Why Startups Fail

Module Description:
Harvard Business School Professor Tom Eisenmann shares several distinct patterns of why startups fail. Come away from this event with frameworks for detecting when a venture is vulnerable to these patterns, along with a wealth of strategies and tactics for

Full Transcript:
Welcome everyone to today's special Ubiquity Ventures event on "Why Startups Fail." We have a very special guest, someone I've known for 15 years, maybe 16 years, Tom Eisenmann of Harvard Business School, and we're gonna dive into understanding a lot of the depth behind his research and making this practical and actionable for early stage entrepreneurs. Tom was my teacher when I was getting my MBA at Harvard and an advisor to my startup Triangulate, which we'll get into, I won't ruin the punchline there, but Tom's been a mentor, advisor and just a fantastic thinker on this topic of startups and startup development. So, Tom, please take it away.



Thanks, Sunil. I never thought of myself as being nerdy about startup failure, but there you are. Thanks for the opportunity to talk about this work. It dates back about 10 years, I got started on this project. I was the course head. All 900 1st year Harvard Business School MBAs take a required course, "Introduction to entrepreneurship and its case method." And back in those days, we would march in front of them in a 30 session course, 30 brilliantly successful entrepreneurs who'd come in and strut like peacocks, telling their story. And of course, we solicit feedback from the students at the end of the semester and they would often say, "Hey, you told us along the way that two thirds of startups fail, and you showed us 30 successes out of 30. So is there nothing to learn from failure?" And I dug deeper in, first couple of tries were disasters. I was a failure at teaching failure. And then the deeper I got into it, I realized I was a failure at explaining failure, when one of the teams launched by other former students of mine, in which I'd invested and I thought I had done a textbook-perfect job of following all the lean startup protocols in a perfect minimum viable product and so forth. And yet they failed. And I could point to some reasons why, but I couldn't pinpoint the causes. So, failure at explaining failure put me on a 10 year path to read everything anybody had ever written, practitioner or academic about startup failure. And interviewed dozens and dozens of failed founders and the investors who backed them. And the result is this research, which also is currently a course at Harvard Business School. So, "Why Startups Fail," here you go. So if you get into the book, you'll see the big goal of the book is to help founders understand how to not fail. And if they're going to fail, despite their best efforts to avoid it, how they can fail well in ways that preserve their reputation and causes a little damage, collateral damage as possible. We'll only touch on the second point and on the first point I'm gonna emphasize 'cause I think most of the folks on the call are associated with early stage startups, like those that Ubiquity invests in. I will, the book, devotes about a third of the book to late stage failure patterns. And I'll just touch on those. So, you know what's in there if that's more relevant to the work you do.



So here's a definition of failure that we use in the research and in the book. Early investors, like Sunil, did not or never will make money. And of course that's just one definition. You could say, "Well, what about the founders and their hopes and dreams? If the thing was a financial failure, but they accomplished what they set out to do, built a great product, built a great team, doesn't that count for something?" Of course it does. And the issue is, as we get to late stage, by the time you get to Series D, something like 60% of founder CEOs are no longer CEO, they've been replaced by a professional manager. So if we're talking about startup failure, you gotta go beyond the founder's relationship to the lead role. You know, and again, investor lens is one lens. I will say that there are plenty of financially successful startups that from a societal perspective, we might all wish would disappear. They've coarsened democracy, they've harmed the environment, et cetera, et cetera. So I recognize anything as complicated as failure is gonna be defined in different ways. But that's one lens, the investor lens. And for our purposes here, we wanna focus on a corner of a two by two here, where we think about whether the fundamental cause of the failure was avoidable or not. And secondly, whether a reasonable objective and well-informed observers, like, say Sunil would say, this was a promising venture at the outset as opposed to, and there are unfortunately, a lot of startup ideas out there that fall into the hapless and hopeless cell where, you know, if we could just talk the entrepreneur out of actually spending time and effort, we would... It's that upper right-hand cell where the failure is avoidable. It's really due to mistakes that the founder has made and where the concept was at least initially promising. Interesting cell is the lower right. There's some businesses that are just destroyed like the dinosaurs were by an asteroid, forces completely outta their control. If you think of what happened with COVID, you know, hundreds of thousands of businesses, particularly in travel, or restaurants and so forth, through no fault of the entrepreneur, and likewise down in that cell, there's a category of failure that I categorize as a good failure. In other words, the entrepreneur had a good idea, a hypothesis, ran a test, and the test said, "Nope, hypothesis was wrong," but the only way you're gonna find out is to actually build something and test something. And if you did that in a smart way with little waste, we should categorize that as a good failure. And again, from a societal perspective, celebrate that. We want people to do that all the time. So the focus is the upper right. And I had students in this course I mentioned, actually interview failed founders, and they asked questions about what was the biggest contributor to your venture's failure? And the blue shaded area here, almost two thirds of the responses, the main contributor, these are failed founders admitting were mistakes that the founding team made. As controlled to the asteroid strike, the uncontrollable surprise is only 13%. And then a slightly bigger slice of situations where we had well researched assumptions that simply proved false after doing some testing.



