EP 84: Ben Dickinson & Bryan Frantz from NexTier Completion Solutions

00:00:00:00 - 00:00:21:18
Unknown
We are back after a nice little summer break with, this episode of Energy Bytes. I am your co-host, John Callahan. Red dad here with with Bobby Nielsen. What's good? How are we doing back here? Got Bobby to come up to visit us in the big city today. So I appreciate you leave the country, make up for the chickens and goats before I left.

00:00:21:18 - 00:00:40:22
Unknown
Yeah, today we are. We are joined by some some, good friends over at next here, Ben Dickerson and Brian France. Appreciate you guys joining us. Ben, you're the VP of strategy and digital development. And Brian, you are the director of product development. Thank you guys for joining us today. Thanks for having us. Yeah. Excited to be here.

00:00:41:00 - 00:00:54:21
Unknown
Yeah. So we, you know, we were talking right before we got on here, but just, you know, we haven't been on I haven't done a episode and Pride Month month and a half, but now it's like just multiple transformational technologies that's been released in a week, two weeks. So that's five years in the tech world. Yeah, exactly.

00:00:54:22 - 00:01:16:05
Unknown
It's like you literally have to I've started doing that in meetings. It's like how you know, is it two weeks real time or is it two weeks quad time, which is actually like an hour? Yeah, we're almost a dog year. There's no one. Yes. No one knows how to like, gauge and actually be between now and then. Like fable was released, the government said it would be out and then we got a part time.

00:01:16:05 - 00:01:42:10
Unknown
And now if you have the max plan right, you you get it 50% of the time. Yeah, it's, 50% the time. It works every time, right? I mean, yeah, that's that's AI, right? But now, just as I mean, this is what I mean, especially you. This is what you do. You're literally doing AI stuff, and it's like you were saying earlier, you're you have to, like, take a step back and, like, just play around with things because you just you can't keep up with it constantly.

00:01:42:10 - 00:02:01:14
Unknown
You have to like, not only keep a pulse on what's going on, but just knowing about it doesn't really help you. You know, generally, because, you know, whatever it is, if it's a benchmark or if it's an article, it's going to be the best possible version. And the best part is just like, you know, anything, it's just like.

00:02:01:16 - 00:02:19:06
Unknown
And so you always take those with a grain of salt until you actually test them and use them. And so, you know, when all of this started 3 or 4 years ago, it it hit me. I was like, all right, this week I'm going to try and solve as many of my problems using GPT, whatever. Three. Yeah, as I can.

00:02:19:06 - 00:02:39:05
Unknown
And I was shocked by how many things I, it was able to do. And so I was like, okay. And that just rolled in. And now I use it every day. But it's the same thing with this stuff, right? Like harnesses are now a big deal. All these, you know, loop engineering is a big is a big thing today that will probably be something completely different in a month or two.

00:02:39:07 - 00:02:59:14
Unknown
But, you know, if you're not constantly touching and testing things, it's just, is so easy to fall behind. And so I feel like a dinosaur, but but then you have to step back and realize, like, all right. And then I go talk to my family, and it's like, I'm. I'm probably in the top, you know, 90th percentile, you know, of the users.

00:02:59:14 - 00:03:16:04
Unknown
But like the, the gap between me and, you know, the, the top users is just enormous. 100, you know, like on our on a daily basis. The only other person I know that actually uses AI in any kind of frequency is my wife for like cheer graphics. Yeah. You know, like it's nothing remotely close to what we're doing.

00:03:16:04 - 00:03:32:00
Unknown
But yeah, when you're in it, you're like, oh man, I'm so far behind. And then you take a step out, you're like, wait a minute, no one is remotely close. Even if you're using cloud code, the fact that you're in the terminal. But you've had a bunch of people. But I mean, what are you guys seeing like, as far as like, yeah, I can't imagine what it's like on other.

00:03:32:00 - 00:03:53:08
Unknown
And it is kind of a strategy. It just how quickly people who are so exposed get so desensitized as fast as we are. Yeah, that's what I personally, I feel like desensitized at this point. Like whenever I have, I have personal agents that run that literally just update me on news. Yeah. Just because I can't have the capacity to sit there and like, search and testing subs.

00:03:53:10 - 00:04:07:17
Unknown
Okay. It's not worth the risk. You, me, the brain dump for the last 72 hours, I have that run like almost every single morning. And then I just I'm like, oh, another one. Oh another one. Oh this thing now. Oh this thing now. Yeah. It's, it's, it's funny because it's like you said when you're in it, you don't recognize it.

00:04:07:19 - 00:04:24:02
Unknown
But I had a I have a friend who's in robotics and he sent me a picture of he was he was basically building, building an app for himself. And he was like, oh, you were right. You actually can build things. I'm like, yeah, no, no, you can stay in for like two years. Yeah. I'm glad you're glad you're catch it up.

00:04:24:02 - 00:04:44:14
Unknown
But yeah. No it was it's it's funny just how normal it feels now. I think we're the top users are now going from experimenting to like, normalization and it's execution now. Right? I mean, there is a big gap between top users and I would say everybody else who is really just trying to get comfortable summarize this email for me.

00:04:44:16 - 00:05:06:13
Unknown
Exactly. But it it is it is helping us get away from which model, which frontier model is the best and more concerned about how it's used and where it's used, which I do think is the right way. It's going. But it, I think it took people like us who are in this every day to like, get to the point of almost desensitization towards it to like, really kind of okay, move.

00:05:06:15 - 00:05:30:01
Unknown
We're moving. We're moving this thing along. Yeah. Well, that's my personal experience. I'm in. The more I use these, and especially now, I've been testing like the Army's desktop, and I'm almost exclusively using open source models on it as my, like, flip, flip to the other half of the brain. Don't use anything, you know, behind an API.

00:05:30:03 - 00:05:50:06
Unknown
And it's I'm very much so to the point that the, the models have are pretty damn close to each other. And the, the big differentiator these days is the harness and, you know, the external pieces of that, the skills, the tools, how you can figure out your system prompts, your agents. And so like there's all this other stuff.

00:05:50:06 - 00:06:03:16
Unknown
It's that messy middle layer of when you hit enter on the prompt and when you get your answer, there's all this shit happening that the average person has no clue, right? Right. It just seems like magic. Well, it's like, well, yeah, they've got all these integrations on the back of these. You've done a good job. Yeah, it seems like magic.

00:06:03:16 - 00:06:34:06
Unknown
It's simple. Right? Right. But that's like the nuance of this, right? It's like that is good software. If the user if it feels like magic and the user gets their outcome. But the flip side of that is it's like they've made it so general that when you come in to someone who isn't as familiar with it, that is the expectation out of the box on this like custom, super specialized, you know, downhole logging tool, you know, failure analysis method that they want to do.

00:06:34:06 - 00:06:51:21
Unknown
Right, like it's they're really good at general things, but they're not the greatest at these specialized things. And so there's this balance of like, hey, it's really good. But you have to understand how they got it to be that good, to understand how it can apply to other things. When you know your first shot, isn't it? It's an 80% answer.

00:06:51:21 - 00:07:11:09
Unknown
It's like, yes, there's a way to get it to 98%, but there's a lot of things you got to go through to get it there. It's but the models, more and more, I think, are becoming normalized to the point that like, it's going to boil down to speed, availability and cost. And privacy is obviously a big issue when you start talking to enterprise.

00:07:11:11 - 00:07:38:22
Unknown
It's wild though. I tested one last week and it was, distilled. Model F is the company's prism, and they've come up with these new, like, quantization techniques, and they were able to distill Quinn 3.6 27 B down to only needing like, I don't know, 16 gigs of, of GPU space. And it was a I think the whole model was five gigs to download, and I was just like, oh wow, shit.

00:07:39:00 - 00:07:55:00
Unknown
It's running on my laptop on my map. I can have some version of Quinn. Yeah, and it was murdering what I was getting. It was just knocking it out of the park and you're like, well, yeah, that's insane. This is on my computer. Like no one has control of this. I'm not reliant on an API. It's effectively free.

00:07:55:02 - 00:08:21:15
Unknown
And then you're just like man how this like I don't know the way that this scale is, is going to be fascinating. Like it's going to be incredible to watch because there's two very competing ideas. Right. It's like bigger more capable all the tools and all the skills and agents. But it's all private. Or you have open source where you know, you have the bigger models or frontier models, but it's seemingly people are really focused on how do we get this running on consumer grade hardware.

00:08:21:17 - 00:08:42:11
Unknown
And so like and then you see what Nvidia and Microsoft have done where like the next generation of surfaces are going to have sparks in them right now where you have a workstation literally in your. So then it's like, well then why would we run them? Why do you need these big models? Right. And so like there's open AI and cloud just explode into nothing because we've distributed this or is there something else?

00:08:42:11 - 00:08:59:19
Unknown
Who the hell knows. But it's going to be a fascinating ride or insanely scary how good those big models are. It's going to be something. But then, yeah, even every time they come out with a good one, you get a Kimmy case, you get the who is now benchmarked better than fable because it was trained by fable. Yeah.

00:08:59:22 - 00:09:20:13
Unknown
So it's like, yeah. Who knows I don't know. But then there's also the argument that at some point those open source Chinese models start becoming closed source because they're so good and they're just trying to eat away the competition. Who knows? But meanwhile, by Apple Voice attacks in September. Thank you Siri. The things that we deal with on a daily basis, it's pretty wild.

