Lessons From Disrupting Two Archaic Industries with Eren Bali, Carbon Health

Healthcare operations are so complicated. Just giving a piece of software to an existing practice is really not going to cut this. It requires significant operational rethinking as well. It's not just about software adding to an existing location. So to me it is and still extremely obvious that you just need to be a full stack company to be able to disrupt healthcare in the way we want to do.

I need some traction? You need some traction. Let's get some traction.

Hey what's up innovators, entrepreneurs, visionaries and disruptors. This is your Traction podcast host Lloyd Lobo. We're a community of over a hundred thousand people just like yourself on a mission to help you get the methods, the money and the madness to explode your business growth. Featuring stories and tactical advice straight from those who've done it before, like Shopify, Twilio, Asana, and many more, this episode is brought to you by Boast AI. Each year the US and Canadian governments give out billions of dollars in R and D tax credits and innovation incentives to fund businesses like yours. But the application process is cumbersome, prone to frustrating audits, and receiving the money can take up to 16 months. Boast AI gets you access to research and development tax credits and innovation funding opportunities without the headache and red tape. Join thousands of North American companies leveraging Boast AI software to maximize cashback. Check out B Oast AI the this episode is also brought to you by Launch Academy, an international tech hub that provides mentorship, resources, network and the environment for entrepreneurs to launch, fund and grow their startups. Since 2012, Launch Academy has incubated over 6,000 entrepreneurs, of which 300 have grown their startups past Seed and Series A and have collectively raised over a $1.2 billion in funding. To learn more about Launch Academy's programs, check out LaunchAcademy CA.

Fantastic topic here today and I don't think any of us on the planet could find a better speaker for it is how to disrupt two archaic industries and lessons learned from doing that and Aaron Bally is a serial entrepreneur founding multiple companies. Currently he's the CEO and co founder of Carbon Health and and leading its team towards his vision to democratize healthcare and make quality care accessible for everyone. Under Aaron's leadership, Carbon has grown its staff sevenfold since January, increased its clinic footprint from seven to 27 clinics across six states, introduced virtual care to 16 states, and facilitated over 500,000 Covid tests. Prior to that, Aaron paved the way to make education accessible by founding udemy. And UDEMY is the world's largest marketplace for Online courses with over 50 million students worldwide. And in his youth, Aaron was a star mathematician, winning awards in mathematics, physics, computer science and chess, including the International Mathematical Olympiads. And he, he still plays chess today and he loves it. And you unwind with chess after a long day, right?

Yeah, I do actually. It's funny, I hear my rating, so my rating is like it goes up if I can have time to play during the day. But then usually I have three small babies, I can only play after all of them sleep. So if I play after midnight, I start losing my score. Like, it's like playing competitive chess like after midnight is not a very good thing to do. But that's the only way I can relax after the end of the day.

Awesome. And you, I think you have maybe some people saying marhaba.

Aaron, it looks like we have some.

Turkish and as I was saying, I was born in Kuwait, so marhaba is a word we use in Arabic as well. Aaron, super excited because disrupting two archaic ancient industries and getting both those companies to hyper growth, there should have been a lot of lessons and I'm excited to get into all of them. But before that, how are you doing in Corona times?

I think we are doing well. So the reality is this pandemic has been devastating for the entire world. So you have to look at everything. Comparatively. Comparatively we are doing really well. Our company, Carbon Health kind of found itself in the front lines of it. So it has been a very, extremely busy year. So like the things you mentioned, going from seven clinics to random right now, 40 already, sorry, 30 already. And going from 100 people to like 800. Those all happened in the last eight months. It hasn't been easy. And Udemy, my other company, is also doing well. So I'm still the chairman of that. And I also have three small babies. So reality is like we were not going outside anyway. So it didn't change our personal life that much. Because when you have three small babies, like 1 and 2 years old. So that's really majority of your time anyways.

Yeah, definitely. I have two little kids as well. And now we're homebound. Awesome. Udemy effectively pioneered a whole new industry and has been super successful. 50 million students worldwide. What made you decide to start Carbon Health and what else did you consider?

Yeah, so I was always interested in education and healthcare, so the healthcare focus was not. Was already the plan. And my sister is a physician, so that kind of helped. I did a lot of conversations with her. So really I think I had maybe eight to 10 different ideas. In and out of healthcare space. And most of the time like my plan was to just maybe found a company who can, who's working on that field. The reason I decided to start Carbon Health myself is there are a lot of technology companies in healthcare space. But I was really disappointed that they were all exclusively focused on this young, affluent customer base. There are simply hundreds of startups competing to become the healthcare provider of choice for Google employees. But nobody was really building technology to help provide better care to an average teacher, average retail worker in this country. And that was really disappointing for me because those are the people who need the healthcare improvement the most. And in reality, when you provide a better care with a much higher cost to an affluent customer base, that's really a hospitality business that doesn't even require that much technology. The place where you need the technology most is making care so efficient so that within the money already being spent, you can provide a really high quality healthcare. So that is really to me like why technology even exists is like creating reducing barriers to access. It's not just going to provide the more gimmicky experience to already rich people. So I was essentially disappointed with the market, not really not having given a single high quality tech startup targeting this problem that I can invest. So I said, and I realized this is actually a fairly big undertaking. You need the team who is bold, capable of building a lot of hardcore technologies, but also doesn't mind getting their hands dirty. It doesn't like this regulatory complex business. It's operationally complicated. So for the tech founders who just want to just build some gadget like here and there and do some ML project, like there are a lot of ideas for them, but it's kind of required with like the craziness that the obsession to healthcare accessibility in addition to the technical capability. So then I really, I realized there are really not that many teams who have that overlap. So I decided to tackle this problem myself. And here's the other thing like, and I also am a savvy entrepreneur as well. I also realized if somebody can actually build the de facto healthcare provider for the mass market in this country, this could also become one of the world's largest companies. Like somebody, this is what I told like a friend who asked me the same question. As I said, if somebody told me that in 10 years the world's largest company will be a healthcare company, I would totally believe that. Because if you think about it, that just the primary care business in the United States is bigger than the entire smartphone spending. So smartphone has marketed companies like Google and Apple. So the biggest healthcare company could actually become much larger than, much larger than Apple, Amazon, Google, today, hundred percent.

