The Art of Data Leadership

0:00:03.4 Producer 1: Welcome back to Data Radicals. Today Satyen sits down with Taylor Culver, Founder of XenoDATA. Taylor is one of those rare data leaders who is actually successful, so he knows what it takes to drive business results with data. Today, he helps executives and data leaders do that too. In this episode, Satyen and Taylor explore what it takes to be a standout data leader. Why staying agile is a must, and how to be smart about both the promise and limitations of new tech. If you're looking for strategies to tackle data challenges head on while staying curious and sharp, this one's for you. Let's get into it.

0:00:41.4 Producer 2: This podcast is brought to you by Alation, a platform that delivers trusted data. AI creators know you can't have trusted AI without trusted data. Today our customers use elation to build game-changing AI solutions that streamline productivity and improve the customer experience. Learn more about Alation at Alation.com.

0:01:05.1 Satyen Sangani: Today on Data Radicals, we're excited to have Taylor Culver with us, the Founder of XenoDATA. A company that helps organizations drive real measurable outcomes through smarter data management. With over a decade of experience in data, Taylor has worked with some of the biggest global brands. Think Anheuser-Busch, Kraft Heinz, and Ulta Beauty to streamline their data processes and unlock serious value. Taylor's journey started as a data leader where he launched an industry leading analytics program that eliminated customer complaints while saving thousands of hours in productivity. Taylor, welcome to the show.

0:01:36.5 Taylor Culver: Pleasure to be here. Thanks for hosting me, Satyen.

Career journey: From M&A to data management and founding a company

0:01:39.6 Satyen Sangani: So like me, you actually, interestingly started your career at M&A and in New York. Tell me a little bit about sort of how you made that journey from M&A to data management. 'Cause it's not the sort of standard path.

0:01:54.6 Taylor Culver: Sure. So, I'll qualify this with, I am probably maybe a non-traditional data professional. However, I find a lot of data professionals to be non-traditional at the same time. I actually wanted to be an entrepreneur coming out of college, and I wanted to learn how businesses work, and that's what led me to be in a corporate development type role. Because being in those roles, you get to understand how the CEO is thinking, what's going on in the market. You get to understand how businesses are delivering value to customers. So it gives you a real high precipice to look at what your industry is doing. So I saw it as a great opportunity to learn. Now how did I get to data management? Is it's just a journey of curiosity and for me it was, okay, so in finance you need to get good at analytics.

0:02:39.3 Taylor Culver: So eventually you get good at that. Then you become limited by the data available in the systems you're using. So then you learn enterprise systems and then you get frustrated with those. So then you learn SQL, so you can get access to the database. Then over time you learn that the systems aren't capturing the data you need, so then you learn object-oriented programming and then so on and so forth. So for me it was just curiosity. And I think what kind of keeps me in the field is that watching kind of people's eyes light up when you're able to unblock them when they're performing business analytics is kind of what keeps me in the field. Because things that are probably simple for me and simple for you can be pretty extraordinary for folks just kind of dipping their toe in the data. And that's what's fun for me. So that's why I am still in the field and why I enjoy it.

0:03:27.2 Satyen Sangani: So you mentioned a journey where you sort of progressively became more technical and started by to your point, following your curiosity, develop this passion of of data management. Did you do all of that while you were in your first job at Hertz? Or was it a progression through different roles that you had had?

0:03:46.1 Taylor Culver: Yeah, it's become a career progression. So like, think about it, when I was in finance, I realized that okay, my work was stopping and starting in PowerPoint and I kind of wanted to roll my sleeves up a little bit more and go into a product role where I could work with a customer and take a little more ownership of the business. But as a product manager, I was non-technical. I didn't have a technical bench. I didn't know how to code, I didn't know engineering. So the only place that I was able to kind of secure a role was as the product manager of data. And then over time it just became necessity to learn. So as I progressed within my own career when I was gainfully employed, was going from basically a BI lead to taking over product, to taking over licensing, to taking over architecture, to taking over governance, to ultimately working for the CEO of the company and running data strategy.

0:04:37.1 Satyen Sangani: And what I find is that those first steps of sort of getting to a technical capability are often the most intimidating because people haven't been academically trained. To your point, yeah, you were in finance, which is numeric, but not necessarily technical as it were, although there were aspects of technical, there was technical aspects of finance. Like what made you decide to take the first... Was it frustration that made you decide to take the first step? Like was there just some, tell us that story, like when did you actually just say, crap, I'm learning this SQL thing?

0:05:06.7 Taylor Culver: For me, it's curiosity. I've got a growth mindset. I wanna learn, I wanna challenge myself. And for me I, there is a point of frustration, but at the same time it's like, I should be able to figure this out. And I just kind of love rolling up my sleeves figuring out how things work. There definitely becomes like a cap on that after a while where it's like, okay, this is a field of infinite possibilities. I'm gonna park my knowledge here and I'm okay with that. But for me it was a combination of curiosity, wanting to grow and frustration with not being able to solve the problems I wanted to solve.

