Episode Overview:
In this 29th episode of the CDO Matter podcast, Malcolm has an engaging conversation with a long-established thought leader and noted author in the data and analytics space, Tom Redman.
Malcolm and Tom discuss Tom’s recently published book, ‘People and Data,’ where Tom shares his thoughts on the necessity of engaging data consumers to address some of the bigger data-related challenges all organizations face, particularly data quality. Tom shares his recipe for engaging as many stakeholders “without data in their titles” as possible, and the need for CDOs to assert their leadership through an acknowledgment many legacy approaches to data management simply aren’t working, including data governance.
Rather than seeing people as the core problem data leaders need to overcome, Tom makes a strong case that people are, in fact, the solution – especially those outside traditional data roles.
Episode Links & Resources:
Good morning, afternoon, evening.
Whatever time it is wherever you are, whenever you are consuming this wonderful content, My name is Malcolm Hawker. I’m the host of the CDO Matters podcast. Thank you so much for joining me and my esteemed guest today.
I am joined by Tom Redmond, aka the data doc, a president of data quality solutions, long time thought leader in the space, I met Tom. Tom, when did I meet you? Like, maybe about a year ago, I wanna say?
When we started chatting a year or so ago, and the first met face to face, I don’t know last month. Yeah. That was weird. I think it was last month in San Diego when we first saw when we were first face to face, but even before the year before we had met individually, I mean, I had known if Tom, if you’re in the data and analytics space, if you’re active on LinkedIn, if you’re going to conferences, you know, if if you’re kinda out there, chances are pretty good. You would have run in a time over the years because he’s long been engaged in in our community as long been held and known as, as a thought leader in the space.
I was just kinda doing some research on you on you, Tom, last night This is why when we were backstage, I referred to you as as Doctor Redman. I I didn’t know, or maybe I I had forgotten you had a PhD, for for Florida State of all places, my wife did her graduate work at at Florida State. I’m I’m a gator. Did we ever have we ever talked about this? Maybe we have that. Almost certainly not or we’d, you know, be fighting now.
So I did my graduate work at at UF, never proceeded as far as PhD, but I did get my masters from, from the University of Florida. So there we go. We have something in common that we that we didn’t that we didn’t realize. But, anyway, I’d been following Tom’s work for a long time, author published all the time in in HBR, slowing management review, other really highly regarded esteemed publications One one of the people that’s actually out there doing research in in our world. So thrilled you could join us today, Tom.
Excited to talk about your book that will, be publishing two days before this episode is is launched. So by the time people are are listening or watching to this. They’ll have the opportunity to go get the new the new book, people in data. That was the working title.
Is that is that is that the final title people and data? Yes. That is the that is the final title, people and data. Okay.
Because I I I was I was lucky enough to have an advanced copy that I, I I excitedly digested enjoyed reading it and could thoroughly recommend it to everybody. We’re going to talk about that book today. We’ll talk about all sorts of different stuff in the in the data and analytics space. Tom and I I talk all the time, and our and our conversations are are generally fairly lively.
So, hopefully, that will, that’ll, you know, carry forward to today. But Enough about me. Tom, why don’t you spend just a little bit kind of sharing the the underlying premise of the book? I think, like, many, I got the sense that it was a COVID book.
Yeah.
I I I mean, in the sense, look, Malcolm, for a long time, I’ve, you know, you mentioned I’m a thought leader. And so, actually, I mean, the the the work I’m most proud of is is my work with clients as as we try to, you know, struggle to figure out how we’re going to make data quality, data and so forth, you know, fit in their organizations and and get them moving.
Right? And so I started my my my career at Bell Labs. And when I was there, I had one foot in these enormous AT and T problems and then one foot in a lab. And and in a sense, you know, the best of real work leading to theoretical understanding leading to real work. And and so, I left Bill Labs a long time ago, but I’ve managed to maintain that back and forth, back and forth.
And, and it is true that that during COVID, that was a great time to synthesize things.
And, and so the the the I made an observation, you know, sort of early there that, you know, once you see it, you can’t unsee it. And and, and it was so obvious once I saw it, and I saw it from from so many perspectives.
And that is the idea, the simple idea everything we wanna do in the data space depends on regular people. Right? Now now, you know, when I’m being good, I just describe regular people as those without data in their title. And, when I’m being a little snarky, it’s, you know, those who are smart enough not to have data in their title.
