#26 – Sam Corcos

Levels has positioned itself as an engineering-first company, so it’s fitting that CEO Sam Corcos has a background in software engineering. He’s able to foster the company’s core competency of product development. In this episode, Sam discusses metabolic health, bio observability, the importance of having a low ego, and how the best moat is the best product.

03:20 – Growth can pause product development

Levels is staying in that product development phase as long as it takes to get the product-market fit right, because once you’re in the growth phase, you start being risk-averse.

Then you start focusing all your end resources on scaling and you start focusing everything on sales. And sometimes if you’re not intentional about it, growth can be the first stages of death for a company. It is when you stop iterating, you stop improving, loss aversion sets in, and you’re no longer willing to take big risks. I try to fight against this culturally within our company by setting unrealistically bad expectations for things. Like I told Marillo recently for a project that … He’s one of our engineers. There’s a feature that had been stagnant for several months. And I really wanted this to get shipped and we kept polishing it and the scope kept increasing. And I basically told Marillo, “I don’t care what it is. I want this shipped by the end of the month, our target is to make something that is so bad that we lose 20% of our users. That’s the goal. I want it to be so bad that we lose 20% of our users. And that would be fine with me, just get it shipped.”

06:52 – Close the biological observability loop

Taking an action and then seeing the effects in your own data is the closed-loop that is necessary for health-based behavior change.

It’s once that loop is closed, that is when behavior change can really start. You can trust it’s in the study, but once you see it in your own data, that’s in control theory. It’s like a closed-loop system where you can really see cause and effect. In observability theory, they call that the tracing layer. Software has something similar where this function called that function, which called this query, which called that function, which then caused the whole thing to explode. Being able to have that for the human body is really the long-term goal of what we’re solving in the product.

10:51 – Improve productivity with small meetings and asynchronous information sharing

Levels almost never holds meetings with more than three people because it’s more efficient, information is recorded and shared, and people can do more deep work.

There are obviously exceptions to these things. There are times when you need more than three, but once you experience the efficiencies of having fewer people in meetings and you reduce the politics of who’s in what meeting, because these things are already distributed and recorded and distributed anyway. You don’t miss out on information. The leadership meetings are recorded and sent to the entire leadership team. It doesn’t really matter if you’re there in person or not. It’s actually way more efficient with your time. You can stay focused on what you were working on. You don’t need to have this interruption. The advantages are for engineers, you will be more productive than you have ever been in your life, because all you’re doing is the stuff that you want to be doing. And you’re shipping lots of code.

13:58 – Find your core competency

For Levels the core competency is engineering. That means the focus for optimization is development velocity.

I think knowing what your core competency is, is a really important thing for every company. There’s really no judgment as to what a core competency is for different companies. A good example of this might be Casper, the mattress company. Their core competency is in growth marketing, and they’re really good at it. It’s like branding, growth marketing, these types of things. They’ve done that very successfully. They are not a product engineering company. It’s actually not super useful for them to build the type of culture that we’re building. Tesla is a very hands-on manufacturing buildings type of company. So is SpaceX. You need people who are in person, who are really building stuff. When you’re in software, you should really just think about what is it that you’re optimizing for? In our case, our North Star is maximizing development velocity.

18:55 – Build the best product, not the best moat

Instead of spending time and resources on building a defensible moat, build the best product. It will be inherently defensible.

Something that is defensible is you just have the best product. You didn’t build the best product to keep other people out. You just built it because it’s the best thing. It just happens to also be defensible. A lot of these network effects are similarly defensible. You didn’t build a network effect to keep other people out of the market. You did it because it makes the best product. This is something we talk about internally in terms of strategy is are we building a moat or is the thing that we’re building just naturally defensible because it’s better? To use the moat analogy even further, the Mongols are defensible because they never had a home base. There’s nothing to conquer. They’re just a roving band. Whereas castles are static. They never move. If you wanted to seize your castle, it’s just a matter of time, if all they’re doing is moat building. That’s a lot of the way that we think about these types of problems is how do we build the NPS unicorn, the next big NPS unicorn that makes a product that is so good that everyone wants to use it?

22:51 – Hire people better than you

As CEO, Sam had to let go of shipping things all the time and learn to be honest about his own failures so that he could build the best team around him.

The reality is that the best way to see yourself kicked out of the company is if you don’t bring on people who are better than you. If you’re trying to build a fiefdom and you don’t let the right person run the company in the right way, in the right role, you are the bottleneck at that point. That’s something that can be really emotionally difficult is to let go of one’s ego. I would say that humility is a trait that you really need to develop as a CEO in a way that you don’t as an engineer. As an engineer, you’re shipping things all the time. In the CEO role, you have to recognize where your own deficiencies are, where the deficiencies are in the team. You have to really be honest with yourself about your own failures and be able to make up for those by hiring really good people who can solve those problems.

