Introduction

Patrick: [00:02:03] My guest today is Matteo Franceschetti, the founder and CEO of Eight Sleep. Eight Sleep builds mattresses with dynamic temperature control and a variety of biometric sensors, and their goal is to make their customers get drastically better sleep. We talk about why biometrics matter, how hard it is to start a hardware company and launch manufacturing overseas, how Matteo manages his own sleep, and the massive potential preventative health companies like Eight Sleep may have in the future. I hope you enjoy this conversation with Matteo Franceschetti.

How Eight Sleep Works

Okay so Matteo, I think this is the first time that I can I am a nightly active user of the product of one of my guests. Your product is Eight Sleep, a mattress that does all sorts of interesting things. I'd love you to begin by describing the origin story of this business. What was the path that led you to found Eight Sleep?

Matteo: [00:02:53] Yeah, I was the typical entrepreneur working long hours, and I started looking into sleep to see if I could sleep less, to have more time to work and do whatever I wanted. And then, I started wondering Elon Musk is taking to me Mars, but I still spend a third of my life on a piece of dumb foam. I mean, if I'm going to live 100 years, I'm going to spend 33 years of my life on a piece of dumb foam, that was not making sense to me. So I thought, maybe I can fix it.

Patrick: [00:03:21] What was the original journey of creating this product like? Because I imagine it's very complicated. There's a cool picture that we'll post in the show notes that shows wires everywhere. What was version one? How did you decide what should go into version one?

Matteo: [00:03:35] It's a cool story, because my co-founder Max, we built the first version in his garage in San Francisco. He just hacked a bunch of different sensors, and then he was also able to create a sort of heating pad. He did everything in three days, but obviously it was a real hack. And then, we invited friends to his place for a pajama party, and everyone came with their pajamas and they tried the product, and they gave us feedback. But at the time, there was no company. It was a side project, it was the third side project we were playing with. One of the people coming give us a check, at the end of the pajama party. And said, "Oh, I want to give you 25K to start the business." That is how Eight Sleep started.

Patrick: [00:04:18] Describe the state of it today, for listeners. What are all of the things that the mattress does today?

Matteo: [00:04:24] Let's start with the vision, and then I'll explain what we're doing today. We want to do two things. We want to compress your sleep, and we want to save your life. Compress your sleep means what if you could sleep only six hours and get more rest than if when you were sleeping eight hours, so technology can enhance your sleep performance. Then, during those six hours, we still want to scan your body to identify if there is any illness, or anything that you should be aware of, so going to bed is more valuable than going to your physician.

What we do today is we already do, partially, both these two things. On one side, we improve your sleep performance, and we do it through temperature. We dynamically change the temperature of your bed and your body, during the night, to help you fall asleep faster and get more rested sleep. Second, we track everything about your heart rate, including HRV, respiration, and sleep. It comes in two form factors, as a whole bed or as a mattress cover that can be installed onto any bed and retrofitted.

Patrick: [00:05:19] Let's talk through each of the reasons why those things are good. Why is temperature an interesting variable to play with, as people sleep? What's the research behind it?

Matteo: [00:05:35] Temperature is the big elephant in the room when you think of sleep improvement, outside sleep medical disorders. There is plenty of medical evidence about that, it's the results of book called Why We Sleep written by Professor Matthew Walker, that has a whole section about temp regulation. But, the bottom line is your body temperature changes during the day, and changes during the night. What we do is we just enhance these changes, to help you fall asleep faster, get more deep sleep, get more REM, less wake ups, and less tosses and turns.

Then, inside the device, embedded are a lot of sensors. We use a technology called ballistocardiography. So you just go to bed, but the bed has all these sensors and is working like a wifi connected stethoscope. Through the sensors, we can pick any information about your heart rate. In the future, we will see arrhythmia, we will see arterial fibrillation, we will be able to predict if you're getting sick because your heart rate at rest changes a couple of beats before you get the flu. And, sleep apnea.

Patrick: [00:06:28] Talk me through the challenge of ... The temperature seems straightforward, you're pumping water that's warmed or cooled through these coils. I think there's been mattress pads that you can cool, through time. That seems maybe less technically challenging. The sensor part seems really hard. What literally are those sensors? And, how accurate are they, relative to, let's say you went to some institute and did a sleep study with the most sophisticated, accurate sensors? How much gap is there between what you can do now, and how quickly will that gap close?

