Perfect is The Enemy of The Good with Ameen Kazerouni

with Ameen Kazerouni, CTO of Orangetheory Fitness

Ameen Kazerouni, CTO of Orangetheory Fitness

Ameen Kazerouni

CTO of Orangetheory Fitness

Ameen Kazerouni operates at the intersection of science, data, and technology to delight audiences and customers with data-driven, consumable, and intuitive experiences. As the chief technology officer at Orangetheory Fitness, he’s leveraging the power of data to drive the consumer wellness journey.

Satyen Sangani, Co-founder & CEO of Alation

Satyen Sangani

Co-founder & CEO of Alation

As the Co-founder and CEO of Alation, Satyen lives his passion of empowering a curious and rational world by fundamentally improving the way data consumers, creators, and stewards find, understand, and trust data. Industry insiders call him a visionary entrepreneur. Those who meet him call him warm and down-to-earth. His kids call him “Dad.”

Producer 1: (00:01)
Hello and welcome to Data Radicals. In today's episode, Satyen sits down with Ameen Kazerouni, the chief technology officer of Orangetheory Fitness. Ameen has spent his career at the intersection of science, data, and technology to create intuitive data-driven experiences. At Orangetheory, he's driving consumer wellness journeys by turning workout data into feedback and personalized recommendations. In this episode, Satyen and Ameen discussed data-driven exercise, keeping humans in the feedback loop and AI data governance.

Producer 1: (00:33)
This podcast is brought to you by Alation. The act of finding great data shouldn't be as primitive as hunting and gathering. Alation Data Catalog enables people to find, understand, trust, and use data with confidence. Active data governance puts people first so they have access to the data they need within workflow guidance on how to use it. Learn more about Alation at alation.com.

Satyen Sangani: (01:03)
Ameen is the chief technology officer at Orangetheory Fitness, an exercise chain that uses heart rate data to achieve a more powerful and impactful workout. Previously he was head of ML and AI research platforms at Zappos. He believes in building experiences — not just cool tech — and focuses on creating scalable solutions with real world impact. His passions include artificial intelligence, quantum computing, and ML. Ameen, welcome to Data Radicals.

Ameen Kazerouni: (01:29)
Thank you for having me, Satyen. I'm very excited to be here and excited to talk all things Orangetheory and data with you.

Satyen Sangani: (01:34)
Let's start there, actually, because a lot of our listeners probably haven't been to an Orangetheory class. I hadn’t, and my entire marketing team was like, “You have to go to the class!” And I was like, “Ah, can I avoid it? Can I get out of it?” But they made me go and it was definitely an experience. Can you tell us what that experience is like and how it came to be?

Ameen Kazerouni: (01:51)
Absolutely. Orangetheory has been around, it's a brand that launched in 2010. It's one of the original pioneers of heart rate–based connected technology and fitness experience. The easiest way to describe Orangetheory is this unique trifecta that blends science, technology, and most importantly, expert coaching, that helps people at any fitness level live longer and more vibrant lives. And the workout at its core is a group workout. It's science-backed tech track. You wear a proprietary heart rate monitor, the treadmills, the rowers are all connected. There's a strength floor where we focus very explicitly on the functional strength and weightlifting portions of a well-rounded exercise and fitness regimen. Workouts are meticulously designed. We have a team of certified fitness experts. Everyone in the world does the same workout on any given day. So there's these micro communities that you develop in your studios that help you stay motivated and connected to each other. But you're doing the same workout as someone in Australia or Japan or Israel on any given day. We're in 24 countries. So it's an exciting brand to be a part of. And it's really cool to see the kind of impact we have on hundreds of thousands of people on any given day. And I'm just humbled by the opportunity to lead the kind of data and technology charge as it pertains to evolving this brand's experience on top of where it's already at.

 
 

The Orangetheory experience

Satyen Sangani: (03:14)
Yeah. So my experience was that I walked in, I got this heart rate monitor and they walked me around and they showed me the machines and nicely kind of made sure that I couldn't get myself in a bit of trouble. And then started me on the rowing machine and everybody else was on the treadmill. And there was that floor and it was a really interesting experience because we rotated between treadmill, rowing machine, and the floor and then had that entire burst on the treadmill. Is that what a typical workout looks like? Or what kind of variation does one see?

Ameen Kazerouni: (03:43)
So that is what a typical workout looks like. What you experienced, based off your description, is what's known as a two-group class where one group is either on the row on the floor and the second group does the burst on the tread. And that's a cool way of describing that workout. Sometimes you'll experience what's known as a run row, where it's actually a kind of run/treadmill/rower transition focus with a dedicated 25, 30 minutes on the floor. We actually also have three group classes that move you from the rower to the floor to the treadmill. So you do each station kind of in a dedicated way with some strength focus with the medicine balls on the floor. And then we actually have an explicit strength offering called Lift 45, where if you want to compliment your kind of Orange 2G, 3G workouts with a lift focus, you can go in and just do 45 minutes of meticulously designed lift workouts as well. So the offerings are getting a little wider to cater to many different needs that we're seeing in the consumer.