I agree and couldn't agree more. And in fact a lot of my family members are in healthcare. My wife works at Stanford. And you see this as healthcare is not democratized in the U.S. what is your vision for carbon health like looking into the future four or five years from now?

Yes. So maybe starting from the early idea. So we are technology enabled primary care provider. So we have clinics with urgent care primary care, we have virtual care capabilities. So if you look at how companies like Amazon, Uber and Airbnb disrupted their respective service industries because disrupting a service industry is very different than disrupting for example the music industry, because service is a different base. So what they really nailed is using a lot of technology and better operational design, services, service design to provide very high quality customer experiences at the same or lower cost with what existed before. But Amazon is both better experience also cheaper than pretty much any other way to buy it for. Similarly, Airbnb use all this like the hospitality of a home, but it's still cheaper than hotels. Uber is the same thing. It's better and lower cost than cabs. So first, like we said, our first chapter is how we provide really high quality customer experience at the same or lower cost to what people are spending. And that's what we have done over the last four or five years. We re engineered every single aspect of care delivery, right. So from the first patient interaction of the patient with the system all the way to depths of clinical decision making, care plan management, things like that, it's a we re engine everything from the ground up and we essentially own the entire technology stack between patient provider and this actually has never been done before. Like a lot of people have failed at things that look like carbon. That's partially because nobody has really tackled like undertook, nobody's really undertook the entire tech stack from the ground up. We did. And the way we actually started is we opened a small primary care clinic inside the office and I actually made it forbidden to even look at third party healthcare software. So it will truly start from the first principles in this design. So over the last four years at this point the company's coordinating to economics are just unbelievably strong. And we had this very massive growth phase. So we had seven clinics last year, we now have 30. Next year we'll have 150. So we have 120 clinic already actually lease signed in the process. So we can launch next year and then by 2025, you'll likely have 1500 locations nationwide. So you'll become by far the biggest primary care provider. So we are surely going through the really rapid expansion. But this is actually what I mentioned is largest primary care network. That's not even the end goal. That's really the infrastructure we are building for modern healthcare system. So the thought was we just imagine what we want healthcare to look like for an average person in this country, not just for the rich. And that meant it's really strong technology driven primary care, but also we are building networks or spatialists and everything. So it will feel like almost like a Kaiser Permanente but design with technology from the beginning. So just we are building an entire network around it. And the primary care layer, we are building platform but then we are building a lot of omnichannel care options from virtual care to at home care. The device hardware, integrated solutions, we are building more and more value based solutions and actually also building the infrastructure for modern healthcare companies to operate on. If you want to build an amazing kidney care solution, you can actually build that with even carbon's underlying essential blocks. Building blocks, that's a feature will provide amazing care directly. We also want to accelerate the transition to like modern care. So like Today it takes 17 years to for diagnostic improvement, for therapeutic improvements to go to mass market. We just really want to accelerate that.

Fantastic vision and I wish you nothing but the best to realize it as soon as possible. But before we dive into some of the more granular things that you've done here, what were some of the key learnings from Udemy that you feel made you a much better entrepreneur today? Because for people who don't know here, Aaron moved from Turkey to the US in 2010 and he was part of Founder Institute and I think that's where you met your co founders. So literally coming up with an idea, going to like a hackathon style pre accelerator with an idea and building a unicorn from there to now building another one. What were your key learnings from udemy that was.