Matteo: [00:07:01] Briefly, about temperature. Certainly, these technologies existed. The big challenge was to make it, to create a great user experience. Make sure that it is not noisy, that it changes temperature quickly, and that it's comfortable. And then, the big challenge is really the machine learning and AI, because we need to keep adjusting the temperature based on your biometrics and your sleep, to really enhance it.

On the sensor, we use these microfilms, so there these sensors in the top layer of the mattress or the cover. You don't see them, you don't feel them. As I was saying, they work like a stethoscope. They substantially pick up the vibration that comes from your chest, and that vibration to be the vibration generated by the heart rate, or the vibration generated by your breathing. Then, together with movement because we are able to track movement as well, we are able to infer your sleep stages. For example, in REM you don't move, but your heart rate is accelerated because you're dreaming. While in deep sleep, you have a different breathing, a different heart rate, and you still move.

How to Measure Effective Sleep

Patrick: [00:08:03] I'd love to talk through, I like the idea that going to bed is more valuable than going to your physician, what the most valuable inputs are for health. You've talked about several of them, heart rate, respiratory rate, things like this. Maybe HRV is a good example of one that people may not be familiar with. Talk to us a little bit about that. Why is HRV valuable, what does it measure? With that feedback loop in place, if I get to see, which I do, using the mattress every morning, my HRV, what can I do with that information? Why is it valuable?

Matteo: [00:08:33] HRV is a great proxy of how rested your body is. Let's say, today you go and you run a marathon, tomorrow your HRV very likely will be really low. The higher the HRV, the better. It means that your body's more rested. Particularly if you're into fitness, CrossFit, or any sort of training, you start measuring how rested you are, based on this metric. When you see that your HRV's really low, that is the moment where you should take the day off, or you should have a very light session. The risk that you have if you train during those days is that you could just get an injury, or your performance would just be below par. Anyone with, again, going to the gym, or in health and wellness and fitness, that is a great proxy.

Another great proxy is your heart rate at rest, which is indicative of what your cardio fitness. So the lower the better, in this case it's the opposite of the HRV. In my case, when I do high intensity interval training, the more I do that, the more my heart rate at rest drops. I have stats now with all the devices I use, over the past probably three years, and I noticed that all the times that, for a couple of months, I do high intensity interval training consistently for a month, my heart rate at rest drops. The other big thing, we don't track that, but it's the VO2 max. That is indicative of the capacity to really intake oxygen.

Patrick: [00:09:59] How much research is there on these variables? It seems like there's good research on the fitness component of this. It's helpful, in terms of its utility to know when to train hard, when to rest, whatever. How much value is there, in terms of more significant health problems or outcomes? Do you think that that's a realm that you'll be into soon? Am I going to get a notification that says, "You should probably check into X, Y, or Z." What are the valuable non-fitness health outcomes?

Matteo: [00:10:27] On heart rate at rest, there is plenty. In particular, there are also benchmarks of where you should be based on your age and gender, so you can immediately understand if you are above, below par, or on par. The heart rate at rest is really important because it's a metric that changes when there is something wrong. For example, two years ago, December 2019, I got the flu. When I started looking at my heart rate at rest, I noticed that it started changing three days before I got the flu. In the future, companies like us, they will be able to have an 80% confidence that maybe you're going to be sick in a few days. Or, if there is any inflammation, or maybe even COVID.

On the respiration, the respiration is more what is really important is to see if you have sleep apnea, because when you have sleep apnea substantially, there is a lack of oxygen going to the brain. If you have a light sleep apnea, it's fine, you want to measure it. But, there is no simple way to measure it unless you use our product. You don't want to wear some of the CPAP machines. But, if your sleep apnea is really strong, you should use one of those, otherwise you will not be rested in the morning because your brain keeps lacking oxygen during the night. HRV, there is not a lot of information yet, in terms of longevity and health. It's a metric that keeps varying every single day, there will be more discovery.

Patrick: [00:11:45] How do you think about the most valuable measureables or data, that it will be really hard for you to capture in a mattress? You showed me you're wearing the Levels patch, which measures continuous glucose, blood sugar in your system, which I think is probably super valuable. Maybe there's other things that can be measure, like vitamin levels, or your urine, I've seen a product that does that. Help us flesh out the rest of what's missing. In 10 years time, what are the other things that we're going to measure that might be valuable? And, which of them do you think you can do, eventually?

