Energy Bytes

Marc Hebert, Managing Partner at Dakota Analytics, shares how data and analytics are revolutionizing the oil and gas industry. Marc discusses his journey from tech to oil and gas, highlighting how Dakota Analytics is driving efficiency and creating value with data-driven solutions. The conversation covers the importance of data quality, the challenges of adopting AI, and the role of cloud technology in modernizing the sector. Marc also emphasizes the need for a strong data culture and continuous adaptation in this rapidly evolving industry.

Join the conversation shaping the future of energy.
Collide is the community where oil & gas professionals connect, share insights, and solve real-world problems together. No noise. No fluff. Just the discussions that move our industry forward.
Apply today at collide.io

00:00 - Guest Intro
00:35 - Mark’s Background and Dakota Analytics
06:02 - Getting into Tech
07:21 - Industry Stories
11:59 - Oil and Gas Industry Challenges
18:09 - Data Culture Approaches
20:06 - Dakota's Business Model
21:50 - Custom vs. Off-the-Shelf Software
26:50 - Closing Remarks
27:40 - The Future of Excel
30:20 - Data Architecture Insights
34:59 - Measuring Adoption Success
39:05 - Artificial Intelligence in Industry
43:10 - Oil-Fueled Data ETL and Databases
47:02 - Operator Conversations
48:58 - Project Failures in Data
54:53 - Getting Started with AI
56:40 - Future Trends in Data
58:10 - Data as a Competitive Advantage
1:02:45 - Programming Languages Used
1:05:10 - Speed Round
1:07:50 - Advice for Aspiring Data Professionals
1:09:45 - Where to Find Mark

https://twitter.com/collide_ai
https://www.tiktok.com/@collide.io
https://www.instagram.com/collide.io/
https://www.facebook.com/collide.io/
https://www.linkedin.com/company/collideai

What is Energy Bytes?

Welcome to Energy Bytes with John Kalfayan and Bobby Neelon. Your essential guide to the intersection of data and energy. This podcast dives deep into the world of energy, shedding light on how data, AI, and technology are revolutionizing this sector. Each episode equips listeners with insights into the most efficient tools and resources, paving the way for a data-driven future in energy. From technical nuances to broader industry trends, Energy Bytes offers an unparalleled perspective on the evolution of the energy industry. Join us as we decode the algorithms of energy, one byte at a time.

Welcome to another episode of Energy Bites. I'm Bobby Neelan. I'm here with my trusty co-host, the Rad Dad, John Calfan. Hey, everybody. And we've got Mark Hebert with us from Dakota Analytics, down from Calgary. He's the I think second Canadian we've had, so. Yeah. Good. But uh welcome. Thank you. Thank you. Fun to be here. Yeah. [snorts] But uh yeah, usually just kind of start off, you know, kind of let people know who you are and then really want to get dive into like, all right, how did you first even get into technology or even or oil and gas or both and just, you know, where did it go and how how did we get to where we're at today and then we can kind of dive in on all the different cool things you guys do. Yeah. Uh so I work for a company called Dakota Analytics. Uh primarily based in Canada, but we are now in the States the past couple years. Uh in large part helped to Funk Future with help with Funk Futures who helped us kind of out to Jeremy. Yeah, shout out to Jeremy. Uh breaking into the space. Uh we're about 50 consultants. We only do data and analytics. We don't do anything else. So people trying to get us to pitch FTC systems and like we don't do any of that stuff. So we're all data and analytics. We work at about 40 oil and gas clients, say like last month. That's how many places we would have invoiced. Uh and like just doing a ton of different projects. So a lot of our focus is like in kind of the basics. So a lot of like mid mid market kind of companies, not the huge ones that have a huge team. Uh not Grayson Mill cuz they are fantastic already, but some of the ones that are just either starting out. Yeah, or or or can't build a basic LOS report without doing data dump stuff. So my history, I started Dakota 11 or 12 years ago, but prior to that used to work at a company called Gordon's. You know the fisherman? Yes. Yeah. Yeah. So Was that up in Boston? Just outside of Boston, yeah, a town called Gloucester. Worked there for 10 years. Started out doing ERP, believe it or not. Um Fun stuff. Yep. And decided that uh it was more fun to work in the data analytics space cuz you got to work with a little more creative people, I would argue. Like trying to solve weird, crazy, different problems. So, migrating to that, did a bunch of Microsoft stuff. This is way back in the day. Yep. And then got recruited to move out to Calgary and work with Dakota Analytics. So, my wife my wife moved to Calgary not having seen correct? Uh, technically my title is managing partner. Yeah. So, there's myself and and two other gentlemen Okay, so you were not the founder. No, the founders so the name Dakota Analytics came from the founder went to the University of North Dakota. Okay. So, we were trying to come up with a name. We're not based in North or South Dakota. We're just that's just where the name came from. Uh, and it's kept and ho hopefully we have a decent brand. Yeah. Now, naming a company is one of those things that I get. It I will spend days [laughter] like going through and iterating products. Yeah, companies. trying to come up with the right name for our dashboard or our whatever it was and Yeah. Well, and it's funny too cuz the oil field is generally horrible [laughter] about it's like this engineering service or or this very ambiguous name like uh what was it? Uh We just go with any of the EMPs that done like the pharmacy companies. Oh, yeah. Yeah. Now, you have the trend of EMPs trying to sound like tech companies, but naming stuff is is a Yeah. strangely unique thing for me. was one of those I think it either just really doesn't matter or if it's really bad, I think it it can be a bad thing, but I mean like in in general some of it's just like don't overcomplicate it and and roll with that. I mean Dakota is a fine name. It's like It's hard to be descriptive with the name. And IT's lots of acronyms. Yeah. So, like let's call it ABC right and [laughter] you're like no one's going to know what that means. Yeah. No, well and like that's the thing. I think a lot of people don't think it's important, but then you get onto the kind of marketing media side and it's like like even Energy Bites for example, there is a some kind of energy supplement company that makes a a product called Energy Bites and it's spelled with a Y just like ours. So, I think there even is an Energy Bites podcast. Like I think I saw that. [laughter] Um, but just things like that, right? Like thinking going and looking and like hey, is there other stuff that pops up when I Google that is going to be competing with me for SEO or for you know even just handles on your social channels and all that stuff. It it's I feel like it's slowly getting more and more important these days but I learned about the lesson and the brand when I worked when I worked back at Gordon's and like they were talking about how long when you go to the grocery store to buy something whatever like a a bag of sugar and how long it takes you to make that decision and it's somewhere between like 0.2 and 1 second or something like it you have a small amount of time you just grab the thing you need and like so they're just looking a lot of times like oh there's the Gordon's fisherman I recommend that that's going to be good and I'll get that and that it's as simple as that it's not what's on sale or any of those other factors. thing I think I was thinking about a couple years ago we were talking and I was always wondering like say dish soap like my parents use Dawn I use Dawn I don't even think about it. Yeah. Like do I even think about trying Joy or the whatever no I just go to grab Dawn like there's some interesting things that have just become So I would like to think that that's how the Dakota Analytics brand is but I'm not entirely sure that's true but So you guys are pretty big in Canada and you're and you're starting to make your way here I mean or of that split you said like 40 companies how how many would you say are you know About 20% of our business right now is in the States so we've been doing we've been doing reasonably well I don't think a lot of people fill the space that we're in we're at a big enough size where we've got like continuity and we can staff a team of 1 to 20 people and then there's just there's competitors out there so not to diss our competition but there's not a lot of people that have the scope and size but then also have the experience of having to the average Dakota employees worked for us for I don't know probably 6 years 7 years so it's a pretty good longevity that like I use this as an example all the time there's oil sands in Canada we did a competing demo 6 7 years ago with a a company which I will not name and they called heavy oils called bitumen and they called it bitumen and I was like we knew at that point we were going to win that opportunity because they they couldn't get the name right. frac with a K on a press release or something I immediately take it with a grain of salt um [snorts] let's back up a minute you kind of grown up or when did you really get interested in kind of tech, software, [snorts] BI kind of stuff? So went to school in New Jersey, went to Rutgers for a couple years, got into ERP, just thought it was really interesting. Uh really like the idea of using technology to change someone's life. Sounds a bit corny, but I really like that. That's kind of the joy in in work for me. So getting into that and then doing more data analytics, which is pretty progressive Yeah. is kind of a good space to be on. So started doing that at when I worked at Gordon's and then for the past 12 years doing that with Dakota Analytics. And what I love about the energy space is there's always this debate, is it simple or complicated? It's fairly complicated. Like relative to other businesses, For sure. it's it's grow or die, right? So those are the businesses that or kinds of businesses that I want to be in for data analytics cuz it's very true. So Gordon's is consumer products, you grow at 3 %, it's a good year. Yeah. Right. On average. I mean to say that for Gordon specifically, but like if you work for General Mills, 3 %. But in oil and gas, you grow 3 %, like you're Grayson Mill won't be around if you're growing 3 % a year. That's not no one's publishing those as their goals, right? So so that's why I really like about it. And doing it for 12 years. What uh

