Evan: Hello everyone and welcome to the Mining Your Own Business podcast. I’m your host, Evan Wimpy, and today I’m excited to introduce Gus Kaeding. Gus is a senior manager for data and analytics at US Ski and Snowboard, which is a super exciting title to have. We’re taking this, taking this interview from inside in a nice little booth here. But, I asked the listener to just envision, we’re on the ski slopes somewhere. Powder flying up everywhere. Hopefully, that’ll make it more entertaining. That’s what I’m gonna be visioning. But happy to chat with Gus today. Gus, welcome to the Mining Your Own Business podcast. Thanks for coming on the show.
Gus: Yeah, thanks for having me, Evan. I’m here in Park City, Utah, and the leaves are popping. That means winter’s right around the corner and we are counting down the days.
Evan: Fantastic. I’m in Raleigh, North Carolina, and it is mid-eighties and hot and sunny today, so it feels like a long way till snow season. Gus, to get started, can you give us a little bit about your background, maybe both into data and analytics and also in the sport in skiing or snowboarding?
Gus: Yeah, sure. So I grew up in Vermont and I was a skier growing up, ski in college where I studied mathematics. And was a professional skier after college and got into coaching. Kind of decided I didn’t want to coach for my entire life, so I went back to school. Got my MBA. Focused on strategy and finance during that time and kind of made a decision that I wanted to work my way back into sport and affect things at a little bit of a different level. Coaches are kind of one athlete or one team at a time, and I wanted to think a bit more systematically. So after grad school, I went to the US Olympic Committee and worked in their internal consulting unit where I kind of learned to apply some of the things I’d learned in school – really formulating a problem, looking at how we want to approach things and solve things. And I mentioned I was a math major earlier, so I had a bit of a quantitative background. And shortly thereafter I made my time – or my way back here to US Ski and Snowboard and really created a position that was about kind of asking the questions of “What does it take to pro improve performance at more of a quantitative level?” And I’ve been in this position now for five years and most of my kind of data science and analytics background has been self-taught during that time just to address individual problems that I’ve needed to work on.
Evan: All right. Super, super interesting background. And sports, maybe not necessarily within ski and snowboard, but sports analytics I think are sort of a window that’s accessible to a lot of people to see what data and analytics can do in some US team sports. You think about the data that’s collected. There’s box scores in baseball going back 150 years. What is the data like that you have in skiing and snowboarding? You don’t have a nice small field to play on and there aren’t easy numbers that anybody in the stands can collect. So I’m curious what type of data you typically have that you’re working with.
Gus: Yeah, we actually do have the ski equivalent of box scores with times and competition results. And one thing that’s nice about almost all of our sports is their individual sports. So what a lot of sports these days struggle with is, they have all of their variables tracked, whether it’s in the gym or wherever, but they don’t have, they have difficult outcomes to associate those two, especially with team sports. With individual sports, we kind of have the outcome piece done, and our challenge is more what variables actually solve for that outcome. And oftentimes that is very challenging because of the conditions and on slope environment that we deal with. So when I first came, the only data we really had in a comprehensive manner was results. And so that’s where I started. And a lot of those, we use a lot of those results to kind of inform the strategic and higher-level decisions we make as an organization. Resource allocation development, pipelines, things like that. And now as I’ve been here a few years, we’ve been able to we use Smartabase, which is an athlete management system, basically like a CRM for sports. And so we collect all of our athletes data there, and it’s just about growing those sample sets, whether they are in the gym or at home, during recovery, or on the slopes as much as we can. And solving for “how do you create those results?”, whether good or bad and incrementally improving those. And there’s still a large delta there. And that’s the fun part is trying to address that delta.
Evan: I suspect this is not reasonable, but is there anything that’s video tracked that you’ve got image or video. A jump or from a course, which is loads of input data that can potentially feed the final result, that outcome that you’re looking at. It feels like that may not be feasible, but, I don’t know if you’ve got anything like that.
Gus: One of the things we actually do in alpine skiing is we have people videoing the entire course, and some of these courses are two, three miles long. So these guys are literally in the trees with camcorders and just getting it done. There’s not it’s not a beautiful thing, but they get it done. And so of course when you’ve got, you know, five, six people with video and in a place oftentimes with no internet, just the manual process of actually stitching together or attributing, first of all around to a single skier and then stitching them together. One of the challenges is oftentimes by that time, that actual process is done. It’s the next day and they’re already training again. So it’s a race against the clock. And, when you’ve got coaches then doing that on the hill, it means they’re not coaching. It means they’re working through video. So it’s kind of a time-benefit analysis there.
