Data Governance is the cornerstone of every successful data management program, but it is often associated with suffocating rigidity, great cost, and unclear ROI. How can we get it right?
Join hundreds of practitioners and leaders like you with episode insights straight in your inbox.
Checkout our brands or sponsors page to see if you are a match. We publish conversations with industry leaders to help data practitioners maximise the impact of their work.
Your ideas help us create useful and relevant content. Send a private message or rate the show on Apple Podcast or Spotify!
Loris: [00:00:44] So today is all going to be about data governance. And data governance is one of those words that can sound a bit intimidating, rigid, sometimes, expensive. The term kind of gained a bit of a bad reputation in the industry. And sometimes we hear experts say, "Don't say the G word when you're talking to the business."
So today I want to deep dive into this topic to understand what is the perception around the term and explore its role as part of the larger data management program and understand what are really the cognitive and emotional challenges and biases that we should expect when we direct a data governance program in a large organization.
And my guest today, you can see here, is the legendary George Firican. George is the director of data governance and business intelligence at the University of British Columbia. He's a passionate advocate for the importance of data and the creator of LightsOnData.com. George aims to provide fellow data management and data governance professionals with practical content in the form of templates, guides, best practices, articles, videos, courses, and mentorship to address key questions and challenges in the role.
And he started creating content, I think in 2018, 2019. And it was really out of the necessity of structuring the conversation and creating order in the chaos. And I think it's fair to say that George is one of the most brilliant, humble, and friendly data professionals in my network. So I'm really excited to have you on my show, George, welcome to The Data Project.
George: [00:02:28] Thank you so much for having me. It's such a pleasure to be here and to have this conversation with you and I'm humbled by your introduction, thank you so much.
Loris: [00:02:37] I don't know even where to start. I think we have so many topics to discuss, but perhaps, let's take an easy start because it's morning here and it's late in the evening there. How did you start LightsOnData? How did you end up directing the data governance program at the University of British Columbia?
George: [00:02:58] Oh, so two different questions. So LightsOnData started, you know, long after I started my career within data governance. And it's just a way for me to give back. Originally it was actually, LightsOnData was geared towards non-profits, governmental institutions, healthcare, but as in nonprofit healthcare not some of the American healthcare institutions, and education.
So it's really trying to service those industries that don't really benefit from a lot of funding that can go into this type of education tools and that other private industries would be able to benefit from. So that was kind of the niche. But as I was putting content out there, and like you said, it was kind of, try to make some sense out of the chaos. As I was putting content out there, I was getting, I started to get good feedback and I noticed, listen, the content I'm putting out there, it's not just for these industries. It's overarching, everybody's benefiting from it. And a lot more people are confused about it than I originally thought. So in a way, you know, that was a good thing.
I'm not by myself in this bucket. And, it's good that whatever I'm providing there, it's helping people out. So it's just a way for me to give back to the community.
Loris: [00:04:14] Yeah, that's brilliant. And I think I want to thank you personally for that because it's helping me a lot. I think you, especially, I don't even know which resource is more valuable, to be honest, between the website and the YouTube channel. The channel has so many videos and really some of them are really short and sweet.
Some got really deeper. Just yesterday, I refreshed my memory by watching your webinar on data governance, maturity models. Yeah. Which is really good and so much, and the research and you put everything in context, but I think I'm jumping a couple of guns here, so let's, let's take it easy and start with the why. Let's start with the why, right? Like, why the heck should we worry about data governance in the first place?
George: [00:05:02] Oh, man. I think we could just spend the entire time just talking about this topic so feel free to interrupt me whenever. Well, first I think, yeah, we need to understand the why and the what--what is data governance, right? And why do we need it so much, like you said? And if any of you are listening out there that are interested in data governance or have an interest in data governance, you know that as soon as you Google, "what is data governance?", you end up with, you know, hundreds, thousands, tens of thousands of search results.
Most of them containing a definition of data governance. And guess what, there are not two alike. They're all so different, which is in a way in my opinion, quite ironic. Its data governance, that's one of the things that it's trying to achieve--all this consensus, as well. And when it comes to definitions.
So as data professionals, as business professionals, I think we can agree that there's no one definition that's alike with the other. So even to me, to be honest, sometimes it's confusing when I'm talking to somebody else, we're talking about data governance. We first need to understand: well, are we on the same page here?
Because when you're being approached by certain vendors, let's say, you know, they're saying, "Oh, well, we have a data governance product." But then when you go more into it, you find a wall, it's just a data catalog or it's some sort of data security or data classification or so maybe parts of it. So there's a big confusion.
And I'm still kind of trying to figure it out in my head. How do I explain that better? But I'll, you know, I'll try and do that here on your show, but before we kind of pick one of those definitions that we can see out there, I think let's understand what it is.
So without necessarily defining it let's understand what it is. I think first we can agree that data can provide a lot of value and because of that it's an asset, right? And it was Clive Humby that, you know, kind of coined that saying, "data's the new oil." And I know there are certain data professionals that, you know, scream, "No, it's not!" And Scott Taylor, if you're listening to this, I know you're one of those.
And I agree. I mean, it's kind of a, a bad maybe comparison? Because there are so many cases where you would say, "well, no, it's clearly not like oil." But really what he was trying to say is that data brings value that's, you know, and he chose this analogy, maybe not a good one, but, that's, that's the whole message beneath the data's the new oil.