It's not very conventional, right? Yes. No, I think probably the biggest learning was really just thinking very realistically about what the real problems are in the world and not getting sucked into the convention capitalist thinking. So when we first started it was a roughly simple idea, right. So there are hundreds of millions of people who just want to better their lives and learning a new skill is typically a good starting point to just improving their lives. It's very critical. But we built the platform so that people can easily teach and learn online and the feedback from every single investor was that it is impossible to build the education business where the consumers are going to spend their own time, own money to learn the subject without a diplomat. So most investors thought this was a crazy idea. Like they were saying, like nobody's going to invest their own time, own money to learn. People only learn because they have to get a certification, they have to get a diploma. That's really the only way to make business. So that's why we had maybe hundreds of investors pass on investing in the idea. And even if some of the investors who did invest actually they thought they should still cut, we wouldn't be able to make it work with consumers. But maybe thought they're thinking, well this might actually work in maybe enterprise. But I think to a lot of people's surprise, we actually built the world's largest learning destination with the exact same idea that everybody thought was impossible. Because here's the thing, all these people who said it's impossible for people to spend their own time and own money in learning, this is very easy to disprove. Books exist, people buy books and they didn't spend their own money and own time in learning a subject and there's no diploma attached to each book. So to me it was crazy that they didn't. Something as simple as the fact that if this is happening for books, wouldn't make sense to have better books, which are online courses that are objectively much better than books or learning a subject. You can put videos in it, you can put articles, assets, you can have a community around it. So it was so obvious to us, it was extremely unobvious to all the other investors and honestly even most of the people in education space like this was a horrible idea. But it ended up became becoming like when everyone thinks it's a stupid idea and you think it's not and you actually turn out the right. That's actually the recipe to become a master to rebuild a billion dollar plus company. So it's. You have to be somewhat controversial and right at the same time. And when I started Carbon, it was actually a very similar situation. I just wrote an article about that that I'm going to publish next year, just going through the details of this body. So even though I had started a multibillion dollar company, we really struggled to raise money for Carbon Health because I was starting from the first principles. The first principle said, okay, you want to start this from the primary care services because that's a consumer decision. If you want to win a market with better customer experience, you need to be in a market which is a consumer choice, it cannot be just provide resignment to them from another angle. And then I was very convinced that there will always be a physical component to healthcare. I don't see pure telemedicine players ever being the dominant player. At least 70% of the care still requires some sort of physical interaction. And then lastly, healthcare operations are so complicated. Just giving a piece of software to an existing practice is really not going to copy this. It requires significant operational rethinking as well. It's not just about software adding to an existing location. So to me it is and still extremely obvious that you just need to be a full stack company to be able to disrupt healthcare in the way we want to do. And then this has to go for mass market which means that you have to be able to accept insurance, you have to be able to have physical location and you have to really improve the efficiency using technology. To me all of these are just again, I don't have a moment of doubt in any of these things. But still, even though to me just nothing else makes more sense. But all VCs, all investors, they hate brick and mortar. They don't want to invest in something which has physical locations. So there's still a really big uphill battle. Despite being the founder of a multibillion dollar company, I just had investors up there, investors pass on the idea because they thought it was capital intensive, which again was also stupid because yes, we had to spend money on opening connects, but in exchange we actually spent almost no money acquiring customers. Our customer decisions rate. There are tile tech startups which spend hundreds of millions of dollars every year digital acquisition to Facebook. Google investors just gladly put money into just investors gladly give you billions of dollars to burn with Google Ads and Facebook ads and they thought it was expensive to build clinics, which in reality is not opening a clinic less than a million dollars. Of course you open it up in thousands of them. It's just still not all that much money compared to what you're building. So essentially this kind of like whenever majority of the people disagree with something which I see as I consider like extremely obvious, I get really excited because that's really a recipe for a massive company.

What were some milestones you said? Because it's a lot of. I'm surprised actually that you said after building Udemy, a multi billion dollar company, you were having hard time raising money from investors. I'm mind boggled here. So that could as an entrepreneur that could set you back mentally in a way, right? Hey, what am I wrestling against, is there, is there really something here or is it a figment of my imagination? What milestones did you set when starting Carbon? Was there like a go no go timeframe or you're like, no, this is all it.

Yeah, it was all it. The thing is, I think between us, the doctors, like people in the company, saw the story clearly. Like we were not really self doubting whether this is the right idea or not, right? So we have such strong conviction there and we just made some iterations. Like initially we were planning to partner with existing practices and realized over time you realize it doesn't make sense. You have to be full stack. So it is less about timing, but it was more about the difficulty is eventually when your company is so large, like people who the naysayers kind of change their minds, right? So like you, they made the naysayers. But at this point, and you have 50 million customers growing so fast and eventually they realize, of course it makes sense to build a massive kind of learning destination. Of course it makes sense to have everybody like instructors. But in the early days, like this whole like nobody's going to spend their own time and money without a diploma, that type of cynical be like disappears once you are growing fast. So at Carbon, the additional difficulty is that we knew that we can significantly improve the efficiency and quality of care delivery using technology. But this type of improvements that you build, you see early signs of promise, but it doesn't show up in unit economics until maybe three, four years of iterations. So for example, the moment we launched after two months, it was obviously an amazing customer experience. But then why do you care about good customer experience? You have to just say good customer experience brings higher retention. But if you only launch a service for three months and it's a healthcare service, right? So people don't use healthcare five times a month, so you just use maybe a couple times a year. So you need to wait several years before you can improve that retention. Like the better customer experience turns into better retention. Also you want to prove that better customer experience turns into better word of mouth growth, which again shows up like several years down the line. It doesn't immediately become visible in numbers. And then what is even retention? With world, you have to eventually be able to argue that all these improvements you're making is significantly improved the unit economics of the individual physical location. Very significant. So initially it's the theoretical suggestion that you'll be able to improve the margins in a significant ways. But then it took really three, four years before you have old operations, supply Chain the clinical everything dialed down so much because when you, what often happens is you build something which is much better in certain aspects but then you're not really optimize like the kind of other like mundane details. Yeah. So if you are worse than other companies in certain other kind of standard areas really. Because MOOC says you do a much better job in the things which are hard to do and then usually do a shitty job in things which are easy to do. Like you're doing a really good job with customer experience, retention, the techno, like the providers are really loving to be a part of it, but you're doing a cool job is like supply management and things like that we should not have really had bandwidth to work on yet. It took several years before we just dialed down everything. And now at this point, by 2020, even if you just don't ever look at the technology, even if you never install our application or talk to a single, you just look at the P and L right now you just understand what did you clear? But that took four years. You just kind of survive for four years until the numbers were undeniable. Because when people don't fundamentally believe you have the right idea, you just need to prove that numbers and that does sometimes take a lot of iterations.