Kinda. So, like, you know, the and and and so sort of, like, you know, one way of looking at this is just take anything we want to do. So, like, maybe we wanna improve data quality. Well, data quality depends on data creators and and data customers, and those are regular people.
Suppose we wanna conduct a data science project or or implement AI. And step one is understand the problem. And so you you as a data scientist or an analyst or whatever, you can’t do that alone. You have to work with regular people. To understand the problem. And and and and so as I went through thing after thing after thing, it became clear that anything we want to do dependent on on regular people.
And then and then there was this, like, second view of this thing, that really sort of empowered the first, but if you look at an organization today and you look at where the data management is actually being done, It is being done by regular people. It’s it’s being done by people with sales and their title or marketing or operations, and And what happens is they need data to do their job. And when that data is not good enough, then they spend an incredible amount of time cleaning it up, fixing it up, verifying, getting other sources, setting up spreadsheets and so forth. Like, the the the all almost all the work, actual data management work is being done by people without data in their titles, day in and day out.
Simply so they can they can do their job. And and it is really important to understand that they’re doing this without any training, without any guidance without any support. And so, basically, basically on their own. And and I hope, people are listening.
I mean, if you have data in your title. I mean, you’re just horrified by this for some right. You know, this is where the data management is being done and and smart people doing their best But, you know, again, I mean, not professional data managers not trained or or anything like that. And and so that was sort of like the second line.
And then the third line is, You know, we’ve been talking about I don’t know since I’ve been around that, you know, data can be transformational.
Right? So, you know, I I don’t know if there is an MBA class, you know, MBA one zero one, transformation one zero one. But if there was on day one, the first lesson would be get everyone involved. Okay.
You can’t do anything that’s transformative and unless you get everyone involved. And and and so, sir, I mean, like like just take these three different streams of thought. What do we wanna do? What’s being done now?
The possibility of transformation?
All point in a bright, bright, shiny light to to to regular people. And and I feel that the main thing, if you’ve take only one thing out of out of people and data. It it is it is the primacy of regular people in the data space and the need to to get them involved in everything we do.
Well, one of the things that struck me, like, one of the, kind of moments to me in the book, which I assume is one of the the aha moments that led you to write it, was this simple assertion or what would seem otherwise simple assertion that for a long time, you know, data people see people as the problem when people are they actually the solution.
And and that that that mindset to me was a was a great kind of galvanizing force to to put my mind in the right frame as I progress through the book. Because I think I think that that that mindset of, oh, people, or they just don’t get it. Or you can lead the horse to water or, like, all of this, all of this stuff that you and I have talked about so much, from the perspective of, you know, kind of, you know, blaming of the consumers of maybe even a faulty product, the mindset shift is saying, uh-huh, they’re they’re the answer. They’re not the problem. To to me was okay. That kind of laid the found the the the path in front of me that allowed me to just kinda keep walking forward because I always I always kind of anchored back to that. And it was that mindset shift that was was to me arguably probably the most powerful thing about consuming the book.
And it was a great way to kind of set things out. So so so so thank you for that because I couldn’t agree more.
Right. You know, you know, Malcolm, I mean, it’s a really great call out a lot of tech people. I mean, even have this word wetware for for people. Right? And it’s kinda like, you know, well, do whatever you wanna do with the tech and then deal with the wetware.
And and and we just, you know, to I I sort of there’s a naivete about it, but but further. I mean, like, there’s you know, it’s kind of offensive.
Right?
You know, things in the world happen through people. And And, and people are amazing. And once you, you know, get them working and and and solving problems, they come up with things that that are just simply stunning.
Yep. Well, you one of the things that I I appreciate also from the book, and I I think this gets back to the very first thing that that you said at the start of our conversation, which is your your your particularly proud not of being a thought leader of things that you’ve published or anything else, but the work you’ve done with clients. And one of the things that I appreciated the most in the book, and I I didn’t know whether they were true aneCDOtes or they were they were maybe somewhere in the middle, but but you would always kind of anchor back to some use cases, right, with with with people who were doing real stuff.
Right? And and and real real work and and and supporting business and that’s something that that I that I actually enjoyed because it kind of always brought me back from the kind of the conceptual, maybe architecture centric or data quality to the, okay, these are the people that are actually really doing the work. So that’s something I actually, I found was also grounding through through the through the path of the book. So I I look, I mean, we did make use of some personas to to sort of illustrate some main points as a com collapse some people. In terms of of things that have actually been done, there’s nothing in there that has not worked in practice.
