Inspired Execution

Chet Kapoor (00:31):

Welcome back to the Inspired Execution podcast. Each episode shares the experience and learnings of a world-class leader on the journey to success. The guests on this podcast are bold, brilliant, and not afraid to change. As you navigate your own path, we hope you feel inspired by their stories, lessons learned, and the vision of the future.

Today we are joined by a tech industry OG, Mark Nelson. Mark started programming computers at seven years old, and his first job was writing software for 911. He turned his childhood hobby into an incredible career, working AT&T, SAP, Oracle, and most recently Tableau. I outside of work, Mark is an avid runner and hiker. I loved our conversation about building a data-driven culture, using data to tell stories, and the power of combining data with human instincts. Mark, welcome to the Inspired Execution podcast.

Mark Nelson (01:27):

Thank you so much for having me here today.

Chet (01:29):

It's absolutely an honor to have someone like you that's so accomplished and has gone through so many different industries within the tech industry. Right? AT&T, SAP, Tableau, Oracle. We will get a chance to go zigzag through all of it, but let me start with one of your quotes. "Data-driven decision making. It's more than a buzzword, it's an approach worth implementing", and we talk a lot about it at DataStax. Where have you successfully implemented this approach? But more importantly, what did you find was hard to implement?

Mark (02:05):

Yeah, it's a great question because this is not an easy thing to do, and because it's not just a technology problem, it's a person problem, it's a process, it's all of it. You might know, I was successful at Tableau, although I will say I had a huge leg up because it was our business, and we happened to be blessed with the best tool in the industry, biased opinion, was free to all of us. So that hurdle was there, it was in our DNA to use the product. The product was there, it was an amazing product. But even there, a ton of challenges. One of the immediate challenges is everyone could create Avis. That's great until it's not. The flip side of that is you really have to have some standards. How do we trust, where did the data come from? What is the right data?

What is the version of profit that we're all going to look at? Because this is one of the things when everyone is a data analyst, that's amazing. That's exactly what you want because you want it in everyone's hands, but it can also lead to the sprawl and different interpretations and the success is really focusing on it. It is making it part of your DNA, and that starts both top down and bottom up. It is walking the walk as a leader, not just, Hey, I'm going to make this decision today, gut instinct because I feel like I need to. Really walking that walk and then bottoms up, enabling everything. Not just the technology in putting in people's hands, but the data infrastructure, the data processes, the governance and finding that right balance of control, and these are things that we all know where it's at and this is where you've brought in what you as an individual working in the business, the very data and questions you want to ask then merging those together into one thing that becomes super powerful.

Chet (03:46):

Have you found that people don't use their instincts because they've over-rotated on data?

Mark (03:53):

That is one thing that can happen for sure is people will over-rotate to believing that data is absolute and that if they search hard enough, the answer is there, and there's a quote I love, which is, "Every model is wrong, but some models are useful". And this is where you get to the right answers. Data is an amazing tool. We're at a point in time where the use of data is nothing short of the invention of the microscope or the invention of the telescope. We can now see things about the world with data that we could never see before, but it's still just showing you one part of the world.

The microscope did not magically answer all of our scientific questions. It just gave us a much more powerful tool to see things and likewise with data. And so yeah, you can get very enamored of the tool and start navel-gazing at the tool. And yes, if I look hard enough, the answer is in there, but in reality, the magic happens when this amazing tool is put in the hands of the human intellect. It is that combination of the tool with a human intellect that leads to these amazing answers because now the human just has more tools at your fingertips. You have more information that helps you make these decisions.

Chet (05:03):

Yeah. I find that because it's not mainstream enough, it's becoming so, but you would agree, even with Tableau, it's not there. Even with Excel from 30 years ago, it's still not there, but it's certainly getting there. We'll talk about AI in a minute. There is no AI without data, and so you need the data driven approach, but I still think it becomes a crutch for people not to make a decision fast enough. And they have to realize that you have to have the data, but you need to use your instincts to make the decision. Otherwise, the data is not going to give you the answer. The data's going to give you a perspective.

