Getting Punched in the Face

People are waiting weeks to months to get to care in the outpatient setting. I'm sick today, my loved one's sick today, and I'm waiting weeks to months to get into the clinic.

I played soccer as an undergrad at Berkeley and I tore my ACL my sophomore year. I called the Palo Alto Medical Foundation to get an MRI scheduled, and they told me to wait three weeks and an appointment would open up — a pretty powerful, out-of-course experience for most patients. And I was that annoying dude who, on a Monday in between classes, because I had some free time and really wanted to get back on the soccer field quickly, called them ten times. Eventually an appointment opened up, I had my MRI, and by Friday I had surgery. So I was on my path to healing in four days rather than waiting three weeks just to get the process started.

The reality for me was, I was in my early 20s — I could have crutched around for three weeks and I'd have been fine. Same health outcome, just three weeks later. But the reality for most Americans, in our experience, is you're waiting two and a half to twelve weeks just to get to care, and every single day that goes by, your likelihood of a good health outcome is deteriorating. That's really the problem.

We wanted to test whether this was unique to us and our own experience, or unique to the imaging world — Tashfeen's a radiologist, and he saw the same thing, patients waiting up to six weeks to see him while every day there were eight openings in his calendar. So, being good nerds on campus, we ran a study with a little over 100 different healthcare providers: doctors at clinics with two or three physicians, all the way up to very large, complex health systems employing thousands of providers, and everyone in between — integrated delivery networks, accountable care organizations, academic hospitals, safety-net hospitals, large multi-specialty private clinic networks, and independent physician groups. What we found is the average American is waiting two and a half to twelve weeks to get to care, and those same clinics — whether two doctors or thousands — are sitting on about 15 to 30% unused capacity on any given day.

So you have people waiting for care, and care available, but no simple, easy way to connect the two. That was the starting point. We said there's got to be a real-time, automated, text-first way — no app, no download, no portal, no login, no friction — to one, identify the patients waiting for care; two, identify the real, dynamic availability in a clinic's schedule; and three, the magic one, make sure you match the right patient to the right care. It was a quote-unquote smart waitlist. That's where the journey began, and now we've grown to focus on that whole set of problems at a much higher level. Patients really struggle to hold onto a clinic and clinics really struggle to hold onto a patient — not just at the onset, but at any point along the patient's journey. So we've become the end-to-end patient communication, or what we call the patient relationship, platform.

Early on, when we were both on campus, part of it was just being at Stanford, where there's a whole culture of innovation and entrepreneurship and so many resources that make it easy to play startup. Tashfeen is a wizard of a human — a practicing physician and a self-taught coder who's built websites and apps for all sorts of things — and he started a prototype. What we quickly did was use what's now Luma Health as the class project across all our different courses. For my computer science classes, I'd build on top of his prototype. For my business school VC fundraising strategy course, I'd use the idea to build up the financial model and justify the investment required.

But the real key thing — and I think we learned this from the design school at Stanford — is that you can think through an idea and analyze it in every shape and form, but there's nothing more valuable than getting it, even if it's a poopy prototype or just a shell, in front of real people. For us that was critical. We needed to get this in front of the clinics, hospitals and health systems who'd ultimately be using it and buying it. That's why we ran the study with 100 different providers across the spectrum.

The great thing is, when you're a student, everyone talks to you happily and freely — there's no agenda, everyone is candid and transparent. So we got invaluable early feedback: "This is a cute idea, but here's what it would take to become real." I remember we were two dudes with an idea, a smile, and a handful of PowerPoint slides, and we got a chance — because we could play the student card — to present to Stanford's health system. Their chief operating officer invited us to present to the leadership team, and that was 45 minutes of just getting punched in the face. But it was the most invaluable 45 minutes.

Exactly. That's where we started to uncover: this is a good idea, and it solves a real set of connected problems, but there's a long list of things we'd have to be sharp on if we wanted a health system — with a sophisticated IT team, a risk-averse technology team, an operational team that's been burned by technology before — to actually implement it. They understood the problem and the solution, but there was a long gap in between that we had to figure out how to cover. And that was the best part: rather than just strategizing and building slides and models, it was best to actually go get punched in the face by real users, because that made us so much stronger when we then decided to launch the company.

That's actually the best time, because the gates are lowered. Everyone enjoys spending time with students — people who are intellectually curious, who have ideas, who are looking for feedback. When you're a student, you're not selling anything. Just by saying "I'm a student working on a startup," people immediately lower the force fields: this isn't a salesperson, this is a sharp, curious kid who wants my advice.

