john-k-thompson.mp3

[00:00:00] This is a great time to be in the field, you know, there's there's a huge number of companies that do get it. There's even more companies that don't. But we don't have to worry about them. You know, there's opportunities for people to be chief analytics officer as chief Data officers.

[00:00:30] What's up, everybody? Welcome to the artists Data Science podcast, the only self development podcast for Data scientists. You're going to learn from and be inspired by people, ideas and conversations that'll encourage creativity and innovation in yourself so that you can do the same for others. I also host open office hours. You can register to attend by going to Italy dot com forward. Slash a d. S o h. I look forward to seeing you all there. Let's ride this beat out into another awesome episode. And don't forget to subscribe to the show and leave a five star review.

[00:01:27] Our guest today is an international technology executive with over three decades of experience in the business, intelligence and advanced analytics fields.

[00:01:39] He has experience building startup organizations from the ground up and has reengineered business units of Fortune 500 firms to enable them to reach their full potential. Is technology, leadership, expertize and experience spans all operational areas with a focus on strategy, product innovation, growth and efficient execution? He's also coauthored the best selling book, Analytics How to Win with Intelligence, which debuted on Amazon as the number one new book in analytics in twenty seventeen. His books serve as guides for non-technical executives and provide a roadmap for the journey of building analytics teams, funding initiatives and driving change in business operations through Data and applied analytical applications.

[00:02:27] So please help me welcome our guest today, author of Building Analytics Teams Harnessing Analytics and Artificial Intelligence for Business Improvement, John Kay Thompson. John, thank you so much for taking time out of your schedule to come on to the show today. I really appreciate you being here.

[00:02:45] Thanks for inviting me, Harpreet. So happy to be here. Thank you for that great introduction. I'm humbled.

[00:02:51] Oh, it is absolutely my pleasure. I'm actually honored to be able to speak with you and talk to you about building the analytics teams as we were talking before the show and in the exact same situation that you kind of describe in your book, in a sense. And this is this is, again, an honor for me. So let's go ahead and get into this. So, you know, a lot of up and coming Data scientists or even organizations that are new to data science. They think that doing data science is as simple as just getting the data and doing the science. But obviously, it's not that easy, is it? And you have a great framework that I'm hoping you can walk us through. And it's the general data science process. Will you be able to walk us through that, please?

[00:03:39] Sure. Absolutely. Yeah, it's it's a very interesting topic. It's a timely topic. A lot of companies want to get involved in data science right now.

[00:03:48] And and there's a number of companies that are you know, they don't really know the guts of it. They they think they do. They they know there's data and they know there's science involved in it.

[00:03:57] And and that's where it often drops off. And that's part of the reason why I wrote my first book, because as I was flying around the world for Dell, I was doing it for about three years. I've been meeting with non-technical C-level executives and I could really see that there was some real reticence in them about trying to understand analytics. I knew that they had McKinsey and Bain and other consultants coming in and talking to them about analytics. Their teams were telling them they had to invest in analytics. So that's really what the first book was about. It was a primmer idea was you get on a plane in O'Hare. I live in Chicago and fly to London and you could read the book on a flight. And at the end of it, then you had an idea of who you should hire, how much you should invest, what you should expect from that team, what you should not expect from that team. So that was really to get executives into it. Now, to get back to your original question of what what how do you do Data science is really what you're asking. And, you know, it's one of those things that that often, you know, you're in the situation where you've been hired in an organization and you're starting to think about building a center of excellence. That's a great way to start. Now, I encourage people to really step back and think about what they want to do with Data and Data science. What are you really trying to understand or you're trying to trying to do pure research and look for things that are five, six, seven, eight, nine, ten years out through innovations and inventions. Are you trying to do incremental improvements? Are you trying to figure out how to get another one percent out of your factory or supply chain or whatever it is? So really try to understand the time frame you're working in.

[00:05:38] But then one thing I often ask people about is what's the Data universe that you are thinking about operating in? And nine times out of ten people will list transactional systems running through SJP, manufacturing systems, e-commerce environments, personalization information. And I ask them and to to take a step back and think about the external data that's out there, you know, what are the external pieces of information you want to use. So really get your mind around the universe of data you can use internally, externally, primary research you might want to create yourself. And once you understand your time frame, your data universe, then you can start looking at use cases and. Personalities and personas that you want to serve in your in your business, then once you understand the time frame, the Data the people you want to serve, then you can start thinking about how it's going to look operationally. One thing that I often find from people who are executives and senior managers and frontline managers is they don't think about the ongoing nature of Data science. One of the things that we often do is that we build, we build models, we build Problem-Solving models, we build linear models, we build predictive models, we build all sorts of models. But often people who are not Data scientists don't understand that just because you built a model, that's just the beginning of it. You know, you need to understand how is that going to impact my business? Is it going to go into production? Is it going to go and be part of a new system that's running continuously in my business? But then the other thing that is often new and thought provoking to them is that that model is going to degrade.

[00:07:27] We as Data scientists understand Data drift and model drift very well, but most people outside the field don't. So they say what I built, I hired these people. They built a model that should be the end of it. It should work into the future and forever. And as we know, they don't. So you have to have teams that manage and monitor and update those models and take them out of production, put them in production. And usually by the time I get people thinking about those things, then they they raise their hand and say, well, that's enough for this conversation. Let me marinate what you said and think about it.

[00:08:02] And then I'll come back and we'll talk more what artists I would love to hear from. You feel free to send me an email to the artists of Data Science Gmail dot com. Let me know what you love about the show. Let me know what you don't love about the show and let me know what you would like to see in the future. I absolutely would love to hear from you. I've also got open office hours that I will be hosting and you can register by going to Italy dot com forward, slash a d. S o h. I look forward to hearing from you all and look forward to seeing you in the office hours. Let's get back to the episode.

[00:08:48] That's very interesting. Point to make is because most software development teams, you know, I'm simplifying it here, but they develop a feature part of an application and that application does one thing with that feature. That's one thing. And it just does that and not much changes. But when we're building predictive models, machinery, models, we're trying to model reality in a sense, regulatory model in the world, in the real world, changes and Data coming into the model is going to change. Right. So it's a whole new set of challenges. Thank you for digging into that when we get more into into that topic later. But I like this this kind of felt like a little bit of a theme going on through your book about the two different approaches for Data science team that your artisanal versus your factory team. Can you talk to us about these two different philosophies, I guess, for lack of a better word?