Mark (05:37):

Correct. That's right.

Chet (05:39):

So let's keep Tableau aside. If you had to give advice to our listeners about them creating a data-driven culture where it's not starting from scratch, if it's much easier to do it when you have two people and a dog in a garage, and you can start from that in the DNA of the company, you are in a company and you want to do this, what would be the two or three things you would tell them to look out for?

Mark (06:03):

A couple of things, as you get started, start small, start where you can make an impact. If you just go, Hey, I'm going to instrument my whole business tomorrow and go, that's one of those two-week projects, becomes a two-year project, becomes a project that never finishes. Pick the spots where you know you have a problem, you know that you have a problem that can be well instrumented, that you can get good data and really start there and then grow out because people will see this.

It will become a very virtuous cycle once people see the power of what you can do when you have good data, when you have a good data culture and a good data process. They'll want that and it will grow on itself. Like your quote, one of the ones I love is, "If we're going to talk about data then let's talk about the data. If we're going to go with opinions then just go with mine", and it's a good way to really cut through that. We're just not going to have a conversation if we're just going to talk about opinions, let's talk about the data and then let's go from there.

Chet (06:58):

Yeah, no, for sure. And so as you think about starting small, is it small data sets or is it small decisions or all of the above?

Mark (07:09):

It's mainly finding small in the sense of the scope of the problem that you're attacking. So yeah, so it's less about even the scope. It's something that you really are going like, "This is tangible. I know I can get my hands around this and I know I can go get the data and drive to a conclusion here because that'll get things going."

Chet (07:27):

That is so true. That's great insight because I think as somebody who wants to do this, should they also focus on the skill sets of the individual who is leading the effort? Is there a specific recipe of if you are going to change the culture and go for a data-driven culture, what are the behaviors or characteristics or skills of the leader for that particular project that you found to be successful?

Mark (07:59):

Yeah. Certainly for the leader, you want someone who's empirical. You want someone who wants to do this, right? That will use the data. This is the human part of the equation and not everyone is wired that way. Certainly growing up as an engineer, you go, "Well, come on, this is how engineers are typically wired", but it's not how everyone's wired. Some people really find data intimidating or even threatening like, oh my goodness, if I do this, my value as a decision maker is going to go down and they feel threatened by that and nothing could be further from the truth. Trust me that again, that human intellect will always be part of the loop and will be always the most important part of that loop. So choosing a leader who isn't threatened by it, is willing to put in the time and the learning, and then ideally, especially if you're looking to spread, who is going to be a good teacher? The perfect first leader for this is one who's going to help this have a viral effect throughout the organization.

Chet (08:52):

That's a great point. That's awesome. You've talked a lot about data literacy and insights. What is your recommendation on how people get there, right? Because I'm sure you got a chance to see many different organizations implement Tableau in many different ways. Which ones did it really well and why?

Mark (09:18):

Yeah. The real key is getting this into everyone's hands because the way those actionable insights come out is when the business user with the problem is enabled with the ability to do the analytics. When you have to ask someone else to come back with an answer for you, that loop gets too long. It is subject to the game of telephone. Like I asked you for this, you gave me an answer. Was it really the question I wanted to ask? I'm sure we'll assume all good intent, but the nuances of that question that I asked as the subject matter expert, did that really carry through all the data machinations that came through to get that answer? And so, one of the magic keys where the organizations that do this really well is that it is very distributed, right? It is not, there is the walled garden of a data analyst and the walled garden of a business expert and you pass notes back and forth.

That does not mean you don't need a data team. You very much do. You need clean data, you need standardized data for the things that need to be standardized, but getting to where, again, and there's a fine balance here where you've democratized that as much as you can. Where you've really pushed as much of the actual questions and answers with data out to the people who know what questions they're trying to answer is where the magic happens. And that is way easier said than done because there's a lot of art in finding that balance, especially around large data environments and highly regulated ones. But I have to say, some of our most successful customers were financial customers because it was so important to them. Finance, it is all data, it's all numbers. And so there's so much value there that they're very motivated to go and do that.

