Building an operations-heavy business and avoiding upside down unit economics  — Thirty Madison’s Steve Gutentag

 One of the biggest traps that people get themselves into is that the unit economics will fix themselves at scale. And one of the big reasons to your point is technology and scale will help improve margin, but it doesn't magically make economics just work. And I think it's really important that when you're dealing with an operational business and one where there's either a physical product or a physical service that the unit economics need to work at the early stages.

 Welcome to in depth, a new show that surfaces tactical advice, founders and startup leaders need to grow their teams, their companies, and themselves. I'm Brett Berson, a partner at first round, and we're a venture capital firm that helps startups like notion, roadblocks, Uber, and square tackle company building firsts through over 400 interviews on the review.

We've shared standout company building. The kind that comes from those willing to skip the talking points and go deeper into not just what to do, but how to do it with our new podcast in depth, you can listen into these deeper conversations every single week. Learn more and subscribe today@firstround.com

today's episode of in-depth. I'm thrilled to be joined by Steve Guten tag. Steve is the co-founder and CEO of Thirty Madison, a healthcare company focused on a widening access to specialized care for chronic conditions. He previously started two other companies with his Thirty Madison co-founder Dimitri CarGurus, including a content recommendation engine that didn't work out and an on-demand cleaning platform that was acquired by home joy in 2014.

After spending time at Google, Steve and Dimitri left to start Thirty Madison in 2017, launching with keeps a men's hair loss solution. I always admire teams who are just heads down aggressively executing, even if that means flying a bit under the radar. And I would definitely put this founding duo in that.

As seed supporters of Thirty Madison here at First Round, we've gotten to watch this operational precision up close. They've gone on to launch several new brands, including Cove, which focuses on migraines, evens for GI issues and picnic for allergies and the move towards telemedicine during the pandemic has really accelerated their business.

They've tripled both their revenue and their team size in the past. And back in June, they announced a $140 million series C funding round pushing the company's value past the billion dollar mark. All of this means that Steve has tons of lessons to share with founders and operations leaders. We start our conversation by getting into the challenges of building an operationally complex business, especially one with a physical or real world component.

Steve shares the lessons he learned from building his first two startups and talks about how founders can figure out if they're the right match for the product. They're trying to build something I found really interesting is how Steve said they wanted to pick a business where the unit economics worked from day one.

He then walks us through their methodical approach to validating the core idea for 30 minutes. He explains how the approached their core business model and assumptions. And he also shared a lot of unlocks that came from their more inventive moves, like their asynchronous approach to telemedicine. Steve shares a ton of pointers throughout, including why you should think of your own internal operations as a marketplace.

And a warning about how unit economics won't magically fix themselves at scale. In the last part of our conversation, we get into building the team. That's pulling this all together. We talked about when to hire for industry experience versus fresh perspective. And Steve walked us through two examples of senior leaders.

They've hired recently to illustrate these points. He also shared more granular hiring tactics, such as the interview questions. He asked to learn about a candidates journey as a manager. I really hope you enjoyed this episode. And now my conversation with Steve. Well, Steve, thank you so much for joining.

 So I thought an interesting place to start might be some of the lessons that you've learned in building a fairly operationally complex business. And I think that unlike most predominantly software businesses, whenever you're dealing with things in the real world, it's sort of like the double black diamond of startup building or entrepreneurship.

And so maybe at a high level, it might be interesting just to start with some of the ideas. That drive the way you think about building a startup from zero to one. And there may be one to 10 that is much more operationally complex than maybe just an app or a piece of enterprise software. I 

 actually think it might be helpful to give a sense for how.

Both myself. And my co-founder figured out that this is where we are good at, and it actually starts with the first idea and company that we had worked on was not at all involved in the real world. It was a content recommendation engine back in 2009, 2010, and it was designed to help people find. Content that they might be interested in based on their browsing patterns.

And we launched it as a Chrome extension at a time that the well was starting to go mobile. And so we were very much, we were both business and operations guys and very much the wrong people for that idea had, you know, the wrong team, the wrong idea, the wrong approach, as you can imagine, did not work during that time though, we had raised a lot of angel capital at the time and had.

Been working on it and had a bit of it left and step back, and really started to do some introspection and evaluation of what are the things that are really interesting to us and the intersection of that with what are the things that we think we're good at and what we realized and came to was that.

Things that are operationally complex, where you can find efficiencies, you can find faster or better ways to complete a task or drive and experience and ones that really connected the online world and the offline world where, where we were both most interested and where we. Had quite a bit of skill that we started to hone and develop over that time.

