Core Data Principles That Drive Business Value

In this episode of Data Analytics Chat, we welcome Shourabh Mukherji, a data and analytics expert with 25 years of experience at major firms such as Chubb and JPMorgan Chase. 

Shourabh shares his journey from starting a small enterprise in India to working with senior executives in the United States. The discussion examines his career milestones, key challenges, and the vital leadership skills required in the industry. 

The data topic explores essential data management principles, the importance of aligning data strategies with business objectives, and the role of AI and new technologies in modern enterprises. Shourabh emphasises the significance of building strong data foundations and fostering empathy within teams to drive successful data-driven decisions.

00:00 A Defining Moment: Leap of Faith
01:19 Introduction to Data Analytics Chat
02:08 Shourabh Mukherji's Career Journey
07:07 Challenges and Defining Moments
09:47 Building Trust and Stakeholder Management
15:57 Key Skills for Leadership
22:32 Core Data Management Principles
25:38 Understanding Business Needs Before Data Management
26:42 Centralised vs. Embedded Data Quality Ownership
28:09 The Importance of Needs Analysis in Data Management
31:54 Identifying and Prioritising Business Needs
34:15 Building a Strong Data Solution Framework
40:05 Effective Data Distribution for Decision Making
42:53 Aligning Business Objectives with Data Initiatives
44:42 Evolving Data Management Strategies with AI

Thank you for listening!

Transcript

shourabh mukherji: 0:00

It was a defining moment for me a leap of faith. My journey with data, having my small enterprise started in India and that time I had a team of 11 people working. The defining moment was, that I took it on myself as a challenge and even more so working together as a team of experts and educating each other skilling up each other so that we can. Make it work, make it a win. Everybody wants to feel cherished and important. So one is people skills, but from a personal side, and, if I'm addressing this to even the newcomers to this industry, I would say be bold. There is a merit to being bold. And if you are, if you have an opinion and you feel confident about it, then speak up. Because if you don't have a strong why even the latest technologies will fall short. So basically what I'm trying to say is, bring the why and the what before the when and how.

ben parker: 1:20

Welcome to Data Analytics Chat, the podcast where we discuss the world of data, AI and the careers shaping it. Today my guest is Sarah Mukherjee data and analytics expert and Pfizer to the C-Suite with corporate expertise with the likes of Char and JP Morgan Chase. In today's episode, we'll explore his career journey, the challenges that have shaped him, the in insights he has for anyone looking to grow their own career. And also we'll be looking at the data topic on core data management principles. Sarah, welcome to the podcast.

shourabh mukherji: 1:56

Thank you so much. I look forward to this podcast, Ben.

ben parker: 2:01

Brilliant. It's a pleasure to have you on. So I guess, obviously you've worked for some Yeah. Leading companies and obviously doing fantastic work. Do you, do you wanna give the listeners a share your career journey?

shourabh mukherji: 2:12

Certainly. So I've been in the space of data and analytics for 25 years now. By education. I have a bachelor's in electronics and then, computer software back when the dotcom boom happened. So I did my master's in software engineering at that time, me and some of my batchmates. We decided to, explore the journey of data warehousing in it's very nascent phase. And we were using. Crystal Reports and that time Visual Basic. Sql, Oracle, these kind of technologies. And I started actually working with manufacturing companies, helping these firms as well as their distributors on deciding on things like chemical dosages and industrial applications. So they were kind of decision systems meant for industrial operations and, then I moved to us and got into consulting and that brought me to financial services where I worked with some of the largest banks and helped them build data warehouse platforms ground up. As well as reports for various financial instruments, hedge funds capital markets, so and so forth. That was a good I would say nine to 10 years in my career where I worked with banks. In 2008 I was called upon. To undertake a large data warehousing operation for a merger and acquisition. And that's how I got my hands on with how an m and a situation works. Then fast forward 2015, I got into insurance in large insurance, PNC company, which was yet another m and. In 2015, so kind of history repeats itself. And what I did was basically data migration, consolidation, building warehouse master data management, which is MDM data quality, the six verticals of data management. That really expanded my horizon with data data science and data expertise. And towards 2023 after having large number of years in financial services, banks, and insurance I decided to impart some of my knowledge, skills and experience to the broader industry. And that's where I am now working with senior Executive C-Suite. Be that CEOs or CIOs in helping them modernize their data ecosystems.