This is the definition of entrepreneurship we use at Harvard Business School, "Pursuing novel opportunity while initially lacking the resources you need to capture that opportunity." And if you think about this definition, it captures a lot. You're doing something new without resources, that is really a prescription for failure. So we should expect entrepreneurs to fail because what you do on this call is just incredibly hard and we love you and we're proud of you and we'll try to reduce the failure, but it's sort of built into the job description here. And if you sort of take that definition and break it into two pieces, I'll get into the early stage failure patterns here. One of them is essentially the entrepreneur, straight outta the box has found the right opportunity, but they haven't mobilized the right resources. So right opportunity, wrong resources. And by resources I mean the founders themselves, the rest of the team, the investors who are gonna put in not only money, but value added time and connections and strategic partners either on the marketing side or operations and technology side. And so even though you've got a good idea, you just never get the capital team together to do it. And then there's the polar opposite of that, which is a very good team, just doesn't find the right opportunity and can't pivot away from the wrong opportunity quickly enough to have the venture survive. We'll see examples of both of those.



So the first one is this example of right opportunity, wrong resources. And these are former students, these are the former students in which I invested, the venture here is Quincy Apparel, Alex Nelson on the left, Christina Wallace on the right. And the concept, they were both tall and they, after business school went to work, like so many MBAs into consulting and found that it was difficult for them to find work apparel, professional work apparel that looked good, stylish, fit well and was affordable. You can usually get two out of the three, but it's hard to get all three. So that's what they set out to do. And the concept was a suiting scheme, a bit like men's apparel, if you're familiar with, men's apparel comes in many more dimensions than women's does, you know, so you just don't have a size 10 or a size 12, we have sleeve lengths and et cetera, et cetera. So that's what they did. And they ran textbook perfect MVP trunk shows, if you're familiar with fashion, where they created samples, had women in their target market try, had a super high buy rate and raised a million dollars of venture capital based on that. Positioned the thing as direct to consumer, which, this is 2013 when they got going, which was a hot concept at the time, lots of VCs wanted one of those in their portfolio and so off to the races and within a year had burned through the capital and failed and shut the business down. And so, you know, this was the situation where I could point to a lot of things that went wrong but couldn't pinpoint the causes until I dug deeper. And this is a situation where poor founder fit cascades right on through the other categories of resource providers. So neither Alex nor Christina had domain experience, experience in the design and manufacturing of apparel, which turns out to be really, really important 'cause it's a complicated thing to design and manufacture apparel. As a result of that, they had to go find people that knew how to do that and they lacked a network, a personal network, professional network rich with candidates, and they lacked the ability to actually assess candidate quality. And they reached for people with skill that came from big companies who had the wrong attitude for an early-stage startup. They went for venture capital, which is a classic MBA move. Not all the money on the planet comes from people like Sunil and very few venture capitalists actually invest in apparel companies for I think a pretty good reason. And so, but they dressed it up as a direct to consumer and VCs were doing DTC. And so they got investors who were not domain savvy, added little value and second tier VCs who were too small to actually bridge them when they got into trouble. And then like a lot of young apparel companies, they relied on third party manufacturers to actually make the stuff. And guess whose orders got moved to the back of the production queue if Ann Taylor sort of stepped in with something that needed to be expedited. So, poor team fit, poor funder fit, poor founder fit, resulting in all of that.