you've cuz you've had I was looking at your your LinkedIn earlier today. You've kind of done like BI stuff with a lot of different kind of energy companies facing folks as well as obviously at Dakota, but kind of talk us through, you know, what are some interesting stories or even horror stories that you've come across cuz it's really interesting in my opinion to kind of see like the evolution of of a topic or a subject in an in the industry over time just cuz I think sometimes it comes full circle and other times it's like you can see, you know, you can see where everyone else was doing this one way and then somebody else came and did it differently and really kind of leapfrogged or started setting the bar much higher and For sure. Like I'll give you a couple of stories come to mind. So, one and this was pretty repeatable, but we'll engage with a company to do some work and one or two, I'll call them super users or power users will latch onto like seeing the vision, seeing how fast and how much better that they can get at their role. A couple of them reported it right into the CEO. So, they were like [snorts] sort of the right hand person to the CEO in terms of getting information, whatever it might be, okay? And they latched onto these tools and said, hey, we need, you know, the automation and the integration and all those things that reporting, slicing dice ability to do ad hoc analytics. Latch onto it and inside of 12 months two people in the exact same role, they just moved one to a totally different role, sort of not reporting to the CEO, and they promoted this other person. And like they'd all worked in the same role for the past three or four years and in the course of 12 months, this person, only because I would say they were leveraging the tools and saw the vision, just took it kind of to the next level. And I see that happen a lot. And I use the analogy all the time like people are building houses and you take a hammer and you know, a bunch of nails and you start pounding them in. And data analytics specifically and maybe decode analytics, it's like the nail gun. Right. So, I'm like if Bobby's got a nail gun and John has a hammer, this race is not going to last very long. Like in a couple days, the foreman's going to be like we need another Bobby. End of the day. You know, [laughter] even. So, it's always interesting to me is when we come and say, hey, we've got the nail gun, right? Or we're trying to describe that vision, how many people are resistant to it? And I think for me, that's the biggest thing about data analytics is like we're really enablers of change. That's what we're That's what we're trying to do, right? And the hardest part for every data analytics project is the people. It's not really the technology. We can talk about technology today and stuff, but that's not really the challenge. Like behavior. Yeah, Bobby and I were talking on Monday and it was like yeah, I mean, we were talking about Snowflake and Databricks like which one's better? It's like uh it's going to depend if you like driving a Toyota or a Ford or whatever, right? Like they're just cars. But it's really about that people. And what gets me jazzed about it is just having people that are getting excited about it. And then similar story. We did an implementation for a company and they had a bunch of accounting people, pretty tight deadlines around their accounting close process and that stuff. It's taking them somewhere between like 8 to 10 days to do a close. Lots of process in there really difficult. So lots of them had like young kids. So for those, you know, 3 4 days before you publish your results, like they're in the office till 12 1 2 like and it's always funny cuz I tell the junior people in our group like I don't know it really matters. I'm like, well, why does it matter? I'm like you don't get to call the street tell them we're going to publish our results a little later. [snorts] And implementing some of the solutions to automate all their reporting processes like they're home at 8:00 p.m. And I had a woman like almost in tears Yeah. a couple of years ago that was like you've changed my life. Yeah. And she told me some stuff I won't share on the podcast but like this is really like impacted my family life and it was tough and we were getting babysitters and like doing all that stuff. And now I'm home. Right. I was like, well, that's the stuff that I mean the bills are going to get paid but that's the stuff that really gets you jazzed about doing the work. No, I mean I think that's the that's the the dream state I think of most kind of tech especially on the data side, right? Like the data analytics, business intelligence stuff is make your life easier, allow you to focus more on what you know they hired you to do, which computers can't do, what uh AI can't do. And ultimately the goal of that is to allow you to have more time to do whatever you actually like doing and being at home, spending with your family, all of those things. So it's uh and and I mean our industry is just so ripe for that across so many Yeah. so many spaces. So it's it's it's cool. I have question for you guys. Like do you feel like I would say a decade ago the sentiment that I heard was like oil and gas was they're not data savvy, they're behind the times, they're kind of Neanderthals when it comes to the data analytics space. Yeah. I feel that way anymore. Do you guys feel that way? I don't feel it. I mean, I think it's even gotten better in the last couple of years. I mean, I think even like 5 years ago, people were like, oh, we're so far behind. [clears throat] And I think we've had the conversation before though too, but I think when you throw in how much risk is involved in some of the stuff, especially on the you know, actual physical side of it, like you can't run machine learning algorithms in the field that you can run, you know, on a marketing campaign like you know, because the risk is way bigger like um but no I mean I I think it's definitely caught up. I mean I and I I think people always undersell it. When you talk to someone in healthcare, oh, we're so far behind. You talk to someone in another industry, oh, we're so far behind. Everyone thinks that they're behind. Like I mean, I think it's relative to your industry too. Like are you solving the problems you need to solve you know, with the technology that's available? Yeah. And I think, you know, the other another component of it that I've kind of thought of over the years is I mean, oil price drives a lot of this. And I don't mean that to say that high oil prices means we throw more money at software and analytics and stuff. It's actually typically the opposite, right? Like it downturns force people to find new ways to be creative and cut costs and automate things and streamline things and try new technology because they have to because oil is $35 a barrel and their break even price is $30 a barrel currently or whatever it may be. Um I you know, I've seen I came in in 2010 as a field engineer and you know, when prices are above $80 $100 a barrel the at least historically the mindset has always been get it out of the ground as fast as we can before it drops back below that price, right? And so people aren't necessarily as prone to do, you know, testing or data spend money in areas that isn't deemed necessary, so to speak, because you know, if I can shave $2, does it really matter if price is $100 or $102 a barrel off my operating costs? It doesn't. Uh it's more about getting it out as quickly as, you know, we can. And so these downturns actually end up in my opinion spawning a ton of innovation and and change in the industry because we're essentially forced to, right? And so that's can do more with less. Right. That's you know, like you're saying before we got on here, the best way to do that is with computers. That's what they were made for. We get the most interest in that like $50 to $60 oil range. You know, where it's like we're profitable, so we're not trying to squeeze a penny. But it's not Yeah, when oil was $140, everyone was like Why am I going to spend time I'm not going to throw resources at anything else. to say that's the other problem. Yeah. They're like let's go for let's go for long lunches at Morton's cuz we're whatever anywhere we stick a hole in the ground, we're going to be profitable. and it also gets to the point too where resources become a problem, right? Like if oil is so high and they can't get enough people in seats, you know, two years into my career and from 12 to 14 I mean, they were just hiring kids straight out of school into completion engineer roles to basically manage paper and documentation stuff because they had so much activity going on. And then they slowly they weren't able to find it like you need more resources the more higher oil prices are because the more wells you're going to drill and all this stuff. And so that means you don't have the time or the resources, personnel primarily, to focus on a new thing or trying out a new technology. It's just blow and go versus to your point once prices kind of get in that stable $50 $60 range, people are making money, they're not making a ton of money, so and the activity has slowed down, so they actually have time to potentially go do try things out, try and figure out how we're going to implement these new things and it actually makes a bigger impact on the business, you know, percentage-wise. It's a It's definitely an interesting [laughter] industry because of that for sure, but I agree with Bobby as well, you know, prior to coming here I worked at a at an edge computing startup and sold into, you know, retail, manufacturing and and some even cellular. And a lot of people are in that same boat where it like you know, we're decades behind or decade behind, or we're slow, or whatever. It's I think, you know, the biggest thing to me is like, it really just depends on when the industry started and how long they've been around, and you know, is the company that you're working with in that industry a fairly new company, or have they been around for 30 years? Because that changes what their tech stack's going to look like. It changes how they approach technology and all that stuff, right? Like, if you're a plant and you've been operating for 50 years, you're probably running on a rather old version of Windows on virtual machines on prem somewhere, versus if you're a brand new one, you have a completely different arsenal because you're starting fresh with modern tech instead of having to figure out how to rip and replace and This is what I find most interesting. I remember working at about 40 different companies. I have exposure to maybe 20 of them. The companies that are really progressive on like data, and I mean this from a culture standpoint, not like working with the that are really progressive, they're the best ones. Yeah. Like I'll write them down for you guys after the session. You can file it away and go look at it in 24 months, Yeah. and you'll see that those are the companies that are doing extraordinarily well. Yeah. Like the executive teams and leadership teams that really focus on that, it's such a huge differentiator, right? And I don't mean IT data. I mean like a lot of mostly engineering and like who can produce a barrel a little bit cheaper. It's those companies that are being really progressive that ultimately are going to win. And I've seen the opposite side of that Right. where everyone's like, you know what? Just dump some data out and we'll do our LOS reporting at a, you know, area level or state level and we won't worry about it. And then soon as prices go down, it's really hard to change the culture Right. of your group to do stuff a little bit differently. Well, then on top of that, you don't have the money to implement any change because prices are too late. Yep, absolutely. At that point, someone's making an offer like, you know what? We can operate that well Right. or this group of wells is much cheaper. So let's just buy them and that makes sense financially. No, I definitely agree with you and that I think that's been a consistent theme across most of our our guests is it's got to be almost organization wide. Like it's a culture, it's a mentality, it's a top down and bottom up. It can't just be one or the other. Um because you know you start getting thing is it can't come from the middle. Yeah, [laughter] absolutely. It's tough to come from the middle. I mean No, it's usually a leadership thing. Yeah. Yeah, I I I can there's a company that we work with or or actually we don't work with them now, but we worked with them for three or four years. And their CIO at the time was like super invested in data. He knew that was the differentiator and had a great conversation a few weeks ago with kind of the leader of that data team. And I was saying we were we're talking about a project. I was like this could save you guys a million dollars, okay? Just as an example. And he's like yeah, I know my boss won't even want to talk about a million. Like it's not even on the table. He's like unless we're talking about like 10 to 20 to 30 million dollars, like take the expense out. It doesn't matter. Right. Unless we're trying to like catch a big fish, it's like I don't want to do it. And it was like I almost fell out of my chair, right? Cuz I was like hey, let's let's spend a little bit of money and we'll save a million bucks. Who's going to not want to do that? He's like oh, that's not even close to big enough. not five, six salaries and benefits. I mean like for minnows. Yeah. No, I mean that's what happens when price gets above $80 $90 a barrel. People people start Well, the other thing that's interesting less about it, you know? the other