[00:09:44] For word, yeah, exactly. I think about things usually in clarities. You know, you've got your technical team or artisan team and you've got your modular your factory approach. And and it really comes down to it as an organization. What can what can you do? How how does this fit with the culture that you work in? Personally, I resonate really well with the artisan model. There is no model. I like to hire people who are highly educated, highly competent, multifaceted in their approach to Data science. You know, those those people are expensive. They're hard to find, but they're very, very talented. So an artisan Data scientists has all the responsibilities for a project. They write the charter, they talk to the subject matter experts. They engage with the executives. They obtain the data. They integrate the Data new feature engineering. They try out all the different models they actually test and produce and then present to the stakeholders and then eventually they work with it and the subject matter experts to get models in production. That's a really I like that way of doing data science. It's a very craft oriented you can almost think of like making craft beer or something like that. You know, every batch, every model's a little different, you know, on the factory side of it.

[00:11:09] This is more along the lines of, you know, you're really just interested in getting it done. You're going to hire instead of hiring three to ten people in an artist, artisanal team or artisan team, you probably can hire thirty or twenty people on the factory side of it and you're going to hire people that do Data was. And Data integration in feature engineering and modeling, and it's going to run like a factory, the Data is going to come in. These people do their jobs, they handed off these people, do the jobs. They ended up they do the jobs. And it just runs like that. And depending on scale, of course, and how many people you hire, people often ask, well, it sounds like the artisanal team is more expensive than the factory approach, or sometimes people say and vice versa. And in the end, they almost cost exactly the same, because if you're hiring seven of these people who are really expensive and 15 of these people who are less expensive, you know, it's really a wash on the cost. So that's why I say it really comes down to how do you how do your executives and managers and organization, how do they like to be interfaced with, you know, do they just want hey, here's the problem.

[00:12:17] Let's work on it. Let's get on it. Or do they really want someone to engage with them deeply and listen to their problems, listen to what's going on and the supply chain or the factory or whatever it happens to be. And, you know, the artisan approach, often people will go in and say that what we're here to look at price performance, you know, how much how much can we discount or what does the price look like? And then in the end, what they really end up doing is opening up the aperture and working on the problem in a more holistic basis. Know the end users might say something along the lines of our distribution is just not as effective, as efficient as we'd like it to be. Well, you find out that your product mix is probably not as good as it could be, and that leads into your manufacturing operations are probably not as optimally is not modeled well, when you get back into raw materials so you end up in an artisanal approach or an artisan approach, you almost end up in solving more and more problems than you do in the factory side of it. But either works either works very well.

[00:13:23] As a craft beer connoisseur myself, I love that reference there. And as you could tell in my podcast, The Artist of Data Science, I very much take that artisanal approach as well with myself. And if you like the the data scientist or listen to my show, they're special breed of data scientists. These are the ones that are definitely for sure going on to become chief analytics officer, chief data officer. So these are definitely that that type of data science, that data scientist that you mentioned, that more that artisanal approach. But it almost sounds like this is something that should be is this artisanal or factory model or does this get decided before you hire your first data scientist or is it a consequence of hiring your first data scientist?

[00:14:08] Well, I like to think that things are more planned than they are just by chance. So I think that you you make a conscious decision. I'm going to hire for a artisan Data science team and I'm going to look for these kind of people. You'll be looking for someone with five, seven, ten years of experience. You know, if you're if if you're actually hiring for a factory model, you're probably looking for people with one or two years of experience. And that's fine. It's just a different profile, that's all. But I do think you go into it with a very specific desire. That's how I'm going to do it and that's how we are going to do it. And more than likely, if you are a chief analytics officer or someone who's coming in and building a data science function, you had to have conversations with the organization about how you were going to do it. So more than likely, you've explained that this is your approach, this is your model. This is the investment portfolio you need in this long it's going to take to build the team. So you've already set that the course well before you hired anyone.

[00:15:13] Is Data science and art or science? How do you view it?

[00:15:16] Oh, it's definitely an art. There's no doubt about it. The name of your podcast kind of leads leads us in that direction. But as you know, you've read the book. You looked at the book that I hold out, and I believe this very strongly, that Data science is a creative endeavor. It is an artistic endeavor. And I think that's one of the reasons why we've seen historically some of the failures we've seen. You laid it out very clearly. You know, I've done a lot of development teams. I'm sure you've run development teams and you get a feature specification. I want this piece of data to be put in this form and written to this database. Pretty straightforward. You know, you get an effort, estimation that's going to take three days or whatever, and you do it in a Data science. It's not like that. Someone comes in and says, I want to understand the price elasticity across the entire customer base for these kind of products with this kind of competition in this kind of market space. That's not the same question. That's not the same development effort that's going it in finding the. Data figuring out the modeling, figuring out how to do it. Price elasticity, we all know pretty well that's that's well worn. But if you're doing some kind of predictive model that takes some creative effort and some creative juices in it and it takes an iterative process, you know, you may go into it and say, I'm really going to try something different. I'm going to go out and try some of this external Data I haven't ever used before. And maybe it works, maybe it doesn't work. Maybe fall back to a tried and true methodology. But I tell people it's it's not the same. You laid it out very well. It's not the same as just developing a feature that executes the same way each and every time. It's a very creative endeavor.

[00:17:06] I absolutely love that. In your book that you repeatedly mentioned that brownnosing. I speak my language because I also be data science as art, and it's because different Data scientists may approach a particular problem in a different way. Right. That's the creative aspect of it. Like the way I approach a problem is can be different than the way data scientist B approaches a problem. But I think the science aspect comes down to the fact that, OK, if I do approach it in this particular way and I lay out this methodology using the scientific method, then what I do should be repeatable by Data scientist. B We may approach it differently, but the way I do it, it should be repeatable by somebody else.

[00:17:47] Absolutely. There's no doubt about that. You know, the the the Data Data science is a process, as we talked about in the beginning of the process. Most of us do exploratory data analysis. We just try to figure out how does this work and what's the real what's the world really look like? How's that happening? You know, we do a little, Ayda, exploratory data analysis. Then we get in the creative part. We try to figure out, you know, is this something I can do a neural use a neural network for or is this a decision tree? Is this a classification problem? This is clustering. You know what? You try all sorts of different things in there. And and you come out and say, well, you know, I thought I could do this, but it didn't really work there. Too much of this or too little of that. That's the fun, creative part. Then you get down to it and then as a team, you can all get together and have really interesting creative dialog. Why did you do that? Why did you trim the tree this way and why did you prune this? And you can really have some interesting conversation. I think that's a very creative part of the process, too.