And over the last 10 years does the type of businesses that we've operated in. We switched from that content recommendation engine. I don't say pivoted because there was a full on switch to an on-demand cleaning business where we would let people. Cleaning services online and have someone show up either scheduled or within two hours of clicking a button and we're able to build and grow that business and eventually sell to a larger venture backed business in San Francisco, which where we got to really hone and learn from folks doing this at a larger scale, I was a company called home joy that we ended up selling to, and that company ran into quite a bit of challenges in the end and went over to Google.

Where are we? Also ended up working on ad products that combined online and offline services. And so we really honed our chops over the last 10 years. How do you create enjoyable online experiences that connect to offline world and for us, you know, as it came to, you know, circling back to your original question around, how do we think about the type of companies that we want to build and operationally complex areas and what are some of the things that are important to consider?

One was. When we started looking at what we want to do next, when we were starting Thirty Madison really important, are we the right people to do this? And do we have the right set of skills? And we really define those skills around really enjoyable consumer experiences that typically start online and typically end offline.

And that's something that. Operates very differently than a business that is entirely online. And so when we looked at businesses and problems, we looked for problems like that too. I think the more operationally complex actually became more interesting to us because that was where we could really spend and put our brains to work was trying to figure out really, really operationally complex problems.

And I don't think there is an industry with more operationally complex problems than healthcare. 

 You mentioned one of these inflection points that you've had thus far in your career was in 2010 or 11, when you realize that what you're uniquely great at was not mapping to the project and product that you were working on.

And so I was interested in going a little bit deeper about what that process of self discovery was, how it all came to. Great question. 

 I would like to say that I had my own epiphanies and there was a moment of clarity and it all happened really quickly. It wasn't that it was a series of events. I do think in hindsight, though, there are moments that become much clearer.

One was we were at the early stages of building our business and our ability as founders to effectuate the growth of the business to really drive what was going to make this whole. Didn't match. So this is about machine learning. There was a not a revenue generating business. And so the way that we thought about growth looked very different and we didn't have the skills, particularly on that side.

So we weren't able to build a product and we're able to grow the product. And we weren't the right people to do that. It led to, we tried many different things and it wasn't working. And one of our. And investors really at the time, hammered that home to us that this wasn't kind of work for a few reasons, but most notably we were the wrong match for the wrong product.

Something that this investor said it was when you're looking to build something like this, it's not supposed to be that hard. And not that it's not hard work to build a business or get product market fit, but when you're. Continuing to pound against the wall over and over for largely similar reasons, both market and team.

You're better use spending your time in an area where you're going to have more likelihood of success given your team and what you're good at. And it really started a process for us to really think through and step back. Okay, well, what are the things that we are good at? What are the things that we have learned and taken away from this process?

We sat down and we were a very small team at the time. I was like three, four a POS, I believe. And we started to map out not only ideas, but requirements for, if we were going to all agree to like, continue to give this one more shot. What are the things that we are good at? And what are the things that are exciting to us?

And what did we require? And one piece of it was. That it had to be a business that every one of us could contribute to. And so given that we had some product experience, but we're not engineers, they really came down to we're good at finding inefficient things and trying to make them more efficient. I think a really silly example of this personally is that, you know, I'm in New York city, I ride the subway is.

Walking to the exact spot to get on the trains, that the minute that I get off the train, I'm at the right exit from the station. And I know there are applications that do this, but like I would do this naturally in my head and try to find ways to make things more efficient. And that problem tends to manifest itself in a lot of industries and technology can certainly help, but the area where we can oftentimes really support it is through operations.

 continuing down this road, I'm interested. Do you have any higher level thoughts or frameworks on what needs to go, right? Maybe in the few different chapters of a business that are specifically tied to these kinds of operationally intensive real-world businesses. What I've noticed is it's exceptionally hard to build a venture backed business broadly, and I think it's even harder when there's these physical components and the number one fail case.

It's basically the business is turned upside down, like the fundamental economics, the machine doesn't work or the core product doesn't work, there's product quality. There's all sorts of issues that generally happen. And so I'm curious what you've learned maybe about first, what doesn't work and what are the most common traps in building businesses with bits and atoms.

And then maybe we can kind of transition, talk a little bit about how you've approached building Thirty Madison in light of those ideas. 

 Love that question. One of the biggest traps that people get themselves into is that unit economics will fix themselves at scale, but by definition, unit economics are units in a physical world.

And so looking at a lot of the marketplace businesses from the last five or 10 years, there's a few that have survived and been successful, but there's so many. I have died along the way. And one of the big reasons to your point is that folks were betting that technology would totally make the unit economics work at scale or just scale would make the unit economics finally work.

