Caroline Carruthers is an award-winning CDO who’s worked across multiple industries. She’s authored a number of best-selling books on how to use data as a strategic asset, including The Chief Data Officer’s Playbook. Today, Caroline guides organizations on data best practices alongside business partner Peter Jackson, CDO of Exasol.
As the Co-founder and CEO of Alation, Satyen lives his passion of empowering a curious and rational world by fundamentally improving the way data consumers, creators, and stewards find, understand, and trust data. Industry insiders call him a visionary entrepreneur. Those who meet him call him warm and down-to-earth. His kids call him “Dad.”
Satyen Sangani: Here in the United States, we’re afraid. Very, very afraid.Of math.Yep, you heard that correctly, data radicals. And it’s a very serious issue. Research in the Harvard Business Review showed that 93 percent of Americans have what is known as “math anxiety.” This anxiety can have a pretty chilling effect. When something makes us anxious, we avoid it, we run from it, we want absolutely nothing to do with it.
But that has to change, because we’re living in the age of big data, and it’s only going to get bigger. According to the U.S. Bureau of Labor Statistics, math related occupations will grow 28 percent by 2030. This will create roughly 67,000 new jobs, and I think that’s a massive underestimate. You don’t have to be a data radical to understand those numbers.
(00:54): If we want to tackle the Goliath of big data, we need to equip our Davids with more than just a slingshot. If you want to be a transformational chief data officer (CDO), then it’s imperative to help your team build their confidence. It’s something that our guest Caroline Carruthers has done many times in her career. In her LinkedIn profile, Caroline calls herself a data cheerleader. As you’ll see in this interview, it’s the perfect way to describe her.
Caroline is also the co-founder of data consulting Carruthers and Jackson. She and her business partner, Peter, are also the authors of The Chief Data Officer’s Playbook. To all current and aspiring chief data officers out there, you don’t want to miss this conversation. So let’s get into it.
Producer Read: Welcome to Data Radicals, a show about the people who use data to see things that nobody else can. This episode features an interview with Caroline Carruthers, co-author of The Chief Data Officer’s Playbook. In this episode, she and Satyen discuss the traits of an effective chief data officer, how to drive transformation within an organization, and the importance of emotional intelligence. This podcast is brought to you by Alation. Chief data officers face an uphill battle.
How can they succeed in making data-driven decision-making the new normal? This state of data culture report has the answer. Download to learn why successful CDOs partner with their chief financial officer to drive meaningful change. Check it out at alation.com/dcr3.
Satyen Sangani (02:31): Before you can give your team their metaphorical slingshots, you have to understand the qualities that make an effective chief data officer.
Caroline Carruthers (02:39): I think the interesting one for me is this one I always call out, which is a little bit of luck. So we actually list that as one of the key ingredients because we know that when it comes to the evolving nature of being a chief data officer, and the evolving nature of organizations to be able to get the most from a chief data officer, you need to be kind of in the right place at the right time with the right support. That is an element of luck. So I think we need to be realistic that you could be the most amazing chief data officer in the world, but if you go into an organization that’s not ready for it, you’re probably going to end up butting heads and maybe have a bit of frustration there. But the other key things I always find fascinating when it comes to the role of a chief data officer and what’s useful, is you need to be credible. So you need to have some background in data.
(03:25): But I think one of the misconceptions I find more than anything about the chief data officer role is people seem to think you have to know everything to be able to do it. So with the best one in the world, if I were the most amazing data scientist and data engineer, and I could do governance, and I could go on and be that unicorn, I’d be training for 50 years before I could become a chief data officer.
You don’t have to understand every single discipline within the data role to become a chief data officer. What you do need to have is a level of credibility so you can talk authentically about the subject you are passionate about. And I have to say that is definitely one of the key ingredients, to be passionate about the subject and believe in the art of what data can do for organizations.
(04:14): Just to do a slight segue, I was in a conference really early on in my career as chief data officer. And there was a gentleman across from me at a table, and he turned around and went, “Well, what we do is kind of boring, isn’t it?” And I got so frustrated and so mad because as far as I’m concerned — and I did have a few words with him — because I find what we do fascinating. I just think it’s incredible what we can actually do with data, what we can achieve, the problems we can solve, the difference we can make. And if you don’t have that passion for it, how on earth can you bring anybody else on this journey with you? How on earth can you convince someone else to change their behavior if you think it’s boring?
