Wie baut man ein Radsportteam mit Daten? Inside Red Bull – BORA – hansgrohe

00:00:00: Because the peloton is getting younger and younger now, you know.

00:00:03: You don't have guys winning grand tours at thirty years old or twenty-seven years old that much anymore where that was sort of considered your kind of peak endurance years or whatever And you would win a Grand Tour Or the classics in that way.

00:00:16: Now if we take Paul Seishas who's going to the Tour de France now He's expected to win so to speak or perform super well.

00:00:26: But what has changed with that, you know if we really take what she said on the moneyball is.

00:00:31: You're now dealing With a kid.

00:00:34: who Is a kid?

00:00:35: There's nineteen years old and you saw your child or seventeen years old but in the cycling environment... ...you want him to be a seasoned professional, an old professional.. ..but when he leaves then still wants them to be kids otherwise he misses out alot.

00:00:50: And thats kind of important point You are now dealing with is that you're dealing with this guy?

00:00:56: Well, you know your paying him X which is typically a lot of money.

00:01:00: We need them to be as immature adults But he's still the kid.

00:01:04: This is AI That Works by Exeter Where real stories of AI and data meet business reality.

00:01:15: Welcome to AI That works.

00:01:17: I'm Johannes A solutions architect at Exeter attack.

00:01:21: in consulting company.

00:01:22: The helps organizations put AI to work.

00:01:25: We see chatbots thrown at everything these days, promises are big results frequently or not.

00:01:31: I want to talk in this podcast who people were experts in the fields and use AI and data in sports business science.

00:01:39: wherever Data matters we ask how our decisions made what data matters What data doesn't?

00:01:45: What works what does it And what can organizations learn from it?

00:01:49: ai sits on three pillars infrastructure algorithms data.

00:01:54: Today, we want to look at data and how it's used to drive decisions.

00:01:58: How is use to get ahead?

00:02:00: In this case in first place We will enter the world of professional cycling together with our guest John Wakefield from Red Bull Bora Hans-Kohl.

00:02:09: He and his team use large amounts of data and sophisticated machine learning algorithms To predict rider performance And uses that build a team From scratch on computer through power.

00:02:22: Before we dive into our first episode, Let's get into today's story.

00:02:58: We're joined by someone who works at the intersection of sports science, data and team performance.

00:03:05: together with John Wakefield from Red Bull Bora Hans Gruller we explore how Data Science can help make better decisions And build a high-performance cycling team From scratch in what any organization Can learn.

00:03:16: form The process?

00:03:22: John is the team's director of coaching, sports science and technical development.

00:03:27: And one other people shaping how The Current and Next Generation Of Riders are identified developed and prepared for the world tour.

00:03:35: You have played a key role in building the Red Bull-Bohra Hansgrohe Development Team that cyclists under the age of twenty three and nineteen respectively.

00:03:45: you built That team in twenty twenty five and you build it from scratch From behind your computer almost entirely data-driven, and you have a lot to say about all components of cycling training sport science.

00:03:58: Data in long term performance.

00:04:01: Thank You for being here because you have along career All the way from motocross And then through accidents not by accident got into cycling.

00:04:12: yes uh,

00:04:13: I guess luckily because now you're successful cycling coach.

00:04:17: You are the director of coaching sports science and technical development which is an interesting intersection of many fields.

00:04:25: i don't want to run employee anybody who's listening too.

00:04:29: use their favorite ai-to have full manuscript your career but feel like that.

00:04:35: last year was a good summary how successful as a coach were in finding talent and building successful teams because Florian Lipowicz, at just twenty four years old.

00:04:50: At his first Tour de France finished as the best young rider for your team And being the best Young Rider isn't even The biggest accomplishment Because he finished third overall at the Tour De France.

00:05:03: It's only the fourth German to ever do so.

00:05:06: So I think that

00:05:07: speaks

00:05:09: for your abilities as a team coach and, I have coaching principles including it must constantly evolve.

00:05:19: The margin between success or failure is razor thin.

00:05:22: performance gains are coming from attention to detail And you're known for patient long term approach in developing young riders.

00:05:32: Would you agree or any of these statements?

00:05:39: No, no correct.

00:05:40: You've definitely done your homework well-done.

00:05:43: Okay I feel like nobody really says oh II just look for short term gains.

00:05:49: everybody would like to claim that they do long-term Like a long-terms approach

00:05:55: with

00:05:55: you.

00:05:56: it does look like It is Really like Your track record shows That that you have been doing.

00:06:03: You are building the rookie team and your coach, what is your daily life like?

00:06:10: if you want to give an overview of these things?

00:06:13: Because there's a lot responsibility in one person – technical development, sports science and coaching.

00:06:22: Yeah it's busy day but everybody in industry consists… If we take those sort three umbrellas.

00:06:31: The coaching side is the team of coaches that I work with, you know?

00:06:36: I'm constant communication with them.

00:06:38: it's overall pro cycling teams athletes prescription if they're coaching making sure their over wellness and performances where should be?

00:06:51: sport science side is making sure that we are up to date on the sports science side of performance.

00:06:58: So new methodologies, new trends what is good?

00:07:03: What isn't good?

00:07:04: and then the technical side of it... Sorry!

00:07:08: And the sport science side there's also from you know the data in that performance software side and then from the technical development everything from tyre compounds tire pressure to aerodynamic frames, to stitching on the skin suits etc.

00:07:28: and different materials... And everything in between.

00:07:31: that I am lucky enough have a really phenomenal team within those three departments which they essentially make me look like.

00:07:41: i know what im doing.

00:07:42: so it's an honour work with all of them.

00:07:45: It is some of their best minds that are definitely there.

00:07:51: Well, the results show that and we get to have you here.

00:07:54: So You can talk about just a few of these things And I guess each of this and you ever appeared on many podcasts already going into detail in Each or some of these These fields.

00:08:06: today We want to talk about data and how especially?

00:08:10: I guess That's the occasion you mean.

00:08:13: Data is your everyday life and using it on race day in preparation The occasion of why you're here is that you built an entire team, basically from scratch on a computer.

00:08:27: Kind off like the movie Moneyball shows which I think it's excellent movie.

00:08:33: and you did that for cycling where yeah You just use numbers instead of gut feeling And then traditional scouting methods to find performance athletes potential in especially young riders.

00:08:48: Yeah, so we still had... We kind of combined the two.

00:08:52: So we did have gut feeling and scouting.

00:08:56: that brought the daughter to us.

00:08:58: And then from the daughter side obviously That's really the final decisions were made.

00:09:02: It was a collaboration between the two But how much labour-intensive The daughter side is definitely I would say Which the gentleman i worked with you know, he brought the human element of hey look at this kid as an example.

00:09:20: and then I kind of took all the data.

00:09:23: And we made a final decision in end-of-the day.

00:09:25: but it was excessively data and labor intensive.

00:09:30: Is that... Was there obvious next step to do?

00:09:34: Like natural conclusion on where things are or is it a little bit of leap technologically and risk making attempt like that Because data is already very present.

00:09:48: It's not entirely new, but building it at that scale was outrageous and risky or just a natural progression?

00:09:59: I think there was an actual progression you know.

00:10:01: yes You always had access to athletes.

00:10:04: doctor.

00:10:05: That's how you would look at an athlete And you'd think Yes he would fit in Or no He wouldn't.

00:10:09: Whatever the case maybe Or may be.

00:10:11: Not yet But i do feel What we have done.

00:10:15: We had kind of delved a lot deeper into it because that was the first year project and we simply have nothing.

00:10:22: You know, It wasn't that you know?

00:10:24: We had our system in place already Especially going to be under twenty three category which is not The same category as what I was two years ago like two or three years ago.

00:10:34: There's still a junior category where now your under twenty-three categories essentially will turn without the death.

00:10:42: So those decisions actually needed make And especially where you're looking for these young talents to know how much further, You can get out of them in the future.

00:10:52: That's when that deep dive really came into it In terms of data analysis and data dump if we want to call at that.

00:11:02: There are other teams doing very good work but I felt what were done because We didn't have a luxury on having this system in place step forward and a little bit ahead of what your conventional scouting in DotA dumps way.

00:11:20: So welcome opportunity to start the process now also, yeah?

00:11:26: And so you already said...so do Scouting first!

00:11:30: You see people who seem show potential yes then go into there.

00:11:36: would it make sense with other round like starting clean slate just get bunch data on different athletes and then kind of filter out the most promising candidates?

00:11:49: or are there too many factors like politics?

00:11:52: And can those people even available, would that not make any sense.

00:11:57: So

00:11:57: the answer is yes because we did run a... Project last year on the sideline and that was very much like that.

00:12:04: So we got daughter.

00:12:05: first We analyze the dots, and then we would go to a discussion or go Hey You know this guy is good writer A B&C.

00:12:11: let's go to refer the discussions.

00:12:12: so you kind of reverse periodized Periodize it in away But you would need to get access First.

00:12:21: all that right to make those informed decisions.

00:12:24: It's not just say that we saw Rider a riding racing and then we think he's good.

00:12:29: We approach him when you get the doctor where on this project that we ran last year in A few other years was kind of doctor first based on one or two testing, but we never physically Met the rider itself.

00:12:41: so we get to dark And then we would bring it in closer.

00:12:44: So as I said is more like a reverse periodizer Periodized approach

00:12:48: interesting and then you probably with your name As a coach and your name has a team have no issue acquiring the talent Once you've shown interest.

00:12:57: Yeah, people are not really saying no if you knock on the door It's a luxury we have which uh... We're really grateful for.

00:13:07: Hype or Helpful?