Chet (11:07):

There are many waves of how companies reinvent themselves. With digital transformation, you would agree was one of the things that caught a lot of people by surprise because mobile really took off very quickly and changed everything that every industry was doing. And I think same thing will happen to AI. We'll talk about that in a second, but I used to always talk about, as I talked to boards and CEOs, I always said, "If you are not thinking about digital transformation every day and it's not part of your talk track in at least two out of the 10 meetings you have a day, it is not going to happen." No matter how much budget you do, you cannot hire a chief digital officer, hire one, that's great, but it's important for you to talk about it and think about it every day. And I think the same thing applies for a data-driven culture. You cannot buy one, right? You cannot buy a team to do it. It has to be something you believe in and you have to talk about it on a regular basis. Is that fair?

Mark (12:03):

100%, you definitely cannot buy it. Yes, you need to invest in technology, you need to invest in, like you said, you should have a chief data officer for sure. That's great. They do not magically make your entire organization data-driven because you have a chief data officer. The world is not that simple. The reality of all of this is it has to get across the entire organization. And so there's this push and pull between yes, you want a driver and you want some parts of it centralized. At the same point, it has to become part of the responsibility of every organization, every conversation where the chief officer, digital transformation, whatever, can become a handicap where you go, "But isn't that person's problem?" It's like, no, no, no, it's not. They're here to help you, but they're not here to do it for you.

Chet (12:51):

That's awesome. Everybody's going to find this so useful. So looking to the future, I talk about there is no AI without data. We were talking about AI earlier. How do you see the world changing in the next few years?

Mark (13:13):

Yeah, it's a really exciting time. We're at this point where the convergence of our ability to collect data, our ability to store data, our ability to process data, and because there is so much data coming in, every day we have more data available to us. And we have more powerful tools that allow us to look at it in new ways. And again, generative AI is just the latest shiny object on those, which is awesome. But like you said, what is that driven by? Oh, because we could scrape the entire internet and we now have both the processing power, and it's fascinating that one of the main constraints now of generative AI is literally just compute power. How many cycles can you apply to it? But you can relate back to we had this problem 30 years ago when I started my career. My first job, we had a 20 gigabyte database, which was massive, and it took a three-quarter million dollar machine to store that data.

That now fits on my phone. I give 10 times that on my phone. We're going to see those same moments right now, these models they're running that take a data center to run, that's going to shrink. If you look three, five years out, that will run on our phones. And I'm so excited to see data driven decisions. Again, these models and this data and AI really taking off as the latest modeling world. I will go back to though that is still just a tool in the hand of humans where the magic will always continue to happen, I believe, is when those tools get in the hand of the amazingness that is the human intellect and people use those to go solve problems. I know there's been debates on has AI become sentient? You look at what the human mind can do. No, it's not even close. And where I believe the power is is when these amazing computational tools get in the hands of the human intellect, amazing, amazing things happen. And that is super, super exciting for the next couple of years to come.

Chet (15:15):

This is great, and we talked about this earlier. We really want the technology to go like we've always had it go. But this time around, I think we do actually need some speed breakers, speed bumps, right? Because I think we are... Not that this will become GAI, it is not going to get down and in the next two years become general artificial intelligence, but there are sparks that actually there's a possibility that we will accelerate that path and we just need to watch out for it. So getting some regulation on how we do it as a society, and I mean that at transcending countries, and it's going to be really hard given the state of the world. But I feel like this time around it's going to be different. You would agree?