And I think it's really important that when you're dealing with an operational business and one where there's either a physical product or a physical service happening that the unit economics. Need to work at the early stages and that's not just, okay, does the revenue make sense for the costs, but also how do you think about whether, if it's a subscription business, lifetime value and how you translate that to how much you're spending?

And so one thing that we would do early on was be really conservative with our STD. Until we got really comfortable with the data on what we were allowed to spend on acquisition, to acquire a customer before we would relax it and, or expand it. And until we really understand what the retention would look like on a subscription business.

So even in an economic. Magically fixed themselves at scale, I think was one really important lesson that we learned the hard way. It was a challenge at home joy. It was a challenge in my first cleaning business, but really making sure that you understand what the unit economics are, understand where technology and scale will help improve margin, but it doesn't magically make economics.

The second then becomes a game of balancing growth and stability, and then capitalization and being mindful that you have the right amount of capital to continue to push the business forward in a responsible manner so that you don't end up in a point where you got economics right early and then threw so much capital on it that the economics floods.

And not just the revenue minus Cox, but down to the line item, go one by one and make sure that you understand where you can get efficiencies at scale, where you're not going to, and really be mindful of that beyond the economic piece. When you go to your point around the service or product. How do you think about making sure that when you scale that service or product, that it's hard, but it's not that hard to get a few great experiences, you know, when you're dealing with numbers in the tens or hundreds, you really need to get that right.

You need to have the amazing experience. That's a starting point, but where it gets really, really hard. When you have to try to deliver that same high quality product or that same high quality experience, no longer to 10 20, 50, a hundred people, but to 10,000 hundred thousand million, however many people you get to in your scale.

And so really making sure that you have the. Control. So if it's a physical product that your supply chain is an area of risk, a lot of companies get those wrong. They get really good prototypes in there. Maybe their demand outstrips their ability to service that demand the product that you created for the first 1000 people.

To the first point around economics is too expensive. And so you have to completely change what you've built to make it more affordable or to get to the volume that customers might want. Really having a strong understanding of, as you think about scale, if it's a service or if it's a product that you have the right controls and that you're balancing and in some ways.

It's like a marketplace. If it's a service, it's certainly like a marketplace, but even on a physical product, you gotta make think of the marketplace as your own internal operations on one side and the customer demand on the other side, making sure that you continue to have the right investments on both sides.

I'll give you an example with Thirty Madison. When we first launched, we operate in a highly regulated industry. And so for us, finding. Really high quality suppliers who would work with us on start-up volumes was quite difficult and it took a long time and we finally found them and were able to get to market.

But then as we started to scale, we would have to then find a totally new supplier or a totally remake the initial packaging, and really try to balance. What got people so excited about our offering in terms of the aesthetics of our product with finding new suppliers who could meet the volume of our scale.

And we didn't always get that. Right. And there were times in our history where we had a stockout, which in a lot of industries can be really good, but in the healthcare industry, you never want to stop. 

 It would be interesting to have you share a little bit more about this set of ideas and bring it to life in your last business around housecleaning.

And what I find interesting about it is like, if you were to explain to someone, Hey, you know, we're creating this product, do you go, you schedule a housecleaner, they show up, they clean your house and that's that it seems intuitively like it would be quite simple. You could have. And you fill out some information and somebody cleans your house and generally a reasonably high cart size to have that person come over and there's enough margin for everyone.

But I think it's one of those examples when reality is much more complicated than we would imagine. And it actually maybe illustrates a bunch of the ideas that you were just walking 

 through. Absolutely. I think you're right in theory, you say, okay, we're going to build something on one end to let people request a house cleaning.

It's a reasonably sized purchase price. And housecleaning tends to be something that is recurring for many people. And so you think, oh wow, this is great. Find someone and they're going to use you every week or every month or biweekly, whatever their cadence may be. And you'll find people to go and show up, clean the house and be done.

I wish it was that easy. You run into a few things that you have to think through. So from a economics perspective, as I was talking about, you have whatever you're collecting from a customer last, the cost of labor and supply. And potentially travel or any other pieces credit card processing. We start to layer in all those things, but there's a lot that can go wrong.

We operate in New York city. So one is you have to manage quality of your network and finding people who are really good. Cleaners who are really consistent and are good with customers is quite difficult and you can start lead, find a handful. But when you think about venture scale and you go from having to my point earlier, five or 10 cleaners to trying to manage a network of a hundred or 500 or a thousand.