ben parker: 5:15

Brilliant. Quite a journey and yeah, no matter what industry you're in, it's everyone's got the same got similar problems, haven't they?

shourabh mukherji: 5:23

that's right. That's right.

ben parker: 5:25

Cool. And how, how did you find, obviously the move from obviously one side of the world to the other.

shourabh mukherji: 5:32

You mean from my roots in India to us. Okay. It was a defining moment for me a leap of faith. My journey with data, having my small enterprise started in India and that time I had a team of 11 people working. We basically batchmates. And then the industry took a turn and and everybody dispersed and I said I should move on as well and get onto the next level and had the opportunity to come to us. I did not have that many coordinates as to what I would be doing in US, and that's why I mentioned Leap of Faith. So I took that leap of faith and, came here to New Jersey, we're still at and started looking for projects and then one led to the other. And slowly and steadily I established myself. So it was an interesting journey, although it was not an easy one.

ben parker: 6:37

Yeah, no, I think you'll see, I think nowadays, I think people, obviously you've got social media and everything's like saying it's an easy ride when it's, like you said, sometimes you just gotta take that leap of faith, haven't you? And just go with it. I think it's, yeah, you've gotta have that risk in you haven't you, basically to go with the flow.

shourabh mukherji: 6:57

Yes, absolutely. And that is so vital. That is very important in life.

ben parker: 7:03

So obviously that sounds, that's obviously quite a defining moment in your life. So has there been any defining moments that changed your career? And I guess what would they be?

shourabh mukherji: 7:13

Yes. I was when I came to us, I was working on a much smaller initiative, which was already underway. And I was new to, I was new to banking and I joined a large bank. This was in. End of 2005, towards 2006. When I joined the team, there was no data warehouse. And there were no, analytical reports within that group. And basically the team composition was mostly Java and, some level of Unix developers. So I had to basically roll up my sleeves and, build the entire platform, soup to nuts. Right from architecture, data modeling ETL, bringing the data from different sources. Running the data quality and then get it onto report a reporting layer where I had to build data marts. And that time we were using business objects for the bi. So all of this was new and more so it was new for me. To have started my life in a bank investment bank. That too where the reports are primarily hedge fund reports, so very high visibility. And I had to present it to the senior vice president and in fact even the chief Investment Officer. So a lot of responsibility on my shoulders. And I was probably in my twenties the defining moment was, that I took it on myself as a challenge and even more so working together as a team of experts and educating each other skilling up each other so that we can. Make it work, make it a win. And it's a win for the team, not just for me. I think that, to me was a defining moment and it has certainly helped me in the long

ben parker: 9:13

It's obviously massive because get, obviously getting buy-in from stakeholders and obviously you are earning your career as well. It's is, it is tough, isn't it, for businesses to actually, it is. Obviously these projects are not cheap, are they? There's expense to be put out to deliver these projects.

shourabh mukherji: 9:31

That is true. That is true. And what I learned through this experience was obviously this is like a massive investment. We are talking about hedge funds. There, there is a. Certain synergy and trust that needs to be built with stakeholders. How does a senior vice president who has been in the industry for 30, 35 years trust a 20-year-old who may be, doing the right things on such a high visibility. So the trust factor is. Extremely important, and that's where the people skills communication, presentation and the right amount of authority is very important.

ben parker: 10:13

Yeah. No, I agree. So how, obviously you've come from obviously I've moved from England to Spain, so it's not as far as you, and it's obviously different cultures. So how did you obviously get that stakeholder management skills? How did you learn that?'cause I guess I see when people are technical expert that move into leadership, that's the big step, isn't it? Is being able to. I guess build, like you said, build that trust with people on board. I.