 
 

AI in the workout

Satyen Sangani: (04:42)
You and I talked a lot about this idea of the workout, having AI assist it, and we talked in depth about this use case of initial training where the first five workouts would give us this understanding of me personally in terms of my heart rate, what the maximum happens to be. Where else do you apply AI in the workout, if anywhere, and where have you found it to be successful? And I guess in the converse, where have you tried to do it and found it to be unsuccessful?

Ameen Kazerouni: (05:10)
One of the things that with Orangetheory that I find particularly interesting is that it's a goldmine of signal-rich information because everyone in the world's doing the same workout, the data is all coming off of proprietary technology. So there's very little cleanup — heart rate data connected by the second to treadmill data — so we can understand the response that the member base as a whole has, but across different fitness levels has to the workout as well. So we do use those data feeds to help inform our fitness design teams, as well, who basically on a kind of two-month lag, are designing 30 days’ worth of workouts. So what we call the fitbook, which is, and maybe I'm getting a little bit into the sausage-making, which is the templates for the month when they're being designed. We're constantly analyzing how certain sequences were received and what the heart rate response to certain workouts was. To continue to make the workouts as effective as they can possibly be to drive kind of material change in our members, whether it be body composition by increasing muscle, maybe body composition, by reducing fat weight loss, speeding up your mile time, training for a marathon.

Ameen Kazerouni: (06:20)
All these different goals require different exercise modalities. So data plays a pretty critical role in helping design that multivitamin workout to give you the most efficient 60 minutes possible. If you get up and drive to Orangetheory studio and do those 60 minutes with us, we wanna make sure it was the best 60 minutes we could have possibly delivered. On the other hand of that, and you asked about where we've seen it be unsuccessful, and I wanna say it's rarely unsuccessful at scale, given that safety of our members is a top priority for us, there's nothing that is data-driven that is 100% automated without a human in the loop.

Ameen Kazerouni: (07:01)
Our coaches are in a real-time feedback loop that's always ever-present. And the guidance we give our coaches and our members is, “Listen to the coach.” The coach is AED and CPR-certified. Every coach has a national fitness certification of some kind. Additionally, they go through what we call OT Fit certification, which is our own certification process. So the goal is to always maintain that human in the loop along with any data-driven experience to where we create this amalgamation of a hyper personalized, almost one-on-one fitness trainer experience, but in that small group setting. And everything we do goes through testing at what we call the lab, which is an Orangetheory studio at HQ where nothing happens for the first time in the wild. Everything happens for the first time at the lab and we make sure that it goes through a pretty rigorous testing process. So many things have failed in the lab, Satyen, I'll tell you that much. [laughter]

Satyen Sangani: (07:55)
I can imagine.

Ameen Kazerouni: (07:55)
And they are immediately turned off. Yeah.

 
 

Data creates “adherence”

Satyen Sangani: (07:57)
You mentioned that there's obviously machine learning in the context of the actual workout. There's also this idea of machine learning across workouts. What's my next best program that I ought to be doing given what I've done, given what my goals are? Has that yielded any fruit? And is that an area where you are able to do work or have been able to make findings?

Ameen Kazerouni: (08:20)
This is what I can share now, and I think it's very exciting. It makes me very happy to tease at it, is that the importance of your workout regimen across the week and the classes that you take and the heart rate response that you elicit, whether you push or you use it as a recovery workout, all of that adds up to the likelihood of you being able to achieve your goals. I think our VP of fitness who designs all our workouts says it the best, is that goals take time and time takes commitment and commitment's hard, and that's why exercise is hard. It's like, "I worked out today, why am I not feeling great tomorrow?" Adherence is what's really critical.

Ameen Kazerouni: (08:56)
So creating adherence, creating stickiness is something that data can obviously play a really critical part in. We've all seen the examples of data and machine learning being used to maximize clickthrough rates on ads, maximize add-to-cart rates, maximize cart size with the people who bought this also bought, what you also like and how do we take all that, that we've gotten really good at as a community of using data to drive consumer behavior and use it to drive adherence and consistency with a fitness regimen to drive some serious change in society as a whole. That's something that we are focusing on very critically. I think that leveraging our pretty extensive brick-and-mortar network of 1,500 studios across 24 countries, coupled with a really strong digital experience that we have available with a highly engaged member base in our digital experience, we are realizing that connecting that personalization from the studio to the digital experience is going to help create that adherence cycle.

Ameen Kazerouni: (10:00)
We are already seeing some of it with experiences that we've launched with this personalized max heart rate actually showing you how that change is happening over time in your app and within workouts, we've gotten even more granular in the kind of replay experience we give you. I don't know if you saw it, if you download the app and log in with the same email, you'll see it for the workout you did. It's a second-by-second replay of the workout that shows you exactly what your heart rate response was to a certain wattage of the rower or a certain speed and incline of the tread. And then using that to let you celebrate micro wins and let you celebrate milestones more efficiently so you can route that energy and excitement back into adherence. It’s something that we're really focused on right now and exciting things are coming for the members for sure.