thing that's interesting is like they believe that those that savings is there, right? So by being innovative, they they think they can reduce their op costs 10 20 30 million a year. Like that's that's stuff that you're going to talk with the board about. Yeah, again, but like but if you guys come in and save a million one time, it's like well, where else can we find another million? Well, well, that's our business model, right? Our business is sort of like we're going to try to build a brand where we're helping and and then that leads to more work for us. So like we're all project based, but we're hoping one project leads to another project leads to another project and so So so let's talk about your the business model a little bit cuz I mean like honestly when I first heard about you guys, I was like that's exactly how I would do consulting if I did it. You know, we've talked about this a little bit. Thank you, Bobby. Yeah. Yeah. Um They no one can see, but I'm sliding a duffel bag of cash across the [laughter] Under the table. Under the table [clears throat] right now. No, but I mean like I mean before, you know, we've implemented some certain tools at GME and I'm like, well man, if I ever got out and did my own consulting, I would do it this way because you can So can you kind of talk about kind of the modular almost approach and it's almost a blend between like a product and consulting and just kind of how that works. Cuz I mean obviously you can come in and do a totally green field thing if you needed to, but you guys have a lot of stuff in place. So just kind of speak to that. So general rule, uh new companies starting out, they don't have automated LOS reporting. I'll use that as an example. Uh so we come in because we've done so many of these projects before, we're going to implement a framework, you know, error handling, reporting, the integration, if you need to do master data or data cleansing, we would do that sort of stuff and we publish the solution. So we would like to think that we could do it, you know, 30 40 50% faster than someone just starting from from scratch cuz we got a lot of code and things we can drop in. So we'll come in and do a project. Again, we call it a solution. Uh it's hard to make it a product product Right. everyone's doing a little different. Just simplicity of like if you're an oil company versus gas company, what you want to look at, Right. transportation, all that stuff would could be tracked differently. So 70 80% is the same. You go implement it. Hopefully the client's happy. They get good use out of it. And then they say, hey, can you help us do this and do this and do this? So we do mostly project work. Uh come in, do that, build out projects and that just sort of leads to more projects. Yeah. So that's kind of our model. Yeah, it's not really software cuz we're we don't sell any software, [clears throat] Yeah. but it's a solution that if you looked at if we did LOS reporting for Grayson Mill and then switched and did it for, you know, company X and you looked at the two, you'd be like, yeah, this looks pretty similar. Yeah. [laughter] Yeah. It's like I mean you're there's not there's there's probably minor differences say on the let's just call it and you guys did a lot of Power BI. So the Power BI report is a lot of the same guts. I mean I I feel like up to a certain point there's a lot of the wiring, but I mean after that I mean and then tweaking some stuff downstream a little bit, but I mean Yeah, how you group accounts, how you group your wells into different reporting areas, like the how you define what gross is or how you net stuff out, that kind of stuff, yeah. Those are all like nuances of it. I hope that's the value of it cuz we're really customizing it to whatever the customer needs. And I think that's maybe a differentiator for Dakota. It's like we really want to build a product that you're going to leverage the most. Right. Again, the idea being around change management is like how do I produce the report that you want with the least amount of change for that business person? Cuz they have to go run their do their job a different way, right? So, that's why I find with a lot of the software packages that implement stuff, it software in general is fairly rigid in terms of like what you're going to get. And if that is too different from your current business process, it can be difficult to get adoption. Yeah. And we've all been through the like going to work around it cuz they need to get stuff done and and we've all been in that world where you know, you're like this software's awesome, you implement it and then like six months down the road no one's using it and you're like wait wait what happened? We and like it's just too different from their business process, right? And everybody's busy, so. Yeah. No, I mean that I've never thought about that way, but it does make a lot of sense on the kind of consulting side. Just you know, you you have a framework, so to speak, right? It's not necessarily templatized, but 80-ish percent of the leg work is mostly done because most people do things in a handful of ways. But the benefit of it is instead of buying a out-of-the-box software that you know, is super structured, locked down and you know, maybe it's got an API or whatever, but even then it's like okay well, now I've got to figure out how to structure it how I would do it versus how they want me to do it um or you know, I need this data to flow into this other thing and so now I've got to figure out the whole API and the databases and all that stuff versus hey, we can give you you know, the same product, but it's actually better because it's completely customized to your architecture where you want, you can get it where you expect to receive it, how you expect to receive [clears throat] it, and how we can have it flow exactly where you need it to go without you having to do all this extra crap. And that's one of for business people, like that's one of the big things cuz we have an engagement, we're on-site effectively for, you know, a month, two months, three months, however long. Being there and having that regular communication, which really is knowledge transfer, training, you but uh and being able to do that versus like when you implement software, it's rare, let's say never, but it's rare for like I use Peloton, Peloton to drop someone into a client site for like, hey, this person will be here for six months just help you if you have questions. That's not everyone through Well, either that either that you're going to pay a lot of money for those guys to do that. happily do it. You're going to pay You don't see it happen often [laughter] is my point. Whereas for us, like that is the model. Like we literally want to sit at the client site. We want to have the water cooler talk. We want to build trust and help you understand like what are your actual pain points, right? For like things that you want to resolve. And that is what brings trust and that's how we build our brand up and that's how we get more work. We don't do a ton of marketing and stuff. It's mostly like through word of mouth. Yeah. I I mean I imagine that's that also results in a much better product too for your end users, right? Like that's one of the biggest things that I mean I'm a mechanical engineer because I I was sick of how poorly designed things were, right? It was like you can tell they designed it on paper in a computer or whatever, but then the person that actually designed it never put their hands on it or they wouldn't have designed it that way because they made some stupid [laughter] thing fit into a spot that was impossible for you to actually get into it to replace or do any kind of maintenance on once it got to that point. And so I software is the same way. I think that's why I like software so much is [snorts] you know the user experience of using a software is so nuanced and customizable or is can be nuanced and customizable specifically for you know these individual workflows at these different departments and all those things that you know if you don't have someone sitting there watching them do what their normal workflows are and peering over their as they do it and taking notes and asking questions, you get a much, you know, it's that turn-key experience where it's like, okay, I think [clears throat] I click this button and then I think I'm supposed you're having to it's that change management. You're having to relearn this entire new thing on top of trying to get your job done like you normally would. So just having just being able to have, you know, it's I mean it is it's custom software development, right? Yeah. And so that's that's a a nice thing because there's so many softwares out there, especially in the oil field that are so poorly [laughter] done from a UI UX perspective. [snorts] Yeah, UI UX is probably one of the hardest things just in general. I mean like Yeah. I mean all the data pipelines like you said like all that we can talk about DataBricks or Snowflake or whatever. I mean like those are really generally solved problems like you can kind of figure out. Yeah, most engineers go in like spend 10 minutes like okay, this makes sense. I got it. Like it's just like do that endpoint then people are like well this doesn't do what I thought it was going to do like or or it doesn't work in the way that I thought or This is the problem I'm trying to solve and it doesn't do that or it's more than three mouse clicks. Or can you give me an export to Excel button? Yeah. [laughter] That's a good good segue. So most of the solutions models things that we build, we obviously Power BI you can analyze in Excel so you can use Excel as an interface. You guys had someone on the podcast a while ago said Excel is never going away and I was I was listening to it and I was like absolutely. Yeah. Right? And I'm supposed to be progressive and a technologist and like a visionary and all this space and I'm like Excel's not going away. No. It's just so easy, right? For everyone to go use and it can solve the problem Yeah. really fast. So all the models that we create if we build Power BI spot for our Tableau reporting, there's an Excel like you can not a download like it's connected to the data set. So you can pull in do all your reporting and that ironically is like in the finance accounting space when we do projects there, Yeah. get all kinds of traction. It's like wait, I can build this report in Excel and then just hit refresh next month? Yeah, that's how it's supposed to work. But a lot of people you're looking at me Bob you're like I'm crazy. A lot of people are like, this doesn't seem like it's possible. I'm like, I can't believe you're doing downloads of CSV files. That's what I can't believe. Wait till you tell people you can put 10 million rows into Excel technically. [laughter] Yeah. We get lots of projects because people are trying to ingest a few hundred million records and ingest rows. people don't realize through like Power Query and like they can load it to the data model or whatever and you can build your pivot table off of a 100 million row table in memory. Like Yeah. The education part is a big part of this job, you know, when you work in data analytics for sure. I mean, but to your point, we had a lot of success at Jimmy just I mean, meeting them in the middle. Like you can you can be the people that say, no, we're not using Excel. We're going to or you can say, this is you know how to do your job with this and like and again, instead of you going to your ERP and exporting this same stupid CSV every month and doing all these things, just hit refresh because I have an ODBC connection or a Power BI data set connection and it just works. it's one of those things too from a just a management perspective, right? Like damn near everyone knows how to use Excel. And so hey, if you leave and we have to replace you with somebody, they probably know Excel, right? Like and I think it kind of ties back to the [clears throat] UI UX stuff, but like serving people, whether it's data, reports, whatever it may be on the software side, serving it to them in something that they're comfortable and familiar with is a much much better experience than trying to get them to learn this new thing or go through this workaround workflow or anything like that. Right? Like I hope all the junior people at Dakota that started listening to [laughter] that. Yeah. Like we have this debate because like, oh, let's do it let's do it on Power BI or let's build it in Spotfire. I was like, oh, they won't use it. Like they need Excel. Yeah. Given who the audience is. Yeah. They're like, no, no, you don't understand. Like you can do way more stuff inside Power BI. Like no, I I know how Power BI works. [laughter] I know how it works. I'm telling you what they're going to want. Well, I think yeah, I mean, that's the other big thing too is like oil and gas is is also pretty unique in the fact that so many different departments need some, all, or a portion of so many different data sets coming from all of these sources. So like just the the data architecture and everything on the background, all your models and stuff, like being able to serve the same to serve information to a completions drilling engineer that also ends up serving some of that information to an accounting finance person, all from, you know, the same group of sources is very important and it has to be, you know, like ideally it's all in one pipeline, it's coming from the same places, it's all been synchronized, the data is structured, it's cleansed, it all your ETL's done. And historically that has never been the case, it's always been all these silos and CSV's or emails or you know [laughter] it's