shourabh mukherji: 10:39

A lot of. Lot of times and even today, I was not born and raised in us. So I, I have differences in mindset, difference in communication, style, accent, all of that. That is true. But there is a common language that threads all of this together. And I've worked with more than 40 countries so far, and that common language is empathy. When I started off and working with the large banks and pretty much every, most of the folks on the floor were older to me. Initially they would look at this new kid on the block and, just starting off. And sometimes they would brush off okay, we don't have time, or come later or ask this person or that person. And that was pretty frequent and I understood that. And especially if you're working in a m and a situation where the uncertainties are very high, it's very hard. And that is where what I would do is I would. I would forget technology for a moment and I would go and speak to the person as a human, to a human right and say, Hey then we get to know a little bit about their family, maybe their kids or their grandkids, what they do or they into sports or their hobbies and stuff like that. And, that's how I would roll the conversations. And be genuinely interested about it. And then, next time I meet them I would go back on that and say, Hey, how did that work out? I saw magically a lot, many times that empathy factor played in and, it transcends the differences in mindsets and languages and cultural background because as human beings, everybody wants to feel cherished and important. So I still hold that in me today no matter what project or where it is or, what kind of industry it is in. Empathy is extremely important.

ben parker: 12:40

Yeah, no, definitely. Everyone's equal. It's, again, you need to Yeah. Treat people how you wanna be treated, isn't it really? Ev everyone's on their own got their own journey, their challenges. It's, you have to, I guess you've gotta be aware and obviously. And help people as when you can.

shourabh mukherji: 12:56

That is correct. Yes.

ben parker: 12:58

Brilliant. So what I mean, what were, was there a toughest challenge you faced along the way that stands out?

shourabh mukherji: 13:05

Again it goes back to a combination of uncertainties and then of course people skills. Now, with technology evolved and matured, a lot of things have changed, but I'm talking like back 20, 25 years most of the systems were mainframes or, legacy systems, if you may which are very much adhered and controlled by, people of demographics who have been in the industry. Much longer than I did. So I did not understand those systems to start with. So secondly, I had to abstract that information to a warehouse. So that was one challenge. Secondly, every industry I worked, even though if it was within banking, we're talking about hedge funds or credit risk private equity, various instruments or even let's just say retail banking. I had to learn the domain very quickly because. In the world of data, it's not just technology it's be able to understand the business language and then translate that into a deliverable. So every time I've entered a project or a particular industry, I had to learn the domain very quickly. And for that to happen, I had to be a learner, be a student every time, have a learning mindset. So that, that, that was useful. And then of course, the challenges of, the. How maybe things were coded previously or you had to transition from system A to system B. There are complexities of that. And as I mentioned, I've worked through two mergers and acquisitions. These are very large m and as I'm talking about the order of, close to$30 billion multitudes of systems on each side that had to be consolidated and migrated or sunset. We are talking about interacting with two very different cultures of companies coming together. So that's where the people skill mattered. So a combination of all this, I would say, these are tough. Tough challenges. And it's not just me alone. It takes a team, it takes an army that's aligned and is inspired to achieve. Yeah.

ben parker: 15:28

Brilliant. Yeah, no, I guess it's, there's a lot you, with technology you use a lot of challenges, isn't there? Like again, it is people, you've gotta change management. You've got, like I said, there's so much that's going on and it's, you have to just. At the time is, it's tough. But then obviously, probably now when you look back at it, you think you, you're happy, you're glad you've gone through that.'cause I think in myself, there's no pain, no gain. Like you have to go through it.

shourabh mukherji: 15:55

Correct, correct.

ben parker: 15:57

So what key skills have helped you progress into leadership, do you think?