[00:18:48] But once you're once you decide decided the model, you've got the models train, you've put it in production, the creative part of it goes away. Then it becomes a very mechanistic process. If you if you listen to me do any kind of conversations about, you know, development and training and testing of models versus models going into production, this is very, very creative and can take time and can be variable. This has no variability in it whatsoever. If you're if you have a model in a financial, let's say, a financial company there, there systems, there can be no creativity at that point. You know, you train the model, you testimony, you make sure the ground is ready when that model and production starts to stray and breaks tolerances, that model comes out in a new model goes. That's why I say that you need to have a buffer between the Data scientists and the production environments because this is too squishy to live in this environment. So you put translators or RSE or whatever you have in there and you keep those worlds a little separate. They're complementary, but if they are butted up against each other directly, there's a lot of friction.

[00:19:59] Have you read the book Moonshots by Sofie Bookal? No, but I heard it before. I think you really, really enjoyed that. So in that book, he lays out the it's a framework called The Bush Fail rules that Amenabar, Bush and then the other guys.

[00:20:16] And there's a beautiful statue of Vannevar Bush in Lincoln Park.

[00:20:21] Nice. And all that one, huh? Anyway, that so there was a concept has called phase separation. You want to separate your artists from your soldiers. Right. And in this way, your soldiers are the they're the guys that keep the bills paid. There's the people that are keeping the wheels moving in the front line kind of people. So they're in this case, they'll be like the software developers and software engineers that your soldiers, then the artists are Data scientists. That's right. And it's it's important to keep them separate, but then keep the flow of information open between them to. Right, right.

[00:21:00] Exactly. Absolutely true. You know, if you're working in a highly regulated industry like pharmaceuticals or financial products or anything that is highly regulated, you know that those systems have to run within the tolerances of compliance and regulations and all those kind of things. So you need you need people who are very process oriented very by the book. Everything everything is it needs to be buttoned up and working in a very prescribed way Data scientists generally don't thrive in that kind of environment. They want to have all they want to have free reign, if you could give it to him. They'd like to have some leeway to make some decisions on what going to get the best model for the environment, the Data, the problem, the different constraints that you're working within. And they do sit in rooms and work together quite nicely, but you need to make sure that both are nurtured and have their their supporting environments and those environments are quite different.

[00:22:04] I definitely think you're going to love that book. Check it out.

[00:22:10] It's an interesting point because I could spend a significant amount of time of my could be a week too deep in thought, deep in research, having nothing really productive come from it except just notes and research. Right. And from an executive perspective, they're like, oh, what am I wasting my money on you for? Just being an egghead? Do notes and stuff. Right. How do we how do we manage that expectation? Like I look, man, like I am deep in the books trying to find a solution for this problem. Yeah. I mean, make it sounds like you are.

[00:22:48] You are. I know. I was just let you play it out. It's it's interesting being in a creative environment, in an enterprise operation, enterprise class operation know there is pressure on everybody in those organizations to perform and there is pressure on Data scientists to perform as well. And I make sure that when I'm starting a new endeavor that we pick projects that we know we're going to be successful at. You don't want to come out of the gate in swing and a miss on one, two, three, four projects. That's not a good look. And the organization will judge you for it. Whether that's right or wrong, that's the way it will work. So when we start out, when I start out with my new team and people come to us and there's all sorts of projects, as you know, you're in a new organization, fairly new organizations. Lots of people don't want you to do work for them, but you look at it with a very critical eye and say, you know what, I know I can solve this pricing optimization problem. We have the Data, we have the systems, we have the people. The data scientists have done it before. You know, that's a clear win. We're going to do that. And we can do that when we in three months. And that's the timeframe we're with. The organization can absorb and understand and accept. If you come out of the gate and say, I'm going to take on this, you know, optimization or simulation problem, it's going to take me a year to get it done and there's a 50 percent chance I'll fail. Not a good bet. Don't do that. You know, go for some projects you can knock out in a few weeks or a couple of months or at the most four to five months, and then be ready to stand up in front of the organization and explain exactly what you did and how you did it and how it's going to be better for the organization. So, yeah, we'd all like to sit around and cogitate and think and marinate in our thoughts, but we got to deliver to this interesting section.

[00:24:47] You talking about good ideas versus bad ideas. Can you help us understand what is a good idea? What is a bad idea? What separates the two?

[00:24:56] Yeah, that's that's actually a good Segway from what we were just talking about. Good ideas are generally something that you have the Data for. You understand that you know, you can solve and the organization finds it strategic and necessary. Usually if you've got those four characteristics, it's going to be a win. There's there's no doubt about it. One of the conversation there, one of the things I've done and I've made it very clear to the entire organization in my door now, my Zoome room or my phone or teams messages or whatever are wide open. You know, anybody can talk to me any time about anything. And I also have made it very clear that I don't want people thinking a lot about these ideas, because if they sit and think about it a lot, they send off a lot of the rough edges and we Data science like rough edges statistician's want to throw out all the outliers and get it down to a nice tight bundle of Data. We like all the noise. Leave it all. You don't think too much about it. So people come and show up at my doorstep all the time and say, I got an idea, we got great, let's hear it. You know, so usually a bad idea is one that you don't have any Data for. You can't model it. There's nothing in the real world that describes it from a Data perspective.

[00:26:22] And I've said it many times. You probably heard me say it before. We are Data scientists. We are. Not magicians, we can't just make stuff up. We have to have something to work with, so that's number one. So a bad idea is something that you don't have any Data to work from. That's that's not good. Another bad idea is something that no one cares about. Maybe someone comes to you and they have a pet project. They're interested in trying to understand, you know, the reason that the two percent sales that to the the native Connerty enter the indigenous population is not going up. I can guarantee you that the CEO doesn't really care too much about a population that buys two percent of the product. You know, they want you to work on something that's probably going to impact 50, 60, 80, 90 percent of the product sales. Not that indigenous sales are not important. That's that's probably a very important segment. And those people are probably very important in using those products. But you want to work on something that you've got Data for, you can understand, and it's strategic. So if you don't hit those notes or if those are the notes that you have in a project, you can easily say to someone who's come to you, it's a very interesting project and I can see why you're passionate about it.

[00:27:35] But it's not something we're going to invest in. And often those kinds of problems are actually solved.