But I will say one of the programs that we’ve been able to improve is we’ve had a kind of a cooperation with the Norwegian ski team over the past couple of years where we’ve shared that video responsibilities and also some GPS with the athletes. And so we’re actually able to look then at athletes as Stitch together video and their speed throughout. It might not sound like much of it. It’s actually pretty complicated to just pull off in our world. And so then we’re able to take individual corners and look at our athletes against other athletes and see like entry velocity and then exit velocity and the correct path through a single corner. And I would say the part we’re missing is the whole run just through one corner. You know, you’d think, Okay, if I exit faster, it must be best, but actually if I’m exiting faster at the wrong line, it could actually be worse. And so those are questions we’re not quite able to answer right now, but we’re getting closer to.
Evan: That’s really exciting. I think ski and snowboard certainly isn’t the only place where the data gets more granular and more granular, and it lets you try to answer the more holistic questions there. That’s really good. So I’m curious, maybe it’d be helpful if you just walk us through an example project that you’ve, that you’ve taken on recently and sort of how that comes to light. You’ve got a background in skiing, but you’ve got coaches, you’ve got athletes that are asking questions. I’m curious how, how sort of analytic efforts get prioritized and maybe an example would be helpful.
Gus: Yeah, so. I think anytime you’re talking about an analytics kind of project, whether it’s in business or whether it’s in, you know an MBA program, buy-in’s always tough and there’s a number of different stakeholders you have to address. So most of the projects that I work on are sourced from a question that comes from the coaches. Sometimes the athletes, but generally the coaches. I’ve found that dealing directly with the coaches is my most successful kind of method. Because if you go directly to the athletes, the coaches kind of feel like they’re taken out of it, and then they’re not the ones supporting that project. So I deal with the coaches mainly, and then I let them delineate the information to the athletes. This particular question, which I thought was a good one it’s really just is our Alpine team fit and the way we. Kind of describing that is, are we the fastest through the last section in a course? So the timing reports for every alpine course show, you know, it can be three and it can be five, six splits throughout a course. And, okay. We know we’re good at the top and second, but really, You’re not gonna win unless you’re the fastest at the end. And really throughout. So are we able, if we’re starting fast, are we able to finish fast? And if we’re starting slow, are we able to finish fast? That effect is compounded by the fact that most of the injuries happen on the second half of a course as skiers begin to get fatigued and ski a little bit above themselves. So we are wanting to look at, yeah, how fit are we? And really that just comes down. Split analysis. So for every one of these World Cups, there are essentially a detailed split analysis emailed out. And generally, they’re in Excel, sometimes they’re in pdf. So we’re still kind of dealing with a little bit of like the 1990s here. So there’s a little bit of conversion at times.
But what I’ve just been working on, and this actually. I know I had said earlier I was looking forward to moving on to some new projects, and this is one I’d like to devote a little bit more time to in the future, but it is creating just more of a comprehensive dashboard with all of these splits together throughout the years just with attribution to date, location, gender, discipline and, and all of those things I’ve spent the last couple months going through and manually adding with the help of a couple interns. Just there is not a pretty way to do it. So the idea will just be to use R, and create a shiny app or something and create a dashboard for the coaches and athletes to really go through this information. I’m thinking so they can compare themselves to other skiers respective to course or compare themselves to the top 10 and see like, am I losing, you know, time at the beginning or is it mainly at the bottom? Where does USA stack up as a nation compared to Norway in terms of splits? And able to make some more kind of macro-trend analysis like that. That’s been a big focus for us, the fitness this year and you know, I would say like qualitatively in the gym, you can see the athletes making some improvements right now. But we’ll kind of compare those results at the end of next year and see really how they’ve done.
Evan: Awesome. Super exciting. Yeah. And that’s really like the input how. Talking to the coaches, coaches asking questions, they’re really your stakeholders. You’re trying to answer a question. You mentioned an R shiny app as probably the end state that you deliver – is that how you would deliver something to a coach as well? I have access to an R shiny app and I think for, you know, the data and analytic savvy that seems like a very reasonable and pretty accessible way. But I think maybe to some folks R shiny app can be a little bit intimidating to navigate. Is that the way that you deliver some insights or deliver some findings?
Gus: No. Yeah. Well, yes, at times it is, but my main platform is, as I mentioned earlier, Smartabase. Even if it’s not, even if it’s kind of a round peg in a square hole, I try and build as much as I can into Smartabase, which allows for, you know, aggregation and input both from the coaches and athletes and then also any sort of visualization. And so the core reason for that is generally compliance. Actually one of – and like the only good part of Covid – is that we did all of our Covid compliance through Smartabase. And so athletes previously had been a bit hesitant or just you know, maybe I did a poor job of selling it or whatever, but we just didn’t have like great buy-in into Smartabase. But once they kind of mandatory had to do twice a day monitor to be on the team, but also obviously if we’ve wanted to stay healthy. Everyone is in Smartabase and now the buy-in has been quite good and the younger generations kind of coming up are more used to these types of processes as well. So when I can, I try and put everything into Smartabase just so the athletes have an app. The coaches have, you know, their desktop or whatever they’re using on the road. And it’s kind of the one-stop shop for all of their performance, competition, clothing, sizing, anything they need.