00:09:20:15 - 00:09:36:08
Unknown
True. I've seen some parallels to people saying that, like, you're kind of in the IBM to Linux transformation of like, you know, sort of your clouds and your, you know, your open eyes are in that space. And now there's all this open source stuff's going to come in and sort of eat their lunch and effectively make them irrelevant.

00:09:36:10 - 00:09:55:06
Unknown
Yeah, totally. IBM had a good yeah. What, 2030 this year or more rather like some of their open AI is going to be like a dinosaur, you know, five years into this. Yeah. Well, I mean, the thing that the average person also doesn't realize is that all the compute is being subsidized, like the true cost of that compute is way more than what you're currently paying for it.

00:09:55:06 - 00:10:16:12
Unknown
And so at some point you start making money. He's gotta pay you. Yeah, yeah. And so either either they come up with these quantities, you know, they're running these quantized versions of their own models that somehow or it goes open source, I don't know. It's going to be fascinating to see. Fun intro. Yeah that's a good yeah.

00:10:16:14 - 00:10:33:20
Unknown
Good topic. But there's no heavy handed topics right. Yeah. So, so, like, why don't you guys give us a little background on, you know, obviously just next year if people aren't aware. And then where were you? You guys fit in that and then what you're working on. Sure. I'll kind of start. I'll give a background to the company.

00:10:33:20 - 00:11:00:09
Unknown
So, next year is the sort of the completions wing of Patterson forgot. Right. So Patterson utai, primarily was a drilling services company, rigs, guidance software, directional tools, all the above. Right. In 2023, we merged together. So now we have both drilling and completions under one roof. Before that next year, was also sort of a company of M&A.

00:11:00:11 - 00:11:31:05
Unknown
So we have fracturing services, wireline services, sand logistics, chemicals. Basically, if you need a frack. Well, we have we have everything under the sun. So you need in order to in order to make it happen. Right. And throughout that journey, where sort of technology and software came into play is that, you know, we had inherently had built a lot of internal tools in order to help basically run our business better, be more effective things like predictive maintenance to take care of our equipment.

00:11:31:07 - 00:11:53:19
Unknown
You know, your general business systems and applying automation and things like that. And that sort of just exploded all the way from, you know, like 2018 up until basically 2021, 2022. Yeah. And then with, with all of the sort of the combined resources of the company, which included our own controls company called MDT, that's based out of Canada.

00:11:53:19 - 00:12:22:09
Unknown
Okay. We're able to Canada for that. It's like every control company. Yeah, like or 90% of them. Controls and gauges. Yeah. All Canadian. Hey, when you're good at something. Yeah. Just be good. Right. And, you know, combining those two things together is really what helped us launch the, the new software products that we're introducing into the into the, basically into the market, which is underneath, a new business line of digital solutions.

00:12:22:09 - 00:12:41:03
Unknown
Right. And that's primarily what Brian and I are associated with now is basically taking the internal tools that we've been developing over the last few years and basically making them available to our customers. And this is also something that we learned from the drilling side of the business. Drilling has been, I would say, more forward with data sharing, automation and optimization with our customers for decades.

00:12:41:03 - 00:13:04:17
Unknown
Right? I don't think that's anything new to anybody. So just sort of learning how they sort of went through this transition and replicating that and trying to improve upon the overall business experience for completion. So essentially what we put together is an end to end data management platform for completions. So, difference being it's built by the frac company, right, by the people who actually are dealing with this data and 24 over seven basis.

00:13:04:17 - 00:13:25:15
Unknown
So I'd argue that there's nobody better that knows more about this stuff than than us. Right. And so holistically, we're trying to basically bring our customers closer to the actual actions and workflows that we deal with on a daily basis through the same tools that we use to run our business. Right? Yeah. And that's, that was really Brian's brainchild that he helped kind of design.

00:13:25:15 - 00:13:30:06
Unknown
I, I put the mission, but I just made it happen. So.

00:13:30:08 - 00:13:59:20
Unknown
It's been a lot of fun putting it all together. It's we've it's been about a year and a half development journey at this point so far. So it's trying to incorporate everything that we've learned across all the different, you know, applications that we've used and bringing all the as much information as we can into one place and putting it in front of the same people across the same data stream was really the main goal, because we had a lot of like, just siloed data copy data, you know, transferred to this one, transfer to that one, and everyone looking at different sources.

00:13:59:20 - 00:14:17:13
Unknown
So bringing it all together. And one of the core things that we did was we built every single piece of it. Like from the moment it leaves the control system to the like, the it gets to the customer's endpoint, it's all the same data stream. So we don't copy things over it all just picks up off that same piece and that keeps everything aligned and, working really well.

00:14:17:13 - 00:14:38:15
Unknown
So we're we're really excited about it because it's we the, the field has had an input on it. You know, the engineers in the office of it important. You know, the executive teams have an input on it. Customers have a big input on it. Like we really try to bring everybody's opinion together. And you know, as much as you can try to make everybody happy and yeah, just go and make notes.

00:14:38:18 - 00:14:54:15
Unknown
But we were joking about you still never will be. Yeah, yeah, yeah. But I think we've gotten as close as you possibly can given all the different, places that it's pulled. Like we need to do this. Okay, well, then we could do that, and then we could also need to do this other thing that's completely different, you know, all the way over here.

00:14:54:15 - 00:15:17:10
Unknown
Yeah. You know, so I think we've done a really good job of finding a good middle ground that allows us to grow and make it even more capable over time. Yeah, I think that's one of the more exciting things I feel like about where we're currently at in the industry is, you know, for the last ten years, people like myself have continuously complained about how little of our data we're using or how siloed the data is, or all this stuff.

00:15:17:10 - 00:15:36:19
Unknown
And it's it's frustrating historically because it's like all the data was there, right? Like, and, you know, we're capturing it in most cases whether we want to or not. And the data is there and it's just there is there wasn't the time the personnel there wasn't the perfect combination of time, personnel and technology to truly capture the value of it.

00:15:36:19 - 00:16:10:16
Unknown
And we're moving into that. I feel like time of, hey, we are now plugging these systems together. We're leveraging AI, we're leveraging whatever data lakes, etc. to start making truly, you know, truly informed, as informed as I can possibly be, decisions based off, you know, all the actionable data that exists. And it's, you know, it's funny to talk about that, you know, transforming data into actionable insights, which is, you know, exactly what I wanted to do straight out of high school, but, it's it's there a have something like that.

00:16:10:17 - 00:16:32:09
Unknown
Yeah. Yeah. Right. And I'm sure that will be on my Father's Day list next year because, I sent it to my wife already. But it's one of those things where it's like, yeah, as an engineer, data is like the easiest way to solve a problem or should be the best way to solve a problem. And we've had it for so long, and it's frustrating when you just see across the board, not just service companies, not just operators.

00:16:32:11 - 00:16:52:14
Unknown
Everybody had this huge data problem. And we're now finally, you know, we've gone through the hard work over the last five, ten years of like, hey, we recognize that there's all this data and it's all siloed and it's junk, and we need to clean it up. And we need all these back end processes. And now we're finally on the the downhill slope of that of where we can actually capitalize on the value of, of that actual data.

00:16:52:14 - 00:17:07:13
Unknown
And so I'm super excited. Like there's just so much opportunity in the industry. I know you guys are well aware. Yeah. I mean, the I can tell you the sentiment that I'm hearing now is starting to just change a little bit. It was it was, a consistent feedback of, I don't know what to do with this.

00:17:07:15 - 00:17:29:18
Unknown
Now people are like, oh, I actually have figured this out. Now I just need something that's reliable and consistent and just let me go. Yeah. And that's what we're hearing from our customers. And it's actually a great place to be. Because that was honestly our first mission. It was even though we've had this big push over the last 5 to 10 years of cleaning up data or just digitizing things that previously weren't sensor eyes, that sort of thing.

00:17:29:18 - 00:17:46:12
Unknown
Right? We've we made a big leap, but then we made a big mess all the time. And we've had people come in and try to figure it out, try to figure it out. And like when, when Brian and I sat down, we're like, okay, what do we really want to do with this? When we when we go and like, if we don't have reliable and consistent data coming from location, we have failed.

00:17:46:12 - 00:18:03:09
Unknown
Yeah. Everything that is I mean, it has to start this one, right? And I know it's I know it's been overstated, but it's like we still hear the customer's fundamental problems. Yeah. Right. So I think the first like 6 to 9 months was just that honestly it was just that because then everything else on top of that was just building off APIs.

00:18:03:09 - 00:18:18:09
Unknown
You're just you're just then you can just go, right. But the key thing was giving our customers the same, the same type of experience. Right? The same thing that we develop off of. We wanted to give them the same access in order to do the same thing kind of coin. The closer you are to the equipment, the closer you are to the truth.

00:18:18:09 - 00:18:32:13
Unknown
Okay, well, you're not getting any closer than this because this is what we work with. So here you go. Right. And it was our job to make sure it was consistent. Reliable. Right. And I mean I'm pretty proud of what the cloud team is, me able to put together and what we're what we're able to push out there honestly.

00:18:32:15 - 00:18:49:18
Unknown
But it's just it's crazy that we're still dealing with some of the same problems, but we're iterating on it. But the good thing is, is I think the people on the other side of that data are now getting to the point where, like, all right, I'm ready to run. Just yeah, yeah, connect and let me go. Oh, and even like across the industry, it's becoming a standard.