 
 

Focusing on your differentiator

Satyen Sangani: (10:45)
Yeah, it sounds like it. You have, to your point, such a trove of data. Super incredible. One of the things you said previously is that if you were trying to build something else that someone is building as their core product, you'll never be able to focus on it as much as the person who's just doing it on a singular basis. You're, however, at a place where you've built this platform, you know, 1,500 brick-and-mortar stores, you've got the workouts available to you. There are so many different things that one could build. How do you think about that now in your evolution? What is the differentiation point and where do you continue to differentiate as you think about your technology organization and where you wanna take it?

Ameen Kazerouni: (11:22)
This is general first principles that I like and I would share as guidance or advice to any technology or data leader is that use your best resources to focus on your differentiator. Because when you're focused on your differentiator, you are basically changing the nature of the game rather than trying to play catchup. Like we use Snowflake as our data warehouse. It would not be sensible to dedicate any kind of engineering resources to replicating what Snowflake does, because they dedicate an entire organization to just doing that. We wanna focus on our differentiators and leverage the tooling, where the tooling's available and not waste a single ounce of dedicated full-time employee energy and resource towards building something that somebody's already built better.

Ameen Kazerouni: (12:09)
When it comes to what our differentiator is, my focus is, without a doubt, it's on this richness of information that we have available to us. Orangetheory has hundreds of millions of workouts’ worth of data of second-by-second heart rate data tied to treadmill data, tied to rower data that allows us to understand the human response to this kind of variable interval intensity, heart rate–based training better than anybody out there can. And I think leveraging that understanding to create an increasingly efficient workout to maximize results is the differentiator, whether it's within the brick-and-mortar experience or whether it's through our digital experiences.

Ameen Kazerouni: (12:48)
I say that when we talk about our technology. I think when you think of Orangetheory as an organization, the real magic is when you take our coaching team, which is a couple of thousand of probably the best coaches — and I'm a little bit biased — that exist, right? They're not only highly trained and consummate professionals, they are so wildly invested in the members, in those relationships that they build with those members, that giving them all this information, giving them all this tooling to make it easier for them to have conversations, to let them also focus on the differentiator, which is that one-to-one personal training being available at their fingertips because they've got that information just like that, is what really becomes a magic of Orangetheory then. And we say this, and we know it to be true, is that our members, they come for the workout, they enjoy the workout, but a lot of them stay for the coaches. The relationships that they build with those coaches and the results that they see with the experience that they get with the coach and the workout working in tandem, the technology at that point just becomes an enabler to that experience.

 
 

Data as a habit

Satyen Sangani: (13:56)
I see a lot of, frankly, parallels between Alation and Orangetheory in the sense that the habit that we are trying to create is the usage of data. In your case, the habit you're trying to create is fitness and long-term health, both of those habits are aspirational. You'd love to be able to use data more in your life in order to make decisions. And certainly in your job, you'd love to be able to work out every day. Certainly, I don't know, whatever, I would — but you know, that's not my reality. And many people's realities. And so there's this kind of aspirational self and then there's this real self and it's really hard to get people to change their habits. So how do you do that and what are the stickiest parts of the experience? Is it obviously first trying to get people to try it? Is it getting people to stick with it after a certain period? Where do you guys see the biggest challenges in sort of achieving your longer term vision?

Ameen Kazerouni: (14:43)
I'm gonna speak from personal experience here because fitness wasn't part of my reality as well when I first started at Orangetheory, and it has become my reality. And I think there's aspects of how the product evolves as well, in which we think about exactly the question you just posed, is how do you break past barriers when it comes to starting something new and staying committed to something new? I think step one is recognizing that there's an aspirational reality, which is what it is and maybe difficult to achieve, but then there's the other reality of no fitness at all. And it's not a binary decision. It's not a matter of I'm not gonna be able to go two to three times a week and I'm not going to be able to complement it with the right strength workout or the right long walks in the evening and the perfect nutritional plan and the right protein intake.

Ameen Kazerouni: (15:37)
So instead I'm just going to not do anything. Right? So I think that continuum, that spectrum of starting the journey, just showing up for yourself is the biggest move that you can make. So keeping that in mind, we make a lot of investment in demystifying the Orangetheory workout. And there's a lot of parallels to data, and I love that because when you think about data in an organization: “Oh well, it's gonna be a multimillion dollar investment. We need governance, we'll need DAG orchestration, we need the right metadata. We need data engineers and business intelligence analysts, and then we need machine learning scientists, and then we'll need a data visualization layer and the right data warehouse behind it.” It can get so overwhelming that instead of being like, let's start piece by piece, the instinct becomes, let's just keep guessing instead. Right? Which is never a good idea.