0:05:37.3 Satyen Sangani: And you said you wanted to be an entrepreneur and sort of always knew you wanted to be an entrepreneur at least as much as you did, sort of leaving college. And of course you are an entrepreneur now you run a company that you started yourself. Tell us a little bit about the company and what does it do and what was this founding story for how you decided to start it?

0:05:55.1 Taylor Culver: Yeah, so, I think it was Steve Jobs that famously said, if you don't know what to do, wait tables. And I spent the first decade of my career hemming and hawing on what was the idea that I wanted to solve. And just no ideas were coming to me. I had no ideas. And for me, when I was working in data governance, specifically at the company I was working at, I was trying the best technologies. I was working with some of the brightest people I know and I just couldn't get things done. And it was very frustrating to me and I thought it was a reflection on my own inability because as I mentioned, I was still kind of learning data at the same time, but I learned that, that's not just me having that problem. And that was kind of my aha moment. And I think the universe was just kind of beating me over the head that, "Hey, this is the problem you need to solve and this is your calling to entrepreneurship and go do it." And so as I talk to other data leaders in different organizations, large and small, they're facing the same problems as me. So I kind of made it my own mission and vision to go out and help people solve that problem. And that's what this iteration of XenoDATA is doing.

The data-business people problem

0:07:01.0 Satyen Sangani: And can you describe for us what those problems are and what problems you wanted to solve at the time?

0:07:07.5 Taylor Culver: Yeah, so I think data leaders face really two core problems. The first one is getting executives to sponsor and sustain sponsorship of the data strategy. And then two is getting Crossfit functional business resources to work together in a way that isn't impaired by resistance. And where people use confusion as an excuse to disengage from the data program.

0:07:29.4 Satyen Sangani: Tell me more about that point. What do you mean by use confusion to disengage?

0:07:33.9 Taylor Culver: So I think that for data people, they love data, but for everyone else it's a pretty obtuse, boring and complex topic. But the business needs it to do well, right? So it's kind of a wishy-washy engagement. The challenge is that a lot of data leaders try to teach business people how to think like data people, which I think is a huge misstep in our space, rather, I think it's more about how do you enable them to do things that are ultimately going to get them where they wanna go without having to teach them every specific detail along the way.

0:08:05.9 Satyen Sangani: And can you give me sort of that exact example in real life? I think I have a sense for what you're saying, but I'd love to just get a little bit more texture and like what does it look like to be a data person teaching a business person how to think like a data person, and what is it like to sort of help them on their path without having to teach them all of the dirty audits brings the data?

0:08:26.1 Taylor Culver: Yeah, so in many cases data leaders are teaching business people how to use tooling, which is often not a needed skill. And it's like, whoa, this is too complicated, this is too far. I think data leaders can really help business people with two specific things. One is problem identification because a lot of times people are at such a high level with their problem statement that it's impossible to execute on. And then secondly it's requirements. Okay, so great, we know what the problem is, but what data do you even need? So before you even start talking about technologies or before you even start talking about data, you need to understand very specifically, one, what's the problem you're solving? And two, what's the data you need to solve that problem? And going through those exercises can be pretty challenging and hard, but also very eye-opening for business stakeholders.

0:09:16.6 Satyen Sangani: Yeah. And are there skills in your mind that help people identify problems better than others? Or is that a learned practice? Like how does one get good at that and how does one think about that?

0:09:29.5 Taylor Culver: Yeah, it was a great question. I think it's experience honestly, and being able to think commercially as opposed to a new insights or something like that. That's probably where I was very much helped working in a corporate strategy role is I understand how the business needs to think to generate and create enterprise value versus how do we improve data quality. But it's really a lot of listening, a lot of active listening to what are they saying, helping them reframe it and trying to link it back to the financial statements if possible. But I think that's the hardest part is I think a lot of people don't know what their problems are or they just see a piece of the problem. So a big thing a data leader can do because they can work horizontally across the business, is align different functions to a common problem, by just kind of bringing clarity to, "Hey, these six people are talking about the same thing, really." So like I said, it's a lot of active listening and kind of bringing clarity and alignment to how people think about problems within the business.

0:10:26.9 Satyen Sangani: It's interesting 'cause people often talk about sort of this blind men and the elephant phenomena where there are different people who see different parts of the business and different people who see the different parts of a problem. And interestingly, you're framing the data leader as somebody who actually has the purview to be able to look at different pieces and sort of kind of envision the elephant, which in theory is what data data's supposed to do. But I don't see a lot of data leaders who are, I see some, but I don't see a lot of data leaders who are really good at sort of saying, great, like here's what we're seeing in marketing, here's what we're seeing in sales, here's what see we're seeing in finance and product and, I'm gonna tell you one story in one picture.

0:11:02.0 Satyen Sangani: That's a pretty tall order. And often, it kind of even comes to sort of the same skill set of a CEO or somebody who's sort of trying to bring together the business in a much more functional way. Do you view data leaders as having to do that? Is that the ultimate sign that you've got a great data leader?

What makes a great data leader?

0:11:17.3 Taylor Culver: You nailed it. The data leader has to act like a CEO without the budget or resources or influence or power. It's a very handicapped role. And the other thing is that data leaders generally have a technical apprenticeship and then when they get to that director, head of product manager, VP level, their job is no longer technical. It's 100% political. And first off, a lot of data leaders are in technical fields because they don't wanna work with people.

