fbpx
CDO MATTERS WITH MALCOLM HAWKER

CDO Matters Ep. 11 | Modernize Your Data Strategy with Samir Sharma

November 17, 2022

Episode Overview:

Our understanding of data and its role within a business isn’t the same as it was 20 years ago. As data leaders embark on a modern data strategy, it is important to continually adapt to change while operating with agility. 

Malcolm interviews Samir Sharma, the CEO and Founder of datazuum, a data strategy and analytics consulting firm. Samir and Malcolm discuss many of the more significant challenges facing Chief Data Officers (CDOs) as they pursue their strategic priorities, with Samir providing a trove of useful and highly pragmatic guidance for any CDO looking to meet both short-term and longer-term goals.   

Early in the discussion, they dive into the unhealthy fascination data leaders have on implementing technology without first understanding the business problems it will solve. Malcolm and Samir also share CDOs need to clearly understand their organizations’ existing operating model and their level of data and analytics maturity, as both are critical in defining a data strategy roadmap. 

Samir shares his framework for working with CDOs whose stated strategic goals require greater maturity but are often starting at ground zero — which involves taking an agile and pragmatic approach to testing the organization’s ability to provide short-term business benefits around certain data use cases within specific engagement models.  

The conversation then shifts to the importance of the culture of an organization and how that impacts the ability of a CDO to deliver a data strategy. However, unlike more conventional approaches to delivering a data strategy that puts culture change as a dependency for CDO success, Samir makes a compelling argument to work best within the existing culture and find ways to deliver value and speak in a common language without requiring drastic cultural shifts. He outlines how developing a common language and a clear understanding of expected outcomes is key for allowing CDOs to work within the cultural constraints of a given business. 

Data literacy programs come under the crosshairs of both Samir and Malcolm, where it’s posited that data literacy programs may often be a symptom of larger organizational dysfunctions. Rather than seeing data literacy as the solution, Samir makes a case for a broader focus on having CDOs and their teams more focused on developing knowledge of business processes and where those processes are architected — all with a deep understanding of the importance of data from the ground-up. 

CDOs focused on defining or executing their data strategies will find this episode of CDO Matters particularly useful, especially those frustrated with the speed or effectiveness of their efforts. This episode makes a compelling case for a more agile, iterative and pragmatic approach to a data strategy that removes major dependencies — such as culture change or a focus on literacy — to one focused on adapting to any existing culture and operating model. 

Key Moments

  • [2:06] Top Takeaways from Big Data London Conference & Exhibition 
  • [7:28] Focusing on Business Outcomes and Building Data Technologies 
  • [11:35] Creating a Data Strategy 
  • [14:32] Data Ownership and Domain Retention 
  • [19:10] Business Maturity and Understanding Data Products  
  • [21:10] Assessing Your Data Strategy 
  • [28:44] Remaining Pragmatic and Operating with Agility 
  • [31:31] Maintaining a Business Culture and Preserving Data Values 
  • [36:11] Knowing Your Business Process 
  • [38:06] Deconstructing Data Literacy 
  • [41:10] Breaking Down Your Business Data Model  
  • [42:52] Making Changes to Drive Value 

Key Takeaways

Building Data Technologies (10:07) 

“I think when I look at it, technology is easier than having to put together the notion of scratching your heads and wondering what you are going to do…I think this is just another iteration of the whole marketing movement for tooling…we had all of the various different terms that we were attempting to implement…all of those areas that which we’re trying to improve how we work, but why have we got all this stuff? Because the systems can’t integrate…we need to start with a data-focused view.” — Samir Sharma 

Is Data Ownership Relevant? (14:31) 

“I think it’s okay to talk about ownership from the perspective of an individual application. But when you start talking about domains that are used everywhere, it’s a horrible label…To me, the notion of ownership…not all data is created equally. I think the notion of ownership is misguided.” — Malcolm Hawker 

Launching Your Data Strategy (21:48) 

“There’s got to be a certain amount of standardization…a certain amount of proof of value that you can start to show stakeholders who are going to invest in this thing long-term. I think that’s one thing that many people forget. Before you go out and start thinking about centralization versus decentralization or a factory model or whatever you might want to have, you got to think about use cases…we want to prove value and we want to show how we can do it and want to show that early benefit to stakeholders.” — Samir Sharma 

Promoting an Adaptable Data Culture (35:06) 

“My view around culture is that there is one. We don’t need to disrupt it. What we need to do is get better at engaging with each other. We need to set a foundation of ways of working that will use business language and be able to talk to somebody about the outcomes that they are looking for.” — Samir Sharma 

About the Guest

Samir Sharma is the CEO and Founder of datazuum, a data strategy, and analytics consulting firm. Advising businesses on how to prioritize data activities, identifying growth possibilities and using data to boost revenue and profit. His clients span the UK, Europe and North America while ranging from medium-sized firms to major multi-national corporations. 

Prior to datazuum, Samir worked at Computer Sciences Corporation, Accenture, Christie’s and Vertex Business Services where he led the development of their data and analytics business. 

He writes on all things related to data strategy, roadmap development and how to execute the data strategy where he shares his experiences and lessons learned.  

He is a frequent keynote speaker and hosts the Data Strategy Show podcast, which was named one of the Top 10 Podcasts of 2022, as well as leading Ask Me Anything events with top data executives. 