shourabh mukherji: 16:03

I would certainly say, and we talked about empathy, that is very important in terms of people skill. I, in my 25 years in, in technology I would say that rate people skills at the top, that without that. It doesn't matter what technology things are not gonna be smooth. So one is people skills, but from a personal side, and, if I'm addressing this to even the newcomers to this industry, I would say be bold. There is a merit to being bold. And if you are, if you have an opinion and you feel confident about it, then speak up. A lot of times, maybe because of communication challenges or, seniority or what have you, people don't speak up because they, they're hesitant or may fear that, they get stood out in the meetings. But if your opinion is the right one in the room, then take that bold step and communicate it. I would certainly say that's very important because ultimately we are talking about software, which is machine. Machine will not have a judgment on who spoke in the room, but it'll certainly result in an outcome. It is important to voice, voice your opinion or your contribution to that particular project. The other thing I would say is and this is from my personal experience, is patience plays a big role. Sometimes, there are discussions and is not evolving and there is things are still, going along then one must definitely try to find a solution, but not be in a hurry to take things on a personal stride, but have that kind of patience, but balance it with acumen. And with agile. Coming in the last couple of years, there is this concept of fail fast. And I believe in that. If we can prototype something that we can agree to, and if it does not work, it's better to fail fast than to. Not speak up and take it till the end and then fail. So the cost is much higher if we fail towards the end than failing right away. So I certainly, I'm an advocate of the fail fast theory, and I tell my team and people I work with is, if you have an idea. feel that it's the right idea, let's just put it to test. And so what if it does not work? And we failed, but it's better. We failed now than we failed after six months, right? Because if we fail now it would be taken as an experiment, as a proof of concept. But if we fail after six months to a year, it would be considered a disaster, right? So that's one thing. And then, I have, I've gone through my 25 years and I've just repeating myself. It took some, hard knocks, so learn to deal with hardship and there's always a sacrifice. There's something gotta give. As somebody is planning to go up the ladder or, advance in their career or even in life but there is hard work involved. It's not a cakewalk. So there's hard work and there's always a sacrifice. So one has to. has to be ready and say, okay, what is it that I'm going to sacrifice to make it to the next level?

ben parker: 19:42

Yeah. No, I agree. I think with any job or anything you wanna achieve, it is always a sacrifice. No matter what you do, it's just whe whether you are willing to do it. And yeah. You mate. You obviously made a interesting point about getting people to speak up and I guess it's a leader that. Because you're gonna obviously deal with people that are super confident, others more obviously, more shy. How do you create that like collaborative environment where sort of any answer isn't a good answer?

shourabh mukherji: 20:13

Yeah, and that happens all the time, there are always a composition of people and various behavioral types. Some folks are more verbose and some of them like to monopolize the conversations. While there are people who, who are more introspective and would wait out and then. I would say the common denominator is it, it's a trade off from a behavioral aspect. One is speaking too much and just putting your viewpoint all the time and this kind of overshadowing what that actually results in is a loss of trust and. Maybe even respect in the person because then the others don't get to speak up. They don't get to share their views, so that's I don't see that as a strength really. On the other hand, people who are quiet and don't speak up and just keep to themselves that's a missed opportunity of voicing your opinion. While, I would say that the person who speaks a lot should mitigate that, but as well, the person who is, very shy or quiet I would encourage them to have their, have them speak in the meetings. But the best way to evaluate both sides of the scale, as I mentioned before, is. To really put things in practice through a proof of concept maybe, and say, okay if you have this, opinion about how we should go about it. I would go around the room and say, Hey, do you, does everybody agree to this? Or if not speak up now and let's understand. And if it's a decision to move forward, I would step into doing a prototype or a proof of concept and then see if it works. It doesn't work. And if it doesn't work, then I would actually challenge the person who voiced it saying, okay, it didn't work. What's your plan B? While communication is important, there's always has to be a ownership to it.

ben parker: 22:26

Yeah, I agree. And that's obviously some great advice there. So we'll move on to the data topic. So we're gonna look at core data management principles. So when we talk about core data management principles, which ones are essential in your opinion to guarantee data quality from a business perspective?