[00:27:42] That address quite nicely with business intelligence. You can send them to a group that will write a report or a dashboard or something that will show them some really simple metrics and KPIs that they can say, yeah, you're right, that's exactly what I thought and there's nothing I can do about it. But at least I understand it better now.

[00:28:02] What about those situations where maybe we have the right Data for it and problem statement on the surface looks good. And this asking for a friend here is racking our brains, trying every possible way we can to get a decent solution. But nothing seems to be working.

[00:28:23] Yeah, there are those. And it's probably a situation where either the Data isn't rich enough, it doesn't have enough variability in it, or it doesn't describe the phenomena as well as you'd think it does. Or maybe you're trying to apply some kind of a technique that doesn't really apply for the Data. So it's usually a Data or a math problem right out of the gate. So what we do in that area is we have our center of excellence operates on we have weekly meetings, everybody attends the weekly meetings and everybody shows their project, whatever they're most focused on at the time. Data scientists have multiple projects. But, you know, and they really they do code reviews, they show results. They talk about their stakeholder management engagement issues they might be having. And everybody has a chance to give them input and creative ideas about what would improve the project. So it's it's always better to have more minds working on it. If you only have one person that's kind of tough, maybe you can get some buddies around or people together in a room, the teams meeting and start beating it up.

[00:29:35] But generally, it's the set up of the project is not exactly right about this concept of the open mindset versus the fixed mindset. I really like that. I'm a big fan of the work that I did with the growth mindset, and I saw the parallels in there. Just for our audience, if you can define these two concepts, first, the open mindset, the fixed mindset, maybe talk about what they are and how are they different.

[00:30:02] Sure, an open mindset is really, in my opinion, is is often characterized and exhibited by a voracious curiosity.

[00:30:13] People ask me, why do you like being in analytics?

[00:30:17] It's one of the careers where I can, when I was working at Dow, be talking to someone about the flow of oil through a mayonnaise factory in the morning and then the afternoon talking to a bank about credit risk. I'm intensely curious about everything I love. I watch the SpaceX launch last night. I love everything in the world and out of the world, too. I just want to learn about everything I possibly can. So I think that's an open mindset. That's a curiosity based view of the world. And there's always more to learn. There's always more to do. Another thing that people have asked me more than once is when are we going to be done? And the answer is we're never going to be done. You know, this is math and Data math always gets better. It always grows you combines and goes in different directions. There's always more Data. There's always new Data to integrate and work on. So, you know, I, I am enthralled and engaged and excited that this doesn't end. There is no end when we get to the best. We know today we start working on what will be the best solution we can do next week or a month from now. So I think that's an open mindset.

[00:31:34] That's a character with an open mind set. I've often seen people through my career and they generally don't do very well in Data science that they think they know everything and they think they know how to approach every problem. And they think that, you know, what was it? Thomas Watson from IBM says, yes, the global market for computers is seven. You know, that's just not the way the world works. We've seen it over and over and over again. And just there's more and more and more to learn and do. We recently had someone on our data science team that left us a few months ago. They just couldn't work with that. You know, everything had to be six or seven or eight or nine. And it was the world needed to be very cut and dried and they needed to know that they were the smartest person in the room. And that's a tough environment to try to be in, in the Data science field, because, you know, I'm a pretty smart guy, but most of the time that I can guarantee you at every meeting we have, I'm the dumbest person in the room I hire for raw intelligence and curiosity.

[00:32:43] So I think that's I think that sums up what I think about a of open mindset versus a fixed mindset.

[00:32:50] Thank you very much. And I really, really appreciated that. And I really enjoyed seeing that section in the book as well. I think you're hundred percent spot on right about that. You have to be very comfortable dealing with ambiguity as a data scientist. I mean, a lot of up and coming Data scientists as part of Data science dreamed up. We've got like twenty six hundred something mentees. And one of the questions I always get is what's the best way to do this? What's the best way to do that or what's the best they want? They want my maps. Right. Seth Godin talks about this very nicely in his book LinkedIn. And he says that people who are artists who are have artist type careers are need to be very comfortable working with a compass and not a step by step map. I think if you want to be successful as a data scientist, you better start getting to use that compass because there's not to be a map to exactly what to do and what to do.

[00:33:46] Yeah, I agree. And I hear that I do some mentoring as well and some work with different groups and people who are just starting to get into Data science. And I think it's a developmental phase. It's not a call out of any deficiency or anything like that, you know, that they're like, hey, I want to know I want to have this crib sheet. You know, when I see this, I do these seven things, and that's just not the way it works. That's not the way the most successful people operate in this field. I went to college a long time ago and I got one of the first four year computer science degrees that you could get in the United States. And I went back to that college years later to help them build one of the first business analytics programs in America. And I asked them because I came from a really unusual background for computer science, at least I thought it was unusual. And the more I read, the more I find, the more it was not so unusual that I asked them. I said, what? Where did you look for students like us back then? And they said, we generally look for people who had a background in auto mechanics or worked in machine shops. And I said, why? What's the connection there? And they said, tinkering. You know, people who worked on cars or worked in factories had to diagnose those problems and how to fix them themselves. And there was any number of solutions that would work. And you had to find one that was the most optimal or the quickest or the most effective. And they said that we found that people coming into computer science and now Data science, we had that tinkering, curiosity, toyin mindset. We're very good at it.

[00:35:29] Is there in computers?

[00:35:33] Yes. You read the story? Yeah, we do think there's a there's a future in those computers.

[00:35:39] But in all seriousness, where do you see the field of Data things headed in the next two to five years?

[00:35:45] It's a great place. It's a great time to be here. Everybody, this is this is this is where the action is at. There's no doubt about it. I you know, often you hear people say, well, you know, gosh, it's all been invented. It's all been done. It's not true. We're just getting to the to the knee of the inflection point at this time. If you're as well read as you are, Harpreet, or if you read the research that's out there. You know, McKinsey has been saying that, you know, the the well worn, well understood marketing model from crossing the chasm, Jeffrey Moore, and allowing those guys, you know, the leaders are out there right now hiring every good. Data scientists, the best data scientists they can get their hands on. Why are they doing that? Because they already know the advantage they're getting from artificial intelligence and data analytics. And those companies are accelerating away from the market and they're going to continue to do so. There's a great more there's a larger number of companies and the early majority, as opposed to the leaders, the early majority is now having the conversation of should we get into data and analytics? Should we use a we have not proven it. We don't know if we really should do it, but the leaders already know and they're accelerating away. So the early majority and now trying to figure out should we hire someone, should we hire a leader or should we read a book? Should we get McKinsey in here and do a study? So while they do that, they're still engaged in the process.