0:11:43.9 Taylor Culver: And they're faced with an enormous challenge. Now, not every data person is afraid of this problem and they're willing to go at it, but it's a big mindset shift and you really need to drop the data at that point and really just focus on people problems and needs. And that's very, very difficult. And it's a first, you need to be willing, you need to be able. And those are two hard things for a lot of data leaders to do. What I like with the data leadership role is taking up and coming ambitious business professionals who are looking to kind of cut their teeth a little bit and go into a data leadership role where they can get a taste for what it's like to influence and lead without formal authority. A good place to really practice their strengths and capabilities without kind of the same level of scrutiny that a GM would get or a P&L owner of another sort would get. So I think it's a great developmental role for up and coming talent.

0:12:40.8 Satyen Sangani: And are you often then selling to data leaders who have these technical backgrounds? And is there a sort of consistent refrain of pushback that you get when you're trying to sell your services and your capabilities into those folks? What's the dynamic of where you're successful and where you're not successful? It sounds like one of the ingredients is selling to people who are business people who wanna use data in order to do something, as opposed to data people who wanna use technology to make something happen.

0:13:05.5 Taylor Culver: So what I find is that I'll work with data leaders who are at their wits end or data leaders who've hired many different solutions and none of it makes sense to them to kind of move forward. So everyone we work with is what I would call a data leader. Sometimes they're less experienced, sometimes they're more experienced. It really depends. It's more about the intent of the data leader than their title or position and their openness to help. Because a lot of data leaders are very smart and they know how to get there at least they think they do. But the challenge isn't about knowing what the destination is. The challenge is about navigating everyone else in your company to that destination. And that is a much harder challenge than just knowing what the right data architecture or strategy should be.

0:13:56.5 Satyen Sangani: And when you come in obviously with an approach and you say, look, the biggest problems you have are both understanding the problem that you're trying to solve and align people to solve that problem, or at least to identify and then solve that problem. That seems like a very obvious thing to say. And yet my guess is that people who are unsuccessful don't always recognize it. They don't even recognize that's their problem on some level. So there's probably, to your point, there's kind of two people. There's kind of the early stage person who's learning, therefore open-minded because they have a beginner's mind. And then there's the people who are sort of experienced in their career and probably a lot less willing to be vulnerable.

0:14:35.6 Satyen Sangani: Is that true or do you find that there are people who are later in their career who sort of tried it one way and say, Hey, I've not been able to be successful by focusing on technology and process.

0:14:43.4 Taylor Culver: Yeah. And then there's a third tier of the people who've done it, suffered the slings and arrows and want someone else to do it for them and they just wanna measure the outcome.

0:14:55.6 Satyen Sangani: Interesting.

0:14:55.6 Taylor Culver: So, yeah, so like a really tenured, successful CDO, it's gonna be really hard to find those lieutenants for those teams, right? And it's like, who's gonna go out and drive those conversations with use cases? I know they get prickly, I know they get political. Can we throw a third party in there to have these conversations and then we can debate the quality of the third party versus our ability or inability to communicate with one another.

0:15:19.3 Satyen Sangani: So tell me a little bit about like, so you have a services capability. Is that primarily you or do you have a large team of professionals behind you?

Growing (and adapting) a data consultancy

0:15:27.5 Taylor Culver: Yeah, so I have a team of professionals who I work with, depending on the scale and size of the engagement. I do a lot of one-on-one 'cause I enjoy that but yeah, it's a mix.

0:15:37.9 Satyen Sangani: And you yourself have developed software called Data Workbench. Tell us a little bit about that.

0:15:43.8 Taylor Culver: Yeah, sure. So, when I first started my journey was XenoDATA. It was primarily time and materials consulting. Over time, I found that those engagements kind of, it's like, oh, interesting new sign insights, but they don't really yield much change. So, I didn't really love that offer. Over time we developed templates to where, here's what you need to do over time, but again, the change wasn't sustained because there was really no visibility or consistency into that. So inevitably our templates became our software solution. So our offer kind of evolved into a done-with-you services model where we're working almost as like a portfolio manager alongside our clients. So helping them tackle and triage real issues on a periodic basis. And then our tool is kind of our collaboration tool, not just with our data leaders, but also with their stakeholders.

0:16:42.2 Taylor Culver: So we can get real time visibility into what's going on. Because if you think about it, when you go to communicate the impact to your data strategy, different stakeholders care about different things. So you can't just be like, here's a PowerPoint for everything that's going on in data. No one will care. So you need prescriptive analytics on each kind of component. Like how are you doing with use cases? How are you doing with governance? How are you doing with whatever engagement roles and responsibilities.

0:17:06.6 Taylor Culver: And then being able to report out on that and share kind of real time status, not only where your data strategy is, but also how your documentation's progressing, who's helpful, who's not helpful. And what I find oftentimes is that gaps in documentation are more telling than anything because if you don't have data stewards for a certain domain or you don't have business people giving you definitions, those are different problems to solve versus I can't get anything done. So we offer essentially a done-with-you services model. Now it's complemented by software, but primarily we remain a services business.