Consistently, it becomes really hard. So you have one issue, which is how do you maintain a really high bar of quality that customers are gonna love. And when you've got a five to 10, you have a personal relationship with these people, you can be more closely monitoring and doing quality assurance. When you get to a hundred or a thousand, you start to lose that personal relationship with people you can't possibly manage.

And do QA as effectively, unless you start to then staff up specifically support services to do QA, and then that's another cost of service that you start to add in. And so you actually don't necessarily get the efficiencies there. And so some things that we did to help us scale and find more efficiencies were one.

We actually invested a lot in the provider, like the cleaner. App in fact, a lot of times more so than the customer one, and that we did things to help them take notes on a client's preferences to help them mark down details about someone's homes. Did they have the right supplies? Were there stairs? Was there an elevator to make it easier for them to write down if they add names of kids or pets so that it was more personal for a customer to really try to.

Find ways to scale the quality as we have more and more people. And we couldn't build that relationship and do a one-on-one. And so software helped us there. You then have a second problem, which is because there are still personal preferences in that type of business. And everyone's had experiences with a bad cleaning service that once you get someone you really like.

You wanna stay with that person? You want them to come back. And so when you're small and you have this personal relationship with your providers and they feel really invested, you can work around that when you start to get larger and there's less connection from providers or whoever it may be to the business, you start to run into issues of disintermediation.

The value really it's that introduction you're providing, helping them find someone. And once someone find someone they want, they really want to continue to work with them. And the value proposition as the company that you're making that margin, you're charging more than the cleaner. And it's just the same one over and over again.

And you already built that relationship. And you start to see one disintermediation, cause there's not as much value there. And so the economics start to change in that. You're still spending money to find customers and clients expecting that they're going to stay with you and you then start to see that lifetime value start to compress more.

And so you then have to spend more to find other new customers or to replace providers. And so then you now are you're increasing your costs on. Providers or you're losing recurring revenue and value on a customer side. And so you actually run into issues on both unit economics and on scaling the service and.

These are issues that are really hard problems because they're not just, okay, well, we can throw software at it. It's human nature. And then you throw in on top where you do add the value as a business is on things like insurance and assurity of quality and that yoke guarantee it. And so as a business, oftentimes costs that you don't think about like free, clean, Or finding a second person because someone's car broke down into the subway, broke down and, or they showed up late and they had to cancel and reschedule.

And you had to give a discount. These things start to add up. And so you have an issue of guests at a really small scale just on the revenue side, but you're spending enough on marketing that you are not necessarily making money on the first job or the second job. And you combine that with. A longer longer-term issue of quality of supply.

Plus once you make a quality introduction, your actual value declines, you end up having a really challenging issue. Really, these are challenges that don't exist in a fully digital world, and can't be fully solved just with technology. And so 

 in the case of house cleaning, do you think that it's the properties of the problem that it's basically impossible to build this style of business?

Or are there certain unlocks if you were to go tackle the problem today, you actually think it is potentially possible to build a good business in that category. 

 I think you have to define what a good businesses in terms of scale. I think that you do see that the businesses that are most successful that have scaled actually remove themselves from the operations and really just focus themselves on the introductions.

So if you look at things like local business ads on the Google website, it's about making the introduction and it's about providing information reviews, quality. And it's about removing Google or from the actual provision or the operations. I think if you look at many of the other ones that have reached a level of scale and really built profitable businesses, it really comes from two things.

One removing yourself from operations or two, it's a much broader basket of offerings, you know, maybe help them with their one time home repair or other types of services where they might be able to cross. Yeah, it's interesting. I was 

 talking to a friend of mine who started a business in the delivery space.

That's at scale. And he was mentioning that if you look at the last 10 years in food tech, it's the companies that have separated the aggregation and delivery function from the food prep that have worked like door dash and Instacart and Uber eats and a handful of others and all the ones that tried to make all the food and deliver all the food are the ones that kind of got crunched.

Mainly I think by this service quality and upside down unit economic problem that you were talking. 

 Absolutely. And I think you see that in a lot of categories and food delivery is such a competitive space. And for many of these businesses not having to manage. All fully integrated vertical, including preparation is why they succeeded because scaling the supply of the food and the delivery of the food, or also by the way, just very different skillsets.

And so it's really hard to, from an operations perspective, do both of those at the same time while you're trying to do. 

 Flipping gears. And going back to something you shared a little while ago, which is basically the idea when you're building an operationally intense or CapEx, potentially intensive business, you really have to be intellectually honest about where you have increasing economies of scale, maybe decreasing economies of scale or things that are cost neutral as you scale.