(05:04): That was a slight side introduction, but I think it’s, for me, actually caring about what data can do is a really key aspect of it. And the other thing that I would probably bring out quite strongly is this idea of being curious. So the idea of being the problem solver, the curious — actually wanting to do something with it and delving into it and asking questions — is a big part of being a really good data professional.
Satyen Sangani (05:34): There was this famous survey that came out — maybe from the MIT CEO Summit — that said that the job had a mortality rate of 18 months and people are having a hard time succeeding in the job. How do you build that balance? Because one’s about business value with all of this AI/ML shiny stuff that’s hard to achieve and fundamentally has some failure rate, right? And then on the other hand, you’ve got all this governance stuff which is hard to achieve for a whole bunch of different reasons, mostly because people just get exhausted from it and it’s compliance driven, and it’s not super inspiring — or at least may not be super inspiring. What do you recommend for people to do to sort of avoid that 18-month failure?
Caroline Carruthers (06:14): I think there were two separate articles. So one article from Gartner said that the average tenure for a chief data officer was 18 months, and that more than 50 percent of them would fail, which I think were the rough figures for a data person. I’m being very lackadaisical about the data here, but I think it kind of gets the point across. I actually think it missed the trick, slightly, because I think there’s two things in that for me. So one is, yes, some chief data officers didn’t succeed, but a lot of that was down to organizations having the wrong type of chief data officer or not being ready. So we saw this whole set of organizations that went out and got themselves a chief data officer because everybody else was doing it. And they came in, and they put them in post, and they hadn’t a clue of what to do with them. Or they suddenly expected miracles because one person came in and they were going to wave the magic data wand and it was all going to be fixed. I remember one wall I walked into where my name — not my job title, but my name — was the mitigating action for every data problem the organization had. Now, as good as I am, I’m not going to solve that in a week. So your expectations need to be reset just a little bit.
(07:34): I think that was partly what was happening when that survey came out. But I think the other side of it is that it was at a time when most of those chief data officers were going into organizations and they were the first chief data officer. If you think about any transformation-type role, the first iteration of the transformation tick is about 18 months.
What I think was actually happening was a large proportion of those people that changed rules in 18 months, they were the kind of slightly maverick chief data officers that liked setting something up, that liked being dynamic, that liked the creation set of the program … they’d gotten the transformation underway. And as soon as business as usual started kicking in, then they were like, “Somebody else can do this. Now I’ve created something that somebody else can look after.” So I do think that part of that 18 months was tied into it being the right time for that first set of chief data officers to move on.
Satyen Sangani (08:35): In Silicon Valley we often talk about founders that innovate and CEOs that scale. It’s the same with CDOs. It’s incredibly important to know where your skillset lies.
Caroline Carruthers (08:45): We tend to talk about the whole idea of the first generation and second generation. The first generation’s exactly that. They’re the first people that walk through the door. There probably hasn’t been a chief data officer before them. They’re really responsible for all the creation, the setup, literally: What kind of teams do you need? How does that work? How does it interact with your operating model? All of that kind of set. Whereas your second generation is the follow-along from that. And the other interesting thing that we tend to find between those two types of chief data officer, is we talk about the data role being on a pendulum from a risk-averse to a value-added side. And the reason it’s on a pendulum is because it does swing back and forth and you do have to take account of both sides. (09:30): But if you’re the first generation, the chances are you’re probably going to have a bit of a heavy thought about the risk-adverse side with a guide toward how you can move to the value-added side. So if that’s where your focus is, you’re probably a first generation. And if your focus is much more on the value-added side, because a lot of the hard yards are being done for you on the risk-averse side, then you’re probably second generation.
And again, you can see why, for both of those roles, you can’t squarely sit in one half or the other. You have to traverse across the two, but your main focus has to be in one of two areas. So I definitely do think most chief data officers now would recognize themselves and that kind of description about whether their focus would be the risk-adverse side of the data role, or their focus would be on the value-added side of the role.