Yeah. And so the key lesson here is when you're tackling and trying to disrupt a regulated industry like healthcare or education, it takes time and you gotta be patient and you gotta focus heavily on delivering a 10x customer experience. I guess. Exactly.

And the reality is generally like you do something which is already hyped. Like you build a machine learning startup today or you build like even in healthcare you build a telehealth startup. Like there's a lot of hype around it. Right. So those, there are certain things that are like hype subjects. If you do one of those things it will be easy to raise money and like the world, the life is not very hard. But if you what you believe like kind of disagrees with the common investor sentiment, then you should just expect that this is going to be an uphill battle. And luckily you only have one or two backers each time. So you have one investor each time and you can keep going. So you won't have people get throwing your term shares and you won't have those really close valuations necessarily while you're pooling things. But none of those matter at the end of the day. What matters is are you right, are you going to be pooled right at the end of the day? That's really what matters carbon. That's what happened. It was four years of like uphill battle kind of fighting our way to it. But then eventually, at this point, it's so obvious. Like now it's just too obvious. Right. To everybody. So, yeah, I think it's again, like sometimes like you, sometimes you do something that agrees with the common sentiment and things are easy, sometimes you don't and like that, then you just have to put on a fight.

Yeah, yeah, definitely. And then I guess comparatively, Udemy, I think by the time you guys were four years in, you were already close to 10 million or more in terms.

Of like the users or revenues. Yeah, yeah, I think exactly like 10 million users. It's a different situation. So like I said, it's one thing to have somebody take an online course, it's something else to have an actual patient.

There's further trust, there's regulation, further regulation and whatnot. In your opinion, what are some key levers that you prioritize to scale the company 10x, like key things you did?

Yeah, I think the most important one was building a product that has true, like mass market appeal. So one of the other early challenges at Udemy was that like when we maybe couple a year after we started, Coursera and Udacity launched and this concept of like massively, massively online kind of university courses became very popular. So essentially the, the super hype subject was how do you just bring the Stanford machine learning course and make it more accessible to everyone? So that was. It's not a bad idea. So there are definitely a lot of people around the world who would want to learn machine learning from a Stanford professor. But what we insisted on was like, let's go to a small percentage of the world. Far more people need to learn some Excel so they can maybe take a kind of office job. They want to learn maybe some amount of like some basic programming skills to work at a local development shop. So essentially the more daily practice the main skills was actually it was appealing to far more people than machine learning courses from Stanford professors. So we kind of. And there was a time where all of a sudden education was very like on 16, 2010 when we launched, but then in 2012, 13, it became a very sexy subject, but only but exclusively around university kind of courseware concept. We insisted that just allowing average people learn average skills was actually a far more impactful thing to do. You just stayed on course focus and that's really, I honestly like there were a lot of growth tactics we applied and everything as such, but really at the end of the day. We grew fast because we had a. We had a product that was appealing to billions of people around the world, not just like a small elite kind of group around Silicon Valley and maybe other places.

What made Udemy and then again at Carbon Health, that you have product market fit. What are some key things you measured?

I think product market fit is one of those things where when you have it, you know it. So if you're not sure whether you have it, you probably don't have it. I know there are some people that just say, ask your customers whether they would be disappointed if you shut down the service. I know there are methodologies these people suggest, but it's actually a little more subtle than that sometimes because you know that your customers really love your product and they're super appreciative of the fact that you're beyond them. So that's really what happens at Udemy. I think we saw it relatively early on, in 2001, 11, one year after we launched. When we first started, people were producing some really low quality, expanding courses on Udemy. So we did not have product marketing back then, but then we launched a course called Raising Money for Startups. It was essentially teaching founders, mostly around founders institute, how to raise your first angel capital, how to build a deck, how to talk to investors, things like that. It was a kind of really tactical, kind of step by step educated course. And we just distributed towards a newsletter and we had I think 100 people that buy that course at $30. So that's, that's $3,000 of revenue. It's not that much objectively, but I just realized that all these people who bought the course, they loved it. Like we actually, like you have ready to refund them. And I thought they would actually maybe watch the first section and just said this is not worth territory dollars and ask for it. And nobody asked for a refund. And a lot of people said this is by far the best $3 they have ever spent. So that was actually the moment I realized, okay, there's product markets where people are. And here's the thing, the reason I think that was particularly interesting is there is ridiculous amount of free entrepreneurial content already. Ton of people do really fancy video conferences and other stuff like published on YouTube. There's so much free content about entrepreneurship, but for some reason people still willingly wanting to pay $30 for a course because it wasn't the kind of motivational bs it was actually tactical learning around how to, like, how to send a cold email to an investor, things like that. So like that actually just that one course we sold like I got 100% convinced that Udemy was going to be a massive business at that moment where 100 people bought their course at $2.