00:13:09: I want to play Hype Or Helpful with You and give you A statement And then Want To Hear Your Opinion On it.

00:13:19: That actually holds up.

00:13:20: No that is entirely Incorrect.

00:13:22: So The First Statement Will Be The More Data The Better.

00:13:26: It's a matter of quantity rather than quality or the algorithm.

00:13:31: Algorithm is important, you can't deny that but more is nice.

00:13:36: as hypo You want quality at end-of-the day.

00:13:40: Okay go in depth.

00:13:41: this one might be more interesting... ...you can predict race results accurately without ever seeing a race from training data alone.

00:13:51: So hold up.

00:13:53: There is truth in it.

00:13:54: I will say The only reason.

00:13:58: I say not entirely because of how stochastic cycling is in nature.

00:14:06: So it's not a hundred meter race, you know?

00:14:10: A hundred meters.

00:14:11: she could pretty much predict that as what it is based off data.

00:14:14: the guy has good day he has good legs and run whatever under nine seconds or whatever it is.

00:14:19: Cycling doesn't work like this Because you have outside factors although is used from a doctor thing, you have weather.

00:14:26: You have Tomac or someone crashing

00:14:30: etc.,

00:14:30: so there it comes down little bit to the on paper yes and you don't race on paper but It does has very good positive influence in potential race outcome.

00:14:44: With sophisticated data driven approach you can take human decision making out of equation.

00:14:51: Is that thinkable?

00:14:53: Well,

00:14:54: maybe in the future

00:14:55: may be in the feature not now.

00:14:57: I don't think you're making taking human side then You would need to.

00:15:05: Then the machine has to have your brain for now.

00:15:08: currently but ii think it will get To a stronger point than where It is at The moment but I don't think it's there yet.

00:15:17: Probably also question of performance level, at an amateur level having somebody you already said that have this automatically generated program might do something-

00:15:27: Correct!

00:15:28: At the world level that your operating at.

00:15:31: its not their yet.

00:15:33: Like programs are being downloadable off the internet for years and years.

00:15:38: so i can download a six week programme to this race.

00:15:45: But it's more than that, especially the higher level you go.

00:15:49: It becomes a bit more fine-tuned in that and the specificity is a bit stronger.

00:15:55: Hyper helpful!

00:15:57: The superpower of data driven approach as you can identify hidden potential for growth rather then riders are already strong but overlooked?

00:16:07: Yes I'd say there's definitely a strong component to this because that essentially what we want do.

00:16:12: so the answer would be The hype may be not a capital H, but definitely fact.

00:16:18: It's the potential.

00:16:20: no strong rider is overlooked because nobody checked the data?

00:16:24: Correct This one maybe about controversial large language models are key to success in future.

00:16:31: AI models can predict performance better than more traditional methods.

00:16:35: Definitely not now at that moment Because there you need teacher to do stuff.

00:16:43: So because some staff, so the answer is yeah I would say it can't or It could help you be quicker.

00:16:51: making those decisions Is what i would say?

00:16:54: I don't think yet.

00:17:00: Yes and no im on the fence on that Because You Can get really...you can fast track What u wanna Do by using AI.

00:17:09: However, some of the stuff that you get back currently is like why?

00:17:13: You need to learn a few bit more things.

00:17:15: so

00:17:17: And I feel like you're already ahead off.

00:17:18: Like your kind of head-of-the AI thing with your very sophisticated specific machine learning models.

00:17:26: Yeah

00:17:26: and actually do see there quite a bit.

00:17:28: an is a good question.

00:17:30: now people think large language models are the thing and then they try to throw it at everything.

00:17:36: We have so many clients actually do that,

00:17:39: can't

00:17:40: we solve any problem with that?

00:17:42: And you're like well this is not solving your problems.

00:17:48: but as I said fine-tunes on fast tracks definitely in other sports maybe like a baseball which statistic value, you know like I know in terms of or football where American Football where it's running yards and stuff.

00:18:08: Like that i think maybe in there definitely um... And I'm sure they are super far ahead on that.

00:18:16: just your cycling..I don't think yet to the but I definitely won't say

00:18:22: Yeah, well I like that.

00:18:22: you kind of helped me on the LLMs are not everything because there is as always.

00:18:29: i've seen a few waves off.

00:18:31: There was the blockchain bitcoin thing.

00:18:33: That was good two years and then everybody tried to find a problem they could solve with that solution.

00:18:38: Yes And now it's AI which is making amazing results but its just not the solution.

00:18:47: It's tool in chain

00:18:49: But also maybe it's just me Maybe I wouldn't trust it a hundred percent yet.

00:18:55: Like full-on trust.

00:18:57: so this week?

00:18:58: Certainly not, especially when looking things up.

00:19:01: i feel like that's the retrieval is still.

00:19:04: if its specific day It's general concepts can do very well.

00:19:07: make specific performance values maybe yeah which start making things up?

00:19:13: Yeah correct

00:19:17: What is what?

00:19:17: Is something that you look at?

00:19:19: maybe to start with a simple matrix obviously writing performance.

00:19:24: What does the data points, and what are qualities within those that you're interested in?

00:19:31: It's a lot because you have different profile of riders.

00:19:34: So whether you see someone as a sprinter You have certain benchmarks.

00:19:39: If somebody like a classics rider or if they were stage racers an example we had certain benchmarks.

00:19:46: we would compare them to.

00:19:49: But also at the same time, you want to understand is just because they don't have a benchmark today it does mean that not good enough.

00:19:57: You wanna look how did I get through and what do think can be out of for coming years?

00:20:04: And within those benchmarks on what we see cause you may see young kid who has phenomenal data in every benchmark tick green boom.

00:20:16: But in terms of when you look at their overall data, they may be literally at the limit or the physiological limit where there are as a human being.

00:20:24: So you're not going to get much out of them In the next one two three four five ten years potentially You know.

00:20:30: so that is something That um it's really important tonight.

00:20:33: So

00:20:33: predicting The headroom Is really what the game and I guess correcting for all these other factors.

00:20:40: How do you build a team?

00:20:41: Do you have certain roles in mind and look specifically for the role, maybe what are those roles?

00:20:48: or is it more that you find the best riders then fit them into categories.

00:20:55: It's a little bit of both.

00:20:56: so you want to build complete teams.

00:21:00: everybody wants to win the Tour de France its kind standard.

00:21:03: So if someone could potentially see being competitive there at the same time in a team, you need a sprinter.

00:21:12: You need like two domestics?

00:21:14: You need mountain support to... There are different elements of what you need so you can't just go well these are five best climbers and stage racers then take them into their teams.

00:21:28: What we did last year was identified where we needed it.

00:21:31: So that's when we could win within our own team And also going onto World Tour Team.

00:21:39: what do we need for the World Tour team in maybe two, three or four five years?

00:21:42: and does this kid follow that progressional pathway all the way into the world tour.

00:21:47: So it really depends on What you need in the team?

00:21:52: And as I said with the rookies last year We had nothing like literally read nothing.

00:21:58: so just building a complete team You know from workers to winners and everything in between.

00:22:02: That

00:22:03: then you try yeah Basically tried to fill a role years from now And have the right talent already working up to that.

00:22:12: Are there many surprises?

00:22:13: Do you find people and they think their one type, then like no your a sprinter or does it change over time also?

00:22:22: It does because what you got to remember is as young kids.

00:22:25: so whether they are seventeen, eighteen, nineteen years old typically not fully developed yet So don't really know where will develop You have good idea But sometimes they do go for it's not that.

00:22:38: They're going to go from an eighty kilogram sprinter to a fifty five kilogram climber, you know?

00:22:43: It's uh... That's not on that but he may be good in juniors at sprinting

00:22:50: etc.,

00:22:50: but then as he develops He maybe come like a sprinter.

00:22:55: thats good In media mountains ,but can win form the small group As opposed to big jump You know.

00:23:00: so that development does Does change and you need to adapt it that accordingly?

00:23:07: So, It's not a sure thing.

00:23:08: when you look at the dot in your car You know rider is definitely our sprinter.

00:23:11: And he's gonna be the next chipolini, so doesn't always work That way.

00:23:15: well

00:23:16: life happens or correct things changed attitudes motivation

00:23:20: they get girlfriends

00:23:24: The most destructive factor.

00:23:33: Let's look at the different data points because we already touched on a few.

00:23:37: There are very easy and obvious ones, you have power meter on your bike or heart rate sensor that is automatically collected through our data.

00:23:46: what some of those classics do?

00:23:51: As I said there were benchmarks in that.

00:23:53: so whether it has powers one definitely you want to track weight over time track progression in power, you want to see how what their heart rate does.

00:24:06: You also want to track wellness.

00:24:08: so is a rider getting sick?

00:24:10: Is it not too often that he's getting sick and need to understand why?

00:24:14: on that you will pick up from the data.

00:24:18: then we also have benchmarks being whether its like your climbing speeds or talk recruitment values, durability etc... We extract all of this stuff out out of the data and then we create a profile of the rider according to that.

00:24:34: A rider getting sick more frequently would probably be a sign of overtraining or could just be the person's individual?

00:24:42: Yeah, it could anything.

00:24:43: It could be simple that his recovery is not good enough.

00:24:47: but he has nutritional intake isn't correct.

00:24:53: he's got bad habits, you know something as simple that or yes it can also be over training.

00:24:59: You have a young kid doing whatever twenty-five hours per week but its really not sustainable and healthy approach to his training.

00:25:09: And that would get me into the, I guess softer markers and more tricky things to measure.

00:25:16: Because there's like self-reporting or it just not as exact as a power meter?

00:25:22: How do you... Is there lot of self reporting?