So 40 clients, 20 years, 50 projects, like we I won't say we connected all of them, but a lot of them. He's like so have you guys connected into Palantir? [laughter] I'm like yeah, it's pretty pervasive. We've Probably two thirds of the clients we were with at least Yeah, [clears throat] the the question was pretty telling, right? Cuz I was like man we connected into it like we connected probably 40 times this year. Right. Yeah. Like [laughter] what what do you what field varcar length do you want to know kind of stuff, right? Like we were pretty intimate with it. He was like you guys connected into that? Yeah, we've connected into it. But there's lots of companies out there just starting down that path. So they just don't live in this space, they don't know about stuff, so hopefully we can keep helping them. Wasn't that I mean like maybe this is slightly off topic, but there was a story it was like maybe I think it was oil and gas executive brought someone in was like hey I got an idea for like a startup and basically your phone could like hear a song and it could tell you who sings it and they're like they're like yeah that's [laughter] Like, that exists. Like, I mean, some people just don't realize what's been solved or what's done, like. In that area era now of like there's more information than you can process, right? Like, I grew up in an era where like you had encyclopedias. Yeah. like you read them all and then that was it. There's no that's all I was learning. yeah, you remember like that's just I thought about that the other day, like how that I mean that experience is basically dead, right? Like, Mike, our kids [clears throat] will never have to go Yeah, I used to sit down with a baseball almanac. more data now. Now it's about like choosing what it's really important to choose which which data or which information you want. Yeah. Cuz there's more information being created than you could possibly digest, right? So It's going to get even And it's only going to get worse. It's going to get weirder with all the AI stuff too because the AI is going to be optimized for being at the top of those results and And that in our space that's the like curating the right data. I mentioned three mouse clicks before. Literally like four mouse clicks and like there was a VP of work who was like that's I don't know how to I was like no, it's one more click. [laughter] Click one more time. That's the answer. He's like I don't know. That's four clicks. So I was like okay, go back to the UI, change it so that it's three. Yeah. No, I mean but you're if you're clicking buttons all day, like it sounds stupid, but going from four to three on a workflow that you're doing multiple times a day or a week or whatever like it has an it has an impact on the user experience, right? Like, I know you know. I'd be retired if I got a dollar for every time that questions [laughter] come up. Yeah, it's pretty it's pretty critical, especially when you're a data worker Yeah. and you're typing and clicking keys all day. Well, that's the thing like developers can get so hung up on like, you know, getting the getting the feature done or checking the box or whatever that a lot of times you know, I I know everyone that's ever had any kind of development work has run into that where it's like give the developer a task and they bring it back and it's like hey, it works and it's like yeah, but the average person wouldn't use it this way, right? Like, or it's not intuitive or it takes three extra clicks or it's like understanding how the the end user is going to be actually doing this and what they're currently doing is so important because it can be a great software, but if it takes them twice as long to use it just to get to the end result, who gives it They're not going to use it. I think that's what's lost a lot is like adoption is the measure of success for us. Like are we getting adoption? Are people using it? So either someone's using it a lot or a lot of people are using it casually. I think that's how you measure it. Agreed. So with that, I mean do you have anything in the stack that you deploy that's measuring like where you're tracking like how much a dashboard or a certain data set or whatever's being used? So so two ways. One just on the data side. So if you're curating like connecting your own tables and building your own stuff, we'll query that Yeah. and then provide stats on like how much it's getting used. Okay. And then specific to Power BI, we have Power BI usage reporting which will connect which is native already, but we actually connect into the uh like the back end of Power BI to be able to connect to you even if you're using Excel. Okay. just doing queries in Excel, you can keep track of like who's using it and that stuff. And it's it's just something we can plug in Yeah. and people can use and it's fantastic because you'll get uh whatever, someone in operations like I need this. It's critical. I got 50 people. They're all going to use this tool. It's going to be the best tool. Yeah. Drop everything and work on this thing. And you're like, okay. And then you go build it for them. And then someone in another group's like, are they using it? It's like, ah they clicked on it four times last Yeah. [laughter]