There's the old saying that when you ask for money, you get advice, and when you're trying to make a sale, you get feedback — but when you ask for feedback or advice, that's when you get the investment, or the potential customer, or someone saying "yeah, I'll try that." For us the most invaluable part was leveraging the fact that we were students who had questions and were genuinely open to feedback. There was no bad data point early on — every one helped us identify trends, what was lacking, what was missing. If I were to give myself advice back then, it'd be to start that process even earlier. The minute you have an idea, get it into the hands of real people. It doesn't have to be a real product yet — that feedback is invaluable.

What we found is this wasn't a new problem — it's been around a long, long time — and folks were solving it in two ways. One, they were just double-, triple-, quadruple-booking patients, because they knew demand was always high and supply varied. When you look at a clinic or health system schedule, everything looks booked, and then within a 48-hour window it's really turning over, because people are people and their lives change. So there's a lot of short-notice change, and that 15 to 30% unused capacity. But on the days people do show up, it's a terrible experience — you wait 45 minutes in a waiting room, then sit in another room and wait another 45 minutes, then spend seven and a half minutes with the doctor and off you go.

Our view was always that there will come a time when the US healthcare system finally starts to think of the patient as a customer, and wants to own and automate that entire journey. That means you have to deliver a good experience alongside a good outcome — because if a patient waits two hours to spend seven minutes with you, they're going to go online and leave a one-star review, and that's a real risk to your business.

The other way people tried to solve it was to throw more bodies at it — hire more people to call patients. But in today's world, no one picks up their phone. The average health system was making 20 calls to reach a single patient, otherwise just leaving voicemails. That's where an automated, text-first, SMS-first approach really made sense, because the average adult will see a text in 90 seconds and respond within 10 minutes. And people aren't annoyed by text — you can text a patient three or four times a day and they're not irritated, whereas if you email them more than twice they're hunting for the unsubscribe button. That was the early power of what we started with, and it let us uncover a broader set of pain points and watch clinics and health systems change their viewpoint: we have to think of the patient not just as a customer, but as a person. They'll reply with paragraphs, they'll use emojis — they expect from their health system the same experience they get everywhere else in their life. That's the transition we've seen over the past six years. We've been pushing this uphill, and the market's finally meeting us and pulling for it — but it took six years.

Portals were supposed to be the magic thing that made everything better, and they've been total poop. No one engages with a portal, no one logs in, you forget your password — the average person engages with healthcare twice a year. So portals failed miserably. Then a couple of years later it was all about apps: every health system wanted to hire a product manager and a handful of developers to build one, or buy one from Epic. When we launched in 2015, we purposefully took an app-less approach, because there's a lot of app fatigue, and healthcare apps are especially hard. It's not like your banking app or your ride-sharing app that you use every week — you engage with a healthcare app a few times during an acute period, or two or three times a year.

Even with MyChart — Epic's big app — well-resourced systems like Stanford or UPMC could spend beaucoup marketing bucks and get a 75–80% download and activation rate, but if they tracked routine engagement it was all sub-20%. In fact, the best way they had to reach a patient wasn't a portal or an app — it was the phone. Their contact center was still the best thing they had, even in a very modern world. That was the moment for us: this makes no sense, we need a text-first approach.

The other part is a desire to treat people like people, and like customers who now have more choice. You no longer have to go to healthcare to get healthcare — you can get it at Walmart, at a CVS pop-up clinic, or from a health coach on your phone. Health systems have finally realized they have to treat these people like customers and like people, otherwise they'll never get the chance to provide care and capture the lifetime value of that patient.

One of the greatest pieces of advice we got early on was two things. First, if you have to explain the problem, you've already lost — whoever you're working with has to feel the problem deep in their core. For us that was a check: everyone knew this was a real problem. They were trying to solve it in all sorts of ways — double-booking, more people calling — and they knew it was lost revenue and missed opportunity to provide care.

Second, it's insufficient to just talk about improvements in health outcomes or cost efficiency. You have to lead with value, and ideally that value is top-line revenue growth, with cost savings and better outcomes underneath. The rubber hits the road when you can deliver tangible financial ROI. By getting patients in sooner and keeping the schedule full, we could show a 5-to-10x return on investment month over month, immediately. People could count exactly how many patient butts Luma helped them get back into seats, and there's real revenue attached to that. On top of that, patients who got in sooner had a better likelihood of a good outcome, and if staff no longer had to make 20 calls to reach one patient, that's a cost to reduce or reallocate to higher-value work.

That was the triangle for us: top-line revenue growth, improved patient outcomes, and some cost benefits. But the latter two were just nice sub-bullets. If we didn't lead with "you'll see a 5-to-30x return on investment month over month," I don't think it would have mattered. There are tons of promising digital health startups that can only make one or two of those claims — outcomes or efficiency — but if you can show revenue, that's what's differentiating and lets you last as a company.

