Investing Billions - E164: How I Built a Billion Dollar Company by Failing Fast w/Howard Lerman

Episode Date: May 16, 2025

Howard Lerman has founded six companies and taken one public—he's a relentless builder with an obsession for speed, innovation, and execution. Currently the Founder and CEO of Roam, Howard is reimagining what the modern workplace looks like by building the “Office of the Future,” where AI agents work alongside humans to accelerate output. In this episode, Howard joined me in How I Invest Podcast to share the frameworks and mental models behind how he leads teams, scales product-market fit, recruits S-tier engineers, and instills an almost maniacal level of urgency and creativity across the organization. We talk about what it means to enter "founder mode," why professional managers may soon be obsolete, and how cultural memes shape everything from product design to company mission. If you’re building, investing in, or simply curious about the future of work and AI-native companies—this episode is a must-listen.

We have engineers that have tremendous egos. In fact, some of the best ones have huge egos and they're huge pains in the asses to work with. And the problem about these guys is that they're a huge pain in the ass and they can and they're right. Today, I'm very excited to welcome Howard Lerman, founder and former CEO of Yext, which he took from an idea to a public company on the New York Stock Exchange, during which a traded at over a billion dollar valuation on the first day of trading. Howard shares his unique philosophy on cockroach resilience and startups, how founders can instill speed into their very company DNA, and why figuring out product market fit is the last thing a founder should delegate. Without further ado, here's my conversation with Howard. Howard, you refer to yourself as a cockroach, which is a term coined by YC founder, Paul

Graham. Tell me what it means to be a cockroach context of startups. It means that when you start a company, you have no idea what necessarily is going to take or not. And you try to make something, you come up with a hypothesis for a problem you want to solve and you make it and it may or may not work, it may or may not be the right problem, you may or may not have the right solution. A cockroach is the indefeatable sort of pest that over many, many millennia is still with

us today. And so I think a lot of the surviving entrepreneurs are a little bit like cockroaches in that they are able to withstand external conditions and evolve in whatever market they see fit. And so some specific examples of that are just like, pretty much when you come up with something being flexible on the details

while being stubborn on the vision and being able to sort of change whatever it is you want to change and evolve over time to make sure that you're not killed. Tell me in the context of Yext, how you were able to apply this cockroachiness, for lack of a better term, to building that business. This was in 2006, by the way. So you got to go back 19 years. We were going to create a lead gen site for the health club industry that was modeled after hotels.com. A little bit like ClassPass, except way before that. The very first thing I did was I realized this was going to be a two-sided marketplace.

We were going to need to have a bunch of gyms sign up. So I emailed the founder of the Trade Association for Health Clubs, and his name is Rick Carrow. He was like in his 60s at the time. And he agreed to join us and make a lot of introductions to the New York fitnesses and the equinoxes and the crunch fitnesses of the world to help us build up the network. And then we were going to drive traffic to it and then people would book a free pass and get paid.

And it turned out that this was not a bad business, but it was not an awesome business. And we got to about 2 million of revenue and then recognized that that was about as big as it could possibly be. It turned out that we realized we could apply the same idea to a different vertical. And then we took that and grew it to 20 million. And then we ended up coming up with a whole different concept later on. As long as you have infinite creativity and are willing to chase new things that are grounded in reality, then there's no problem continuing to evolve over time and surviving over the long run, which is I guess the origin of the cockroach metaphor.

If you take a completely first principles look at it, a startup is a C-corp that bundles together humans and some social relational capital and an actual capital going after a certain goal. So you could actually evolve that structure as the company evolves and as the market evolves. You can evolve a lot of different things. And if you can install the willingness to evolve in the DNA of the company and the genes of the company, that's when the magic happens.

By the way, if you look at every great company over time, they never stand still. In fact, they typically have a founding CEO or somebody like that pushing the envelope. Obviously, the famous ones like Apple and Microsoft and Google and Facebook. Look at how Mark Zuckerberg is pushing meta and AI and all these different things. At the next rung down, you have Shopify, for example. Look at, I mean, Toby is personally coding constantly and pushing things forward. He is on the bleeding edge himself. He just announced that Shopify has become an AI first company.