0:17:38.7 Satyen Sangani: And what do those templates and processes like, what is that done-with-you engagement look like? Where do you start and what are the sorts of things that you are doing day one in order to discover problems for your customers?

How data leaders can tackle business problems in 3 steps

0:17:50.3 Taylor Culver: Yeah, so I kind of see it as a three step process and data leaders jump into governance without doing the first two steps. And this is the biggest kind of value add we see with our customers is data leaders need to get out of their seats and go talk to people and formalize the way that they're engaging with people. No different than a salesperson does. So no different than a salesperson calls 100 people a day. Data leaders should be talking to people all day long.

0:18:14.6 Taylor Culver: And what they need to be talking to them about is what their problems are. And not to set up some gripe session about, oh, this data sucks, but hey, what are you trying to solve for? Like, what matters to you? And then over time, use cases will emerge from that. And then the more people you talk to, you realize that there are parts of different communities and different communities have different needs and those different needs can be aligned to use cases which ultimately flow into projects. And then when you've kicked off a project, then you can start going down data governance work activities. But if you jump the gun and go to the, "Hey so and so executive said, you're the data steward. Go steward this data, they're gonna say, go pound sand. So.

0:18:52.0 Satyen Sangani: Yeah. And so in your case, what you do is you basically try to get the people that are working with you to engage with their business counterparts and really learn about what they do. And are there formal templates about how you interview one of your business counterparts? Or like what sort of questions you ask, or who are you're supposed to talk to within an organization?

0:19:08.7 Taylor Culver: Yeah, so we have a whole suite of templates and libraries that we use, but every customer is different, right? I don't think there's like one universal way to solve data for any company. I think everyone's a little bit themselves, so we've got the schema, but as far as the content and the attributes, that kind of is fluid with our clients.

0:19:28.3 Satyen Sangani: And do you find that customers will take your software and use it after you've sort of departed? Like are you able to teach them to fish? Or are they enabled to go sort of take these skills and do it themselves?

0:19:37.7 Taylor Culver: Yeah, so with our customers, we usually have like a heavy upfront reset. This is what's going on, this is where we need to focus. But then over time it goes more into maintenance. And then from maintenance it will usually go into the software. So some people are gonna be self-sufficient, and then they're gonna be able to move into that. So, like our longest term client we've had since 2018, since we started, they actively use our products. Most of our customers are working with us for three plus years.

The challenge (and opportunity) of data leadership

0:20:07.5 Satyen Sangani: Yeah. Now, one of the things you've said, in prior conversations is that sort of ambiguity and the absence of clarity are sort of the terrible secret in this space. And I've actually separately had people who are maybe a little bit more cynical, say that it's actually not just the terrible secret, but it's actually also something that many data leaders use to their advantage, because something that's sort of arcane and not very well understood can be a defense mechanism and in some sense job security. What's your take on that?

0:20:37.5 Taylor Culver: Yeah. It's a real thing. I think the challenge for data people is that people don't care and don't necessarily understand data outside of data's important. Maybe do something with AI. And the common conversation around data professionals is like, what's the best way to do this? Which is a good conversation to have, but it's not a conversation that the business really cares about, right? So I think a big miss by data professionals is worrying about the perfect way to do things. Because in this space, the bar is very, very, very, very, very low, right? And perfect is very much the enemy of good. And if organizations just did okay with data, it would be revolutionary for most Fortune 1000 companies, right? Because a lot of companies are struggling with very, very basic problems, right? They wanna experiment with nascent technologies, but still departmental silos of information and summing and counting information is still a real challenge for a lot of big companies. It may surprise you, right? And it's always surprising. But then you've got data leaders who don't have the best intents, right? To your point, you've got ones that are defensive looking for job security, right? And you also have some that misrepresent the potential and sell a greater vision and they can deliver. But on both sides of the coin there, their clock's ticking at about, 18, 24 months.

0:22:04.0 Taylor Culver: But the challenge, and I just wanna say this, is that the probability of being successful as a data leader is very, very low, probably worse than, hitting a baseball in the major leagues, right? Because it's not about driving your own success, it's about driving others success. So the compounding effect of being successful at that is very hard, right? So what data leaders should just own right, is the path to me is probably going to be fraught with failure. But I need to be able to pivot and I need to be agile and I can very much serve myself by adhering to a common set of principles, which I'm going to practice consistently and continually adapt and adjust in the way I engage with my stakeholders and identify their problems and lean in or lean out on data management techniques or delivering certain solutions.

0:22:55.0 Taylor Culver: And it's at their discretion, right? It's ultimately that data leader's judgment. So I think the job's hard enough. If you go to be defensive, your path to success is probably very limited. And then if you're gonna tell a tall tale again, your clock's ticking. So yeah, I definitely see that. I definitely see, and like I said earlier, is that it comes down to intent, right? Do you genuinely wanna help people in your business solve problems with data? Do you genuinely want to grow? Do you genuinely recognize that there's not a magic bullet to doing this? Those are the data leaders who will be successful despite facing adversity.