Satyen Sangani (10:24): And you also mentioned that there was this skillset divide that people are not going to be able to have sort of the full gamut of skills required, right? One might have grown up on the AI/ML data science side, whereas in others, very successful CDOs might have grown up on the data governance side. A lot of that comes down to team building. And then also to prioritization setting, which strikes me as some of the big skills.
Do you find that people still are making the mistakes of sort of trying to go too big too fast, or staying within their lane? What are the pitfalls that people fall into when they are not successful or when the 18 months are really due to them having failed?
Caroline Carruthers (11:02): You’ve hit the nail on the head for one of the big factors, which was they think they’ve got to do everything quickly. We talked about data drive and the digital agenda, but we don’t spend enough time talking about how people drive the data agenda. A large part of that is starting small and actually demonstrating value to the organization about what you can do.As you said, they start too big, so somebody goes in and tries to justify why they should have a team of 200. You’re much better off starting with a team of 10 or 20, demonstrating the value, giving the organization a really good glimpse of what’s in it for them to go on this journey with you and building incrementally. I’ve always seen that as a much better model for taking and starting a transformation.
Satyen Sangani (11:49): How can you get your stakeholders on board with your plans for transformation? Caroline starts with something you might not expect, coffee and cake.
Caroline Carruthers (11:59): It comes down to the whole idea of managing your stakeholders. I mean, there’s a big joke when I first started; I was known as coffee-and-cake lady, because virtually every speech I gave was about how you must spend time with your stakeholders, don’t go to them and talk to them about data, and don’t go to them and tell them about what problems you’re having with data. Relax them, take them out for coffee and cake. Sit there and just get to know them as a person and ask them about what they’re trying to do, what hopes they’ve got for their departments for the future. What’s causing them to be awake at night? What’s stopping them from sleeping?
(12:37): Have those kinds of conversations with them, because when you are actually looking at how you make the biggest difference, what you should be doing is solving problems. So rather than going for the part of the organization that shouts the loudest — and I will guarantee that if any new chief data officer going into a role, there will be parts of the organization that will get very enthusiastic and shout very, very loud — it’s absolutely crucial that you take the time of understanding those cross-organizational problems. Because remember when I said about starting small, you need to demonstrate the difference. And if you can find some relatively small problem that impacts everybody, then all of a sudden you’ve got your first what’s-in-it-for-them — “I can take this away, I can solve this problem” — and then you buy the credibility from the organization to trust you to carry on that transformation.
Satyen Sangani (13:30): When you look at the work, what do you find are those common one or two or three projects that are easy ones for people to start with?
Caroline Carruthers (13:37): I think some of the common ones are around the data governance, because virtually every organization I go into they always cite the quality of the data, or people don’t have access to the data, or it’s not in a timely fashion, or they’re not talking around the same things.
Some element of governance around the data is normally one of the first things that organizations have to tackle if they haven’t already started to tackle that. And that obviously comes into the whole idea of the data catalogs and the governance software that can help you with that.The other side — and very popular at the moment — is this whole idea of data literacy. So, actually, improving data literacy across the organization, because it’s one thing to have data that you can trust. That’s fantastic. If you don’t have the skills to do something with it, then it doesn’t matter how much data you’ve got.
Satyen Sangani (14:29): There’s another challenge with data literacy: making sure your organization understands what it is. Caroline says there are a lot of misconceptions about what “data literacy” actually means.
Caroline Carruthers (14:40): One of the problems I’m seeing around the whole idea of data literacy at the moment is people assume it means that you’re becoming “technical.” So it means you have to be able to use a particular piece of software, or you can do brilliant visualizations. And for me, data literacy is something more fundamental. It’s actually: Do you have the ability to ingest, use, make decisions on the back of data? If you have that, then that’s a level of data literacy.
And it’s also not about one size fits all. So I often talk about this as a bit of a spectrum. There’s the whole idea in organizations that some people will be what I would call data aware, or need to be data aware, which is the basic kind of level that you would expect from anybody, which is the kind of thing of, “Do you understand why you need to fill a form in properly rather than just hit the top of the dropdown box?” That is an element of data literacy: that understanding and the importance of it.
(15:37): And for some people that’s good enough. That’s what they would need to actually be able to do their job properly.