This is what it takes to sustainably build a company, not just for today and for something that works now, but for something that's going to work in the future going forward as well. Are these founders actually instilling something in the very fabric of the company, or is it more a lead from the front? There are both flavors that exist.

My bet is that when you look at the kind of people that Shopify has, we just were talking about them for a second, they probably attract a lot of engineers given Toby being an engineer and still in the code all day and that the AI first kind of stuff really speaks to them. But we just retweeted a tweet that he put up the other day about, you know, trust people who make stuff. And I think we're probably going to be looking at a lot of people that make stuff as the people that we're going to be listening to and inspired by going forward. It's like the Elon, what have you produced this week as applied to every single business?

I think he says, what do you get done this week? But you just evolved it to say, what did you produce this week, which is a really important tweak, right? Because going from what did you get done, I mean, one of the things you can get done is scheduling a lot of meetings and having a lot of meetings and doing a lot of whatever, that might be things you think you got done, but it didn't produce anything. And so by focusing on production, that's clearly how companies are going to maximize output.

SEO of Roam, your new startup. What tools do you have at your disposal that makes your employees want to evolve the business alongside you? One of the cool things about running Roam is that we all live in our product every day, all day. We're the ultimate connoisseurs of dog food as a consequence of this. I and my team are continuously sampling lots and lots of dog food.

I like to lead by setting a really specific vision for the office of the future. And we say it's people in AI's working side by side to enable AI first companies. And then the tools you have at your disposal, and again, these are not technical tools. These are techniques or Jedi tricks. I, for example, give a founder mode update every night, literally every night. It usually happens at about 11 p.m. because that's when I look at the final run

of what's come through Stripe for the day. And around 11 p.m., I'll review the metrics of how many inbound new trials we got, how many demos were booked today. The most important job of a CEO, in addition to setting the vision and having the correct vision and making the right decisions is then to make things go fast. And so I try to make sure that the company is operating at warp speed all the time.

And back to my dog food analogy here, one of the neat things about Rome is that the office is always on. An office of the future is 24-7. AIs don't sleep, and neither will a lot of engineers that are working to ship code. I go into our virtual office at 9 p.m., 10 p.m. There are people there at 11 p.m. I'm able to engage people. My job is to make sure that I am never what is holding things up, and then also that I am pushing people to go as fast as possible because there's an absolutely new standard of speed that we're

witnessing kind of break out here in this new era of AI transformation. There's a renaissance happening that is going to leave a lot of companies in the dust if they don't, you know, kind of shape up pretty soon. If tax season felt harder than it should have, it's time for a change. Juniper Square offers a smarter way to handle fund administration. With industry leading solutions and team of in-house experts who understand venture capital,

we simplify everything from LP reporting to audit prep. No fund too complex, no better time to switch. Visit junipersquare.com slash VC to connect with our team. So you Howard,, are a founder, CEO. You have this, this weight on you when you talk to any employee of this authority and this power. How do you keep from going to authoritative and to founder mode and take away people's implicit motivation and talk to me about how you build an

organization where people are both empowered, but you're also in founder mode. I actually think that founders tend to over-rotate on the non-authoritarian side because all the advice out there from quote smart experience people on their board or advisors is to, you know, bring in the experts and let them do their job. And it turns out that that is a absolutely terrible advice. And founders should try to control the details for as long as they possibly can. And I actually have a stack rank order by which I think founders should give up details

by kind of function. I think the first thing they should give up should be legal. You have a lawyer. Okay, you don't need to understand case law to be able to run your business. HR, you don't need to understand the HR laws in the Netherlands. You can let someone else think about those things. And then you kind of get down the chain to pay an executive who you are. And the very last thing, the absolute last thing a founder should ever delegate is product market fit.

The single most important determinant of the outcome of a company is how strong their product market fit is. That is, how well the product fit that they've shaped and continue to shape because it's an evolving process fits in the market and is being received in the market. Frankly, some companies have such strong product market fit that execution can't screw it up, that they succeed despite poor execution. Twitter has always been an example of that. They were not a greatly run company for a very long time, and they just continued to succeed because people loved it.