shourabh mukherji: 22:47

Yeah. So the answer is actually in the question, and that is business perspective. My experience has been this and I've been pretty vocal about this statement that the data actually belongs to business, right? They are the owners of data. They're the ones who create it. And ultimately it is business who has to carry on with that data, good or bad. So it is actually the business perspective. It is the voice of the business that needs to be very clearly understood. Prioritized, right? As technology folks, we often forget that. We think, oh, okay, let's just take the data and run it through data quality and figure out the gaps and then throw it back to business and say, okay, you guy, you guys have this, good or bad quality of data. But I think that's a very myopic view. First, it's important to understand what is a business perspective to data. What are they gonna do with it? And why are they reaching out to it? That means there is, they have a need for this data to maybe use it further or, related with other data sets for business analysis, which could be for revenue or, profit analysis or sales or whatever. So that, that is a good starting point. And then from a business perspective, understand what are the business rules to that data. I'll give you an example to substantiate this. In one of the very large insurance companies I was working with, we got the data from business and this was basically insurance risk ratings, right? With each geography, there are risk ratings, okay. Within us. So when we got the data and we looked at the address and we said some of these addresses don't look right. So we scrubbed the data, we standardized the addresses, and we got it down to the, gold standard of United States Postal. Service addresses and we said, Hey, we did a great job. We cleaned up the data with data quality and we went back to business and business said what you guys have done, you guys have shifted our risk ratings because you moved our zip codes around and this is gonna impact our business. We cannot go with. You may have done data quality on address, but it's actually harming business more than doing good. So we failed to take a business perspective early on. And that is the point I'm trying to drive here is before we get into any kind of data management or data quality or MDM, need to first gather the information from business. What do they want to do? Why? Why are they doing this? What do they wanna do with it? What are the business rules? What are the boundary conditions? What should be done and should not be done? And then from there, do a data quality. Within the permissible bounds and go back to business and review with them and said, okay, this is a before and after. Do you perceive or do you believe that this quality of data is significantly better and is going to contribute to an improvement in whatever the business wants to do? If that answer is yes, then we have achieved a data quality.

ben parker: 26:37

Okay. Interesting. So then would it be, in your opinion. Does it work better when data quality is owned centrally or is more embedded with business?

shourabh mukherji: 26:51

So data management as a practice, and this is where a lot of the chief data officers, it's a new role in the industry and the chief data officer role is in between a business function and an IT function. It's not very clear cut. The data belongs to business. In many respects, data management is more it is more business facing than really technology facing, Technology is the how piece, but the why and the what is the business side of it. So back to your question, if you wanna look at data quality or data management as a practice ideally, it should be at an enterprise level so that the consumption and distribution of the quality data is at one place. But sometimes, especially with larger companies and companies that are multi geography for compliance and regulation reasons, or for audit purposes, you cannot have the entire enterprise set its data quality at one place. And so there, there could be various factors to that. Yeah.

ben parker: 28:07

Okay, brilliant. So look at needs analysis. And obviously it can be, I guess like a technical exercise. So how, how should businesses reframe it to ensure like they're actually solving the right problems?

shourabh mukherji: 28:24

Yeah, so needs analysis is very critical, right? And. It is a language between business and the technology folks really needs analysis is stems from design thinking. It's a part of the design thinking paradigm. So I break break down needs analysis in three parts. One is the stated need, which is to say, we are gathering requirements building use cases. And the business says, Hey, I need this kind of report for my customer segmentation or customer journey, or doing a customer 360. And I, and we are looking for. These kind of data attributes and so on and so forth. So that's pretty clear. That can be documented and well understood. So those are stated needs. The second level is what I call is implied needs. Implied needs are, what would be taken for granted, right? But sometimes not stated. For example, business is expecting reports. But it is implied that these reports should, they're probably real time and, the performance should be like under two seconds, right? But it's not stated. Maybe it's part of a NFR or some kind of a technical spec, but it's, it is not been stated. It's implied. So it is always good to. Figure out what could be implied and convert those implied needs into stated needs. And then the third, which is, the most difficult and challenging are what I call the hidden needs. Hidden needs could be, it could be psychological, it could be political, it could be compliance and regulation related. It could be sentimental. So these are like, most likely these are human psychology related, right? To give you an example. Now going back to the reports, right? We might do a report, which is talking about sales performance. And, the company has their sales for sales staff, across regions or across businesses. And then maybe the report is being viewed by the chief sales officer or the chief marketing officer. But there could be certain criteria where, you know, for a particular group of. Sales or specific individuals there need to be considerations and it cannot be, everything could not be one and zeros, right? It's not always a logical output. Sometimes because we are all humans. Here, there is always that gray area. There's always that. Non-linear zone, which I like to call, that's where these hidden needs stem from. So these are difficult to figure out, but more times than not, these hidden needs surface. When there's, let's say a UAT or a true production happens, and that time we said no, we can't do this. We need to, we need to remove this criteria or we need to mitigate it through some kind of special conditions. So anyways, long story short, that's how I look at, how we can derive various needs through the need analysis.