Okay? So data is an asset. It brings value. That's the key point there. Now, what other assets does a company have? Well, there are plenty, but there are two that, kind of, come up to my mind right away. The first one is the employees and the other one is well kind of any, any financial asset.
So for these two assets of financial and the human resources one, we kind of have two disciplines that manage the assets, right? Human resources that help manage the employee life cycle. So recruiting, hiring, onboarding, you know, retaining, creating those workplace policies. And then you have the finance department who's kind of handling those financial assets.
So payroll, accounting, record keeping, conducting financial controls, audits, all that stuff. So like we said, data is another asset. So we need a discipline to help the organization have the necessary strategy, policies, processes, roles, responsibilities that we need to ensure that data is managed as an asset. Are we on the same page so far?
Loris: [00:08:54] I think so. One hundred percent. And I, I feel the pain of that marketing, you know, the marketing buzz? It really doesn't, it doesn't help. So one of the reasons why I've started The Data Project is to bring, to try my best to go beneath the surface and kind of say, okay, what do we actually mean by that? Let's focus on the what. So this is brilliant. And I could detect it, you wanted to say something else about that...
George: [00:09:23] well, yeah, I just wanted to just conclude this piece and, you know, take, like finance, for example, who is managing these financial assets? Who is doing that, well we kind of have the, I don't know the accountants, right? The accountants are also governed by a set of principles and policies that are, you know, they're also checked by auditors, external parties and so forth to make sure that the books are all right, nobody's cooking the books... So the principles of policies and auditing accomplishes for financial assets what data governance accomplishes for data.
Loris: [00:10:00] Yeah, but that's really well, well said, and I, there's a couple of things that come to mind when you say structure, right. And the first one is, sort of the policies and the procedures and sort of the, let's call it pipelines. It's a term that makes sense in my mind because I'm doing a lot of Notion pipelining stuff and data pipelines. But it's, what I mean is the processing, it's what do we do when something happens? And who's responsible for it and what should we focus on? You know, what's the priority? Where should we invest more? What, how do we track the progress? What is progress in the first place?
And the other one is the sort of the infrastructure, the architecture, around it. So, you know, in the case of finance and HR, we know that there's a difference between the type of software you use to track your employees, to track performance reviews and the overall business strategy around the management of those resources.
And so data it's kind of the same, there is a difference between the tooling and the infrastructure that allows you to manage data and make sure it's trusted, and the people focus, behavioral focus bit that must be there otherwise you can have all the tech you want. It's not, you're not going to be able to guarantee that level of availability and trust. Would you agree with this difference? Or do you see them as a one thing? You know, the reason I am asking is because I am confused too specifically in the role of data governance in data management. How does data governance fit into the broader data management picture?
George: [00:11:50] So, so for example, the DAMA organization, they're kind of seeing data governance as part of data management. And they're seeing, so they kind of have this diagram, which looks like, a Bundt cake. You know, the Bundt cake is like this round-shaped cake with like a hole in the middle, each slice of this Bundt cake represents like a data management knowledge area, that's what they call it. And some of these could be data quality management, that's one area. Data warehousing, it's another one. Data security, data architecture and so on.
I think there are about 10 or so different data management areas. And in the middle, you have data governance, and data governance kind of makes sure that all of these data management pieces are working with each other because there's always overlap.
So if we just take a look at data quality, while you need to have some data quality, if you're going to ensure data security. also need to look at the data architecture. If you're going to architect your environment in such a way that would be able to, support good data quality practices. Right? All of these also tie together with the whole business intelligence, which is another, slice of this pie. And so data governance is in the middle. It's like a data management area but has sort of its fingers into everything. Because again, they're putting that structure together.
They're figuring out the roles and responsibilities, what the standards are, what the procedures are. And then off of this, let's say framework, all these different areas are maybe taking those templates and working it for their own expertise.
It's kind of like, again, back to the HR finance example, you know, HR, again, they're the ones that are creating those policies while, you know, these are the grounds to fire somebody type of a thing.
They're not necessarily going to do the firing or the hiring. They can help with that. They're providing the templates, they're providing the help with it, but then you have the individual manager that is still managing their own employees, their own staff by using these guidelines and the templates and tools and standards provided by the HR group. So kind of seeing what the data governance team...
Loris: [00:14:20] Yeah. So, you know, in a way, data governance is like the foundational element of data management and data management is the foundational element of anything that you want to do with data. Like if you, if you want to walk it backward, if you want to do machine learning, advanced analytics, BI, you gotta have data available that you can trust that scale and safe. To do that you need to have an overarching data management program in place.
But data management is useless if you don't govern it well if you don't worry about that people's side, the behavior around it. Would that be a fair statement?
George: [00:14:52] Exactly. And yeah, you know, some people say like, David Mark was kind of saying that these two are two different sides of the same coin and you can't have one without the other. And sometimes whenever you have data management without data governance and some do, they actually might have some data governance elements within each one of those data management programs or projects, but it's not really structured or recognized as its own program. So, and you know, like data quality you can't have good data quality without data governance either.
Loris: [00:15:25] So let's, I want to take the perspective now of the skeptical business leader, the one that goes like, "Oh yeah, you know, we've got, we really spent a ton of money on IT. We have systems all over the place. We do have data, right? That's not a problem for us. Now you're saying that we need to implement this governance thing.