I call this pattern "Bad bedfellows" and it doesn't have to kill a startup, but it surely boosts your failure odds if you've got as Quincy did, a good idea, but just not the capital team to execute on it. So I'm gonna show you along the way some data from a survey I did of almost 500 founder CEOs of early-stage startups. They'd raised a seed round of a half million to 3 million in the period 2015 through 2018. I asked them as of December, 2019 before COVID got rolling, how had the value of their seed equity changed? I mean, they obviously don't have, you can't transfer seed equity, so it's not like they had a market price, but a good founder will have a sense for whether their seed equity is worth more or less. And a lot of cases these companies had gone out of business, so the seed equity was usually worth zero. And I asked them a zillion questions about how they ran the place. So I'm gonna show you some of the data from that survey. So, remember, this includes successful founders and failed founders and founders trending toward failure. So the way you read these charts is I ran a regression model like a good academic that predicts the probability of failure based on all this stuff. And then we vary one thing at a time holding everything else constant. So did your founder CEO have prior experience in the industry in which the startup was operating? If zero, your odds of a low valuation, or think of this as the odds of failure were 13% compared to 8% if the founder had 11 or more years. So, that's a statistically significant impact, but not a huge one. If their co-founders were fighting over, lacked clarity on who was doing what, that was a pretty good marker for trending toward failure. You see on the right there, I'll show you some others. This is Quincy, example of overemphasizing skills when hiring the team, way too much emphasis on skills as opposed to attitude boosts the failure odds. This is an interesting one, a structured HR process in an early stage firm, we think of early stage firms as being very fluid, but these firms had thought through onboarding and compensation approach and a set of policies that the other firms hadn't. So best in class really reduces the failure odds, whereas no structure at all tends to boost it. And then the last one here, if folks like Sunil were adding a lot of value, that reduced the failure odds, very little value added, boosted them. And as with a lot of statistics, you gotta be careful with causality, on the right you see if there was frequent intense conflict with investors over policy and strategy, that really boosts the failure odds compared to everybody getting along well. Now you can ask "Whether, are we fighting because we're failing or are we failing because we're fighting?" And that's a fair question. This type of statistic doesn't let you sort that out. How do you avoid this pattern? Well, you can try as Quincy did to recruit experienced co-founders and advisors, doesn't always work. In this case, they could have done a lot more research on apparel before they quit their job. And these two founders essentially shared the CEO role. And that's a prescription for slowing things down unless you have a crystal clear delineation over who's gonna do what. Important in an early stage firm to hire for skill and attitude. With investors have a stopping rule if your fundraising falls short, the Quincy team wanted to raise 1.5 million. They only raised a million, the extra half million, they were actually making progress to getting their operations under control, really contributing to the failure of that company was, the promise was good fit, but a sizable fraction of the apparel just didn't fit well. The return rates were normal for a e-commerce company. But if your promise is good fit, you don't want normal, you want far below normal. And then if your partners are ornery, there's not a great deal you could do. You can't sue them. So you can yell at them on social media, you can get your advisors to lean on them, you can give them equity to try to get the goals aligned, but now you have them on the cap table forever. So if it's not going well, that's probably a bad idea.



Onto this guy. He may look familiar to a lot of you who know Ubiquity well, this is Sunil. I would guess about Sunil, 13 years, 14 years earlier, hasn't changed much. So, Sunil is my example of the right resources, but never being able to get those resources to steer the company to the right opportunity. So wrong opportunity, right resources, the business was an online dating business. When we get to the interview portion here in a minute, you'll hear a lot more, I think about Triangulate, which was Sunil's firm. He'd raised $750,000 from a very good VC, top tier VC. So right resources, recruited a cracker jack team that could crank out stuff, data scientists and great engineers and launched on Facebook's platform when the platform was brand new with a lot of enthusiasm for Facebook. So good partners, good team, good investors. And Sunil himself knew his way around startups at the time. So kinda the polar opposite of the problems with the resources that we saw with Quincy. But data-driven dating, I'll describe it a little more, turned out to be not such a good idea. And even though Sunil's team did three big pivots in a very fast fashion, responding to customer feedback, they couldn't get to where they needed to go quite fast enough to save the company.



So, this, a diamond in the middle is a framework we use at Harvard Business School for thinking about a business model and it's about customer value proposition, technology and operations, how you build it and service it, how you make customers aware, and then how are you gonna make money from doing all that. And so Sunil's idea was initially essentially algorithmic matching. The original idea behind Triangulate was gonna get into your patterns online, especially social media, see the pitter-patter of what you were doing on social media and other online activities, and then use that to match you to somebody who looked compatible based on their pattern. And he was gonna build an , and did build an algorithm that did that and licensed it to existing dating sites. Now he got advice from a different academic advisor who said, "If you're gonna do that, you gotta launch it yourself to prove that it works," which he did. So, Triangulate became Wings, which was an online dating site. This other academic advisor was fascinated with social things and there was this concept of social proof. So the idea was Sunil was gonna bring your friends online with you so they could vouch for you, just like a wingman. If you saw the "Top Gun" movie, you know, the lead pilot always has a wingman that's protecting the pilot as they go in for the mission. So that happens, as you all know, in bars. And Sunil wanted to bring that activity online. So a lot of bad things happened here. One thing was it turned out that daters didn't need an algorithm to pick who they would like to meet. If you showed them five profiles, they didn't need the algorithm to say, "That's the person I would like to meet as opposed to these other four," nor did all of them want their love life unfolded before the eyes of their good friends. So that was a problem. The wingmen made navigation and engineering complicated. 'Cause you had to build essentially two versions of every feature because there were regular daters and their wingmen, the thought was the wingmen would make the thing viral, but that didn't happen at the rates that Sunil was hoping, and it was a free to use service. He was gonna monetize by selling the ability to send digital gifts and the ability to send messages and so forth. But that just never happened at rates that made any kinda sense relative to the cost of acquiring customers.