And then similarly at Carbon, when did you realize you had product market fit? Because comparatively at udemy it took four years to get to 10 million users. Carbon health, it's been a four year grind and trying to get investors on board. But when did you realize your product market fit?

Yeah, the four years. Carbon is a larger company, much larger company than medium wells after four years. So technically Covid has grown a lot faster. But I think again one month into our first product launch, so what we did is we started a primary care clinic inside the office that was like the first version of the product. So in 2016 we started working on this. In September we launched a clinic and obviously like the minimum viable product in healthcare versus education is very different. At Udemy you put some online videos and then you charge people with a credit card transaction interface. That's the kind of viable product. At Carbon we had to have a physical location. Doctors, nurses, able to write prescription, able to submit lab requests, referrals, imaging, like it's a far more kind of complex and new viable product. But we launched in nine months regardless and we had both ourselves and our friends and family use the service. And like the moment even you use the first version of Carbon's product, it just becomes like you use and realize there's no way I'm going to go back to just normal healthcare delivery. Because even in the first version we made 100% of the onboarding, receiving your care plans, requesting like talking to doctors, all dumped through your application. And that experience, the moment you get it, you realize it was so stupid that we had to fill out those long forms. Those long forms is not really going to be still ask the same questions again and they tell you what your care plan should be, but you don't even have a good place to reference that. So these things are relatively simple. At Carbon, the team that spent a lot, that took a lot more time was dialing in the unit economics. So that's not just like product market as one thing, but the other thing is do you have a viable kind of business, kind of like business? Because look, it's not too hard to. I could easily hire a bunch of doctors and have each doctor see five patients a day and provide amazing care. That actually is not very hard. So what is really hard is like having each doctor see 40 patients a day and still be a really compelling healthcare. So that's like when you into economics and the customer experience business, everything is overlap. For that to happen.

Would you say this is like a three sided marketplace in a way kind of thing? Because you're like tackling the patient problem, you're tracked tackling the provider problem and then you have the insurers also. Right. Who did you first go after?

Yeah, that's actually a great question because so sometimes when people think about Minimum Vital products or Lean Startup, they're thinking about only like just launching something in two weeks. But when you do healthcare, as I mentioned, you cannot just launch something like half ass and put it out there. But that said, we still got really crazy about it. So what is the order of things that we have to prove? The first thing we had to prove was that by using this really technology driven healthcare experience, we can create a much better customer experience for patients. That was the first thing to be able to prove. The second thing to prove was we can do this while still maintaining a really good work schedule for the clinical staff so that they have to be also be happy because there are finite number of clinicians. If you're not better at hiring them than other people, you're not going to be able to grow anyways. That was the second one. And the third thing was prove that you can do this activity with the unit economics. That actually would make it easy to be in network with every insurance company. Those are really specific orders because if number one did not work, number two and three wouldn't matter at all. If number two didn't work, number three would still not matter. So we just had to put them in that order. So what we did is like we first built the end to end customer experience and the patient side, but our provider side was really rudimentary in the beginning. It was really kind of unbaked, unfinished, but it was fine. We can hire the physicians. We told them that we'll improve your platform in the future, but for now just, can you just, I know this will be a little harder than usual, but can you just get along with it and try to make things work while we're trying to build the provider side of the platform. But even with that, we essentially faked the backend with doctors doing more work than what they would normally do just to make sure the patient front end was amazing. And then once we finished figuring this out the next year, we spent a lot of time optimizing the doctor's workflow. So we made that work. And when we first started, we did not even have issues contracts and also issued billing is a very complex piece of operations. So it's complex in terms of work, but it's not rocket science. It is things that you can do when you have enough resources, but it is time consuming regardless. So we did not really want to spend any of our time dealing with issues building kind of problems because again it can be a huge time sink for a small company. So our solution was. This is going to sound crazy but we in the first year we actually saw all the patients for free. They didn't know it was free, but we were actually collecting insurance because we didn't want to pitch this as a free healthcare is not gonna work in the long run. We just acted like they are gonna charge their insurance, but we do not have the capabilities to charge insurance. So we're essentially like eating the cost for every visit. So we were just kinda learning how to provide better care. So once we figured that out, then we just figured out how to then do the insurance billing so that like we don't have to eat the cost of every visit. And then eventually like the operational things got optimized. So it was just like a step by step process to kind of prove each critical because you want to start with the more the biggest risk is like you cannot improve the customer experience with technology. So we had to prove that first and then provide the workload and then you need to go to oral clinic.

I want to get into how AI and machine learning fits into all of this and how are you guys using it to improve better customer or create better customer experiences?