Ameen Kazerouni: (16:24)
You should never, “Let's just revert to not using data at all because it's gonna be really difficult to use data perfectly.” It's the same thing with fitness. You don't revert to doing nothing at all because meeting all the requirements will be hard. Showing up and getting started is what gets you going. So demystifying the workout and making it known that Orangetheory is not hard. Orangetheory is what you need Orangetheory to be for you. If you show up, you can power walk, you can run, you can jog, you can ask for modifications on the floor. You can choose to do every movement, but just body weight instead of weights. You can use a strider or a bike if you don't wanna be running on the treadmill, you can modify it to whatever you need it to be to get started on your fitness journey.

Ameen Kazerouni: (17:10)
I think once you get started and you continue to show up, the results start speaking for themselves. We often survey our members and the bottom line is the answers that come back most often are, “The coaches are great, it works. I don't have to think, I can just show up, do the 60 minutes, do what the coach tells me and leave. And within weeks, months, I'm starting to see tangible results.” So it's simple, it's fun, and it works. And the hardest part I think is showing up. Once you show up, it is a domino effect after that. And I think it's the same for data. Once you get started, even if it's the right cleaned up dataset in Excel that helps you make just one decision, that was intuition coupled with information, you realize that's a much more powerful decision than just guessing. And you know, I've said this many times before, is that people jump into AI, jump into machine learning, jump straight into the really expensive stuff — and just creating a shared language, creating the right layers of data governance, creating the right accessibility and democratization of information in its simplest form is going to start driving material change in your organization.

Ameen Kazerouni: (18:17)
And that's what's gonna allow you to start that journey on becoming a data-driven AI-first, ML-driven organization. It's a journey. You don't just jump to the finish line.

 
 

Complementary competition

Satyen Sangani: (18:28)
It's a habit. And like with all habits on some level, perfect is the enemy of the good. And you'll never get to perfect because there'll always be some other level to get to, no matter how in shape you are or how much data you happen to have, which is super cool. This entire revolution of information technology and fitness has happened in so many different places. We just happened to hire the former CFO of Peloton, and through getting to know her and as a Peloton customer previously, like I had seen and learned a lot about that. Is something like Peloton, your primary competition or how do you think about your primary competition? And it would seem that there's actually a lot of complementarity with some of these other technologies. And so how does one think about the vision and the work in the context of differentiation, but also in terms of extension of the product line?

Ameen Kazerouni: (19:17)
Competition is a way that I try and not think about it. You can't help it, but you do start thinking, “Oh, that's a competitor, that's a competitor.” But there's a stack, I think I'm getting it right, but only 20 or 25% of the entire US population works out or has a gym membership or something like that where it's a small number of people that are actually exercising and working out and invested in fitness. So I feel like competition, there's enough to go around, it's not ubiquitous enough. So the more concepts there are, the more access there is to ad hoc content, live content, at-home content. The more accessible in-studio fitness and boutique fitness is to more and more people, the entire industry is gonna benefit and society is gonna benefit as a whole. I was reading an article recently where the... U.S. is one of the only countries where life expectancy is actually down, and that's not fun to see, ever.

Ameen Kazerouni: (20:10)
So I think investment in self and investment in wellness and fitness is just critical as a whole. And the more companies focus on making it accessible, the better. I do think the point you make of it being complementary is absolutely right. There's a time for at-home fitness, there is a time for on-demand fitness, and then there is tremendous value in in-person motivation, in coach-led community-based fitness as well. That is what our contribution to the industry is, is this extremely scientifically backed, meticulously designed in-person, community-based experience that brings technology to the table to help you measure your improvement and leverages that technology to show you how you are growing compared to self and bringing that expert coach into the equation to make sure that even though it's in a group setting, you're getting that one-to-one attention and using technology to make it easier to personalize, make it easier to feel and be what you need it to be for you while you are in a group setting, getting the motivation and community of your peers and people who are on the same journey as you.

Ameen Kazerouni: (21:16)
So while I do think we bring something distinctly unique to the table, there's nothing wrong with there being a lot more at the table because I feel like this is something that we just need to keep investing in as a society more and more and more. And I mean right now I'm wearing a continuous glucose monitor… this is a bionic arm at this point. It's not for any medical reason. It's predominantly out of curiosity and trying to understand how different foods impact it and how exercise and consistent exercise changes the glucose response to different foods.

Ameen Kazerouni: (21:49)
And it's been eye-opening to me. And now while this is a relatively not easily accessible or even something that is easy-to-commit-to piece of technology, I think five years from now, even 10 years from now, 15 years from now, whenever it happens, the watch in your hands gonna be able to give you a continuous glucose reading. And that's huge because we're gonna keep learning.

Satyen Sangani: (22:08)
It's huge.

Ameen Kazerouni: (22:08)
And we're gonna keep growing and we're gonna keep either getting healthier as a society or knowing how unhealthy we are as a society, one of the two. Those are the options.