24:11 – Observe what’s happening inside the body

We need a sensor system for observing the inside of the body the way we can observe the health of machines.

When I think about biological observability, I think about in almost every other system that we have, especially mechanical systems, we have a lot of things that we’re measuring. The example that I think most people can relate to is we have many orders of magnitude more data about the health of our cars than we do about our own bodies, which is kind of a problem. The goal is to figure out a way to add sensors and add ways of measuring what’s going on inside the body. I think right now a lot of these metrics are, and I mean in a literal sense, they are superficial, meaning outside of the body. And that is a good starting point. It’s better to have at least superficial data, than no data. Where it really starts to get interesting is when you can measure molecules. Glucose monitors are really the first. It’s the tip of the spear for that. Where it really gets interesting is in multi molecule.

27:59 – We need transparency in the biometric industry

Each company has its own proprietary way of interpreting data which makes it hard to understand. Transparency across the industry could improve everyone’s data.

You get your Whoop score and it says, “Great, you did a 90.” Then you get your Aura, it’s like you were a 40, and Apple says you were at a 26. It’s like, “Wait, which? Those are all completely different things. I have no idea what this means.” We’ve been talking about ways to either through an industry consortium or by open-sourcing our own algorithms for these things and just making it very clear what data we’re using and why, and trying to make it so that anyone can use the algorithms that we’re using and making it really transparent. We don’t want it to be a black box. We want people to understand why the decisions, why the numbers are coming back the way that they are. Some of that I think is just cultural, which is people think that keeping things secret is useful. And I think it’s actually entirely counterproductive to the entire industry. This ties in a lot to our culture of transparency. I think that’s a really big part of it.

30:27 – Have great people come to you

Levels aims to be transparent about company culture and how things work within the company so that the right candidates will find us.

I think this is similar in many ways to our NPS unicorn growth strategy, which is what is the hardest thing to do for recruiting, which is how about instead of we find them, they find us? How do we put enough material out there to convince all of the best people in the world that we are the company they want to work for? It’s obviously way harder to do that, but if you can pull that off, you’re in a much better position. A lot of our content efforts, a lot of the transparency of publishing all of our investor updates, which we do publicly, it’s on our website, publishing all of our weekly team all hands, publicly, giving people a sense of what the team culture is like.

35:18 – Look for people with low ego

Sam said that Levels looks to hire senior people with low ego. What’s most important is finding people who will add value wherever it’s needed, regardless of title.

Most recently, David, co-founder who ran product for the first couple years of company had a similar recognition. David is really good at scoping and specing and doing a lot of the nitty-gritty stuff, but he lacked a lot of the skills on the operational side of a product org. We started an external search and realized we actually have somebody internally who was the best person for it. We elevated him a couple weeks ago to be the new head of product and super low ego. It was just the thing that was the right thing for the company. Nobody here is building a fiefdom. That’s one of the biggest things that we look for is people who are willing to add value wherever value can be added.

Alex LaBossiere (00:00):

Sam Corcos is a co-founder at Levels, a company on a mission to solve the metabolic health crisis. Most recently, he was the co-founder and head of technology at CarDash, a company making automotive repair and maintenance convenient and transparent. He previously started Sightline Maps with Ben Judge, where he worked with special operations command to develop mapping technology under a cooperative research and development agreement. He’s also written a book on building distributed concurrent web services using react in Phoenix called Learn Phoenix. As if that wasn’t enough, he worked at the oncology center at UC Davis for two summers. Co-authoring a paper during his time there, which has too many words I can’t pronounce in it. I’ll leave it to you to read more at Corcos.io Today, we talked about metabolic health, running a remote company, bio observability, and more. Hope you enjoy. All right. Do you want to kick things off here?

Sam Corcos (00:53):

Yeah, sure.

Alex LaBossiere (00:54):

To start things off and before we get into the details, I think it’d be a good time to set the stage here. There’s a stat that Levels put out that 88% of Americans have at least one marker of metabolic dysfunction, meaning only 12% are considered metabolically healthy. For anybody listening, who isn’t maybe familiar with Levels or what you guys do, I’d first like to ask how you usually describe the company and very broadly why metabolic health matters?

Sam Corcos (01:23):

Yeah. Right now our focus is on showing you how food affects your health and the mechanism through which we do that is using biosensors. We’re currently using continuous glucose monitors, because that’s the best technology that’s available to show you what is happening inside your body in real time. We also do supplemental blood work and nutritionist and functional medicine doctors through a marketplace as well. But the real focus is on biosensors. The goal is to show you cause and effect. When you eat something, most people rely on dogma or some dietary philosophy that’s loosely grounded in fact, but it’s hard for people to know is keto the right answer? Is vegan the right answer? There’s truth to all of these things, but it’s very hard to know unless you actually have a feedback loop that shows you how your choices are affecting you.