Matteo: [00:12:16] To me, the most important thing we will do, and I really care about, is to really scan your organs. I want to be able to detect cancer. There are already technologies that do it, so we're not talking about inventing the wheel, just no one was able to bring some of the these technologies to mass consumer at a reasonable price. Imagine if, with some of this technology, in three to five years from now we will be able to scan your organs and see sorts of cancer in a couple of different areas of your body. We will go beyond that, because maybe we'll just be able to see how your liver behaves after you have a couple of glasses of wine. Maybe it's not an illness, but right now no one knows does your liver become bigger, how much bigger.

And then, there will be the last step of predictive health, where we will not even wait for you to develop a cancer, we will be able to know a couple of years in advance, based on certain trends and patterns, that you have a high likelihood to develop a certain type of cancer. That is what's going to happen in 10 years.

We will change the form factor of the beds, we will introduce many more sensors. The beauty of what we do is...there are three. Because of the price point, we can add sensors. Adding a 100 buck sensor, it's something we can do, it's not going to meaningfully change the MSRP. But instead, in a wearable you cannot add a 100 bucks in cost. Second, we have a lot of space because there is all this dumb foam that no one is really using. And third, you use the product every single day, for six to 10 years. That type of retention becomes really valuable in terms of longevity, because your heart rate today is different from your heart rate in six years from now.

Early Days and the Business

Patrick: [00:13:54] I love the idea of the space, the frequency, the MSRP as variables that make the bed the uniquely right form factor to measure this stuff. Let's talk through the business, now, and we'll probably keep coming back to why this information's valuable, so many neat things to dive into. What were the interesting early challenges of getting this off the ground? It strikes me that there's a heavy R&D function in your business. It's physical, so it's manufacturing. The marketing aspect is really interesting. Talk me through the early days of the business, and the biggest challenges that you faced.

Matteo: [00:14:25] Let me start with a funny story. We went to Y Combinator, we did YC Demo Day, we raised money so it was a good moment, and it was time to start manufacturing. But, things were not happening in China. I go to my wife, who is one of the co-founders, and I say, "Manufacturing is not happening, I have to go to China and fix it." She says, "Oh, when do you go?" I say, "Oh, I leave tomorrow." And then she says, "And, when do you come back?" "Oh, once I have fixed it." So I go to Shenzhen for a couple of months, one way only. Yeah, I started meeting manufacturing, I started meeting contractors, and we figured it out. But, the first batch ... Everything is started with an Indiegogo campaign, it was a crowdfunding campaign. At the time, we sold 8,000 units. None of us ever built a physical unit of anything in our life, so we didn't know much. We had to build 8,000 units. That was really painful, we were late in shipping. And, everything costs more than you expect, in particular when you don't have experience you're not able to predict the mistakes that you will make.

Patrick: [00:15:27] Talk me through the early manufacturing lessons. So, for those that want to go manufacture something overseas. You land in Shenzhen. Who are the players? Who are you talking to? What are you literally doing? What are the steps that matter? What have you gotten better at, since that original trip?

Matteo: [00:15:43] The real bottom line, if I was redoing this again, is I would just hire people like our SVP of hardware. It's not something that you learn overnight.

If I was giving you any advice, it's hire the best person who knows this shit.

Patrick: [00:15:57] Yeah.

Matteo: [00:15:57] Find the person, and they will figure it out.

Otherwise, the learning curve is too slow because any mistake can cost you months.

But, if you're going to do that, then what I did there. Through contacts and friends, I got connected with a couple of contractors, identified one. This person, who was Chinese, was taking us around to see all the different manufacturers. But, what you really need to do is to identify who is going to be manufacturer, negotiate with them. Based on the product, you might need more than one. For example, for us there is one manufacturer only for sensors, another manufacturing for the thermal engines. And then, there is one that assembles everything. You need to run a process, and there is where someone with experience would be way better than you, in order to get the quotes. And then, you need just to be really good at timelines. Again, any mistake can cost you a quarter.

Patrick: [00:16:49] How did you think through the different potential business models here? Obviously, it's a big physical product, you have to charge people for the product itself. And, there's lots of ways that you can earn margin in any business, I guess. I think, your model right now is you just buy the bed, and the application that I'm using every day is currently free. What have you considered, in terms of potential business models to make this thing, one, the most valuable to customers of course? And then, two, the most valuable business?

