Fireside Chat with Tanay Tandon of Athelas

So I would love to welcome Tenae Tandon onto the stage. Uh, Tenae is the CEO and founder of Othellis, a digital health company that you're going to be hearing all about. YC first met Tenae when he was 17 years old when he first won YC's first hackathon. Now, Othellis benefits hundreds of thousands of patients across the nation, and the company is worth more than a billion dollars. Welcome, Tenae!

Oh, today, thank you so much for being here!

Thanks for having me!

Yeah, of course. So, let's start with, um, can you tell us a little bit about what Othellis does?

Yeah, um, Othellis builds remote patient monitoring and revenue cycle management tools for healthcare systems and healthcare practices. We started primarily when we went through YC; we were primarily a diagnostics company. So we built a device for in-home diagnostics associated with febrile neutropenia patients who were on immunosuppressive medications, and that's really what got our start, and it's still one of the larger revenue lines in the business.

Okay, amazing! So, um, today, how old were you when you started the company?

Um, I did YC after I was 19, during YC.

Okay, okay, all right. And so, um, I hear a lot of folks saying, "Hey, you have to be like an executive at a biotech company before you can start, you know, a startup." What gave you the confidence, you know, at 17, 19, um, to jump out there and start a company?

I think a lot of times being an executive in a biotech company could have helped, honestly. But I think the reality is that, you know, things like YC exist now, and these things didn't exist, or, you know, exist in biotech 10, 15 years ago. And also, I mean, the internet—the fact that you can like learn almost anything. Like, we learned how to do our 510k submission by Googling. Like, and the first draft of it, the FDA could tell it was by Googling. And so, I think this concept that you can actually iterate your way to get good at something is fairly new, and it’s definitely new in biotech and healthcare. Um, and so, I—you know, I think something that's deep in Othellis's culture even today is just very first-principle thinking. There's no market you can't actually build in, and if someone tells you you can't build in it, they probably run it currently, and they don't want people to enter it. And that's what it comes down to.

Right, right. So, tell us how you got started. How did you find your co-founder? How did you find that first diagnostic concept you wanted to work on?

Yeah, so I, um, I was a research intern at the Stanford AI lab, and my co-founder, Deepa, one of my best friends since, since high school. Um, we competed against each other in science fairs, um, and her research was at the Stanford Multi-Modality Imaging Lab right across the street. Uh, and she was also a research intern there through high school and then afterwards as well. And I think it was really interesting because one of the data sets that I came across, we trained, you know, very simple, uh, you know, CNN on was essentially classifying red blood cells that had malaria versus those that didn't have malarial parasites in them. And I think the idea was like, okay, if we can get high-resolution enough images of a small volume of blood, we could probably do this for, you know, any type of blood cell, and as a result, you could generate a CBC from a small volume of blood very, very quickly. And that was really it; like I think it was clear from a first-principle standpoint that, look, pathologists, you know, used to be the primary way of doing a cell count, and if you can automate that with computer vision and microfluidics, you can generate this really important blood test in a point-of-care setting or even in a patient's home.

Okay, so what was your co-founder doing when you tapped her on the shoulder to begin this thing? How did that conversation go? What was her initial reaction?

Yeah, I mean, she was in school as well, and I think she sort of saw it as like a summer research project, and that's how I pitched it as well. And then at some point during the summer, it was like, well, now we can’t go back to school. And so we, um, I think what interested her about it and also what interested me about it was it was a really interesting application of, you know, two research fields that we both worked in and we both loved in a very tangible way. And I think that’s it really started as a research project, and then, you know, once we noticed it was working, we realized it was important to build a business here and actually turn this into a product.

Okay, how did your family or other people around you respond to you wanting to drop out of school?

Yeah, I think it was definitely a battle, convincing, you know, like both of our parents in terms of, hey, we're going to take some time off. And similarly, it was pitched as a quarter off and then a year off, and then we're not going back now. And so, but it I think once it was even for us, like what gave us conviction to leave was we could see that it was starting to work. Like the, you know, our very first bench trial who worked incredible. Our first clinical trial, you know, we were able to do during YC, and also, you know, had some really incredible data, and it was clear that there was something here.

Okay, great! So at what point in those early days did you discover YC, decide to apply to YC? What was your thought process there?

Yeah, so I mean growing up in the Bay Area, YC was sort of ever-present. Um, and I, you know, I went to Startup School in 2013, I think, and it was, um, you know, PG and Sam were interviewing Zuckerberg, and then they had Jack Dorsey come and speak. And so it was—I mean, it was just part of the ether here. Um, and I think it was so exciting that you could see—I remember I saw Flexport’s YC interview on stage. Um, and like, I think Paul Graham made him do it live in front of everyone. Uh, and it was just such an exciting, exciting moment. And so we ended up doing, like, the first version of Italus was actually built at YC Hacks when it was purely focused on can we detect a malaria parasite versus, you know, just a binary classifier. Um, and so that's how I got introduced to YC, and then we applied once, and it was too early, so we got rejected, and then we applied the second time, and we got in and then we, you know, we did YC.