I I I don’t wanna say that anything in the book is typical. Right? But frankly, I mean, I think the book is is is ahead of its time, and it’s I mean, you can’t really say them about your book, but but everything in that book has worked, and it has worked well.
And the thing I’m most proud of was it made people’s lives better. Right? It it made their jobs easier and simpler and and and more powerful and enable them to take control.
Yeah. One of the one of the things that There’s there’s there’s a few things I I I would like to press on because I’m I’m a bit of a provocateur myself.
And and one of them is that you that you that you say in the book. And and it’s really and you say it in a few different ways.
Which is is arguably what is a shortage of leadership in in our in our space. And and I and I would really love to shine a little bit more of a light on that because CDOs are our our our audience here, but but also because I’m not entirely sure that the idea of leadership in the data space is well defined.
I’m a little concerned that that how we’ve been historically defining leadership is maybe slightly a little bit mis guided.
When you say the time for better leadership is now when you say it a couple of different times in the book, point blank, frankly, What what what does that what does that mean to you?
Well, so so look, I first of all, I mean, the way I think of of leadership, it is, you know, he, she, or they, who are out in front.
And and I I, you know, it’s something that it can be individuals, it can be people, I I, I have, I talked in the book a lot about, I think, like, senior corporate leadership so far in the data space has set stayed on the sideline.
Right? I I think we yeah.
Right. We, in with data in our titles, have have made it really, really hard for them to know what to do.
But but I when you’re talking about data leaders, I mean, I I thinking we’re talking about what can we with data in our titles do better. And and for me, the first thing that we’ve not not done is we’ve not recognized reality on the ground as it is. Alright? And and the unfortunate reality on the ground is is most things in the data space are failing. Right?
Now, I don’t know, I, you know, your eyes got, wasn’t it, but it it’s really it’s it’s not that simple. Because there’s these amazing points of light where things really got better. It’s okay. And I, you know, I spend most of my time on data quality.
Well, data quality, if you do it right, if it as described, In the book, it’s a it it is amazing. Right? You can you can eliminate orders of magnitude of error and cost associated with that and empower people. And, and but but that’s not the way most attack the problems.
I most attack the the data quality problem by going in and and buying a tool and and seeing what they can do with the tool. And they don’t get at the root causes of the data quality issue. So I think the first thing is is too many people in the data spaces. We we are not seeing the full reality, and we are not acknowledging the the the reality.
And and in particular, like, I, you know, I I talk to people in governance, and I I kinda say, yeah. You know, seems like that’s not really very popular. I’m I’m kinda dancing around. And then and I always get this nervous giggle and it’s, oh, yeah.
There’s that. Right? Well well, you know, the the reality is is governance is not working. Right?
Governance as performed today. Is is not working.
And and so if if we want to lead, we have to recognize, okay. What we’re doing is not where we have to do something different. And and maybe that involves dissecting things in the database and figured out what really does work and and how to address them or maybe it means something else. Now personally, my view is is that is that we have we are trying to work too much in in a current organizational structure that is just not fit for data.
And and we have shied away from the hard problems.
You know, another example of that is culture So, you know, we talk all the time about data culture, data cultures.
Well, you know, like, where are you really? Where are we really addressing that?
And and so so I mean, this is what, you know, we have not done the kinds of things that we need to do to recognize the current reality.
And and and get beyond it and move beyond it.
And by the way, I hope people in data makes it easier for leaders to do this, right, to see the okay. We did not know. We needed to get everybody involved. We get that now.
Right? We did not know. We needed to attack quality in this way. We know that now.
And, and and we can get on with it.
One of the things that you you touch on, well, one of the things I also appreciated when it comes to specifically to data quality There’s no shortage of metrics that you’ve thrown out about the opportunities that sit in front of us for when we actually do figure this stuff out. Right? Like, the the the jaw dropper was the three trillion dollars of negative impact. I think it came from an from an IBM study, but you you threw out a bunch of other, a bunch other data points, as well.