ben parker: 31:54

Obviously, do you get times where stakeholders come to you with something, a need? They think they need, but it's not actually what they need?

shourabh mukherji: 32:04

Yes, that happens. That happens quite often. So one way to, besides the three categorizations of needs it is also very important to rank. The needs in terms of priorities and urgencies, right? So business may say, I need these 50 things. Okay, so out of these 50, which ones are most important, most priority in terms of business impacts, and which ones do you need? On immediate basis versus can wait for two to three months or maybe even six months. So I always like to, put this kind of a template on the needs in terms of priority and urgency. And then once that is set, what I like to do is look at it from a feasibility standpoint, right? Like from a technology feasibility, right? If something is, for example, if there is a need, which is not important and can wait, but it's highly complex, I would not take that up first if possible. I would rather take something that's. Urgent and important business and has a low to medium complexity, which are quick wins. And a lot of times I have seen, including in my approach in projects that we miss that because to show the quick wins that impact business that are important to business, that builds trust and that builds a tempo and that, that, that creates a good, a good perception from the start, you know otherwise, as technologies we get into the complexities of code and everybody's doing their things. Sometimes it becomes a tunnel vision and after six months. We deliver something, which wasn't as critically important, but it was heavily complicated. But business now says, Hey, I needed that other thing, which I'm already delayed now. And it's in the red. So that's what we wanna avoid.

ben parker: 34:14

Okay, perfect. And then so what does a strong data solution framework look like?

shourabh mukherji: 34:21

So

ben parker: 34:22

I.

shourabh mukherji: 34:24

I have been, blessed in a way and also practice this as much possible, I believe in frameworks, right? If one can come up with a framework, in this case, let's just say it's a data framework that can be reusable, scalable, and simple to understand. And I think the third. Point is important, simple to understand because it's got value. Then, it becomes a kind of a cookie cutter model, right? As rinse and repeat, can, and that is where, I learned the art of data warehousing, how to build those kind of frameworks. Especially for enterprise warehousing or building data mods. And after a couple of years, it becomes second nature, right? So you understand, okay, there is going to be a concept of a product or a region or there is a temporal view to it and transactions and all of that. But you can. You can pretty much put it into a framework. And if we can't have that kind of a mindset, that is enormously useful, right? The same thing, I worked on MDM and we took some time to build those frameworks. We had to rehaul the way we look at. The data model per se, but it's all hinged on business definitions again, right? What is it that the business wanted? But that drove a lot of understanding of frameworks. And then we could scale this framework from, one hub to say 10 hubs in four years, which was impossible to do, right? Even now. When we are talking about AI or, the other emerging technologies, the concept of framework is very important. Otherwise, what happens is everybody's doing their own things. A seasoned developer has a different style of building a particular solution versus somebody coming out of college or maybe the senior person retired and things are just unattended. If we don't have a framework, then you know that institutional knowledge is lost and things get very murky. But if there's a solid framework which is rinse and repeat and understand. Somebody coming in new or maybe even, joining in new in the industry can understand this framework because it's simple to understand and there is, and it's well, drafted, documented then that solution can live for much longer than a random ad hoc kind of a solution.