If you look at Google, Google was obviously very well run for a long time under Eric Schmidt, but they frankly could have put anybody in there, and I think the thing still would have exploded. Probably the case is true with a lot of companies. So a founder's job is to latch on to product market fit and make sure that whatever the company is producing and putting out into the world has a very snug fit with a group, a market that loves it. And all the you know, sales and all that stuff matters. But you can give that up, you know, first, don't give up the product market fit with regard to controlling the details.

I don't actually produce anything, I guess. So that gives me an advantage in not being overly tyrannical. Like I don't code and I don't design. Everything I do is a collaboration. Um, but I'm the orchestrator of putting all the pieces together in a, in a symphony that sounds good. And as the composer, it's my job to make sure that the violin plays well with the viola, playing well with the bass, playing well with the

wood instruments, the whole thing, and that it's conducted properly. Foundermode is actually maybe a little bit a response to over rotation of companies delegating to professional managers. Professional managers can't do anything. In fact, that might not even be a job at some point going forward. Professional managers, you know, it's like the famous example of course is like what happens when Apple brings in John Scully.

I just read Alex Karp's book, The Technical Republic, Technological Republic. Many of our founding fathers who by the way were super young like 20, 21, 22, they were scientists. Organizations tend to promote people that are like more on the lawyer side than the science side and who are capable of putting together and figuring out the secret statuses within meetings. And you know, there's this whole elaborate bureaucracy.

You have a top engineer, this mythical 10X engineer that comes to Rome. And part of what drives really top engineers is their autonomy, their ability to produce what they see as the right approach. How do you balance that with founder mode? And what are some best practices? You call them S tier. Like if you check out the Rome website, you'll see we hire S tier people. 10X, 100X, whatever. S tier in Japan, by the way, they have a grading system. And you can get an A, a B, and many American students are surprised to learn that in very

rare circumstances, a Japanese teacher for something exceptional may award an S. S stands for shoe, which means like extremely good in Japanese basically. And so this kind of S tier term took on a meaning within video games, like S tier is like the best player in the game. And actually it turns out that in the Gen Z folks, a lot of folks use S tier to describe people who are, you know, the best in the game at any sort of domain that they're playing in. We want to appeal to those kind of people. So we built up, we basically require at Roam to get hired or to have an interview that you submit your S tier achievements to us.

And we look for S tier achievements like being a national merit scholar, winning a computer science competition, winning a math competition, placing in a math competition, these sorts of things. My goal is to not be the smartest person in the room. In fact, my goal is to be the dumbest person in the room. I surround myself with these incredible S-tier individuals who are way more capable than me, particularly with their respective domain. And so again, I'm trying to kind of put the whole thing together.

So if we bring in an engineer, engineer doesn't just get to make whatever they want. Like they don't, we're building an office of the future, like powered by AI. Functionally, there's, there's a clear thing that kind of you have to do. How you implement it, the specific architecture that you put in to, you know, achieve the result, like how we implement that. There's a lot of kind of kicking around. You mentioned the S or the S tier. If I was head of department of education, I would actually institute an A plus

and A plus plus grading system in order to codify within the schooling system. This reward for excellence. When you can't get anything more than A, you default to doing just enough. We don't reward the over-the-top exceptionalism quite enough, who by the way, if you look at the history of the world, are the people that tend to be moving civilization forward. It's the small number of few people who come up with the breakthroughs and the innovations that benefit everyone, yet we're very focused on cutting you know, cutting you off at AA or whatever. You know, why would we do that?