Matteo: [00:17:15] There are a lot of features that are coming, so you can think of us like a Peloton of sleep. That is how some of our customers, they call us. Right now, the subscription is not paid yet, so we don't charge you for that. That is going to happen. We just want to make sure that we deliver enough value to our customers through health notifications and content, and other data and insights, and automatic temperature adjustment.

We call it sleep fitness, and the reason why we call it sleep fitness is, to me, going to sleep is like going to the gym. It's something that you're doing for your body, health, and longevity. In the same way that you pay for your gym, there must be a small amount that you pay for your sleep, in order to have deep sleep.

Patrick: [00:17:56] When you talk about sleep fitness, you mentioned at the beginning one goal being potentially reducing the amount of sleep. Why would the goal be less sleep, versus just more good sleep? Just want more waking hours in life? Talk me through that as a North Star, and why I care about that.

Matteo: [00:18:11] Our hypothesis is that what really matters, more than anything, is deep sleep and REM. Those ones, we want to increase them. The point is, roughly, you still spend 50% of your time asleep in light sleep. Our hypothesis is that light sleep is substantially an inefficiency of your body to transition to these different stages. Again, we don't want to touch deep and REM, we actually want to increase them. But, we want to compress light sleep, that counts for around four hours a night. What if we could cut it in half?

Patrick: [00:18:40] I'm going to pull mine out, because I'm just curious to ask the question. I'll give a real data point. Hopefully, I'm not a terrible sleeper here. If you look at my data from last night, 60% of my sleep was light sleep. And then, basically split between REM and deep sleep for the rest. What is good? What is a target? What percent should be light, in the best sleepers out there, that you've studied or measured?

Matteo: [00:19:01] Yeah. I'll be super personal, but rule of thumb, I think you could see your REM and deep anywhere between 15 and 25. For me personally, for example, if I got less than 18% deep, I will not feel good during the day, I will not perform well. Anything above 18%, I start being good, anything above 20%, I'm a killer of a day. That is when you start really thinking that, together, deep and REM, they are in the 40%-ish. It could go really between 30 and 50. And, all the rest is light sleep, but light sleep is not what your body really needs to recover at its peak. Because if you get the full night only of light sleep, you will feel terrible in the morning even if it was eight hours of light sleep. What you need is deep and REM, more than anything.

Behavioral Changes

Patrick: [00:19:44] Yeah, really interesting. What behavior changes have you gone through, or maybe you and your wife gone through, as a result of sleeping on one of these and having this data every day? What are the biggest changes that you've made?

Matteo: [00:19:57] Obviously, I'm super obsessed with my sleep, so I can tell you all the hacks. The most impactful to me is what I call a thermal shock. Before dinner, I usually take a shower and I keep switching between hot and cold, 10 times, 30 seconds each. The thermal shock really relaxes me. If I'm in a hotel or if I have the opportunity to do a sauna and an ice bath, that is the same as actually bathtub. The other big thing, if you talk to any sleep doctor, is consistency. Go to bed at the same time, and more than anything, wake up at the same time every single day, also during the weekend. That way, you train your body to wake up naturally at that time, and you will wake up without an alarm or without feeling groggy.

Temperature, as we were saying, is the big elephant in the room. Key thing is when you hear people saying, "Oh, you should sleep at 68 degrees, the whole night," that is BS.

The reason is your body temperature changes during the night. So 68 degrees could be right for 30 minutes, an hour, but not for the whole night. That is where we come in, and with our technology we keep changing the temperature of your body, based on what are your needs. You can try supplements, you can try Normatec boots to compress your legs, mobility exercises.

Patrick: [00:21:08] Have you learned anything about what you consume and how it effects sleep? When and what type of food, alcohol or not, these sorts of things? What have been interesting lessons that you've learned there?

Matteo: [00:21:19] Yeah. I stopped drinking two years ago. Alcohol has a huge impact on your sleep, and your recovery in general. Sometimes, you don't notice it though. You think the opposite because this just slows your mind and it might relax you. And you think, "Oh actually, a couple glasses of wine, they will help me sleep better." The reality is, if you look at your biometrics, you will see that your heart rate at rest will go up, which is bad. Your HRV will go down, and in general you will get less restful sleep. I stopped drinking completely. I was not a heavy drinker, even before, but now I'm good. Then, also coffee. Coffee, what we have seen is the rule of thumb is to stop drinking coffee eight hours before going to bed.

Patrick: [00:22:01] Anything on food specifically that's interesting, for you personally, that you've learned? Types of food, or proximity of eating to bed, or just anything like that. Anything on the consumption side of food?