Sam Corcos (02:26):

That’s the goal of the company is really to solve this. We’re still pretty early. We’re still in closed beta. Hopefully we’ll open up the beta by the end of this year. We’re not in any particular rush. It’s really just making sure that we’ve nailed product market fit, but we also want to make sure that we’ve solved fundamental incentive alignment. This is something that I … One of the benefits of being a fourth time founder and having been through this the hard way is that growth is not always a good thing, which sounds counterintuitive to a lot of people who haven’t been through it, but growth can be an intoxicant. Once you are in growth mode and the line starts going up and to the right very quickly, it just sucks up all of the energy of the company. Everyone wants to be on growth projects and people don’t want to do product development anymore.

Sam Corcos (03:20):

Then you start focusing on your end resources on scaling and you start focusing everything on sales. And sometimes if you’re not intentional about it, growth can be the first stages of death for a company. It is when you stop iterating, you stop improving, loss aversion sets in, and you’re no longer willing to take big risks. I try to fight against this culturally within our company by setting unrealistically bad expectations for things. Like I told Marillo recently for a project that … He’s one of our engineers, Matt. There’s a feature that had been stagnant for several months. And I really wanted this to get shipped and we kept polishing it and the scope kept increasing. And I basically told Marillo, “I don’t care what it is. I want this shipped by the end of the month, our target is to make something that is so bad that we lose 20% of our users. That’s the goal. I want it to be so bad that we lose 20% of our users. And that would be fine with me, just get it shipped.”

Sam Corcos (04:36):

And of course we got a shift and it wasn’t great. We didn’t lose any users and then we quickly followed up and made it better. But you can see how the loss aversion kicks in. We’re like, “Well, what if something bad happens? What if it’s not as good? What if it’s a regression?” When you’re early, you just have to be okay with that. And you have to just move quickly and you have to try as many things as possible. A lot of what I’m doing right now is just culturally fighting against that kind of thing. The inertia of early indications of success can really be toxic to a company. We’re still very early. We’re about 35 people on the team now. We are hiring, I think, 30 more engineers in the next 6 to 12 months. It’s going to be a wild ride in 2022.

Alex LaBossiere (05:27):

Exciting stuff. Second and last question to set the stage here. You seem very, very, very focused on products. What’s your vision for the future of Levels? Is that something you spend a lot of time thinking about?

Sam Corcos (05:39):

Yeah. The vision for the future of the product is to solve the tracing layer of biological observability. And what that means is once multi molecule sensors become a reality, which will probably happen in the next roughly five years, being able to show you how your choices affect your health, and that can be anything from food, which is a huge part of it. It can also be parts of lifestyle. One of the values that a lot of people get from things like Whoop and Aura … I used Aura for several months and it was super helpful. I knew this already from reading books where I knew that drinking alcohol before bed is bad, but it had not really clicked for me until I saw the data come back from my Aura Ring, where I realized like, “Whoa, this is actually really bad and I need to stop doing this. This is not worth it.”

Sam Corcos (06:52):

It’s once that loop is closed, that is when behavior change can really [inaudible 00:06:58]. You can trust it’s in the study, but once you see it in your own data, that’s in control theory. It’s like a closed loop system where you can really see cause and effect. In observability theory, they call that the tracing layer. Software has something similar where this function called that function, which called this query, which called that function, which then caused the whole thing to explode. Being able to have that for the human body is really the long term goal of what we’re solving in the product.

Alex LaBossiere (07:28):

Really interesting stuff. Another thing that’s particularly interesting about Levels and something that’s gotten a lot more attention, especially as of late, is the fact that you guys are totally remote company. Having been through the ringer with online college classes myself, I can say pretty confidently it’s not something many people have been able to execute on properly, but you all were remote even before the pandemic hit. Is that right?

Sam Corcos (07:52):

Yep. That’s right. Since day one.

Alex LaBossiere (07:54):

It seems to have worked for you all. What would you say are the biggest advantages of remote for Levels?

Sam Corcos (08:00):

Yeah, t’s definitely a trade off. I would not strictly define it in terms of advantages. I literally got an email one hour ago from a great engineering candidate who passed on moving forward in the process for Levels because he said, “I really appreciate all the material and the transparency. I’m really looking for a co-located team where I can be with other people.” This is not a failure. This is actually the success of the process is to make sure people understand the implications. There are major positives, which are especially for engineers or for people who need lots of deep focused work time. There is no better way to do it than remote because I think remote and async are really intertwined. If you want to do remote successfully, you have to embrace asynchronicity.