Episode Links & Resources:

Episode Transcript:

Malcolm Hawker 

Hi, I’m Malcolm Hawker, and this is the CDO Matters podcast. The show where I. Dig deep into the strategic insights, best practices and practical recommendations that modern data leaders need to help their organizations become truly data-driven. Training for thought provoking discussions with data it and business leaders to learn about the CDO matters that are top of mind for today’s chief data officers. Good morning, afternoon or evening, everybody. I am Malcolm Hawker, your host of the CDO Matters Podcast today. I am excited to be joined by Mr Samir. Sharma, who is the CEO. Of data zoom, a data consultancy specializing but not limited to but specializing in the area of data strategy. So I’m I’m excited to talk about. Data strategy today. Anyway, so there nice to. See you thank you for. Joining the podcast today. Like I said, it’s excited to talk about data strategy. On session. But before we get into the heavy, meaty stuff I I know you attended a. In conference last week, the Big Data Conference in. New London with most of your kind. Of your high level takeaways from from that experience. What are some of the two or? Three things that really. Kind of stuck with you coming out of that conference. 

Samir Sharma 

Sure, sure. Well Malcolm firstly, thank you for inviting me on. I am privileged and you know quite this is my second podcast. In the last three weeks and which you know, which is quite cool. So thank you for having me on. My top takeaways are of big data London one. Is that we can all be back and see each other. In in the flesh. Which is quite an extraordinary thing when you think about what we’ve had to endure over the last two hours. Is the other thing I think that that I did see is a lot more vendors coming up with a lot more tools and specifically. Uhm, I was having a conversation with somebody on Friday evening. Actually, a group of us were out after the. Conference and dumb suddenly put a word cloud together. And on that point. 

Malcolm Hawker 

I believe that was what once equated. 

Samir Sharma 

Once a quarter yes, yeah yeah yeah yeah yeah so. From like I’m getting that well. 

Malcolm Hawker 

I loved it, yeah. 

Samir Sharma 

We we were out on Friday night discussing. The Red Cloud that he put together and actually the the top word on that was billed so interestingly enough, when you when you walked around a lot of the vendors, it seems that everybody moved into this sort of build phase. Uhm, but equally. I think you know everybody and and. And the interesting thing. Is there was there was no one that actually stated mesh in any of their taglines? So that was another thing that one picked up, which was quite interesting too. So so I think you know for me I’m I’m very tech agnostic when it comes to data and working with clients. And and you know. Getting their data strategy up and running, but. Actually, what we are. Seeing is that. Our clients experience data strategy. Obviously from a you know we create it for them. We help them. You know surface what their use cases are. We put together their their sort of future operating models and we you know, we we. We look at things like data architecture and so on. Obviously the conceptual level, but you know what, what, what’s the immaturity in our organization? So I think one of the things that I see is the parallel is people are moving into that build phase. Probably trying to get there quick. Yeah, and and vendors are pushing for use. The tools use the tools, use the tools, but actually I think still there’s a lot of people. If you you know there was a data strategy stage and there was huge huge queues getting into that stage, so I think it tells you one. A couple of things, one that that people are still. Some struggling too. Embed and even create in some instances their data strategies and I think the execution is not as easy as people thought it was and just hard, you know, just just getting a bit of tech and and implementing that is probably about 10% of the equation to be fair, and it will only get you so far. It will, you know you’ll probably get value for a certain amount of time less than 12 months, perhaps not even that. But if you don’t know what to do with, you know the actual, uh? The actual data that you you generate store, you know you you you source from external sources. Then I think you know you you’re still missing out on the point that that tech is not fundamentally the final final sort of solution for a data strategy. So I think I think there’s that and I. I think just the proliferation. Of of a lot more vendors. You know on. The floor space. Was quite huge things. I think that I I you know, the other. The other thing that I think was that the trends that we’re seeing are data catalogs. Data observability Again on Kwanzaa list observability was up there and so I I think there’s a lot more about how do we. How do we build a better ecosystem? Thinking about lineages because as you know, data governance struggles with that and as you also know, data quality struggles with or you know things like observability and understanding where the the the defects are in the process. So I think there’s there’s. There’s those types of tools that people are. Are moving towards, which I think you know. Again, is a good way because then we can start to automate some. Of those tricky and difficult things that we couldn’t necessarily do. But I think it you know the the the industry is. Still developing and I I I think we’re finding so many things day in day out that will will help in terms of tooling. But I I. Equally, I see that we’re still seeing a. A need for data strategy in the main to really plot out where we’re heading and and you know where companies are going in terms of. Road maps and in terms of value generation at, you know that that that word value is always bandied around. But really, how do they get business results? What is it that they need to latch onto? In order to drive, you know increased acquisition or you know, increasing revenue, optimizing efficiencies and cost savings, and so on so. I think. You know, we need to remember that those are the kinds of things that we’re all you know challenged as data. As the data people to do as data practice. 

Malcolm Hawker 

Well, it’s interesting because in one post about the word cloud, which I found fascinating. What he what he highlighted? What is what he didn’t see was a focus on outcomes, AKA business value? That yes, we were just talking about, which is arguably the other 90% that is in technology. 

Samir Sharma 

Yes, yes. Yes, correct yeah yeah. 

Malcolm Hawker 

Or you get into people process culture elements of these things but yeah, but but the point here really is, is that. All about vendors. All about capabilities, all about functionality, all them building. Which which for. IT people like me sound fun. 

Samir Sharma 

It’s it’s, yeah great, let’s let’s yeah, let’s build some stuff. 

Malcolm Hawker 

Well, yeah, I like. I like to build things. 

Samir Sharma 

And wow, dazzle everybody, yeah? 