Yeah, so I'm glad you were asking because Carbon Health is one of the companies which actually leverage machine learning AI the most. We actually don't like. It's not our branding. Our branding is being like amazing healthcare. It's not that we are going to be an AI driven healthcare company, but behind the scenes. So the first thing you might realize as a cop patient is when you book an appointment, typically there are all these questions that Dr. Would be asking you during the visit in the first 5, 10 minutes of the visit. So we ask all of those questions to our bot. We call this Carbon Bot. Carbon Bot asks all those questions to you and it takes all the answers and put into our health record system automatically. And the main advantage, there are two main advantages. One is like productive did not waste much of time like asking and inputting routinely standard questions. But the second advantage is now we have all these things in heavily structured data form. So we can now use this data. So we have the data there. We also have the previous care platform that our doctors generated, we have the orders that generated. So essentially everything in carbon is as structured as humanly possible. So our machine learning algorithms can now do multiple things. For example, they can predict how long the appointment is going to be based on all the information we collect. So if you for example, say you are booking an appointment for a sinus infection and let's say you're a 20 year old person, don't have really any kind of major issues, and from the answers you gave about your sinus infection, nothing is really alarming. Your visit might really very well be like five minutes at all. So you don't really need anything more than 5, 10 minutes with the doctor. So that might be very fast. But imagine the same reason you are still sinus infection, but you're a 70 year old person with several, several kind of chronic conditions and you're not being in the clinic for the last 12 months. You have some multiple medications. And then when you say sinus pain, like you might have told that like you're also coughing really bad. There's some other things that kind of like maybe let's say you have confusion, you're like double vision these days, right? So like all of a sudden that makes the really first of a high risk sinus pain issue. But also you have chronic conditions so you will likely want to refill your medications today and you're likely because you're not come to Dr. In the last 15 months, you might have other issues that you bring up during the visit. So essentially this might end up becoming like 67 minute visit interaction doctor with a lot of complex diagnostics. So essentially our machine learning system is the one connecting all of this. We actually predict a lot of these things, like how long is the appointment going to be, what kind of diagnosis orders are we going to make? Does the patient have to go to ER immediately? So all these things actually we learn from existing data and we automate it. And that's a big part of how we make things efficient.

So what I'm hearing here is start with the customer, identify all the manual breakpoints, frustrating pieces of their experience and remove that piece by piece. And, and technology, particularly AI machine learning, are great ways to automate a lot of that friction and you'll be on your path to delivering a better customer experience. But it all starts with the customer.

I would just add one more thing. It starts with customer problems, like identifying what problems there are first. And second thing that you have to do is a big part of the investment goes to just having high quality data about the problems you are trying to solve. For example, if we did not have the platform that tracks every visit, we wouldn't even know how long the historical visits were. Like essentially a lot of the investment initially was like I said, workflow platform which knows exactly how long each employment took, know what type of diagnosis orders we did for every visit. So if it does not have this data cleared data quality, you couldn't really be machine learning. This reality is if your input output data is really clear and structured then like Dan the attributes, it's really not that hard to build a machinery model that kind of connects everything.

I'm getting some questions here on what AI and deep learning tools do you recommend or is that too nitty gritty for this conversation?

Yeah, like our engineering team makes final choices there, but I think like Spark is the one they use the most. So we're not using any neural networks yet. So like my co founder CTO is looking at neural networks. It's not so far we haven't used them. But if you want to use them, I think there are different solutions for that. But like Spark has been probably the the catchy awesome.

And none of this is possible. You said it right there that it's all dependent on the engineering team. So I want to dive into team and leadership. You've done this now few times for people starting a company today. How would you build your team? Who would be the first few people to bring on board to help get you to the next step from product market, fit to scale?

I think it goes back to kind of the ordering I mentioned. You have to figure out like so every business have a bunch of open questions. Like each open question is the kind of risk item and you need to first decide at what order are you going to de risk those things. And based on what you are going to risk at this point you have to focus on different hires. So as an example, a lot of for example people just kind of over generalize this idea that you should not focus on revenue in the early days. That's a very common advice you'll hear because you know that Facebook did not monetize for so many years, but then once they monetize it became a massive business. So I think people overgeneralize. The reason it makes sense for Facebook not to worry about monetization is Facebook's monetization is very obvious. It's just advertising. Advertising has been done so many times before. Once you have eyeballs, it's very obvious that you can monetize it through advertising because it's so obvious, so straightforward. It's okay to delay that question to much later in the company, much later in the company's life cycle. But if your problem, if your business actually relies on this kind of unclear mute economics, then you have to prove those things as soon as possible. For example, if you're a movie subscription play, you charge somebody a monthly subscription fee and then you allow them to just see as many movies as they want. Now all of a sudden there's a pretty obvious risk on does the math work here? Is that the first thing you have to just, you have to just start monetizing as soon as possible so you can prove that the math actually does work. So it's just really about the you always want very strong kind of engineering, product design kind of people. That's always for starting point. In my case I'm also an engineer product person myself. So we hired a couple really, really strong people and then I think that the first team was literally, that was actually funny was me as a co founder, entrepreneur, one very strong engineer, one very strong designer, one very strong physician that was really founding team. So we started with that. Then we hired an amazing, very well experienced nurse who understood the hospitality part of healthcare delivery. So she had worked at a lot of concierge care companies. She understood how to build a really high quality in person kind of experience. So she, she was really instrumental. We hired people who understand the branding. So again it depends on again like what risk you want to risk at what time.