Not really — when you're a student, people want to give you good advice; no one's trying to steer you wrong. One piece that was really helpful was that speed to market and speed to impact really matter. The scarcest resource for an early-stage startup isn't capital, it's time. Our very first pilot was with the University of Maryland's academic medical center, and that was the worst place to start — a very complex, bureaucratic organization, through no fault of their own. As a team of two or three, we'd have spent 18 months just getting to a formally approved, deployed pilot, and by then the product would have been customized so much to that one system that we couldn't replicate it anywhere else. The advice we got was: either raise a buttload of capital and suck it up for 18 months to start with health systems, or start with clinics and physician groups where you can get the product into real patients' hands within weeks. That was really good advice.

As for bad advice — I wouldn't call it advice, but folks often highlight the positive outliers. Everyone talks about the Steve Jobses and Bill Gateses and Musks, fine — but even setting those iconic names aside, you hear these stories: "I was on the Caltrain from Palo Alto to San Francisco, sat next to this VC at a leading venture firm, handed him one of the two coffees I was holding, and by the end of the 45-minute ride I had my investor, and the next day I got my term sheet." You hear a lot of "it all just kind of happened." For a first-time founder, that's damaging — they're not advice, they're just anecdotes. The reality is it's a slog, and you'll get punched in the face day in, day out. There are a lot of valleys and few peaks. It takes real stick-to-itiveness to get through the early stages, and I think people should talk more honestly about that rather than about the bright, shiny outliers.

Endless examples — any honest founder would tell you the same. It's always easy to say "we're crushing it, we've got so much demand, I've got VCs I'm fighting off." There may be truth to it, but a lot of it is founder bravado. Most founders have sleepless nights agonizing: "Oh my god, that one shaky part of the product — if it doesn't work, this whole thing unravels." You honestly have to fake it till you make it, and that's okay, because that's how you get better.

A couple of examples. One, from a product and patient-impact standpoint: early on we pushed ourselves to work with larger and larger clinics. We'd been working with maybe ten clinics across psychiatry, dermatology, asthma and allergy, physical therapy, and then we went to our first multi-specialty federally qualified health center — a sizable organization treating underserved populations in the greater New Orleans area, about 20 locations doing a few thousand appointments a week, whereas we used to see a few thousand a month. We were eager, we wanted to show we could handle larger customers, but we weren't ready from a product standpoint. We turned the switch on and the product fell over — we messaged one poor patient literally something like a hundred times in a fraction of a second. It was just a bad loop we'd gotten into; not a rocket-science fix, just something we'd overlooked.

That was very telling. Any product, but especially one that directly impacts patients, you have to measure twice, thrice, before you cut. Your classic Silicon Valley product person says "you've got to break it, break it, break it" — but when it impacts real people's lives, be very thoughtful about how you deliver change. I remember going out to visit that patient with a handwritten apology card and a gift card. I took it to the clinic, and they said "well, the patient's here — you're the guy, you go give it to them in person." And I did, hat in hand. That in itself was invaluable, because you see who's actually using your software, you feel the frustration it costs someone, and it drives you to build a better product.

The other example: I talked about being in the business of benefiting patients, not the business of benefiting from patients, which unfortunately a lot of healthcare has devolved to. That's meant we've sometimes made quote-unquote bad business decisions that are the right patient decisions — and that's hard when you're a super-small company and every dollar of revenue counts. We had dermatology clinics that wanted to use our now-sophisticated messaging platform — deeply integrated with the EHR, AI and NLP built in, with stellar response rates, 65–70% plus, which you can never match with a voice call or a portal. These clinics, run by good business people, said "let's use this amazing platform to send out Botox Tuesday messages" — purely spammy marketing stuff, and they were willing to pay top dollar. We had to step back and say, that's not the company we set out to build. Our north star is connecting more people to the right care quickly and keeping them coming back. A Botox Tuesday message is, at best, tangentially related to care. So we said no.

That was hard — I had a well-intentioned director of sales saying "that's a huge opportunity we're saying no to." But our perspective was that we had to maintain the purity of that channel with the patient; if we polluted it with spam, we'd lose the response rate that's so important to guiding patients along their care journey. We left a lot of revenue on the table, and when you're at 10, 20, 30K of recurring revenue trying to get to a million, that's a really hard decision. Part of it was Tashfeen — our chief medical officer and a practicing physician — whose job is to keep us focused on the north star and not get pulled into areas that take us away from our core mission.