I also appreciated that that that, you know, you you lay out a lot of reasons for why we got to how we are. One of them to me a little bit of an aha moment. Something I guess we’ve always known, but just kind of like the the the function specific, the manufacturing specific, maybe even the kind of the manufacturing lines specific way of how we allocate work within an organization makes it makes it harder for CDOs to succeed. Role ambiguity is certainly something that that you that you touch on is one of the four core reasons why we got into a pickle, roll ambiguity.
It’s just something that you just touched on and has been a really a common theme at a lot of the conversations that I’ve had with people about CDL related challenges. So, obviously, there’s something there for for folks to to figure out they’re being recruited for a CDL role really, really, really drill down on what that role means and how successes, are defined Another area that I that I that I enjoyed, with the book was, discussions related to, I think what you call consumer and supplier type relationships and how important those are. They they seem to me to look a little bit, like kind of more of a data product type approach.
Would would you agree? Although you didn’t really kinda say that directly. So so let’s go into this discussion a bit slowly. The the the first thing that that, I, I, I wanna do is, is I wanna make clear that everybody in all and every corporation, everywhere, and practice in another day in and day out lives too.
Everybody serves in two roles. First of all, they’re a data creator, and and they’re a data customer. Alright? And and so I’m not I’m being very precise with the word.
Right? I mean, I’m not using the word consumer. Data doesn’t get consumed with use. Right?
You know, it it it it’s still there afterwards. Consumer is not right. So And and and so the the the fundamental idea of these job that you come into your job and as a customer, the stuff is not, you know, good enough for you to use. So you have to clean it up, fix it up.
Right? And if you get into that trap, then that becomes the the the the the thing that goes on every day. And the people depend on who I talk to, you know, a a quarter to a third to a half. In some cases, three quarters of people’s day is spent cleaning up.
Cleaning up data. And and the way you have to break that is you say, well, gee, why do I keep getting bad stuff in? And and the answer, the the the resolution that is to reach back to the creator, right, and say, you know, hey, look, here’s what I’m seeing. Let’s get some metrics around this.
Can we, you know, put some some cost on it? Can we can we identify and eliminate the root causes of error? Can we put in some crawl. So so things don’t come back.
And and and so fundamentally, I just went through a few things. People understanding themselves and roles is as customers and creators, right, articulating the customer’s needs, measuring against those needs, finding and eliminating root causes of errors and control. So root causes don’t come back. That’s the way quality works.
Right? I mean, that’s the only thing so far that has has has led to two big improvements. Alright?
And and so now and and by the way, I mean, I I I, it’s this has proven itself an industry after industry. We’ve gotten lots of case studies in in the public domain. So with AT and T and with Chevron and air energy and and and so forth. Right?
I mean, this is this is out there now. Now you ask the question about product management. So the product management is is is kind of interesting. Right?
So, you know, just this idea that you’re gonna come in and and and and do quality of those those steps. I mean, it may not translate into an overall program for the company. Right? And and so how programmatically are we gonna describe this?
And so you may wanna put a wrapper around these steps that that I just talked about at a good wrapper would be data supply chain management or a another wrapper might be supplier management, external supplier management, or a six sigma is a wrapper that people in that, you know, with with that set of tools might using. And so as far as I’m concerned, like, the the, the data product manager wrapper, Right? That could be an effective wrapper. I I’m I I wanna see it happen, kind of thing.
I was I was just at a conference recently people talked a lot about it. And, and there was a lot of, you know, good ideas and, you know, and and not too many success stories.
It it seemed to hinge on whether a set of customers, a group of customers needs were similar enough that you could satisfy, a bunch of of of light needs with one product as opposed to sort of the one on one that the customer creator model leads you to.
Right. And and I think there’s a subtle difference here between what you and I are hearing at conferences, which is mostly in my perspective data as a product, meaning as as a productized widget, something you sell, versus product management as what you would call a wrapper. Right? Or as what I would call more of a mindset or a discipline potentially as a way to kind of look at your role as a creator of data.
Yes. You’re a consumer. We all consume reports, but you’re also creating it day in and day out. And if you took more of a product management approach, meaning thinking that you had customers to keep happy, that that could be useful.
And that’s slightly different than saying, okay, data. A product, and we need to put a SKU on it, and we need to market it, and we need to put it on a marketplace, which I think is a lot of the conversations today. There’s a nuance y difference there I think. And what I read in the book kind of aligns more to the the idea again as product management as maybe a of overarching discipline.