It sounds really complicated and expensive. I wonder what the ROI is. Is it going to even payback in a year, in five years when we have other pressing priorities?" And so when the board needs to decide how to allocate funding, inevitably, you know, if you don't have a strong business case that, and maybe you do because, you know, we both get it, but there are people that don't necessarily see it that way.
So I think it's interesting to explore what happens through the life cycle of a piece of data. Like, you know, from the second that you get access to a spreadsheet or a CSV file, or like some blob of binary information on a disk or in the cloud, right? Let’s call it T-zero, you know, the first instant in time you have access as a single person to that. Then why is this, why is the behavior important? Well, what happens to this piece of, this file over the next six months or a year or five years across thousands of employees?
George: [00:16:48] Well, if there's no governance in place, first of all, it's kind of like the Wild West. And this is what we're seeing in a lot of organizations, a USD employee, you have to do your job. So you get access to this Excel file and you kind of have to take matters in your own hands. And you got to work with it.
A lot of data scientists are doing this because they have no other alternatives, but it's also one of the reasons why they are spending like 80% of their time cleaning it. For starters, right? So, this simply would do that too. So first, they will try and understand what they're looking at and how they could tailor it to their needs. And it would do so, you know, they would do some data cleansing, they would do some data transformations. Maybe they would do some definitions in their head, at least to try and bring that data into the context that they're working with. And they're doing all this work to prep that data for whatever project they need it for. Now multiply that, you know, to another hundred employees that might require the same data file or a similar one, and they replicate the job in a similar fashion, but not the same.
Because of their own context, their own understanding would be different. So they're looking at it, you know, with different eyes and because there are no standards, there's no definition, there's no clear pathway that they could approach and know who to go to ask or what, I don't know, Intranet page they could go and check this against, they will make their own assumptions. They would replicate the same work that their coworker did on transforming that data, cleanse that data, and then bring it into a different context. Even though it's the same piece of data, they might have a different definition for it.
So in their own project, let's say they're putting together some sort of analysis or they're creating a report, even though they're working off the same numbers, they could reach different conclusions. And different, you know, sum totals whatever. So this is like one issue, of course, but it's also that inefficiency that multiple people are really doing the same thing about the same data source.
Loris: [00:19:05] Yeah, and good luck tracking the lineage. You know, what, what does a piece of, like a number, where does it come from?
George: [00:19:11] Yeah. And that, and that's another issue. You have the lineage and the security, privacy concerns of it as well. Absolutely. And especially with GDPR, and that's why kind of data governance, I think came into more of a focus when GDPR came into effect because GDPR was requiring for this to happen. Okay. Well, it's that lineage. If we have that right to be forgotten from the customer point of view, can you actually detect all of the pieces of data sources where you have this customer's data that you can remove?
Loris: [00:19:47] Yeah. And the other side of the coin is the carrot, right? I was reading the other day, this article brilliant article written in 2009 or 2010 by James Price and a co-author here in Australia, they did some qualitative research and they went around.
I'll probably do an episode on this because this deserves it. They went around and asked them, a whole bunch of organizations, C-level what are the impediments to proper management of information and data in your organization? And they talked to CFOs, the CEOs, the CTOs. And what emerged is that the fact that data is intangible, that you cannot see this, and that's why kind of all these analogies break down, right?
The oil analogy the water analogy, all these things are made of matter, atoms that we can touch. We can, and we have an intuition for how they behave. With data, not really, right? Because it's typically, a tech thing, you know, there's a database, you need to know engineering, basics of IT, to be able to manipulate it and access it.
So a normal person in an organization, someone that is just doing the work, dealing with customers, they deal with spreadsheets and the way that I understand it, "Okay, there's a table here, I am in control of this, what's the matter, right? What can possibly go wrong?" And that relationship that you were talking about, the lineage, the dependencies, the "what's going to happen tomorrow", the interpretation around the number, because numbers are just numbers, right?
But the real world is made of people, what does a number really mean? Is that important?
And so I think here we could talk about metadata management, but before we go there, I want to still go back to the “why” in the C-suite. Imagine you are, you're a CEO and you're pretty aware that this is the situation that you want to do something about it. the first step into a data governance program?
George: [00:21:52] Well, I think it's to first determine the driver. So why do you want to invest in data governance? And for different companies, it's different reasons. And there is kind of maybe three different paths as to what the driver is? At least from my experience. The first one is it's the stick, really. It's ensuring that regulatory compliance. And for a lot of them, this is a great incentive. Citibank, they were fined $400 million, I think it was last October, November? Where they actually cited a lack of data governance is one of the reasons why they got the fine, which is like, that's a great business case for a lot of companies that just, "Hey, you know, reference that example, that fine, and, maybe you would have some power to convince your CEOs to invest into it". So that's kind of the stick. If we don't do, if we don't implement this, if we don't invest money into this, we will pay a bigger fine than if we would've invested into it.
Okay. That works for some. The other one maybe is really more of some sort of a, there is a business imperative, so they do want to do better, right? They want to meet their business goals. They want to serve their customers better. They want to provide better, you know, services, better products. I don't know. They want to grow in a particular market. So it's a specific tied to a business goal. I think this is one of the best ones. And these ones then translate into a bunch of other data areas.
So it means that "Hey, we need a better data science program, or we need to invest in this AI to automate certain things. So we would reduce the, you know, manual work and get better results that way, or be just one better reporting to at least see where we stand from an operational point of view because we're relying too much on hunches, not on data".