00:25:25: what are some maybe surprising factors that look at?

00:25:29: Do we look at mood or sleep will be an easier one outside of actual racing?

00:25:36: What is somethings going through?

00:25:39: Sleep is one, definitely on how much quality sleep they're getting.

00:25:42: Mood is another one.

00:25:44: that's typically linked to their sleep profile.

00:25:48: Stress is another and your weight and heart rate variability are really good.

00:25:55: but you know white.

00:25:56: you can also track if a very stochastic in the weight the best health or best recovery.

00:26:05: But as I said, if it's sleep mood stress values is really good.

00:26:09: This like secondary and tertiary things Like.

00:26:12: obviously you will get a three D profile of course And wind data Or weather data The heat of day Yes All that plays a role to judge the performance.

00:26:28: There You would go.

00:26:29: You know if you want to extract those external things, we also just on the writer profile.

00:26:35: The data weed extract is also... On their website as an example, pro cycling stats?

00:26:41: You literally have everything of the writers kind-of race results

00:26:45: etc.,

00:26:46: so we can extract that at certain times in a year and understand why he did or did not what's happening with regards his overall performances.

00:26:56: but you know, stuff like weather is really important for our stages especially for sporting directors.

00:27:03: You know in terms of having plans and tactical and etc.

00:27:06: performance approaches to the races.

00:27:08: so whether that's just temperature or humidity... Whether it rains no thunder everything else gets thrown at you.

00:27:17: That all comes into effect For everyday racing And thats very important Especially as I say to our sporting directors.

00:27:24: So Thats also a large volume of data that we do grab.

00:27:29: Nutrition and hydration, obviously key factor.

00:27:32: is there a better way than standing with the clipboard?

00:27:35: And writing down one gel two liters of water?

00:27:39: Yeah Do you have something that solves our problem?

00:27:42: yeah so We Have A Software System That Tracks All That With The Riders And It Also Gives Them At The Same Time Their Meal Plants You Know.

00:27:53: So Whatever whether it's X amount of carbs x proteins x fats, you know Whether that's breakfast lunch and dinner And then also feeding on the bike depending on this stage.

00:28:04: Okay?

00:28:05: Yeah is there?

00:28:06: because I remember from like The two-hour record attempt in the marathon by Elliot Kipchage.

00:28:13: They have like while they were racing he would drink and Then throw away the bottle.

00:28:18: somebody would pick up the bottle measure It and then they would calculate how much carb and water mix needs to be in it for the next bottle to drink.

00:28:29: Is that level you do at every race?

00:28:31: Or is during a race more by feel, they know what their plan is or I guess its always more carbs than they need take-in?

00:28:40: That's

00:28:40: not a point!

00:28:43: So yes so... What you described as an answer was Yes They have their nutrition per stage But nobody's throwing a bottle and no one is collecting the bottles to you thing, but all those bottles are there for the day.

00:28:56: And they know exactly how much they need to be having typically grams per hour.

00:29:01: uh...for their race in that discounted accordingly on that is updated at the end of this stage For dinner for the following days stage At The Same Time No one's running behind the cyclist collecting bottles.

00:29:16: It would be yeah, it might be excessive

00:29:19: guy.

00:29:20: if he is Yeah

00:29:21: And that's a lot of bottles then

00:29:23: and you also need to fight with a lot Of spectators on the road.

00:29:26: I guess that's the trophy everybody wants out right?

00:29:29: So worth more than gold

00:29:30: Is like his self-cleaning system isn't it?

00:29:33: Like you don't need a cleaning maybe for the gels or

00:29:37: something stuff but they're off kind of waste zones now that are allocated during stages.

00:29:43: So you can only really throw away in those, those zones especially packets and net.

00:29:48: but if your on a mountain or something and you throw the bottle off it's yeah It still happens might.

00:29:55: Yeah

00:29:55: has Your approach to data And your desire To have as much information at The highest quality Of information changed?

00:30:04: The training Approach like did it feedback That You're trying to make adjustments better or more data?

00:30:11: Yeah, our data definitely has changed your coaching prescription.

00:30:17: I would say from my side personally and the monitoring side to understand what that rider is done.

00:30:22: then why do we need it next day on a couple of days training-wise so you know... Is there load and stimulus in the correct path ?

00:30:30: Or does it not ?

00:30:31: Do We Need To Back Off ?

00:30:32: Do You Need To Add More Load Etc?

00:30:36: But you are definately as As dietary has become more freely available or more available, your coaching definitely has become I would say specific to that specific rider.

00:30:50: You know it's not like there is one coach for the whole team and you can coach everyone essentially the same Like all the sprinters are the same All of the climbers they're the same.

00:30:58: So yeah its definitely helped And elevated The game For More precision.

00:31:05: in terms Of what u do.

00:31:10: We talked about quality earlier regarding quality for the data.

00:31:16: When I look at my sports watch or your sports watch, what any amateur can do?

00:31:21: How far away how much more sophisticated is that you're collecting?

00:31:27: Is it far off and not even comparable?

00:31:31: Or is within reach?

00:31:33: would any data that i'm getting on my way to work every day be good for prediction.

00:31:41: Yeah, so like that's... So I think where there is a ying and the yang with it is you can collect everything i collect You can theoretically have available to you And you can collected.

00:31:54: obviously we mentioned him earlier Florian Lipovitz' daughter definitely not but what he gives me You know, those things are super freely available.

00:32:12: It's not that our palm each is any more special to what you buy in the shops?

00:32:16: Maybe they're more accurate because of our calibration system.

00:32:19: so we have.

00:32:20: but it's not That We Get Something Special Out Of It That You Don't!

00:32:24: Its What We Do With That Data That Will Separate Us From The Guy In The Street.

00:32:30: So let's look at briefly At The Blocks That Were Discussing.

00:32:35: First Block Data Collection.

00:32:39: key takeaways You have to get all the data together.

00:32:44: quality anything else where you would say in data collection Profiling a writer.

00:32:51: Those are those other The important factors.

00:32:55: Yeah, so like what's really important is getting good data.

00:32:59: So I'd like but when I call clean dot I don't know if that's a thing and the dots hold But I called it clean daughter.

00:33:05: Where?

00:33:08: those decisions when you don't get accurate data, which is sometimes the case kind of often like if you as an example just on a head unit.

00:33:18: If there's maybe water damage or something your daughter that she gets wrong.

00:33:22: and a lot of time some athletes don't all coaches whatever they clean up.

00:33:29: so look at her ID and see really bad numbers are using.

00:33:33: good numbers can identify It's right or wrong, it's high low whatever the case may be.

00:33:41: You kind of make a false assessment on the rider and you either go I he is not good in your letter talent disappear.

00:33:49: all you guys phenomenal?

00:33:52: And then get like bit of donkey.

00:33:55: so it some in terms that its really important to look at this.

00:34:01: Ninety nine percent actually know i don't believe will be one hundred percent for couple factors but You know, ninety percent is also good enough.

00:34:09: So that to me it's really important on how you receive and what you get.

00:34:16: Biases in blind spots Do have any examples of maybe biases?

00:34:23: Metrics that you looked at... ...that didn't matter as much or were unreliable because the data just isn't there or quality is not good?

00:34:33: Yeah, definitely.

00:34:34: You always want to improve that in terms of what metrics you look at.

00:34:38: so as I call a cleanup also with how cycling is progressing and what are called modern cycling today's term which probably last five years coming forward.

00:34:50: till now it changes constantly.

00:34:54: The reason i say this we did research published Well, I think twenty-twenty two or twenty three.

00:35:03: I won't tell the other guys.

00:35:04: I forgot that and last year we realized it was completely null and void from a new research that we did on literally the same topic.

00:35:13: so there are definitely metrics and data points that i wouldn't say not valid because everything is valid but not as of an importance of what house cycling has changed.

00:35:26: So as it changes you look at different things.

00:35:30: I'm just not going to say what those metrics are in case other teams is still looking at them.

00:35:38: But definitely, as it evolves and even dietary evolves you have to evolve with that the same time then be up-to-date

00:35:46: and maybe the other way around.

00:35:47: Other things that you find very... And I understand if you might not be able to give away specific things, but there are a few times just attitude and mood like personal life at level of performance playing big role.

00:36:07: Other thing from Moneyball movie that the scouts were talking a lot about like personal stuff.

00:36:13: Like, oh this guy has... Oh he smokes or he has personal issues?

00:36:18: He gets angry or something other.

00:36:21: many surprises you had along the way where you discovered all of these have surprising impact.

00:36:27: Yes because the peloton is getting younger and young enough.

00:36:30: You don't have guys winning grand tours at thirty years old Or twenty seven years old That much anymore Where that was sort of considered your kind of peak endurance years or whatever and you would win a grand tour.

00:36:43: all the classics in that we now.

00:36:46: You know, you have an nineteen year old.

00:36:48: if he take Paul Seishas is going to the Tour de France.

00:36:51: nine these you know expected to win so to speak will perform super well.

00:36:56: but what has changed with that?

00:36:58: If really take which I said on moneyball as you're not dealing Nineteen years old, you saw the child or seventeen years old but in a cycling environment.

00:37:12: You want him to be your seasoned professional all professional But when he leaves that you still wanting it to be a kid?

00:37:18: You know.

00:37:18: otherwise he misses out on a lot and That is kind of really important.

00:37:24: Point that you are now dealing with is at you're dealing With this guy.

00:37:28: you going well, you know you paying him x which Is typically a lot of money do need them to be as this mature adults but he's still a kid.