00:18:49:21 - 00:19:08:13
Unknown
Not not the data type or the format or anything like that. But you don't don't go down that rabbit just. Yeah. Just the fact that you can monitor, measure something. Right. Like because even when Bobby and I were together, right, like we'd go out and, you know, one of our new products was, you know, we were doing stuff monitoring frac fan data, where we were monitoring work over rig data.

00:19:08:13 - 00:19:31:22
Unknown
We were just monitoring data from sources that had some kind of digital output. But it wasn't going anywhere. Right. And so being able to unify that data immediately provided additional incremental value and stuff. But, you know, historically speaking, though, you might be the problem with office is that it's never it's normally not just one company providing all the services on a site.

00:19:31:22 - 00:19:59:10
Unknown
Right. And even you guys. Right. It might be like water transfer or sand or whoever. But so now you have another company who has a different format, who has a different standard. They report at different frequency, like all this stuff. And so then everything breaks because. Right, you can't capture that one thing. But now again, if it's just instrumented, which I feel like most things are in the field now, thankfully, it doesn't matter what the structure is, you can transform it, you can map it.

00:19:59:10 - 00:20:17:23
Unknown
However, as long as the data is there. But for the longest time it was always, well, if only we had something monitoring this, then we would have a fuller picture, right? Like you can't just look at a frack data set without the wireline data set and have a full picture of what's actually you can make a lot of inferences, but you know, you're still missing pieces.

00:20:17:23 - 00:20:39:17
Unknown
And so I think that's one of the the more exciting parts just holistically is it's becoming an expectation that everything is outfitted, everything is censored. And you should be able to monitor that data. Right. It's not a question of can you connect how fast right. What's it look like on the other side? Yeah, right. Which I do think is a healthy way to for us to continue to, to go with as an industry.

00:20:39:17 - 00:20:59:07
Unknown
Honest. Yeah. So I've got a lot of things going on up here. And I don't want to make sure I don't totally word vomit, you know, a million things at once. But, so first things first, like this product is exclusively for if they're using your, you know, services, your field services. Correct. Like you can't connect into if they were using a different provider.

00:20:59:09 - 00:21:18:15
Unknown
This is for this is for you all or you can't iOS sit on top of. So yeah, it's built to be agnostic. Yeah. It is built to be. Okay. We, you know, we built it with the original intention of having it be provided with our frack fleets. It's a it's a plug and play thing. Yeah. As soon as we arrive here you go through it again.

00:21:18:15 - 00:21:38:08
Unknown
That's part of the, the, however, we do recognize the fact that our, our customers don't only use our equipment all the time. To your previous point, right. So the, the architecture is agnostic, and that is the intention is that it can it can survive without, pieces. So it is like a totally separate ish product that, you know, can stand on its own.

00:21:38:08 - 00:21:54:15
Unknown
Yeah, exactly. Because then you're just getting into data transformation and collection and you go from there right now, the, the part of how it interacts with our control system, that is the thing that stays obviously with next. Yeah. Right. That gets into vertex, which is our automated controls. That's, you know, how we how we run our fleets and things like that.

00:21:54:15 - 00:22:12:18
Unknown
So that part obviously stays with our control systems as of next year. However, the software, how we move data, how we understand it, how we work with it that is actually agnostic to that okay, control system itself. Cool. So yeah. Yeah, I mean, I want to like anchor on that if it was purely either way, but then it kind of funnels into the next part.

00:22:12:18 - 00:22:32:18
Unknown
But so I think it's interesting because you're creating a digital product that's customer facing, but it sounds like you all did the work kind of upstream of that, like for the business. And I think especially where you have like the hardware and stuff, like so many companies, they need to be using data internally to do their work better, work on the business or in the business.

00:22:32:18 - 00:22:52:21
Unknown
Right. But then it's like, how do you prioritize that work there? Then also creating like the, the customer facing thing or, you know, who gets the most love? I even think about it like, say, Rob Stover when he was at and various he was working on like the the data side but for like them helping understand their business, not the whole data data product.

00:22:52:23 - 00:23:06:07
Unknown
You know side was an intern. It's almost like a, this kind of meta kind of thing where it's like, all right, I'm doing data work, you know, for this company. But then they also have a data product and like, do you share resources even, like, or are they two totally separate sides of the house. You know. Yeah.

00:23:06:07 - 00:23:24:20
Unknown
As far as like talent or, you know, people in the business. That's a good question. Honestly, I don't think everybody's ever asked me that question before. So we're here for the hard hitting questions. I mean, I think the way the team is structured, we we kind of have a team that specializes in what, I'll say, our internal workflows, things that just make the business run.

00:23:25:02 - 00:23:53:00
Unknown
I think that's incredibly important to have somebody who understands is embedded within the business itself. Things like supply chain forecasting. Right. They understand every touchpoint. They understand every, every user's pain point, and they sort of specialize in that. Once you get into the outwardly facing product lines, you're talking frack, you're talking wire lines, cement. It's funny from from my view, I get to experience the feedback from everybody.

00:23:53:00 - 00:24:12:02
Unknown
So we're good in the bad from both our business users internally and our customers. There's so much overlap. Yeah, end to end of the actual problems. It's crazy. And sometimes I don't think either side really knows because I mean, but this is a human thing, right? Everybody's number one priority is what's in front of their, you know, their faces on that day.

00:24:12:02 - 00:24:34:12
Unknown
Right? But take frack, for example. A lot of the same issues with data management are internal teams, like our reservoir engineering teams. We're having with the data that was being produced. It's the same issues that a customer or a customer reservoir engineer is having. So why not build a tool that solves the problem for both? Sure. And then you're just managing access and authorization.

00:24:34:14 - 00:25:05:01
Unknown
Yeah, it's really the same thing. So I think the fun thing has been working with our solutions engineering team, who's responsible for designing all of this and understanding that you guys are actually are saying the same things. You're just saying it from different perspectives. So from from that side, it's actually been quite easy. Yeah. Honestly, I was going say that that's one of the biggest benefits of building software for yourself is that typically if it's making your life easier, it will also make the client's life easier, right?

00:25:05:01 - 00:25:23:05
Unknown
Like it's doing a big post job. Analysis project for a client. And it's like, okay, I've looked at a million post job reports and it's like, okay, when I click on the cluster that I'm trying to extract, take me to the page where it extracted it from so I can validate that it's there. Right. And it doesn't matter if you're a consultant or like manager.

00:25:23:05 - 00:25:45:11
Unknown
Guess what? It does the same work, right? Yeah, right. If you're familiar with the workflow and you're building it for yourself, you understand like, oh, at this point I should go and check this. And it'd be really convenient if that was just right there. When I got to that spot. And so it's, I see that all the time and just our stuff, you know, we're building stuff all the time, but it's it's it just makes it that much.

00:25:45:12 - 00:26:07:20
Unknown
You're to production and and to happy client feelings faster because the people who are using it are also building it alongside. Right. And it's also very helpful. This goes back into the field feedback. I mean, it's really easy to get good feedback from field users. Whenever you have a mechanic who is using, let's say, our mobile app to enter in, what's going on with the bottom in the company.

00:26:07:20 - 00:26:24:12
Unknown
Man is sitting right there going, what are you doing? Oh, it'd be nice if I saw that, right? Right. I mean, it literally can be that easy, right? We just need to give them the voice. Right? Which is, you know what? We've been trying to do with this whole internal transformation that we've been going through, basically taking ownership of the of the full tech stack that we have inside the company.

00:26:24:15 - 00:26:44:10
Unknown
Yeah, I think for the other, the other good thing too, and we were talking about open source a little bit ago for the internal workflows. There's so much stuff out there that's already been done. Like we don't need to be experts in, in, in logistics or inventory management or stuff like that. There's so many open source libraries out there that we can pull from, customize and then deploy within our own applications.

00:26:44:10 - 00:27:16:08
Unknown
It just makes it so much easier these days. I don't know, oil field didn't invent 5 or 6 Sigma, but we did adopt it, in 2016 or whenever that article came out, it made the front page. It still never ceases to amaze me. But no, I completely agree with you. That's historically been one of my gripes within the industry, is that we think like we're really good at what we do, but then for whatever reason, we start thinking that that means we're the only ones that are good at that thing.

00:27:16:08 - 00:27:38:12
Unknown
And it's like logistics or inventory inventory management. Let's just let's just play with it. A solved problem that is literally that is the vein of a frack company's existence. Is fucking inventory like that. You all, as a field engineer, that was 95% of my job, was managing and making sure we had everything to pump a job. And so it's like, but yeah, you're not starting from zero.

00:27:38:12 - 00:27:59:13
Unknown
You don't have to reinvent the wheel. We've been doing like all these other industries have been perfecting this for decades. It's like, I like drawing on the auto industry. Yeah, I like drawing on. I like drawing on the medical industry because the the way that they handle inventory is pretty critical, right? If you don't get critical medical supplies to a patient, that's a problem.

00:27:59:14 - 00:28:15:16
Unknown
Right? Right. There's a high risk high reward scenario, very similar time critical. We've we've we've hired a lot of people, over the last, you know, six months. And, I'm honestly proud to say they all come from a such a very different variety of backgrounds because I didn't I didn't really want to set out to hire people who are.