Mark (15:59):

I would agree. And I hope that it is different because there is no transformational technology that cannot be used both for good and for bad. It is, right? It's a tool that you put in the human intellect and humans will find different things to do with it. And then the other quote that I love, "A sufficient enough advanced technology is indistinguishable from magic." And we're at this point where most people in society view generative AI as magic, and it is not magic. It's amazing. And again, it's transformational, but it is not magic. And we need those controls. We need those things to really keep it in its box. And you can go back to choose your favorite example, maybe the invention of nuclear weapons where yes, at first it was just a technological thing, but very quickly the world realized, okay, nuclear power, it can be amazing or it can be destructive.

And the answer is yes. And so back to having those controls around that, because we are at a moment where it's amazing the power that is in these tools. And that can go well or that can go badly, and as AI in particular gets further and further along because it is indistinguishable from magic as images become indistinguishable, as we talked about deep fakes and other things. Yes, as a society, and you're exactly right, as a global society, we need to figure out how we get our heads around that and make sure that it stays in the bound of good uses.

Chet (17:28):

Yeah, no, for sure. That's a great perspective. I'm going to switch gears to something more personal. You've done a lot of different things. What's next?

Mark (17:35):

That's a great question. I am still trying to figure out what I do when I grow up. I'm definitely at the phase of my career where it's a little more reflective and a lot more about teaching and giving back than it is doing both for how fast I want to run and also what I think I have most to offer. So it's a great question. I don't know. I'm a little reflective at the moment on how best to do that.

Chet (18:01):

That's cool. Well, I'm sure people will track you and try to figure out what to do next, and as always, people will flock and say they want to follow you. I'm going to go to Rapid Fire, obviously brought to you by GPT, so it'll be quick questions, quick answers. The first one is your favorite childhood hobby?

Mark (18:23):

It was computer programming. My father brought home a PDP-1170 when I was in Heathkit, like put together, so I had a computer in the basement when I was seven years old. I spent a lot of time in the basement with a teletype machine and that PDP-11.

Chet (18:39):

You are definitely part of the OG.

Mark (18:42):

Yeah.

Chet (18:42):

Definitely part of the OG. Most inspiring moment you've experienced recently?

Chet (18:47):

Oh, well, I had the good fortune in February of taking a two-week trek through Patagonia, through Torres Del Paine. Amazing. Amazing, amazing, amazing. And just being out in nature just in general, but I would say the moments you come around the backside of Torres Del Paine and you come across a mountain pass and you look down the other side and laid out in front of you is one of the glaciers that makes up the Patagonian ice field. And it's, yeah, that is an awesome and awe-inspiring moment that reminds you of our place in the world and the beauty that is out there in the world.

Chet (19:27):

That's awesome. That's awesome. If you could try and travel to any period in history, which era would you visit?

Mark (19:34):

I would've loved to have lived through and been part of sending the first human to the moon. The technology involved, the exploration involved because it was just like people sending out ships for the first time, they didn't know if they were going to come back. They didn't know it was going to be there, and I would've loved to have been right in the middle of that, in those rooms. Would've been a lot of fun.

Chet (19:57):

One word or phrase that best describes great leaders.

Mark (20:02):

Trustworthy. It all starts with trust. Yes, we want to be inspired, we want to have mission, but ultimately you want your leaders, you want to be able to trust them implicitly. And know that there are good intentions and good results on the other side. Yeah, it all starts with trust, I would say.

Chet (20:21):

One piece of advice you would give a younger version of yourself.

Mark (20:26):

It's a long journey with lots of paths, and so try out those different paths. I can now draw a linear arc on why my career got to where it was. It was not that way coming the other direction. Not being stressed about that. Not being stressed about, oh my God, I don't have a five-year plan. Yeah, that's okay.

Chet (20:51):

Like you said, you're still figuring out what you want to do when you grow up, so I think that's a feature, not a bug. Right. Mark, this has been absolutely phenomenal. Thank you so much for your time.

Mark (21:02):

Yeah. Well, thank you so much for having me. It's been a great discussion.

Chet (21:06):

Thank you.