Sam Corcos (09:00):

Our engineers average, I think two meetings per week, which is … Yeah. When we joke about this internally, when we hear companies have this revelatory thing where they’re like, “We’re going to do no meeting Mondays” and they have one day per week of no meetings and we’re like, “Yeah, we have some meetings, rarely.” It’s like we could have no meetings Monday, Tuesday, Wednesday, Thursday, and optional meetings on Friday. And that would be just normal for us. That’s one of the big advantages, is it forces good habits of documentation. This is I think one of the reasons why companies are floundering, some companies are having a really hard time adapting it’s because they’re taking old systems and they’re trying to just jam them into what remote is. You end up just spending eight hours a day on Zoom calls and that’s worse.

Sam Corcos (09:55):

The way that you solve this … In fact, I was talking to somebody the other day who their leadership Zoom call is like 70 people. They’re all on that Zoom call together. And my response to him was what if you just did the one person who was presenting and they just recorded it, and then they sent it out to everyone and they could watch it on their own time if it’s relevant to them at 2x? What if you did that? Because that’s what we do. We have a rule internally that we have no meetings of more than three people. And all of our meetings are recorded and distributed internally. The no meetings of more than three people was more controversial when it was proposed than it is now. I think people realize now that there is always some arbitrary cap to the number of people in the meeting.

Sam Corcos (10:51):

There are obviously exceptions to these things. There are times when you need more than three, but once you experience the efficiencies of having fewer people in meetings and you reduce the politics of who’s in what meeting, because these things are already distributed and recorded and distributed anyway. You don’t miss out on information. The leadership meetings are recorded and sent to the entire leadership team. It doesn’t really matter if you’re there in person or not. It’s actually way more efficient with your time. You can stay focused on what you were working on. You don’t need to have this interruption. The advantages are for engineers, you will be more productive than you have ever been in your life, because all you’re doing is the stuff that you want to be doing. And you’re shipping lots of code.

Sam Corcos (11:41):

If you’re in other roles, it really enables you to leverage your time. These are things that could ostensibly be done in in person teams using some of these same asynchronous tools. It’s an entropy problem where it’s so much easier to just tap someone on the shoulder when they’re right next to you. And it’s like, do I really want to record a video of myself when I’m right next to this person and then send it to them through email and get their response asynchronous? No, I’m just going to tap them on the shoulder, because they’re right next to me. And a lot of this comes from my … I largely wanted to build the company culture around what I wish it was when I was working on different engineering teams. This is my first non technical role.

Sam Corcos (12:33):

I’ve always been on the engineering side. I really just thought as an engineer, what would I like to be the case? It’s like, “Well, you know what I really hated in previous companies is when the ops team or the sales team would break into the engineering room and grab me by the shoulder and force me to work on their project, even though it was not priority. And I was supposed to be working on something else.” They’re physically right next to me. They’re like, “No, no, no, you have to work on my thing.” It’s like, “But I have to work on this other thing.”

Sam Corcos (13:02):

“No, no, no, you have to work on my thing now.” And you don’t really have that challenge when you’re working remotely because they can send you an email and you can be like, “Cool, thanks for the email. I’m going to to go back to what I was doing.” In fact, I won’t even see the email because I don’t check my email more than a couple times a day. There’s really no way to strong arm people into doing other things. It allows for a level of deep focused work that you really can’t get in any other environment.

Alex LaBossiere (13:33):

You’re very much if it hasn’t been apparent already, an engineering first organization. And naturally, like you mentioned before, being an engineer before moving into this role as CEO, you kept it that way. I think Tesla comes to mind as another company that leans into this philosophy really heavily. But I’m curious to hear you dig into that a little bit further. Why is that the case?

Sam Corcos (13:55):

Yeah, I think it’s something that we’ve been thinking a lot about as well. I think knowing what your core competency is, is a really important thing for every company. There’s really no judgment as to what a core competency is for different companies. A good example of this might be Casper, the mattress company. Their core competency is in growth marketing, and they’re really good at it. It’s like branding, growth marketing, these types of things. They’ve done that very successfully. They are not a product and engineering company. It’s actually not super useful for them to build the type of culture that we’re building. Tesla is a very hands on manufacturing buildings type of company. So is SpaceX. You need people who are in person, who are really building stuff. When you’re in software, you should really just think about what is it that you’re optimizing for? In our case, our North Star is maximizing development velocity.

Sam Corcos (15:04):

You’ve probably seen some of our Friday forums and our monthly updates. We ship a lot, we ship a lot of code. We bring on really excellent engineers and we stay focused on how do we maximize velocity? I think in our case, recognizing that our core competency is in consumer software and data science. Leaning into that as much as we possibly can and building the culture around that. Every other entity within the company supports that, which is our core competency. I think that’s the real focus is that it’s a similar thing with people often ask around our marketing strategy and they ask like, they’ll say, “How do I build a great content machine like you guys did? How do I do SEO?”