And I think they’re there may be something there. The only the only thing that that I I wonder about, because if it was a product and I was a product manager, I’d I’d be selling things and I’d be making money One things that I that I would wonder about is, you know, if I’m just a, you know, I I’m doing, I’m I’m putting entries into a general ledger or I’m or I’m doing credit checks for, you know, a new supplier on onboarding. I may not be thinking that I’m a data creator and I may not be thinking about those down stream impacts. And I’m and I’m probably not paid on data quality.
I’m paid on how fast I put stuff in the GL. What what what do you what do you see there’s potential bridges to to bring those worlds together, particularly around maybe the idea of incentives for for for doing better data quality? Is that is that what do you think about that? Well, I I I I think you’re, you know, you may be taking two steps and we’re, you know, let’s go one at a time.
Okay. Right. I I mean, the first thing is is just, you know, well, like, when what is what are the experiences of people when they reach out to data creators and say, hey, I need help. Right?
You know, you you know, the stuff you’re providing me is is is not getting the job done. And and and look, all almost every data creator wants to help. I mean, particularly if somebody is wearing the same badge, I find that, you know, some managers are are horrified, right, that the that the data that, you know, their teams are creating would you know, would not be meeting, would not be meeting the needs of of people inside the the company.
I wanna build on this too a little bit too. It’s like, you know, with people who are on the front lines and maybe having the first contact with customers, or it’s a maybe they’re in a sales role or or or or something like that. But, you know, getting that data right, like, it it sets the tone for a lot of things. Throughout the company. And and it’s really, really hard to recover if if that data is is not right. But I don’t know. And oh, I you know, in almost every situation, right, explaining.
Right? Well, okay. Yes. You gotta make the sale. Yes. You have your your your productivity quotas and so forth.
But also, look, you’re part of the enterprise. And and if we’re gonna keep this customer for a long time, we gotta get certain things right. And and and and let me explain how, you know, what happens when you get stuff wrong, how that impacts the rest of the company.
Over and over again, people have been so respectful of that.
I I wanna go one step further by I mean, if you think about it, like, like, think about how dehumanizing it is to say Malcolm, your job is to fill in that field.
Right? As opposed to Malcolm, we need you to get this field right because if you don’t get it right, then here’s what happens in regulatory reporting. And here’s what happens when we try to sell more stuff to the customer. Here’s what happens when we try to try to build them. Right? I I mean, providing you the context for doing your job, that’s just treating you like a, you know, like a human being.
And, and in that context, I don’t know. I find almost almost everybody responds well to that.
So so look, I mean, I think that’s you know, I think that’s step one.
The the idea of data is a product. I mean, it’s a go ahead. You have a question. Yeah. It’s not not a question, but what what I’m what I’m hearing you say, the the way I’m internalizing what what I’m hearing you say is that we may not necessarily need or even get to a point where, like, transactionally, I’m given, like, a nickel every time that I enter in something correctly because I don’t wanna be a machine and I don’t wanna be a data entry robot. I wanna be a part of something bigger, which is which which is what I think I’m hearing you kinda say. I’m I’m I’m paraphrasing you.
But getting back to the incentive thing and realizing that we do have to pay mortgages and feed our families and all those other things. I’m I’m wondering how much of this could be solved by you know, maybe what would be more of a kind of a private company approach where I actually have some ownership in the broadest ticket company. I know I’m going way off way off track here. But I I had a little bit of an aha moment there because I’ve always I I I tend to think a little transactionally about these things.
And what I just heard you say is don’t think transaction really think bigger picture, think broader scope than just capturing and entering data. And that in and of itself, I think could be useful in helping maybe data leaders figure out what incentive models might look like here or what process owners might might, you know, from from the perspective of of some sort of incentive. So I I I cut you off, but I’m I’m I’m thinking of talking at the same time. Not a good combo.
Yeah.
Look, let me make a an important point. There’s some stuff going on and sort of like, you know, data mesh that’s, you know, that’s it’s got a lot of people excited, and it’s sort of, you know, like, on these incentive things and stuff like this. But we’re the way most data quality most great data quality programs start is not with the creator, but it’s with the customer. Right?
So being a good customer is really, really the key here. And and most of us come in and we kind of think, well, I’m entitled to, you know, whatever I need to to to do my job. Right? And then, well, if it’s not here, okay, I’ll fix it up and And and, you know, and I’ll get through my day and I’ll get my productivity stats.