00:28:15:16 - 00:28:32:00
Unknown
I've only ever worked in one guy. So I wanted I wanted a variety, different experiences, because I know we're not the best at this. Yeah, right. I mean, we we had a recent hire came for the broadcasting industry, and it was funny. You want to tell a story about. I'd like the light bulb that went off as we here as we were talking to him for the first time.

00:28:32:03 - 00:28:50:14
Unknown
Yeah. So, yeah, this was on our, on our DevOps side. So the, the guy was still kind of just taking us through his history, and he talked about is the time he spent at a broadcast company where they basically take these vans, drive them to a stadium, drive them wherever you plug them in. They have a whole they run the operation, unplug them, drive them to the next thing, like everything back in.

00:28:50:14 - 00:29:07:22
Unknown
And here's the challenges he had with edge deploying and making sure that we're not updating a game or like, oh, that's exactly what we need to do here, you know? Holy crap, this is the same thing. Yes, you would have relevant experience for updating fragments, you know, like it's just not ever about that. Yeah, I spoke to the sense there are a lot of weird parallels out there.

00:29:08:02 - 00:29:24:12
Unknown
We've found a few people that like like, oh, yeah, no, we did something similar and I can help, you know, solution that for you guys, it's just like, we're not the only people that have these problems. They have that conversation a lot. I was like, we don't need to like inventory. Logistics, like UPS probably has more logistical problems than us.

00:29:24:12 - 00:29:42:23
Unknown
And, you know, there's research out there that can help us out, you know, for, you know, like, yeah, there's there's a lot of I mean, that's yeah, that's that's how it should be. Right. Like, I, I think people get so hung up on like the nuances of the industry that they, they seem to think that the business problem itself is you.

00:29:43:01 - 00:30:00:05
Unknown
It makes it unique because of the constraints. But it's just like it's just like physics, right? You have your givens, your knowns and your unknown, and you have the constraint of whatever world, the vacuum of physics that you're living in. And that's it. Right. And so it's like it's the same thing. You know, the business constraints might be a little bit different.

00:30:00:05 - 00:30:25:23
Unknown
You might be in the middle nowhere instead of the middle of a city, or you might not have reliable internet or whatever, but the problem you're trying to solve, generally speaking, is pretty replicable across multiple industries. And so it's nice, it's refreshing just to see people, thinking that way, because that is just, again, one of those. I went I came from manufacturing directly into field engineer at Cudd in the Fayetteville.

00:30:26:00 - 00:30:45:23
Unknown
And like, I remember the first month, sitting in the van, you know, I'm still training. I'm with my, my p and, I look out the window, there's like ten sand trucks just sitting there. It's like, what am I like, what are they doing? He's like, oh yeah. Well, you know, you know, when we screened out two days ago or yesterday or whatever, it's like no one canceled those.

00:30:45:23 - 00:31:04:20
Unknown
And so now we're paying to merge on all of those trucks for the next two days or what? However long or nothing worse for a driver than sitting there. Not. Yeah. Yeah. Well, and it's like, so I came from manufacturing where I literally had a stopwatch and I was measuring every single step of the process to the minute, you know, like that it was that important.

00:31:04:20 - 00:31:18:10
Unknown
And then. Yeah, just in time, all the tools are organized or where they need to be. I say, you brought up Six Sigma. You should have just got into action right. And so for me, it was just a lowly field engineer. I wasn't going to be doing it. You know? I didn't know what the hell a frack job was at that point.

00:31:18:10 - 00:31:41:13
Unknown
And so it, But yeah, it's just like we, we think the constraints are the are what make it unique, which they are, but they're not the root of the problem. Right. So when I was similar vein, I think we touched on a little bit. But like the side about like the tech stack because again it's I'm like, you guys are doing all this internal work before, you know, creating this product for the customers.

00:31:41:13 - 00:32:19:22
Unknown
So are, are you all utilizing the same tech stack internally and externally? And if so, like obviously you start to say 2018 whatever. Like the, you know, different tools like again snowflake, Databricks, these were kind of maybe in their infancy if you were even using this at all. And now there's been advances. So like, have you had a chance to start almost kind of greenfield for the product, the customer facing product or, you know, how have you guys thought about that or what do you start with, where you, I also want to know, how do you feel that that impacted how it's architected today versus what most people do historically, which is how

00:32:19:22 - 00:32:47:23
Unknown
do we just fix all the holes in the boat instead of redesigning the boat from scratch using modern day tools and just be good with it? That's good. So, I'll, I'll answer historically and I'll let Brian talk about from this point forward, because, we did go through a significant transition. Yeah. Since when we started, let's say, the point in time when we started our digital transformation was like 2017, 2018, we set up our first data warehouse.

00:32:47:23 - 00:33:04:14
Unknown
We we set internally or cloud internally. Okay. I mean, it wasn't or was it. But it was like, no, it was on prem. It was with GCP. Okay. So we've always been traditionally a Google house on the next tier side. So, we we got you know, we had some great people that we worked with that helped us stand up.

00:33:04:14 - 00:33:22:12
Unknown
We really didn't have any resources internally. I mean, this is literally ground zero. Right. We were and the original idea we had two things that we were working on. Number one, we wanted to digitize our time logs because we were still dealing with Excel sheets being emailed to somebody that basically put them into a nice SharePoint that got connected to Tableau.

00:33:22:12 - 00:33:40:09
Unknown
Yeah, magic. So we wanted to get rid of that. And then we also wanted to start, databases, our work orders for maintenance and then pairing that with our real time sensor data. And then that was actually the start of our whole preventative maintenance program because just that initial step. So that was the first version of our data warehouse.

00:33:40:11 - 00:33:59:13
Unknown
At that time, all the architecture was built to only ever be used internally because I was we were right. You know, this was big query hindsight. Yeah, exactly. Hindsight being 2020, we just never thought anybody else outside the company would care. I was going to say the timing of that also matters, right? Oil field in that era, no one was sharing anything.

00:33:59:13 - 00:34:21:18
Unknown
No one had APIs. No one wanted anything to do with anybody else. Yeah. So any any sense of like external user security or that was. Yeah. It wasn't even a, it wasn't even a consideration right now. Yeah. So now you fast forward till now and I'd say the it's pretty much now. Okay. It always must be thought of from that perspective.

00:34:21:18 - 00:34:40:00
Unknown
Right. So not even just the tools that we were using, but even how we were building things and like how we were referencing Right Well's API numbers, like originally, all of our architecture was based off of our quote to cache system, right? We wanted to tie everything back to a job that made money, which makes sense if you're only looking internally.

00:34:40:01 - 00:34:57:12
Unknown
But that's not what our customers care about. Yeah, they care about the well, how are we connected to the well? How are the services combined at all? Well, so we've had to completely revamp our architecture that is more centered around the customer viewpoint, but it's actually worked for us internally because it's allowed us to connect more service science back to a singular well.

00:34:57:17 - 00:35:17:16
Unknown
So it's like little things like that that we've noticed through this transition have been huge. So I'll, I'll let Brian kind of take it from there because pretty much this point forward, it's all hands. Yeah. I mean, so we originally set out to like remove the need for any third party software to transmit and report on and basically generate the end deliverables for a job stage.

00:35:17:22 - 00:35:34:06
Unknown
Well, right. So that was the initial push for the for the platform. So we had to was that all the way down to like in the van from the from the software and all the way down from the control system to the customer? That was the gap. We were, you know, trying to fill. And it did ultimately do.

00:35:34:06 - 00:35:52:00
Unknown
But the so that came with a lot of challenges, as you might guess. Imagine. Yeah. So like, I mentioned earlier, the first step was just sort of making sure that we had a good foundation of streaming data time series data, because that's what everything else is built on top of, right? Can you briefly just hit on some of the the control system problems that you guys ran into?

00:35:52:00 - 00:36:09:07
Unknown
Just because I feel like that is such a it is like it's an issue and it isn't, but they're basically everything in the field has a control system, generally speaking. And I feel like there's a lot of, people have a lot of interest in that, but they're very hesitant or scared because it looks like this. It is.

00:36:09:07 - 00:36:34:14
Unknown
It's a big, scary, messy, right? I'll say honestly, we didn't run into that many control systems issues. The MD controls are very solid when it comes to the actual operation of the equipment, and they're very, robust in that sense. Where we ran into the most issues is mapping that equipment data into a job semantic layer, right. Where you have it's not just a bunch of pumps and a blender following a schedule.

00:36:34:14 - 00:36:52:10
Unknown
It's taking certain sensors off of the pumps, certain sensors off of the iron, certain sensors off of the blender. And combine that into and making, you know, what we call our job data schema internally. And so that you, no matter what kind of job it is in, it was intended to be well focused, kind of going back to make it a little bit.

00:36:52:12 - 00:37:11:08
Unknown
That's something that they figured out before us. Is that like if you focus it around the well, it makes everything a lot easier. Instead of focusing around a pad or a job. So kind of transformed it to be well focused and what so effectively setting up the configuration to know what fluid composition is going down every one second slug of a well was the challenge, right?

00:37:11:08 - 00:37:34:00
Unknown
So in accounting for you know, zipper frack similar for tracks twin frack all you know all the fun little combinations of building all the Frack Pro. Yeah. Yeah. It was used to. Yeah. So that was that was the major challenge in the field was just getting all of that configured so that we got a relatively consistent data stream from job to job to job that didn't have to be rebuilt every single time if it was a twin or, right.