That's great! And so what was the batch like for you? What did you spend your time doing?

I think it's interesting because it was really, really helpful that there were software companies in our batch that had week-over-week growth metrics. And the thing I loved about it was that if you're a competitive person and you see someone else's graph going up, and you're sitting here with this, you know, shitty Raspberry Pi, and the thing isn't working, you're gonna get really mad at yourself. And I think the fact that there was this pacing, that there was someone who paced too, uh, was incredible, and like Scalia was in our batch. I still chased that company, and, you know, I think the beauty of having people that are operating faster than you is you will— you will try to acclimate to their pace. Um, and that to me was one of the most incredible things about YC.

Uh, and so, you know, during our batch, we started—we set really aggressive milestones around, you know, yes, you know, we didn’t—we wanted to make sure that the fact that we were a biotech or digital health company was not an excuse for delivery results, which it can become because you convince yourself that there's you—so many—there's just so many roads, right? Which there are, there are a lot of roadblocks when you're working with hardware and you're working in a lab. And so, before even starting YC, we said we're going to finish a clinical trial, and it is going to have FDA grade results by the end of the by the end of these 10 weeks or 12 weeks. Uh, and I think using that as our North Star just forced a level of speed and execution that wouldn't have been possible otherwise.

Okay, so you're saying that during the three months at YC, you wanted to finish your clinical trial. That means you must have started long before you entered YC, right?

No, not quite! So we, um, we definitely—we had early versions of the device. We had a, you know, a pretty simple partnership with Stanford where we would get their residual samples, and we'd be able to run—run tests. But the trial itself came together during the batch. Um, we basically, you know, we emailed dozens of labs. We emailed dozens of local clinics. Um, you know, and the brand names, like I think Stanford quoted like half a million dollars to run the trial, and like we had 120k from YC at the time; it’s just not going to happen. Um, and also, I think Stanford's timeline was like, this will happen like in Q3 2018, which is like, it’s just unfathomable when you're thinking about, you know, a 10-week timeline. Um, and so I think it was a lot of it was a lot of outreach, um, and we found, you know, a lab in Mexico that was actually run by one of Deepika’s friends and friends' families from college, and we were able to very quickly get something set up. We scoped the trial, you know, patient selection, IRB approved it in like a week, and it was just a lot of pinging. Like, we had to get people to conform to our timelines versus working with their timelines.

So you started and finished a clinical study during the three months?

Yeah, and for what it’s worth, this is, you know, a Class II diagnostic. You know, if you're working with like implantables or you're working, you know, with a drug, it's a very different—I'm sure there's an equivalently ambitious benchmark, but it's probably not going to be the whole clinical trial. But for us, we thought this was the right level of heart, where this was a trial that could actually be done in—you sit with the device; you run a comparison study against the lab; you can do it in—and you do it with stations; you could do it in a couple of days once you have all the necessary access.

So were you also interacting with the FDA or doing anything on the regulatory end?

We—so at this point of time, we had—I think we had sent—we had one or two pieces of communication just trying to honestly just understand and clarify the requirements for our device, and we identified that, look, as a Class II device, you know, there’s a pretty clear, you know, predicate in the form of the Sysmex CBC machine, and there were some clarifying questions, but that was it. Like, if you message the FDA, they tell you that we will schedule a pre-submission meeting in like six months, and that's like, it's just not good enough, right? And the reality is that as a startup, you're a heat-seeking missile for a very specific goal in a very, very fast timeline. Um, and so we scrounge together what we could, and we knew the trial probably wouldn’t be perfect because we hadn't done a pre-submission, but we, you know, leveraging what they had publicly available, we scoped and designed that on our own.

Okay, and what went into the thought process of, you know, submitting a pre-sub or not? Isn't conventional wisdom to submit a pre-sub?

No, yeah, I— I think pre-subs are, like I said, a waste of time when it comes to, um, it’s just the way I really feel like it's a way for the FDA to buy time. Um, and I think the reality is, is that just get your submission in. All the feedback you're going to get will be in that response because they're legally obligated to give you a response within 90 days total. And you will then, during that time, you will be iterating on, you know, your trial design, your patient election, probably even parts of your device. And I think that it's really important to find wherever you can overlap timelines. In healthcare, you have to take advantage of that. So we don't do pre-subs; we submit the thing. We know we're going to get, you know, lengthy additional information requests, and then we run with it. We did the same thing at Envision because, uh, with a pre-sub, it's 90 days before they even schedule a meeting with you, and then it's more time after that, and the feedback they give you was in binding. So 510k thing is faster, so I prefer it just going straight for the kill.