Evan: Awesome. Yeah, that’s, that’s very good. Forced buy-in is still buy-in. Yeah. And if you can get it, you’ll take it.
I’m thinking, I’m trying to draw parallels to sort of analytics in the business world where oftentimes the hardest time to get folks to consume analytics is when it challenges the norms, it challenges the narratives. And I think probably, a pretty ubiquitous example is, for a long time it was, hey, stock the shelves with as many products you can have. The more variety, the better. But then we’ve got sort of a paradox of choice and the data suggests maybe fewer products, fewer offerings are better. And that’s really a data-driven thing. And so I’m curious if in the sport and the ski and snowboard world, if there’s anything where you felt like you’ve had to challenge norms, like the data suggests taking this corner, you should be cornering this faster. And everybody else says, No, we should be more conservative here. Or, you know, fitness is good and we should focus on something else. Maybe that’s not a good example, but is there any time where the data has really challenged what people typically believe?
Gus: Yeah. I think there have been more micro examples, but right when I began with the team, five years. I had kind of been collecting a whole bunch of results data and looking at development pathways. And the best ways, you know, the ultimate goal of the US ski team is to have medal winners at the Olympics and World Championships. And so it’s “what does it take to be the best athletes?” And you know, the conventional wisdom was we kind of take the best athletes from around the country and we helped them become a little bit better. And that basically meant just you know, the elite athletes. What the analysis that I worked on really showed is we were kind of ignoring that development portion. And so, probably the first step there was helping athletes. We would, we were taking younger athletes and helping them develop more quickly and that was okay. We developed a number of athletes quite quickly and we actually, I kind of had said these will probably be showing up at 2026 games. We actually saw a number of them in 2022 either just off the podium. I think we had a number of kind of teenager-to-early-twenties in the five to 10 range across several sports. But I would say the second iteration, and now what I’m more interested in of this, of solving this problem is instead of, you know, taking the athletes out of where they’re comfortable and fast-tracking them to help with development. We are really trying to use that athlete to bring their friends along as well. And so it’s more of a scalable solution.
And the kind of the data behind all of this is if you have some of the senior athletes who have, you know, a two or 3% chance of a medal you know, kind of a long shot, but they’re a strong competitor. You might actually have athletes in the junior ranks who have really high potential and have a four or 5% of the medal. They’re just kind of visually a lot further from it. But they might have some of these young indicators that mean they are going to fast track to very quick results. And that’s where we were a couple years ago and now we really wanna make sure we’re scaling. And when we identify some of those athletes, to support them in a way where they’re bringing their friends, their community, and the sport along with them as well.
Evan: Awesome. Yeah. Super exciting. Scalable solutions. It’s a push, push everywhere. I wanna get your take on – I don’t know if you, you’re still active skiing and snowboarding personally. But I think practicing data science, I think about when I’m part of the data generation process. I think some data scientist is gonna be on the back end of this collecting, analyzing this data. Does it change the way I think about and like from using the shopping card to the way you drive or anything? I’m curious if any of your work analyzing and looking at ski data has changed the way you think about your own skiing. When you’re on the slopes, do you think, ah, I need to be more upright here, or different positioning or change my approach on the course?
Gus: So we have some really cool projects kind of in, I would say, beta with biomechanics, those sorts of things. And I wouldn’t really say I’ve felt a personal connection to any of those. Certainly. I mean, I used to think I was a good skier until I skied with some of the guys on our team, and it’s just, it’s in completely another world. I’m not even close. The part I would add looking at the data and then doing the sports are, I also manage our EMR, the electronic medical record, and look through all of our health data and we look of course to injury rates, re-injury rates best practices in terms of rehab. And all of that, you know, we look at snow types when people are getting injured and I can’t help but think about that stuff a lot more. I am an insanely conservative skier now just because yeah, I don’t want to get hurt and it’s amazing to watch like an elite athlete work through some of those rehab processes and see how focused and dedicated they are. And I just know I am not that, and like even seeing. You know, pretty rare that they’re able to make it all the way back to where they were before. With that dedication, I just I’m skiing the back of the pack from now on.
Evan: Well, I hope you, I trust that you can still have fun out there on the slopes. And hey, look, if you ever – I fancy myself a pretty good skier because I live in North Carolina where there aren’t many ski slopes and they’re pretty small. So if you ever want to feel like an elite athlete again, just come to the southeast and ski and you’ll feel comparatively pretty good.