00:37:34:02 - 00:37:54:18
Unknown
You know, zipper, once we kind of sorted that out, building everything else is, you know, the biggest challenge is probably the chart, right? So just making sure that that's very snappy, up to date that, you know, and where we're looking right now, we're at like a sub two second delay from the control system to the chart. And it can be better than that depending on like the network.

00:37:54:21 - 00:38:07:17
Unknown
Yeah. Yeah. Yeah. So you're you're bounced you know, like I always like to say like we want real time. It's okay. We'll define real time. Right. Very important. I mean, I can preach that in later on. I mean how some people real time is like, oh can I see this chart every week? Yeah. Yeah, I got that. Yeah.

00:38:07:17 - 00:38:25:00
Unknown
It's totally different, you know domain. It's all over the place. It's it's a definitely relative term. You know, you're bumping up against physics eventually. Like it just can't move faster or like. No. My personal favorite is, you know, trying to visualize one second data, right? Like, they want to see all the data. Yeah. Pixel. You sure you want to see all the data?

00:38:25:00 - 00:38:44:19
Unknown
Because there's only so many pixels on this page that we can display data in. Yeah, yeah, yeah, we face that one to it to you right now. Bobby and I are right there with you on. Yeah. Yeah. Okay. But, what's we. So we got the turning and all that sorted out, and then we built sort of the back office to support all that.

00:38:44:19 - 00:39:01:11
Unknown
So where we can throw all the well, static information, treatment schedules, all the stuff that our engineers handle on a daily basis and kind of like, I don't know if this what you really do or before with external versus internal, but like we the platform everyone's working on the same system. Right. So the chart that the engineers uses the same one that our customers see.

00:39:01:12 - 00:39:18:01
Unknown
Okay. And that was, that was like 100% on purpose because and we've even found where like feedback is like a customer class or something. It's like, oh, that actually helps the engineers and vice versa. Right. So we get and you just do that feature development once and on both sides get right. Right. So and it's much better for both you know for both sides.

00:39:18:01 - 00:39:40:08
Unknown
So and then I would say the last like major challenge on that was just the ethics side of everything. Right. So the authentication is fairly straightforward. We just use a you know off zero kind of off the shelf. And then the because it's sort of this multi-tenancy, third party. There's a lot of, I don't know, it's like a matrix of permissions you have to think about.

00:39:40:08 - 00:40:00:00
Unknown
Right. Because you might you know, you have customers who can be tenants and tenants who can be customers, who can also be vendors, depending on what job perspective you're looking at it from. So we put a lot of work into how that's divided up and how it manages, like what you can see, both from a like a feature perspective in the UI and then all the way down to the data.

00:40:00:00 - 00:40:22:15
Unknown
So the, the authorization service that we went with actually, integrates with. So we were fast API SQLAlchemy. Yeah. For most of our APIs, it actually integrates all the way down to the SQL alchemy level at the for the, and it reads in the policies and can actually write your base filter for that particular request based on the end user that's requesting it.

00:40:22:17 - 00:40:41:06
Unknown
So you can like put like, this is baked into like basically every query that a customer makes into the system, it automatically puts like where customer ID equals blank into every single base query, and then everything's built on top of, okay. So our developers don't have to worry as much about getting all of the authorization. Correct. Right.

00:40:41:06 - 00:40:59:05
Unknown
Because it's just baked into the policy. And if we want to change it somewhat on the fly, you can just update the policy. And then that enables something or changes the way that that respond. So that's been a huge accelerator for us just not having to like, be like I say, not that we're not worried about it, but you know, like we're not there to be like an admin.

00:40:59:06 - 00:41:18:20
Unknown
Yeah. Like that's particular. Make sure everything no one change the base query of the base filter. Everything everywhere. Right. Yeah. So that's been a huge help. But that's that, you know, those are the I would say the major challenges given going through it. And then, you know, all of the when it was finally put into the hands, everybody all the edge cases that you just don't think about of like, well, when it's this and then, oh, we're doing a side.

00:41:18:20 - 00:41:33:06
Unknown
Well then a zipper, then a zipper, then a side cycle, then like switching back and forth and that'll be account for that. And yeah. Yeah, I mean it would not be tenable or a lot of creative job configurations that we ran into as we were rolling everything out. And then I guess the final piece was the deliverables.

00:41:33:06 - 00:41:51:13
Unknown
And, you know, obviously there's a lot of customization that goes into that, but we've built a pretty robust system that generates like an end of stage data set, and then we build everything on top of that using templated, reports. So they're very easy to change. We've built into the back office the ability for our engineers to change them on the fly.

00:41:51:13 - 00:42:09:00
Unknown
So we're, we're able to make corrections or add fields, to reports for our customers very, very quickly. You know, could be. Oh, it didn't show up on this stage. It shows up on the next stage where historically, we're having to deal with our third party vendors. Right? It could be days. Yeah. You know, you never really know because you're on their schedule for development.

00:42:09:00 - 00:42:28:10
Unknown
That is the worst feeling in the entire world. Yeah. Like, oh, that channel wasn't on that stage and now it is. And no one knows what it's meant to. Yeah. Yeah. So and it reads in like fairly readable English the way that the data structure is. So and that was intentional so that our engineers can go in and kind of like discern, oh, that's treatment pressure average right okay.

00:42:28:12 - 00:43:02:12
Unknown
No one. Yeah. Yeah. Right. Yeah. All very you know spelled out you know what it is. Yeah. That's that was going to be the question I asked after after you breaking that down for us was one of the most infuriating problems of Brac data. Is that the the guys in the field ultimately still have control over the naming conventions that they use for the channel in the mapping, and that causes so many problems, even though all of our software's have this idea of a template where you can just import, yeah, template and that that's all it is.

00:43:02:12 - 00:43:17:01
Unknown
And it's always the same thing. And it doesn't change between day crew and night crew or you know, what day of the week it is well known as a developer. If it's a consistent format where underscores are in the same spot or this that you can. Yeah, you do have some kind of structure to it, but I think it's more the two.

00:43:17:01 - 00:43:35:16
Unknown
I mean, I know, it was it, GitLab I think that published out like a, like a, naming convention. You know, how they name things, but like, they, they were very big on, you know, X plus X plus is better than implicit, you know, 100 and again. But I think that some of that it's a very seemingly small thing.

00:43:35:16 - 00:43:52:01
Unknown
But like what where technology has gotten to like you can have a 15 character column header naming like but used to be like, I think one of the guys I used to work with was an Oracle developer back in the day, you know, from like 20 or 30 years ago and like, you could only you only had so many characters you could put in, like I really had to hammer on there.

00:43:52:01 - 00:44:11:19
Unknown
Like, no, you can use longer like I'm like, yeah, you see what it is? Yeah, yeah. But that way when someone steps into it, they know exactly what they're looking at. Yeah, there's no reason not to. But now, I mean, that that historically speaking, that is the biggest issue with frac data. We didn't we don't have well, annoyingly, we basically have an editor.

00:44:11:20 - 00:44:30:11
Unknown
We just never came up with a normalized structure that everybody agreed on for that. And so you know, that's that might be something good to get in because you hit it on at the very beginning. Talking about Patterson being more on the drilling side, how drilling was had an effect on it some. But it you also were kind of cautious to get in on the like a standard conversation.

00:44:30:11 - 00:44:51:19
Unknown
And I mean, ever since we were at Ards, they were talking about completions with some mole and I'm sure it's gotten to some point. But but again, whatever you think about what's the mole on the drilling side? It is a standard and people have been able to develop against it. And then you kind of almost get similar to the banking APIs, where now you get like a, you know, stripe or clover or all these things that you can build these FinOps things off of there.

00:44:51:23 - 00:45:09:00
Unknown
People have been able to build, you know, your core Vas and different folks build off of a, off of a standard. Whereas like completions has been the Wild West and it's been channels and stuff flying all over the place. So, you know, I guess what have you seen on that side? You know, or where were you at on a, on a standard?

00:45:09:00 - 00:45:23:05
Unknown
I mean, I know you you've obviously created internal standard. Right. We've built a width one for and a width to one. I'd say we didn't want to but we did it. Yeah. It was just by requests. I think it was a good exercise to do, honestly, just to see if we could fit into a drilling style standard.

00:45:23:06 - 00:45:42:19
Unknown
Yeah. I think what we found was, is the thought process and methodology is probably where we all need to head. Yeah. Exact technology that we do to do that doesn't necessarily need to be the same thing that they've always done. Sure. I think we've been able to do more with kind of almost customized Json packages than anybody else's stuff.

00:45:42:20 - 00:46:03:22
Unknown
That's the thing, right? At the end of the day, as a data person, I don't care what the format is, as long as it's structured in the same way and it's repeatable, especially when I basically I can have an agent do the entirety of my writing process. Right. That's the thing I like to do with myself. Now, if it's in Json, I guess what I can make a CSV like, I can make that into anything.

00:46:04:03 - 00:46:22:19
Unknown
It doesn't matter, right? What matters is that there's a structure and a naming convention and normalization that sticks to it. Is there every single time. And I think depending on the I would say how far along the service line in general is within the industry. That's why you're seeing frak probably get there first. I think waterline is going to follow right after.