 
 

Surprising Orangetheory insights

Satyen Sangani: (22:14)
Well, it's the knowledge of lack of health that often becomes a motivator to do the opposite thing. It's pretty cool. I also love this notion of investing in your differentiation, but also realizing that since 75% of the people don't have an active fitness regimen, that's really the problem. I mean, we say exactly the same thing. I say exactly the same thing to my team and to my customers about data. The problem isn't, there are some competitors that do things slightly better or slightly worse. The real problem is people don't use data as much as they ought to.

Satyen Sangani: (22:42)
I always get this question like, "Ah, is the industry big enough?" I'm like, “I think so. I think data's kind of useful and if more people used it would probably be better, but a hard problem to solve.” Nevertheless, before we get off of Orangetheory, what's the one “aha” that sort of captivates you with Orangetheory in terms of your favorite data point, your favorite insight, what bit of data has stood out to you?

Ameen Kazerouni: (23:03)
So there's this one piece of insight that kind of just blows me away. It's that 70% of members that have spent at least three months with Orangetheory on average have increased their for-class tread speed by half a mile per hour, their tread distance by a quarter mile, the average wattage on the rower by 30 watts, and they've done all of this at a lower or same average heart rate as when they started. Seventy percent of members, more output in terms of speed, distance, and wattage on the rower for a class without increasing the cardiac output needed to achieve that work. It's just so astounding to me that there's such consistent results in not like a super-caveated thousand-person population, but 70% of all people that just stick with the program for three months. And that's a large chunk of our member base.

Satyen Sangani: (24:01)
And how many workouts are they doing per week in that three-month period on average?

Ameen Kazerouni: (24:04)
Two to three workouts a week, which is our general prescription.

Satyen Sangani: (24:07)
Nothing critical intended here, but I don't understand what the other 30% are getting. If they're showing up two to three times a week for three months, how is that not happening for them?.

Ameen Kazerouni: (24:17)
It's a really good question with a very simple answer, is that driving cardiac efficiency like that, it really depends on what your window of opportunity is. If an elite athlete comes into the studio that has already got a relatively established fitness regimen, the window of opportunity to drive more distance at a lower heart rate, everyone's increasing the distance that they cover the speeds. But doing it at a lower heart rate means your heart's getting stronger and that window of opportunity tightens as you get healthier. So it's not that the other 30% isn't gaining anything, it's that the window of opportunity had probably already been pretty tight for a large chunk of our member base that has already got a fitness regimen, for one.

Ameen Kazerouni: (24:56)
And two, consistency and maintenance as you age is something that the fitness industry as a whole just doesn't celebrate enough. We always talk about improvements, we always talk about weight loss, we always talk about things getting better. What we don't talk about is maintenance over time is actually the real goal. You will eventually arrive at that fitness level that is sensible for you to maintain, given your lifestyle and whatever that level is, maintaining that is the real long-term investment. And it's just not as sexy as razzle dazzle stats on improvements, so it's not as exciting to talk about, but maintenance is just as important and not sliding backwards is just as important. And we find that a large population of the Orangetheory member base uses Orangetheory for that as a maintenance regimen, and that's totally fine as well.

 
 

“AI data governance”

Satyen Sangani: (25:45)
Let's switch gears a little bit. I mean, I think having listened and spoken with you now, it's pretty clear that you are so thoughtful and so scientific about how you approach the professional problem you're tasked to deal with. And you gave a talk about this idea of AI data governance and talked about it as being a force multiplier. I think that was the specific term you used. I haven't heard about the idea of AI data governance. Can you tell me what that is and illustrate what that is first of all?

Ameen Kazerouni: (26:11)
I think that it's a phrase that I've thrown around and the perspective there is that when you think about machine learning and statistical data-driven decisioning and the umbrella dome of artificial intelligence, you wanna get stronger at maintaining the right data governance that helps you maintain trustworthy reproducible models. What I don't like the idea of is not having the right governance in place to where your feature space can drift, where your understanding of what the model was trained on and what the model's being used to make predictions on have diverged. The idea of a black box that you create a large amount of like business dependence on where if it stops working, now what? So it's nothing new, it's nothing fancy.

Ameen Kazerouni: (27:00)
It's maintaining a shared language, keeping track of your version control. And higher models have been trained, higher models have been deployed. Looking for model drift and divergence in your feature spaces and just staying on top of generally maintaining ethical use of data, maintaining reproducibility of the outputs that you expect with your artificial intelligence model. Having a general understanding of why a model's working and not just deploying it because it works are the kind of thoughts that come to mind when I think about AI data governance. But I think that field as a whole, explainability, ethics and AI, monitoring your feature spaces, the amount of investments that so many different companies are making in just the enterprise software available to abstract out a lot of that for you, it just goes to show that focus on your differentiator, but make sure that you cover those components when it comes to data governance, data lineage, the right metadata for your information, the right version control for your model, the right model drift and feature monitoring and things like that.