Sam Corcos (15:53):

And sometimes my answer is like, “Well, maybe you shouldn’t do it.” We knew before we took on that strategy that it was going to work because we talked to people about where they get their information from. And we knew that that would be an avenue that they would use. Once we knew that it would work, then we built it. I think sometimes people approach these things the wrong way. They often end up building a product that they assume people will want, as opposed to figuring out what people want and then build the product.

Alex LaBossiere (16:29):

You’re anticipating my next questions really well here. I was going to go on to say you haven’t spent any money on marketing, right? In that same vein, word of mouth seems to ensure you’ve got a great product. Is your reasoning as simple as that? Or is there more to it?

Sam Corcos (16:44):

Yeah. I mean, the classic YC motto of build something people want is definitely applicable. Our growth strategy, our head of growth that a podcast on this called building an NPS unicorn. Our growth strategy is to build a product that has an NPS over 90. Which is pretty insane. I think only Tesla and Peloton have successfully done this. Those are the only two companies in history that have maintained that. Our North Star for growth is how do we build the best possible product? If your NPS is over 90, all of the other problems basically solve themselves. If almost everybody loves your product and is recommending it to people, you’re in really good shape. We’ve had some internal conversations around performance marketing and all these different ways that you can spend money on marketing.

Sam Corcos (17:46):

And it’s just very hard to justify. Would I rather spend a million dollars in ads to juice our growth numbers? Or would I rather hire several engineers, like 10 more engineers? It’s like, well, obviously I would rather hire 10 more engineers. Let’s just do that and make the products 10 times better. Then people will want to use it more because it will be better. I think about this stuff a lot. We’ve been talking about things that are defensibility as a company, and people often use the word moats and I’m not a big fan of moats per se. I do think there is a difference between moats and defensibility. It’s maybe a semantic distinction, but a moat is something, if you imagine the actual metaphor of like you have a castle and you dig a ditch around it, a moat is something that you intentionally build with the intent to keep other people out of your market, keep them away from your castle.

Sam Corcos (18:55):

Something that is defensible is like you just have the best product. You didn’t build the best product to keep other people out. You just built it because it’s the best thing. It just happens to also be defensible. A lot of these network effects are similarly defensible. You didn’t build a network effect to keep other people out of the market. You did it because it makes the best product. This is something we talk about internally in terms of strategy is are we building a moat or is the thing that we’re building just naturally defensible because it’s better?

Sam Corcos (19:34):

To use the moat analogy even further, the Mongols are defensible because they never had a home base. There’s nothing to conquer. They’re just a roving band. Whereas castles are static. They never move. If you wanted to seize your castle, it’s just a matter of time, if all they’re doing is moat building. That’s a lot of the way that we think about these types of problems is how do we build the NPS unicorn, the next big NPS unicorn that makes a product that is so good that everyone wants to use it?

Alex LaBossiere (20:11):

I mentioned earlier, as did you, that before running Levels, you came from an engineering background. I’m curious. Are there any big lessons or surprises you’ve taken along the way in making that transition?

Sam Corcos (20:22):

Yeah, One of the big ones would be, it took me close to a year to be able to emotionally grapple with not having deliverables anymore. That was a weird thing as when you’re an engineer, you ship stuff all the time. And it’s really gratifying to have deliverables that you’re shipping that are in the product. You always have something to show. And often times as a CEO, my role is just like shuffling information around. I have between 1% and 20% responsibility for basically everything that happens in the company, but I don’t actually … I’m not actually responsible for anything, but also I’m ultimately responsible for everything. It’s a very strange role where it’s like we made the decision to do this thing and I helped coordinate some stuff, but I’m not really responsible for execution.

Sam Corcos (21:34):

When it’s done, I’m not the person who’s presenting it to tell the team about it. It’s just, I’m the guy in the background, shuffling information around, doing a lot of email. That took a while to get used to, and just emotionally accept that that is my role and that is what I’m supposed to be doing. I found that doing long form strategic writing really helped a lot with that same sort of sense of deliverable. I’ve been doing a lot more writing, which has been really helpful for me, which I think for some people that might be a similar thing that they’d be able to get to similar value out of.

Sam Corcos (22:09):

I would say that one of the other big things, this is maybe just applicable in general is to really put a lot of effort into bringing on the best people. This is something that I see as as a failure in first time founders especially, is a fear of giving up one’s LEGOs. I advise some early stage companies just informally, a lot of them are friends of mine. And a common thing that I see is there’s a fear of bringing on people who are better than them. Because they might get replaced and they might kick them out of the company. That’s the fear.