00:46:22:19 - 00:46:31:20
Unknown
I mean, you're seeing different service lines at different maturities, which is perfectly normal. But I do see us going in that direction at some point. Yeah.

00:46:31:22 - 00:46:52:01
Unknown
I started to a consistency and reliability game after that. Yeah. Well that's the thing. Like the industry doesn't have to come together and come up with a new format for, you know, frack data for the next five years until the next one comes. Like the format itself does not matter anymore. The structure and the naming conventions, however, would be incredible if everybody was just like, yeah, this makes sense.

00:46:52:03 - 00:47:08:13
Unknown
Let's try and do this right. Like who cares? The naming the names could be different, but the structure and the order and all of that, like it would, it would change everything. It would change so much about the fact that it would. So let's move a little, maybe a little more downstream in the app. So we talked a lot, a lot about the back end stuff.

00:47:08:13 - 00:47:26:06
Unknown
Now. So I the two sides like so you had customers and some of them are going to want to use your front end interface, and then other ones want access to the data. So like versus like outlook. How can people access the data coming out it like is there a public API or I mean it sounds like using GCP, I'm I'm not as familiar with what the options are there.

00:47:26:06 - 00:47:45:10
Unknown
And a lot of people say use snowflake. Now to, you know, share data, you know, kind of pseudo instantaneously with, customers. So like what's available. Just if someone said I just want the data, does it give me the data? But then also curious on like the, the web interface side or whatever interface there is and like how opinionated is it?

00:47:45:10 - 00:48:02:21
Unknown
Or like, you know, kind of what's available to people. Yeah, it's a good question. So from the from the data side, it's, it's a product that we call Fleet Stream. Essentially it's a, it's a sub, pub sub subscription that we can set a customer up with. Okay. And it basically subscribes to the direct feed that we used to build all of our web applications.

00:48:02:21 - 00:48:20:21
Unknown
Okay. So like the iOS web applications, are our developers building that, that subscription in order to visualize and manage data that I mean, it's honestly it's as simple as that. Okay. There's a lot of work that goes into it. We just talk about simple design. Easy, right. But like but but simple does mean good software. Yes. To your, previous point.

00:48:20:21 - 00:48:40:17
Unknown
So that's what's available today. It's available in a variety different protocols. Per second via data warehouse. We built an empty protocol as well. So all of that is available today. And that's something that a lot of customers are actually subscribing to. And they're using that to build their own internal web applications. Now, we're streaming that into a snowflake database.

00:48:40:17 - 00:48:58:21
Unknown
And there are engineers who are literally vibe coding their own apps using our real time data that's coming into their system. Can can you all talk to what tech you're using for that? Or, you know, Pub sub is literally the name of the product, right? Yeah, it's a Google Pub sub. Yeah yeah yeah. Okay. Like that is the GCP.

00:48:58:21 - 00:49:19:14
Unknown
We're using that to kind of as a, as the real time backbone to the, to the whole system at the moment that like a Kafka kind of alternative, it's similar. Yeah. It's in that vein. You can go back in time and like Kafka, but it's probably closer to RabbitMQ or like just an mQTT style protocol. But it does have like history and stuff to it.

00:49:19:14 - 00:49:35:14
Unknown
It's fairly, you know, I mean they built all their stuff on it. So, it's fairly it's fairly solid product across the board. And there's a ton of, SDK language support for it, which is one of the reasons we've sort of stuck with it is there's just if we can say, oh, you want to use it, here's a ton of documentation from Google.

00:49:35:14 - 00:49:50:06
Unknown
It, it tells you exactly what you need to provide that endpoint to like. So there it your developer and like yeah we provide a service account that has access to that that topic. And they can go and build and do whatever they want. You know. And the secret sauce is really like how we enrich it along that. Yeah.

00:49:50:08 - 00:50:05:22
Unknown
Transformation path. Yeah. So it's not like you're getting raw sensor data because it's doing it right. Yeah. Well there's an enrichment processor's a cleansing process that we go through, which is the same thing that we do for our own internal users. Yeah. Right. So going back to you're literally working with the same data that we work with on a daily basis.

00:50:05:22 - 00:50:27:05
Unknown
Yeah. Right. So that's what that is. Did you say did you offer like BigQuery access as well or not. Yes I would say that's not nearly as popular as just give me the per second data please and then go from there. But yeah. So there you go. Yeah, exactly. I do see that evolving over time as people are using more of the, the web applications that exist within iOS itself.

00:50:27:05 - 00:50:44:10
Unknown
Yeah, because there's always the curious engineer that wants to just kind of dive query and then just do some analysis on their own. I mean, we have engineers who do that today literally within BigQuery as well. We use it for today. So I do see a a customer facing front end. Yeah, eventually. How would that work? I mean, do you all charge for that service more?

00:50:44:12 - 00:50:58:21
Unknown
I was thinking about that like people go and start hammering your your BigQuery, especially BigQuery is one where it's like they do per data scan. So if someone just like a select star, you can really get screwed pretty quick. Well, well, good thing we can sit in a lot of different policies to shut the out of it. You don't.

00:50:58:22 - 00:51:20:14
Unknown
Yeah. Yeah. The the data set that's currently available, you'd have to do a lot of work to, to spend a lot of money. There's just not a ton of data because it's it's all the summary data. Okay. It's not the one second data that's in the data warehouse. Depending there's different tiers to it, but the, the traditional just data warehouse is just the summary data of the chemicals prop end, you know, end of stage.

00:51:20:14 - 00:51:39:08
Unknown
Okay. Summary, that kind of stuff. That's the bulk of it. So again, you'd have to do a lot of work. You'd like to get burn through, you know, terabytes and terabytes and terabytes that stuff. And even then that's, you know, 20 bucks. Yeah. Yeah. Like that Google BigQuery pricing is actually, I think fairly, pretty good I don't know, can you can do it super affordably.

00:51:39:13 - 00:52:00:10
Unknown
Yeah. You know, but it's just a different model than like, say, snowflake where it's compute time, right. You get, you know, charge for where it's like it's more data scanned and BigQuery. Right. Yeah, yeah. Data processing we've had we have to teach that process that lesson to a few people internally. Come on. You know, we've got some dedicated user, views to kind of help curtail that a little bit.

00:52:00:10 - 00:52:16:21
Unknown
But it's just if you come from a, just a traditional, that traditional, Postgres background or, you know, MySQL, Microsoft SQL server, it's a just a different way of thinking about it because you're just like, oh, it's like start just me, join everything in it, see what I want. You know that like, like, well, that was a that was a 16 gigabyte or terabyte query.

00:52:16:21 - 00:52:32:00
Unknown
You just did like, like to get two columns that like, I think people have been burned. I've heard like where they'll do like a top 100 or limit or whatever that doesn't actually do anything. But it it has to scan it all first and then it. So then you have to pay for all of that scanning but only returned.

00:52:32:02 - 00:52:49:12
Unknown
Right. Yeah. It's awful. Yeah. All that's wrong. Yeah. Just truncates the response. That's all it does. So yeah we the other you know it's not it's never been that bad. But we have. No it just runs forever. And then you get you know we see a little spike in the billing or like what was going on there. And we can go back and kind of, oh, let's go have a little, little chat.

00:52:49:14 - 00:53:02:19
Unknown
Interesting. Right? I don't want to go too far down a rabbit hole, but like it was on the snowflake side, but I saw an article or LinkedIn post and it was it wasn't clear. I thought it was like someone on the way out said like, screw you to the company. But I think this person might got let go after they did this.

00:53:02:19 - 00:53:23:02
Unknown
But I think he set like the warehouse like a triple XL, but then turn the auto suspend off like set to zero or like so like that just shut the lights off. So then you just ran like, I think they like $150,000 bill or something like that. Yeah. Okay. So yeah, you put a backdoor in when they lay him off, he could just hit that and get one last few in.

00:53:23:02 - 00:53:42:12
Unknown
Oh man. I do have one. One question that I wrote down that I want to I want to ask you because I do think it's, it's interesting, and kind of important as we kind of come to the end. So, like, in the past decade around the frac data space, there's been a ton of focus on, like operational efficiency, right?

00:53:42:12 - 00:54:01:22
Unknown
Which was, as we've already discussed, with all the 5 or 6 Sigma stuff, long overdue. Right. Where do you see where do you see that moving? Because I feel like most service or most companies, both operator and service company, have really focused a lot on the ops side. Right. And justifiably so. That's where all the money gets spent.

00:54:02:00 - 00:54:21:22
Unknown
That's where the big, you know, time is running out and. Right. Like and time is money could not be more impactful than in that area. But you know, like when I first started, it took 30 days to drill, 10,000ft lateral. Now we're drilling three mile hours and, you know, less than two weeks. And so, you.

00:54:22:00 - 00:54:45:18
Unknown
Yeah, at some point you hit an operational cap, right? Like there's no more juice, or there might be a few percentage points left to squeeze by right. Exactly. Where do you see things going? Right. Like that kind of, you know, just loosely from your experience, your your thoughts. Where do you see things going? Right. We've seen there's a lot of innovation around the control systems happening, which I think is fascinating.