We don't have the reports to tell us, "where do we stand right now?" This, and this is actually true, with even big companies. So that's sort of the business imperative and the last one kind of ties it all together, but it's really, it's not really a reason. It's maybe the intermediary between all of these, but at the fact that they don't have good data quality. So because they don't have good data quality, they can't meet those regulations, regulatory requirements very well, and they can't meet those business imperatives. So for a lot of data management professionals, this is kind of the first hurdle that they want to approach. Okay. We need to improve the quality of our data. So that based on good data, we could build everything else.
Loris: [00:24:36] Yeah, and there's another aspect that comes to mind here when you talk about efficiency. And I think it's still in the same paper, which I'll think I'll have a link in the description I think Tom Reedman was talking about this figure as well , it's estimated that the average enterprise loses 20% of the yearly salary basically, the average employee because of poor practices in the management of information. And we're not talking about, you know, complicated pipelines. We don't, we're not talking about data ops or, you know, snowflake data warehousing, we're talking about simple rules around how do we handle Excel files? When do we attach it to an email, whether we create a copy or not, . And 20% is a lot of money. And so that made me wonder, first of all, there's a lot of low-hanging fruits here that we can, that we can tackle so that our governance program can be very simple.
It doesn't have to be this boiling the ocean thing that you mentioned, as you mentioned in your videos. And the second is this kind of gives us a scale of how big is the impact of a properly funded data governance and data management program.
And so in marketing, we get in, if you ask the average marketeer, they would tell you, yeah, it's expected of you spend 5% of your revenue for marketing. We get that, you know? But data management and data governance, somehow we don't.
And so I wonder, would that help? And once we secure the funding, what is that first step? Do we hire, do we put together a committee or do we first run those interviews, as you explained in one of your videos, trying to talk, take a slice vertically across the organization and talk to anyone, understanding what is current, you know, real practice in the field and not going through the shallow IT and actually getting to the ground truth? How are people behaving around their data what'd you think? What, how did you do it?
George: [00:26:41] So, let me add to that example first, just with another example. And I cover a lot of this in my business glossary course, into how we can put a business case together on investing in a business glossary. And a business glossary is really this tool that the data governance program would deliver, one of those artifacts that it would deliver. And it's really a, kind of think of it of a dictionary for the company that would have their business terms defined and explained to make sure that there's no assumptions, no ambiguities. If we're talking about, this is what we mean by customer. Well, this is what we mean by customer.
This is how it should be calculated and reports and there's no, you know, confusion about it. And there is a lot of confusion about a bunch of business terms, but anyway, to cut the story short, the IDC, which is the International Data Consortium, I think it stands for, they did a survey a while ago and one of their findings was: they found out for middle to large companies in North America, the average information worker spends about 8.5 hours each week, kind of chasing information and not finding it.
Loris: [00:27:55] That's crazy.
George: [00:27:56] And they defined the information worker as like anybody in a managerial position that needs to access like reports. You know, unless of course not just your business intelligence development team and all that stuff, but anybody that needs to rely on reports, on documents to do their job, they're searching for it. They don't know where to find it and, they're just wasting time just looking for it. Exactly, you know, looking, because they don't have that process, that they need to talk to their colleague because a lot of that information lies in their heads. And it's a lot about networking and personal relationships. And not that that's a bad thing, we need more of that, but we should also have it centrally available to people to consume information.
Loris: [00:28:45] I mean, it kills innovation too, because the speed of thoughts, I mean, there's an estimation that in the brain, the round trip time for a signal, for a spark from a neuron to, so the time that it takes to propagate it and back, it's a hundred milliseconds. So a 10th of a second, that's kind of how fast we think, but if it takes, if it takes you a week to get a bloody Excel file, then you know, that thought is long gone. Frustration is sky-high, you did the whole hormone imbalance is flipped in your system. And so you go from, "I have ideas. We could do stuff" to "I hate this place. I want to get out of it."
I've seen it happen a number of times. So it should be very intuitive. You know, this is an easy case, I suppose, for someone. Do you want to have more time per day? You know, it boils down to how much energy do you want to have left when you got back at home to your kids? Do you want to be flat because you dealt the whole day with impossible, like mazes that you feel like nothing is progressing, or do you want to have everything one click away?
George: [00:29:55] Exactly. Yeah. That frustration is definitely tangible in a lot of companies. And even the time that you spend on it, that's costing you, that's costing that company. It's time that's paid, that's wasted.
Loris: [00:30:14] There's an aspect to that, I think I agree a hundred percent on and is seeing data governance as a change management project, like a new challenge I suppose.
George: [00:30:26] Yup.
Loris: [00:30:27] Gotta be careful not to say project and say program. But a change management challenge. And the importance of focusing on short-term wins as well to sort of demonstrating to the business that there is actual progress and there are tangible outcomes, even if they're small. But there are, you know, that it's progressing.
And one of the things that you touch on in your videos is the problem of under-communicating their vision. So let's focus now on this aspect here, the people to people interaction, when you've got the funding, you convinced the C level you've, there's enough funding to run for at least a year, but you want to focus on, you know, getting a deliverable within the first three months.
And, now you gotta start. You gotta start building a picture of what is the reality of information management in the org. How does that work and what are the human challenges involved there?
George: [00:31:31] Yeah. And thank you for asking that because the human challenges are maybe the biggest challenges with data governance and maybe not just with data governance, but in general, I feel that we're not doing this enough, we're not practicing that change management aspect enough. And with this, I've made this mistake, so that's why I think I can talk about it, and I think a lot of others did as well. You know, as soon as we get that funding, we kind of want to just run with it and work and do stuff and improve data and, you know, set all this up. And we forget about the people side, which is the most important piece.