Then you move up the spectrum to a more what I would call a data competent. So that’s the people that are working closely with the data teams, they’re asking questions in the right way of the data teams, they’re getting visualizations, reports, dashboard, metrics. And they’re knowing what to do with them. They’re knowing how to drill down into that data. So that’s a much higher degree of data literacy.
And then, obviously, you go through the spectrum into the data scientist, data engineer, all the people for whom it is their day job, so they probably better be very literate with it. So I think those two, the whole idea of the data governance and the data literacy. So you’re tackling the basic trust and the data and helping people understand the skills they need to use it.
Satyen Sangani (16:32): Data literacy also has a different definition at each level of your organization.
Caroline Carruthers (16:37): I guess the difference for me is there’s a difference between a person being data literate, or a department being data literate, or even an organization being data literate. And we need to kind of be clear about what it is we’re talking about when we talk about these kind of terms.
I mean, in organizations… And again, we see the software that you work with. This would be completely in line with who you’re talking about. But too often than not organizations have different parts of their company that talk about the same widget, but they’re actually talking about apples and oranges. And in some cases, they’re talking about something so different. It’s actually apples and orangutans.
It’s so fundamentally different from what they talk about and they don’t realize it. And I would hate to think as data professionals, we start and put terms in place and we do exactly the same thing. So that’s why I think we need to take some time out and get these terms, right?
Satyen Sangani (17:29): This idea of organizational literacy, departmental literacy, and then obviously individual literacy: What do you think the characteristics and the hallmarks of each of those levels would be like? What does it mean for an organization to be literate relative to a department? And how do you think of those differences?
Caroline Carruthers (17:48): So I think if you go back down to the person, then I think that’s about, Can you use data to do your job? That’s the very simplest level I could think of. That’s the person side of it.When we go up to the organization, sort of go to the complete opposite end, I think the big part of organizational literacy is around, Are you all marching to the beat of the same drum when it comes to data? And I don’t mean from a robotic “we are borg” type fashion of everybody having to repeat this in a fashion – but that they get that it’s safe to ask questions and they know how to interrogate data and build stories, and that using data is second nature.
(18:31): I think that’s where the organizational literacy comes into play. I think departmental literacy, that for me is where you actually work well as a team. So you may have two or three highly data literate people in your team, but everybody knows the role they play and how they get the best from the team, how they get the best from the department, without it becoming a bottleneck.
And I also have to add that last bit in, because it’s too often than not you have this one valiant hero in the corner battling against the tide of all the data requests that they’re getting from their team. And that just stifles them. So I think the departmental literacy is probably the hardest one to get right.
Satyen Sangani (19:11): Right, because everybody’s using this sort of expert as a crutch, as it were. And rather than thinking for themselves, they’re expecting this individual… Like at our first customer there was a guy named Caleb. Everybody just asked Caleb questions and he had this one datamart, and everybody would go to him with all of the answers or try to go to him with all of the answers.
Caroline Carruthers (19:33): Exactly. It’s so many things in organizations. It’s the whole 80/20 rule. So it’s where the department knows that for 80% of the things they need to do, they can do it right, but they’ve got a few experts that can help them with the other 20%. And that’s the way they balance it out rather than, “I’m not going to bother doing it. I’m going to go straight to Caleb.”
Satyen Sangani (19:52): I want to switch gears a little bit because.. I love your LinkedIn page. It has this four cheerleaders with a word data on them, give me a D, give me an A, give me a T, give me an A, which I’ve never… Those data and cheerleader don’t seem like words that I would naturally align. And maybe I don’t understand as one as well or the other, but how do those two things relate? And what does it mean to be a data cheerleader?
Caroline Carruthers (20:18): I genuinely get really excited about what data can do. Let’s be honest, it’s not data itself. Data is a building block and an atom of business. It is a small component. It’s not the data itself that excites me realistically, it’s when you put lots of bits of data together. It’s what kind of problems they can solve and what kind of difference they can make.
(20:43): And I know that I can get a little bit evangelical about this, but for me, the potential of what we can do if we actually start and use all the data sources we have morally, safely, but pull it together and start and use our creativity and our imagination, I genuinely think we can solve some of the world’s biggest problems by doing that.And that’s why I’m a cheerleader when it comes to data, because the more people that I can encourage to be passionate about this subject, the more people that I can help build the skills of, the more people that I can excite to come into this space and help me solve these problems, the more that I think that we can really start and tackle some amazing things.