Sunil hadn't raised enough money to play the game that really drives success with online dating, which is by subscribers and sort of harness a network effect that's so essential to the success here. So he pivoted away from the algorithm, pivoted away from the wingman, was making a lot of progress, actually had discovered a way to expose people to profiles before they saw a photo. A big problem in online dating is physically attractive people get deluged with requests and others do not. And Sunil had actually found a formula for evening out the attention in ways that kept everybody satisfied. But by then he burned through a good chunk of the capital and just wasn't clear he was gonna be able to raise enough to sort of take the business to the next level. So he shut it down. And I call this pattern a false start, just like in racing, swimming, whatever. And here the entrepreneur builds version one of the business without doing enough upfront customer discovery research or MVP testing. And Sunil was a pretty good example of this. He was an engineer, knew how to build, he'd recruited other great engineers and what could be more natural for an engineer than build the thing and get it out there. Every entrepreneur wants to build and sell. So that's what happened with Triangulate and we can see why, right? It said, actually, if you want the number one killer of early stage startups, this is it, the false start. Entrepreneurs have a bias for action. It's again, the job description, make things happen. The lean startup movement actually, I think encourages this behavior inadvertently, you know, lean startup wants us to fail fast, launch early and often engineers like Sunil have a passion for building. So they're gonna do what they're great at. It turns out most of my MBAs are non-technical. First off, they're really good at, they hear correctly that you need great product to succeed in a startup. How do you get great product? You have great engineers. How do you get them? You exercise the networking skills that so many MBAs are good at and the world has just made it so much easier to go on any number of marketplaces where you can find engineers overseas who are very competent to build the thing for you within your budget. So you see this pattern a lot and here's a little bit of the data that's relevant to it. So did you do any research before you started your engineering? If the answer is nope, zero, it really, basically doubles your failure odds. And then there is the issue of how frequently did you pivot. If you got it right, that's great. There is such a thing as pivoting too often, the founder equivalent of ADHD and at the other end there are stubborn founders who should be pivoting. The universe is telling 'em their idea is flawed, but they never do it. So those patterns contribute to failure.



The fix is, this comes from the British Design Council, this concept of double-diamond design, where you very carefully define the problem and then develop the solution. And these patterns of diverge and converge, divergent thinking is you make the problem space bigger, wider set of potential customer segments and those potential customers have very different needs, but you have to converge on the segment that you're gonna target and the needs that you're gonna fill in a distinctive way. Likewise, with solutions, there's usually a lot of ways to solve those problems and you've gotta explore those before you converge and actually start testing something, a real product. So if you look at where MVP testing is down below, it's at the end of a process, where there's a whole bunch of work that a founder can and should be doing before they actually build a thing and put it into potential customer's hands. There's all this ethnography and customer interviews and survey work you can do and developing personas, journey mapping and so forth. And what you see with a false start is that work to the left of MVP testing, in a software business it might take only a month, like four weeks. And essentially if you follow the false start pattern of building and selling the thing, it's probably four months to build it, put it out into the marketplace, figure out it's not working, and then figure out what's gonna work better.