We forget to engage with them. We just tell them after the fact, "Oh, this is what we've done for you." And they're like, "that's not what I want. Why would I want that? How is that helping me all?" Like from the start, most of them would just be reticent and standoffish and because they were not involved in the process. So my first recommendation is first, you kind of need to have that, you know, education / listen to complaints / find out what's troubling them. So then you can build a case for each one of them, for each department or each individual there you're talking to. How would data governance help them with some of the challenges that they've expressed?
And that way, you know, they're also part of the journey along the way, right? So you always have to communicate what you're doing, what you've done, what you're planning to do, and how all of these things tie into the "what's in it for me".
Loris: [00:33:09] Yeah.
George: [00:33:09] And this can be nerve-wracking at times. And of course, you've got to deal with different personalities. There are different priorities, there are different communication styles. So yeah, it can be a bit of an energy drain at times, but I think it's very important to do. I think I was at this conference years ago in Florida, one of the panel speakers there, one of the sessions, he was saying that you know what, data governance, in his opinion is 90% communication. Maybe 90% is a bit high, but still, I agree that communication is a very big part of it. And I feel it's not done enough.
Loris: [00:33:51] This is like, this is the part that really excites me because I've, I feel like, there is a huge opportunity to have a real impact at scale in an organization if we understand how to communicate and how frequently as well we should communicate. So my experience is not directly in data governance, but this episode that I'm telling you now, in a moment, it was in a context of a data engineering team, where we were building a new system, more scalable, more powerful.
The company used to have a legacy system, they just outgrew it, but they developed three years of analytics and BI on it. So there was a lot of like reverse engineering. There was a black box, really. So you have to start from the plot and understand what data models needed to be ready to reproduce it in a new front-end system.
I was, kind of at the beginning, just alone demonstrating, you know, can we even do it? And then the project started to pick up momentum, three people from the engineering team joined and we were full-on, on the, you know, the deadline. The focus was sprinting, sprinting, sprinting, writing as much code as possible, getting it done so we could switch the old system, turn on the new one and save some money in the process.
But what happened with that really output-focused approach, is that we completely ignored the communication with the business. We knew what was going on, but the business didn't know. And we assumed that they knew, right, because they gave us the funding, you know, there was a thumbs up,
And I'm mostly responsible for this. What I didn't understand is that they put down the funding, but that doesn't mean that they understood the impact and the value. They were exploring. They were like, "let's see if this is actually the right way of investing this chunk of money right now."
And this is something that, you're right, like, we need to do a lot more. To the point to that I'm wondering if it makes sense to have someone within the team whose primary goal is to communicate because when you're down into the woods when you're writing the code, you're worried about the data type for that column. You're worried about the test for that table. It's a very tiny little detailed thing. You're not in the headspace to talk strategy or to necessarily write a newsletter and let everyone know why you're doing it because the “what” is the number one priority? And have you seen an example of that in place? Do you think that could help?
George: [00:36:42] Yeah. So, you know, a lot of times, let's say the lead of that data science team, they're kind of looking to have this ability and desire to communicate and they have the skills to do so. But oftentimes, as you said, you know, we like to keep our heads down. We want to do the work and we want not to have as much engagement from that point of view, we just want to get stuff done, right?
So what I've seen that works very well was two things: one is either to the couple, some of these, you know, keen workers with like a business analyst type of an individual who is very well at articulating the business needs and translate those into technical requirements. But also the other way around, they're able to engage with the business stakeholders to raise different technical constraints or to make sure that communication exists between the two. So not necessarily like a communication specialist, but the business analysts would have those skills to do so.
Or you actually just embed that as part of the process. That, you know, kind of programmatically, you have to do this before you start working type of a thing. Or, every two days you have the sprints or whatever project management methodology you're adopting, then that can kind of force you to do it and remind you to do it, take you out of your comfort zone. Yes.
Loris: [00:38:13] Make time deliberately…
George: [00:38:14] Deliberately. Yes. Yes. But you know, you have that structure, you have the framework that you know, well, this is how it works here. And the other one is, actually when you do put, let's say for the data science example, when you put the data scientists within the business units and there are different models for this like they're either like part of that business unit and they report to the head of that business unit or some sort of like a consultancy model where you have a central data science unit that is then lending, let's say, its data scientists to the business and they become embedded into the business unit so they can learn the context, understand the problems.
But when you do that, you're kind of pairing them with the business stakeholders and their task for poking you for communication, so to speak.
Loris: [00:39:07] Yeah, and I think there's a, I remember you saying, and it made me laugh because, I think it's important, but it's also subtle, that there are three communication steps for successful data governance program. And one is, as we already mentioned to focus on the business outcome. And so align the program to the strategy and the vision of the business, then talking to the departments. So that's, you know, kind of completes the slice top-to-bottom and then communicate the results, maybe, if you have them. Apart from the difference between active listeners and passive listeners, it can be a bit intimidating for the data governance lead to communicate results three months into the job, right?
Because you feel like you don't have results. What results you're talking about, this is long-term stuff. Do you feel that that is indeed a problem? And perhaps what is a tip that you could share for the data governance lead to focus on something that is still valuable? And it is in a way a deliverable, even though it's not the finished product.