[00:37:16] And at some point some part of those companies will figure it out and they will hire data scientists and and they will do more with data and analytics and they will accelerate to and then the late majority, they're skeptical. They're not really doing as much. And then when you get to the laggards, they're actively campaigning against data and analytics because they don't believe in it. So this is a great time to be in the field. There's there's a huge number of companies that do get it. There's even more companies that don't. But we don't have to worry about them. There's opportunities for people to be chief analytics officer, chief data officers, data scientists. The way we've set it up in our environment at CSA is that we have an entire and this is kudos to our H.R. department. Really good job from Elizabeth Walker and her team where we have set up our career framework. That you can come in is, let's just say data scientist, because any technical profession fits here. But let's just say with data scientists, you can come in as an intern data scientist and be hired as a full time junior data scientist, become a data scientist, a senior data scientist and a principal data scientist, all without having any managerial responsibilities whatsoever. You can keep climbing the ladder, growing your skills, getting greater compensation and doing the things you love. So I know that's a long answer, but I think it's a great time to be a data scientist.

[00:38:47] So talk to us about your journey from individual contributors to where you are now, all that wonderful experience you've amassed and this wonderful depth of knowledge you have. What was your journey like going from from an individual contributor to executive level?

[00:39:03] It's in everybody's different. So listen to the story and take it and make it your own any way you can. I was a developer way back when I wrote Assembler Code in and worked on a bunch of different systems and a variety of different areas. And it is early in my career that I immediately gravitated towards data and analytics even before it was a field. I would talk to people about the fact that the most important part of of our mission, no matter where I worked, was to be a good steward and understand the Data. And people just look at me like I had six AIs and they're like, What are you talking about? We have to write this features to take this Data from here and put it there and write a report. And I'm like, I don't really see that as being important. It's the Data that's important. And when we started to see companies come out that were interested in data and analysis, I immediately went from being on the the big company, developer and user side and immediately went over to the vendor side. And they got involved in data warehousing and and and data mining.

[00:40:14] And in the early days of that and then business intelligence and then state statistics and advanced analytics and A.I. and I was always very interested in that kind of stuff. I was always pushing the envelope to try to be involved and use new techniques and technologies as they were being developed. So once I found that field, then I moved back and forth between being at a vendor, creating innovative technology or like I am now. I'm at CSL, which is a biopharmaceutical company building an analytics team. So I'm a little unusual in that respect, in that I've done both. I've worked on the technology side and creating products and technologies, and I've worked on the end user side implementing and using those technologies. And I've always enjoyed the technology. I always say I'm a lapsed technologist and then I've always enjoyed the team's part of it as well. So it's been one of those that have gone from being a developer to being an executive. And and I've taken kind of a curvy path to get there.

[00:41:17] And through that that experience you had and that journey you've taken, what would you say makes analytics teams analytics project.

[00:41:26] Just all things related to the lyrics so unique, it's it's the ability to look at different phenomena through the lens of improvement, you can look at anything and it can be improved. Now, when you talk to some of the people that are running those business units or running those functions, often they don't believe they can be improved. They think that they're optimized and they're the best they possibly can be. But I don't really think that's the case. I think everything can be improved. So it's one of those things that as an analytics professional, you're always, you know, you're not really the owner of the business. You're kind of always a consultant. You know, whether people like that label or not. I always liked being thought of as a consultant. You know, if I was a consultant or if I was working as an internal operator, I was consulting with people, trying to help them understand how to get better. So it's always through the lens of how can I make this? How can I improve this function? How can I get a better level of efficiency of effectiveness? So that's how I look at it.

[00:42:33] Thank you very much. I appreciate that. In your book, you talk about how it took two years to get your head wrapped around the fact that the executives not only do they not know what you're doing, but they didn't really understand it. But how would you recommend the first Data scientists in the company handling a situation where you have an executive who's maybe read a couple of blog posts? They know the names of a couple of things, but they don't really understand or know machine learning. How would you how would you handle that situation where you have somebody who thinks they know your job and is trying to tell you what to do and not to do?

[00:43:12] Yeah, that's it's always fun when when you have the the people who don't know what they're talking about, trying to direct what you're doing. Humility is always important to patients and taking a deep breath always works as well. One of the things that I talk with everybody that I talk to about if they're in Data science especially, is that when we sit around the table and we discuss problems and things that we're working on, we can talk any way. We can talk math, we can talk algorithms, we can talk Data, we can just all out geek fest all day, all the time. And that's fine. But when we leave that room and we're walking down the hall and we encounter, let's just say it's the CFO and they want to she wants to talk to you about how Ehi is going to be. It will help their area. You need to immediately switch into financial discussion. You need to use the language she understands. You need to talk about the cadence and the terms and the measurement and the things that connect with her. You can't you can't get anybody to understand what you do by talking in your language, talking in our language. You need to talk in their language is the first thing. And then secondly, the thing that always works. And my wife taught me to do this. Over twenty six years we've been married. She's a career development coach.

[00:44:41] I am her longest project, that's for sure. And the one thing that she taught me and I finally, finally got is that questions are your friends. So when you have someone who is really giving it to you about how you should do your job, just ask them questions, you know, hey, that's great. I'm really excited that you're interested in this area. And do you understand? You know what? I want to understand your views on how we can improve cycle time efficiency or how we can make more money in this area, or how can we increase increased gross margins in Western Europe? So you just start asking a bunch of questions. And I can guarantee you by the third question, they're going to leave you alone because they're going to realize I really don't know how to talk to this person. I really don't want to get into this conversation. And we're more likely they hired you because you're really smart guy anyway. So I don't really want to do your job. They just want to have a conversation about it. So often we will get a little defensive thinking that people are trying to tell us what what to do. Really, they're just excited. You know, they just want to talk to you for the most part in it and questions well, always in that conversation in a way that you both feel good about it.

[00:45:57] Thank you very much. I appreciate that. I was asking for a friend again, so to speak, that I have a philosophy that as a data scientist, you should learn to build and learn to sell. Being able to do both is an ultimate superpower. When you're the startup founder of a Data science team, how much of a job is selling?