What what what’s really required, you know, I mean, the way things have really gone well is when somebody said, you know, this is never gonna work. I’m gonna continue to work twelve hours a day. Right? I’m gonna continue to use more.
I’m gonna continue to spend more of my time fixing up data. I have to get in front of it. Right? And and the first step is for a customer to say this process does not have to work this This is no way to run a store.
And then they have to do the hard work of figuring out what they really want.
Okay. And and articulating that in a way that then you can go talk to a data creator and say, hey, these, you know, the the these are the kinds of things I I I really need. This is what good data looks like. This is what I’m getting. How do we work together to, you know, get me better data selfishly kind of I’ve not seen anything that didn’t start with the customer.
Right. Recognizing that there was a better way. Right? That there had to be a better way and and and and reaching back. And And so, you know, the incentives around the the creators, I mean, I haven’t worried about those. The incentives around the customer are You’re working twelve hours a day now. Six of it is non value added work, cleaning up junk that you shouldn’t have to clean up.
Right? Make that so then it’s no longer six hours. It goes to four or two, and maybe it’ll stop at one. Right?
And now you can actually spend a greater of your day doing your job. Right? And and people don’t need to be incentive on on on that. That’s That’s what they wanna do.
That’s why they took this job in the first place.
Now you do highlight though in the book that the people who drive those changes, right, are unique. Right? Yes. And and and the and that they are they are unique.
You you call them provocateurs. Right? Yes. I call them provocateurs. I mean, look, these to me are the heroines, the hero heroes and heroines in in the data business.
Are those who recognize that the path we are on does not go anywhere good. Right? And we have to get on a better path and they figure out a way to, you know, to do that.
Right? And they’re, you know, and the reason they’re provoke provocateurs is is they figure out a better way and they provoke the rest of the organization to do better. Finding those provocateurs, helping them and maybe even incentivizing them or at the very least acting out or creating partnerships with them because in theory, they’re going to be in those roles in supply chain in finance in wherever these roles are because they’re they are the customers of the data. It’s finding those people and really kind of promoting them and helping them and working with them So do you see a role from within the data and analytics team then to to maybe find those folks and to to to help them to to you know, act as in a consultative way with those folks?
Well, look, I think it’s the most important role. I mean, it’s, you know, it’s it’s it’s the most important thing I do. Right? I mean, the way, you know, I, obviously, I I don’t work for a company.
I don’t go look for them, but when people reach out to me, and and, you know, basically, and, you know, hey, I’m wandering around this. It doesn’t make sense. I mean, can you help me sort it out? Right?
You know, provocateurs, no things about finding oil and and and, you know, meeting regulatory requirements and and selling stuff the people. They don’t, you know, they don’t know things about data, but what they are is they’re smart enough to have a sense that what they’re doing doesn’t work.
And then, yes, these are exactly the kind of people we we we wanna help. And, you know, it’s I don’t think I used the term in the book, but, like, these are the high leverage people.
Right? And and what’s really interesting, Malcolm, is is that organizations are loaded with people who have a great idea and are afraid to speak up.
Mhmm. Okay.
And and so, you know, I I mean, if we could do anything like, you know, get these people speaking up a little more and and and seeking help, then then that would be great. Now, by the way, this is gonna put real high demands on most people in the data spaces. So most people in the if you’re working on quality, Mostly what you’re doing is is you’re implementing a tool to find errors. You’re not figuring out what’s really needed. You’re not figuring out up simple measurements in place. Right? You’re not figuring out how to make root causes go away.
You know, you’re probably thinking more about, you know, can I use AI to better find errors than you are about how do I get to the root causes?
Well, you just touched on something. I think that’s that’s that’s is really, really relevant for the average CTO or VP of data and analytics. And if you’ve got somebody in that data quality role who is working to define rules or manage, you know, ex exceptions to the rules, maybe instead give them a week a month to go sit over the shoulder of somebody working in supply chain or in finance or in marketing of the very groups that they’re setting up the rules for or managing the exceptions for because I I think just having that insight, that that process level insight could be could be hugely valuable. Do you agree?
Yeah. Absolutely. Look. I mean, I all the real I mean, we’ve talked about this before.