00:54:45:20 - 00:55:16:00
Unknown
We've also seen things like, you know, the real time frack monitoring. I worked on a real time screen out project five years ago. You've got all the stuff that, like, the guys over at Devin did with FDIC. Is it do you think it's moving more towards that, like an optimized real time design, or is there something else that we're totally missing and do you think is is coming up with a yeah, I, I personally think there's there's two things that's happening in the background because it all comes down to time.

00:55:16:01 - 00:55:42:12
Unknown
Right? There's there's only so many hours in a day when we have frac crews that are pumping 20 to 24 hours a day now continuously. I'm ops, you know, you can't get another minute. It literally doesn't exist. So then it comes down to, okay, what decisions were made either in two places. One, the decisions that made that actually affected what went in the well and then what decisions were made to get everything to that location to then do that job.

00:55:42:12 - 00:56:04:18
Unknown
Yeah. So going back to, you know, we're building things for internal users and external users. Our external users are asking us, how do I know that every decision made in that 24 hour period was the right and best decision to make that, well, work better or to produce more? And then our internal users are asking, how did I know that every decision, up to the point where we were ready to say go was made optimally?

00:56:04:20 - 00:56:37:10
Unknown
Yeah, all of that is a software problem at this point, right? And it's all how users are interacting and how people are making their decisions, either based on just historical inference or do we have agents actually helping make these decisions or actually running these decisions? I think that that's what I'm saying. And I do think, from the, well, perspective, it is how do I make sure that every decision made either by the by the company man that's on location, by our engineers or in the field, our customers who we're working with in their centers, they're making the right decision every single time across that 25 hour period.

00:56:37:12 - 00:56:55:12
Unknown
Because even though we have 24 hours, not every hour is created equal. Yeah. All right. So I think that that is where we're going. It's first it's how do I know that we made the right decision. And then it's like, okay, how do I make that decision faster? I think those are the that that's the thing that I'm seeing most, is just validate that we did the right thing then.

00:56:55:12 - 00:57:17:16
Unknown
Okay. Now just do the right thing every single time and make it repeatable. Yeah. Right. That's that's the biggest thing you think in the long term I, I personally keep coming back to you. Yeah. There's still 90 to 95% of that oil in place in the, in shale. Right. And so it's like at some point we've got to revisit that because, yeah, because we're not finding it down there.

00:57:17:17 - 00:57:41:21
Unknown
We're finding new sources of oil nearly as as quickly as we were ten years ago, even. And so I'm very curious, like, that's just my personal hypothesis is that we're, we've we're hitting our peak on the operational efficiency side. Right. Like we're doing tri ops sim ops. That's right. You can't fit more. Yeah. In a day. I mean, we were we were elated if we did 4 to 6 stages in a day when I was doing this.

00:57:41:21 - 00:58:00:10
Unknown
Right. Yeah. Daylight work. And, I was in the field during those times to save the day was cooking. Yeah, yeah, yeah, we were getting stage bonuses. You're imagining the stage one. This would be right now. If you do, like, 20 stages in a day. Incredible. And so I'm just I'm very curious to see, you know, also, even just from a frack perspective.

00:58:00:10 - 00:58:23:05
Unknown
Right. Like a conversion, a conventional, conventional frack job, the the consultant or the engineer, whoever's running the job is going to change that design based off what the well, how the well reacts. Right, is a reactive thing and the design is just there is a suggestion ultimately, if you like unconventional frack work is the complete opposite. It is only the design.

00:58:23:06 - 00:58:47:12
Unknown
We are right and we only pump two plan and that's it right? Like and yeah but also detrimentally so because like I remember in the fable we, we were, we were looking the wireline was in the hole. We were looking at the schematic, the well survey, and we were like 200ft lower than the Fayetteville. And they were like, frack it or perfect.

00:58:47:13 - 00:59:08:15
Unknown
We'll try and pump it or whatever. We spent like four hours or it spin up slick water, 30 pound cross-linked, you know, everything. We're just burning money to frack a zone that's not in the reservoir that we're actually targeting. And it's like, how many times has that happened? Or does that continue to happen where it's like at the end of the day, there were just dumb decisions like, that's a lot of that.

00:59:08:15 - 00:59:36:12
Unknown
But ultimately, I do feel like all this starts accumulating towards how do we get more out of the same rock? And I think it's going to be fascinating to see where the service companies play into this, because I do think there is a lot of interesting stuff, whether that's, you know, the offset monitoring with brackets, there's going to be a bunch of real time things that start getting rolled into the ops that really start enhancing both not just operations, but also also like how I think I asked earlier.

00:59:36:12 - 00:59:58:22
Unknown
But that's a good point. Like how opinionated do you want to be? Like you, you have frack engineers, I'm sure they work collaboratively with like the the completion engineers at the operator. But at the end of day, like they give you a design and you've got to pump it to their specs, right? But like, how opinionated can you be on the front end or are you already where it's like, hey, I think you shut it down now or like, you should.

00:59:58:22 - 01:00:16:18
Unknown
Yeah, you should pump some diverter. Yeah. I mean, that's another thing, right? Like is it we're pumping too long. We, we hit a offset. Well, we should shut down and move to the next stage. Right. There's all cutting decision like our ten years off. Does your software like, you know, try to have an opinion on those things and obviously is there their choice or or is it just for reporting the news right now.

01:00:16:18 - 01:00:31:14
Unknown
And yeah, this is I mean, this is clear. This is one of the biggest topics of all of our customers, right. And I think it's driven a lot from the fact that we have been fortunate. I feel that most of our customers have had a lot of significant tier one acreage that they've been able to put out designs that are just repeat, rinse and repeat, rinse and repeat.

01:00:31:14 - 01:00:49:02
Unknown
Rinse means fine. You're talking about Fayetteville. It literally that reminds me of, us moving from Marcellus into like deep Utica for the first time in the 2000. I literally went through the same thing in like, Eastern Ohio. Yeah, we're just we're just slinging it like it felt like wildcats to a degree. We were just like, add six clusters, stage five clusters.

01:00:49:02 - 01:01:11:15
Unknown
This stage, let's just pump it and see what happens. Is it really just dealing with the pressure response is funny. But I think what I think our customers are asking of us specifically is just number one, be open minded. Yeah. Because I think we're we're learning what works in that regard together. I think everybody is going to be looking for some type of real time operational change.

01:01:11:17 - 01:01:29:08
Unknown
The I think the key difference is going to be what triggers that change. Right? Right. Well, who is going to find or what is going to be that ultimate solution that will say, okay, if I if I see something happening on a different well, what does what response to that Garner? I mean, I think the workflow is important.

01:01:29:08 - 01:01:47:04
Unknown
The causality and the thing that actually they determine is this is my flagship thing that then creates a response is going to be the differentiation across the most of our customer base. I think our responsibility as a service company is just be versatile enough, and how we build and interact with our software that we can service. All right.

01:01:47:06 - 01:02:07:13
Unknown
You know, our equipment strategy at our company has always been, you know, our fleets are going to work wherever they are, whatever the capability is. Right? We have hybrid fleets. We have next gen equipment mixed with tier four dual fuel equipment. Our controls work with everything. Our software works with everything because we're not here to box our customers into a certain set of, you know, you must use this.

01:02:07:13 - 01:02:28:16
Unknown
You must use that. Right? No one would ever do that. And so, I mean, we have to have our, our software be responsive to any type of input. I think what our customers are now asking us to do is with what we have access to, what what is our opinion? I think our opinion is actually being asked more than ever, which I think is incredibly healthy.

01:02:28:17 - 01:02:45:16
Unknown
I mean, you've seen a lot more than I have, right? Like, I mean, like that's always been the thing, right? Like I was I was, you know, support engineer or a technical engineer selling frack jobs and supporting frack jobs to an A&M grad who just graduated and had never been on a frack job first. And I was I'd been doing it for three years.

01:02:45:16 - 01:03:08:12
Unknown
And it's like that person has no clue what is going on, no less outside of their lease, right? Like what? As a service provider, you see everything, you see every what everyone's doing, what's working, what's not right. And yeah, I feel like historically speaking, the like power, not the power, the knowledge, the trust in the knowledge has always been at, you know, oh it's operator knows.

01:03:08:12 - 01:03:38:22
Unknown
Right. Like that's that was always my mentality. Right. I'll make a suggestion. But if it's your call ultimately I mean we've seen like especially around the reservoir, we've seen data driven analytics models start to take over some of the physics based maps. Right. Because we do live in an imperfect world, even in the reservoir space. Right? So now that we have companies like next year who have massive amounts of data that we can do accurate statistical based modeling to actually get some of these outcomes and predict these outcomes that could have an effect on the reservoir.

01:03:38:22 - 01:03:57:08
Unknown
It's incredibly powerful, right? You don't necessarily need the full physics background to do so. What you need is a data science background and clean data to work with. Well, that's that's always my like. And I don't have I'm not smart enough to know one way or the other, but I always feel like, yeah, there's always the physics driven folks.

01:03:57:10 - 01:04:16:20
Unknown
But if I'm recording data in the field in real life and we're not living in a simulation, then that is physics driven sensor data that I am capturing. So the physics is built into the statistics that I'm now running on, and you're entering in that gray area where I like, there's this myth about which chicken in the egg situation.

01:04:16:20 - 01:04:37:21
Unknown
Right. And so but it's you know, anyway, I do there's a world for both. They both help. But I do think people have discounted the statistics more than they need to. I had agree because of that physics nomenclature, but it's like the data was generated in the world with physics and gravity and everything that existed. Right? Like it's physics derived data.