Satyen Sangani (21:25): Caroline is also helping to cultivate the next generation of data cheerleaders.
Caroline Carruthers (21:30): We also run something at Carruthers and Jackson called the Chief Data Officer Summer School, the Data Leader Summer School. So we run this school every single year. I think we’ve run it four years now. So we just finished our fourth year. We run it once a year because, quite frankly, it’s a ton of work and I could only do it once a year. And we weren’t free. And it’s for the next generation of chief data officers. It’s for the next generation of data leaders. It’s for people who want to be doing the stuff that I’ve been doing, that other people have been doing, and hopefully doing it better than I have because they’ve learned from my mistakes and they won’t make them, they’ll make different ones.
(22:06): So, that’s where a lot of the enthusiasm comes from because I can quite frankly say, “Me on my own, I can only solve so much.” If I can spread this at each year, create another 400 people who are as passionate about this as I am, and they go out, and they can be enthusiastic in their organizations and their companies, then there’s no end to… It’s boundless about what we could actually achieve.
Satyen Sangani (22:29): What do people that come into the school look like? And then what is the outcome of what they become afterwards? And what are the key skills they learn during this course that you teach?
Caroline Carruthers (22:38): So the kind of people that we look for are the people who are getting ready to make that step. So they’re probably a head of, they work across different departments. They get the data’s important, maybe their organizations are about to go on a journey and they’re getting ready to build up and be instrumental in how that does it.
So that’s where they start, but we literally take them through the summer school, through the basics of how to create a team, create a data strategy. What does that mean? How do you look at the maturity level of your organization? And what does that mean going forward in how we create those programs? But we also do things like, how do you tell stories with data? We talk about softer skills like emotional intelligence, resilience, coffee and cake. How do you put people at ease?
(23:22): So we cover all of those kind of skills as well as creating the team. How do you create the right kind of team dynamic to help you solve the kind of problems that you want to solve? What are the key skills that you need to make sure that if you don’t have you have around you? And we make them do homework as well. It’s not a free ride. You have to do homework.
So the whole idea is that by the time they finish summer school, they’re ready for the next task. They’re ready to either take the step up to being a chief data officer, or they will be within a year period. And we have found it takes about a year from what we’ve seen so far of… It’s about a year from being on Summer School to achieving the next step up.
Satyen Sangani (24:04): One thing that you mentioned that I wouldn’t readily think about putting in the curriculum would be sort of emotional intelligence. Tell me a little bit about that. What are examples of why somebody might need emotional intelligence and doing this work with data? And what experiences have you had where that is even relevant in this work?
Caroline Carruthers (24:26): I could only see this as a data person myself. Ordinarily, we get caught up in the hole and what data can do, facts, figures, numbers. Oh, does that chart look amazing. We can easily slip into that kind of way of thinking.
Whereas it’s so much more powerful to be able to bring people on the journey with you, get everybody marching to the beat of the same drum. And that’s why the emotional intelligence side comes of it, because a data leader is a business leader, so you’re not focused on technology, you’re focused on outcomes, strategy, transformation. And to be able to do any of that you need a very clear set of emotional intelligence so that you can understand what’s happening politically around you in organizations and how do you bring people on the journey with you?
(25:14): I mean, in any organization, I would always say bringing people on the journey is probably the hardest part, because we are incredible as human beings. We love our comfort zones. We sit there in our nice little circle. This is our comfort zone. And it’s called the comfort zone for a reason. It’s lovely, it’s nice, it’s warm, it’s cuddly there. I know that if I do X, Y will happen, because my experience over the last 20 years has taught me that.
When you are coming in as a data leader, and you want them to use data in a different way, you need them to change their behavior. You need them to do something in a way that they’re not comfortable with. So you need to be able to convey what’s in it for them: Why should they? Is it worth making an effort about changing behavior? And all those things start with emotional intelligence.