I still think it's execution. What's the famous saying — luck is what happens when preparation meets opportunity. Yes, we had a good idea and a real pain point, but lots of people have good ideas that address real, broad-market pain. It was really the ability to execute, and execute quickly. As I said, the scarcest resource a startup has isn't cash, it's time — speed to impact is critical, because every hour, every day counts in those formative periods that determine what type of company you grow up to be.

Now that we're a later-stage startup, soon to be Series C, with a couple hundred people, it's still execution, execution, execution — speed to impact, speed to impact. A big part of that is maintaining a culture that encourages and celebrates smart risk-taking. Early on, you expect everything you try to blow up in your face, and sometimes spectacularly — and you want that to stay true when you're much larger. We celebrate when something blows up, because we learn something and we'll be better next time. Do more of what works, less of what doesn't. It all leads back to execution.

20-year-old Adnan was lacking in a lot of ways — probably most of all in professional maturity. I'd been on teams and led teams, but in full candor I hadn't really managed, hired, or built teams. At the end of the day, every organization is about people. And people remember how you make them feel — they don't remember what you did, they remember how they felt. So it becomes really important not just to "do things right" but to do the right thing. 20-year-old Adnan wouldn't have known that.

It came through a set of experiences in my 20s. Grad school in Cambridge was a very different approach to communication and learning than I grew up with in the United States, and 90% of my classmates were from different countries — so how do people communicate, problem-solve, where are the nuances? Those were invaluable. Then growing up as a healthcare and management professional at Genentech was invaluable training, because when you have an idea, you have to convince really good, all-star, A-plus people to join your journey — people who have alternatives, who could be making more money somewhere safer. Many of our first ten employees are still at Luma; we call them force multipliers — they come in to do a job and then 10x it.

It was really understanding how you make people feel, and how you do the right thing for their career growth. If someone's with you for six, twelve, eighteen months and then takes the next step in their career, that's okay — we should encourage that and have honest conversations about it, rather than "oh, we can't lose our good people." You want people to grow and do the right thing for themselves, and that's critical to continuing to attract the best talent. That's the one thing that sticks out most: being able to build, hire and grow an all-star team, because that's the only way you execute and bring a vision to life. 20-year-old Adnan did not have that. He had a lot of energy, sure — he could pull all-nighters at zero cost. 31- or 32-year-old Adnan could still pull all-nighters, but they had a cost, because I was married, and we now have two daughters.

Thank you — that's very kind, I don't even know if anyone watches those. "Respond, not react" really started from my personal life. They say marriage is the best of mirrors, because it lets you see yourself through the eyes of others, and you learn a lot about your deficiencies. That's where the talk originated — rising up and responding rather than reacting, because we can all react to the things we get bombarded with at any given point.

In the frenetic pace of a startup, with a lot of ambiguity and chaos, that respond-versus-react framing is incredibly important — and it's most true for a CEO, where the buck stops with you. You make a lot of decisions, in conjunction with your co-founders, investors and advisors, but ultimately there's a point where a decision has to be made, a crucial conversation has to be had, or something goes terribly wrong and you have to respond. When time and resources are tight and the pressure's on, the natural thing is to react — "these are the data points, this is what we need to do, let's go." My role, in particular, is to step back and say: let's take the extra 30, 90 seconds — because we don't have 24 hours, we might literally have 30 to 120 seconds — and respond in a more thoughtful way rather than just reacting.

Great question. There's a difference between being an American-born Pakistani versus a British one — I saw this when I was in grad school in Cambridge, where I experienced discrimination in a way I hadn't growing up. Part of that is I grew up in a very diverse community in Silicon Valley with a lot of immigrants. But the biggest difference is you can see US companies — Microsoft with Satya Nadella, Citibank at one point, PepsiCo, GM — that have non-white leaders as CEOs and CFOs. The head of Alphabet is Indian-American; the head of Microsoft was actually born in India. Or someone like Jerry Yang, who co-founded and led Yahoo for a long time. You have more examples of Asian Americans, Latino Americans, African Americans, people of different origins already in those positions. That's helpful when you're growing up, because you think, okay, it's not impossible — those people exist. That's a big difference between the American academic system I grew up in and the British one.

That being said — and this may not be a popular opinion, but I think it's the honest truth — it is much harder to raise capital if you're not a white guy. That's just the honest truth. Look, I've had a life of privilege: I went to Berkeley for undergrad, Cambridge for grad school, Stanford Graduate School of Business. I played a collegiate sport. I check all the boxes of that classic profile a VC looks for — other than my name and how I look. And my name is a very easily recognizable Pakistani Muslim name. So I've had crappy experiences with the venture community.