[00:46:20] Most of it. You know, you're a lot of people in our field think, oh, I. Here to build things, I have to sell anything that's not really the case, I've been the CEO of seven different start ups. I've worked at Dell and IBM and other large organizations. And there's always someone that holds the budget above you that's making a decision on how much budget to give you and how to support your function and whether you're going to grow or someone else is going to grow or something like that. So you do. And there's a fair amount of it in the book, as you know, about how to connect with C-level executives, senior executives, senior managers in understanding their viewpoint and convincing them that your organization and the things you do are the best bet for that, that pot of money. So I would say my job is probably 60 percent selling at this point. The first year that I was here was probably 80 percent selling. Now I've got a number of people who understand what my team does and what we can do. I don't need to sell them anymore. They they just want us to work on projects for their areas. So as you go, is the people that are listening to this podcast go through your career and you move from being an individual contributor to a leader of people, to a leader of functions, to a leader of multiple functions. As you go up that ladder, you're going to have to do more and more of that if you don't like to call it selling more and more of that, connecting and convincing many resources on how we can get better at that.

[00:48:01] I don't know. You talk about in the book as well any any courses or books that you recommend, anything that helps you communicate better.

[00:48:11] That's usually the foundation of all of it. I consult with Oklahoma State University, University of Texas at Austin University of Michigan, University of Illinois on all their analytics programs. And the one thing that I have told all of them, and I'm sure they're sick of me saying, is that you don't have enough communications in your programs. You're not teaching people either verbal or written or presentation communications well enough. And starting there, if you can't stand in front of a group of people, C-level executives, senior managers, whatever it is and with confidence, get your point across quickly and easily and convincingly, then that's where your Achilles heel really is. You need to start there. And then after that, you can take all sorts of classes on. We just at CSL had our entire entire data science team go through a presentation skills class. And I had taken it before and I didn't join the class. And I think the team was very happy for that. They were kind of surprised in the beginning, but they all had to write presentations, they had to give presentations and they all benefited from it. Then after that, we went through. Interesting enough, this is kind of people say, what are you talking about? We also took them through visualization, training, how to build compelling visualizations, because what we found was we could build really cool and interesting models that were incredibly impactful. But if the visualizations weren't engaging, then users didn't use it. So you have to communicate written verbal presentation. You have to take your your results and turn them into something people are interested in looking at or engaging with and then whatever. There's a million classes out there and know persuasion or communication or connecting or social I.Q. or emotional intelligence. There's lots of people to talk about the stuff. Now, once you get your communications to the point where you feel you're good, you work on your emotional intelligence because that's one area that we use. Data scientists generally need improvement.

[00:50:29] It's one of my favorite areas to study up on and get better. I mean, I often say that being a data scientist is like the least interesting part about who I am, what I happen to do for work. I thoroughly enjoy it, but I love learning about how to deal with people and human psychology. A couple of books that I found really helpful was that Daniel Pink to sell is human. There's another one, the art of selling anything. And then even just this book that I keep right here at hand whenever I need it. It's just business writing for dummies.

[00:51:01] It's right there. Well, well earmarked there.

[00:51:06] Yeah, I got everything nicely color coded with what I need to get to it. So what are the first Data scientists in an organization? How can we ensure that we're building or at least cultivating a culture for analytics to thrive?

[00:51:23] So it's a great question. One of the things that. We is is if you're the first and only data scientist in the organization, as I said, being successful is important. It can't be overstated. We've talked about it. We're going to talk about it again. Pick projects that you know you can win and pick projects that the company cares about and then shout how successful you were, communicate that, hey, we got a win here and this is how it works. And then lay out your case for why you shouldn't be the only data scientist. That should be a function. It should be something that serves the greater good. So I think early success and high probability of of being successful is important. And that doesn't come by chance. That comes by design. That comes out of the projects. You pick the Data, you work on the people, you connect with. Everything that you're doing as that lone data scientist in your organization is you're almost like a mason. You're setting the foundation that all the other data scientists and the functions and in the center of excellence will be built on. So if you're building on a swampy ground, you know, it's not going to it's not going to work. You need to build on bedrock. So you want to be in it to win it and engaged and doing good work and then talking to people about what you did.

[00:52:50] So it's definitely like the Data scientists responsibility to help make that culture the one that they can thrive in.

[00:52:58] Yes, absolutely. Because the organization doesn't know you know, you can't expect the CFO or the CIO or the CEO didn't know what you or the future Data science function is going to need. They just don't know them and they've been exposed to it. It's like asking someone to speak Hindi when all they speak is English. They can't do it. It's just they don't know what's going on. So it's it's it's incumbent on you. And it behooves you as the first person to build a function that's going to be repeatable and extensible and scalable.

[00:53:31] Thank you very much. I appreciate that. Next, I have going to dig a little bit deeper on the creative, innovative aspect of data science and how it contrasts with more of that factory kind of production system, process oriented system. How can we balance and this creative, iterative, unpredictable process of analytic discovery with those environments that have these operational or production process oriented characteristics?

[00:54:04] Well, the way I think of it is that we have the data scientists who are doing the model, building the creativity side of it, the EDA and those kind of things.

[00:54:13] But once you get through that, then you're building a model that you're going to you're going to test and validate and prove that it works. And really, at that point, your work is here, is done. One way to say it, because you're really then becoming an adviser. You know, the the work of putting this in the models and the applications into production systems are going to be you as an advisor. Then there's going to be the subject matter expert who's going to understand what the business result really should look like. And then the person who's going to be primary in that function is it they're going to have to understand that. Let's say that this is a model that's going in to predict where to to put the next facility. Well, that needs to go into the system where people make decisions about where they're going to buy land and make decisions about new factories or new retail stores or something like that. And the people are generally the ones that know that. So you're going to go in, you're going to talk about the screen real estate, where the end users are going to see the results. You're going to talk about the operational systems and how they operate, what the cadence of those operations are. And you talk about how that entire application or those models are going to be inserted in there. You're going to give them all the inputs and outputs and the contingencies and the constraints. They're going to build it. So it really isn't the data. Scientists have to get their hands dirty in that kind of work because they're not good at it. They don't know what to do. You become an advisor. So, you know, you want to make friends with all those people because you're going to be working through some interesting and possibly tense situations. But that's the way it works.

[00:55:59] How have you seen Data sons play out in, like sprint environments where you're working in sprints? What's your experience with that? How can we how can we do our work in that kind of an environment or system?