You've managed quite a few of these 10X or 100X engineers at Yext now at Rome. Is part of what makes them so great, their lack of ego, is that essentially one of the features of a truly elite engineer? Thank you for listening. To join our community and to make sure you do not miss any future episodes, please click the follow button above to subscribe. I think that's not really related. We have engineers that have tremendous egos. In fact, some of the best ones have huge egos and they're huge pains in the asses to work with. And

the problem about these guys is that they're a huge pain in the ass and they can and they're right. They're correct. And so I don't like I'm okay with that. Like I'm okay looking stupid. I'm okay looking dumb. You know that by the way, like, you know, when I talk to young younger people that have joined our company that are like, you know, just graduating college and they school me on like a demo and crush it where I couldn't have done it. I'm sort of looking at these people like, holy crap, like I have nothing to learn, nothing to teach you. They get everything to learn from you. You know things I don't know. Ego is an independent dimension to talent. And part of being a leader is knowing how to kind of put the right people together and let people cool off and heat up. But I have no issue whatsoever dealing with someone who has a huge ego when their talent is also huge.

You mentioned these nightly 9pm founder updates and the sense of urgency. Is that something that you could instill in a person or organization, or is that something that you have to hire for as an intrinsic trait of a top performer? We look for people at Rome who are obsessed, who are able to get obsessed, who have a crazy streak and who want to be on a journey and that when they wake up in the morning, the first thing they think about is the journey they're on and then the last thing they think about that night is the journey they're on. That is the level of obsession.

It's tough to teach that. So you have to kind of look for that. You can certainly nurture it, but if they don't have that right out of the gate, that crazy streak, it's tougher to find the right people. So we look for people who have a crazy streak that, and I think the right word is obsession. And that obsession, that crazy streak, that's mostly an individual trait or is that they love the mission so much they go into obsessive? It can be either.

I think you can have just crazy ability in general. The thing that tends to bring it out of people more than the mission itself is the people they're surrounded by. And if you're born with obsessive crazy, if you're born with the ability to get obsessive about something, and I think most people are by the way, depending on what it is, if you're born with the ability to get obsessive about something, it doesn't mean that you will get obsessive at work. It has to be the correct conditions for it to come out of you.

So we try to find the people who are obsessive and then create the circumstances by which they will become obsessed. The dirty trick here is that it's really not about the content of the company so much as the people. And the people around you are the ones that you never want to let down. And when you get on a journey and you feel like you're at war together, that is the kind of mentality that brings folks together in a super awesome way. The culture itself is a product. It's a product that's consumed by the entire company. How do you know that someone might have that obsession in

them through the interview process and how you recruit? It's hard. You can kind of tell from how they talk about things they have done and the level of detail at which they talk and then how fast they talk about it or if their eyes begin to light up. I like to talk to people that have, by the way, it doesn't have to be like, oh, I was working at Facebook and it was so exciting because XYZ. In fact, if you're working for Facebook, you may not be obsessive at all unless you were there in your very early stages. It's almost a disqualifier right out of the gate unless you're like a senior researcher in a specific area.

The thing that I like to see are people who have been obsessed with something in the past and have a bit of an involuntary response to you asking them about it and then they can just keep talking about it and talking about it and talking about it. And they can talk about it at different levels of detail. That's the other key thing. They can talk about the big picture of kind of the idea of it but then also get into the weeds, get into the very specific things that show that they were really, really, really deep diving into this subject or project or dance competition that they were trying to

win. And by the way, it can be getting obsessed with a dance competition or it could be getting obsessed with playing an instrument. It could be any of those things. They're all S tier types of achievements that we look for. And you have to be a little bit crazy to invest that extra time into winning something. It reminds me of a book, The Hypomanic Edge, written by Dr. Gartner about two decades ago, which delves into this advantage that some entrepreneurs have

where they get really elevated, really excited, almost manic, not clinically, but slightly in order to become an entrepreneur. So there's this like edge case of people almost being manic and wanting to drive a vision forward. I'm certainly not a psychiatrist. So I can't speak to what actually is or is not clinically manic.