Gus: Yeah, sounds good. There’s a few mountains out here in Utah for sure, where you go out on, you know, any given day and you’re like, Oof. There’s some good skiers out here.
Evan: Yeah, I used to think I was good, I guess, until I was around just mere average folks. So I guess I wanna ask you one last question here we’ve got time for, you’ve alluded to a couple of interesting projects, but let’s just say you’ve got unlimited resources, you’ve got time technology, you’ve got people and a team, and all the coaches and athletes are bought into Gus’s vision on what you want to do, how you can improve us ski and snowboard. What’s the project that you want to tackle? What’s the question that you want to point your analytics efforts towards?
Gus: Mm-hmm. Well, I mentioned earlier that we know the outcomes of our competitions and we even know the outcomes of our training runs. We time them top to. What we don’t, and we know, we know what happens in the gym. We know what the athletes are doing in the dry land. What we don’t really know is what happens from when an athlete leaves the start gate to when they finish. So this has been our challenge and there are sensors out there, but there are not really sensors that granular enough with the battery power to last through an entire training session, ah on the slopes and even when there are those sensors that we are, yeah, in, this is one of the things that I mentioned. We’re in the beta stages of developing, so kind of looking at forces throughout a turn. We don’t have the video to go along with it. And that’s often just because of kind of the challenges, my outlaid, you know, just stitching it all together. And is it a worthwhile process for it? And the way we do it at the moment and not sure it is. So, you know, if I had unlimited resources, it would basically just be setting up an entire course. That video and each athlete was auto-tagged and they had sensors where you could capture everything that’s happening and really understand it. Tag all of this with any sort of ai just to piece together, you know, what happens, what leads to a good turn, to a bad turn to an. And then of course we could overlay with those statistics that we keep during the summer, you know, how fit is this athlete? How recovered are they are what, what are their scores in terms in the gym or whatever. But it’s just those, those granular pieces that are on the snow right foot versus left foot force. Because those forces that they generate going around the corners are really. and that’s just the kind of thing that we don’t understand yet. And even at, you know, one or two turns would be amazing, but to think of like a gauntlet through an entire course, to just see the athlete at every second and track what they’re doing on the snow would be, that’s the dream.
Evan: Awesome. Yeah, that I feel like that’s a common thing of I wish. Better, more reliable, more granular data. So maybe we’ll have you back on the show in four or five years when technology allows for that.
Gus: We’re getting a little closer. There was, if, you know, we’ve kind of been down this road for a couple years now, and I would say waiting for cost to kind of come down and, and we’re getting there. And I think it’s a realistic goal in a little bit more closed environment, like a snowboard halfpipe right now, an alpine ski course that’s a lot longer. Probably still a little ways off, but something that we can control and see from top to bottom with 1, 2, 3, 4 cameras.
Evan: Yep. Awesome. Yeah, that’s exciting. You referenced battery life there. Quickly, I just got is, is that I assume the cold weather makes this a lot harder of a problem.
Gus: Yeah, I mean, those are just, it’s, those are the challenges that you just don’t think of, or it’s condensation or it’s anything outside. And when you’re dealing with variable temps and international travel where you can’t bring a lot of these things on planes. And honestly, some of the competitions don’t allow us to use some of this technology either. So this, the sport still is, needs to do a little catching up on. Side of things. Sure, sure.
Evan: Makes perfect sense. Hey, hey, Gus. I’m gonna break my promise here. One last question for you. For the fans here of US Ski or Snowboard fans in generally, what should we be looking out for? What should we be excited about for the US team?
Gus: Well, I think I mentioned it earlier, a number of young athletes that have kind of shown promise at the last Olympics. And also a number of young athletes that probably should have qualified for the Olympics but came onto the scene so late last season that they weren’t allowed to go or just didn’t qualify. Moguls, aerials, cross-country skiing, alpine all have some of the best development programs right now in the world. And so we have kind of been building and looking for that next group and I think they are arriving. When you look at the gym this summer I’d mentioned free ski and snowboard, but those guys are kind of perennially at the top. Yeah. We’re all trying to chase the standard that they’ve set, which is awesome.
Evan: Super exciting, exciting time to be a fan. Exciting time to be working with the data. Gus, thanks so much for coming on the show. If you enjoyed this content, make sure you follow US Ski and Snowboard cheer on the team. Subscribe to this podcast for more fun episodes and learn how you can use data in your own organization to help drive more success. Guests. Thanks so much.
Gus: Thank you, Evan.
Joining Evan on the data science slopes is Gus Kaeding, Senior Manager of Data & Analytics at U.S. Ski & Snowboard.
Evan Wimpey | Host