I mean, but the important action with data is happening with or people. Right. And and I mean, it’s not to say that, you know, there’s not cool stuff in AI and great potential there. But right now, right, the problems and the opportunities lie in in hundreds of of thousands of of mundane little data quality issues between, you know, this customer and this and this customer and this creator, right, and getting them together, right, to helping them to is it’s what’s transformational.
It’s what builds the culture. Well, maybe instead of a week a month in in in governance committees, you’s you spend a week a month in, you know, some sort of like design type sessions, Anyway, I’m I’m I’m thinking and and and problem solving at the same time here and and talking.
Right. Let let’s transition a bit to to one of my favorite areas. So one of the other kind of you you I highlight four kind of root causes of a lot of our data related woes.
We already touched on the one which is role ambiguity. Another one is is silos.
And and and one of the responses that that you highlight and one of the things that that data leaders can do is is to, develop what you call kind of fat pipes.
I’ll leave the depth of that for the for the book.
But one of the the kind of the deliverables there kind of goes to to a place near and dear to my heart as as you know, which is our beloved common language.
Right. And and we’ve really struggled with the common language, a lot over the years. I did appreciate your your insight in the book with a recommendation to kind of focus from a data modeling perspective on a bit of a more of a party, and and kind of data modeling perspective because I think that’s that’s huge. Right? I think we we muddle a lot of things. We we we we muddle concepts of an entity with relationships and entities one of the areas that you that you outlined in the book where we model things as roles.
And we we kind of model role with entity. And simply just just extracting these things, you know, visa v better data modeling can go a long way to creating that that that common language.
But but I do know that that common language is is still problematic.
Do do you see and and I’m being provocative here because you just mentioned AI. So let’s go there. Do do you see maybe AI helping to play some sort of role in the in in helping solve the the common language problem?
Well, as I should so again, I mean, let’s back up and come to that question a little bit slow. Okay. Alright. Okay.
I mean, you know, the the the the first The first is, well, why do we bring up common language at all? Right? And and and the reason, you know, I sort of bring up common languages is, you know, organizations are loaded with technical debt. And and and it is making it harder and harder to, you know, do things like implement AI models and and there’s silos which make it, you know, harder and harder to for people to work together kind of thing.
Right? And and so So at the root of this is, you know, to to the degree that data models are reflective of language that people use and the ability of people to work together depends on, you know, us establishing a common language. Right? Common language lies at the root of of a a a lot of really, really hard problems in in organizations that are slowing them down in terms of, you know, just working together And then in terms of really doing anything big with AI that, you know, depends on on getting across domains, The, and and but but I think the the bottom line is is common language is is is for almost all organizations now, they cannot attack it with their current organizational structures.
Right? They they they they will make the mistake of viewing a systems don’t talk as a computer problem and assigning it to their IT departments, and IT departments have been trying. I don’t know for two generations, you know, to just to try to bang stuff together and, you know, do it in the warehouse and whatever. It’s a really, really fraught. And And so, you know, the the the, the book introduces this notion of a fat organizational pipe that gets around these things and and there’s a federal organizational pipe to, you know, to get at the common language thing. I do not recommend that anybody in their current, you know, in their current organization go after the FAD Organization, excuse me, go after common language without putting this structure in place. Right?
Now your question about your question about me. Right?
So so I mean, it’s it’s really interesting. The the the hard part of common language is not language.
It’s common.
Right? Yeah.
Right? And and so So you and I only have a common language if we agree on. Right?
So So can I, you know, imagine, right, you know, something in the background that’s that’s scoping around? So, you know, coming up and said, well, you know, Malcolm, here’s a hundred terms you use all the time and kinda what you mean by them and Tom, here’s a hundred you use all the time. You know, what you mean by them. And, you know, here’s the overlap of thirty five or something like that. Right? I I mean, I can imagine that happening.
Right? But then getting us to agree on what those terms are really gonna mean. That part is is the hard part. Okay. And so, you know, so other than, you know, in some dystopian, you know, scenario where computers have taken over and And so, well, Tom and Malcolm, you don’t have to agree. We’ll agree for you.
Kind of thing. I I I don’t see the common part you you know, and and I don’t know, the next five years or so. Okay.
Okay. Well, it five years is a long time. It it would appear in in the world of AI because things are changing drastically. When it when it comes to the when it comes to the common language, I mean, I I agree with you.