So essentially to avoid spending four weeks, you've wasted four months and if you've only got a year's worth of capital in the back, that's a bad trade. So do that upfront work if you're just getting started, the last early stage pattern is what I call a false positive. And we're all familiar with false positives from the last few years. These dogs are doing what is called a baroo. There's a word for when a dog turns its head to try to understand what you're saying. That's Lindsay Hyde, whose startup was called Baroo. It was a dog walking service. The wrinkle was she was gonna do deals with luxury apartments to basically be the preferred dog walker for a luxury apartment building, so that would be her built-in go-to market approach, which turned out to be a pretty good idea. The first customer was the Ink Block in South Boston. If you've ever been to Boston, this is the old headquarters of the "Boston Herald newspaper," which exists no more. It was turned into luxury condos all on the same day. So 100% new leases. A typical apartment building is gonna turn over about a quarter of its residents every year, of its renters. In addition, when this complex opened, they had a lot of units that were filled by a Hollywood movie crew that was in Boston shooting 15 hour days with their pets for months at a time on per diem. So they had money to spend, they had pets and they needed a dog walker. And if you lived in Boston in the winter 2015, we had an amazing eight feet of snow in 30 days. So if you were Baroo and you could walk dogs, you were very, very popular, very successful. They blew the doors off the Ink Block and within weeks had other luxury apartment buildings in the surrounding area demanding, "We need this concierge dog walking service too." Signed up a bunch, Lindsay had launched the business thinking she was gonna do it from a small amount of angel capital, get Boston profitable before she went to a new city. She changed her tune because she wanted to avoid the hypergrowth pressure that comes from raising venture capital. That changed when these buildings, which operated in Boston said, "Hey, we have other buildings in Chicago, why don't you come there and do it?" So she entered Chicago within a year, tried to raise venture capital, also went to Washington DC, also went to New York, raised a little bit of venture capital, but just basically overextended. And the problem was the Ink Block was a false positive, the success there because they turned over 100% of their new leases all at once and everybody was new to the neighborhood and needed a dog walker. If you do this in a typical building, you'll get new customers as they arrive, but they arrive, they dribble in over the year. So when she went to Chicago and DC the formula was working, but it just wasn't working fast enough to justify the complexity and cost she'd accumulated. So, had to shut the business down. And the fundamental issue here is a difference in needs between the needs of your early adopters versus your mainstream customers. In some happy cases, they're identical, like, the needs of early adopters and mainstream customers are the same. But in a lot of businesses, the early adopters are really different. They can be power users and when they're different, you've got a really hard product strategy problem. Are you gonna build for the early adopters or are you gonna build for the mainstream customers? So good example of that was Dropbox. If you think of probably a lot of Dropbox users on the call. Dropbox is, when Drew Houston applied to Y Combinator, he said, "I wanna build something so dead simple to use, my mother can use it to store her recipes." And he did that. His early adopters were software engineers with incredibly sophisticated needs for file management. Multiple machines, big files, collaboration with a lot of other people, tunneling through firewalls and so forth. They had feature requests that were nothing like he would see from his mother. And he had the discipline to not build for them because what he built was still good enough for them, better than anything they were using, which was probably carrying a thumb drive around. So that's not always the answer, you know, there're really three choices, you can tailor for the early adopters and migrate to the mainstream. You can tailor for the mainstream and assume it's good enough, like Dropbox did, or you can actually launch two products, which can be more costly, but sometimes it's the right solution.



So there's no right answer here. It depends on your situation, but the key point is you must understand whether your early adopters and mainstream customers have the same needs or different needs. 'Cause that's gonna drive this crucial issue on strategy. Sunil, I'm almost done. I'm gonna whip through the late stage patterns so that folks who are in the late stage know that there might be a section of the book that's relevant to them. So one pattern is a speed trap, where basically you get off to a good start and you keep going really fast, but you lose product market fit as you go and you burn through capital faster than you can adapt. And some cases it's fatal, like with Homejoy, in some cases it slams the company, but they can rebound like Blue Apron or Groupon. So that's a pattern that's very kinda, this is the number one killer of late-stage startups. Sometimes you keep product market fit, but one of two things happen and call this pattern "Help Wanted." This is Anthony Soohoo, who built Dot & Bo. It was a online furniture company and did a really good job of sustaining demand as he grew, but he was missing two things. Help wanted, missing a vice president of operations that could actually get furniture shipped across the country on time, undamaged and missing the information systems that are needed to actually keep track of this stuff. So sometimes help wanted, takes the form of a mission critical role inside the company. And it took Anthony three tries to get the right VP of operations and by then he'd burned through a lot of cash. The second thing, by the time he got things under control, the capital markets for e-commerce, online retailing, just slammed shut in 2015. Some folks on the call may remember that. Even basically babies went out with the bath water by then Dot & Bo was doing great, but even a healthy company couldn't get funded because nobody was funding e-commerce at that point. So we've got some thoughts in the book on how to manage through a help wanted situation.



The last late-stage failure pattern is cascading miracles. And in these cases the entrepreneur has a big, bold, audacious plan for changing the world, dent in the universe kind of thing. And lots of things have to go right. Fundamental change in customer behavior, cooperation of incumbents, changes in government regulation, huge amounts of capital, et cetera. Think of all those things as probabilities being multiplied together. If any one of them turns out to be a zero when you multiply things together, if there's a zero in the expression, the whole expression goes to zero. So, and a lot of like Segway or Webvan, sometimes you just can't get it together. So that's the third pattern. And then the final section of the book for anybody who's been through it will vouch for how painful this is and how little advice there is out there on how to fail well. Last part of the book gives some advice on how to think about when to pull the plug on a struggling startup. It turns out to be a very hard decision. Secondly, how to actually manage that shutdown in a way that preserves your integrity, your relationships, make sure people who are owed money get paid money, not your investors, but vendors and employees get some severance, how to communicate with various stakeholders, really important. And then most crucially, how as an entrepreneur having failed, since your identity is so wrapped up with the venture, you are the venture, the venture is you. That's a very painful episode, as Sunil can attest. And how to get past the pain and resist the very human temptation to blame other people or blame the universe for your misfortune and actually learn from the experience and be able to position yourself for what comes next. So that's it.