ben parker: 37:09

Interesting. Yeah. So do you think a lot of businesses get this wrong in terms of the architecture and. The framework in general.

shourabh mukherji: 37:19

Yes. It certainly starts with, right from the architecture, right? How are you solutioning from a solution architecture standpoint. I'll give you an example in terms of building a framework in any industry for that matter, there are various systems, various lines of businesses, maybe even various regions to, in that if it's large company, so the definition of a customer. Different in each department. It could be different by lines of business. For example if I speak of insurance one of the companies I was working with the definition of customer could have been a policy holder, but the claims department, for them, the definition of customer is a claimant. For the marketing guys, the definition of customer could be a prospect. Or for a broker or for a distribution department, the customer could be the broker dealer, right? So everybody is defining customer differently. if we can come up with a common business definition model where we can have. Whether it's an individual or an organization defined as a role to that customer, then that role defines, how we view that entity, right? So then that becomes part of the framework. Now, what that does is it brings in scalability. So then this framework can be used by the broker dealers because the role is a broker dealer, although it could be the same. Sure. But he's a broker dealer in that department, or Sure. Could be a claimant in the claims department or, sure. Up could be a prospect for marketing, but sure is. So I as the entity, have not changed. So that is a good starting point. Then having my name as in 20 different ways and 20 different systems, right? So that is where, we build in this common framework, a common definition, and then from there, build it out into a solution architecture, bringing in data management and, mastering the data and all the good stuff and maybe have the processes for. Ingestion and distribution be that batch or real time and what have you. But we paved the foundation in a way that this solution can work in us and it can also work in Singapore, right? Because the foundations are solid, right? So that's the kind of framework.

ben parker: 40:03

Okay. Brilliant. And then I guess, what are the most effective ways to distribute data? So it's not just collected, but actually used to drive decisions?

shourabh mukherji: 40:13

Yeah, that's a good question. So I often ask the business, because that's the origination point. They're the ones who own the data and who are making the request. And the question is about why? Like why do you wanna do this? The why is very important because that would lead the business to answer the question, how do they want to use this data? Like, why do they want to do it? And what would be the end? What would be the north star, the end result, right? So with that, why. We get a better appreciation of, what do they want to do with it. And what business drivers are, as you said, what would drive those decisions? Because if they don't have a strong why and it's not often done, but I strongly advise it. The technology folks should actually challenge that. And I'll give you an example. Most times you would see business, they're very happy using spreadsheets and Excel has been there for, I don't know what, 35, 40 years now, right? It's I call it the most durable. Tool there is, right? A simple spreadsheet. It's not fancy, but it's been used for the last 40 years, right? And business loves it, right? So I always like to ask business this question, why do you wanna move out of spreadsheets? And they gotta give a compelling answer. And maybe it is because there are too many of them. It's all over the place. They want to unify the data, they want to look at it at one place, right? But let business come out with those answers. Like they might say, I want to do financial analysis. I want to do, customer segmentation, what have you. Once those answers can be elicit from. From business, then we have a strong case because there is a buy-in versus, some IT guy, sold to business that hey, this is the best shiny toy that you must have. And because business got sold into it I don't think that's gonna cut it. Because if you don't have a strong why even the latest technologies will fall short.

ben parker: 42:45

Okay. Interesting. So then I guess ha with business then. So you've got leaders. If they're struggling to, connect business objectives with data initiatives, how can you align across stakeholders?

shourabh mukherji: 43:01

Yeah it aligns to what we discussed. I would say build a picture together. The business stakeholders and it stakeholders business obviously has a why to start with, and they have a vision, like what the end picture. What does the picture look like? It's almost like manifestation, right? They want to do some. Let's say analytics or, digital transformation and they have some idea about it. They may not be able to articulate to the dot, but they have some ideas about it, right? So when they have those ideas and those desires that is where I think we can build the picture together and it folks can come in and say, okay. You wanna run these kind of analytics well we have done something similar. We can show it to you or, demo, maybe do a quick PO and make that picture more aligned, right? Which, in, in the architecture, if you look at it from an architecture angle, we call it the target state architecture. So in the target state architecture, you say, okay, this is your current state and this is where you wanna go from point A to point B. And in your point, this is how your world is gonna look like. So that buy-in. Is very important. So basically what I'm trying to say is, bring the why and the what before the when and how. Right?