Sam Corcos (22:51):

The reality is that the best way to see yourself kicked out of the company is if you don’t bring on people who are better than you. If you’re trying to build a fiefdom and you don’t let the right person run the company in the right way, in the right role, you are the bottleneck at that point. That’s something that can be really emotionally difficult is to let go of one’s ego. I would say that humility is a trait that you really need to develop as a CEO in a way that you don’t as an engineer. As an engineer, just shipping things all the time. In the CEO role, you have to recognize where your own deficiencies are, where the deficiencies are in the team. You have to really be honest with yourself about your own failures and be able to make up for those by hiring really good people who can solve those problems.

Alex LaBossiere (23:47):

I want to take a minute to dig a little bit deeper on bio observability and what that looks like in the future in your mind. We talked a little bit about vision at the beginning of this conversation, and you mentioned we’ve seen a lot of great consumer products in the space. The Whoop band, the Aura Ring, I know Eight Sleep is doing with mattresses now. What does bio observability look like and say 5, 10 years time and how do you see Levels fitting to that more broadly?

Sam Corcos (24:11):

Yeah. When I think about biological observability, I think about in almost every other system that we have, especially mechanical systems, we have a lot of things that we’re measuring. The example that I think most people can relate to is we have many orders of magnitude more data about the health of our cars than we do about our own bodies, which is kind of a problem. The goal is to figure out a way to add sensors and add ways of measuring what’s going on inside the body. I think right now a lot of these metrics are, and I mean in a literal sense, they are superficial, meaning outside of the body. And that is a good starting point. It’s better to have at least superficial data, than no data. Where it really starts to get interesting is when you can measure molecules.

Sam Corcos (25:13):

Glucose monitors are really the first. It’s the tip of the spear for that. Where it really gets interesting is in multi molecule. There are a lot of companies working on that right now. I expect we’ll see multi molecule sensors on the market probably within five years. And once you can start correlating outcomes, decisions with outcomes, I’ll give you some examples. Glucose is interesting, but it’s optimistically 10% of the equation if you want to figure out what your choices actually mean for your health. That’s very optimistic, 10% would be … It’s like maybe even 1%, because there are false negatives and there are false positives when you only have the one metric. If you have a really intense workout, you’re going to see a glucose spike, but it it’s not caused by food. It doesn’t trigger an insulin response.

Sam Corcos (26:06):

That’s just your liver dumping glycogen. It’s totally normal. If you down a liter of fructose, you’re not going to see anything because it’s a different monosaccharide. It’s a different sugar. It doesn’t show up in a glucose monitor, even though it’s way worse for you. If you drink a bunch of alcohol, it’s going to cause all these different problems, but it’s not going to show up on your glucose monitor. “Oh, it turns out alcohol is fine.” Well, actually it’s not, but it’s not the only thing that matters. It’s just like one of many things that matter. Being able to pull in more data from different sources and especially at the molecular level, that’s what matters a lot.

Alex LaBossiere (26:47):

To add on to that, doing bio observability well seems like a really difficult thing. And you mentioned the differences between a lot of these products and your approach to that. I’ve seen some videos comparing some popular products that actually end up getting concerningly different readings from time to time.

Sam Corcos (27:04):

Yeah, sure.

Alex LaBossiere (27:05):

What do you think specifically makes it so hard to do well?

Sam Corcos (27:08):

It’s something that we’ve talked about a lot internally. I think some of this is cultural. Are you talking about the sensors in particular? Are you talking about like sleep metrics and stuff like that?

Alex LaBossiere (27:21):

Well, both, both.

Sam Corcos (27:23):

Yeah. We’ll take the sleep metrics as a really good example. Because we have actually several internal documents talking about this. The challenge is that the science is still in development. It’s still very cutting edge science and where I think a lot of … I think that whole space has really shot itself in the foot because everyone comes up with their own top secret, patented, proprietary metric. When every company has that, you get your Whoop score and it says, “Great, you did a 90.” Then you get your Aura, it’s like you were a 40, and Apple says you were at a 26.

Sam Corcos (28:08):

It’s like, “Wait, which? Those are all completely different things. I have no idea what this means.” We’ve been talking about ways to either through an industry consortium or by open sourcing our own algorithms for these things and just making it very clear what data we’re using and why, and trying to make it so that anyone can use the algorithms that we’re using and making it really transparent. We don’t want it to be a black box. We want people to understand why the decisions, why the numbers are coming back the way that they are. Some of that I think is just cultural, which is people think that keeping things secret is useful. And I think it’s actually entirely counterproductive to the entire industry. This ties in a lot to our culture of transparency. I think that’s a really big part of it.