Malcolm Hawker 

Yeah, but the the lack of the focus there on on outcomes potentially is slightly problematic and not surprising. I mean, we’ve never really been focused on outcomes that much, I would argue. When I was Gartner analyst, it was it. Was the number one topic of conversation the the number one driver of program failures? 

Samir Sharma 

Yes, yes. 

Malcolm Hawker 

The number one driver of lack of funding, stakeholder support on and on? It always traced back to value. But I yeah I. 

Samir Sharma 

Absolutely, absolutely. 

Malcolm Hawker 

I find the build motion a little. Curious how much of. That, do you think potentially is a function? In that so many of the technologies or architectural patterns like the mesh and the fabric, maybe even data virtualization. Maybe some of these other technologies? How much of that build do you do you think is? A function of. The fact that a lot of what’s being. Sold you can’t actually buy. 

Samir Sharma 

Right? Meet with me to meet meeting. 

Malcolm Hawker 

Hey, let’s let’s do a mesh, but you can’t go by 1. 

Samir Sharma 

No no. 

Malcolm Hawker 

Even sum up. What would it acknowledged that? 

Samir Sharma 

Well and she did, and she did acknowledge that. There is, there isn’t, there isn’t a. Specific tooling, and I guess that’s. Why she’s now instead? Mode and possibly, you know. 

Malcolm Hawker 

Well, granted, to help go solve for. That I I would I. 

Samir Sharma 

Yeah yeah, yeah. 

Malcolm Hawker 

Would love to talk to her more about this. This notion of what she calls an operating system for for an edge, which I would argue is. 

Samir Sharma 

Miss you. 

Malcolm Hawker 

Only kind of a. Piece of it, which I. Think kind of roughly aligns to perhaps what government would call a data and analytics governance platform, but it’s the this. This semantic layer plus. Quality Plus integration plus brains plus graph AIML. 

Samir Sharma 

Yeah, yeah. 

Malcolm Hawker 

It’s it’s the thing that would kind of be the. Cockpit as it were. 

Samir Sharma 

He he. 

Malcolm Hawker 

I think that’s what she’s working on, but you can’t buy that and you can’t buy a data fabric. 

Samir Sharma 

No you can’t. No no. 

Malcolm Hawker 

So the people who are pushing a fabric narrative out of Gardner potential, especially, uh, Mark Byer, who’s a friend of. Mine are really. Going guy would tell you you can’t buy one. 

Samir Sharma 

Yeah so. 

Malcolm Hawker 

I wonder how much of the notion of going to build is is very much a function. Of hey look. At all this. Cool stuff you can do. 

Samir Sharma 

But you can’t buy it. 

Malcolm Hawker 

So you’re. Going to have to build it. 

Samir Sharma 

You know, I I, I think that’s always. Been the case. Though hasn’t it? True, yeah. You know, we’ve always been in that in the moment of hey, let’s do a data. Program, whoa, OK, let’s go get see what tooling’s out there. You know, but not asking the right question. Why are we doing it? For what reason? Who’s going to benefit? You know, the how is how is fine, but actually ask those questions up front and I think you know when I look at technology is is is easier than having to put together a notion of you know guys scratching heads. What are. We going to do here, you know. And so. Yeah, sorry, go ahead. 

Malcolm Hawker 

No, please go ahead. 

Samir Sharma 

No, I I I. Literally were saying, I think you know this is just another iteration of the whole marketing movement. For for tooling so you know and and it will morph because maybe we had virtualization and then we had no, you know we had node computing and we had all of these various different terms that we were attempting to implement SOA and you know all of all of all of those areas which are trying to improve how we work. But simply if you look back at it. Why have we got all of this stuff? Because their systems can’t integrate. They’re not very good at doing that. You know we’ve layered on. You know corporations have layered on so much on top of it that they never started with a. Let’s say they never started with the data focused view, and when I say data focused view, I say OK, we’re going to implement a CRM system or you’re going to implement an ERP. How do we get our data model sorted before we actually implement these things and why you know why do we need to be forced into adopting a data model? The data model from a software program. You know, so there’s those types you know so so build out what I say. What we always say is build out your business data. Model and then start to. Layer in what you need to. On top of that, but come from a sort of view of the business architecture as opposed to fundamentally, you know, going straight for data architecture and you know. Buying, you know a a platform and and hoping that you know it’s going to give us some value. So I think all of all of those things and look, you know we’ve gone through so many iterations from the data warehouse, the data lake, the data Lake House. Now the data mesh. Centralization to decentralization and you know and it it’s. It’s just like. A you know someone constantly going round a roundabout and you know just coming back to their their own tail and they keep doing that so it’s it’s it’s one it’s you know. I I do see some principles that are good. Uhm, you know, in in the fabric or the mesh or you know equally I I, I think they’re just an evolution because of the you know such tools like Hadoop which were very difficult to to to use. And you know a lot of people. But Hadoop is in as a data strategy and. You know that that didn’t necessarily work, so. I I think we’ve. Got to move away from one thing as well. You know most people. Will Will will put together a data strategy, but when I’ve. Gone in to review the data strategy that somebody has created either with an external company or whatever. It’s more a technical data strategy. And that’s the bit which I think is is is. It’s not necessarily wrong, but they’re missing that whole bit, which is, hey, OK. You know we can put the plumbing in, we can, you know, get a platform we can do that. But actually, what are you know what? What are the things that the business needs? And maybe that’s the bit which you domain specific you can. You can understand a little bit more about what the domains require. And I also equally think that you know the decentralization and getting governance closer to the lines of business actually is probably a good thing and not having this huge big central team with policies and you know hitting you over the head if you don’t do something. But actually if we can. Build in with. You know the. The current tooling that there might be available things like observability, things like lineages. Well, that’s gonna help, isn’t it? And then you can put your policies. Sprinkle those privacy policies, but you know, as near enough to the business as we can and and and I think we also need to sort out the age-old problems of ownership and and stuff like that. You know maybe that’s another, yeah, that noise. 