0:23:29.8 Satyen Sangani: Yeah, I think it also depends on how people both define their jobs and are given the definition of their jobs. I can imagine and know that a lot of CDOs are hired under the sort of guise of or under the problem statement of, "Hey, you need to come in here and help us organize our data." Which is sort of a never ending specifician, like impossible task in any organization of any real scale, as opposed to, you need to come in here and help us drive these three strategies forward which is far more rare because nobody ever thinks of the CDO as somebody who's gonna drive a strategy forward. They think of the CDO as somebody who's sort of, to your point A, enabling function, maybe in the best case scenario for skills and capabilities. But often the people who are doing the hiring aren't viewing that person in the way in which maybe the best intended and best, most experienced people would like to be viewed. And so there's kind of this strange structural dynamic. Have you found cases where people have successfully changed that mindset and really sort of re-define the role for themselves when they've been hired in one way and are trying to get to the new model that you're espousing?

0:24:37.0 Taylor Culver: Yeah, I think that data leaders have a definition problem, which is kind of ironic. And if you look at a finance professional, career path is pretty straightforward. Data professional, no way. So a CDO has a different challenge. So when they come in, they have to immediately go on the offense, which is, this is the data strategy, this is my job, this is what we do, this is what we don't do, here's our metrics for success. If you don't like this, please fire me because this's gonna be a giant waste of everyone's time if we continue down this path. Right? And those are hard conversations executives need to have. On the other side, data leaders who tend to be the senior most data professional, usually promoted up from the bottom all the way to middle senior leadership, who have not had that experience of a CDO, who do not know what it's like to be at that executive level. It's up to them to really just make that conscious choice of, "Okay, do I really wanna take this leap?"

0:25:27.1 Taylor Culver: Which is usually the first question I ask 'em and a lot don't, right? Which is really telling now, if they do, we need to reorient what the data strategy is, show what the current capabilities are, show what they're currently focused on or over hedged on and show where the gaps are and show where we need to go. Now, executive leadership could come back to that data leader and basically say, no, we just need reports. But that's a good outcome. Maybe not for the career of that data leader because they wanna try different things, but it gives them clarity on what their mission and charter is, which is, Hey, the data strategy is BI serve operations. All we need. We don't need cross-functional transformation. And I think that not every executive leader is thinking data is a innovative function. I've talked to a lot of CEOs and a lot of people see data as just a reporting function still. So those expectations need to be set and reset and visited often with senior leadership, where possible.

Is data a strategic function or an enablement function?

0:26:22.9 Satyen Sangani: How many companies in your experience, or what percentage of companies truly do see data as sort of a strategic capability and are actually acting to further that view? I know it's a subjective question 'cause you haven't done any informal surveys, but like, is it 20% of the time? Is it 10% of the time? Is it 50% of the time?

0:26:38.6 Taylor Culver: I think there's talking about it and there's doing something about it. I think every executive leader is talking about data and AI as far as doing something about it between small POCs or testing it in different areas of the business. Few and far between are really committed to data as a transformative function. I've worked with one CEO in my career who is by all means a thought leader on the topic, but for a lot of businesses, data is enabling function that produces analytics. And even though there's a pain point around siloed information and inconsistent quality, which can become red herrings for data leaders to try to jump into some kind of cross-functional champion executive leadership probably does not want that because it can become disruptive and frustrating for people involved and it can put the data leader in precarious positions.

0:27:29.8 Satyen Sangani: Yeah, if you're listening to data, you have to be willing to be wrong. And that's obviously a hard thing to do.

Data buzzwords: data mesh, data products

I wanna get your take on a few sort of data terms and data trends and just sort of maybe thumbs up, thumbs down in a little bit of why. So data products and data mesh. What's your view on that? Important, relevant, useless, stupid? I don't, like what's your take?

0:27:52.0 Taylor Culver: On the data products? These are all interesting words. I posted one day, like I don't get like data mesh and I got dog piled by someone saying, well, as an expert, you should care about this stuff. And I go, in three years people aren't gonna care, right? And, since then, data fabric is now a Microsoft product and not even a product, but a package of products and data mesh is kind of okay, interesting. So these things come and go at the end of the day, you've got a database, a reporting tool, some kind of copy paste ETL like tool and then some sort of data catalog on top of it, whatever you want to call it and how you wanna use it doesn't really change and hasn't really changed in decades. But I think data product management, I really like that more so than like what is a data product? Because I think that data people are coming out of like a BI function, which is less customer centric, whereas a data product manager is gonna be more customer centric.

0:28:46.2 Taylor Culver: So it's more in line with problem identification and engaging stakeholders. So I really love the idea data product management because you can weave governance and architecture into product management and do kind of like data strategy light within a product management function. So I really do love product management as far as what a data product is. I think it's subjective. And as far as data mesh, I still don't understand it, nor do I honestly care too. So.