00:37:37: Hasn't had his heart broken, hasn't gotten drunk whatever the case may be and that comes with its own fair share of obstacles you now need to navigate in professional cycling and believe it or not is quite data controlled.

00:37:53: yes we have the personality there are trends that we track within terms

00:38:01: We have a few little games that we want to play or repeating segments, and the first one will be this-or-that where you wanna get your opinion.

00:38:12: Which do you prefer if had choose one?

00:38:16: The first would quality over quantity which is always interesting with regards data Would rather go more in depth than higher Quality of the range of data that you already have.

00:38:30: always quality over quantity

00:38:32: Okay.

00:38:33: Yeah, and that means your already Have just so much or?

00:38:37: Just because it's much more relevant

00:38:39: It's more relevant.

00:38:40: So you can have a whole lot of data.

00:38:41: That essentially does nothing for you just Because you have a high volume of it where you want to do quality data at The end of the day.

00:38:49: You know if I give an example you Can have A huge like whatever three terabytes of an athlete Data whether it's training files as an example, you know You can have six years of data But you only want to take out that quality of data.

00:39:05: That's in itself at the end of the day.

00:39:07: You want quality over quantity because you wanna make informed decisions which are pretty important especially at In performance environment

00:39:16: and your level.

00:39:18: yeah obviously It's like the margin for error is very small okay?

00:39:21: Would you rather do faster development cycles or have more time per development cycle to go in depth and perfect things, so higher iteration speed over getting things exactly right.

00:39:36: Can I choose both of those?

00:39:39: Not in this game!

00:39:42: Okay run there find me again So

00:39:45: you can iterate faster.

00:39:48: You could try out a thing And then just run the new thing every day versus having a metric and just really go into the detail, have a long cycle.

00:39:59: I guess in your field it's like there is season.

00:40:02: maybe training cycle.

00:40:04: would you rather have them shorter?

00:40:06: And do more or Do you wish you had more time for seasons?

00:40:13: Yeah You always want more time at the end of the day because if you have time yeah The quality that she can do so you don't rush to things.

00:40:20: So If you didn't have that luxury you wanted to be done as quick as possible with that final outcome.

00:40:26: And personally, would you like more days in the week or longer days?

00:40:31: Longer days is what I would like

00:40:33: so You can get an extra training session in Or Get some sleep.

00:40:37: Yeah

00:40:37: correct.

00:40:37: So by Monday afternoon i'm not already doing Tuesday's work So I could finish Mondays' Work and finish off on then at Tuesdays.

00:40:45: I'd rather have longer days than Extra Days.

00:40:50: Okay, is there a big difference throughout the week for you or?

00:40:53: Is it just Training days and then training cycles.

00:40:56: Yeah

00:40:56: every day's essentially the same Monday to Sunday.

00:41:00: It that's not death.

00:41:01: You know.

00:41:01: I know some people go home on a Friday And they go to The Beach and stuff.

00:41:05: i may Go To The beach but That's Not like that Thing you Know.

00:41:09: so you could go to the Beach Equally

00:41:12: Each Day.

00:41:12: Correct

00:41:13: if you only had one metric Or even one value, one number that

00:41:18: you

00:41:19: needed to use to predict if a rider is going be success or not.

00:41:25: Do you know what the number would be?

00:41:28: If it's one metric... It's a tough one.

00:41:32: I will go with the compound score as well.

00:41:39: So the compound scores are a bunch of metrics.

00:41:42: I feel like that's almost cheating on this answer.

00:41:48: Yeah, so it's a value you know?

00:41:51: It is different per rider.

00:41:53: So its not just...you have the number so whatever Its two thousand and that determines success.

00:42:02: You know..it based around specific rider weight, rider type etc.

00:42:06: So it isn't as easy or simple thing.

00:42:09: but Yeah, it's a value that you do in the calculation.

00:42:18: That is pretty much what I would do if we put a gun to my head and use the compound score.

00:42:24: That's your go-to?

00:42:25: Yes!

00:42:25: One more question to jump all of way back into the start collecting data because as you said there are these very young riders who compare them.

00:42:35: How does that I assume?

00:42:37: That's going to be tricky.

00:42:38: how do you compare riders at different ages, which i think You have to then correct and benchmark differently depending on their age.

00:42:47: um so there it comes down To sort of identification or what sort of rider type u Uc in them, you know are they a sprinter climber etc.

00:42:57: Classics rider?

00:42:58: And then you would group them accordingly.

00:43:01: But that job is for our expert scouting groups, so that's Tim and Michael.

00:43:06: they need to identify that... ...and when they bring it right in say hey here is Michael Smith.

00:43:13: we see him as a GC climber or classic guy.

00:43:19: then we extract data and look at the data.

00:43:21: Then, you know where they are in terms of benchmarks and processes?

00:43:24: Do have a lot of historic data already that you could compare against.

00:43:29: I don't how long data has been collected at this level.

00:43:32: if can go back to the old guys see what were doing like when there was seventeen.

00:43:40: Yeah so you can.

00:43:41: So I worked closely with an example.

00:43:47: That dot has kind of changed over time with especially with the young kids.

00:43:52: The twenty-twenty two, a winner probably their second biggest amateur race in the world.

00:44:04: he won it and you wanted super convincingly like really dominant performance.

00:44:09: He wouldn't have finished... You would've been lucky to finish on top ten to twelve at last year's Baby Jiro.

00:44:16: so You know, from that side you kind of see how things have developed.

00:44:20: So to go oh we have this rider who's a pro now but he was an amateur ten years ago is completely null and void.

00:44:30: it's not even the same category anymore.

00:44:33: so as I said that research study we did at the time really based on under twenty three there was phenomenal data kind of you're looking at that a lower age group category now, which isn't even really relevant.

00:44:51: And that's within the two-three year span.

00:44:54: The overall performance is

00:44:56: just going through the roof and because of accessible data, accessible metrics

00:45:03: etc.,

00:45:05: so all these kids have... All their toys available where only few had them before.

00:45:10: Coaching is better now because of that.

00:45:12: Because you have that data acquisition,

00:45:14: etc.,

00:45:15: so you can make informed decisions

00:45:17: with regards to age.

00:45:18: Is it just the performance and success pressure?

00:45:22: That makes you want to and I see that in other sports with very young football players Already getting into the pro teams yes at a performance pressure mostly or is it also that you Can predict better who is worth investing in?

00:45:38: Does that play a role at all?

00:45:41: Yes, there's always performance pressure.

00:45:42: You know any professional sport is high-pressure and because you kind of grabbing these kids or skating and identifying them younger and younger... That pressure for them to perform is immediate now!

00:46:01: It's tricky with young kids but as we said earlier is to go, well this is what we see today.

00:46:13: But do we see that in two three four five years?

00:46:16: You know and the kids are being signed for two Three Four Five Years at a time And their pressure does come.

00:46:22: so it's quite how'd you call it.

00:46:26: It's a balancing act.

00:46:28: Do feel like there might be cut off where you just can't predict anymore because There too much volatility.

00:46:35: Definitely

00:46:38: super young.

00:46:38: so I still remember school times.

00:46:41: It's not that long ago, i'm getting old and this light is no doing me any favours.

00:46:48: Is there a cut off at some point where you just the volatility gets too high?

00:46:52: And then to much happens between I guess the ages of fourteen fifteen sixteen.

00:47:00: Yeah, definitely.

00:47:01: You know there's been some young kids that have being scouted That were predicted.

00:47:05: you know they've been six year contracts and They remain to go on and be super successful.

00:47:10: in after two years it fell apart unfortunately for whatever reason.

00:47:15: but then you Know the one.

00:47:17: when got a young guy that we started last year He was like he is okay nothing really on paper daughter Nothing to write home about And he has excelled exactly from where, you know in a very short space of time.

00:47:30: We gave him a good environment.

00:47:32: we give him what he needs and use really jump jumped over the fence so to say I can.

00:47:38: as you predicted

00:47:39: probably

00:47:42: We knew that he was going to be good.

00:47:44: There is no doubt provided we gave him the platform and this support, but didn't know He'd Be That Good.

00:47:50: You Know so.

00:47:51: But then we've had guys who were like hey This guy's gonna be phenomenal And unfortunately in two years that that Was it?

00:47:57: So I don't believe No matter how much Daughter you have or clever brains that she can predict that far into The future.

00:48:04: especially with cycling.

00:48:06: you know It's just too stochastic In nature and its Just Too Tough Currently

00:48:11: Even outside of I want to say like life events or injury.

00:48:15: Anybody can break, especially on fast bikes.

00:48:19: you could have a major accident and then be thrown off.

00:48:22: but even outside these big somebody fell off the wagon.

00:48:28: there's still varieties.

00:48:30: that fifty-fifty opportunity is too strong.

00:48:34: i won't say maybe today and tomorrow with how like ITN stuff is changing.

00:48:42: You know, maybe there you can get more of an inaccurate trajectory providing everything goes well in a perfect world.

00:48:52: but the one thing that you do need to remember on cycling it's like... Everything on paper looks amazing But you don't race on paper.

00:49:00: That's where The Winecatch comes in I think to predict and exact trajectory To-the-Point Providing he stays healthy doesn't crash, etc.

00:49:09: I think currently it's you can do It.

00:49:14: You Can Get A Very Good Trajecturing Because We Do Have That But Its Definitely Not A Set In Stone

00:49:22: Racing What?

00:49:22: We're Going To Quote That Racing.

00:49:25: The Racers Are No One On Paper

00:49:26: Yes So You Look At Riders On Paper Essentially But You Don'T Race On Paper

00:49:33: Attitude And They'Re Still Humans.

00:49:36: I want to get into decision-making and calls for action.

