underwritten by Mic Labs. I the parent company of marketing Sherpa. You can get conversion focused training from the lab that helped pioneer the conversion industry in our AI guild, along with a community to collaborate with. You can grab your free three month scholarship to the AI Guild at join Dot Mech Labs ai.com that's joining Mech Labs ai.com to join the AI Guild community.

All right, so let's take a look at some of the lessons from people you collaborate with. The first person you mention mentioned was Shashank Saxena, the CEO of Pinto Math. And from Shashank you said you learned that helping others creates genuine connection. Yeah, but how did you learn this?

Prateek Shrivastava : So Shashank is a very interesting man. I have not met, a lot of people like that. And he's he's very successful in his career, so, when I was coming to University of Cincinnati for my masters, he actually started a program at University of Cincinnati mentoring, new folks like, whoever I had coming from anywhere.

And, fortunately, he got assigned as my mentor, and he just went out of the way to help me out. Right. So I was I was taking a that was my first flight, like first flight in my life. And that runs from Delhi to Chicago. And I took that flight. I came to Cincinnati, and he was there, like, he directly came to meet me at my apartment, on the first day itself.

And I was like, okay, this person is like a VP at the Citibank. He was he was VP at the Citibank, like at that point of time. And then he just went out of the way to help me out with everything. So he took me shopping and I, I came from India, so I had like, no, no warm clothes, etc. so he took me to shopping to buy all of those things and, and throughout my career he always mentored me and he always helped me out and, just made me wonder, like, what is the impact?

Like, why is he doing that? And then when I asked him, because he was my mentor, I asked him, like, why do you do that? Is like he said that while there is no benefit that you can see right now, the benefit is in long term. You don't play for short term things. The genuine connections that you develop through these kind of relationship, these organic relationships that sort help you go to grow faster.

So he was he was not just helping me, he was helping out his colleagues. He was helping out the seniors. He was just he would just go out of the way for anything that he would believe in. And that helped him create like, his company created a company called Windley and that got sold to rugby. And now now he's, he's he's he's a very rich dude.

No. And so, he just he just created those values organically throughout his career and that, that helped him grow so much. And that's what I learned from him. So if somebody asked me for doing any work, I wouldn't usually say no, if I can, if it is in my if it is in my possibility, if I can help out anybody, I would help out.

So and that's what I learned from him, specifically.

Daniel Burstein : Well, is there anything you proactively do to build genuine connections in Cummins? Because one thing I've learned, like a big company like Cummins, like in big companies, they're sometimes so hard to navigate, so hard to really connect. And definitely people in data science or analytics roles, I mean, from what I've seen, they have, kind of more tendency to kind of be heads down and to focus on their models versus kind of pollinating throughout the company.

So is there anything specific you do to, to, to forge genuine connections?

Prateek Shrivastava : Yeah. So at Cummins we do have a mentorship program. So you can register for that. And then the assignment are even outside Cummins. I am part of Informs community. So informs is an organization for all our professionals, like operations research professionals. And, they are also I do a lot of pro bono work as well as, like, mentorship programs there.

So so that's all I'm doing to do my part in this, in this whole world. And that's something that I learn from him. Yeah.

Daniel Burstein : And is there anything you do when you bring your team on to mentor them and help them? Because, again, as you mentioned, it's a it's a global team. It's a big company. How do you how do you help get them to for genuine connection.

Prateek Shrivastava : So I usually set up so when, when, when anybody joins my team, I usually, understand what they really want to do and their career. And then I try to connect it with people at Cummins who are, masters of that area. For example, like my manager. Help me do that. So my mentor, is a director in a really completely different organization.

So I would talk to him for getting any advice and I would suggest, for my, my peers to do the same, like my, my, my, direct approach is to do the same.

Daniel Burstein : Okay. So this next one is kind of controversial, okay. Because I've definitely talked to a lot of people of my career who specifically try not to do this. They try to undermine this because they're worried they're going to get replaced by a vendor or someone else. So you said you learn to automate your job and you learn this from Nick Hamilton, the SVP of Kroger Precision Marketing.

So how do you learn this from Nick?