Malcolm Hawker 

Well, I’ve made a few videos about this stuff you write on LinkedIn. Actually even wrote a research pieces at Gartner. About ownership, I think I called it. Who owns master data? 

Samir Sharma 

Right? 

Malcolm Hawker 

Right, and you know, I think it’s OK to talk about ownership from the perspective of an individual application. But when you start talking about domains that are used everywhere. It’s so really, it’s a. Horrible label right? 

Samir Sharma 

Yeah, yeah. 

Malcolm Hawker 

Because because because ownership of what? 

Samir Sharma 

Yes, yeah. 

Malcolm Hawker 

Right? Like ownership of the creation. OK, maybe ownership of the governance policies that define the rules around creating and updating and reading and and as security archive all that stuff. Well, probably not. ’cause a lot of people are going to have a say in how customer data. Is retained, they won’t define or modeled particularly at higher levels of the organization, so. Yeah that but, but to me the. The notion of ownership I. I know that like if we applied like Adama damn Adama, where you want. To save model, yeah, it’s all about ownership, but. 

Samir Sharma 

Dharma yeah yes. 

Malcolm Hawker 

First, not all data is created equally. Right and the I think the notion of. Ownership is isn’t is a bit misguided, but. Yeah well yeah well yeah. 

Samir Sharma 

I mean, I don’t. Want to go down that rabbit hole? 

Malcolm Hawker 

Oh, it’s uh, it’s Well, it isn’t. It isn’t right, I think. If you if you talk to. Kind of people who live in green in. Governance the notion. Of ownership, Oh yeah, it’s it’s critical you have. 

Samir Sharma 

Yes, Oh no, no. 

Malcolm Hawker 

To define your e-mail, yeah. 

Samir Sharma 

I’m not saying it’s I I’m I’m saying there’s probably different ways to do it and I think the other thing. Is what you know what things like the mesh try attempts to do is to say, well, actually, let’s have a data product owner. And now we’ve gone down to a level which is a data product term. And therefore you. Know I, you know, and and zamac defines data products as a as a pure data set. You know on a node you know, and that’s what we what you’re basically managing in your. Responsible for so. Uhm, but some people. Will will actually say a data? Uh, uh, data product is more than that, you know data product is at the output of an algorithm. A data product could be something like a dashboard, a data product you know could be a data set and and equally all of these will have different types of value streams associated with them. So I think we’ve got to even get down to what. You know what the definition of a data product is possibly, but. If you read. Some Max book. It’s very much. It’s a data set and it’s owned by an individual data product owner in the domain. Team and you know, hence the reason that you go down to the domain specific level. They’re able to them to think about more data sharing parameters with others, which I think is a decent thing. You know, I think we should be be data sharing across internal parts of organisations. It’s it’s clear that other datasets will be. It’s full to someone else, and I’m not saying for any moment. Am I saying that there shouldn’t be? Perhaps a very thin veneer of central data governance? Uhm, that will you know oversee policies and certain types of procedures but but it doesn’t necessarily need a big machine and and a massive program that goes on for, you know, years and years. So I think. I think we do have to think differently in some ways and I’m not I. You know, it might not be the mesh. It might not be fabric it. It might be, you know, I don’t know. It might even go back to the data warehouse. Probably not, but you know. It’s it it. It’s worked, but there are different ways of doing it and I think you know we we need to. Uh, that the age-old conversation about data modeling now as well has has popped its head above the parapet so you know now we’re talking about that. So I I think we need to. We certainly need to discuss these things and not saying right that that yeah, the as you said earlier, there is no way that we can have a a pure mesh right now, absolutely. In a few weeks. 

Malcolm Hawker 

Or maybe we could like you could have a hyper hyper Federated environment where individual data products could be truly. Like hey, that. Meaning I define the policies I define them, honest I. Find the business. Rules I’m the product where I would. Find the business case. I I define. How it’s packaged? Go to market the. Training relating to it. I’m responsible for the literacy out of my data product. And how people use it? You could do very cute. 

Samir Sharma 

Yeah, you you can’t do that. Yeah, yeah. 

Malcolm Hawker 

But then we’re back to, you know, you’re essentially completely breaking. The benefits of centralization. 24 

Samir Sharma 

And and I think this is where we have. To look at the. Maturity of an organization. 

Malcolm Hawker 

Right gonna be more. 

Samir Sharma 

I don’t think we, we, we, we are. You know many, many organisations out there are not ready for any. Of this new. Wave of thinking. They’re just not, and you know that, and I think lots of people know that, so there’s. There’s definitely an acceptance of hey, are we mature enough? To put a mesh or. Fabric, probably not. Maybe if if we’ve got lots of tooling and maybe the fabric. It can be, you know, essentially molded, which I guess is a mesh. Anyway, you know, but you know, I’ll probably get. I’ll probably get thrown out with the bar of saying that. 

Malcolm Hawker 

There’s some Pearl clutching going. 

Samir Sharma 

Yeah exactly yeah. 

Malcolm Hawker 

On right now. Of course, and you know. 

Samir Sharma 

What you know is it. Is a fabric like a mesh as a match, like a fabric? Yeah, right? Well, you know. 

Malcolm Hawker 

Maybe tomorrow. 