Matteo: [00:22:11] Food and sleep, specifically, there are some studies that prove that carbs, they can help you fall asleep faster. But then, you will have less restful sleep. Usually, you should stop eating a couple of hours, at least, before going to bed if possible. And obviously, a lighter meal is better than a heavy one. The reason is that if you involve digestion too much, that is where all the blood goes in your body and that makes everything harder. Then outside sleep, I have a lot of stories about my nutrition, because I monitor my glucose. For example, for me eating berries is worse than eating a pizza, in terms of glucose spikes. Well instead, maybe because I'm Italian, an ice cream, a gelato has no impact at all on my glucose. Zero.

Patrick: [00:22:57] That's great news.

Matteo: [00:22:59] Yeah.

Patrick: [00:23:00] Why is glucose spikes bad? Is there good science behind what is going on in the body? What's downstream of that? If that's the indicator, you want to stay in a low variance band of glucose, why is that? Why is that true?

Matteo: [00:23:12] The simplest answer is glucose spikes, they create inflammations. Inflammation is the problem that then drives any sort of illness, including cancer, and has an impact on your longevity. Then, if you keep being at high levels of glucose, you develop this insulin resistance, so you will start creating more and more junk food, or food that can generate these high spikes, which is then what can lead to diabetes.

The Business: Models, Manufacturing, and Marketing

Patrick: [00:23:39] Yeah, that's fascinating. Going back to the business, you've got the early days, you had the 8000 orders on Indiegogo. You struggle up the hardware learning curve in Shenzhen and elsewhere, and that's the early stage of the business. How would you describe the next chapter of the business, after that? What were the challenges in chapter two of the business?

Matteo: [00:23:58] We finally ship the 8,000 units, I think it took us 18 months overall, was 2016. At that point, we finally are in real live orders. The number one challenge you face is then, it's not that you sell or you need to build 8,000 units anymore. The big problem for hardware companies is do these successful crowdfunding campaigns, you build these large volumes. But then, when you start selling on a day by day basis, then each month you sell, I don't know, 300 units. That is going to screw up your whole supply chain, and the manufacturer isn't happy because he says, "What's happening?" All that kind of thing.

But, things started going well. At the end of 2017, Khosla Ventures came in, they invest in us. Keith Rabois is the one leading the round. They give us the money to build the pod, which is the current technology. Because the first version didn't have cooling, but we knew since day one that cooling, again, was the big game changer, it's what can really enhance your sleep performance. We just didn't have the money to build it. So Keith and Khosla, they come in, they give us the money. At that point, we spend around 15 months to build the pod. By then, we had the SVP of hardware, so he was able to do it properly and on time.

We shipped that, and since then it has been, I would say, things went pretty well. After the launch of the pod, three, four months later, is when Founder's Fund led our round with Trae Stephens, and that was our last formal round. Things are going really well now, people love the product. If you follow us on Twitter, you would see probably many people talking about that.

Patrick: [00:25:34] What have you learned about marketing with this product specifically? What has worked well, what has failed? When marketing a product that's a high price point and very different sort of thing, it's an infrequent purchase.

Matteo: [00:25:46] A lot of things there, as well. I would say the biggest thing is we probably tried to go on too many marketing channels too fast. The result is this is not an impulsive purchase, so our customers, before buying, they want to hear from other people, secondary voices that they trust, that the product is good. They don't want to spend blindly, 3K.

So what we understood is to focus on a smaller community, and really own that community. We try now, for us, is the tech space and that's why you see me so active on Twitter. It's better to own small community that is still large, because it's still millions of people, and make sure that people keep talking about the product, because that will create the snowball effect. Instead of just doing TV, radio, this, this, and that, but you don't create multiple touchpoints with the same customer.

Patrick: [00:26:37] Yeah, it's a fascinating density or heat of the comments that is always really important. What was the most exciting day for you, in the research part of your division, across the firm's history?

Matteo: [00:26:48] I think two things. First, the first prototype that was really cool to me, that was an amazing experience. Once you start using it, you start thinking why there is nothing like that. It's like living in a house without an AC, or in a car without an AC, once you start using it, it's so normal and so obvious that you can't imagine living without that. That was one. The other one, it was we picked the first episode of AFib in one of our clinical trials. You really start understanding that we can really save lives. That is when everything started having a bigger meaning in what we do. Then, one of our employees, a year ago, died with the heart attack in the early morning. Things like that, they shouldn't happen. We could prevent them, hopefully one day.