00:49:42: What level do you have?

00:49:45: automatic decision making that is entirely data driven yet?

00:49:49: Or, Do You Get Input and Inside And Then You Make the Decision Probably Together with The Writer?

00:49:56: Is There A Level At Which You Can Just Output Instructions still very much that the human needs to stay in the loop and aid in decision making.

00:50:10: Is that on race performance or as an example hiring a talent, or scouting a talent?

00:50:17: Well actually both would be interesting.

00:50:18: maybe hiring a talented is like one of decisions.

00:50:22: so I guess you want everybody.

00:50:25: look at it then supported by data I assume.

00:50:29: So the data part is really, really important and you get very good information on it.

00:50:35: as i said earlier with if your have benchmarks then what are looking at?

00:50:40: or a team has specifics?

00:50:45: that's really important.

00:50:48: however there isn't an element because its still sport in human body connected to.

00:50:54: You do need to make a decision based on the human side.

00:50:58: So you could have... The best data, everything is great and the guy's just unbelievably difficult to work with isn't friends with anyone or whatever?

00:51:09: Then decide okay well did we win races but he would just have constant fight for them?

00:51:14: Or do step aside look out from personality which as a human draw card were not going that route.

00:51:22: You know, an example I had.

00:51:25: I worked with a rider who phenomenal guy and um Phenomenal daughter like he should have won multiple grand tours And all the physiology or the doctor everything showed at The testing dot in that you put him on the road?

00:51:40: He just couldn't deliver.

00:51:43: So which was super I wouldn't say Super strange because he did deliver but not to the potential That our daughter showed.

00:51:51: We're going to go with him anyway, you know?

00:51:54: Because we know he can do something.

00:51:55: But... He definitely boxed under his weight category.

00:52:00: put it that way

00:52:01: Do have a guess.

00:52:02: what the factor would be is an attitude motivation.

00:52:05: I mean race like sports psychology at your level where everybody's going at human limits probably plays very big role.

00:52:14: Yeah i think kind of kind of buckled under big match temperament, personally.

00:52:19: You know when you took that grantor aspect away from him and he told them to go stage hunt at a grantor who'd win stages because the pressure was significantly less so where that...you can track it with data from a psychological point by stuff they do and you can get trends in processes.

00:52:41: but Yeah, he was a perfect example of it.

00:52:45: You know when you looked at him on paper your like mate we're gonna make me and win races.

00:52:50: And yeah He just didn't deliver on race day.

00:52:55: Perfect segue because I would like to get further into how you doing something with the data and what?

00:53:02: It would be interesting.

00:53:05: Maybe first who gets access to the data?

00:53:08: do you does every rider get their report Or maybe even action already from the data automatically, or do you get to look at it?

00:53:17: And does it take your experience to make a call-to-action reason on what is presented and then have adjustments made.

00:53:29: who was involved in this cycle.

00:53:31: So Ryder has access to his data.

00:53:33: so if we called His access is freely available to training files, etc.

00:53:38: So whatever he does on the bike or off-the-bike He has full access to it.

00:53:43: essentially its his data You know.

00:53:44: so we can't stop him looking at that dot and we never will.

00:53:48: It's what?

00:53:49: The coach.

00:53:50: if we take a what I take Florian's daughter I analyze her daughter On my own And then i would report back To Him If need be on.

00:53:59: hey this Is A B O C Or D E N F. If it's a report, we do have obviously reporting but that is typically for management and stuff like that.

00:54:11: We do obviously report back to riders at the same time.

00:54:13: so It's not that they have the ability to analyze To the degree that we do or create reports That we do because we have that on this On a separate software system.

00:54:25: But they have access to you know all their sleep data there wellness data training files erase files etc.

00:54:32: How interested are they?

00:54:34: I guess it might be different because is an interesting sport.

00:54:37: And many colleagues, I feel like techie people also bike riding and the data collection's fun when you look at them instead of getting better sleep or doing the obvious things that you tweak to get little things in your training.

00:54:51: how interested...how involved are athletes?

00:54:55: Some are very involved!

00:54:58: You say oh that was really good.

00:55:00: then thanks That's where it ends, but then some you really do go into deep dive because they have an interest or a passion and super involved in that.

00:55:10: And if I trust your feel good on the bike so don't need to worry about myself sort of takes an element away from them.

00:55:21: So it really throws between.

00:55:25: Do you feel like it makes a difference?

00:55:26: Like, do people who care about the data maybe do a little bit better because they...do their own reasoning?

00:55:33: or is that actually better to just not care as much and focus on- On the ride.

00:55:39: It's both ways.

00:55:40: cause sometimes You know..you want athlete understand what he's doing Just from training prescription And then he gets He does this session and say This why we did this?

00:55:52: However, at the same time some of them that are super over involved.

00:55:55: You know their brains got a hundred thousand kilometers an hour.

00:55:58: There's a hundredth thousand different questions and it sometimes works backwards against.

00:56:03: then because today they did five hundred watts tomorrow They do four ninety And there's a big problem.

00:56:09: But they don't understand it.

00:56:10: that how your sleep school?

00:56:11: you didn't sleep last night?

00:56:12: Do had to fight with your girlfriend or whatever.

00:56:14: so there's an outside variable but Then they only fixated on that one performance.

00:56:21: So sometimes it's a bit of game with tennis, you just need to manage the personality correctly.

00:56:27: Do have an issue interpreting or do feel like fall into looking too close at data?

00:56:34: I could see that as you said there is different factors maybe go up to altitude which will be big impact for few days.

00:56:42: how do prevent being too closed incorrectly interpreting the data, because a data is right.

00:56:52: What you take away from it might not be correct

00:56:54: and then that's what you do with it.

00:56:55: so if I take myself i have kind of certain metrics that I look at within the window?

00:57:03: And yes, I have some that I don't look on very sped like every day basis If we want to call it That.

00:57:09: So I Have sort Of Like A Second Data Range Of What I Look At But which it is, we had spoken earlier that there's so much information available you do get lost in translation.

00:57:24: An example I had with a coach the other days.

00:57:26: he was a writer who didn't perform one hundred percent but really went off on something completely else and did this huge deep dive into data... Which actually very good work!

00:57:41: kind of he went so deep that he missed the point off exactly where the problem was.

00:57:46: You know, and it's just well let us go backwards to see if we can move forward.

00:57:50: but here is already three weeks up there on very good data.

00:57:54: But That way you get lost in translation sometimes.

00:57:58: Okay hit these biases maybe blind spots.

00:58:01: Yeah

00:58:01: correct We could talk about.

00:58:02: this will be interesting a little bit later.

00:58:07: And you already said like the having people be able to do analysis, how do you see technology in your field?

00:58:14: Maybe even just personally Do feel it's better if we get out of there way.

00:58:19: I want results.

00:58:20: Do have fun playing with tools.

00:58:23: where did u land on that?

00:58:25: I love playing with tool.

00:58:26: i'll be honest On that side, as I said before myself.

00:58:38: I do have a cut-off point where you don't want to kind of go so deep down the rabbit hole because there is so much freely available.

00:58:46: but for me...I really like their software tools.

00:58:51: Our doctor scientist on team at the moment works super close with them and he's creating phenomenal stuff in the background And i'm like kid in candy shop with it.

00:59:04: It is good, but again it's.

00:59:06: you need to know what you're looking at.

00:59:08: Does that make a decision?

00:59:10: Can you make an informed decision of it?

00:59:12: can You understand What is going on?

00:59:13: if the answer yes then you go forward.

00:59:15: But as I said some people do so much available That they get lost and thats where i think becomes dangerous in terms Of too Much data.

00:59:26: No stupid questions.

00:59:28: We are actually gonna Do another little game A category that I like to call no stupid questions and then immediately have a stupid question.

00:59:39: Sure,

00:59:39: so if i made my Sports social network data available in public And then you could see My performance data that I put out on my daily commute or my bike

00:59:51: Yes

00:59:53: Is there chance that I still get discovered for the age group?

00:59:58: That I'm in?

00:59:59: or would you need significantly more data than what this sample every day?

01:00:08: It would definitely kind of bait me.

01:00:10: So if I saw, say you were putting all your training doctor on Strava as an example and... ...I saw it in a kind of caught my attention then the answer will be yes.

01:00:24: so As an early factor The answer would be Yes.

01:00:28: You know If i saw just even what is available on Strava, which they have quite a detailed info and stuff.

01:00:37: Yes that would attract my attention.

01:00:38: but then I'd want more from you after this.

01:00:43: so once i make contact with them like hey pretty good for guy commutes to work in back whatever one or two races will go into deeper dive.

01:00:54: You're being impressed when

01:00:56: are.

01:00:57: Yeah, I've read it all over social media.

01:01:01: It's right there?

01:01:04: Does that help?

01:01:05: to have the data?

01:01:06: to convince riders was a trickier before.

01:01:12: is that tool helps you

01:01:14: To come over to the team?

01:01:19: Yes and no?

01:01:20: so obviously look at the dice.

01:01:22: then You can present to the rider but within that realm presenting an overall view of why we would like you to be on the team.

01:01:32: So where do see what your role is?

01:01:36: Do get some freedom, only set as a worker which also fine for some riders.

01:01:41: so from that side but not in every case say this this with you, your loss will become a better rider.

01:01:53: So we have room for growth whether it's performance or just from the mentality side.

01:01:59: however we can support you but its not that as an example I do analysis of view and then when go to call with manager now present full data analysis back too.

01:02:12: It doesn't work like that.

01:02:14: We say hey We see you can improve these areas or hey, we could make us stronger in this.

01:02:21: Or This is really good But it's not.