Sam Corcos (29:00):

The second piece on the accuracy of the sensors themselves, this is something that just improves slowly over time. They they tend to track the relative difference pretty well, but the [inaudible 00:29:13] differences can often be plus … I think the mean average variance is plus or minus 10%. You can get sensors that are like 20 or 30 points off out of a hundred, just because one’s 10% high, one’s 10% low. And that’s just the way it is right now, but they’re getting more and more accurate every year. Another reason is that they are calibrated towards people who are diabetic. Because that’s the on-label use case is for diabetes management. They’re calibrated for people that have much higher ranges.

Sam Corcos (29:46):

Over time, these things will become more accurate. Every year they get better. They are currently sufficiently accurate and sufficiently affordable to be used by a subset of the population. This is the nature with all products like this, is that they start out pretty expensive and then over time they get cheaper and more accessible so that everyone can use them.

Alex LaBossiere (30:12):

You know, as an early stage startup, you’ve mentioned this a little earlier, you obviously have a lot of focus spent on recruitment and hiring. Broadly, just as start, what’s your approach to recruitment and finding really great candidates?

Sam Corcos (30:27):

Yeah. I think this is similar in many ways to our NPS unicorn growth strategy, which is what is the hardest thing to do for recruiting, which is how about instead of we find them a find us? How do we put enough material out there to convince all of the best people in the world that we are the company they want to work for? It’s obviously way harder to do that, but if you can pull that off, you’re in a much better position. A lot of our content efforts, a lot of the transparency of publishing all of our investor updates, which we do publicly, it’s on our website, publishing all of our weekly team all hands, publicly, giving people a sense of what the team culture is like. We’re actually in the process of commissioning a content series with another content creator that we did one recently on team culture.

Sam Corcos (31:25):

That was really fun. And we’re doing another one now on the super tactical nuts and bolt of execution of building remote teams. People have asked us, how do you do knowledge management within the company? And we’re going to go through step by step, how to build a Notion database and how to manage knowledge and how to distribute that knowledge within your team and all of the policies, really nitty gritty execution of how we operate as a team. Part of that is internally, just so that everyone on the team knows how all of this works, but really a big part of it is externally. One, I’d like to up level people who are starting companies and just have more companies exist in the world, but also people who … It’s like Tesla’s AI day.

Sam Corcos (32:14):

I don’t know if you watched that, but yeah, I watched that video and I remember thinking if I was working in AI I would work at this company, no question. There on a sales pitch at the end, they weren’t saying, “Yeah. buy a Tesla.” It was basically like, clearly this was built for the purpose of recruiting. This is a recruiting pitch for all of the smartest people. We’re thinking about it very similarly, is how do you build a team and a culture where all of the smartest people … The term in network theory is preferential attachment. How do you make it so that all of the best people preferentially attach to you?

Alex LaBossiere (32:55):

It’s funny, it sounds like your marketing strategy and your recruitment strategy are pretty much the same thing, which is just build the best thing. I think-

Sam Corcos (33:03):

Yeah, exactly.

Alex LaBossiere (33:03):

I think that’s pretty cool. To dive a little bit deeper into that, what kind of person makes a great hire at Levels? You mentioned this conversation you’d had with this prospective engineer earlier. What kinds of qualities do you really look for and why might Levels be a great place for them?

Sam Corcos (33:18):

Yeah, I think one is an emphasis on deep focused work time. For some people, it can be anxiety inducing to not have that external structure imposed upon you. We treat people like adults. That’s probably our top value within the company is treat people like adults. As a consequence, we also expect people to act like adults. People who like having that flexibility, we’re not going to check in and see how many hours you’re working. Nobody cares as long as you get your work done. If you only work two days a week, nobody’s going to know or care as long as you’re getting your work done. If you want to work weekends, great. If you don’t, whatever, nobody cares. If you only want to work weekends, that’s fine too.

Sam Corcos (34:14):

As long as you’re getting your work done, that’s great. I would say another thing is we tend to hire more senior people, and I think a lot of that has to do with ego. We have very low ego people on the team. Good examples of this are, if you just look at the transitions we’ve had within leadership in the last year, my co-founder Josh was running operations for us. Then it became very clear that the operations function had grown beyond what he was comfortable with. We had a conversation about it, and we brought on a new head of operations who has been really excellent. Josh and I actually did a podcast on this, on a whole new level, the Founder Dynamics podcast. And he talks about his own emotional struggle, getting over that ego barrier to be able to do that.

Sam Corcos (35:18):

We similarly, most recently, David, co-founder who ran product for the first couple years of company had a similar recognition. David is really good at scoping and specing and doing a lot of the nitty gritty stuff, but he lacked a lot of the skills on the operational side of a product org. We started an external search and realized we actually have somebody internally who was the best person for it. We elevated him a couple weeks ago to be the new head of product and super low ego. It was just the thing that was the right thing for the company. Nobody here is building a fiefdom. That’s one of the biggest things that we look for is people who are willing to add value wherever value can be added, and who are not attached to … Ben is our head of growth right now.