Patrick: [00:27:36] I love that. Defaults matter, and people are lazy, people don't take active steps often. But, if it's done passively, and collected passively, which a bed is uniquely situated to do, it seems like a cool business angle or idea. What else have you learned about, Peloton I guess is a great example, of turning an infrequent purchase into a frequent purchase? Any nuance to that business model that you think the audience would find interesting?

Matteo: [00:27:59] Yeah. What we did, because we knew and we know, we'll keep releasing new technologies. We didn't want people that bought a bed last year to then find a thing that, immediately obsolete seven months after. So the way we designed our beds is that you can upgrade them at a fraction of the price.

Let's say, tomorrow we release a new pod. You don't have to buy the whole new pod, you just buy the top layer and you replace that in your current product. The foam part will still last 10 years, as it has always done for the mattresses. But, the technology at a fraction of the price, can be updated every year. A large part of our monthly revenue is still people just upgrading. Which, if you think, is like the iPhone model, it's the same concept.

Patrick: [00:28:44] Yeah, that's fascinating. Obviously, layering a subscription like Peloton has done, content for them, probably insights and data for your customers, is another way of doing it. It's pretty fascinating. What has it been like working with your wife? That's a fairly unique co-founder story. I'm involved in another business where that's the case as well, called Bottomless. I'm just fascinated by it. I don't work with my wife. What have you learned about doing that effectively, and not killing each other?

Matteo: [00:29:07] I'll share a couple of stories, because they are funny. The first one is we use WhatsApp for our personal stuff, we use Slack only for business. Sometimes, I might be screaming on Slack about something that needs to be fixed today. Then on WhatsApp, I'm asking what we are going to have for dinner, as we have these two dimensions. And then, she's the one ... I would always talk about work, I would always be working, so she was the one who has set the boundaries. She'd say, correctly, "You need to treat me like a colleague. So it's not that at nine PM, randomly, where we are on the couch, you can start talking about work." But then, once I did it, it was fine, I got that. I cannot talk about that, but I can still Slack you because I would Slack a colleague if I have an idea. So sometimes, we are on the couch at 9:30 PM, and I start writing her. I can hear her phone vibrating on the couch, and it's me, but I cannot talk about that.

Future of Health and Self Movement

Patrick: [00:30:02] That's really funny. What else is exciting to you about the future of this quantified self movement? If we were to just start to dream a little bit and think about what might be possible in the future for us, for our kids, what is the almost sci-fi like potential outcomes that you think we might see in our lifetimes?

Matteo: [00:30:22] Preventative health, I'm so excited about that. We have this vision, let's assume we nail our vision, as we will. I'll give an example. There was the bed of a person, she's a founder in tech. She lost one of her parents in 13 days, because they discovered she had cancer all over her body, and then the person was gone 13 days later. This can be solved, there are a bunch of different types of cancers that can be detected very early. And actually, through if you detect them very early, they are not even that complicated to be solved. The point is it's really hard to detect them early, so by the time we discover them it's usually too late.

The end goal, as we were saying, is not even to detecting day one of cancer, to me it's detecting the fact that you have a high likelihood that that will happen, two years before it even starts. That will happen, multiple companies will be able to help us achieve that, not just Eight Sleep. We have the data, with all the wearables, to make it possible.

Patrick: [00:31:23] What has been the hardest challenge, overall, in the business so far? What was the hardest hill to take?

Matteo: [00:31:33] Obviously, we started in 2015 when hardware, there was a lot of interest around hardware. Then, a lot of hardware companies had problems in the following years. So for a couple of years it's been really tough for us, because investors didn't want to invest in hardware anymore. We were lucky to find the right visionaries that were able to bet on our vision, and take both bets. But, it was not an easy time.

Patrick: [00:31:58] Yeah, it's fascinating.

Matteo: [00:31:59] Hardware is hard.

Formula 1 Racing

Patrick: [00:32:00] Hardware is hard, yeah. Yeah. I have a relative who is building a hardware business, and it's like building a house. It's twice as expensive, and twice as long, with way more problems than you anticipated. It seems to be the rule of thumb. I'd love to hear, switching topics a little bit, the most valuable things that you've learned as a huge fan of Formula 1 racing. This is something I know is a passion of yours. F1's obviously fun and interesting to watch, and exiting, and precise, and high paced. I remember watching the Senna documentary years ago, it's one of my all-time favorite documentaries. But, beyond just the enjoyment of it, what have you learned about the way that business operates that you find most interesting?