ben parker: 44:28

Interesting. And again, just it's all, all about, again, you've touched on it before, is building up, being able to build up the trust, isn't it, between the teams.

shourabh mukherji: 44:37

Absolutely. Very important. Yes.

ben parker: 44:40

Okay, cool cool. So let's move on to the final question. So with AI automation and sort of the, new data tools changing the way businesses are run how should businesses evolve their data management strategies to stay ahead?

shourabh mukherji: 44:57

Yeah, I might sound a bit controversial here because everybody seems to be on the ai AI boat nowadays. But the truth is that the success of ai truly in terms of, production, deployment of ai. As well as the usage of AI or gen ai, what have you. The percentage does not look very encouraging. And it's probably improving, which is a good thing, but a lot of it is also hype. And what I would say is a combination of what we have discussed. Number one before we even talk about ai we really need to understand from a business lens. why, what is their why, right? That's the most important. Like do the needs analysis, understand what does their end picture look what does a picture look like? The target state, right? And in that. To reach that target state most certainly data is involved. So how does, how is the data looking today? Is it's probably fragmented all over the place. The quality of data is not good and all of that. So before we even look at ai, there has to be a very strong foundation on data. And that is where, the data management comes in, consolidate the data, cleanse the data, master the data, curate it, have a data stewardship, have a governance charter, right? Make those kind of data policies, regulatory policies, make the data foundation very strong, right? Because. We could have ai, but if it has, if it's sitting on a on poor quality of data or lack of data management, then it's almost as good as saying that, I have a beautiful looking cake with beautiful frosting. But the cake underneath is, has gone sour is bad. Who's gonna eat that cake? So that's the analogy, right? So here the foundation. Is that of having very core data principles and good quality of data that's aligned to what business wants and why they want it, right? And from there we look at where AI can played part right now. And again, I'm not discounting AI because I myself am very keen on where AI is heading towards. There are certainly a lot of advantages of using ai, especially for procedural operations. Things that are very, routine that can be automated through ai. Even some level of scripting can be automated through ai. So AI is a very good enabler. But if, let's say, to my earlier point if we are running some kind of a, statistics with ai or, building LLMs and, driving decisions off of ai. If the underlying data is not coherent, if the data quality is poor, then you know the output is going to be very noisy, and that's more harm than good because that's gonna drive wrong business decisions or maybe create more confusion and ultimately the project gets scrapped. Where I see value with AI today is certainly, in terms of and I've been part of some of these projects we are talking about AI driven chat bots and AI assist in, especially in different for let's say education or for back office operations. And then, AI coming in for content management or, video rendering, a lot of that is a good thing. But when it comes to tier one business critical decision making I would say in my humble opinion we cannot discount data management.

ben parker: 49:09

So I guess in essence, it's not about having the latest technologies, again, it's getting the foundations right so people and processes are ready to adapt.

shourabh mukherji: 49:20

Yes, that is absolutely accurate.

ben parker: 49:23

Yeah, and I think obviously, I think it's obviously everyone's always pushing for the late, for the tech, new tech, aren't they? And it's, I guess the found data foundations is massively overlooked in business and I think it is. I definitely, there's more noise around these topics happening on the mar, on the marketplace.

shourabh mukherji: 49:45

Correct, correct. Yes.

ben parker: 49:47

Brilliant. Cool. It's been fantastic having you on. Sure. You've provided some amazing insight and yeah, loved your career story and obviously thank you for your time and I obviously wish you the best in your future career.

shourabh mukherji: 50:02

Thank you so much, Ben, and I appreciate it.