Perfect, thank you Tom. That was a good start for our conversation here. I have maybe 10 minutes of questions, then we'll have some time for questions at the end. What we'll do is we'll keep everyone muted, but if you have questions, please put 'em in the group chat. I'll be monitoring that and then I can post them to Tom as we go through this. But again, everyone on this call, I emailed you this morning with your prepaid copy of the book, highly recommend it and I reread it before today's event. And I was struck even at the beginning, Tom, this idea of having to define failure. You know, I want to talk about that for a minute. This anecdote you have there that many times people will say a startup failed because it ran out of cash. And that's just like saying a patient died because of a loss of blood. Can you talk a little bit about that and this like, horse jockey, these kind of oversimplifications of failure?



Yeah, I mean, there are a lot of, when you study failure, you end up studying a lot of, some psychologists or economists would call 'em behavioral biases, sort of wired into human nature. And one of them is a penchant to look for simple explanations. We like one cause and when something fails, we like to point the finger at someone or something. And Sigmund Freud had a concept called "Over determination," which is, if you were nutty in a way, there probably was more than one thing going on. And with failure there's usually multiple things contributing. Sometimes any one of those things is sufficient to end the venture, but it is really dangerous to oversimplify and we tend to oversimplify the other big human bias. If you took psychology 101, it is so fundamental. The psychologists have called this the fundamental attribution error. And in psychology it's essentially, Sunil, if you did something wrong, dropped the ball in the outfield, that was because you're not trying very hard or just not very good at baseball. If I did it, the sun got in my eyes or some, you know, there was some loud noise in the bystand. So we blame other people's misfortune on their lack of skill or will. And when something bad happens to us, we blame the things out of our control things from the universe and founders are really subject to this. So when you look at failure, you gotta ask the question, you gotta get past the question of who's to blame and look for complex causes and there's plenty of complex causes to be found. There's also this question that I went into in the intro of, failure from whose perspective? From the team's perspective, from the founder's perspective, from the investor's perspective, from society's perspective.

So, of the three patterns, we'll go through them, I have questions on each of them, bad bedfellows, false starts and false positives, on bad bedfellows, you know, I think some of the idea has to be is around watch out for who you're getting into bed with. And I'm curious if you have very practical advice for that conundrum of how can you know what they're gonna be like during tough times or other quirks before you actually get into bed with them. Like, it might be partner specific, but I'm curious.

Yeah, two things here, I mean, one is, you know, yeah, the example I gave, we trace part of the problem with Quincy to a lack of domain experience. Alex and Christina really had no experience with apparel design and manufacturing. I wouldn't over-index on that. Like, that's important in apparel. By the way, if anybody on the call has ever done food and beverage entrepreneurship, there's certain businesses out there where understanding how things work is just super, super crucial. I mean, in food and beverage, you know, can you trust your co-packer? Will your wholesaler actually ever pay you? They tend to hang onto the money that's owed to you. Should I pay for end of aisle displays? It's, you know, what should my packaging look like? Zillions of decisions that are really crucial that if you don't have experience, you know, it can be really expensive to learn, you know, trial and error and apparel is like that. I don't think at Instagram those founders needed to have worked for 20 years at Polaroid or Kodak in order to come up with an amazing... So the key point here is, sometimes domain experience is important, sometimes it isn't. And you just need to know whether you're going into an industry where that understanding of the domain is crucial. The second thought, what you're saying about sort of matching, is, I mean, there's nothing like when picking an investor, talking to failed founders that are in that investor's portfolio. I mean, there can be some "Jekyll and Hyde" behavior out there, you know. The investor will steer you to a happy founder, probably a founder that's doing really well. But what you really wanna meet is somebody who has seen the colors of that investor when the business gets in trouble. So that would be my advice on how to assess funder fit. If you're an entrepreneur, you know, talk to somebody who's failed while being associated with that investor.

Perfect and I love that recommendation. So moving on to the second pattern with regard to false starts. I do wanna dig more into that idea of the irony of if you know how to build, it feels and I'll speak personally, you know, being an engineer at HBS, it felt like an advantage and an entrepreneur should play up their advantage, so other folks couldn't build, but I could, so let me go ahead and kick it off. Or some other classmates had a connection to a team offshore that could spend $10,000 and get $100,000 of work done. So these feel like advantages. How do you balance that? You know, is it an advantage or is it a liability if you know how to build or you have a quick connection to someone who can build, why can't you leverage that?