Yeah. And I mean like I mean I get them we're and we're that's probably an evolution that we need to get to at GME. We have some stuff where we can see, you know, which reports are being used most, but I remember actually my my buddy Alex I think we're I've talked about trying to get him on to the podcast. He works for Oxy. He's a AI manager now, but he was doing some really cool Spotfire stuff, but he and another guy they they wove Google Analytics into their Spotfire projects so they could see where people were clicking and when they clicked this where they went to the next thing. It was it was pretty cool to like then they had the stats in house to see like how people were even using the reports. Yeah, I think it's great to be able to drive like where should you be spending your time and your money next. Yeah. Yeah. If you're getting good traction, I I I mentioned before those power users that are going to go really move the business forward. Yeah. If you can figure out where they're going and and help them, you know, or or even see them doing something that's not that efficient and be able to be proactive to go help them, like make that faster or better. And then for you all even to say on the sales side, you can find your champions. [laughter] Like in a way, I mean like [snorts] Knowing who's using it the most and how they're using it. you know what I mean like nefarious big brother thing, but I mean like you can identify these are people that you know can help us, you know, in the business are you know or that believe in us. Yeah. You can tell that in a five-minute conversation too. People's enthusiasm for it, right? It's pretty evident most of the time. Well I mean I got people in our accounting team it's like there's a two or three like those are the people if I know I want to get some adoption like they're all in like and the other people there's we're still working on them like you should probably do this a little better. But like there's like two or three in there in there higher up anyways so it's great you know and they have that kind of push down effect, but yeah. One uh thing I'll throw out there cuz it's one of the points of the podcast, but uh one of the tools that a lot of people use on kind of the web marketing side is called Hotjar and I highly recommend it, but it'll give you heat maps of it's cloud-based, but it'll give you heat maps of where people's mouse is, where people ran into you know they fell off essentially which means they couldn't figure out what they were supposed to do next. Um it'll give you anonymized recordings of the actual session to show you like okay they landed on your home page on the website and then they clicked this button and then they went over here and then it looked like they were trying to click on something there, but it wasn't actually a link or it was a dead link and then they went back and so it's like oh that's a really good way to to [laughter] figure out you know what where people are getting stuck you know a lot of times or where people are spending a lot of their time and stuff. It's uh it's something I originally found out about when I was having those guys look at redesigning our RDS one um I had never heard of it, but you know like the the UI UX guys love it because it gives them direct insight into what how the user's navigating it, what they're doing you know and cuz that's one of the again one of the hard with developers, even on the front end, is hey, this looks like this really obvious button and this is the workflow and you know the workspace, but you're so into it that you don't realize that someone who's completely fresh to it has no idea what they're supposed to be doing just because you designed it for yourself almost, right? [snorts] Um I want to get into a couple of things. One AI and two kind of the the joys of of oil field data ETL and databases. Um which one should we Well, jumping on AI, I mean, the first thing I want to know is when is AI going to replace dbt Analytics? [laughter] Hopefully not soon. Yeah. Uh

