Data Leadership Unleashed: Navigating the Evolving Role of the Chief Data Officer with Salema Rice, Founder and CEO at CDO Today

Announcer (00:03): Data Masters is the go-to place for data enthusiasts. We speak with data leaders from around the world about data analytics and the emerging technologies and techniques data savvy organizations are tapping into to gain a competitive advantage. Our experts also share their opinions and perspectives about the height and over height industry trends. We may all be geeking out over join the Data Masters podcast with your host, Anthony Deighton, Data Products General Manager at Tamr.

‍

Anthony Deighton (00:38): Welcome to another episode of DataMaster. Today's guest is Salema Rice, chief Executive Officer and founder of CDO Today. The recently launched CDO Today is a think tank for CDOs by CDOs, a private community with executive mentorship, board advisory and thought leadership. Salema has over 25 years of experience mentoring, evangelizing, and directing data management strategies, AI transformation, digital innovation, and advising large, complex Fortune 500 companies such as Accenture, Bain Capital, Allegis Group, and Ally Financial. She's also the AI expert commentator for cnbc. Welcome Salema..

‍

Salema Rice (01:26): Thank you for having me this morning.

‍

Anthony Deighton (01:28): So you've been working in this world of data and analytics for newly 30 years and at this point, so and you've acted as a mentor, but who would you consider sort of your mentor and how is that the nature of the mentor relationship changed over time?

‍

Salema Rice (01:47): Oh, that's a good question. I have many mentors, actually. I have several people throughout my career who really I looked at as people who I wanted to be more like, right? I mean, I think as a woman kind of coming up in the world of, you know, I started my career in it. I was one of three women in the entire department of 200 at one time, right? And so it was, there was a gentleman at the time at, at B M W really, who by the name of Greg Lambert, he put a book in my hand called The 21 Irrefutable Laws of Leadership from John Maxwell. And it really, and he said, you know, people naturally follow you and you need to decide what type of leader you wanna be, you know, and so it meant a lot to have somebody like be there for me in a way that, you know, I wasn't the CDO at the time that he put this book, but he recognized a leadership quality in me that I don't know that I would've seen in myself even on that day and over the years, I mean, I keep literally keep the book on my desk <laugh> and often referenced it after all these years just because now in different points in my career I've referenced different sections of it, right?(03:10):To where you now it's really relevant to think about the law of legacy. And you know, as I look at mentoring and what I'm doing with CDO today, I have almost a hundred chief data and analytics officers around the globe that I get to spend time with and I get to be a part of their evolution and where they're going as chief data and analytics officers. And it's the, the best part of my whole career, right? I mean, like to see them evolving and to see their progression and, you know, the type of legacy that I wanna leave in this industry, it's been a big game changer. There's been others, you know, that I've looked at in terms of the, you know, leading by compassion, leading with compassion, obviously, you know, again, like the world of coming data wasn't sexy when I started in this industry.(04:10):We were a byproduct of the IT department and most of my offices the first 10 years were in the basement, literally. And so, you know, for me to be building teams and to be the type of leader, even of my own teams, not just leading my internal and external type customers, but to be the type of leader that I wanted to be to the people in my organization, to create a culture of innovation, to create a culture of experimentation, to win it together, to lose it together, to be able to not just hire and, you know, but to retain talent because they wanted to be a part of something that I was leading. You know, they helped me to really bring out that quality in me that a lot of my peers at the time looked at as a sign of weakness. And I looked at leading with compassion and leading with, you know, building relationships with people as something that was the right thing to do. And even if it wasn't the, you know, the status quo, it was how I was gonna lead it. And as a result, I think in probably a 10 year period, I had less than 1% attrition at a time when, you know, especially in like the data science area where, you know, a lot of people have, you know, a consistent turnover. I feel that that was probably one of my biggest success factors, and I owe that a lot to the mentors that I've had in my career.

‍

Anthony Deighton (05:46): Yeah, they say that people leave bosses, they don't leave jobs. Would you agree?