Yeah, you know, I think it's the single strongest piece of advice I give to MBAs and I think it's very counterintuitive to 'em because basically you hear like, "Just build it, get it out there, get feedback fast." But once you bring the entrepreneur, particularly if you're a non-technical founder, once you bring those engineers on board, they are expensive in one way or another. If you're paying 'em outright, they're really expensive. If they're co-founder, they expect to be used at what they're good at, which is building. And if you haven't figured out what to build and you give 'em a half-assed version of the idea, they're gonna build that, you know, and pretty quickly you're gonna find out they have to unbuild it, you know, and replace it with something better. So, yeah, it's tricky. It's really hard to tell an entrepreneur to slow down, take, I mean, no one wants to study, no one wants to do research. I mean, it feels like you're really not doing it if you're out there just talking to customers in a structured and rigorous way, you know, when you could be building. So, you know, despite all of the very best rhetoric of lean startup, you know, I don't think it's hit home. You know, lean startup really has two sides to it. I think of it as the Eric Ries side, which is the emphasis on build, measure, learn, you know, get that MVP out there there and get it out there fast. And the Steve Blank side, Steve Blank is big about customer discovery and Steve's stuff is harder to read than Eric's and I think everybody wants to build. So we've really focused on the MVP side of lean startup without hearing and clearly enough the Steve Blank advice to get out of the building and talk to lots of potential customers before you get going.

Mm-hmm, yeah, I love that idea. In the book you mentioned that a lot of people, myself, definitely guilty, like, partially implement the lean startup stuff and it's probably more of the Eric Ries and it really strokes the ego to like hurry up and get something done and that quick dopamine hit of being able to show folks an early alpha. I wanna talk a little bit about deep tech. You know, Ubiquity is a seed deep tech firm and in deep tech, which I would in some way categorize Triangulate as like trying to do something that had never been done. It's not even clear if it's possible. That was one of the first conversations you and I had in the fall of 2008 actually, was like, is it gonna produce matches if you know the digital breadcrumbs of someone's online behavior? So if these are the two questions, like do people want it or is it even possible? Like, doesn't it make sense to do some feasibility validation before you go ask people if they want it?

Yeah, I think so. You use the term deep tech, East Coast, we tend to use the MIT term for the same thing, which is tough tech or some people are hard tech, some people are emerging tech, some people are frontier tech, it's all the same thing. It's robotics, it's space tech, it's clean tech, you know, and some bleeding edge, nosebleed kinds of software applications. And exactly what you say, the two attributes are, there's uncertainty both about technical feasibility and market acceptance. If you think about it, I think this is worth a detour, most software businesses, most SaaS businesses or B2C businesses, you usually can build the thing, but you don't know if the market's gonna buy it. And many life sciences businesses, you know, if you invent a low cost cure for cancer with no side effects, there's a market for it, but you just don't know if the science is gonna work. So deep tech has this really brutally challenging problem of being high, high on both, high technical uncertainty and high market uncertainty. And I think you are right. I mean, a lot of deep tech comes out of university labs, material science and so forth. And a lot of what goes on in those labs is establishing technical feasibility. So if you can get that done with government grants and so forth, so much the better. And that can work in material science, it can work, you know, if you're gonna get a DOD grant or an NSF grant or so forth. I don't know that it works as well for some of the software stuff that Ubiquity's focused on. So I do think that technical feasibility is crucial. I mean, the other complication here is very often the case that deep tech can be applied in a lot of different markets and the entrepreneur's task is to figure out what should be the lead market. And the complication there is that the lead market will often shape the technology in some pretty fundamental ways. And so your R&D path, your innovation path is gonna be determined by where you go first. And so the entrepreneur has this coupled design problem, you know, figuring out where the technology's going and figuring out what market to target. And it's why we call it tough tech. That's why I think the tough tech is better than deep tech.

Yeah, so I'll add two words on this from my experience investing in Ubiquity, including one kind of early big failure. I think you're looking, I'm looking for two contrary things, like a little bit of a paradox. I wanna find folks who know the deep tech area so well but are not so enamored with it, they're gonna always return to the lab whenever things get tough. 'Cause often it's often about putting the pencil down and thinking, and that was one of my first startup investment failures at Ubiquity. And then that path of non-dilutive financing, SBIR and grants I think is useful, but it can often, I'm gonna offend someone on this call, I know, but it can often imply a certain pace of development, which is completely at odds with startup and in theory my failure at Triangulate was racing too fast. But SBIR and sort of grant stuff can be even too slow. So I need to find folks who maybe can draft off of that but are willing to kick it into high gear and that's also been sort of an ongoing tricky issue.

Sunil, so good insights.

We have a few questions from the group. So, Neil had a question I think that'll be of interest to a lot of folks. So, "would an early premature exit be considered a failure? And then more generally, any rules of thumb that would indicate it's time to exit?" We've talked about struggling founders and one of them pulling the plug.