Like the hardest thing to do is find good people. Any hacks along the way, running few companies. How do you find these people?

Sure, I think the biggest, the most critical one is understanding that hiring is an arbitrage game. So there's definitely like the uber engineer who has a lot of experience who went to Stanford cs. Like that engineer is a very obvious engineer. So because it's very obvious, every single company on earth would love to hire that person. So in that case like your competitive, you don't really have a competitive edge hiring that person over companies who are willing to pay five times what you are willing to pay and maybe it has just more things. So the hiring is all about like who are the kind of who are what, where can you find skill set which is going to underappreciate them. So at Udemy I hired a ton of people from Turkey because I'm from Turkey and like Turkey has like a very strong young population, extremely smart engineers. Like I was one of them. But There were almost no companies hiring, there were no sinking rail companies hiring in Turkey. So that was a really obvious competitive edge. We hired a ton of great engines from Turkey and then these days now there are maybe like 5, 10 companies you compete with even in Turkey because people realize there's like a huge pool of really high quality engineers there. So I think similarly you just need to find pockets of really strong engineers. Like sometimes there's amazing engineers, but they haven't gone to Stanford, they haven't worked Silicon Valley. But you can identify those with them because they contribute to an open source project. Right. And they are very sharp from their code and, but they don't have their traditional backgrounds. We essentially want to find those gems that most other companies like that most the companies will not be able to see. And usually one of the best tricks is just go to where you came from, like go to your university, go to your hometown, go find places, go find like people who you can get one on one feedback from. But it's not gonna, does not necessarily have a very kind of polished resume. Yeah.

Because LinkedIn outreach is getting oversaturated. I like the concept of looking at who's contributing to open source projects.

That's great.

And of course going back to your hometown, I think I might ping you after to look at building a dev team, hiring some developers in Turkey.

One of the tools that worked well for us was triplebytes.

Yeah.

So I think triplebyte worked like decently. Sms, hit and miss, like any other kind of. When you don't know the person you're hiring, there's always going to be some success rate. But the advantage was that it works a lot of people who don't have very strong resumes necessarily, but might be great engineers. That was one of the kind of tools we leveraged pretty well.

Awesome. Was most of the dev work at Udemy and Carbon done in Turkey or has it been a combination of things?

At UW it was mostly Turkey and then we moved some of the team to San Francisco and that was because we had raised a lot less money. So that was one of the bottlenecks there. At Carbon, it has so far been all like San Francisco based, but it was like since the pandemic started, we decided to hire people outside the Bay Area as well. So our team is now going to all remote, remote setup. So now we are actually hiring people in Turkey and other countries as well.

Awesome. Awesome. Yeah, because I feel like, I guess you guys have raised a lot of money, almost 200 million or so. But like when you're a younger company in the Bay Area, it's hard to have people stick around. Right. Like, they leave after two years if you've not raised another round of funding at an up round. Because they're like, what is the opportunity cost of my stock options.

Yeah, exactly. Look, I think like, getting engineers or like designers, any kind of like, product person is a similar work with fundraising. If you look at my time covering is almost a thousand people company right now, they're a ton of operations, HR stuff you have to deal with. But still the majority of my time is spent between fundraising and hiring. Those are the two things. I spend 80, 90% of my time.

And that is, I think it summarizes, as a CEO, you're effectively evangelizing the company and to investors, to customers, and to potential employees and team members.

It's not just like, I'm not just talking about hiring, like, executives. I'm actually, like, still involved in hiring, like, designers, engineers, like, any in demand skill. I'm in the interview list just to be able to attract that campaign. So I just told people, like, if there's somebody you want to at Rec, like, I'm here to help. So. And they use me. I've done like three interviews this morning. Wow.

So if they basically really love a candidate, they want them evangelized. There's no better voice than one of the founders.

Exactly. So that's like still 30, 40% of my time.

Interesting.

Yeah.

Because people make the whole thing go around, right? Like, without people, you don't really have a business on either end, customers or team members.

So that said, you said you have.

A thousand people here. How do you make sure everyone on your team is well aligned with the broader company strategy and goals? Like, how do you basically train people to make better decisions when you're not in the room?