That's the other thing—the pre-sub feedback isn't even binding. So, like, they can just tell you something and then, like, completely reverse it like six months later, and I've had companies go through that where they do the preset, they say, "Oh, we kind of like this trial design," and then later they say, "Just kidding."

So, um, and so if, so you had all the clinical trial data; you submitted the 510k. What was it like kind of after you left YC those first initial years, and what caused the pivot as well?

Yeah, I mean, I think what's interesting is that the years immediately following YC, though it's like a year, year and a half, two-year time period was really challenging, and it was, um, there were like moments of serious despair just because I remember we got our first additional information request back, and it was pretty much like getting an F on a report card. Like, it was just—it was 40 pages or like there were 40 additional information requests; most of them were like redo this, redo this, and we had like a year, year and a half worth of burn left. Um, and so it wasn't—we were not in the most comfortable position. And I think the um, the thing that kept us going was, you know, when I first got that additional information request back from the FDA, it was very much so like, okay, like, they're not going to clear us; that's kind of what it felt like. And then when you broke down the information and put it on a to-do list, I remember like our CTO, Dhruv, and I, we stayed up till three right after we got the response, and we just made a, you know, 30-page Google doc with a to-do list of what are the things the FDA wants from us, and honestly, the morning it felt a lot more attainable. And so then we sprinted for six months, and, you know, in 2018, we were in summer—16 was our YC batch; in 2018, we finally got the clearance. Uh, and so I think the two years in between doing YC and getting the clearance, something that was really important was we kept the pace of YC alive because there were times where it felt like, especially in the six months after, it felt like we were going slower, and I couldn't quite fathom why. And I think part of it was that the batch didn't really exist in, in like—and I still saw, you know, other companies, but it wasn't like this like mad rush to go, um, you know, to go deliver results and make your group partner happy and, you know, beat your other, you know, fellow companies.

And so I think we figured out ways to replicate that. You know, I started doing check-ins with my group partner, um, again, and, um, and that that sort of helped reset the tempo. Uh, and then I think the next thing you asked about is how the sort of product transformed and evolved. So, I mean, today at Othellis, you know, the initial device, it's a segment, right? It's a—it's part of our revenue. We, after clearance, we—in the first six months, we scaled to about one, two million in ARR, and it was, uh, it was great! Like, we had patients that really needed the device; we had clinicians that were relying on the system every week for their patients on clozapine and their patients on immunosuppressive medications, and we quickly realized that like, look, we can scale this business probably to tens of millions of ARR over the next four or five years, but if we want to build a business that is, you know, doing hundreds of millions or billions in run rate, we need to expand what we offer these—these—it's almost like the same way fintechs expand share of wallet. We realized we needed to expand share of patient in a clinic, and what we did is that we introduced, on the foundations of that first device, we introduced hypertensive monitoring programs, medication endurance programs. Uh, we, you know, started expanding the revenue cycle in the practice, and when you go into a healthcare clinic, they need so many tools, and everything is broken. And I think the biggest realization was that once you have the trust of the doctor and the trust of the patient, you can be that technology provider, that technology partner that actually builds so much more than just that first device for them. Uh, and that's—that's really how— I wouldn't even call it a pivot necessarily because we just added product lines and, you know, our ACVs went from ten thousand dollars per clinic to now hundreds of thousands of dollars per clinic, uh, and— and that's—that's kind of the equation we had in our mind.

So, so clearly that move was fruitful. Going from tens of thousands to a hundred thousand, hundreds of thousands of dollars' worth of revenue per clinic. At the time that you wanted to expand, how did your existing investors react? What was that conversation like with them?

Honestly, if it wasn't a conversation, I think the—our approach was, look, we have delivered on the initial—like you, I think you earn the right to both with your customers and your investors and honestly your employees. You earned the right to to build more as you deliver customer happiness on on the first pieces, and for us, there would have been, you know, no remote patient monitoring suite or revenue cycle suite that is now like— that are now the core revenue drivers for the business without the first device. And so I think there was a lot of trust, you know, we got the FDA clearance in in record time. Um, there was trust that we could execute, and I think the like roll off from from Sequoia says is, you know, is that the TAM is—TAM is like a fake number. There's no such thing as a total addressable market; TAM is just a measure of the ambition of the founding team; that’s it. And I really—I really believe that, which is if you, if you execute again and again and again, you unlock more TAM, and if your customers need you, and you continue to deliver products that help them, you will grow as a business.