Sam Corcos (36:10):

He hasn’t done anything that would be considered growth function in probably six months, he’s been working on content creation. He’s been working on internal company culture stuff. It’s like where can people add the most value? And that is where we put them. There’s no politics around titles. Some of that I think is company stage. We’re still 35 people. It’s still pretty small. You can get away with that for a while. But most of it I think is the culture that we’ve built and the types of people that we bring on.

Alex LaBossiere (36:47):

To start to wrap us up here, I thought it’d be fun to go a little bit left field. I’d like to hear a little bit about what keeps you curious. In one of our email exchanges before we sat down, you mentioned you host regular salon dinners with founder friends to talk about everything from the future of education, to the value of art. How did that all start for you?

Sam Corcos (37:08):

Yeah, it’s been a real source of personal intellectual value, hosting these dinners. It started accidentally. I had gone to some things that were similar. They tend to have a speaker though, which is not really a format that I like as much, because one of the downsides of reading as many books as I do is that usually I will have read the person’s book before I go to the event. Then they end up spending the entire event just summarizing their book because nobody else has read the book. They have to just summarize the whole thing. And I started hosting small dinners with a couple friends and these types of topics would organically come up and I would bring them up to people, other friends of mine. Like, “Yeah, I had this really great dinner. We talked about X.”

Sam Corcos (38:06):

And they’re like, “Oh man, that’s so cool. I’d love to come to one of those.” Then it expanded into then there were five people, then there were 10 people. At some point we got to about 20 people, which turns out to be way too many. Then at that point I was like, “All right, I need to cap the number of people that come to these and start writing this down.” Because it certainly gets a little bit out of hand. I actually have a Notion doc on the whole process that I have for hosting these things, but as you’ve probably come to expect, I have documentation for every part of my life. That’s been a big part of it, is just they are purely selfish. The salon that I hosted on the future of K-12 education, I have some open questions myself.

Sam Corcos (38:57):

What is the point of education? What does any of this mean? And I try to get an intellectually diverse people with different backgrounds that can represent whatever topic is there. I had teachers, I had former teachers, I had administrators, I had ed tech founders, just a really diverse group of people who can bring different perspectives into the conversation and really just ask these open-ended questions and just listened to the people who are smarter than me talk about it in my presence. And I learned a ton from that particular salon. That was the most recent one that we did. A lot of these is just me trying to pull information from people who have been way, way deeper on these topics than I have.

Alex LaBossiere (39:46):

Do you have any favorite topics?

Sam Corcos (39:50):

Yeah. I mean, very few of them are recurring topics. There have been a few that were surprising and interesting. The future of K-12 education was really good because it’s a topic that I really care about. We did one on the declining role of institutions, which was interesting. And also sometimes the subject you just spend the entire time to trying to define what the words even mean. It’s like, “Well, what are institutions?” Then the question is like, “Are they declining?” And we spent basically the whole time just talking about that. And that alone was sufficiently interesting of like are communities institutions? Are companies institutions? Are government institutions? Where do you draw the line for that? Then you often end up drawing something that’s kind of arbitrary and then are they in decline? Well, for most of the things that we defined as institutions that we thought were declining, the revenue numbers are going up.

Sam Corcos (40:55):

Does that mean they’re declining? Certainly trust in a lot of these institutions is declining, but they seem to be making a lot more money. Maybe they’re not actually in decline, maybe that’s just in our own heads. Another one that we did that was really interesting was anonymous, unlimited political contributions, which is another one of those topics where I had a really diverse group of people in the audience. And I wanted to better understand whether it’s good or bad and pretty much everyone came in with one set of assumptions, which is that obviously it’s bad and came out the other side with the majority of people in favor of it.

Alex LaBossiere (41:37):

Wow.

Sam Corcos (41:39):

Yeah. Once you really think through the implications of what it means to ban anonymous contributions and to limit the amount that you can contribute, it’s one of those things where in a utopian ideal, you could do this, but it’s like the James Madison quote. “If men were angels, we would not need government.” You can’t start with the assumption that we’re all going to be operating with positive intent. You can’t assume that. You have to think through like, “All right, practically speaking, how would this actually work?” The answer is it would not. You have to think through what compromises you’re willing to make and it is messy. There have been many salon dinners like that, where there were a lot of learnings for me personally, around different issues that I was grappling with.

Alex LaBossiere (42:33):

That sounds like a really great time. Listen, Sam, I appreciate you doing this.

Sam Corcos (42:36):

Awesome. Well, we’ll talk soon.

Alex LaBossiere (42:37):

Appreciate you making the time.

Sam Corcos (42:39):

See you later.

Alex LaBossiere (42:39):

Bye.