Prateek Shrivastava : Nick, Nick is someone who who grew up very quickly in his career. So to Nick, Nick went from like an analyst to like a VP and like to, like SVP. Now, in, in matter of years, at Kroger and it it's a very hard thing to do at at 8415 and not anywhere I think.

So, when he became my vice president, when I was, I was at 84, 51, one of the key things that he said, during one of the town halls was that he once and maybe that's only true for data scientist. Maybe. No, that's not true for every every career. But at least it is true for data science.

Definitely. So when we are building. So he said, that, what I want you guys to do is to, is to automate whatever you are doing. So if you are doing, if you have created a model, you I don't want you to stay with that model throughout your career, that you just hold on to it, and then you are just, like, making little, little small incremental changes.

I want you to automate the whole thing so that you can move on to the next interesting problem. If you just spend all the time just, keeping in keeping with, with that particular thing in mind, then then you're not going to go anywhere you want to new learn new things. You won't learn new data, you won't learn what is the business.

You just stay with one thing and that that would easy. That would become much easy. I mean, for me, it would become much easy to lay you off in future because we already know that what you do and what happens there. But if you keep on moving to different interesting problems, then you will generate a personality who knows everything about the business and who can create solutions in different spaces.

And I think, with this and what I learned and pivoting in my career, I realized that that's what that's what I have to do. So every couple of years, you keep switching. If you don't want to switch companies, switch teams, at least understand different areas of business and don't stay with one problem itself. When you get a problem, create a solution for it, and then move on to the next problem.

The solution needs to be self-sustained, though. I mean, I'm not saying that you create a solution and then just forget about it. You create a solution. Automated, and that's what that's what my calls aside. And if I look at Nick's career, that's what he do. That's what he did. And then he moved, he became like, a really big person at, 84, 58.

Daniel Burstein : Okay. Pretty good. You talk about automating your job. And so just, you know, the water was swimming in his eye, and I can't I can't ignore that. So I want to ask let me ask about what we should be automating and what we should never automate known as humans. And I'll tell you why I ask, especially so Cummins, you are in the industrial sector in manufacturing and automation came probably first in your industry, right.

We talked about automating the factory floor. I remember, you know, I was in high school, college in the 90s. We saw all these people losing their jobs because of automation, outsourcing. I'm like, I want to choose a career where that's not going to that's not going to happen to me. And now the crazy thing is automation is coming for the writers.

You know, I I'm a writer. I'm immune. Automation is coming for the writers. The marketers are creatives. So I think you're a little more on the leading edge. And some of us for two reasons. Like I said, you do work at Cummins, you do work in this industrial atmosphere where automation has paid off in a lot of ways very early.

And as you mentioned in 2017, you know, as a data scientist, you're looking at these things, you're looking at, you know, we probably didn't see the ChatGPT style till many years later. So as we bounce our careers, as we bounce our teams, as we bounce our marketing departments, what should we be using AI for and what should we make sure?

And how can we learn as humans to keep something for ourselves that's going to still be providing value?

Prateek Shrivastava : So I think like first of all, the understanding of AI is a little bit important here. So what we need to realize about AI is that AI is only learning from our collective human experiences. And those collect. I know now the collective human experience is at a large like we have like, we have from Shakespeare to like now everything is in there, everything is uploaded.

And that's what the AI has learned from. And that's why it's so good and it is going to get better. But I think what we also need to realize is that it still requires a human to understand, output of that AI and convert that into a business outcome. Yes. For writers, I think it is. It is a very tricky situation.

Right now because we are in a very early phase of it. We still don't know what exactly I could do and what exactly where it would lend us to, but for for everyone. My suggestion is that don't be scared of AI. Use it as a as a productive tool. As you said that, it takes you 85, 80% over there already.

Like when you start using an AI, like for, let's say for a marketer, when you are building a plan, it will take you there. But I think like AI at the end, like the, human understanding of it is not going to go away. And, and the connections that you will have with people that would always be required.

So once you take those two things in consideration, I think that that's when I can be used as a tool, versus as, as something that would just truly replace everybody.

Daniel Burstein : Sure. Yeah. All right. Here's another career lesson you said you learned be your own brand. You learned this from the I love Nathan, the CEO of cardinality Dot. I, how did you learn this from. Yeah.