Satyen Sangani (25:59): I agree with Caroline. Emotional intelligence is something that’s essential for chief data officers. That sense of anxiety and fear that I discussed in our opening monologue can come from anyone at any organization, even someone at the very top. To close our episode, Caroline has a fascinating story that demonstrates why it’s so important to meet people exactly where they are.
Caroline Carruthers (26:23): It was a major international bank we were doing some work with, and we were finding out a large portion of resistance coming from one particular area. And we happened to be running a workshop and the director of that area was in the workshop. And I have to say he was a little belligerent. It’s probably a really nice way of putting it in the first part of the workshop, really undermining what we were doing.
So at the break, I took him to one side. And kind of you know, being a little bit straightforward, shall we say? So it was like, “Come on. You and me. I want to go and have a coffee.”
Satyen Sangani (26:55): And what was this workshop about that he was being belligerent? Was this some controversial topic that you were trying to proselytize?
Caroline Carruthers (27:03): Not at all. Not at all. This is what I would’ve called data 101 for executives. So this is trying to teach some base skills on how you want to use data and the kind of language that was coming out was “I don’t need that.”
In fact, I think one of the comments he came out with was, “I don’t need to do this. I don’t use data. I just look at spreadsheets, because it feels like you and I are sort of going to have this conversation. It’s like there’s a communication link not happening.
Pulled him outside, went away, took him for coffee. I sort of extended the coffee break and let everybody else have a half an hour. Very long story, or very short, was that when I started asking him questions and had a really honest conversation with him, it all came down to, he didn’t get it. And what I mean by that is that if we were using terms like machine learning, artificial intelligence, he’d heard a lot of terms that to him were very scary because he didn’t understand them at all. So even big data was terrifying.
(28:10): So it’s things like that. And because he was incredibly senior in the organization, he didn’t feel like he could admit that he didn’t understand. So what he was doing was just putting roadblocks in the place and actually blocking any proposal that had anything to do with data, because then he didn’t need to feel exposed that he didn’t understand.
So when I explained some of the terms that he was frightened of in that kind of way, he literally pulled that, “Is that it?” kind of face and went “That, that is what you’ve been frightened of.”
And as soon as he got over that idea of fear, as soon as he got over the hole, “This is not as complicated as I thought it was.” It was brilliant. Actually, he turned out to be one of our biggest supporters. And all it took was a little bit of honesty and a “Do you know what? We’re going to have a frank conversation here, because this is not working.”
Satyen Sangani (29:00): So how do you conquer math anxiety at your organization? Apparently, it’s by becoming a data cheerleader. What I love about Caroline’s use of that term is how rich it is. There’s a lot more to being a data cheerleader than telling everyone how amazing data can be.
It also means that you have to understand the relationships others in your organization have with data. So, going forward, let’s make sure we’ve got our pom poms out ready to support our teams and their data struggles.
Thank you to Caroline for joining us on this episode of Data Radicals. This is Satyen Sangani, co-founder and CEO of Alation. Thank you for listening.
Producer Read (29:42): Alation empowers people in large organizations to make data-driven decisions. It’s like Google for enterprise data, but smarter. Hard to believe, I know, but Alation makes it easy to find, understand, use, and trust the right data for the job. Learn more about Alation at alation.com. That’s A-L-A-T-I-O-N.com.
Season 2 Episode 25
Sanjeevan Bala, ITV's Chief Data & AI Officer and DataIQ’s most influential person in data, embraces the ubiquity of data in the enterprise by embedding domain-specific data ‘squads’ within business units to localize decision-making. He discusses how, unlike monolithic data teams, meshy data organizations are the best way to align data initiatives with business value.
Season 2 Episode 24
What does baseball have to do with data? Ari Kaplan, head of evangelism at Databricks, was instrumental in bringing a data-driven approach to a previously gut-driven sport and inspiring the Moneyball book and movie. Ari explains how businesses can learn from sports analytics, why a data culture is so critical to success, and how AI and generative AI are, literally, changing the game.
Season 2 Episode 8
How can a software engineer create the next big thing? According to Matei Zaharia, creator of Apache Spark and co-founder of Databricks, it demands a single architect to build the cathedral – and an open bazaar to empower the masses. In this conversation, Matei shares his startup philosophy and reveals exciting advancements with Databricks Unity Catalog and Dolly 2.0, an LLM for enterprise.