I think just like you're joking, but yeah, it'll augment the work that we do. Like we're already obviously leveraging AI for a bunch of stuff that we do making us more efficient. Again, AI is a great tool. It's I don't believe it's the answer to everything, despite what most people say. So yeah, I think there's a place. It's growing. I think to the It'll grow at the pace that humans can adopt it. Like it's there's no limitations in terms of what we can do. So What are some interesting things you're seeing kind of focused around the the data analytics and BI space with that? You know one of the most interesting ones is just like a migration of data platforms and being able to leverage AI to do a huge part of that. There used to be a lot of heavy lifting. Right. If you wanted to swap out from like uh traditional, you know, Oracle, SQL Server database and like move it to the cloud and go to SaaS and do that kind of stuff that used to be a, you know, two, three year project at some some of our clients to migrate everything. And I think now there's just so much opportunity to be able to leverage a bunch of the AI tools Yeah, what you're saying. I think Andrew was saying the other night at dinner like, you know, he wants to he wants to convert you know, Microsoft SQL Server T-SQL to Databricks or to Snowflake. Like you just pop it in and say convert this SQL dialect, another one. Like, which sounds ridiculous. I mean, they're and they're very similar, but there's enough. careful, because like, people are, like, there's people in the office getting shamed. Like, why are you recoding that? Like, toss it into ChatGPT or Bing or Yeah. It's like, hey, they don't know. Like, [laughter] you know, show them the first time. Don't make them feel bad about it. So. But again, like that augmentation that, you know, I just It's again like like Microsoft's calling it a co-pilot. Like, allow it to help you. I mean, like, it's there. I mean, don't spend, you know, 3 hours banging your head against the wall. Like, go ask or I mean, and that or I mean, ask other knowledgeable people like you were saying the other night. You have a lot of really knowledgeable people at Dakota, so talk to each other. But also, like, you know, let the computer help you. Yeah, exactly the same way as, you know, like the simple analogy is [clears throat] you used to do, like, arithmetic on a piece of paper and stuff. And now it's like just learn how to use your calculator. Yep. Yeah. Right? You don't need to know what the actual math is to do it. It's the same thing with the AI. Just which tool do you need to use? Because there's a lot of different tools, Yeah. co-pilot and things that you can leverage, sort of exploit doing work. But fascinating to me because obviously because we work across a lot of different clients and some people are really really good and like they're churning out work at 300% of what Dakota could do, and some are at, like, 3% of what Dakota would do. So you get the whole spectrum of, like, what people are leveraging. And that why I say, like, those companies with the data culture that are really progressive are doing it at the 300% and like, well, you got to win at some point. Like, if you're running that much faster than everyone else, at some point you got to win this race. So I think one of the interesting things for us, and we were doing more of, like, advisory services of late where everyone's like, I just want to hear your perspective. Like, come in for a week, let us show you what we're doing, and then compare us to lots of your other clients Right. Sure. to see, like, where they fit. Because I don't think data people in general are that collaborative. I think we're collaborative in things we want to talk about, but not the things like the, you know, dirty laundry on the floor. So, like, we're not collaborative about the things we don't want to share, and we don't get good perspective. Yeah. Like, you know, Bobby's at Jimmy, like they're the best data company in the whole world, right? I'm not going to say they're not. But there's probably like things where like, oh, you should look at that. This is kind of different or you're really you are really good at this. You said before, most of our clients think they're terrible at everything. We're like, you're actually doing really well. Yeah, yeah, you know, when you don't have perspective on like where other people are at. I mean, one thing that kind of came into my mind there too is like, what about like on the documentation side? Cuz I mean, and so do y'all provide like say data documentation like a data dictionary glossary and like have you even looked at some of like the AI stuff for augmenting that as well? Yeah, for auto creating documentation on stuff. Well, we already have a bunch of stuff published, remember, frameworks, right? So you're going in, you're creating a standardized solution. So we're taking the general stuff, making a few tweaks to it, and then it goes. Um in terms of like functionality, like of the tools and stuff, the vendors like, you know, Tipco already provides a whole bunch of information like here's how you do it. Sure. Um so we're seeing a little bit of that around the AI, but yeah, cuz documentation is one of those no one really wants to do. No. No, it's it's like the perfect use case for uh a language model if if if it's good at it, right? Cuz no one wants no one likes doing it unless you're like you're building something for yourself that you know you're going to need to go back into multiple times in the future. Just dump it and click the button. I mean, [laughter] it's Yeah. And most of that's the other thing too, most of the time it's better than what a normal person would have done because they wouldn't have documented it as well potentially. Mhm. Um Or it's hard to have Well, I mean, I guess It depends on the person of course, but Well, that I was going to say it could be maybe a little more standardized cuz the model produced it. Right. But the model's also very uh can be a little what the creativity level is. ones for us, Bobby, is just creating documentation like inside your code. So like as you're doing data integration across sources or like putting business logic into the code, Yeah. typically tends to be like one, two developers. So it's not a lot of people doing it. Just being able to leverage AI to document what's happening Yeah. and kind of like why it's happening Yeah. so that, you know, six, 12 months down the road, when someone calls the code and says, hey, can you support, this isn't working, like you can go in and look and find out really really quickly what it is. And that's it's changing the amount of time to do a support ticket from like 15 minutes to like it could be days, right, if you don't know what's going on. Yeah, you got to dive in and you're like, what the heck's going on here? Yeah, so very useful in that space. Yeah, I think it's also, I mean, just in general, like for the the uh generalists like myself out there, being able to take, you know, go find something on Stack Overflow and dump it in here and have it explain exactly what I'm looking at is just such a better way to learn, you know, the different things, the different parts of code and you know, different ways of doing uh of programming because, you know, it's like your typical workflow is very slow and and kind of arduous at the beginning when you're learning to code and stuff, but most people just want to jump in and solve the problem that led them to try to figure out how to code to begin with. no no, how you format it matters a ton. Yeah. Right. [laughter] And so, you know, just being able to have that where it 's like, hey, tell me what this is doing or, you know, explain to me in common language instead of you know, all [snorts] the the loop and the class and the libraries and all that stuff, or even just like can't tell you how many times I find write-ups, like full blog articles, how to do this with Python, blah blah blah. And then you go and you know, you copy the whole piece of code over and you hit run and you're like, oh, well, they forgot all the libraries that I needed to use to run this code. And then, but you know, now with AI, it can tell you that. Before, if you didn't know what you if you were brand new and you didn't know that it needed all these libraries, you would spend hours, I know this from personal experience, trying to figure out why it wasn't working even though it's exactly what's in the stupid write-up. Yeah. And then I call Bobby and he's like, oh, well, uh did you put did you install the the libraries? Oh, no. No, I didn't because I was rushing and trying to copy and paste and get this done, right? Like just being able to have those little things that will prevent you from wasting so much time is is I think going to help a lot of people quite a bit. Yeah. Do you see any uh what tell us a a little bit about what the kind of conversations with different operators have been like or clients um around the AI stuff? Are they still Are you getting any questions or anything like that from them yet or is it they're kind of hands off and watching? I I think there's lots of interest. Uh I think there's lots of like kicking tires. Yeah. I think trying to decipher the value proposition of doing some of that stuff can be really tricky, right? It really is until a bunch of places have done it, published the results, talked about it conversationally and said, yeah, if you invest a dollar, you'll you'll return $2 you get a lot of just kicking tires around stuff. So I I wouldn't say there's a I don't know. How do I describe this? There's some going on for sure, but I wouldn't say it's pervasive yet. I I think it will be the next few years. We just look at the general innovation like curve, what it looks like. You've got a few early adopters that cycle. willing to take the risk and go do some stuff, but it's also a fairly expensive proposition to sort of get into to start doing. So you're not talking about, you know, a $5, $10, $20,000 investment. You're potentially seven figures, right? To go do it correctly. So I think people are kicking tires. I also think the other thing is a lot of people are underestimating the amount of data quality work Yeah. required not like required before you go do it. Yeah. So we all know the adage, 80% of every AI machine learning AI project is just doing data cleaning and cleaning, right? And that's entirely true cuz most [clears throat] of the data you want to use is from sources that don't require really rigorous data quality or there's lots of gaps. We're looking at like, you know, public data and stuff like that. So that's been the hard part. It's like, let's go do it. Yep. Data scientist is like, yeah, it's going to take me 3 months. And then like a week in, they're like, it's going to take 4 years. Yeah. [laughter] Well, I think you're just like narrowing the scope. I mean, right now you just see all everything that people are doing with LLMs and like there's any number of spots in your business that you could try to apply it, but like you really got to narrow a scope and say like I want to use AI to solve this, you know, like I I think the other thing is like and maybe we shouldn't say this as data workers, but like most projects fail. Yeah. Yeah. Depending on how you define the criteria for those kind of advanced analytics projects, most of them fail. Yeah. Like they do not yield the result. And it's always interesting to see people spin it publicly Right. that like, oh, we had all these successes. And I'm like simple example, someone's like, hey, we did this project. It's saving us $20 million a year in OpEx. Okay? It cost us whatever, a million dollars. So that's fantastic. Yeah. Let's go look at your financial statements. Look at last year, look at this year, and it's not $20 million different. Right. So I'm like, well, Yeah. did you just decide to spend $20 million more somewhere else? Is that the survivorship bias? That's the one that worked. So I think that's the [laughter] And And again, like for us, I think it's like that trying to be really candid with our customers and our clients to be like I haven't seen anybody do it yet, John. Like sure, like we want to do it if you want us to hire us, but like I haven't seen anyone be successful doing it yet. I think a lot of executives appreciate that candor. for sure. Yeah. When it was $100, maybe you want to roll the dice and go do it, but But I mean like think to your point. I mean like you guys can do that, but you got to lay the foundation of what you do as your core business first because you guys focus a lot on data quality and getting the data in a good spot as well. Like Yeah, so we'll do a lot of projects where it's just like, hey, we just want to do data cleanup. And when you're ready, we'll turn it over to the data science group. Like we don't have We're not engineers, right? So we don't have like production engineers or reservoir engineers on staff to do that work. We don't live in that space. Yes. That's a profession. But I mean and one thing I think just kind of coming back a little bit, you talk about like these basically experiments failing. I think one thing that always gets lost when people talk about data scientist is that it's data science. So you have a hypothesis and then you're out testing it. And often times you don't you know, it doesn't work the way you think it was going to work. Like it's not a linear like, oh, I do this, I'm going to train a random forest model and it's going to yield 30,000, you know, $30 million. It it's not how it works. Like, it it is a, yeah. to it. It's very iterative. is a play. Yeah, well, what kind of artifact are you trying to get to? Like, you know, that's those are the things that you're trying to Yeah. And when you ask someone like, well, how accurate or how precise do you want to be? Yeah. And the executive's like, I don't know. What like, when would I make a decision differently? Right. But then it's really good and it's over fitting and then it has problems like played around. Yeah, you you can As we all know, you can iterate through models for a long time. So, it is hard thing for us to estimate that work as well cuz it's like I don't know when it's good enough. Right. Well, then you have the flip side of it too where it especially in our industry is it's like you've got to how many companies you've seen where they've got their black box AI that does whatever and then the biggest hold up is they can't explain the results to the operator or to the client. Like, there's no transparency into what's happening behind the curtain, so to speak. And then the operators are like, well, it can't be right cuz you can't even explain how the hell it produce these results. And that's millions of inputs and There's also lots of vendors, certainly for like subsurface, you