01:02:24: that doesn't really go back to the rider because That's also what?

01:02:28: We look at and we don't always want stuff like that.

01:02:30: public You know all teams in general Don't want that.

01:02:33: you know cuz if they're either turned around says well I don't think it's a good fit which is fine but then He lets out information art kind of.

01:02:43: no It's not always the best

01:02:44: he gets.

01:02:45: free evaluation.

01:02:46: yeah A thing I see very often is that and even in personal life, we use data to confirm our own gut feeling.

01:02:54: Do you have a way of trying especially when making this big decision?

01:02:58: You probably have... When you look at the rider, do you feel like?

01:03:01: well i think it could be candidate or maybe your skeptical how.. Is there good way to avoid That- It does play role and expertise doesn't matter.

01:03:12: How can you avoid having just looking for data that confirms your belief right away.

01:03:20: Do you see that as an issue at all?

01:03:22: No, not really.

01:03:23: to be honest.

01:03:23: it's not like we'd say again... We use a rider what we think and look at the data And then once you as a rider.

01:03:46: Then we go, okay can he full the role like that within the team?

01:03:49: You know so then it's nothing there.

01:03:53: I don't...you know not always but sometimes you just say well are going to move on from here and will find look for somebody else or otherwise if they do get data.

01:04:07: this is really big surprise.

01:04:12: Do you still have a lot of spreadsheets?

01:04:14: Is that what, like just in practical terms.

01:04:17: You already... I guess you had the software suite.

01:04:19: Yeah so... How often do

01:04:21: spreadsheets appear

01:04:22: on your day-to-day?

01:04:24: Not that much to be honest.

01:04:25: they definitely faded out but you still do have spreadsheet.

01:04:31: I think Excel will never die.

01:04:33: i'm sorry it's to say that.

01:04:34: But uh..I would probably say any.

01:04:37: no Probably seventy percent of our stuff is now software data related, and then you can print what you're looking at or what you need versus working only on spreadsheets.

01:04:51: But I will say that you do still work on the spreadsheet but it's definitely a lot less in terms of overall time and volume.

01:04:58: We see a lot of Excel sheets being sent around and then they're off just one column.

01:05:06: And even if you ingest them automatically, there's the spreadsheet that is like oh it... The color needs to be the same every time or I didn't know?

01:05:14: We actually had that the other day with an Excel document where the sum of was literally K-nine or k eight but through everything out when now we are in a process of automizing.

01:05:27: Yeah no i already asked you right spreadsheets because.. You said what were going to ask excel somehow It just doesn't die.

01:05:35: It won't

01:05:36: die, I mean like and i'm not a Microsoft guy at heart?

01:05:41: They've gives me PTSD windows but from that side XLI don't think we'll ever die.

01:05:47: yet it's definitely less.

01:05:50: But I can't see it dying yet At the moment.

01:05:54: We were talking about And the key topic is prediction.

01:05:58: You have always worked in rookie development and finding young talent, developing it... Is there a special source to how you predict and benchmark if somebody can keep growing?

01:06:14: Yes or no!

01:06:17: If someone tells you that they will be X for four years then they are selling your story personally.

01:06:23: but look at the doctor.

01:06:27: How I personally do it is, see where they are today.

01:06:31: So how did they get here?

01:06:32: Today and then what i believe They will be in you know can we get more out of them?

01:06:38: so that growing to hear Can they grow further Is at a hundred percent accurate?

01:06:43: I don't think saying today yet In today's terms.

01:06:46: I think We'll Get A lot More Accurate With It.

01:06:50: But That is essentially what you want to do, understand where they are today and will be in the future.

01:06:58: Do feel like technology can still leap your head?

01:07:02: And I believe there has been a phenomenal development recently with abilities... ...and I bet that you have the best engineers!

01:07:11: Do you still feel or so much more could be done?

01:07:15: Yes definitely because now whatever AI and those sort of platforms.

01:07:24: You know, it's an ongoing learning as I always say is machine-learning but there's definitely an on going learning for both Machine and you're a human being.

01:07:34: so i definitely believe It'll improve.

01:07:37: that will definitely help.

01:07:39: Will it be perfect?

01:07:41: As I said right now II don't think so no But he would definitely bear with to get closer eye.

01:07:47: feel

01:07:48: the corridor, I mean it is a prediction.

01:07:49: The prediction is inherently...

01:07:51: A prediction?

01:07:52: Yeah you know so- But the

01:07:53: corridor i guess gets more narrow.

01:07:55: Correct

01:07:55: yeah i think you'll be able to funnel that a lot closer.

01:07:58: So where are going today?

01:08:00: oh i think these six guys will go all the way.

01:08:03: It'll probably only be three You know or two Or even one which is also completely fine as long As you manage to secure That One or Two.

01:08:10: but i do believe that um Even from last year To now It's definitely become a lot more streamlined from what you've learned, and how are doing it.

01:08:22: So I feel that is the pathway forward.

01:08:34: to see which one is the most important, and we'll start with long-term training data.

01:08:54: How relevant that for your decision making?

01:08:56: It's important.

01:08:58: Can you give it a... At

01:08:59: moment one I would say one This

01:09:01: number One Okay

01:09:02: For now

01:09:02: Okay Recent Training History The last few weeks maybe Two.

01:09:07: for now I'd say that

01:09:09: Long term history still more important Actual race data from The last races actually.

01:09:15: so I would run it.

01:09:16: Um, i would run two to go to one race data we'd got.

01:09:22: We still stay at three and then long term would be one okay?

01:09:26: Long time would be too much number.

01:09:27: Okay.

01:09:28: So recent history long-term history raised data.

01:09:31: yes

01:09:31: after that recovery And I think your proponent of recovery yeah,

01:09:38: I would go to Number two on this

01:09:41: and the writer's attitude, personality

01:09:46: motivation.

01:09:47: Yeah it's important.

01:09:48: I suppose that would be...I will probably have that as number one because thats' important.

01:09:52: our writers personality is really how we wouldn't execute his job or execute stuff.

01:10:00: That depends on the data you get back based on that as well which should short term data cause.

01:10:05: there'd a rolling average of stuff.

01:10:07: so yeah i'll go with short-term data to long term three, race data

01:10:15: four.

01:10:16: So this is not prediction anymore just factors in general for the performance of a rider new category.

01:10:23: we have six.

01:10:24: this time.

01:10:25: nutrition

01:10:27: yeah and today cycling super important.

01:10:29: so i'd go one for now.

01:10:31: on that

01:10:32: recovery

01:10:33: they go hand in hand.

01:10:36: what's the third one?

01:10:38: consistency.

01:10:40: Well, okay.

01:10:42: These are all number one for now.

01:10:43: next

01:10:45: time The coach

01:10:50: You're super important.

01:10:51: also give me all of them and then I'll put this puzzle together

01:10:54: in mindset.

01:10:55: Wow This is hard.

01:10:57: these would literally go to number one.

01:11:01: I Would say

01:11:04: recovery, nutrition consistency.

01:11:06: The coach and the mindset

01:11:09: need consistency first and foremost.

01:11:11: so I would go consistency first.

01:11:14: you need a good coach.

01:11:15: um So i would say that second.

01:11:18: And then but the nutrition is also goes on the consistency You know?

01:11:25: Two point one two point three.

01:11:30: It's a tight race.

01:11:32: Yeah, it's the tight race.

01:11:33: that because all of those are really super super important and interjoined.

01:11:37: without The one you can't really do the other.

01:11:39: so you need the consistency for progression You need your nutrition to be able to progress?

01:11:44: You need a good coach on And off-the-bike so to speak.

01:11:47: So prescription and personality is really important to get the best out the athlete.

01:11:52: Unfortunately I wouldn't really give that an order would almost put them all horizontal though its really any.

01:11:58: That is essentially how you get an athlete to be better.

01:12:03: It isn't interplay of all these things

01:12:05: together.

01:12:05: So, yeah athletes in the middle and then everything else it's just kind of circling around this.

01:12:10: Then your optimize whichever one is correct lacking right now?

01:12:13: Yeah We have talked about the collection All the things that go into the machine that you just put on top.

01:12:23: And then you stir and

01:12:25: do

01:12:25: some calculations predictions, diagnostics.

01:12:31: I mean you always start with description diagnostic prediction and then prescription.

01:12:36: try to find the action that is correct?

01:12:40: And right but it's not everything.

01:12:43: there are only...I guess You can get out of data.

01:12:46: what in the data

01:12:47: Correct?

01:12:48: There still probably so much human element especially at this level of performance, so that will be very interesting to maybe hear a little bit more about the relationship with the rider and coach.

01:13:02: You already said there's sometimes people who look good on paper but I guess even the connection coach can play a difference in performance.

01:13:12: if somebody hates where he is?

01:13:15: He won't perform as well or someone loves his team gets along really well with everybody which probably has an impact As well as decision-making, where the human still matters and experience.

01:13:28: And that would be very interesting to get into.

01:13:33: how does coach trainee relationship play out?

01:13:40: Is it a fight?

01:13:41: Who is maybe, and I was wondering you working with young riders.

01:13:45: Yeah are they easier to go with because they don't know yet?

01:13:49: or is it harder Because They Don't Know Yet And They Probably Have All Kinds Of Ideas

01:13:53: Again?

01:13:54: so the answer's yes and no To everything of what You Said.

01:13:57: So It Really Depends On The Rider Itself In On That Personality.

01:14:02: So Whether Its A Young Rider Coming Through Or Whether Its Season Pro you know, that it doesn't really change.

01:14:10: The part that you do see a lot with the young riders now because they've grown up in this kind of information technology rich world and... You know?