[laughter]

0:29:12.4 Satyen Sangani: Curiosity extends in sub directions, but maybe not others. That's totally awesome. By the way, like I think the product managers and product management, like people are always like, ah, get out of the building, you gotta go talk to your customer, which is kind of where you started. So I would expect you to say that about data products. And by the way, I see it the same way. I think there are people who take this view of data products that they're rigid and well, and they're very strongly typed. Like it's a certain type of data set with a certain type of interface, with a certain set of things and attributes to it that contains a certain set of things. And I think maybe there's value to that. I like, I haven't seen a lot of that in reality at an action. Some, but not a lot. But I think the sort of broader idea, which is, look, this is a product, it's got an end user with a set of use cases, it creates value, it needs to evolve and be versioned. You have to have a customer for it. I think that's super valuable and super useful.

0:30:03.1 Taylor Culver: Well if you think about it by its very definition, a database could be a filing cabinet. So, it's an abstract thing. It's truly an abstract thing.

Why hire a data engineer?

0:30:11.6 Satyen Sangani: Yep, for sure. So you mentioned data fabric. Is this like, random product, so I won't ask you about that. Tell me about data engineering. So what we are finding is that data engineers have a lot more influence and buying processes and running and owning the data strategy. What's your take on that? How do you feel about that function and role relative to the overall data function?

0:30:32.8 Taylor Culver: I think they're an important member of the team specifically when it comes to integrations and data modeling. Usually the organizations I work with don't need data engineers, but they hire 'em anyway, which creates enormous complexity where you find ETLs finding their way to Python and people building analytics and object-oriented programming. So it's a lot of horsepower. A lot of companies are still doing very, very basic reporting capabilities outta enterprise and client facing technologies. So I think it's a very valuable skillset. But I would put organizations are probably underweight in focusing on product and probably overweight in focusing in technical capabilities. So I think it's an important role, but it may be mishired to solve the problem that the organization needs solved.

0:31:21.0 Satyen Sangani: So you would advocate for hiring fewer data engineers and hiring more of what?

0:31:26.1 Taylor Culver: It depends on the organization. So as long as there's a clear cut portfolio of use cases linked to measurable business outcomes, then I would staff to those projects with data engineers. So I see it as an enabling function for a project and a project as a function of need. So for me, I think organizations need to do a better job identifying who needs to participate in the data strategy and what those problems look like than, technical capabilities. Just in general.

0:31:57.2 Satyen Sangani: On some level in a product organization, you've got this ratio of product managers to engineers and often people over hire engineers, but, under hire product managers, I've actually been personally guilty of that, different story for a different day. But in your case, what you're saying look is like, look, if you don't have a library of problems and a clear set of use cases and clear understanding of customer discovery, you might wanna hold off on pumping a whole bunch of money into engineers that are gonna build stuff that may or may not be useful for the end customer.

0:32:29.3 Taylor Culver: It's where data leaders get into trouble. How often is the story of hire the data professional? They launch a multi-billion dollar project. They hire a bunch of people, the value's questionable and then everyone moves on.

0:32:42.9 Satyen Sangani: And gets frustrated 'cause They're like, yeah, data sucks. And like, we're never gonna get this right. And we've already tried. We spent so much money if we couldn't get it right, nobody can get it right.

0:32:50.8 Taylor Culver: And there's so many smart and talented data engineers that would kill to have that business buffer or air cover to be like, just tell us what we need to do. Engineers eventually when they become PMs will face those challenges, but a lot of engineers are, five years of experience, may not wanna even do that. So they're hungry for people to be that frontline leader between the business and then the technology solution. 'Cause they wanna build cool stuff. And without someone bringing clarity to what the problem is, it's very hard to build cool stuff. And then data engineers or data scientists end up becoming essentially business analysts or BI analysts. So they become overkilled for what they're trying to do. And they probably keep themselves busy by creating complexity, which adds to confusion, which creates a different kind of problem. I think it's a super important skillset and I think a lot of people should learn the ins and outs of it, but there's probably more supply than demand to be perfectly honest.

Strategy: Data offense vs. data defense

0:33:50.2 Satyen Sangani: So maybe switching gears a little bit. Data professionals often talk about sort of defensive governance and offensive governance and a lot of, I think this business value conversation of trying to solve your counterparts problems kind of veer into the world of offensive. Do you do work to help customers, comply? Is that a part of your portfolio? Do you find that's a step that's either before or after what you do? Tell us a little bit about that work and does that work differ or is it all kind of the same thing?

0:34:16.9 Taylor Culver: It's funny you asked that question because just the nature of people we attract tend to probably fall into that offensive category. And then for whatever reason this year we started working with a financial institution and we've been approached by a pretty large bank who wanted to participate in the business as well. And you gotta think about it, rewind, gosh, 10, 15 years to the financial crisis, which basically created a lot of the demand for what is now modern day data governance best practices, right? But now you've got a bunch of banks who put in place the data governance, which we would call defensive data governance to meet regulatory reporting requirements. But they still have to have a portfolio of use cases to drive business value and the opportunity for driving efficiencies out of banks is in the billions.

0:35:06.0 Taylor Culver: So they're trying to figure out how to do that. But they cannot let go of that risk side. So, where we've needed to evolve when we started working with financial institutions is adding risk as a pillar to our data strategies, which is essentially, putting controls around the methodology around, hey, here are the policies for how you work with the business. Here are the policies for what a use case is. Here are the policies for categorizing a risk-based use case versus a value-based use case. I think they can be managed in the same thing. The area where you need to be touchy as a data leader. And I was talking to one of my clients this morning about this, is you cannot be a policeman, a data leader when they go out to the business saying you have to do this, you have to do this documentation.