Prateek Shrivastava : So this one, I think this one is more controversial. So, what, more.

Daniel Burstein : Controversial than automate your job? Okay. That's it.

Prateek Shrivastava : Yeah. No. So this is, this is when, we were working on one of the projects, and then we were we were in San Diego, presenting to one of our clients, and I was I was a very I was a junior analyst back then. So for junior analyst to go into, major majors, sale, where we are trying to sell our product to them, we, we developed a POC for them, and then we're just going to show show what we have created.

And then ask for more, bigger projects. And, I did not know what to wear, in that presentation. And then they looked at me and then the accept that, batik. You realize that you are not, representing me. You are not representing the company. You are representing yourself, the way you want the world to look at you.

Not just think about, like, how, how it would be looked upon on the company, but think about how it would would look upon on you. And if you want to wear shorts, go ahead with it. If you just want to come in T-shirt and short, but but then give a presentation that that matches that, you know, so it's up to you.

Like you are building your own brand. And when you are building your own brand, the way it changed me is that, it made me realize how people would look at you with, with what you do and how how you behave yourself in public. And I did not had that kind of self-consciousness before he gave me this, knowledge.

So, since then, I'll be I'm, like, very cautious of caution, like, I take caution, with before saying something, before understanding, like how the world would react to what I'm saying or what I'm not saying. So. Yeah. So that was that was an interesting lesson from him.

Daniel Burstein : I mean, in terms of building your own brand too, like, do you create content and put it out there? Because I've got to mention, like, my hats off to you. You are you apply to be on the how I Made it marketing podcast. Like we've talked about, we normally have cameras on who are used to speaking. I you know, not many analytics managers data scientists, analytics leaders wouldn't feel comfortable doing an interview like this.

There's these other things too. Like I'm a content guy, so it's natural to me to do things like LinkedIn and, you know, speak at conferences and presentation and, and teach and stuff. But there's so many other roles where that's important for branding, but people don't feel comfortable putting themselves out there. So is there something you've learned? I don't know if you do LinkedIn or speak at conferences or other other things you do, but obviously on a podcast there's something you learn and you can teach to other, more analytical, data tech oriented peers about building a personal brand.

Yeah.

Prateek Shrivastava : Yeah, definitely. So. Again, what I have realized is that, your brand matters. Like, if I, if I'm looking for a job outside somewhere, like, if I'm looking for any job, people need to know what I have done. Like, I can tell them what I've done, but if I show them what I have done, that that's a completely separate thing.

Going on podcast, building your own personal brand. That that all helps you in, like, building your, building your career to a bigger level. Right? And, of course, I like talking to people. And of course, I like meeting new people. And that's that's why I do it personally. But definitely I would advise everybody to, to have to have your, your career in your own hands and, and do these kind of activities.

Daniel Burstein : I think if you had to break it down, what are the key qualities of an effective marketer? And I know I've asked a lot of marketers that question. I appreciate your perspective now because you're kind of, you know, coming from a bit of an outside realm and you're looking at what's are externally and what are the key qualities of an effective marketer.

Prateek Shrivastava : I think for a marketer, the data, I mean, so I think we talked about, talked about it earlier where we talked about, data, like what our data cannot tell to a marketer, but what I realized as data with storytelling is, is the really important thing that that marketers should, should strive for. Data alone doesn't tell you anything.

But if it is combined with a storytelling approach where you can segment through, like why you doing something and how it is, backed up by data, that's when you become a very successful marketer. That's when you are able to sell your product both to the customer as well as to the leadership. And I think that's, important lesson as a data scientist talking to a marketer.

Daniel Burstein : Not street data, but story data on its own is just numbers. But, all right, thank you so much for your time, for.

Prateek Shrivastava : This was a lot of fun. Thank you. Thanks, Danielle.

Daniel Burstein : Thanks to everyone for listening.

Outro : Thank you for joining us for how I made it and marketing with Daniel Burstein. Now that you've got an inspiration for transforming yourself as a marketer, get some ideas for your next marketing campaign. From Marketing Sherpas extensive library of free case studies at Marketing sherpa.com. That's marketing rpa.com and.