‍

Salema Rice (05:51): I agree, absolutely. I had a dear mentor, he was one of the CEOs at Bain, art Knapp was his name. And you know, I remember when he hired me, he said, you know, you are the c e O of the data company, right? And so regardless of what, you know, other, how others lead, you know, because you don't always have the perfect leaders in any organization, especially as you get into larger organizations where that feel, you know, very siloed in some cases, right? And as you're building out that environment, it's very easy to fall into that, you know, I have to follow the leader pattern rather than, you know, take the approach of, you know, if I'm the chief executive officer of this department, how will I run it differently than maybe the way the rest of the organization is running? Or be an example to other parts of the organization and how we could do it differently to have a different outcome, to have that type of work environment for your people where, you know, you are creating an environment where they're a part of something and you know, people wanna feel like they belong.(07:06):You know, I, as I look back on my career, you know, so many times I thought, you know, I came up through a number of different industries, you know, I spent a lot of time in financial services and a lot of different areas of analytics, so marketing analytics and data warehousing, and like so many different areas that would get put into my lap. And I kept thinking like, you know, as a believer I was like, you know, why am I, why do I have all these different components? And then as I started to evolve into becoming a C D O, and I'm kind of what the industry calls a fifth generation C D O, so you know, throughout my career I've been kind of the bailiff, I've been the lawmaker, I've been the builder, I've been the value driver, the strategic driver. And so as I look at that now, I think wow, like all being able to and have all those different opportunities across all the dif different functional areas, it really prepared me for the future and for the roles of chief date officers, which I did for almost 20 years. So,

‍

Anthony Deighton (08:16): So in the context of mentorship, I wanna separate two ideas. One is the mentorship to the person and the other is the mentorship to the company. And I'm curious, and I hate to go negative, but what are some of the big mistakes you see? And let's start with a person, a new leader in general and a new CDO specifically, what are some of the mistakes that that person makes when they take on that role, especially if they've been sort of promoted into that role, this is a step up for them. Are there common pitfalls and mistakes you see?

‍

Salema Rice (08:52): So every industry is different and every company is different. Right now, you know, when I came up as a, I'll call myself a legacy C D O, there's not a whole bunch of us out there that are legacy that kind of had it all right? Like had both the IT and the business and sort of everything in between that looks or smells like analytics or ai. And so for a lot of chief data officers, there's this concept of, you know, we're either gonna thrive or we're gonna dive, right? And so I think that it's important because, you know, when I was coming up, we had this mentality of trying to boil the ocean sometimes, you know, the idea around products wasn't there, you know, it wasn't until just the last probably five years of my career that I sort of added the hat of chief product da chief data product officer.(09:44):And so as you're trying to, I identify, you know, what is the most critical things for us to monitor, to measure, to make sure that in order for the business to make quality fact-based decisions, you know, what's the best thing we can do for them, right? What's the, and it's not, you know, the business coming to you and saying, I need a table, I need a server. It's what is the problem that you're trying to solve? And so trying to help chief data officers that have been in that role and new ones understand that by taking a more dual velocity approach and not doing it the way we used to do it, right? I mean, we used to reverse engineer s e c reports in financial services to determine the 20,000 most critical data elements of the bank, right? And then you spent how many years doing all the data quality and all the components, but now if you take this more product based approach and a dual velocity approach, you can narrow down a very specific problem.(10:47):And I think that that's probably been one of the biggest ways that I've been able to help chief data officers is because I did it the hard way where we spent years trying to show the value of data quality of master data, of these very complex things that do take years. But when you narrow that down into a very finite problem and say, as a result of answering this problem, we're also going to make sure that it has the best data quality that it is mastered, but you're not trying to do it to 20,000 or, you know, attributes or trying to do it to a very much smaller number. I would say, especially in like organizations that don't have policies and procedures around doing these things, a lot of times they'll build the policy or procedure just for the sake of we check the box rather than who are you really having kind of go behind you and make sure, like do we have data observability, do we have data quality?(11:51):Do we, you know, like are we really confident or do we really trust the data? Right? And I think that for a lot of CDOs, that's been probably one of the biggest things. You know, the, maybe the second to that would be around using other data from around the organization. A lot of CDOs are either in the business or in it, and they're very focused on one area based on the type of industry, right? So for financial services, you know, they have it the easiest cuz they've got the regulators kind of with the gavel, but take like, um, take like a pharma, right? Or a biotech where it's not as much regulated, but the focus has always been around testing, right? Or identifying, you know, if they're looking for cures for things or, you know, identifying new medications. But it's very, very siloed. And so what happens is that there's great work done, but the whole organization is not sharing that.(12:55):And so by helping the chief data officers to understand that, even if their area does not cover like the whole world of the company to, you know, direct them to kind of build a way to build an operating model where they're sharing the best of what each other is doing, the best practices, like learn from each other's mistakes and then work to integrate that data, right? Because a lot of times, you know, we see companies that have kind of done it the same way for so long, they've really struggled with like a rear view mirror approach of, you know, reporting and analytics. Like, here's my report, here's what happened. But to get to that evolution of here's what happened, here's why it happened, and here's what's gonna happen next. You need to have a more holistic view of the organization because something that happened in sales is what's gonna impact finance. And something that happened in is what's gonna happen, what's gonna affect hr, right? And so, so many of these things cross-reference, but the CDOs today are really, you know, a lot of CDOs really struggle with like, how do I narrow that down into a very small minute product that we can quickly show the value so that I don't have to wait the three years or the five years to justify the investments.