George: [00:40:26] You know, I think there are all kinds of metrics that you can track your progress against. And some are, like you said, like low-hanging fruits. And I think in three months you can do a lot of work. Even just putting a metric out there on how many meetings you've had already to engage business stakeholders. I think that's already a great indicator that, okay, something is happening. Like people are getting engaged. People are listened to, that's something that is worth reporting on. I think during those three months, you'll also start some sort of a data governance committee, council, advisory group, whatever people call it.
And, and again, that's to me a big accomplishment and that really sets the stage for, who is this group that will set the strategic priorities and will determine, like, you know, put the stamp on, "yes, this is a standard that we will adopt" and all that kind of stuff. And even reporting on that, who your members are as part of the council. How many times have you met? I feel there are so many low-hanging fruits that you can report on already, even just identifying, "Hey, we've actually just taken a map of all the systems that we think we have that, you know, capture data or work with data."
Loris: [00:41:41] That's amazing.
George: [00:41:41] All of that it is easy to do at least in the first three months, and it's something that nobody really has. If you're asking one person, "well, what are all these systems?" Everybody would give you different answers depending on their view of their work.
Loris: [00:42:00] Yeah, there's so many... the conversation could take very different routes now. I'm trying to, I'm trying to keep it focused on the human challenge, but perhaps.
George: [00:42:11] I think the human challenge, you know, I mean, there are a few human challenges, but one is the different personalities. Working with certain people, it's definitely hard. So I think you need to learn how to navigate. There's a lot of political challenges. It's usually a consideration of what people should you start bringing into the same room for certain conversations.
It really serves as the basis on how you select certain members for your data governance council, even how you select some of these focus groups or working groups. So that's why it's great if you, if your data governance lead comes from within the company and not as an external source, because they kind of know all these ins and outs.
Loris: [00:42:56] And in this way, the lead and the sponsor are still different people, right?
George: [00:42:59] Yeah, yeah.
Yes. So the lead, I would call that like your data governance manager or director or whatever title you want to give them. And to be honest, in a lot of organizations, data governance is kind of like a one-man show in terms of the title. And this is like a one-person department, many times. If you're lucky, you might get, an analyst, you may get a business analyst, maybe a project coordinator.
But yeah, usually it's a one-person team and yes, you do work with IT. You do work with business analysts, but they don't report to you. So it's always also a challenge on competing for their time and resources and all stuff. So it is a struggle, but yeah, if you're lucky you're getting, you know, you have a couple of people in your team and you also have great partnerships with you, like a data model or from the IT team, for like data architecture team or the business analysts from the project management team and so forth and so on.
Loris: [00:44:00] I wonder if that thing that you just mentioned, the fact that people perceive the data governance lead as a tax on their time essentially, is because the vision hasn't been communicated at probably at the level of the CEO, you know. So the next question is, what is the role of the sponsor, who is a data governance sponsor?
George: [00:44:21] Well, ideally it's somebody that's at a very high up that has a lot of influence that people respect and they listen to. And it's a person that, you know, when you, and when you get a communication from them, when they're stepping in front, you know, that meeting room you'll want to listen to her or him. And when they say we gotta invest stuff in data governance, you're like, "Oh yeah, we have to", just because they said it type of a thing, they don't even need to give you an example. So that's your like best case scenario type of leader to sponsor your program because people listen to, they follow their direction and they're also ideally able to secure resources for you.
That's why you want the sponsor to do. You want to be able to say, "Hey, dear sponsor? I need another head on my team". They'll say, "yeah, I'll get that done. I'll figure it out. I'll get that done for you". So they can help secure that. They can also help secure, sorry, convince other leads like directors or VPs that, "Hey, you got to, you know, your team needs to have an active role in this. We need these type of resources. We need your time. We need your involvement". So there, you know, the sponsor is able to bring people on board. And usually, sometimes the sponsor doesn't just need to be one person, but multiple people. And sometimes that's even better because if the sponsor leaves and it's the case that does happen at the C level, you know, they can rotate every couple of years. And every time they rotate, you kind of have to make your business case again. Or you've got to convince them if they've, it's something that haven't dealt with before you got to go through the motions. And it happened to me too, and it's a pain.
Loris: [00:46:11] I can imagine.
George: [00:46:13] If you have more than one, you kind of, well, you know, one fell, but now have somebody else as a backup, and they could kind of have, they can make that business case for you. And it's also the power of example, which we love, I think, as humans. You know, we see that "Oh, somebody else bought this. Yeah, maybe I'll buy it as well", or "somebody else is vouching for this and I'll trust it more".
Loris: [00:46:38] Yeah, it's a very good point. That also requires for the two sponsors to kind of align on what the priorities are, but in a way kind of diversifies risks, because you know that there are different perspectives of already aligned at the very top before you even start spending a single hour, interviewing people, creating a model.
George: [00:47:00] And yes, as you said, there are definitely pros and cons. You definitely don't want two sponsors that butt heads.
Loris: [00:47:07] Yeah.
You know, I find this topic very interesting. Because if you think about an organization from the outside without any involvement, emotional or political, like added noise to the train of thought. And you're like, okay, what's the purpose of an organization is to deliver value.
They typically want to do it at a scale. They have a bunch of customers, a bunch of brands. They really have one mission, but there are different departments and that is natural because it's a way for us to manage resources and tackle complexity. We divide things into boxes and we'd say, okay, you do marketing, I do sales, I do engineering.