I think the venture community has finally realized it has to change its product if it wants to keep finding the best founders and companies — otherwise it misses out, and what they hate more than investing in bad companies is missing out on great ones. The truth is I could have raised twice the money in half the time if I looked different and had a different name. That's also true based on gender — men raise more money faster than women founders — and then religion, ethnicity, demographics all come into play. Those who want to hear it, I'll share my perspective; those who don't, that's fine. But the VC industry as a whole is being forced to transform, and the ones who win will be the ones who adapt in the right ways.

There's real truth to that, and it's okay for people to acknowledge and talk about it. It's often seen as "oh, you're making an excuse" or "this is a meritocracy, what are you talking about?" One of the biggest things I learned through the fundraising process is that it is not an objective, merit-based process — it's incredibly subjective, with many things you can't control. If a partner walked into a partner meeting that morning having heard something terrible about digital health, or how hospitals are the worst people to sell to, you could have the perfect presentation with all the right metrics and fundamentals, but what they remember is "selling to hospitals is hard," and back to torpedo. Or someone had a fight with their spouse and comes in Monday and takes it out on you. That's true in any professional interaction, but especially in fundraising, where so much is pattern matching. It's not the sophisticated science VCs would have you believe.

As a founder, I'll candidly say venture is often a necessary evil to bring your vision to life. There are great and not-so-great investors out there, but even the great ones — there's not a whole lot of science to it. It's a lot of pattern matching, looking for similar traits to other companies and founders they've seen succeed. That's limiting, especially if the entire partnership looks the same and they're all men — you've put yourself at a disadvantage, because your unconscious bias and your pattern matching are limited to a sample set that isn't truly representative of what success looks like.

When I speak to folks at Luma, or people who want to join, in more of a mentor-mentee relationship, I share some advice I got as a grad student in Cambridge, from the CEO of a biotech there. He said: what you do in your 20s doesn't matter — what company it is, the logo, your role, your title, none of it. What matters is, one, are you learning something every single day, are you in full growth mode? The minute your learning starts to taper, find something different. And two, are you going to work for a manager who'll pound the table for you and get you opportunities you couldn't get for yourself — and who'll stretch and challenge you to do things you didn't think were possible? It doesn't matter if the logo is Facebook or Google or Amazon.

I see that now from the other side of the table. When we interview people, I don't really care which company they worked at or where they went to school. What I want to know is: tell me what you actually did. Tell me the real story of when you were in a place of challenge and had to solve a problem, and had to do it with people who were different from you. People who can articulate that clearly, go deep with lots of examples, and then still summarize it succinctly — that's what you're looking for. Especially for a startup: how did you do it when resources were constrained, when there was ambiguity, with a cross-functional team in a short amount of time?

The other habit: my first job out of college was management consulting, and that gave me the ability to communicate and articulate a lot of complex data in a simple, digestible way — whether in an email, a ten-minute pitch, or 15 minutes with a potential customer. Lead with the "so what," the thing that's most important, and provide enough color that the person understands it — but don't lead with a lot of storytelling and data and justification before getting to the punchline. Lead with the punchline, be crisp, then fill in the details. When you're a data nerd like we are, the inclination is to start deep in the data, but most people don't have the time or the aperture for that. Those core consulting skills — sift through the data, then quickly articulate here's what matters, here are the gaps, here are the risks, this is my recommendation, and if you want to go deep, I'll take you deep — have been invaluable.

There are tons of books about founders and starting businesses, but I'll start with one that isn't that at all — I just enjoyed reading it. It's called Walking the Nile by Levison Wood. It's about a dude with the dream to walk the entire length of the Nile, and I found it had a lot of parallels to the startup journey, even though I don't think the author intended it — you start in the swamp, in the muck, trying to get to the end of this long, winding journey with all sorts of trials and tribulations along the way. For me it was about how it didn't go according to plan — it was by no means perfect, but they still got to where they wanted to get to. There are a lot of parallels for a startup leader.

The other, which I think is a must-read, is Zero to One by Peter Thiel. It's not a gripping novel, but it's very authentic and walks through all the things you have to do in terms of preparation as a startup founder. I wish I'd read it much earlier — we definitely walked over landmines and made boneheaded errors, and even skimming it before we embarked would have been really helpful. There's very tactical, tangible advice in there.

No, thank you so much. I hope this wasn't boring, and I hope your audience finds it interesting and helpful. And look — I talked about playing the student card and benefiting from it myself. If I know you're a student, and I know many of your listeners are, if I can be helpful to anyone in your network, let me know. You have my contact information, and people can drop me a line — whether there's something interesting they want to challenge me on, or just to get feedback or advice on something. I'm happy to help.