Yeah, so, I mean, to me an exit, if you exit for proceeds that get the investors back their money, that's a success of some sort. You know, could it have been a bigger success if you'd sort of dug in and kept building the business? You know, every entrepreneur that sells to a billion dollars like the YouTube team or 1.6 billion and you know, is now worth, I don't know how many tens, hundreds of billions is gonna ask themselves that question. But you know, this is where I think the founders preferences, you know, the founder gets to make that call. So yeah, if it's premature, if it's an acquihire for example, and the team got placed and the investors got their money back, that's fine. To the question of how to make the decision to pull the plug. I do go into that in the book, one of the most important factors is just basically how miserable are you, you know, and a lot of founders will have a lot of bad days, you know, but if you wake up in the morning and dread going to work for day after day after day for weeks on end, it's time to do something different. The other thing that founders get wrong in deciding to pull the plug is sort of hoping for a miracle. We hear these stories of 11th hour rescues and you know, some investor will swoop in, some acquirer will come and you know, the real reality of failing is you tend to try a pivot and it takes time to figure out if that's working. You tend to try to raise more money from new investors 'cause that's where your existing investors steer you first and they don't buy it. And then you try to get your current investors to bridge and they say, "No," but it takes a long time for them to say, "No." You try to sell the company and you get a lot of interesting nibbles. 'Cause what competitor doesn't wanna see, close up what you're doing, but they sort of dither and the longer they can string you out, the cheaper it's gonna be if they do buy you, but most of them aren't really interested. So each of these nos increases the odds that it's gonna be negative with the next move. And the whole thing can play out over in, it's usually six months or a year of sort of slow motion trouble. In meantime you're burning through capital and another mistake that founders make is misestimating the amount of cash you need to have in the bank to shut down a startup responsibly. Takes way longer to actually unwind the thing. And you wanna have some good cash flow planning so you can figure out if vendors who are owed money are actually gonna get paid, that's a good way to trash your relationship. I mean, the lawyers that helped you, that were sort of racking up fees, it's nice if you can pay them. It's nice if you can get some severance to your employees and so forth, and obviously help them find jobs. So knowing the timeframe for a graceful exit is important. Knowing how miserable you are is important and knowing that there's this very predictable rhythm of moves that a company goes through when it's in trouble and how, you know, every time you go down that path, each step on that path makes it harder and harder to succeed.

Mm-hmm, yeah, your definition of, or at least your recommendation of it's time to pull the plug when you feel really miserable, really resonates with me. At Triangulate it was a very tough decision, but I remember specifically, I would start to notice what time it was during the day and I never noticed what time it was in the day, earlier I talked about nerds losing track of time, and I would go into Triangulate next thing you know, it'd be midnight and I wouldn't know. And what ended up happening is I'd check my clock at 1:00 PM and then I'd check my clock at 3:00 PM, and I started to notice that I wasn't sort of... That was one of the sparks. Okay, there's another question here about a pattern of role clarity on the founding team. I mean, you touched on it briefly, co-CEOs with Quincy, but how critical is that? Are there tips on role clarity? Is it always better to separate or that ever make sense to collaborate a little more closely?

You know, I mean there are successful examples of co-CEOs out there, but I would say, and I don't know that anybody's ever run sort of proper statistics on it to sort of see if it fails more often than it works, it's just a matter of speed. You know, the problem is the transition from truly early pre-seed stage where you typically have two or three co-founders who sit in the same room, sort of shoulder to shoulder and talk through everything, you know, so even if you haven't formalized the roles at that stage, you know, every single decision is getting talked through with this entire small team. And, you know, before long, if you've got any kinda traction, you've now got a marketing unit and you know, somebody running engineering and so forth and it feels weird to take that collaborative pattern of talking through everything that you had not too long ago and sort of abandon that, you know, and have very crisp rolls. You know, so that's one aspect. The other aspect is sort of temperamental complementarity, you know, a lot of the best founding teams I've seen have got somebody who's the bold innovator, who's willing to take risks and somebody else on the founding team who's the ballast, you know, sort of, "Well let's talk that through whether we can really pull that off" and those individuals can, I mean that can lead to conflict obviously, but it also probably more often than not leads to good decisions. It can slow things down too. So it's complicated.

Yeah, well, this makes sense, Tom. Thank you for taking the time to... Time means, 10 years to sort of dive into this topic of startup failure to produce the book. Everyone on the call gets a copy, courtesy of Ubiquity Ventures. Thanks for taking the time today and I hope to continue on this journey with you sort of making startups stronger and more successful in all the different ways that we approach that problems-

Thanks, Sunil. And good luck to all the entrepreneurs on the call. Thank you.