Satyen Sangani: (28:00)
What is your approach to data governance then at Orangetheory? How many people are involved in that work? How many people are using data? How do you think about using data at Orangetheory? Tell us about your data state and usage patterns.

Ameen Kazerouni: (28:13)
So when you think about data governance and particularly at Orangetheory, instead of thinking about how many people are involved, I like to think about which roles are involved. Because I think there's a certain accountability that everybody has to data governance, whether it comes to records retention or whether it comes to access control or the way you disseminate, the way you share what you take out of our ecosystem, out of our network. There's these kind of policies and governance and just general corporate hygiene in place that everybody participates in together. But when you think about the kind of data governance office that we have at Orangetheory, it's an amalgamation of our analytics leader, our engineering leader, a security leader, and data privacy and protection and policy from the legal team as well. And then that group together is effectively got shared accountability of effectively a two-pronged approach of maximizing the value that we can extract from our information while minimizing the risk associated with the data that we store.

Ameen Kazerouni: (29:15)
Whether that's a matter of using the data intelligently, democratizing access to the data, making sure that we have strong stewardship and governance in place so that everybody accesses the same information and talks about data the same way, or it's the right security protocols, the right roles and authentication protocols, the right encryption layers in place, to ransomware prevention, etc., to make sure all that's in place to minimize the risk the data is at. It's just the kind of balance across those two layers of optimization.

 
 

The Orangetheory data governance workout

Satyen Sangani: (29:40)
Do you think of data governance as a singular process, multiple processes? Like is there a single committee, is there multiple? Like how do you actually organize around it?

Ameen Kazerouni: (29:50)
I see data governance as a cultural shift more than a kind of process that you put in place. There's of course tools and approval workflows and checkpoints that we've built over time, like DAG orchestration that we've instituted with the right failure points with software, like great expectations to make sure that once data refreshes, it makes sense and it meets certain conditions. Yes, there's process flow where failout and roll back a DAG or rerun it, figure out what went wrong, etc. All of that's an investment that you make gradually as you evolve as an organization.

Ameen Kazerouni: (30:24)
And there's, again, focus on the differentiator, trust the experts, these all solve problems, but the cultural shift is what will need to be unique to each organization. One thing I've found with data governance in a more abstract sense is that you can lay down as much rules and regulation and process as you'd like, but without buy-in across the various verticals, across the organization, data stewardship becomes really hard to enforce. Like making sure that when we talk about a metric, that definition of that metric that's driving strategic decisioning is shared across every vertical in an organization. Everyone has the same understanding of what a particular metric means is hard. We have a dedicated committee to where everyone has a voice in establishing a new data definition, the business logic that goes behind a core metric, like how do we define leads? How do we define conversion? How do we measure the number of workouts a member is taking per week? There's different implications from financial reconciliation at the franchisee level to try to measure improvements in fitness regimen and response for the fitness team. And we wanna make sure that metric can either be shared with a shared understanding or to new metrics support, something as simple as that requires a cultural shift to where now we've got our CFO...

Ameen Kazerouni: (31:40)
Our BB of fitness, our COO, our CMO all either in a room together, all the right delegates in a room together once a month to sign off on new data definitions that are gonna go into place, a standardized glossary that everyone has access to. And then that kind of information is what helps on the more linear, process-oriented, data flow, ETL, DataOps part of the equation. But that buy-in comes from the fact that everyone feels like they're part of that decisioning process. And similarly, for data privacy policies that are constantly evolving, we're in 24 countries, so GDPR, PIPL, the APPI in Japan, the Indian data privacy protection laws, the constantly evolving landscape across the various countries.

Ameen Kazerouni: (32:26)
So maintaining a strong touchpoint with legal and outside counsel and making sure that we are compliant wherever we need to be compliant as data flows, not just in America but outside of America as well, is something that requires a cultural shift and strong collaboration that goes beyond just the tools that you put into place to enforce data governance.

 
 

Driving Orangetheory’s evolution with data

Satyen Sangani: (32:46)
How big was Orangetheory when you joined it?

Ameen Kazerouni: (32:48)
I joined Orangetheory in October of 2020. So that's a difficult question to answer because there was a pandemic that was in full flight. So when I joined Orangetheory, the brand was strong and the dream was real and I was very excited to become a part of it. We were about close to 1,500 studios across all 24 countries.

Satyen Sangani: (33:07)
And the reason that I asked that question is you mentioned the words “cultural shift.” Was it a shift when you came in to sort of implement some of these processes? Because one would look outside-in at something like Orangetheory and say, “If that company's not data-driven, then what company would be?” And I would imagine yet on the inside there's still changes to go do because data's not just about having data, it's about using it appropriately. So what shifts have you seen or been presided over or what were the changes that you've seen since you've been there?