But it's far too easy to forget that we're all doing the same thing. And so this competition for funding and for promotion and for headcount within the organization, I wonder how much of that is helpful when you think about a program like data management and data governance that impacts everyone in the organization. There is no one person left out. The data. And it's why I keep banging that introduction that data is valuable because it connects all business functions. And we don't necessarily think about that. The real world is made of objects and we make, we are good at making sense of objects, but we don't think in terms of relationships that much. Apart from our network of, you know, really close friends, our family, the people that interact with us or, even on social media.
There are people that have hundreds of thousands or millions of followers, but there's only a thousand, right. Research shows it's only a very limited number of people that we can manage. And in the sense that we know what relationship we are with these people and in a business, in a dataset, we have millions of data points across many, many systems across regions.
So it's normal. It's to be expected that people lose track of the web of how interconnected the systems are and they're focused on their thing, what do I have to deliver tomorrow? Okay. The report. Great. So I've got a data problem. Give me, give me good data. I want, you know, one person in my team just focusing on the data quality because I want to look good in the eyes of the CEO.
Do you think this is something that it's a problem, or oh it's a problem, but it's only at the theoretical level because you'll never be able to really change the structure of an organization so that everybody is really on the same page. And so we have to find as data governance, professionals, ways to get around it.
George: [00:49:55] Yeah, I think it is a problem. And I think I'm theorizing here, but I think data or data, for the most part, it's still seen as a by-product and not the product itself for a lot of companies. Like, you know, Google, Facebook, Amazon, maybe. Well, yes, Amazon, they do see data as a product as well, but for a lot of these other companies, they see it as a by-product. No, what we do is we sell some sort of a physical product.
Yes. We collect data then sold it, but that's a byproduct of our sale. And because they are seeing it as such, they're not investing as much resources into managing that by-product as they do and, you know, selling their physical product, if that makes sense. So, because it's seen as this sort of afterthought something that gets created, but it's not the intention necessarily, it's not what we're after, then it's harder to convince why resources should be dedicated towards this and not marketing, not sales. Makes sense?
Loris: [00:51:03] Absolutely. No. Yeah, I was drifting. I was thinking about, a similar problem, when you talk about change management and is, climate change and you know, all of the risks associated with rising temperatures globally. It's something that for a long time suffered of this very same perception problem where people are like, "this is not impacting me today, tomorrow, maybe in 10, 20 years. Who knows the scientists are correct in 50 years by, you know, 50 years time, I might not even be around. So that's not my problem. You know, I'm not going to prioritize it". Then, and this goes on for 20 years, then Swiss Re, the insurance giant comes and does a really good piece of research.
And they put a dollar value to the rising temperatures of the ocean. And they estimate, you know, $20 to $30 trillion lost in the economy by 2050. And immediately you get, you know, the US, Europe, Canada, everybody got, yeah, we got it. We gotta double down on our efforts. We gotta reduce CO2, and we gotta become carbon neutral.
Okay. So you understand this is something a bit humbling for us, especially for me. And every, you know, for all people that like to think in abstract terms and they like the philosophy, they like the sciences. All this thinking at a high level is nice, but we need to remind ourselves that what really, the majority of human beings understand is money, right?
How much is it gonna impact the bottom line? And so, in the sense, the work of folks like Doug Laney with Infonomics is really trying to approach a problem in that direction. Let's establish a framework for valuing, for understanding the monetary value of your, intangible asset, to the fourth intangible asset, and let's, and let's really be serious about it.
This is not just, you know, science talk. This is not just nice to have, but this is a must and must have. I feel like, starting... even when people understand, because that's the other side of the coin, right? When you have a strong business case and people are like, Oh my God, we can't live without this, we need to do it, and it's going to be a priority in the business, Then you have a problem that expectations are sky-high. Did you feel that in your current position or was it something more of an organic kind of growth in terms of understanding the importance and also becoming familiar with what could actually deliver?
George: [00:53:46] I think it was a mix and it was a mix of people as well, seeing it differently too. For a lot of them was they weren't necessarily looking at their data, but what they could do with it. So for example, the fact that they had to send all these marketing materials to prospect customers and they would waste or spend a lot of money only to find out that a bunch of those materials will return back because it was like an invalid address or because they weren't following like postal office standards.
So the mail house would charge them more for this bulk delivery or the fact that, hey, they only got like 1% worth of sales off of spending so much money. So then they understood, okay, well, what can we do to reduce some of these costs and what can we do to best target those that we think to have a higher chance of becoming our customers?
So that was like, you know, one of those needs and that already starts from a very positive place, because you know they're on board of understanding, "okay, well that means we need to invest in data management and data governance", right? And so starting there with the business approaching you or recognizing the issue I think it's already a win and you're starting from a good place.
Loris: [00:55:09] Yeah, absolutely.
George: [00:55:11] I just wanted to, you know, it's funny cause you give all these great examples and it makes me, you know, think of others as well. So I just wanted to add this one piece on what you mentioned on the value of data and Infonomics, and the example with the rise of the sea level and, what that's costing and we have this company in Canada. They sell basically like hiking gear and outdoor stuff. And like, I'm an outdoor guy, so I like frequenting the store. And in Australia you have something similar, like Anaconda or Kathmandu, right?
Loris: [00:55:50] Yes. Yes.
George: [00:55:51] Okay so think of this store as one of those.