Samir Sharma 

Yeah, yeah, exactly so. You know, let let’s not. Get down to those those those levels but. I think it’s you know. I think it’s important for for organizations to understand that maturity and not jump on everything that they can really think of implementing. You know, come from out there, in, in, in, in, the tech landscape because #1. There’s nothing wrong with their hub and spoke model. You know there’s nothing wrong with that, and I think maturity comes with let’s phase our way through. The types of operating models we want to have sometimes. Actually, if if it’s a Greenfield site or if it’s a, you know that there’s a difficulty with perhaps a lot of legacy systems. Maybe centralization is good to begin with. You know, I don’t know it again. You know if there’s no data team. If there’s no real, you know. I mean, some companies do, have, you know, very little. Very lack of resources or skills that that can actually help them to build out. You know, these kinds of structures so dependent on maturity depending where they are depending on the types of capabilities and skills depending on the operating model you know there could be a there. There could be benefits of decentralization. 

Malcolm Hawker 

So what do you recommend you’re you’re? At the table with a. Client, CEO, or maybe a CIO. And you’ve done a little bit of Recon from the perspective of understanding the operating model, understanding the existing culture, understanding the existing maturity levels around governance and data management, and what you find is a gap between where they say they want to go from. From a strategy perspective, we we want to be, you know, way way we want. To drive a Ferrari. 

Samir Sharma 

Yeah yeah, sure yeah yeah yeah yeah. 

Malcolm Hawker 

As it went from a strategy perspective, but what they’ve got is it broken down Honda across operating models across maturity levels. But what do what do you recommend there? Do you? How do you go from A to? You know, from from 60 kilometers an hour to. 

Samir Sharma 

Yes, yeah. So, so there is a there. Is a mechanism by of doing that so you know when we do look as you said in the preamble, absolutely we know where they are. We know what they want to do, where they want to get. To and there’s a. Massive gap, so I think when we start to speak to the data teams. We look at the. Operating models and we look at and we assess. Capabilities and skills. ’cause that’s important. You know you can’t, just, you know, just say right? We’re going to start putting data scientists down into the into the lines of business. That doesn’t make sense, right? So there’s got to be a certain amount of standardization that you need. Yeah, there’s got to be a certain amount of proof of value. There’s that term again that you can start to show stakeholders. Who are going to invest in this thing long term? Yeah, so I think that’s one thing that many people forget before you go out and start thinking about you know centralization versus decentralization, or you know a factory model or whatever you might want to have. You gotta think about, right, OK? We know that there’s some the use cases that are important. You know, we know that they’re attached up to the business strategy and the levers that the CEO needs to to basically work on in order to drive that over. You know, maybe in the in the interim, but broadly horizon scanning three to five years, even longer in some cases so. We want to prove value. And we want to spin up and say this is how we can do it and we want to show that early benefit to stakeholders. Now they’re actually. One of the companies we’re working with. We have decided. Not to embed a a sort of you know, centralized team. Right now they’ve got. They’ve got policies, procedures. And so on, but they actually had a. Well, that actually. They’re still going through the iteration of their data governance program and and we came in after two years when they started to do that. But what we’re actually doing is spinning up something called virtual squads for them. And they’ve got a use case and they’ve got capabilities, but those capabilities are in in particular teams around around the the organization, and essentially we work around that purpose, which is the use case. Delivering that out to a bunch of stakeholders who are actually a part of that squad or commando team or whatever we. Want to call it? And they come together for a certain period of time. You test the parameters, you test the how that works and what you’re doing is you’re testing how you can actually start to to drive out people coming together, and essentially they may we. We may at the end of it fit. The resourcing said now we need to actually have, you know, a a a bunch of people that come together around use cases almost like domain specific. Yeah but. But essentially we’re just testing it, so we want to see if it works and it stands up within the processes that we’re going to embed. The procedures that we’re going to, you know, layout and the ways of working in the engagement framework. So one of the things we also find is that the the the gaps between business and data on it. Teams is normally about engagement. How do we speak to each other? How do we work with each other? How do we? Start to to communicate. It’s our own set of values in terms of cultural values and how do we start to then talk about data at a base level and what you know we need to get and drive out of that then and why the marketing department is X and they, you know, need to drive that out. But again, we pin it up to. What the you? Know the The the CEO wants to drive out in the C-Suite are looking at and the border are driving out for revenue. And so on and so forth. So I think building that and testing it is the first iteration that you need to do. You can’t go all in and say. Right, we’re all. Now we’re you know we’ve got we’re going from zero and. We’re going to implement a hub and spoke model. Out right because you still need some stability from a center and maybe you might start from the central area and then you need to think about how do we start to, you know, sort of move down to those spokes and start to have the the dotted line and and working with those. So I think there’s got to be a testing phase, a proofing phase, a real understanding of why. But how much agility there is in the organization to move? Because we know that they’re immature because we know that they don’t know perhaps how these teams come together, and culturally, they’re all very different. So how do we aspire to one purpose? How do we move? Those teams to really embedding certain principles. And how do we just talk about business? You know, in business language and not in technical or data language or you know, but actually we’re all on the same page. So coming together with a virtual view of it, testing that proving it. And then saying right it works. These are the things these there were some kinks. There were things we need to iron out, but essentially we know it can work and operate. And then we can start to roll that out. So I don’t think there’s a you know, even if you’re going into an organization where they’ve got very mature data, people who have got. All the tools they’ve got all the. The the the access to data and so on. It’s not always so clear that they’ve got the you know they’ve got the right perception with within the business, you know and and within the lines of business and communication. And how do they set up, you know ways of of of of working together, which sounds strange, but still, you know engagement is is is a massive thing. I think that’s the cultural aspect which everybody raves on about data culture, which I don’t fully agree with. I think more so I you know, there is a cultural fabric which needs to be worked in and operated in so. Uh, you know to to to cut it short. I’ve gone on for too long, but. I think the I think it exists on. You know there are many operating models. There are many ways of testing. But I think you’ve. Got to work through what is the best option at that point in time for that organization. You know you’ve got an area to get to, you know you’ve got the view of the world and. And that future state. But you can’t jump to it immediately, so you need to have phased views of that. And sometimes having a centralized or, you know, like we’re doing, you know, spinning up these virtual squads or or literally, you know, in some cases you know the some organisations do have analyst. It’s in their lines of business who are not really aligned to the data folks, so maybe it’s just getting standards and you know making sure that they understand what what’s possible and what’s not and how to onboard data. And you know how to to integrate it into their into their tooling and so on, and then start to maybe work backwards to say. How do we need to start to embed the right principles. So there’s. There’s different, just different ways of doing it. There’s never 11 way of of operating with a company which every you know, I’m sure you’ve seen that. 