Ameen Kazerouni: (33:35)
That's a great question and that's why I like breaking it up into maximizing value and minimizing risk as two separate pieces of the puzzle. Orangetheory did a great job at minimizing risk because security, data privacy, those are your instinctive top of mind when you're a data-driven technology-tracked organization. But when I came into Orangetheory, I was the first chief data and analytics officer that Orangetheory had ever appointed. And that's that shift in, okay, we've, from an engineering and security perspective, locked down the data that needs locking down. How do we maximize the value from the data we've collected to now drive the next evolution of not only the member experience through machine learning and personalization, but also business planning and strategic insights and data-driven decisioning from an operational perspective with the rest of the organization, including the ops team and the marketing team, etc.?

Ameen Kazerouni: (34:28)
So I think that shift of, it's not just about minimizing risk, but it's also about maximizing value in a way that keeps the risk where we need the risk to be, required the incubation of a data organization whose sole purpose was the data. And I think that was the investment that our CEO decided to make when they brought me on in that role. Since over the last three years, we've incubated and grown that function and built that muscle and kind of merged the data and technology organizations into one because now that it's built up, these teams are able to move with a lot more velocity and efficiency as one team, which is why I'm in my new role as chief technology officer with those functions amalgamated back into one organization.

Satyen Sangani: (35:12)
Brilliant. What changes would you like to now make going forward? What are you seeing as the challenges in front of you today with data and what's the next horizon or strategy?

Ameen Kazerouni: (35:21)
This is an interesting one. It's like we unlock the data and we opened up the candy shop, right? And now it's like, everyone wants the data, everyone needs access to the data. So now it's become, it's almost like a rate limiter at times because we can't unlock the data and make new assets available fast enough to stay ahead of the need for data to be part of the decisioning process. So it's almost backward in the sense that it went from, okay, data's not accessible, new data assets are coming online, one run for the time to, hey, a new feature launch, the associated data asset needs to go live at launch because what's the next decision going to be?

Ameen Kazerouni: (36:01)
And that's a great problem to have because the next iteration of the kind of data journey at Orangetheory is going to be driving efficiency, driving optimization, creating the kind of guardrails to which we can move faster while sticking to our first principles on how we want to minimize risk, maximize value.

Ameen Kazerouni: (36:20)
And I think that's gonna be an exciting journey to go on. That's one part of it. The second part of it is kind of what we alluded to and teased a little bit earlier, is taking that data that has been really, really powerful at optimizing that 60 minutes to make it the most efficient 60-minute workout there is to across workouts, evolving the product, the digital product, and evolving that digital experience to leverage data to become even more prescriptive and predictive in how we go hand-in-hand with our members on their fitness journeys.

Ameen Kazerouni: (36:54)
So those are kind of the multi-pronged approach. And Orangetheory has only got a couple of hundred international studios. So as we turn our eye to growing even more aggressively globally, and there's a lot of articles out right now where we've started making some serious moves in various countries around the world, the data landscape and the governance architecture needs to be ready to evolve at the pace at which it'll need to evolve to keep up with that. So there's work to be done, Satyen, that's what I'll say.

 
 

“Creating a mess but controlling the mess”

Satyen Sangani: (37:20)
Yeah, for sure. I mean, this premise that as you have more data out there and as you unlock the candy shop, there's more data is what we see time and again. And what's interesting is you get more data out there and people are like, “Of course I have a question and there's a question behind the question and there's a question behind the question and therefore I need more data.” And then the tension becomes that you produce all these data sets and in fact that you think that more control and more governance leads to less sprawl, but it leads to more sprawl, more curiosity, which is exactly the conditions you wanna create. And yet there's this sort of tension of like, creating a mess but also controlling the mess. That I think is inevitable and constant.

Ameen Kazerouni: (37:58)
I think it's a success measure. Almost like each time you unlock a layer of efficiency with governance and a process which makes more data available faster and more easily, you don't arrive at the finish line, just like your fitness journey. You’re just like, “Oh, there's another destination to get to. Cool. Let's get working on that.” And I really enjoy that part of this entire conversation and I'm almost borderline embarrassed that as the chief data officer or chief technology officer of Orangetheory, I had not drawn the parallel to a data journey with the fitness journey. So thank you for that because that's something I wanna lean on, I think, more and more as I talk about data.

 
 

The future of “connected health”

Satyen Sangani: (38:36)
And as a leader at a data company who's trying to get more fit, it's a parallel that came to me, certainly at the end of the workout. So right back at you, before I let you go, I'll ask a couple of things. You came to the United States from India when you were 16 years old and started in college at 16. Is that right?

Ameen Kazerouni: (38:54)
That is correct, yes. Yeah.

Satyen Sangani: (38:56)
And as I understood your journey, you went into a masters/PhD in bioinformatics and then after that sort of ejected, decided to go do machine learning, but now are sort of in this bioinformatics domain, broadly described. It seems like it's such an incredible full-circle journey. Where do you more broadly see the world of fitness and information going? You mentioned the blood glucose monitor is one great use case, like it'll tell us what to eat, it'll tell us what to do. How much do you think this connected health world has yet to do? And where is it yet to go?