0:35:50.1 Taylor Culver: The person's gonna wanna stab you in the back the second he leaves the room. It's like bringing me work without solving a problem. Thank you. Let's never work together again. So as data leaders within financial institutions have to go to a value-based approach, they need to change the way that they work with the business, which is, "Hey, my hands are tied here, I have to do this because the regulators wanna do this, so I need this from you." But at the same time, "Hey, I also wanna help you solve problems." And that's a tight rope to walk and not having been in that role, I can't speak to any experience, but I would assume it'd be very, very, very difficult.

AI and data leaders: Resetting expectations

0:36:22.0 Satyen Sangani: Yeah. So interestingly in this world of like Whizz Bang AI and everybody trying to talk about and do AI, you are finding that in your business is at leasing some, if not a lot of literal traditional defensive data governance. Are you seeing people trying to do AI and call you up and say, "Hey, I need to do AI and like let's do AI and AI's an important thing." Like do you get that problem statement at you? Or are people well enough familiar with your brand to know, like your response is gonna be, "Hey, what's the problem you're trying to solve?"

0:36:51.0 Taylor Culver: It's funny, where data leaders need the most help with AI is getting the C-suite off their back and finding use cases to find AI. And to be honest, like I think AI is super cool, right? But I'm bearish relative to the market sentiment. Not because I'm bearish on AI, but people are way, way, way in the stratosphere about its potential for impact. I am not with them, right? I see it as a, something to work alongside you if you're a developer, a copywriter. Super useful, right? But as far as revolutionizing the way most people do business, no, it's a compliment. It's gonna enhance your comparative advantage period. So yes, we do see use cases for AI, but they're very much in the minority. And I feel the same way about AI that I feel about data monetization, which is another, a trend from a couple years ago I've worked with enough hedge funds trying to push data sets, trying to get people to buy and license data. And the commercial upside there is not huge. And the complexity of meeting those needs based on the available data that a lot of these companies have, again, is also limited.

0:37:57.3 Taylor Culver: So I think there's a lot of big talk around AI and while I think it's very interesting, I think the hardest thing for data leaders to do is kind of reset expectations around, "Hey, I worked with the business, here's the problem they need. Here's where AI could help. Here's the ROI, is this compelling?" And eventually executive leadership will be like, oh, how many a hundred thousand dollars use cases do we need to make, certain staff more productive? And how many times do you need to beat over the head to learn that you probably cannot build your own product off a generative AI API? There's not much value in that exercise.

0:38:32.3 Satyen Sangani: Yeah. Although it's a really tough situation to be in because I think a lot of these executives are under pressure from their board who themselves are listening to investors and then everybody's sort of in this hype cycle where if you're not doing something, you're seen to be laggard and real decisions get made off of the fake news. And so, you're in this position as a data leader where you're like, well, I'm still learning about this stuff. I don't even know what use cases we have. I don't even know what like, how this stuff can help the use cases that we have. But if you're not doing something, people will perceive you as being sort of behind. It's a tough thing to manage because even if you do have real conviction of have learned all about this stuff, it may not be the case that you are listened to, which is, often of course a problem for data leaders.

0:39:14.5 Taylor Culver: I think a powerful way to be heard sometimes is saying, let's stop or I don't know. And I think it's a scary thing for a data leader to say, because they're expected to know everything about everything. And sometimes if you say, "Hey, my use case portfolio kind of stinks, maybe we stop, or hey, I really don't know, maybe we bring in a third party or another AI expert to take a look at this. Let's see what they have to say." I think that data leaders need to be humble in where they are and what they know because it's a pretty vast topic where no one knows everything to reset expectations. That would be kind of my pearl of wisdom for that group.

0:39:56.4 Satyen Sangani: Yeah. By the way, like, I would, I'm long term bullish on AI, but I'm short term to your point, relative to market sentiment and relative to where the hype cycle is pretty bearish. 'Cause Like, if people come up to me and they say, we should just have gen AI models write up everything that's in the catalog and were it's so easy then of course you wouldn't need all these people to describe all this very sophisticated, nuanced data. And that has a whole bunch of invisible complicated process behind it to produce it. And yet, people are like, oh, we're just gonna turn GenAI models onto every table and column and every single thing in our database across all of our databases. And what you find is that it's sort of a garbage In garbage out problem, but it makes your discoverability problem actually worse, not better. So it's super interesting, even in our very narrow slice of using it for something that I think people think that is like so obvious to use it for.