Malcolm Hawker 

Oh indeed, so a lot to unpack there. And and I want to I want. To circle back to the culture. 

Samir Sharma 

Right? 

Malcolm Hawker 

Discussion ’cause ’cause I think that’s it. I think that’s important, but. There’s a few things that I just want to tie off on that that I really love. By the way, one. It’s refreshing to hear a strategy consultant. Talk about it. Uh, I in my personal experience, I’ve I’ve had a lot of frustration with. OK, here’s your strategy by. 

Samir Sharma 

Yes yes yeah. 

Malcolm Hawker 

Right? 

Samir Sharma 

There’s history behind that as. 

Malcolm Hawker 

Right just good. 

Samir Sharma 

Well, I’ll tell you that in a minute. 

Malcolm Hawker 

Luck with the implementation. 

Samir Sharma 

Yeah, not bad. 

Malcolm Hawker 

You know where it is one type of of consultancy to do the strategy. That is, you know generally runs from 5 to 700 an hour in and another type of consultancy to actually start putting shovels in the ground, which generally costs a little bit less than that. And and it’s refreshing to hear a business model for most from a data strategy perspective that is divorcing both worlds so that that that to me, is refreshing and comforting. That’s one thing that that that I heard you say. Another thing is I I’m just paraphrasing here, but is really kind of the agile. Be pragmatic, I I heard. 

Samir Sharma 

Think yeah. 

Malcolm Hawker 

I heard pragmatism for sure. 

Samir Sharma 

Yeah, and you know. I will always talk about pragmatism because. I don’t, you know, sometimes I will advocate for being agile, but I think there’s another thing. There’s another word that I use, which is probably the the word agility. 

Malcolm Hawker 

But have agility. 

Samir Sharma 

Have flexibility in the way that you operate, test things, learn, and then be able to move in a nimble fashion. Don’t be wedded to a particular process. But you know, for years, and then realize that it doesn’t work, because then you’ve just wasted so much money. You’ve come to a conclusion that you could have come to way earlier and the CEO is looking, you know, or whoever it is that your sponsor is is looking and saying how much money have you you know costed me now so. That’s that’s a. That’s a tough it’s. A tough gig so. 

Malcolm Hawker 

Three day agile versus little egg at I’ll. 

Samir Sharma 

Yeah, yeah. 

Malcolm Hawker 

Or or agility? Because yeah, I let’s not be confused here that if if I say agile I mean like big A or you’ve got all of a sudden everything. 

Samir Sharma 

  1.  

Malcolm Hawker 

Everything is expressed in user stories and you have. 

Samir Sharma 

Yes yes yes yeah. 

Malcolm Hawker 

Sprint meetings and. You have you. You know you you have. Uhm, scoring in stories. 

Samir Sharma 

More boards than one could imagine. 

Malcolm Hawker 

And yeah yeah, yes, exactly, but maybe. 

Samir Sharma 

Yeah, I mean, yeah. 

Malcolm Hawker 

A part of it. Maybe it was agreed, but but I think that’s a better way to express his agility, pragmatism and agility. 

Samir Sharma 

Yeah, yeah. 

Malcolm Hawker 

And respond, learn, respond, adapt right so so, so. 

Samir Sharma 

Yes yes yes. 

Malcolm Hawker 

I I certainly heard that I, I heard. About making incremental process right, this ties. Back to agility, don’t. Don’t don’t try to boil the ocean. Don’t go fully. 

Samir Sharma 

Correct, yeah? 

Malcolm Hawker 

Tank, but I heard you say is focus on specific use cases like move the needle from a use case perspective. So getting to the notion of culture, you said you’re not a big believer in that I I want to drill down a little bit because what I think I I’m hearing you saying is that culture can be molded and adapted. Through results is that? 

Samir Sharma 

Yeah, yeah, so I think. 

Malcolm Hawker 

What we want to say. 

Samir Sharma 

There’s a couple of things. You know most. Organisations have a culture, have a culture. They have a set of values, they. They operate within a particular context. Yeah, they they. They have a particular set of ways of working. Yeah, they they. They will often have, you know, we’re collaborative where they’re cooperative. Or, you know, we we have a set of values, etc. So we already have that fabric that exists. And on on each of those queues on each of those those levels we have lines of business. And you know one of the lines of business is data. Let’s let’s just say that you know that’s true. Right that you know. It could be. In some cases it could be a shared service. It could be a you know, particular just just another line of business profit center. It could be. You know cost center in some cases. So anyway, it’s a it’s a line of business, right? So we and that line of business. Works within the cultural context. And I think what’s happened over time is. Through several. New fads that have occurred over the years. Things like data, democratization, and. And you know the elements of some of the early ideas around big data and and and decentralization as well in some cases. Is that people have been pushed. Have been given stuff without really being told why. Yeah so. 