01:14:21: It's like second nature to them To what they do And they're super involved.

01:14:24: Like every kid now barring one out of a group of thirty has a power meter.

01:14:29: They have training peaks or whatever performance software.

01:14:35: Information is readily available, but some of them do think that they know more than you and a very I won't say opinionated because that's in the strong personality.

01:14:45: You know?

01:14:45: And then like well i knew this was right should be any grams not one hundred twenty grams or cobs on your life.

01:14:51: okay so you can have with them.

01:14:54: it's more an education process within.

01:14:57: back your connection to the coach.

01:15:01: It has to be close.

01:15:03: you know its not just talk to them.

01:15:06: Well, at this level you need to communicate every single day and you need have a really close connection.

01:15:11: There needs to be a lot of trust from the rider-to-the coach.

01:15:14: And that coach also needs to trust the rider.

01:15:16: when that element is gray It's potential for not The best outcome.

01:15:23: put it that way.

01:15:25: You know if you go down the coaching side too?

01:15:28: You know like an age group or so If I had to work with you that level of communication isn't really warranted because you also have, it's not your career.

01:15:38: You have a job

01:15:39: etc.,

01:15:39: but hey let's catch up on coaching and you're doing this well or

01:15:42: whatever.".

01:15:43: But at this level even from the under-nineteens to under-twenty three is that human connection is really important.

01:15:53: That is why I think at the moment AI Coaching It's good like if i download it then can ask it give me six week program better at the end of six weeks.

01:16:06: But, the problem is when things start going wrong or there's outside factors like I've fought with my wife for whatever case may be it doesn't understand that.

01:16:15: yet you know and i think where super high level kind can go very

01:16:23: don't quite show up in the data.

01:16:25: or do you maybe even catch riders having, because there's a subjective experience and I know that is metric.

01:16:32: In itself, the perceived exertion... Yes Do see big differences?

01:16:37: And it can help.

01:16:39: Or yes See if there are misunderstandings where they're like.

01:16:43: No!

01:16:44: The data isn't supporting what your saying

01:16:47: Definitely.

01:16:48: So how i explain this?

01:16:51: Correlate the two.

01:16:52: so you correlate to data and you correlate the athletes itself.

01:16:55: So we have objective in subjective.

01:16:58: The subject of is the athlete an objective?

01:17:00: as a daughter, so I could open the file And go.

01:17:04: well this was amazing You know like excellent good numbers good sessions everything with spot on?

01:17:10: And then iPhone new and your like.

01:17:12: oh mate i was on the couch last night I've lost my job or whatever.

01:17:15: you know that That is, and then I need to correlate those two.

01:17:19: To make sure that tomorrow or where you are today's good to go forward You know.

01:17:23: so all you see really bad darts And your like.

01:17:26: oh what wrong?

01:17:26: You're on the couch again Your like no everythings best in world.

01:17:30: So it definitely subjective an objective.

01:17:34: when looking at

01:17:35: Yeah, and sometimes you get like an angry ride.

01:17:37: And then you see all this.

01:17:39: what?

01:17:39: What do you did yesterday to that again?

01:17:42: Well

01:17:43: terrible fight with my girlfriend Angry road

01:17:48: message.

01:17:49: your sister again go go do the interval.

01:17:51: so yeah

01:17:54: Do you feel decisions are easier to make With the data or Are they just better decisions?

01:18:01: like does it help you Make the right decision?

01:18:04: or maybe both.

01:18:05: Yeah, it's both is definitely so.

01:18:07: you can make Better precise informed decisions based on the data especially in today's times if you want to call at that.

01:18:17: So that that is definitely without a doubt

01:18:21: The.

01:18:22: I think we already touched on this.

01:18:24: but again moneyball and data versus personal experience.

01:18:30: In the movie they kind of dismissed having these softer qualities and somebody having issues at home or something like that.

01:18:39: There it was kind of dismissed, uh...that is not the most relevant factor I feel.

01:18:45: here you have already kinda touched on It being Not The Opposite but those softer quality's attitude mood are really a big factor.

01:18:55: Do You Evaluate That?

01:18:57: Or do you Feel Like The Writer Himself Has A Good Feeling On That?

01:19:02: How do the software qualities inform decisions?

01:19:07: So again, personality wise some riders have a really strong personal feeling on that.

01:19:14: You know and they can express it also in an positive way.

01:19:18: Some to be honest cannot do.

01:19:20: let's call itself reflection if you want use that word.

01:19:23: but there like as example If your tracking heart rate variability so you have data there's maybe a trend is wrong or they're kind of high, low etc.

01:19:34: You can also then go to the athlete and say hey you know I've picked this up whatever your sleep is less Your stress is higher Your mood is down Your RPEs are up And your heart rate variability Just on what i see from the doctor give me What's going On?

01:19:50: Then sometimes things aren't good.

01:19:53: So it Is definitely Especially with how high stress, procycling is especially in today's peloton and stuff.

01:20:04: It really isn't an important factor but you can kind of sometimes pick stuff up quicker before the athlete says something to you.

01:20:12: The saying would be interesting how important it still a conversation if phone call.

01:20:18: I guess You're not always like hand-in-hand with every athlete.

01:20:23: Yes

01:20:23: How important Is that Still In An Age Of Data That You Just Like?

01:20:29: Really super important, you know because sometimes doctor will say.

01:20:33: You know he's not good or there is a problem of whatever.

01:20:37: we wakes up.

01:20:38: any example.

01:20:39: I wake up and I'll get told my sleep wasn't going to end my HIV is down etc.

01:20:46: And i'm like well how do you feel?

01:20:48: they're like no fine?

01:20:49: Then you said Well then your bike can go to work You know, so from that side it is super important but you can also get on the same site where they're saying oh no I'm fine and your going well.

01:21:00: The trend is definitely not showing.

01:21:02: you find.

01:21:03: So you kind of work at both sides But its really important.

01:21:06: That does help with decision making And you'll also find out if an athlete's honest or not.

01:21:12: How

01:21:13: did this happen?

01:21:14: Yeah, because they often say oh everything is great if it's not.

01:21:16: Because I don't want to lose a training day cause i think They go backwards or they loose form and fitness where sometimes It's not.

01:21:23: then you said well let's just do three easy days maybe A day off in two easy days And we can get back on the program one Thursday?

01:21:31: Then You see all that dotting All those trends completely changed In a positive direction.

01:21:38: Now you ain't touching.

01:21:38: You did super well, so yeah there is that.

01:21:41: do you ever catch riders just like skipping a day and Like not telling you?

01:21:46: And then you're like well this is like does no workout here.

01:21:50: It's not a lot but it is I mean.

01:21:54: But i can tell them i'm lucky enough to hear That.

01:21:58: um, you know Athletes will say looky i'm Just not feeling it today or whatever and its very few in far between days.

01:22:04: But if you log on and see there's a red day, what is going on?

01:22:09: Sometimes they just don't upload which sometimes within twenty-four hours or whatever the battery goes flat in their common unit.

01:22:20: They really hardly ever go out training and low to fall.

01:22:29: So the dog ate.

01:22:30: my homework version of procycling is, The battery went dead?

01:22:35: Yeah

01:22:35: pretty much that's it yeah.

01:22:37: Oh my Garmin batteries flat

01:22:41: again but

01:22:42: very few days of dog ate My Garmin.

01:22:46: Dog ate my Garmen.

01:22:48: That should be the headline or podcast.

01:22:54: Do you do a lot querying, analysis technical work yourself and how much do you enjoy that?

01:23:03: I do.

01:23:03: We also have our doctor-scientist who does a lot more now for me but i do do a lot of it.

01:23:10: um...I do enjoy it!

01:23:11: Um..i will say that It's not like ...you know ..I open the computer and am like ohhh I've got to do this.

01:23:19: yeah....I do enjoyed.

01:23:20: To a point.

01:23:22: Do you feel like your being unable to get all the information?

01:23:25: Because I feel that's, uh... That is big factor.

01:23:28: now and things have improved so much.

01:23:30: You can have very non-technical people And management and CEOs Get inside!

01:23:38: And there are still key factors.

01:23:40: do you have similar situation?

01:23:42: Just understand it.

01:23:43: Yes because when we said managements sometimes need You take the technical side of it and then you need to almost lower It into everyday language, I mean handed over to them.

01:23:58: So you got a kind of translate from A-to B like its definitely a thing.

01:24:03: our doctor sciences an example just dumps The highest level of like.

01:24:08: I mean you'd need to put some of it in to chat gpt or something too even for myself to get their answers out of it.

01:24:14: but Yeah, yeah, you are definitely the answer would be yes.

01:24:18: And yes i enjoyed.

01:24:20: And then that does sound like technology hasn't enabled you to drill down deeper and the axis is widened through technology.

01:24:31: We talked earlier about manual data input, do you see it as a big factor or issue when people have put things in?

01:24:43: Yes-no also again Like knowledgeable or do they just have no interest in it?

01:24:52: You know, nine out of ten times now.

01:24:55: It's okay But in the past we have had times where I mean this is back and i can actually tell you a funny story on This.

01:25:04: I was in South Africa at that time And I was told to fly to Italy To a rider because he hadn't uploaded data for A really significant amount of time And we were working with the company.

01:25:23: The unit was a little bit tricky, it's still in its early phases but wasn't that tricky and he was kind of challenged to save at least in terms of uploads.

01:25:34: as I said I had to sit on a twenty-four hour flight to go upload his daughter for him so It does happen.

01:25:42: however now things have progressed how things talk to each other is not really that much at the moment.