Okay, absolutely! So you mentioned employees a little bit, and a lot of early-stage entrepreneurs or people who have been thinking about starting companies wonder how they're going to grow their initial team or find the right people to get the company off the ground. Do you have any words of advice for them?

Yeah, definitely! I think—I think the—our initial team, I mean, you know, Drew, who joined our founding team as CTO, um, just this like hacker mindset, right? Like the—the two of us could work on problems for endless amounts of time, Deepa, you know, innately a hacker mindset. And I think there's a lot of push that people get from investors or people get from, you know, even just conventional wisdom that they need to hire the, you know, the biopharma exec or the med device exec that that went and did XYZ—it's a very different skill set. And one example that I use is even in like—even in sales, right? A lot of times you'll hear from folks like you need to go hire the person that's selling like four million dollars' worth of product a year. And I remember we hired someone from like Salesforce, right, that was like selling tons of product every year and we're like, this guy's going to be an awesome, awesome sales leader, etc. The reality is like Salesforce is like kind of like a tax on startups at this point of time, right? Like the money's going to show up whether that guy was selling or not. It's like it's like hiring someone from the IRS. Like it doesn't really matter because like the money will show up. Like people need will always be buying Salesforce. You—our most successful sales reps—and this is a slight deviation from what you initially asked—but our most successful sales reps, for example, like one of them was a door-to-door pesticide salesman, like like a really hard product to sell. Like that guy was grinding in order to sell one. You know, one person. And our belief has always been slopes versus intercepts, which is find people that will work really hard to learn the underlying thing, um, and don’t, you know, haven’t had to coast at any point of time because, like, again, like there's products that just sell themselves, and they're not going to be good salespeople. And I think it's similar on engineering—the same philosophy. Like honestly, people from the Googles of the world just haven't been great fits at a thousand. And I think, you know, we've chatted with other startups—this is often the case—unless you pick from a very specific team that actually has a fast shipping cadence.

Yeah, so that's—so you think it's the fast shipping cadence that makes all the difference?

Yeah, anyone that's had to fight for survival is going to be good, and if they've lasted long enough, and anyone that's had to deliver results on a, you know, on a weekly or monthly time horizon will be good. And anyone that, you know, doesn't have to do that if their job would exist if, if, you know, they were working at 25, it’s not going to be a good fit early on at a startup.

Okay, okay, fantastic! Tenae, can you give us a snapshot of where Othellis is today—revenue, how much money have you raised, dear investors, that sort of thing—and maybe what your future plans are for the company?

Yeah, I think—I think so, you know, the high level we want to end the year at a 50 million run rate. We are, you know, in total, 300 people; uh, you know, 150 in India, 120-ish here, and then like we have 30 that are kind of remote spread around the world. Um, the, uh, you know, today our product is is essential to the clinics and healthcare providers that we work with. And I think for me, it's very clear that there's so much to build for these customers, which is why we're all here today. Uh, whether that's in devices, whether that's, you know, in software, whether that's in other forms of tooling, whether that's bringing care directly to patients and cutting out middlemen wherever you can. Um, I think—I think for Othellis, really our end goal is if you look at our product pipeline and you look at, you know, what we're shipping week over week, it’s very clearly just like what is it that makes the doctor's life 10 times easier, right? Like what is it that they currently do manually that we can use software to automate for them, or on the diagnostic side, what is another test that we can launch that eliminates, uh, you know, a two-hour drive back and forth for a patient to go into clinic and then, you know, come back home and then get results two weeks later? Uh, and so like honestly, our product evolution is somewhat boring, I’d say for the next like two years with the exception of probably like one or two areas where we're starting to apply things like large language models, and it's like very exciting, and you can suddenly do things that we weren't able to do a year ago. Um, so I, I’d say like just more infrastructural back-end, you know, admin tools for healthcare providers, and you know, in-home diagnostics and sensors for patients.

That's fantastic! So, Tenae, in the audience we have a lot of entrepreneurs a couple of years into their journey and some folks who are just on the fence about whether or not to take this leap of faith and start a company. Any advice for budding entrepreneurs?

Yeah, I think I think the—there's the, um, like the Sam Altman saying, right, of like, uh, like the real risk is not taking a risk, and I think it's super true and it's it will be very, very hard. Like if you're if you're gonna take the jump, be ready for like 10 years of pain, um, and—and—but but with pieces of fun and, you know, just like really exciting moments that make it all worth it. And, and so I think the—the other thing that I tell people is like it—the only things that are really worth doing are the ones that you can compound in for a long time. And so pick problem spaces that you're genuinely excited about, and, and then just, you know, I think the score takes care of itself after that. Like if you if you work hard every day, nothing else really matters.

Thank you, Tenae, for sharing your journey with us. Incredible! Thank you very much!

[Applause]