Malcolm Hawker 

Here’s your data, yeah? 

Samir Sharma 

And and so. That’s it, right. Democratizes democratized your daytime in Pi was all about that right? So there you go. Democratize it, give everybody you know everybody’s gonna be. 

Malcolm Hawker 

You’re logged into Tableau. 

Samir Sharma 

Yeah, exactly and then what they do is they they look at it and they you. Know build something ’cause they they’ve maybe gone to one or two sessions, but do they need to? And then they’re downloading. Excel, and they’re comfortable back in Excel. OK so. So I I mean. That was, in my view, that’s why we’ve now got this big issue of data literacy right? Because the tool sets have created that the democratization has created that a lot. Of the the. The the push in the industry has been to, you know, talk about visualization as if you know you switch it on and I literally can start building charts. Well, you know it’s not always the case. Yeah, and not everybody needs to do that, so I think I think those have attributed to this issue and this. Idea of data culture so I. I think you know. As as the line of business, just like all other lines of business, they have a they are unable to do certain things for the organization and therefore. It’s it’s incumbent on the data. Professionals, practitioners you know and the teams to work with everybody you know to to work with the business. The other lines of business, what they want. They have to learn to work with it. ’cause that’s another line of business that they have to cooperate with. So everybody in the organization needs to cooperate with each other. This is how organisations are. And now there is an organizational underpinning. There’s an organizational psychology that is actually the way that these these. Lines of business and and their sort of ecosystem. They work together. They live and breathe, yeah, so my view about culture is really there is one we don’t need to disrupt it. What we need to do is it get better at engaging with each other. We need to set a foundation of of ways of working that will. Use business language you know and be able to talk to somebody in the outcomes that they are looking for, rather than saying you know. Uhm, presenting for example a the output of an algorithm in in, you know, in a very raw format talking about outliers and this and that you know and and and and specifically going into you know how they trained it. Does a business person really want to hear that? No, the business person wants to know what is the decision? Like you know with this output and and can you tell me that? And can I then take the action that I need to in order to drive my marketing campaign for more acquisition? Or, you know, reengineering my processes ’cause I need to to do that. And you know, get more more efficient in that. So I think. That is one part of it, and I think. Also, as data people were, you know. We’re not really good at looking at process, and I say that because. I’ve seen it. In so many conditions where if you look at a set of use cases and I’m, you know, walking into an organization where they say right, we need to deliver these things and you’ve got this wonderful documentation laid out. Not all the time, but you. You know you have sort documentation laid out. Then you know the first question I ask is OK. You want to do this, but what’s the underpinning business process that drives this? And then when that questions put out on the table, everybody kind of looks at each other and says, well, that’s the business concern. Well, no, it’s not actually right, you know, so I think that’s the bit which we need to get better at, and therefore is it the illiterate part of that data? I think it’s really just, you know, getting people. Involved with each other. Learning more about you. Know the principles that we work through and you know the the reasons I I’m very passionate about this. ’cause I’ve seen people shouting each other in meetings ’cause they’re not speaking on the same sort of wavelength and just bringing it down to. What is it you want to do? Why do you want to do it? You know who do you want to do it for? And then being able to say right and and you know, finally, when we get to the solution, this is how we’re going. To do it. I don’t need to know technical components, you know? I don’t, but but really I need to understand at every point where I’m going to be needed, how I’m going to be testing that, what I need to learn about the output. Maybe you’re going to test it and say. Work as a hypothesis, but you know, I think there’s no ways of working and you know what we did was to build something called the data Strategy Canvas, which actually helps to have that conversation with each other. You know both from and. I just talk about the triangle of. You know the the sort of IT data and business teams and it’s just a a meaningful conversation focused around business objectives rather than it being, you know, around technical objectives. 

Malcolm Hawker 

Well, and this sounds mostly for a data team. One IT team. This sounds more like a listening exercise. Than a training exercise, yeah, and where I’m kind of going with this is is this notion of of literacy and I really dislike that term. 

Samir Sharma 

Yes, yes. Same here and I I saw you rinse at one month at at one point. 

Malcolm Hawker 

Yeah, it’s it’s it’s, it’s the. It’s the literacy. Term right because? What we’re what we’re. Building here, this is not rocket science. 

Samir Sharma 

No no. 

Malcolm Hawker 

Right? I’m sorry data folks, it’s just not we are trying to provide. 

Samir Sharma 

Yeah, yeah. 

Malcolm Hawker 

Like data or systems or processes or tools that help fuel better decision making and they provide insights that can reduce supplier costs or increase customer experience. And you know, without building rockets, so if they’re listening and you’re you, you are able to translate what you’re hearing. Into actionable, valuable solutions, then the literacy required should be marginal. Right there should be. 

Samir Sharma 

Yes, I agree. 

Malcolm Hawker 

You know there. 

Samir Sharma 

I agree. 

Malcolm Hawker 

There should be like some instruction booklet. Sure, you could have, uh, maybe some groceries or some other things. 

Samir Sharma 

Yeah, yeah. Huh, huh? 

Malcolm Hawker 

But if we have to embark, I’m getting a little philosophical here. But if we have to embark literally on a on a program, which for some companies could be a multimillion dollar program of this, this top down. I’m gonna make you more data literate initiative then I think. 