01:25:51: You know, more it's potentially a bug fix or something from a manufacturer if there's something going on but typically from riders and stuff It's not that much of a problem.

01:26:01: at the moments

01:26:02: Do you have like manual processes where somebody has to rate something?

01:26:08: Or send something over because

01:26:11: Yes, we do especially on like a subjective information from a rider.

01:26:15: That's kind of manual input if you call it where that sometimes becomes the challenge is not done automatically and become a bit of a challenge

01:26:24: because thats kinda my philosophy an hours that any manual step is mistake or potential for error.

01:26:32: as data engineer I'd like to treat this as if its manual i don't trust which can be at certain level maybe And sometimes that's all you get, and then you have to deal with it.

01:26:43: But anything that involves a human is inherently

01:26:47: flawed

01:26:48: in a way.

01:26:48: Fair play

01:26:50: for me as time consuming.

01:26:51: so even if I take something as simple as our weekly reporting document we've literally ninety eight percent optimized the whole process So far.

01:27:04: us coaches are easy relative terms.

01:27:08: yes you still have to manually put in, hey this guy did nothing and he lay on the beach all week as feedback or he was phenomenal.

01:27:16: But I mean that takes six seconds in terms of typing.

01:27:19: but everything else we had to manually insert is now fully optimized You know, so from that side it's significantly helped us.

01:27:28: Okay and you would say that is important because super-important?

01:27:32: Yeah!

01:27:33: It's time saving.

01:27:34: cause before just time consuming.

01:27:36: um...you know I'll say everythings in one place etc.

01:27:39: So its definitely improved.

01:27:43: Just that one small area In terms of what we do as coaches From the outside..I

01:27:48: have One more question about race day Is a Because are your moving along with.

01:27:54: Are you in the car with a laptop on your lap or?

01:27:58: In the control room, how can I picture data-driven decision making?

01:28:02: so that's normally pre and post.

01:28:04: So it's not.

01:28:06: It's not always in race although sometimes yes i get a phone call or coach gets a phone called.

01:28:11: hey Give us this Performance outcome You know And you would give it in the race to the sporting director.

01:28:21: however A lot of it is done pre and then directly post, you know.

01:28:25: So the pre race would be.

01:28:27: this is what the game plan is.

01:28:29: these are the scenarios.

01:28:30: we'd give them information whether it's something as easy as climbing times on tourmalet or whatever maybe crosswinds coming up etc.

01:28:46: but then post-race It's full analysis, feedback on the performance side of it from that data and then we make decisions for next day based off that.

01:28:55: We just talked about decision making how Data is used for it.

01:28:59: Do you have a good summary?

01:29:02: what are key factors?

01:29:04: How do use Data to Make Decisions?

01:29:07: We made decisions based obviously on what we see.

01:29:10: so whether its benchmarks, performance factors Um, historical data as I said on how we got you what we see etc.

01:29:19: that is really important in terms of What are benchmarks or?

01:29:25: How do we analyze there not really say on the podcast now um but it Data driven.

01:29:35: Is a big decision-makering In terms of what we do.

01:29:37: and then we couple that You know again As i said to to the human side of the athlete And what our scouting group, as I said is Tim and Michael what they see in the athlete.

01:29:48: Do we correlate it?

01:29:49: And then We go from there because sometimes As i said you See really not The best data but the riders performing or vice versa.

01:29:58: You know.

01:29:58: so those Those to make Really important decisions But essentially hand over What I seen the data and Then They Go From There.

01:30:07: How would you say has scouting the role of Scouting Changed since you've introduced more data-driven decision making.

01:30:16: I think it has, its fine tuned to make a lot more... How do say?

01:30:22: Just because ten kids are winning bike races doesn't mean that those ten kids have the future.

01:30:28: where now we take these ten kids out of winning a bike race but probably only really takes four of them or interested in four and watch others five or other six kind of progress over time.

01:30:42: So it's helped, I wouldn't say soft art non-talent but just make a more correct informed specific decision especially on how modern cycling is in todays times

01:30:58: and the traditional scouting off having people assess riders Is still there?

01:31:05: Yeah

01:31:05: hundred percent.

01:31:06: But you definitely need the two to talk to each other and come to an agreement.

01:31:12: Like, I don't believe you can only do one without the other

01:31:14: now.".

01:31:15: Is there ever a big disagreement?

01:31:18: Does it not align... ...the assessment...?

01:31:22: Not a lot but sometimes yes!

01:31:25: You know.

01:31:25: example of the day when we said hey guys to our scouts are really nothing here that blows my hair back.

01:31:33: what did he see in the

01:31:34: races?!

01:31:35: And they were like no he's really performing well.

01:31:38: So there you, then it goes to a bit of deeper discussion.

01:31:43: We had that last year as well where it was almost opposite.

01:31:46: really phenomenal numbers.

01:31:48: but from the human side we didn't feel he could do what nothing can with numbers.

01:31:55: just for me humans I don't think he would carry this through all the way though.

01:31:59: so... That marriage needs be good.

01:32:05: And how can you, if a cyclist is not already on your roster assess I guess that's through conversation at some point.

01:32:12: That you can assess attitude motivation mindset?

01:32:16: Yeah correct.

01:32:17: so typically How like the process would work?

01:32:20: they'd be identification of the athlete.

01:32:23: then You know The the scar could say hey we interested whatever.

01:32:27: Can We get access to Your data?

01:32:30: We Get Access i give my feedback back and Then from there it normally goes to, should they be interested?

01:32:37: Then it goes the actual scouting one-on-one phone call or meeting with a rider parent agent whatever whoever's in the call.

01:32:44: And that goes into more deeper dive In terms of That human.

01:32:50: I can ask you questions all day and we can go on because also You cover all the fields.

01:32:55: Or what was your current carb number?

01:32:59: Because its funny how cycling, I feel like is influencing running because the runners for many years were still on like thirty grams

01:33:07: more

01:33:07: or something.

01:33:08: And then it's like you guys are taking how much?

01:33:11: Yeah and now i think through triathlon yeah that information on The carb intake uh Is like moving over.

01:33:21: and

01:33:21: yeah It's definitely increased even in cycling from what it was before.

01:33:26: You know, even in cycling before it was a fairly low volume of per hour.

01:33:33: But what you've got to also remember whether your triathlete or runner is cyclist and that is... ...you can say do one hundred twenty grams an hour which I mean if its high-volume of carbs but if the body doesn't actually absorb it then use as fuel source kind of just going to the toilet afterwards.

01:33:53: So it's really there again, really specific to the athlete.

01:33:58: yes you can train your gut to absorb more but they will be a ceiling where you just don't take it in.

01:34:05: on that note how what are you iteration like?

01:34:09: How often do you adjust?

01:34:16: program for six weeks and then I see you again.

01:34:19: Is it daily, hourly?

01:34:20: What's the cadence?

01:34:21: Typically a week but you change to like... So i'll build your program for a week And as an example You come back from training today.

01:34:34: It really wasn't good.

01:34:35: We understand why.

01:34:37: Then we will change that tomorrow.

01:34:39: so its built For a week But it is adaptable day by day Correct.

01:34:45: Things occur?

01:34:46: I

01:34:49: have one more question, because our slogan is high tech with a heartbeat and that's why my company so much there like this personality thing in the excitement for technology we built.

01:35:02: Yes

01:35:03: What something you are very excited or maybe even most excited about In your work.

01:35:10: I like seeing people progress, you know.

01:35:13: so it's not yes.

01:35:14: You want to look again?

01:35:16: i love winning bike races.

01:35:18: okay simple

01:35:18: as that.

01:35:19: first and foremost there is no always about winning bike racers.

01:35:22: you also want athletes to perform in and progress So you want them be better whether or not everybody can win a bike race but somebody If he's going to carry bottles, it needs be the best bottle carrier.

01:35:36: Unfortunately that is what it is.

01:35:38: and if you go from a medium-bottle carrier to the best domestic that world has like if we take Danny Van Poppel I believe this is the worlds' best lead outrider.

01:35:50: so on that side...I haven't worked with him for super long but he definitely grown into that role.

01:36:00: That essentially is something for my job And that's what I like seeing the most, and winning bike races.

01:36:08: Obviously!

01:36:09: That is the

01:36:11: goal.

01:36:12: Where do you see the most potential for improvement in your work?

01:36:17: Is it technology algorithms getting better data quality?

01:36:21: where do feel if we can make more gains there would be a big

01:36:27: improvement.

01:36:28: A little bit of sports science side, you know to make informed decisions or Decisions kind of easier.

01:36:39: And with that we need whether it's algorithms platforms etc.

01:36:44: from a data aspect You know to be able to I'd say analyze or read that data quicker Or more precise.

01:36:51: I think that that That will help us make decisions A lot easier, you knows How to tactically win a race or whether that's signing a certain rider?

01:37:03: I think they're all you know training Prescription to make the rider better.

01:37:07: I think that it all falls with within that realm of work,

01:37:11: I think That's a good sign.

01:37:13: Thank You so much for taking the time.

01:37:15: It's a pleasure

01:37:16: understand this time over here is Critical for you and your eager to go back but very happy To get all his insight and information from you And share your interesting story of data-driven decision

01:37:29: making.

01:37:30: Yeah, thanks for having me and also as I say thank you being a part of the team.

01:37:35: it's been great.

01:37:39: We've got to go there once we have our

01:37:41: third member in the cast.

01:37:43: now Let us know in the comments how do YOU handle large amounts of data?

01:37:51: What problems do you typically run into... ...and How Do You Solve Them?

01:37:55: This was AI That Works.

01:37:58: Real stories from the people making data and AI work in real world.

01:38:03: See you next time!