And they recently sold because of COVID, they had issues with their revenue and everything. They sold to this American company. Well, ends up the American company, the reason why they bought it is for the data. It's because they had this loyalty program. So they were gathering a lot of different information about their customer trends and what they're more interested in and when they're more likely to buy and all these things. And this company, which is part of a bigger consortium that has different types of stores such as this one, they're basically buying information on how you can better succeed in the world of retail, you know.
They didn't care about the products. They didn't care about the stores and the tangible assets, but really the information that they could mine out of this data. Which obviously the company didn't really do a good job at taking advantage of it. So I really found that to be very interesting, and you know, such a great opportunity. And again, an example for all these companies that are collecting data that are maybe not utilizing it as they should. And other organizations kind of seeing the gap and capitalizing.
Loris: [00:57:14] I'm excited though, in the space, because I see the rise of a number of courses that focus on data literacy, not so much for people directly working, the practitioners. Because there's a ton of courses on Python and Excel, warehouse design, but I haven't seen a lot of data literacy programs for senior leaders, for people that sit in the C level.
And it's incredible how even those that actually are active on the topic of data governance and data management and information management, are the first that have serious troubles with basic technologies, even just like a Google doc, for example, or the idea of an online version of an Excel sheet that you don't have to keep bouncing, you know, via email, you can just collaborate and comment and keep it as a single.
So there are, so there are so many low-hanging fruits here and I can see something that's happening in that data literacy for C-suite space. And I'm excited to see more because I think this is the spot that you want to really tackle first, because if you've got that support and people feel that it's okay, right. It's okay. If you're a 60, if you're 65 years old, 10 years ago, you know, the technology was completely different. 20 years ago, Google didn't even exist. I mean, Gmail, right? So it's okay to be sometimes out of sync with the trends.
The first step is to recognize that. The second is to feel okay with that. And the third is to have someone or some organization that can guide you through a very simple process to go from this level of maturity to the, you know, what's expected to be a data-driven organization without necessarily charging, you know, millions of dollars because this is basic education. But I'm excited about the impact that something like this can have. So we'll see, we'll see.
George: [00:59:27] Yeah. And full disclosure. I was kind of thinking about this, that I should focus on my courses towards like upper management as well. Offer, you know, short introductory lessons on that. So yeah, because I definitely feel there is a high need there for that audience.
Tell me I feel like, the data world now is kind of like where IT was maybe back in the early 90s.
And maybe in some parts they are still like that, but just from the point of view that they always had to make the business case, "Hey, you got to give us money so we could help you better". And they always had to make a case for themselves. And why do we need to invest in security? And why do we need this equipment? And why do we need the software? And let's move on from, you know, recording stuff on paper and move it into Excel and all that stuff. So I feel like we're kind of following the same pattern and same challenges that IT did, you know, 20, 30 years ago.
Loris: [01:00:26] Yeah, with the added advantage that 30 years ago, we didn't have social media so developed and people shared a hell of a lot less today. I mean, it's enough to go around and look what remote-first companies have been doing, especially the tech companies, because, you know, they're the ones that use the latest, cool tools available.
But it's enough to just look at what they're doing. The first one that comes to mind is GitLab. GitHub is another one. There's WordPress, you know, examples of large organizations that operate across many different time zones. So they have thousands of people and they're fully remote. They literally leverage every single thing that modern digital technologies have to offer.
And, somehow managed to keep growing and keep people happy, fulfilled. So I'm not saying it's easy, but there are many, many examples we can take from.
George: [01:01:22] But you know, what else sucks that I've noticed when COVID started in the first six months and some of the companies were closing where they were having issues. Some of the first positions to let go were in data management, data governance.
Loris: [01:01:38] okay. Yeah.
And at the same time, the topic now is dinner table topic like Pedro Cardoso said, the master data marathon 2.0 where he was giving a roadmap for the MDM promise land. And what unpacking data trust and obviously data governance is a huge part of that. So yeah but I wanted to like dive into your courses and promote it to everyone that's listening because I think the stuff you're doing is incredibly valuable.
So what are top two or three learning outcomes that you have in mind, that you had in mind when you designed the courses and what's going to come next?
George: [01:02:21] Well, first is to try and simplify it as much as possible. Make sure it's clear, not so much theory, but actual practical stuff that you can take and apply. So that's why I'm providing those templates and actual examples that you could follow. I feel that it's a much easier to learn when you're looking at it like a real case.
Yes. I like I had to take names out and things like that. Yes. You know, I respect the privacy of my work and previous clients and stuff so really practical takeaways that you can learn from. And as, you know, my expertise lies in data governance. So a lot of my courses are on that topic, but then I also have one that I've co-developed with my colleague on data visualization for data storytelling. So that's also an area that I'm very excited about.
Loris: [01:03:09] So LightsOnData.Com is the place to access this content? Absolutely. Cool. Well, we have some time left, but, at the same time I feel like we covered so much. So I'm gonna leave the ball at you. Did you want to add anything?
George: [01:03:25] I think more people should watch your show because you're doing such an amazing job and I'm looking forward to hearing and seeing more content coming from you. And, you have a lot of great things to say. So, looking forward to having you on my show.
Loris: [01:03:38] Yes, that's going to happen soon as well. And I have my t-shirt ready. It's nice and ready for me. Well, George big, thank you for taking this time. I know that it's late, it's the end of the weekend, it's the end of the day for you. So double thank you. And, I think I'll see you, I'll see you soon on the other side.
George: [01:03:58] Sounds great. Thank you so much.