know, the drill complete, that kind of work that lots of vendors that have invested tons of time and energy and money in already solving the problem. And I really think that it's probably that is a product. Like, I don't think it's a competitive advantage for you to go try to build something that might already be built. Yeah. So, and certainly like if you're talking about like, you know, you've got some AI vendors that focus primarily on Permian, so like it's they're kind of the same wells. So, like to take that IP that someone's built and just buy it versus trying to build it, a lot of times is going to be a lot cheaper and and better assuming you have data quality. We're talking to you, operators. No, but I think that it kind of made something click in my head. I mean, like cuz we I think there's a lot of this talk about operators, you know, build versus buy or whatever. But I mean, like operators, you know, what say what I do who what your company does for operators is build systems. and I think any good company should have good systems. You know, but like they shouldn't necessarily be product companies because I mean once a product, someone has to maintain that product, someone has to support it, someone you know like um there's going to be confusion. the end, you know, you can argue a Power BI report or Spot you know Spotfire dashboard or I mean those are products in a in in a way, but not in the same sense of like this is a software product where we've built this solution and now it's only ours, but now you have to support it for all the customers inside your business. Like um you should be building systems that allow you to plug into you know solutions like that and Yeah, that's where we would lean much more towards that, right? Like we're doing the data integration, providing the data, providing the tools. We don't actually want to go solve the business problems. I know it sound a bit weird, but we don't want to solve business problems. that the SMEs can I I think that all AI machine learning stuff requires a ton of business knowledge to be effective and to iterate through stuff. And we're focused on data and providing integration and doing that work, the foundational stuff. And I don't to date like we're not delving a ton into anything subsurface. Like I can tell you what a of like this is how you should be analyzing it. is, but I don't I don't ask me where you guys should drill your perforation. but I think a little bit on that side there too. I mean I think don't undersell yourselves too. Like I mean you guys understand oil and gas workflows and that's important as well. Enough to have a meaningful conversation and also I would say maybe more importantly, enough to know that we don't know. Yeah. Like don't don't ask me. Just just as important. More important to be honest, right? Like there's a lot of people that would pad their resume Yeah. around like what they're good at and what they don't do and I mean all the LLMs, right? They'll return something that could be completely wrong and be very very confident about their answering it, right? Yeah. [snorts] No, that's a I think I I kind of agree. I think the the approach in my opinion for anybody looking at language models and especially on the operator side is find a really small niche part of the business that ideally has really quality data that's structured and you have a high degree of confidence in and might not need a lot of additional human context and input and start with something small just to see because to your point if an operator wants to come and put their entire data set of all of their ops and stuff into an LLM that's going to take forever and it's going to be this giant monolith monstrosity and like you're going to end up spending tons and tons of money just on getting on the ETL cleaning the data, processing it, transforming all of the stuff that we know historically oil field has been pretty bad about. We have so much data, yes, but historically most of it has been unstructured, unstandardized and it's you know sitting in a [laughter] and they're either in a file cabinet or in their data lake which I know Bobby knows my favorite day like meme which is just a dumpster fire and they're like yeah it's in the data lake. We just you know pointed at that. It's like well just cuz it's there it doesn't mean it's worth a [ __ ] and it sure is [laughter] if it's not structured it's definitely going to hurt the model more than anything else right and so think a lot of people don't a lot of people don't talk about how fundamental the data input is for the language model or any of the AI stuff right it's like yes these things exist and yes you can train them and yes there's potential that they have pretty good impact but there's a lot of leg work up on the front side that you got to do in order to do that. I think that's the next two to three years coming back to your question Bobby's like getting so a lot of the obviously one of the bigger players or people really invested into data will already be down this path so they can discount the next 30 seconds but for the small kind of mid-range players for them to be able to compete they're going to have to delve into this space in a meaningful way and get access to data and do the data quality and the cleanup if they want to compete in that space to like you know how do we optimize plungers or whatever the whatever the problem you're trying to solve is. Yeah. So I think that that'll be the next few years for us as as the accessibility of tools. Right. And really for us, I would say the past three to five years we've seen tons of cloud adoption. So because you've gone to the cloud, now you have unlimited horsepower. You were always constrained when you were on premise, mostly constrained. Now you have this ability to really If you decide you want to, you can do it tomorrow. you can do anything you want to, right? You can you can spin up a whole new environment to do a POC in an hour. Right. Yeah. Whereas like that used to be wow, we got to order a rack [laughter] for the service. So being able to be much more scalable, you're seeing a lot more of those projects not suffer the technical constraints of like getting hardware to do work. It's just like okay, what's the business problem? Can we justify it? And then can we go to a POC and try to solve it? I think you'll see a lot more of that. Yeah. I've got a The cloud allows you to iterate so much faster like Well, and spin it up and spin it down like it's just the the cost is that didn't work. Now it's gone. Like Right. I've got a question for both of you actually. What is y'all's do y'all see given I think we all kind of agree that yeah, these next these coming years are going to be really focused on that foundation uh the data piece. Do y'all see it as a potential competitive advantage especially for the smaller mid-size you know PE backed companies um having good like having a good data set so that when they do go to ultimately get acquired which is what most of their goals are that ends up being a a differentiator for them to you know their customer. They either get a premium or they have a lot of more interest because ever gone into a data room and said it was good? [laughter] Yeah, no. heard that before. Yeah, I don't I mean like to we we brought on um the guy Sean from Ensign and then we talked to Andrew on the and like they had some really I thought they did had done a great job from everything I heard. Marathon basically threw it away when they acquired them. Yeah, we same experience. all the stuff I mean like we don't care. We just want the we just want dirt and we want the oil underneath it. Like Yeah, we had a client that bought another client of ours and and they they did some really cool stuff. I thought it was really cool. And it just didn't fit into their data strategy for what they wanted to do. So they basically just turned it off. So they didn't get value but I think the value I think more of the value was like pre-purchase for those small to mid markets where like someone's going to come in and say, hey John's got a bunch of data. Let's look at his data in the data room and run it through kind of our analytics and our processes. And oh, we've realized that we can carve out 50% on or 50 cents on transportation or if we group these wells, we'll get better throughputs or whatever it might be, right? So I think there's more value in that in trying to find value and stuff and I think if you're not doing it, back to competitive advantage, if you're not doing it, you're potentially missing out on those on those opportunities, right? You're like, nope, it costs you know, six bucks a barrel to get it out of the ground and stick it where I want to stick it. And I think if someone can do it for 5.50, especially at scale, that's where you're going to either pay a premium or realize value. you make it that obvious. You know, I mean just like you said, the usability of it, right? Like if it if you're looking at two companies and one has just a bunch of junk, you know, it's not structured, there's errors and missing data and all this stuff all over the place versus another that has everything clean and nice and ready to go. It's like it's almost like what, you know, there's almost a prejudice that gets put you know, you're setting the the floor essentially with what the people that are looking at that data are going to expect, right? And so you know, you're you're not doing yourself any favors by having a messy data set versus your competitors being Yeah, and the game the game is simple, right? Like you you better rock or you produce it cheaper. Yeah, so I mean there's two things here. So one of them is yeah, I think I think just good quality data in all these systems that we're building tie into operational operational excellence, you know, and when you are a better operator, you can prove that there's more value in this acreage. So then you're more attractive. Like you know, we came in and there were things that seemingly or clearly we did better than what Equinor was doing. And, you know, oil price played a big deal early on that really helped us. But like even before we drilled a well, we added tens of thousands of barrels just to our base by remediating wells and identifying those candidates. So I mean like all all of a sudden you just prove that there's more value in this than there was when we bought it. Um but to your point about the data room, I don't think it matters like in the data room. But I think you know, there were some data qualities especially say on our land side like and we were looking at maybe marketing some stuff or whatever like and when you have your ducks in a row and you can provide that data quickly, get every all your ducks in a row so you can go to market faster like while prices are in the environment cuz like you can this can swing. We've seen it especially lately like you know, we if you want to market you know, an asset at $90 oil, that's a lot more attractive. But if you don't have your data ducks in a row the way you want to be able to provide it and then two months later, three months later, the prices go down to $65, well now you're not getting the price you want and the opportunity's gone. So I mean how many deals are going to due diligence and 30 days go by and they're like eh Mhm. Yeah. We might need to change the price. I mean I think there is that side of it where if you really have your ducks in a row and and I would say we could probably spend a whole another podcast talking about land data and you know, the use of data in the land side. It's just I mean it's [laughter] um but if you I think especially on that side I think if operators have their ducks in a row on the land side, I mean there's a lot of a lot of value there and that more kind of time is money side of it to your point of like a data room kind of stuff. Yeah. What uh what kind of languages do you all mainly use? Do you all use using R and Python or Not a lot of R. Not a lot? Sorry, Jess. it's more Python. Yeah. Yeah. [laughter] Uh more Python. I I Again, part of this is just our preference. We're doing the same stuff, try to create commonality across code us so we can support stuff. Uh some more technical people than me would say they like Python better. Um Python, SQL like M, DAX, Power Query Yeah. those are kind of the It's the only thing I don't like about Power BI is that I have to learn two new languages instead of just one with Yeah. You It's okay cuz you can use ChatGPT. Yes, you can now. [laughter] That definitely helps. the tooling's gotten a lot better around it now some of the tabular editor and DAX studio and everything. made a lot tabular all the time, but they made it better. It's still I mean it is what it is, right? Like I mean I mean they're different purposes. I mean like and there is a little crossover, but DAX is you know for the kind of metric layer almost really. And then you know I mean I love Power Query for massaging the data and when we were talking earlier about you know uh Excel, I mean when you can get Power Query into some of these accountants people's hands and like then there's some traceability, a lot more traceability to what they're doing than like trying to figure out oh they have this X lookup that has this sum if [laughter] and you know like I mean we can go through and see like each step like oh they filtered this and then like cuz there's sometimes I mean again it gets out of my hands at that point. I provide them the data set out of Snowflake and then they connect to it and then they're doing their Power Query stuff and they're like hey Bobby, this broke. Like usually it's user error, but at least I can go and within 5 minutes we're stepping through and like let's just step through the steps. Boom, this is where it is. You had a filter on to July and it's August. You know like you're missing a whole month of data. Like that's the problem. You know, but like it's traceable You say that like it might have happened recently. yeah, I mean [laughter] No, I mean that's I won't name names, but no I mean um but again like I yeah, Power Query is probably my favorite part of the Microsoft data stack honestly. I think I don't know how we are already at an hour, but And you feel like I got to just I mean on the Python thing, Python and I think we said before is just such a Swiss army knife. I mean if you got people doing advanced analytics, but I mean between any of the cloud vendors they have a Python SDK. I mean it's probably the most worth learning. I mean like Commonality across all platforms. If you're working in data, I don't know how you get away from Yeah. Yeah. That would using it, so. But yeah, I guess we're running up on the uh Go go to the speed round. round. So we're just going to pepper you with some some random questions. Okay. Uh favorite board or video game.