I think this is what I'm learning on scale right now, because you did not have thousand people when I was there. So in the earliest, it was easy. You just had to set up some values and you have to really champion them. So at Carbon, for example, I actually drafted values before I even started the company. So the first one, the values ideally are not just like things that anyone will say yes to without a question. If your values like integrity, honesty, hard work, like those are just like bland. Like, just everybody could say they have. Nobody would say. So you radius ideally like your stuff a lot more like, specific to you. So like the. For me, the critical things was one pretend karma exists. So essentially the idea that sometimes you don't even know what the right strategy is. Sometimes you don't know what's about a commercial idea. But doing the principle thing, the right thing to do consistently will reward you a lot more than any kind of business calculation. That's one of the key kind of key pills of ease of Carbon is really you don't need to calculate every move. This is not necessarily chess. Sometimes you just do the right thing consistently and then it comes back to you like a hundredfold. So every person who drew Carbon in the early days was really like when people talk to it, they realize that's really a strategy. They're not, I'm not asking them to just put data behind every decision. And similarly we are respecting the crowds is very critical at carbon and hiring people who genuinely care about the problems you're trying to solve, those are very critical. So it's first important that you have a very authentic, very specific DNA that is not really something that every other company is going to align with. So we have for example, very intentionally not data for them like Carbon. Udemy was a very data driven business. Carbon is far more into design driven than data driven. So I think setting those up is great, like great. And you have to talk about them. You have to just get everybody to onboard with those ideas. Early days, which I have done successfully but I think as you go from a couple hundred people to thousands of people, you have to do a transition now the videos in your head is no longer like enough. So you have to just now I'm turning these ideas into writing. So I have this actually one of the interesting hacks for if somebody has a scaling company. So I have a writing coach. So we meet once or twice a week. So we sit down together and we write together. Essentially kind of make it a regular application forces me to write because otherwise I always want to write something but I get busy and just procrastinate. But when you have somebody else that you are hiring to help you. So you have to do and convert what I have in my head into kind of actual text and words and then we may later make illustrations about them. So I think that type of like writing is as you're scaling writing is probably the biggest leadership skill you have to gain and hiring a writing coach is a really good idea.

I never thought of that actually I hate writing too personally. And I have a lot of content just sitting there on the back burner and you procrastinate on it. It's not in your head. I should look at getting a writing coach. Do you have a Business coach at all.

I received a lot of coaching in different parts of my life. Like at Udemy for example. I still have a heavy, especially in the early days of Udemy, I had a very heavy accent. So I took excellent coaching, communications coaching. These days I'm taking media coaching as well. So if I go on like I speak fast so like they have to train me to speak more slowly. Speaking sound bicycle. If you are doing a five minute like broadcast interviews, very different than talking to in a casual manner like this. So whenever I'm like not good at something and it makes sense to hire someone to help, like I always go for it.

Definitely want to take the last few things in sort of rapid fire fashion. One is what is your biggest fear as a leader?

I think my biggest fear is like as the company becomes larger, having this disconnect between your people on the ground and your executive team. Because when you build 30% people, I talk to everybody in a daily basis and now I can no longer meet everybody or even know their name. So just having this culture disconnect is my biggest fear. And we are trying to be very proactive about making sure that it never happens. Yeah, it's important to give, sort of.

Align people with the mission and the metrics and at the same time give them the autonomy and responsibility so they're executing when you're not there. Key skills all entrepreneurs need to have like your top three or five.

Yeah, I think number one is you have to, you have to be able to see ahead of the curve a little bit, but also be realistic in doing this. Like sometimes I see ahead of the curve or what most people understand is just suggest that everything is going to be replaced with AI and everything is going to VR and AR and so saying go stuff like that's, that's cheap, that's easy. But you have to understand what you have to say is in three years this type of AI is going to replace X part of that. And these type of companies will appear like the business model will shift that way. So you have to essentially see what is like next six months later, one year later. And very much in a lot of details. It's not about telling what's gonna happen 20 years later. It's about what's gonna happen like one, two, three years later in a lot of accuracy. That's really what you have to do. So one thing you have to do be good at is like really managing stress. So no matter what you're doing, like taking responsibility has a toll on you. So I think not Everybody's well set up to be able to work under stress like that. So you have to be able to practice working under stress and not losing it. And I think lastly, you just need to. You just need to build that. Right. So I think like almost all the good founders are like builders in nature. So you need to just understand like how to go from nothing to something because that's really what most of what you. Most of what your job is like taking something which hasn't happened before and having an idea about what the first version of that looks like. So that's critical skill. Yeah.

So in summary, I think looking into the future, but like more granularly rather than high level and being able to map it today and get the steps there and then inherent passion for building in general, I think if you're not passionate about your craft, it all becomes a job. And that differentiates from a 9 to 5 to like wanting to do this for a lifetime kind of thing. Like Michael Jackson found his passion and a lot of entrepreneurs find their passion. You found your passion in healthcare. What are your book recommendations? Any favorite books?

Yeah. So this is okay. I should enrich my memories. I do. I think one book I read a long time ago, but it's still gonna. I reference is like Good to great and Crossing the Chasm. Like those are definitely. Those are like timeless books about business that I Definitely, I think 0 I'm starting with the technology like startup books. Like I think Zero to one is still great work despite even if you disagree the parts of it. The hard things about Hard things from Ben Horwitz was a great book in the healthcare space. I like the Patient. We'll see you now has a great book. I actually read a lot of like I read the book about like how to be a nourish and how to be a medical coder. I actually sometimes enjoy reading like how to books because it gives you a sense of what the fundamental skills are for different jobs. And I think there's a book called the Autistic Brain which is like very fascinating. Even if you have no interest in like autism or any subjects, use a different lesson like how your brain works.

This has been great advice.

Thank you for joining us.

I learned a ton and it's always good chatting with fellow immigrant founders who are absolutely killing it. And you're also giving back a lot to the community through events like this and angel investing. Wishing you all the best. I am looking forward to buying the stock at the ipo. I need some traction.

Thank you for listening. And we hope you enjoyed this week's episode of the Traction Podcast. If you enjoyed the show, please leave us a five star review and you can find more information and all the resources mentioned in today's episode at Boast AI that's b o ast AI forward slash blog.