One of the things that AI seems to be particularly good at is is is language. Now the the the commonality part, we we could cert we could certainly talk about. But one thing that I’m excited about is that historically, now now challenge me on this if you disagree. A lot of data quality measurements and a lot of data quality tasks, a lot of data stewardship tasks are are highly rules based and are highly, deterministic.
Right? Either something conforms or or it doesn’t conform. Either it says customer or it doesn’t say customer. Right?
And and I am optimistic about a a world where you could overlay probabilistic measures in the in the world of data quality that are not just deterministic but are also possibly deter are probabilistic based off of a whole bunch of other things that we also know about the context of how that data was created or how it would be consumed. So I think I I do see some opportunity there. I I do think we’re, to your point, I think we’re, I think we’re years from it, but I remain optimistic that that that I think that the computers can help. They’ll never fully ever solve any of these things.
Well, should it look look, I do wanna do wanna get at this a little bit. I mean, look, data quality, I I mean, it is a broad topic. Exactly. You know, data quality data quality, you know, to to to a drilling engineer, right, what what he share they require is very different from somebody who’s doing regulatory reporting in a bank.
Right. I mean, so so so there is this incredible range of of of requirements.
And and I one of the problems I you know, so we’ve had as too much is is that we’re gonna find this this, you know, sort of, you know, one direction for solution.
Kind kind of thing. Rich, I mean, the only thing that has consistently worked is what I talked about before about, you know, customers, creators, understand your needs, right, measurement, and and and and finally eliminate root causes of error. But then when you get to controls, right, you know, inside is this piece of data correct or or or something like that? That is broad and and fascinating. Alright?
Right now, we’re dealing in the situations of typical data quality measurement for, you know, a typical, just typical somebody doing it in any kind work, sixty percent.
Right? And and we’re not trying to go from sixty percent to four nines. We’re trying to go from sixty percent to ninety six percent kind of kind of thing. And and and and and so that is much more about simple tools consistently applied management.
Right?
If we get to the area where, you know, now we’re at ninety nine point two, and we need to get to ninety nine point six. I I have no doubt that we will need far far more powerful tools.
Than, you know, than than than we’re using now.
Got it. And and and I I think your point is is valid And and I did applaud you in the well, as in my mind as I was reading the book, I mean, you you you did define data quality. I think accurately so, which was kind of this idea of what I would have said at Gartner’s is kind of loosely defined fit for purpose.
And and you and you just highlighted that. Like, what what is date what is quality to one consumer is different than is quality to another consumer.
And there are some basic things that we can be doing to improve a lot of life, whether you are a creator or a consumer of that day. So so point extremely extremely well taken.
On that note, Tom, I have enjoyed the conversation as I always do. I hope our listeners and viewers also enjoy it. People and data now available at a book seller near you because it will have released two days ago. So congratulations for that. I enjoyed the book. I I would recommend it to all of our listeners.
Really, really help you kind of think in a different way about the role that people would play.
And maybe even get you thinking in a little bit more of that consumer supplier, type mindset, which I think is is is very valuable. So so Tom, thank you.
Yeah. Thank you. But by the way, I I I do wanna bark at you just a little bit. I I want Wouldn’t be a conversation between us if you didn’t.
Yeah. I I do want you to adopt the language creators and and customers. And and Creator customer. I love it.
Yep. I love customer. I mean, right. I mean, And and it it just, you know, thirty second lecture.
Right? Either you create it correctly or you have to deal with the downstream.
Right? And, you know, you may pick it up from a supply from some place, you know, different. Right?
So this, you know, that point is not some you know, the old man’s gone off the rail on language. It really is reflecting what you need to do to get get this stuff right. I love customer. And and if I misspoke earlier, I should have said customer because I’m all about that and we take more of a customer centric view and everything we do.
Life’s gonna get better for data leaders out there. So with that, my thank you, Shannon. Congrats on the book. Thanks for your time.
And thank you to our listeners. We will see you on another episode of CDO Matters sometime very soon. Thanks, all. Bye.
ABOUT THE SHOW
Malcolm Hawker
Benjamin Bourgeois
Ben Bourgeois is the Director of Digital Marketing at Profisee, where he works with product, engineering and pre-sales teams to develop useful content and resources for prospective and current customers. He has experience managing content teams across the B2B SaaS, healthcare, specialty chemical and manufacturing industries.