But I would not be surprised if I would be categorized as occasionally manic. So taking a step back, tell me about your high school experience and how that formed you as an entrepreneur. I was very lucky in that I was not only born in America, I was born in northern Virginia, where there was an exceptional high school and it still is called Thomas Jefferson High School for Science and Technology. And, you know, growing up, I was sort of

a weird kid, I, I played baseball, I sang opera at the same time. And, you know, you can imagine what kids would say when your mom kids at baseball practice would say when your mom would, you know, pick you up to take you to opera practice. It was not the coolest thing possible as a nine-year-old boy. But I think it really taught me to deal with being made fun of, being teased, being rejected. I'm very good at that kind of thing because I just really don't care. And I grew thick skin from being on stage. As a child being thrown into the opera, singing opera on the stage at the Kennedy Center at 9, 10, 11 years old in front of 2,000 fancy people every night because I was not a fancy person. I mean, lower middle class, do on stage singing opera in front of fancy people.

You get to, you know, you get comfortable on stage pretty fast. And so that was a really formulative thing for me. And then I was lucky enough to get into TJ, a STEM high school where, you know, the only way you get in is if you get a good score on the math test. You can't charm your way in, you can't essay your way in, there's no political influence, you can't buy your way in, you can't give five million dollars and get in. It's just you can take the test and then that's how it's going to be and they take the top 400 people. And going there, I met other people that were like me, that were a little different.

me that were that were like a little different. And in particular, had like a love of science and math and technology and STEM. So this is where you know, the school of supercomputer. At the time, this was like 1995. Like, so I met some people there. And that Tom Dixon and Sean McKay, who I've been starting companies with for the last, now, I guess, 20 plus years, on their, you know, at Yext and at Rome, I was very with for the last, now I guess 20 plus years. They were at Yext and at Rome. I was very lucky to be able to meet folks in high school

that were like-minded and I was kind of able to just get started early. You've told me about your superpower of being able to look stupid. How did you cultivate the superpower? I'm reading a book right now called improv and it's written. It's a 1972 book written by a famous London improv troop leader. This is like folks on stage.

And you'll hear me coming back to this idea of a stage, because I think that's a recurring theme in my life. My favorite Shakespeare quote is, all the world is a stage. And if you think about it, like, you know, I'm on your stage right now. When I'm at Rome, it turns out one of the room types we have is a theater, because every company has

all hands presentations. So we have this gorgeous theater with all these like little details, like stereo effects with clapping and booing and hissing and all these different types of things that happen only in a real theater that we whispering to people next to you that we've tried to recreate in a virtual environment in an AI type of theater.

And being on stage, like, you can't hide anything. You just kind of get used to having a hard shell about looking dumb. I've been really interested in this concept of mimetics, mimetic copying, how ideas are transferred within society. I think Elon has really mastered this. He went past using words and he realized that memes are harder to block out of the subconscious and they're easier to share and lower risk to share and all these things that make these kind of memes and mimetic behavior really powerful. That just really gets to David Deutsch's book, right?

The Beginning of Infinity. All species are capable of transmitting information one way, which is through their genes. But we as human beings and thinking machines have a different way that we can transport information amongst each other, which is through memes. And memes are empirical. They have to be self-experienced and all knowledge is empirical, has to be self-taught. But the idea that you can kind of create higher chunks of ideas I think is really important.

That, you know, there's so many things, going back to like the kids of today, if you will, you know, this idea of these Gen Z younger people that I'm myself I'm learning from instead of the other way around, they're, what's happened learning from instead of the other way around, what's happened is so many of the concepts that I had to learn over time have now been memified. So what that has done is it has increased the speed at which people can get really good

at something, kind of just with higher order chunks. Specific example of that is this idea of product market fit. You and I both know what that means now. But when I was 20 and I started my first company, the concept did not exist. The concept may exist, but the meme did not. And so you had to kind of find your way to product market fit or growth marketing or all these things that now you can write a textbook about, someone can read it and start from a jump, you know, begin from a jumpstart of these higher order ideas, which, you know, are transmitted through memes in our brains. And what's funny about memes is the way that they evolve too, just like genes

evolve accidentally and through mutation and the better ones are selected, the same exact thing happens for memes. I'll give you a very specific example of Rome. When we started Rome, I branded it as the office of tomorrow. And I'd tell people office of tomorrow and everyone would come back to me or I'd hear them saying someone else, oh, that's Howard, he's making the office of the future. And so it was the same idea of the meme, but I was witnessing it live, being modified. And the surviving version was not the office of tomorrow, it was the office of the future. Recognizing that this was happening, I actually changed it to the office of the future, because