Matteo: [00:32:39] It's really an engineering business. Everything is about gaining a tenth of a second. They spend 300 million or 500 million in a year, to just gain a couple of tenths. They need to be reliable. There are a lot of things that really common to what a hardware startup does. You need to almost match perfection with a product that doesn't break and really performs at its best. They have some of the best minds in the world for software. There's all these software engineers. And sometimes, you see the drivers that they start having software problems, and they just change their steering wheel and the problem is solved. Or, you see these engineers connecting to the car with their computer, to reset all the key specs of the car. I think, for the average Formula 1 fan who's not in tech, they don't understand that because they are still very old school F1, it was very mechanical. Well, instead now, the F1 is really a software controlled device, driven by these supermen.

Bringing Software to the Mattress Industry

Patrick: [00:33:41] What do you think will happen to old mattress companies, as you and maybe others are successful in modernizing the bed? What's interesting to you about, everyone knows Sealy, or Serta, or some of these brands? You go in that stupid showroom, and lie on these things. That's got to be a pretty big industry. Purple and Casper are the modern equivalents. What have you learned about all those companies? What else is interesting about the mattress ecosystem or industry to you?

Matteo: [00:34:07] The problem there is foam is really a commodity. If you wanted to start a dumb mattress company today, you could do it today, and you could be able to ship probably within a week. You just call a couple of these large foam manufacturers, you set up a deal with them. Because there are so many companies, they have probably a template, you sign it, and they start shipping for you. There is no technology, so it's really a matter of how do you want to pay for this foam. Do you want to pay 100 bucks, 300, or 1000 bucks? It's really like going to the market and having a menu. And then, you start shipping it, and it becomes just a branding exercise. Obviously, I still have respect for these companies, they're still great people working there. They're big, they're able to move large units. I'm not talking about the people, I'm really talking about the type of business.

I think, if you go on Google and you just Google beds in the Middle Ages, it would still look the same bed. Or, if you look at beds at the time of the Romans, it's substantially the same thing. This means that, for the past 2000 years, no one brought any sort of improvement, technological improvement, to our sleep. Which again, is 33 years of our life if we live 100 years. The last innovation in sleep was memory foam, in 1966, and memory foam is not really a technological innovation it's just a different type of foam. That's the part that I honestly don't like, of the industry.

Patrick: [00:35:28] What other businesses have you learned the most from? In terms of lessons that they've offered you in building Eight Sleep.

Matteo: [00:35:35] I think Tesla, electric cars. From a certain perspective, I'm being ambitious, I think of us like a Tesla of sleep. The reason is when Tesla started, obviously electric cars existed but it was very niche market. And no one was betting on that because the user experience was not good, it was too painful, the batteries were not lasting enough, all that. Elon doubled down on that, and now every other company is trying to catch up with Tesla but they cannot. The reason is Tesla is a software company first, then they also have great hardware capabilities. But, it's not a mechanical company. I also think they built a brand that really stands for electric vehicles, so it will be way harder for BMW, and Mercedes, and all the others to catch up with that. My dream is that the same thing will happen with sleep, where Eight Sleep will be hopefully the Tesla of sleep, and all the other companies, they will try to catch up. But, because it's not part of their DNA to have that hardware skill, and that software and machine learning skill, they will always be behind.

Data and Preventive Sleep

Patrick: [00:36:40] I'm sure, also, that you're collecting valuable data that only you could have, by virtue of having people on these things every night. You've got a dataset, and there's a compounding data advantage of sorts, that you couldn't buy this data from some other vendor. I guess, you could buy sleep study data or something. But, you can't do it in a way that helps you iterate the quality of the product via software, in such a way. Anything you've learned about that part of the business? I'll call it the pure data part of the business, and how to manage that piece well. So that one, you're getting good lessons, and two, those lessons are feeding back into the quality of the product.

Matteo: [00:37:16] Our data is really helpful in two ways. The most obvious is for us to keep improving our product, so more data we collect, the more we can improve our algos. Our algos, they learn. And sometimes, we also ask our users feedback in the app, based on the feedback, we keep improving our algos. But, the second big thing is the amount of data that we collect in the night is probably more than what some of the best sleep professors in the world, they have seen in their whole career. The reason is because, in the past, there were no devices that were able to track sleep, every single night, at home. Five years ago, you could go to a sleep clinic, but a sleep clinic is a very foreign environment, a hospital. They cover you with all these sensors, in this room, and you pretend to sleep, which is not going to happen. Particularly, if you already have sleep problems, and then you add this anxiety, nothing is going to happen. The way we can help doctors and research is by providing three years of your sleep data at home, in your normal environment. And based on that, they might be able to help you to solve some of your sleep issues.