And I'd agree with the idea that a lot of founders basically trick themselves into thinking that all their unit economic problems go away when they operate at scale and as just a broad, maybe. The statement that effectively software will provide more leverage or software will magically take a mediocre services business and make it a fantastic software driven business.

And in most cases you end up, at least in our experience. I think getting. Incremental changes when you apply software to some of these traditional business versus transformative changes. And so it might be fun and interesting to talk maybe a little bit about your process of building Thirty Madison when you were getting going.

And you thought about this, you had just had this really difficult experience building your last company at home cleaning and learned a lot of these unit economic lessons, quality and service level. And I assume when you were building the Thirty Madison model and playbook, you were thinking a lot about where can we actually get leverage and what's probably not going to radically change as we scale.

And so it might be interesting to talk a little bit about that. Maybe walk through the model or some of the areas in the assumptions that you tested to give you conviction on where there was actually leveraged as you grew the business, and maybe what was more of a linear cost as you scaled the. 

 I think a lot of people go in with first, the mindset of, I love this idea or to your point, I think that with a little bit of software, will it all fix this or it'll make a bad experience.

Great. And look, you can build really good sustainable law. I think businesses with a strong software component, but for us really, when we thought about Thirty Madison or really any new idea, we would ask ourselves a different question, which is what do we think is the number one reason this business is gonna fail.

And that question is really helpful because it really requires you to take a critical, sober view. First before you fall in love with your vision for taking over the world, that process could be started by listing out. Why do you think this is going to fail? And what do you think is going to go wrong?

And then trying to pick from that list. And then we would ask ourselves, how are we going to solve this? How are we going to overcome this? And, and what's the level of degree of challenge and pain. We're going to go through to overcome this. And we would do that if. Um, until a certain point, you know, you get to like number three or four and you feel like, okay, these are the reasons why that's going to fail, but like they're more obviously overcomeable.

And so then we would flip to, okay, how are we going to make this thing a massive success? How are we going to build something that people are going to love? We would go into that with a better understanding of what are the watch outs and the catches first, as we think about. Thirty Madison I'll use Keeps, which was our first product line that we launched at 30 minutes, our platform.

And the big thing that we were doing was we were connecting this patient journey, finding inefficiencies of. Getting a treatment plan recommendation for men's hair loss combined with actual treatment finalized with helping those guys build habit, adhere to that treatment tools, to track their progress and really doing that in an engaging way.

And so when we look at the areas of costs and the areas where our revenue and we look at, where can we get scale? There are. Physicians who are providing care for patients. And that's an area that you really need to make sure quality remains really, really high. And so you think about, okay, well, how can you manage quality?

And one could be what we're going to do, live conversations one-to-one and we're going to try to scale that, but that means you're going to need so many physicians who are really comfortable treating this. You're going to have. To figure out and scheduling. And so this is a really big, complex operational issue.

And there's companies that have solved that there's large telemedicine companies that are built large businesses on that, but I was just one piece of what we're going to do. And so then we started to look at other ways with which you could approach telemedicine and learned about the asynchronous or a store and forward model, where you could actually collect a ton of information upfront that a physician would then look up.

Asynchronously could ask questions and respond with a recommendation afterwards. And. We realized was well with that model, you can scale. It takes physicians last time. It was not the scheduling challenge. And as we learned more from patients and we worked with physicians, the questions you ask get better and smarter, and then you can use software to help physicians get more efficient in terms of how they review.

Look for things like contra-indications and support physicians as they are evaluating patients. And you can invest in the software to help on both the patient side and the consumer side. And so it's actually, we had to look at different models and figure out how do we actually implement this as we grow.

And as we grow, we realize, okay, well, the number of patients a physician can see it in the given time. Can evolve based on the quality of the software, the quality of the questions, diagnostic frameworks, the clinical guidelines that are being established. And how do you codify that into the experience can actually help provide scale.

So the cost of a visit can go down and can go down pretty dramatically from starting with a very live one-on-one visit to an asynchronous one. As you grow, those two costs are going to scale very well. And we knew that it doesn't go to zero. And so you have to think about, and again, quality of care is critical.

And how do you think about doing this as you scale? There's other costs that go in in terms of making sure that there's consistency and support along the way. But that was one thing. When we looked at the business, we really took a critical eye of, are we saying this up the right way from day one so that we have a baseline.

Steve Gutentag shares the lessons he’s learned building three startups, including how to figure out if there’s founder/product fit and how to build an operationally complex business with unit economics that work from day one.