[00:56:14] I think you're referring to agile for the most part, unabashedly. I'm not a fan and I've said it many times, but continuing to say it, I don't like an. I don't use agile, I think agile forces, people to do things I think is great if you're actually developing very linear features. OK, you need to do it right. 10 lines of code for this feature. You wrote five of them. Why didn't you get the 10 the six through 10 done? When are you going to get six through 10 done. OK, fine. And that works, OK, for those environments. That's not the environment we're in. And Data science that would work in the later parts of the cycle where we're talking about where it's turned into it and integration project. You can certainly use agile there and those kind of environments. Most of the people that are there accept that and understand it and find it comforting that they've got daily check points and daily stand ups and stuff like that. We actually tried it early in the covid crisis to see how it worked for our team. I was pretty sure that it wouldn't wouldn't work well.

[00:57:23] And we were about four months into it doing daily stand ups. I guess it was about two months into it seemed like four months. And I just stopped one of the meetings and I just asked the team. I said, how does this feel to you? How does these how do these meetings, how are they working? And I think someone said, well, it feels to me like I'm pulling my fingernails off every day and I'm like, OK, great. So we immediately went to three days a week and then we went to two days a week. Then we went to back to our weekly meetings. It was one of those things that we did it because we had to show that we were dedicated to the cause, but it didn't improve our productivity, it didn't improve our models. It didn't get anything done any faster than it would have been the other way. So I don't think Agile fits very well. On top of Data science team. I never use the word agile. I actually describe our team as being very nimble.

[00:58:20] I like that that word nimble. You talked about in your book as well, this concept of linear, non-linear thinking on it just now as well. How can we leverage this understanding to set our analytics teams up for success?

[00:58:37] Yeah, you have to realize that if you are dealing with a linear thinker or a non linear thinker, you will not change the way they think. So if you're on some quest to convert someone to the way you think you will fail, that is not the way it works. You need to understand where people are coming from and you need to meet them where they are. And if if someone is a linear thinker and you're a non linear thinker, then you need to take a deep breath, center yourself and work with them to understand that they only see it in the sense of one, two, three, four, five. They don't see it. Like I get the sense you're probably a nonlinear thinker. I'm a nonlinear thinker. I often go from A to Q in my Data science team laughs about it all the time because it's like, well John just skip thirty two steps and that's just the way I am and I'm not going to change. But I can understand where other people are coming from and I can, I can relate to what they're doing and I can change how I communicate so I can have a meeting of the minds.

[00:59:48] It's not the way I like to operate. That's not my preferred mode. But I know I have to do it, especially when you're working. You're on the the production side of someone who works in a factory, who is a process engineer. Their whole life is process. You know, to to not go down a process is to invalidate their reason for being so we can actually be really upsetting to people and not realize, you know, I've seen that a few times where I've gotten up, but I've been in the moment. I've been in the flow. I've been on just all about it. And I leave the meeting and some people come up and they're just really upset. And I'm like, why? What happened in there? Like, you just didn't follow the process. You were all over the place. Like, yeah, that's kind of what happens when I get into the flow state. So you just need to understand who those people are and where they're coming from and and be someone who can shift gears and get into their world and make sure that you can connect with them.

[01:00:46] And it's not necessarily one way is better than the other four for Data. But I think if the matter were working together.

[01:00:53] No, no, it's there's there's no judgment there. They're not at all. It's just a nonlinear thinker. That's all there is to it. And my wife is a non linear thinker. So when we sit and talk, it's we're all over the place. We're like Betis in a tin can, we bounce all over. In-process people are fantastic. We need process oriented people to set things up in a way that it's controlled and it works and it works every time they're there. Just some processes that AIs depend on it. So you don't really want me designing that kind of process.

[01:01:23] Speaking of lines and linearity one. See, a lot of that isn't like an org chart, a new experience. Where does it make sense to have a data or analytics teams? I've seen them anywhere from I.T. to finance to products to ops teams. What do you think it makes sense to put the data science team?

[01:01:44] It's a great question and it's something that, given my thirty seven years of experience, I've seen almost every permutation known to man. And I come down on this and this is where I come down on this. And I'm pretty I feel pretty strongly about it. The Data science leader, whether it's a chief data officer, chief analytics officer or whoever believes that function should actually report to the CEO. And if they can't report to the CEO, they should or to the CEO. And if they end up reporting to a VP or a line of business function and vice president or something like that, that's usually a that's usually suboptimal. And why do I say that? Because as a data scientist and a data science leader, you need to have the parity in the organization. You need to have the sponsorship from your management to go anywhere in the organization and be able to recommend new solutions to those groups. So you end up in a lot of people are here and I'm not judging anyone for where they report into, but I know what happens and I've seen it happen over and over again, that over time, the Data science team, let's say that you work for the CFO in the Data science team, has been working on supply chain issues and manufacturing and pricing and all sorts of different things.

[01:03:10] But over time, your group, the Fine Data science team, is going to end up just working on finance projects and you will be measured just like every other piece of the finance organization. And that's not the best way to measure a data science team. You need to have your independence from the functional groups. You know, if you end up working for the CIO, you're going to end up working on IT projects. That's just the way it is, because you people, you can't reprogram people. A CIO is probably been in the I.T. function for twenty five, thirty years. That's the way they think. And they think they are doing the right thing by having the data science team or a server or optimization or cyber, something like that. But I can guarantee you most data science teams don't go in wanting to be pigeonholed into doing either finance or I.T. applications. They're there for the bigger picture. They want to work on things that aren't going to have a strategic impact. And I'm not saying that working on applications in finance or I.T. are not important. They are. But that's not what a Data science team wants just to do all the time.

[01:04:23] What would you say are some unreasonable expectations that executives and management have of these startup Data science teams?

[01:04:32] It often it's the unreasonable expectation of time. You know, if you hire someone, let's just say you hire someone. And I work in a biopharmaceutical company, let's use that as an example. You hire people into your biopharmaceutical company who are artisanal Data scientists but have never worked in biopharmaceuticals. They've worked in finance or technology or transportation or utilities or something like that. They need time to learn the business. So you can't just drop someone in and say, hey, in two weeks, I want an application that predicts, which is going to be the next covid vaccine doesn't work like that. They can be productive and they can certainly network around and learn the organization and meet people. But I say that you really shouldn't expect a Data scientist to really produce good, solid results for about six months. They've got to learn the organization. They know where their desk is. They've got to figure out. They've got to set up their stack. They've got to learn where the day the data is and access the data. There's lots of things that people have to figure out. So they usually the time expectation is one that's pretty consistently unreasonable. Another one is asking people to do the impossible because you don't know what is possible or impossible. They'll ask the data science team to predict something that is there's no data. You know, you can't predict it. If it does, there's no data just doesn't exist. And it's very hard for people outside of Data science to know what is possible and what's not possible.