01:04:37:22 - 01:05:11:02
Unknown
Right? It's we didn't make it up. Right. Well, I mean, there's a reason why our data science team has literally doubled in the past. Yeah. And it's driven by this, and it's really coming from the customers wanting to ask us, right. What else can we figure out here that maybe nobody else has yet. Right. And I think as we see the tier one acreage start to fall off and we start seeing people having to go to tier two, tier three acreage where it does matter every single decision that you make, because that could mean a couple percentages of production that the merge charge on those trucks is models that are going to be better.

01:05:11:02 - 01:05:36:15
Unknown
Right? They have to be right, or else we're all going to be in trouble. Right? So are you saying that you guys can take them from tier one to the next year? We're saying, you gotta be okay. That's all we got, folks. That was perfect. Oh. Thank you. That was good. No, it's a that's a it's a it's a crazy world.

01:05:36:19 - 01:05:57:18
Unknown
I mean, like, it's just like it, it's so multivariate because, like, we've talked to this before too, but it's like we can do all this and optimize it here. And we still don't know if that was the thing because, like how that increased you. Right. Like, because then they put it back differently or the, you know, that's 20ft away that no one saw on the logs is impacting it, like, you know, but the good thing is people are actually trying this stuff.

01:05:57:18 - 01:06:13:15
Unknown
Yeah. Yeah. Like they're they're actually open to it. Yeah. Which is where the new collaborative environments come from. Five years ago, that probably would have been the case, or at least the requests weren't as loud as what they are now. Yeah, right. So you have legitimate leaders now on the amp side who are saying this is the future, we must push for this, right?

01:06:13:16 - 01:06:30:16
Unknown
Which is really great for us because now we can be like, we have a lot of smart people that really want to dig their teeth and just let them go. Well, that's, that's that's what's happening now. It's a win win, right? Service companies don't exist if the operators all shut down. Yeah. So let us help you help everyone.

01:06:30:18 - 01:06:59:09
Unknown
But yeah, I mean, that's the thing that like a 5% increase in current production is remarkable. And yet, like in that, like it could be some little design change. Right. It could be clusters but it could be anything. And even then you still have 90% of all it. Yeah. But I mean in some of those things that are like the change in the paradigm that keeps you in zone, I mean, like, I know we did some stuff at Chris and Mila where we proved out more of the western side of the box and where we but we did some, I think would it be more clustered about where it kept that more near

01:06:59:09 - 01:07:13:22
Unknown
wellbore. So you didn't frack out into certain zones where you're getting more water, and all of a sudden you have a logistics issue with water hauling because you know how the infrastructure out there, but if you can keep it in zone now, you actually get the oil out. That's right. Yeah. There's a lot of like we're 50 years of engineering and science, right.

01:07:14:00 - 01:07:29:12
Unknown
That went into fracturing before we started the shale stuff. Yeah, yeah. Made a lot of sense. But even like for me, like I've got a customer up, you obviously leave their name out of it, but we just really talked about yesterday. I have two new wells came on on a pad and they're in the same zone, same pad.

01:07:29:14 - 01:07:51:22
Unknown
One of them is getting iron sulfide and they have to ask. The other one doesn't say so. They must have fracked into some zone on the one where they were exposed to. Yeah. Or either there could have been something geologically. Yeah. Yeah. No, I mean, at the end of the day, we're doing stuff 20,000ft from the surface that we have no idea what's going on, and we still somehow figure out how to do it.

01:07:51:22 - 01:08:14:21
Unknown
It's crazy. Yeah, it's truly crazy when you think of, like what? All the medical imagery and shit doctors have now, and it's like we're perving 40,000ft diagonally that way on a, you know, 10,000ft vertical and three mile lateral. Yeah. And we somehow we're right there like, it's crazy. It truly is nuts. Yeah. This has been awesome. Yeah. That's been a lot of fun.

01:08:14:21 - 01:08:31:23
Unknown
Yeah. Like I mean we always say yeah, but we we could have felt like yeah, yeah I could, I could honestly do this all day. I really could. I mean, just the last six months alone and some of the things that we've been able to see and do with some of our customers has been insane. So like it it's getting faster, which is what we talked about I earlier, we didn't even talk about how they're using.

01:08:32:02 - 01:08:41:20
Unknown
I know how are you using I real fast we use it right.

01:08:41:22 - 01:09:05:05
Unknown
You know, I represent you know, speed one speed question. Yeah, that one's good question. Brian, what is your go to, coding tool? AI coding tool. Do you have an IDE or. Oh, I mean tool that you like. Yeah, we all use, I've been just in JetBrains, paired with cloud code has been pretty good.

01:09:05:07 - 01:09:24:20
Unknown
It started out I tried a few of them at the beginning and just at the time, at the beginning of the year ish. Yeah. Know really got into it. Cloud was just ahead and we just kind of kept running with it, as a team. So we, we put a lot of our stuff to vertex AI and GCP, so we could kind of not have to worry too much about what's going on on the, on the cloud side of the world.

01:09:24:22 - 01:09:44:23
Unknown
But, that's been immensely helpful for me, especially somebody like the 16 years of software developer experience, being able to see what it's doing and sort of just like, I'll go do that and I'll go do my other it is, you know, thing that I'm working on. It's just so much fun. Yeah. Like I was telling Bobby, he's got a download, Hermes desktop.

01:09:44:23 - 01:10:01:23
Unknown
Just so he can, he can get into the open source world and harnesses all that stuff. But also it's like, yeah, I built a game that I put out on LinkedIn last week. Yeah. Using Hermes, using an open source model on my laptop while I was doing something else. Yeah. And it's like, it's crazy. Yeah. No, it's a it's a lot of fun.

01:10:02:03 - 01:10:20:21
Unknown
I didn't notice it. Like, you get a little less of the dopamine hit when you solve the problem than you did when you were like after, like struggling things that you're desensitized now chasing the next one. Yeah. Yeah. It's a very weird thing. Yeah. But it's also like, super exciting with all that problem solved, I can move on, you know, like, yeah, I go like, that might have taken me five days or.

01:10:20:21 - 01:10:35:03
Unknown
Yeah. Yeah. Oh, yeah. I was just talking to the one of my guys on my team of like, you know, I mean, cloud slow today. It's it took like ten minutes for it to get this done, you know. But you should probably like this would have been two days worth of work, you know, like, well now. So yeah, APIs are slow today.

01:10:35:06 - 01:10:56:09
Unknown
You know, like actually I mean, we have a we have a quad alert channel in our slack, but every time they make an update on their API, make sure it's up or down, right? Yeah. No, but it's a lot of fun for sure. You guys are both from Western PA or. Yeah we are. Yeah, I'm from southwestern PA.

01:10:56:09 - 01:11:14:23
Unknown
He's from northwestern P.A.. What's the go to cheesesteak? I know that's the wrong side. I know, I know, it's like, oh, she's. Thanks. Yeah, yeah. All right. Do you know cheesesteak is a Permanente sandwich? Yeah. Yeah, that was a perfect answer. I was going with that, but I was I was going to ask more like, have you found a, good place to get pierogi down here?

01:11:15:01 - 01:11:32:19
Unknown
Know that in that interview? Oh, yeah. That's like that in wings. Yeah. I miss my brother in law. He's from Connellsville area and like. Oh, yeah, he's like the one we have. We have a shop near there. Yeah. So what, like the wings are not on here just to deal with the wings I what do they do different I don't know I, I can't find the sauce variety.

01:11:32:19 - 01:11:49:11
Unknown
Yeah I feel like every wing places like the same six sauces. And you go to any place in the northeast, they've got like 20 or more or they're not legitimate. Yeah. And I think they're just they're just not as crispy. I don't know how they cook them or you gotta have doubts. So I get puckers in some of those places.

01:11:49:11 - 01:12:10:13
Unknown
You can tell I'm like, double fry. Or like you just northwest. You just you just judge the sauce really harshly. Yeah. Like my my favorite wing place in Pittsburgh. It's it's like they've got like, 30 plus wing flavors and they're all named after, like, local things that. Yeah, you only know what I'm saying if you're from there, like, sure.

01:12:10:14 - 01:12:32:09
Unknown
My favorite wing sauce is called a Beaver Falls. Does that mean anything to you guys? Exactly. But it's probably the best wig sauce I've ever had. I can't get it anywhere but there. Yeah. That's perfect. Yeah. Guys, thanks so much for, for coming on today. Where can where can people, get in touch with you? So you can find us on LinkedIn?

01:12:32:11 - 01:12:48:05
Unknown
You can find us on our website, either next to your Office.com or pat energy.com. You can request a demo of the platform through there as well. But that's usually the best way to reach out to us. Or if you're a customer, call your local sales for yourself so they know how to get Ahold of Brian and I.

01:12:48:05 - 01:13:06:07
Unknown
Yeah, yeah. Cool. Beautiful. Wraps up. Yeah. No, no. Thanks, guys. But, you know, if you like this, make sure you, like, subscribe all time. Yeah. Like subscribe. Watch this thing. And, wherever it is, you know, then it's on us to actually start recording more. Yes, we will do better, I promise. I've got an I've got an agent that I'm working on.

01:13:06:09 - 01:13:08:04
Unknown
Take it easy, guys. You guys. Thank you. So.