Data is a people business: the value of trust

0:40:45.4 Taylor Culver: I've had the same experience with my customers where we can build AI into our product and every time they're like, meh. So if my customers are saying it's not valuable, then it's not worth building. Right? And that's fine with me. What I think that data really comes down, and especially the world that you and I are in, it's all about trust. And people trust people. No executive is gonna go to a data catalog and be like, this is the definition. They are gonna call up John or Sally and say, is this right? And as long as John or Sally says, is this right? It's right. So data is a people business. It's not a technology business, especially the data management side of things. And I think that it's that trust that AI will need to leapfrog somehow to where people trust a machine more than a human being when they go to make a decision, which may happen, at which probably will likely happen with a lot of low hanging fruit decisions. But if you're running a multi-billion dollar company, are you gonna go to the GenAI output or to the person who's been working there for 20 years, who knows every exception, who knows what it should be, what it is, and can look at a report and tell you if it's right or wrong without even knowing how the data's put together based on how well they know the business.

0:42:00.8 Taylor Culver: So a lot of the knowledge, which I keep telling people, is that the most valuable data is inside people's minds and hearts. And accessing that information is what's gonna unlock data potential within your business. And if you cannot get to that data, which AI will not be able to get to, you're not gonna be able to get the leverage out of data that you're looking to within your organization.

The future of data management: What’s made an impact? What’s exciting?

0:42:22.6 Satyen Sangani: So, you've talked a lot about different technologies and I think earlier you were like, look, it's kind of, everything's sort of like basically just, a whole bunch of scripts with a spreadsheet, with a database, with a catalog, with some usual visualization of, and these technologies have kind of existed since the beginning of time, but it's kind of the same stuff. Is there stuff that you look at that you're like, oh wow, this is actually really useful, and are there tools and capabilities that are emerging or existing that you're excited about?

0:42:50.0 Taylor Culver: I was 10 years ago because things used to be a lot slower. And for me as a analyst, it would be like, why does it take 10 minutes to get a report? And now because of cloud computing, everything is so accessible, inexpensive, fast, and scalable. I think technology's way ahead of the game when it comes to data. Like technology is beyond the needs of most organizations now. Top 1%, yeah, sure, technology probably has a long way to go, but most everyone else is just looking at, basic analytics, staying organized, common modeling across departments, stuff like that. So is there anything out there that really wows me? Not anymore. I think cloud computing really revolutionized the data stack and I don't think we talk about that enough, to be perfectly honest. Is that that's what's been so incredible to watch in the past 10 plus years.

0:43:45.6 Satyen Sangani: Super helpful. Tell me a little bit about XenoDATA, your practice. Where do you see it evolving? How do you see it changing, given the current world of data and AI and what are things that you are super excited about and directions that you're super excited about moving in?

0:44:00.7 Taylor Culver: Yeah, so it kind of goes back to my own experience as a data leader. Which is, it was probably one of the most isolating and frustrating roles I've ever had in my professional career. And I think the data leadership role is harder than being an entrepreneur as far as I'm concerned. And people say entrepreneurship is hard and it is very, very hard, but being a data leader is even harder. I like working with data leaders who are looking to get their organizations unstuck, reorient their data strategies to value, get their teams around them, working with data in a way that makes sense.

0:44:31.9 Taylor Culver: Or just helping organizations get started with their data programs in a way that makes commercial sense so people don't get frustrated or confused by what's going on. And when you get stuck, you get to make actual decisions about how to get unstuck. So my vision is just helping as many other data leaders as I can, who may feel frustrated and alienated even by their own organizations looking to drive increased value and help their companies, see the potential benefits of data because I know what it feels like to sit side by side with a colleague and teach someone who's already very, very smart how to become 1% smarter and that's the best. And for me, that's success. And I still have friends, from when I was a data leader who like you sat down with me for five minutes and taught me this thing and has changed my career.

0:45:20.0 Taylor Culver: And for me at the time, it didn't mean so much to me, but I didn't really think about it. But it's been really cool to watch that and foster those relationships. So if there's a network effect of me teaching data leaders how to do the same thing, I feel like that's my mark as an entrepreneur. And it's all about helping people solve problems. And it kind of goes back to why my mission is to doing what I'm doing.

0:45:41.4 Satyen Sangani: Yeah. Which is an awesome mission and it's a really unique and cool position to be in the market. So we're excited and curious to see what you end up doing. If you have not engaged in or heard of Taylor, one of the things that I deeply recommend is that you go follow him on LinkedIn. He's got some of the most insightful, pithy and fun content out there on data. So I super enjoy it and often I'm a little jealous 'cause I'm like, ah, I wish I said it that way 'cause it's so well done and so well put. So really good stuff. And Taylor, thank you for coming on.

0:46:14.3 Taylor Culver: Yeah, Satyen and thank you so much. I appreciate the chat today. It was a good one.

0:46:20.5 Producer 1: Taylor reminds us that being a data leader isn't easy, it's complex and often political. But by staying focused on the real problems data can solve, it's possible to balance both defensive and offensive data strategies and drive real results. A little curiosity and listening can make all the difference. Be sure to follow Taylor on LinkedIn for more insights on data governance. Thanks for tuning in Data Radicals.

0:46:44.5 Producer 2: This podcast is brought to you by Alation. Your boss may be AI ready, but is your data. Learn how to prepare your data for a range of AI use cases. This white paper will show you how to build an AI success strategy and avoid common pitfalls. Visit alation.com/ai-ready. That's alation.com/ai-ready.