Oh. I'm going to say Roblox because my daughter's a huge Roblox player, so I'll go with that. My daughter is on that all the time. [laughter] It's a problem at our house. Yeah, I'm I'm going to pull my hair out, what's left of it, if she asks me for Roblox or I've still Robux? Yeah, I still refuse to buy her any because I know it's just going to be this [laughter] fuel the addiction. But Uh, favorite cloud data platform. Today it's Databricks. Yeah. What about favorite uh

favorite open source anything?

It doesn't it doesn't have to be it doesn't have to be related to what you guys do at all. I don't know if I can pick one. Give me a couple that you you like or even just that you stumbled upon recently that you thought were kind of cool. For open source stuff?

That would be tough. I'll tell you one to to check out that I'm just kind of uh

blown away by. It's called Open Interpreter and someone I don't know how they do this, but someone found a way to get GPT-4 access to the internet and then created an agent on top of it, so you can like I literally spin it up and say I want to pull in my fantasy football team from ESPN and then it'll say, okay, here's three libraries that I found that can do that. Which one would you like do you have a preference? And then you literally just it'll just go through and start pulling the code and writing the code for [laughter] you. And what's it called? Open Interpreter. Open Interpreter. Okay. It's kind of a it uses the GPT you need a GPT API uh key, but I think it uses part of the GPT Interpreter add-on as well for the coding side. It's pretty slick. You can check it out. There's so many. Uh, who's your favorite follow on social media?

These are tough questions. Um, [laughter] you know, in our space? Sure. There, whatever. I mean, We had someone say Snoop Dogg on LinkedIn. I mean, like going to say Art Berman. Art? Yeah. Yeah. a lot of good He I don't think I understand half of what he's saying, [laughter] I'm being honest. But But I really like It's very interesting what he's talking about. very data-driven. Forward-looking stuff, yeah. Nice. Um, what else? What is What is um what is your advice to people getting into the kind of either technology, energy tech, or Let's go with just like the data space and stuff, whether that's an operator that's like, man, enough's enough. Like we need to go do this, or just an individual that you know, is looking to to get into that side. Start. That's one That's A lot of barriers to like just like not sure where and how. It's always It's not a straight road. Yeah. Just start. It's an education. It's It's experience. No, that's one of the most common answers, which I think is crazy because it's the simplicity of the answer is also the beauty. Like it It can be scary. It's terrifying. It can be terrifying. It's hard. You don't know where to start. You don't know who to talk to. You don't You know, Which cloud provider do I use? There's three of them. I don't You know, matter. Do it. Do it. Just do it. Yeah. Use people are passionate about which providers and tools and all that stuff. Well, it's like I know, for example, AWS is like the the most popular one. But like personally, I like Google's interface. I don't do anything on GCP other than I've got a BigQuery server, but you know, just like little things like that, right? Just Yeah. try it. You can't do that with other stuff. this era where things are rip and replace, I mean, spin it up, try it. Half time is a free trial. Yeah. Yeah. Build something with it. Spin it down. Yeah. You got 30 days. Yeah. Or, you know, for a year of Amazon RDS or whatever. Right. Um, let's do Last one. What was I going to Houston or Calgary? Ooh. Don't be wrong. Now. [laughter] Will you guys let me leave the room if I say Calgary? Yeah. I grew up there, so that's going to be a home for me, but Calgary in the summer, Houston in the a lot of time in Houston and Denver of of late, and it's ironic how many parallels there are. Yeah. Yeah. Where uh where can people find you guys, find you? Uh if I mean on LinkedIn, obviously website decodeanalytics.com. Mark with the C. Mark with the C. Yeah.

Our website, yeah. Those are kind of the platforms we don't tweet. [snorts and laughter] You should try it. Uh You'd be amazed at the uh the things that we can do from there. podcast on the value of tweeting. [laughter] It's probably not for us. Yeah. No, fair enough. Appreciate it. Thanks for coming on. Thanks for having me. Yeah, absolutely. This flew by. Yeah, it did. [laughter] Means it was good conversation. thank you. Thanks, guys.