I knew that that was a more powerful meme, and that it would actually drown out the office of tomorrow. And so that's an example of, you know, you can almost think of a gene, a better version, a more adapted version, a set of genes, basically beating a crap near version because it just landed with people. The reason why I think a lot of businesses compound and improve is that founders and people at the company are able to kind of spot the evolution of certain things and incorporate into the business. How you phrase something, you kind of just observe

how people are using your product. And you're like, what did you do? You like put this and you emailed it to your grandma and your grandma went on her iPad and clicked a button. Like let's codify that and let's turn that into a product. And I think that's why there's such a value to being able to, the duration of your business and being able to survive because the longer you survive, the more these things start to

compound. It's a really great example. And I'll give you an example at Rome. I mean, right now we're witnessing this mullet in AI and you could say, oh, it started a couple of years ago. The fact is that the AI technology that's out there is now good enough to do a lot of the things that people have been talking about doing for decades and was never before possible. And we sit at this unique point in time where the things that are possible and the adoption

has not caught up yet. And that gap is the opportunity for many entrepreneurs right now to seize the moment and be able to bring these new capabilities and superpowers to businesses around the world. Let's take businesses out there, tens of millions of businesses out there that are either gonna basically become AI first or be disrupted by a company that is AI first.

That's the only choice. We've been lucky with timing, but been able to build in so many incredible AI features into our platform in the office of the future. Starting with Magic Minutes or AI note taking, building agentic workflows off of the AI note taker so that after a meeting is over, you can see what are the action items and then also give some of those action items to AI agents. So, you know, I think a professional manager may not have jumped on these opportunities in the same way or been able to evolve things as fast. You mentioned agentic workflows, AI agents, oftentimes talking to AI agents. This is a little bit difficult to graphs.

Walk me through what AI agents are and how they're being used early on in the early use cases and how you see that evolving. There are different definitions of agents. Some people might say agents, a single entity. Other people are spinning up pools of agents to do things at the same time and then wrapping that up into a higher entity called an AI worker. I would just say that all these definitions are kind of in flux right now and are going to evolve over time.

But what is clear is that there are specific actionable things that drive incredible amounts of productivity and speed and cost reduction and accuracy that has never before been seen. I'll give you a couple of really specific examples. Say that an agent is capable of examining the text of a meeting and deciding who's going to do what after the meeting and assigning those things to you. So David is going to follow up with Sally and Sally is going to write the report on this feature and Joe is going to file the JIRA ticket. Some of those things now can actually

be handled by an agent. So an agent can look at, oh, Sally was supposed to schedule a meeting with Frank. I have a skill where I am able to examine calendars and schedule meetings. And so an agent can jump in and do that. An agent can jump in and answer, follow up with people. It can make sure David sends the creative assets to Sean so that the correct thing can happen. We've already seen success with agents in customer support, answering customer questions based off of

knowledge feedback. These are these are the kinds of things that like are not there. They're now possible and they're all happening in the real world. Yet most companies have not adopted them yet. And that's why we've seen in the past few weeks, a number of companies starting with Shopify, then just last week, I saw a box. Then I just also saw Spotify. And then I also saw Duolingo announced that they are now AI first companies, which means that, you know, and each of them has slightly different interpretation, but they're basically not like, they say managers are not allowed to hire people until they check to see if an AI can do the job first. Howard, I only got to about half my questions.. We're going to have to do this again soon.

Where could people follow you, and where could people keep up with Roam? I'm pretty active on X. I like to post what I'm reading. So I'm at Howard on X. I was very lucky to have gotten that as an early user, at Howard on X. And then Roam is also at Roam. And our website is RO.AM. You can check out the office of the future there

and step in for by the way, a live tour. You can literally just come in right now. And that is because we have all the presence and the live feeling of being in a real office together with AI co-workers. Well, thank you Howard for jumping on. Look forward to seeing you down soon. Thanks for listening to my conversation with Howard. If you enjoyed this conversation, in a real office together with AI co-workers.