‍

Anthony Deighton (14:27): Got it. Yeah. And, and so in quick wins is clearly important. So I do wanna talk about this idea of data products, but before we get to the sort of the output, the data product, let's start with the role of the chief data officer themselves. And you've had the benefit of many years of experience, both as a CDO and advising CDOs. So in your view, how has that role changed over time and where do we find ourselves today as it relates to chief data officers?

‍

Salema Rice (14:59): I think it's changed a great deal. I think that, you know, one of the reasons that CEOs and boards look for people like me to be on their advisory board is that, you know, they wanna hear from legacy chief data officers how data and analytics impacts their C e O strategy, right? And even if they have a C D O, the C D O today may or may not even have had experience in data, right? Like they might not know about, you know, if the CDO is on the business side and only comes with business expertise, have never been a part of the, you know, never own the ingestion, never own the teams that bring the data together, that architect the data for analytics, then there needs to be a connection there, right? Because, you know, in my world we were early and we kind of had it all right?(15:59):There's a handful of us that had to do it all, but now you've got CDOs who live in IT who don't own the data, right? So I'm the chief data officer and I don't own the data. And then there's CDOs in the business who own the data but don't own any of the processes and don't have the financial capabilities to architect the data properly for analytics, right? And may or may not have, you know, the global responsibilities, they may be very narrow. So I think that, you know, I even have a couple of chief data officers that I mentor that have no experience. They came from VP of supply chain, right? But because they owned analytics on that side, the C E O or C F O or the ccio, you know, has decided that this person who has a very broad spectrum of knowledge of the company needs to be groomed to be our chief data officer.(17:02):And so that's, you know, very, very different from, you know, I think, you know, having that technical background, you know, at one point I remember saying to my team, there's not a role in my organization that I haven't personally done and that's very different than what it is today. And I think probably the other really big item is ai. Like, you know, I mean, you know, when I first started right, it was like reporting, then it was, you know, then it was analytics, then it was big data, right? Or big sandboxes, I dunno what you wanna call it. And then, you know, and now the idea that not only can we, we have evolved from, you know, to where we're now not just predicting, but ting, you know, what is gonna happen in the future. And we're seeing things with generative AI that, you know, chief data officers were not prepared for, right?(18:04):If you were a chief data officer in, you reported up to the chief risk officer, you probably have an 80 20 rule, right? If it's 80% right, it's probably okay for risk, right? But you know, if you're, if you're the chief date officer in a a retail company, you might report to the chief marketing officer. And so they've kind of been, the industry has kind of like positioned the chief data officer. It's, it reminds me of, um, you know, I spent a lot of time in the talent area and I love talent data because I feel like it's one area where AI could really make a change. Early on, we were able to use AI to predict the right talent for the right job at the right time. And, and I saw a lot of times the idea around a software developer, right? You have software developers that make, you know, 55,000 a year and you have software developers that make 300,000 a year. And it's all based on the skillset. And I think that that's one of the things that's happened to the chief data officer is that, you know, depending upon what skills are required to be the chief data officer in that organization, it can look very, very different.

‍

Anthony Deighton (19:19): So I wanna get to this question of AI and generative ai, but before we do that, I want to go back and as I promised ask about data as a product and data products. To your point earlier, this feels like something pretty new. And this idea of treating data like a product, it's very common in an organization to have somebody responsible for a product line. I mean, that's a, a general manager, it's very well understood role in this context. We could think about a chief data officer as actually responsible for a product and think about rather than treating the data an organization generates as an exhaust or something that needs to be dealt with or managed or protected and, you know, but think about it actually as a foundation for real value. So I'm curious if you would agree with that and then, you know, more generally your view on, well,Salema Rice (20:13):I love the idea of data products because I think that it allows, the role in general of chief data officer is very broad to me. I think that if you are the chief data officer, you have responsibility to the board to make sure that whatever the company's goals and objectives are, that you're helping them achieve those goals by using data and I ai within your organization, right? For me, a lot of times it's been how can I help the company grow or can I come help the company save money? And so how can I build products to do those things? Because by building small products, right? Like we learned early on around, you know, I loved when the idea of agile data management came up, right? It was probably 10 years ago and everybody said, oh, you can't use Agile and data management. And we're like, oh, sorry, can, like we can do anything in small chunks.(21:17):It may not look like, you know, it may not look like everybody else's, but we're gonna use Kanban boards and we're gonna make sure that we have, you know, parts of the business identify as owners, just like we did identifying data stewards to own the data. We need parts of the organization to step up and own the products. So not necessarily the chief data officer, but it was my role to make sure that any products that I built that were going to be used internally or externally that, you know, that we had somebody identified for. And so I think that because you know, there is a structure in place then where you're bringing those people together where you have somebody who, you know, it, it's kind of like the old story of, you know, if I just need to get to work faster, I can buy a, a skateboard, right?(22:14):But I mean, when you're delivering value to the organization, you, you're creating even for yourself, right? Like you're, you're going to be recognized more often. You're also gonna be more visible, right? More often. So careful what you wish for. But I think that this one area is the single most critical way that chief data officers get a seat at the table. It's the easiest way for them to show value and it's the most beneficial to the organization. You can take something big and make it a product or you can take something small and make it a product, right? You can customer can be a product, right? It can have the same components as a widget, right? That's used to help deliver value within the organization. But the common areas of that are that we treat them all the same regardless of what type of product. We can easily categorize products, but it's a way to bring sort of that, that three-legged stool together. I need it, I need the business and I need data to really make this sick, right? To make it happen. And so, yeah, I think that, you know, for chief data officers that in, at least in the last two years, that has definitely been the number one topic.