Patrick: [00:38:25] How do you think about gathering other useful data as frictionlessly and passively as you gather sleep data? I'm sure it would be valuable for you to know when somebody's had a glass of wine, or something like this. Or, eat some fried chicken at 10 o'clock, or something like that. How do you think about that? Getting more data into the ecosystem that helps your customers, but maybe doing so in a way that's not invasive, or requires them to open an app every time they do something.

Matteo: [00:38:51] We're pretty obsessed about this passive data, so we want to have as much data as possible without you doing anything. Because I hate doing things, where things are challenging. So, we do it in two ways. On the sleep side with our device, but then we are connected to Apple Health. Through Apple Health, if you use any other device, there will be the data there that then we can pull. We are already running studies internally, where we see when you train, or what time of the day you train, what kind of training. Was it yoga, was it tennis, was it high intensity interval training, whatever? And then, we start working on correlations. That is something our users really want, because they want to know, "Okay, should I train in the morning, or should I train in the afternoon? Or maybe, I'm agnostic and I can do it whenever I want." In my case specifically, as Matteo, if I train in the late afternoon, I struggle to sleep. My sleep quality drops 10%. The reason is I tend to go to sleep early, so if I don't let at least a couple of hours go by before I go to bed, then my sleep is very nervous.

Patrick: [00:39:56] What have you learned about what's going on at Apple, that people might find interesting? Because I'm very interested in what I'll call protocol businesses, where it's the standard that everything pipes into, and data can be pulled out as well. What's interesting about what's going on at Apple Health?

Matteo: [00:40:12] I think Apple Health is becoming this first platform where they are really collecting all the data, plus they are developing their own devices. The Apple Watch is really a medical grade device. Maybe not now, but it will be in a couple of years. Today, it is already, to a certain degree. They created this device that, if you think it's very similar, they use a form factor that people are already using, as we do. They are reinventing the watch, we are reinventing the mattress. The end purpose is to build medical grade devices that can help you live a healthier and longer life. But, the strength that Apple has is that everyone else is pushing data to them, from other wearables or devices. I'm sure they are developing a very large gen database, health database. Myself, I keep using my Apple Watch, I have been using it for three years now. Not so much for the single day, but because I'm really building a dataset about my heart rate. In 10 years from now, we'll look back and I will see what was my heart rate today, and how that is changing. Hopefully by then, machine learning will be at a point that it will be able to tell me I can develop arrhythmia, or I could be subject to a heart attack.

Patrick: [00:41:24] It's a fascinating future. The preventative health side of all this is what's most exciting, as you laid out. I'm an early adopter of all these things, typically, so I recognize it'll be some years before this is everybody. But, you have to figure that it's going to be a generally good thing for national health. What is the thing you're most excited about for the longterm future of Eight Sleep? I realize we've talked about some of these aspects already. But, as you look forward and just put your vision hat on, what thing are you most excited about?

Matteo: [00:41:51] Compressing sleep and preventative health are two things that, probably even today when I share them, people still think I'm crazy. That gives me more energy to really make them happen, and I'm pretty confident that maybe to different degrees, but both of them, they will happen.

They're already happening with our device, because our customers are already falling asleep faster so they are gaining time there. They are already getting better sleep, so they are improving their efficiency. And, we are already detecting some diseases.

Patrick: [00:42:20] Fascinating stuff. I've loved using it, and I think people listening know I'm into all this measurable, wearable, quantified self stuff. But it's not just for fun, I think the impacts can be quite interesting. It's been a pleasure to meet you in this format. I ask the same closing question of everybody. Which is to ask, what is the kindest thing that anyone's ever done for you?

Matteo: [00:42:38] I would guess my wife, she has taken care of me, moved both times. Even recently, I had a really bad day at work, and I was crazy, I was a bit down. She made some dinner for me, and gave me energy again. That just happened last week.

Patrick: [00:42:53] Very simple. Matteo, great to see you, thanks for doing this.

Matteo: [00:42:57] Oh, my pleasure.

Created by @mercedeschien