[01:06:15] So I think it's usually time, expectations and then production of results or predicting phenomena. Those are usually the two big ones. The two big problems that I. You know, we had a situation at its Xcel Center, my tenure, he was frustrated not with us, you just frustrated. He had a challenge that was given to him by the organization. It was in the research and development area of things like I'm really just at my wit's end. I don't know what to do and I can't figure this out. And what he was tasked with doing was to look at the entire universe of pharmaceutical research, to try to figure out the most important site. The citations are sources for information about a certain therapeutic area and talk to him for a little bit. And then I said, well, give me a minute. And I went and talked to one of the Data scientists and I said, Can I call you back in about a half hour? So I called him back and talked to him some more. And I said, OK, you know, I'm going to send you an email and I want you to fill out everything I'm asking you for. And he did. And I gave it to a Data scientist. In the next day. We had set up we had written a few lines of code and, you know, went out and did some crawling on PubMed and various places. And we came back and we had given him a taxonomy.

[01:07:39] You know, here are the ten most important sources you could have here. The next hundred most important. Here's the next thousand years, the most next ten thousand.

[01:07:49] And he looked at us like we were we were walking on water for him. That problem was insolvable for us. It was like two hours, you know, so people don't know. They just don't know what they're asking for. And you have to be the person that listens very carefully and says, oh, that's the I have that to you this afternoon. And people are very grateful or a little harder to deliver is that's not possible.

[01:08:15] Thanks so much. I appreciate that. So with the last formal question before we jump into a random round, it is one hundred years in the future, the year twenty one twenty. What do you want to be remembered for?

[01:08:28] Oh, that's an interesting question. I think that's where you're going. And anybody asks me that. I always want to be remembered for being a good father, a good husband and a technology a technology leader that actually made the Data science field better.

[01:08:46] Definitely have. Many books have been amazing. I really enjoyed going through this one. I'm looking forward to to the next one has me back on the show when that time comes, those jumping to a random round here. If you were to write a fiction novel, what would it be about? What would you title it?

[01:09:04] Fiction novel. I really love Quentin Tarantino's a couple movies, Once upon a Time in Hollywood and then Inglourious Basterds, where he's taking history and changing the story. So I would probably I would probably write something about the history of rocketry and change the exploration of Mars or something like that.

[01:09:30] And I don't really know what the ideal would be, but that's what I would write about.

[01:09:35] When do you think the first video to hit one million views on YouTube will happen?

[01:09:41] And what will that video be about and probably happen in the next year or two? And it'll probably be something ridiculous, like a kitten stuck in a tube or a baby pig kissing a cat or something like that. Oh, I'm that's my guess.

[01:09:59] What are you currently most excited about or what are you currently exploring?

[01:10:04] I'm really excited about the future of Data I'm thinking about. That's going to be my next book. There's some really intriguing developments out of Europe. We're all I think we're all aware of GDP and what that did to privacy around the world. There's a new directive coming out of Europe, about Data Commons, and in paying every European citizen the Data dividend based on their Data. I think that the control of our own data is very fascinating and really exciting.

[01:10:36] What are you currently reading or what book do you recommend?

[01:10:41] Well, let's see. And I have some books here that are just off off the field like Fichter Frankel's Man Search Man's Search for meaning. Not a very light book, but something that I really enjoy reading. My son gave me this one string theory about tennis, and then this is a book my wife gave me, which I'm really enjoying. Excuse me. Think like a rocket scientist.

[01:11:12] Nice, nice ass. Actually, I listen to a man's search for meaning on Audible a couple of days ago.

[01:11:20] It's a heavy book and whenever you start feeling like, you know, your your life is. Pick that one up and read 10 pages, put it into perspective.

[01:11:30] Yeah, yeah, I read a lot of stoic philosophy as well. And so there's a lot of that present in Nancy as well. Brian Holiday.

[01:11:38] Yeah, I read his book about Marcus Aurelius. And then there's another one that I want to get. I think it's called Daily Stoics or something like that.

[01:11:44] Yeah, yeah. Daily Stoic. I've got a couple of episodes on my podcast. One is an introduction to Stoicism. And then I recently interviewed Donald Robertson, who wrote the book, How to Think Like a Roman Emperor that should be in a couple of weeks. Speaking of holiday, I just recently interviewed his mentor, Robert Greene. So this is a huge, huge moment for me because a big fan of his books. What song do you have on repeat?

[01:12:13] There's a there's a there's a guy named John Astley, not Rick Astley. And you don't remember the name of the song, but it's about walking around and enjoying life, but always coming home and, you know, just being satisfied with his life. So that I've been listening to that a lot. And then I have been playing on vinyl.

[01:12:41] It's hard to kick against the Pricks by Johnny Cash to look both those up when I'm going to jump into a random question generator for a few questions here. What talent would you show off and talent show Ultimate Frisbee? Oh, I love Ultimate Frisbee. I miss playing that so much. Yeah. In your group of friends. What role do you play?

[01:13:06] Smart.

[01:13:08] What's your go to dance move and cakes or waffles or pancakes. Do one more. What's the worst movie ever seen.

[01:13:21] Oh this is a great one. Have you ever seen or heard of the movie called The Abyss?

[01:13:25] Oh, that's an old one, right? Yeah, yeah, yeah.

[01:13:29] Four years. Four years. My friends and I watched it and then we would always rate the horrible movies we've seen as nothing could be as bad as the abyss. It's the worst movie ever made. So it was like, how close to the Abyss is this movie? So that's when we played that for decades.

[01:13:50] How can people connect with you and where can we find you online?

[01:13:53] Linkedin is the best place to find me. John Thompson, fairly easy to find. And John Thompson, I don't think that actually works on LinkedIn. But as John Thompson would be and you know, I post to LinkedIn frequently and I always double post on LinkedIn and Twitter, but I rarely ever go to Twitter LinkedIn as best plays that include a link to your profile there.

[01:14:19] The show notes. John, thank you so much for taking time out of your schedule to come on to the show today. I really appreciate having you here.

[01:14:25] My pleasure. Really enjoyed it, Harpreet.