‍

Anthony Deighton (23:43): Interesting. Yeah, no, I would agree. And I think what I appreciate about the data product concept is that as you point out about Kanban boards or other concepts in traditional software product management, we can now apply those same techniques, those same ideas to data and also think about connecting it more directly to value within traditional product management. The idea of connecting to value. What is it we can charge for? What's the pricing on that? You know, why does the customer value this? Those concepts could easily be applied even within an organization to the data that the organization's using. So as we sort of wrap up here, I wanna make sure I get back to this question of generative AI chat, G P T, the question of the moment, a question I'm confident you get every five minutes, but from your perspective, large language models, chat sheet pt, is it over-hyped or over-hyped?

‍

Salema Rice (24:49):I don't think it's over-hyped. I think it is our first true inflection point into the world of ai and I think that it's, you know, when we look at chat CTP and we see, you know, I think they had a hundred million users in the first 30 days and now it's roughly 30 million a day. I think that it's, it's scary though for chief date officers. I don't think it's scary for the world cuz it's a fun tool. As long as you recognize that you probably trust it 50 50, right? I mean, it's, it's not something that you should, you know, I tell my my kids all the time, they like, they love it and you know, it's kind of like Alexa and steroids, right? <laugh>. So, um, but I think that for Chief data officers, it's opened up a world of, you know, the importance of privacy of quality and how we use it because I think that it can, it can help make the world a better place.(25:56):Definitely. There's so many areas that I've been working with chief data officers on of, you know, how they're gonna use it within their organizations. But I think the importance of protecting the data, knowing who owns the data, where the data's coming from, where it's going, putting guardrails, you know, um, it's just a, it's a new shiny thing that if we're not careful, you know, I think that with, you know, I think I said recently, like with great power comes great responsibility, right? It's like Superman, I know the Spider-Man <laugh>. Um, and I think top DTP is just like that.

‍

Anthony Deighton (26:35): Yeah, no, I think that's a good point. And in particular, I appreciate this idea that these generative models force us to take, put a premium on the quality of the underlying data. The idea of ground truth with these generative models becomes much more suspect. People type things into these, they get back these answers that feel very authoritative and they're based off of bad data. And so, you know, making decisions from that risky.

‍

Salema Rice (27:05): Yes, very much so. But the good news is that there's lots of, you know, there's lots of ways to integrate the data and have your own private instances within your organizations that you can make it safe and you can make it fun for the organization and you can make it, you know, to where organizations are now going to really start to use data as a way to monetize within the organization itself. And I think that it's, you know, that's a turning point for a lot of organizations, right? I mean, they, they want to use data as an asset, they want to be a data driven company, but now I think even more so we're gonna see, you know, that increase as they want to move into this generative AI and things like data quality are coming right back, right? Like how do we ensure that the data that we're using now in generative AI has the quality controls around it that we're making sure there's no, you know, privacy concerns. Like those are all things that are gonna be, you know, top of mind before they start allowing those models to be used company-wide.

‍

Anthony Deighton (28:10): Right? Exactly. Uh, well look, so the fantastic to have you on DataMaster. I really appreciate the perspective both in terms of how to be an effective mentor, how to be an effective mentee, how to think about the changing role of the chief data officer and data products, and now the next revolution in G P T and large language models. So thank you very much for joining us.

‍

Salema Rice (28:35): Absolutely. Thank you Anthony.

‍

Announcer (28:38): Data Masters is brought to you by Tamr, the leader in data products. Visit tamer.com to learn how Tamer helps data teams quickly improve the quality and accuracy of their customer and company data. Be sure to click subscribe so you don't miss any future episodes. On behalf of the team here at Tamer, thanks for listening.

‍