Samir Sharma 

Easy easy. 

Malcolm Hawker 

Something else must be broken. 

Samir Sharma 

I think so I I think so, and I think. 

Malcolm Hawker 

  1.  

Samir Sharma 

It’s, you know, I think you’re. Right, but let’s 

Malcolm Hawker 

Yeah, right? 

Samir Sharma 

I think you’re right that there. Are many things broken and again, you know it. It does go back to how we work with each other and and you’re right around the philosophical questions that you’re asking. You know, I I, I don’t think there’s anything wrong with it, but I think we’ve got to be pragmatic. Now we’ve got to sort of say hey guys, look, you know this has gone on long enough. Is the well going to dry up? Is is the sponsor or the CEO going to say enough is? Enough, and we’ve been there before, you know so and and and so I think. We’ve got to really. Think about how much we drive out capability wise, but equally you know as we talked about before driving those outcomes driving those results, you know whether it, whether we can attribute, you know to the overall objectives of maybe improving revenue by 5%, I mean one of one of our clients. When we started to work. With them They they were they they gave us a an overall view of what they needed to double margin in in 10 years in increased revenues by 5% in, in, in in six years an American client and. 

Malcolm Hawker 

There’s certain things. 

Samir Sharma 

About culture, which they. Hadn’t got right yet and I think it was more about DVI and. You know, diversity and so on. But when we started talking to them, we said, look, you know one of the things that we. Really like to do is. To break your business down into its business data. Model and and that actually. Really got them interested and it got the business people interested. It got the lines of business really interested. Because they said. You know that’s interesting. No one ever sort of. Come at it like that and you know I’m not saying we’re the only people who do that there’s. Lots of people who do it. So what I’m saying it though is that they had been through so many iterations, so many. Consultancy’s had come in and built stuff and you know delivered it and it wasn’t being used, but it was really going back to the fundamentals, right? How do we build out your your how do we? How do we cater for your business? It’s the context that it’s working in the environment, the marketplace. You know the political sphere, the regulatory aspect. And so on. How do we start to get all of that stuff included? So we’re building your business data model. You know, as things as a foundation, but as things move as markets change as political regulation you know is enforced, how do we then have that agility? And again, it’s about that you know so so and and. Then it’s it, you can. Have those conversations at a lower level around you know the data product and the owner and the business process owner and is the business process owner. You know the person that actually does you know for you know along that business process end to end, own that and the data in it. You know there’s lots of different conversations that you. So I think it’s important for us to think about that designer business that is going to operate within the context of data. So we’re architecting the business with data, which is the way I think is. Perhaps a better way to say it instead of. Talking data models, you know. But anyway. 

Malcolm Hawker 

In my experience, any time you’ve asked the business to fundamentally change whether that is culture, whether that’s technology, whether that’s process. If you as a data leader are asking to do that, you have increased your risk tenfold in your ability to actually drive value versus what I heard you say, which is kind. Of deal with. The you know part of the cards. That you’ve been dealt. To a certain degree, like and and make incremental changes do what you can within the environment that you can give and do it again with the technology. 

Samir Sharma 

Yeah yeah, yeah. 

Malcolm Hawker 

Maybe even processes existing government rules and make slow incremental. Half full. It is I. I love it. So we went from at the beginning of the conversation to talk about kind of technology leading and technology driving to do now to this. To this to more of a perspective of the outcomes driving what you’re trying to achieve. Driving and making incremental process from where you are to where you want to be. I love it, similar Sharma. This has been a fantastic conversation. I know we could keep going. 

Samir Sharma 

Great, I know we could, I. 

Malcolm Hawker 

For a holiday. 

Samir Sharma 

Know, I know. 

Malcolm Hawker 

And we need to absolutely, positively meet at a conference someday at some time. In the future, preferably over. 

Samir Sharma 

Yeah, I’d love that. 

Malcolm Hawker 

Or over applied. We do want to note you are the host of the data strategy show, so most certainly check out singers podcast on your broadcast provider of choice and I know that there will be a fantastic episode upcoming that that that is hosting. 

Samir Sharma 

Yes, I am yes. Thank you Yep, Yep, Yep. 

Malcolm Hawker 

If it is featuring yours truly. Here press. 

Samir Sharma 

Yes, yes there will be. 

Malcolm Hawker 

Ah yes, what is? It what is it called that out? 

Samir Sharma 

Fantastic, of course it will be. 

Malcolm Hawker 

Yes, yes great but again Sir, thank you so much for your response. 

Samir Sharma 

That’s my pleasure. Thank you. 

Malcolm Hawker 

Really appreciated it and thank you for everybody tuning in today to CDO matters and look for another episode shortly. Thanks again. 

Samir Sharma 

Thanks Malcolm bye bye. 

ABOUT THE SHOW

How can today’s Chief Data Officers help their organizations become more data-driven? Join former Gartner analyst Malcolm Hawker as he interviews on thought leaders on all things data management – ranging from data fabrics to blockchain and more — and learns why they matter to today’s CDOs. If you want to dig deep into the CDO Matters that are top-of-mind for today’s modern data leaders, this show is for you.

Malcolm Hawker
Malcolm Hawker is an experienced thought leader in data management and governance and has consulted on thousands of software implementations in his years as a Gartner analyst, architect at Dun & Bradstreet and more. Now as an evangelist for helping companies become truly data-driven, he’s here to help CDOs understand how data can be a competitive advantage.

REGISTER BELOW

MDM vs. MDS graphic
The Profisee website uses cookies to help ensure you have the best experience possible.  Learn more