Ameen Kazerouni: (39:31)
I think it's here to stay. There's decades and decades more of transformation after transformation that we can expect to see. I think that there is a nearsighted view of a lot of the wearables that we have giving us more and more information and a better understanding of our health and wellness levels. And I think similar to data access, leading to more curiosity and more questions and questions behind the questions as we realize just how little we know about our own health and wellness and how far we have to go as a society to get back to being healthier, connected health, connected fitness at home and brick and mortar experiences are gonna continue to become more and more a part of the consumer's day-to-day life. I think that the idea of a person like me, whose Garmin or a Whoop CGM being the abnormality rather than the norm, is likely gonna shift as these pieces of technology get more cost-effective to access and the insights get more consumable to understand.

Ameen Kazerouni: (40:35)
I think what's missing, and I think this is missing across the board, is a lot of the tech is getting very good at telling you what's changing and whether it's good or bad. What's not happening is no one's able to efficiently tell you what you can do about it. And I think that's the next inflection point that we need — not just the numbers are moving up and the numbers are moving down, up is good, down is bad, down is good, up is bad. We actually need to know what we can do about it. As this information becomes more readily available, I think there's a growing hunger on, well, what's the prescription? It's like going to a doctor's office and they're like, oh, well you have the flu. See you in a couple of months. You're like, “Well, how do I feel better? What do I do to improve it?” And I think that's a big focus that we have is not just about telling you what's going on, but giving you the prescription to impact it. And I think that's the next kind of evolution that is to come.

 
 

The data journey: keep it simple, keep it inclusive

Satyen Sangani: (41:28)
Makes a ton of sense. I talk to a lot of data leaders or a lot of leaders who are wanting to embark on a journey of using data more often. What case would you make to them in order to get started on that journey?

Ameen Kazerouni: (41:41)
I think that the two things that I would stress, without a doubt, as being the most critical to keep in mind when getting started on a data journey is one, start simple. Data governance and access to information, identifying areas of opportunity where you can automate already time-consuming and expensive tasks with data is going to very quickly show value, rather than being, oh, we need X million dollars for this generative AI solution that's gonna deploy this conversational agent to do X, Y, Z. Start simple, build up. And the second one I would say is bring your domain experts along for the ride. We very often, when we think of AI or machine learning, think of these AI first organizations like a Cruise or a Tesla or a Google Photos, even, or whatever these applications that use AI natively as part of the core product offering.

Ameen Kazerouni: (42:38)
What we don't think about is that 99% of other companies who are not AI first, they're not even digitally native, they're not even tech organizations, they're just companies that heard about data, collected a bunch of it, and they're data rich, but they're insight poor. And I think those companies are where the true data revolution is gonna happen. Those companies unlocking that potential and just getting wildly more efficient at what they do is where that revolution's gonna happen. And for that to happen successfully, the data leader and the domain expert need to be working together in tandem rather than the data leader coming in as a threat that's going to replace parts of the domain expertise. I think creating a collaborative environment there, just like we did at Orangetheory with the fitness team, that could have been a contentious relationship or that could have been a beautiful one. And we've landed in what I believe is a beautiful relationship because we complement each other rather than trying to swim in each other's lanes, we enhance each other and I think that's the kind of collaboration that will drive that kind of multiplier. So those two would be my key piece of advice. Keep it simple, keep it inclusive.

Satyen Sangani: (43:49)
Keep it simple, keep it inclusive. Nothing more to add to that. Thank you for your time, Ameen. This was awesome. I really appreciate it. I think our listeners will, as well, so we'll all try to get a workout in and maybe use some data today.

Ameen Kazerouni: (44:01)
Love it. It was great talking to you, Satyen. Thank you.

Satyen Sangani: (44:08)
When it comes to data and even fitness, perfectionism is enemy number one. Perfection is impossible because we'll always be striving for that next level of performance no matter how much data we have or what shape we are in. Let's take a page from Atomic Habits, the bestseller by James Clear. Clear recommends that you focus on creating individual habits that will empower your customers to make better decisions around data usage — or in a means case, long-term health. You also don't need to make huge investments in things like AI right away. It's much more powerful to start small, ensuring you have the right level of accessibility and democratization of information in place. Then, you can start your journey to becoming a data-driven and AI first organization. Thank you for listening to this episode and thank you Ameen for joining. I'm your host, Satyen Sangani, CEO of Alation. And Data Radicals, stay the course, keep learning and sharing. Until next time.

Producer 1: (45:01)
This podcast is brought to you by Alation. The role of Chief Data Officer CDO is more vital and challenging than ever before. Alation offers a vision for building a strong data culture that empowers people to find, use and trust data. Download the CDOs Toolbox, Seven tips for building a successful and sustainable data culture. A white paper available at alation.com/cdo-tools.

Other Episodes You Might Like :

Start with Story, End with Data

Ashish Thusoo

Ashish Thusoo

Founder of Qubole and Creator of Apache Hive

Subscribe to the Data Radicals

Get the latest episodes delivered right to your inbox.

Marketing by