March 8, 2024

Cost-Traps that Cost Your Company Millions David Heinemeier-Hansson, 37signals

David Heinemeier Hansson is the inventor of Ruby on Rails and founder of 37signals, a company that has developed products such as Basecamp, and more over the last 20 years.

However, apart from Jeff Bezos, David and his co-founder Jason Friedman didn't take any money and bootstrapped the company.

We talk about the biggest cost drivers for companies, how 37signals managed to develop multiple products in parallel and how David thinks about the differences between Europe and the US when it comes to starting a company.

What you learn:

  • When might it make sense to move away from the usual cloud services and choose your own hardware resources instead?

  • What advantages does the USA offer compared to Europe if you want to set up a start-up?

  • Who should raise money from venture capital investors and who shouldn't?

  • How do you learn to correctly assess which topics (e.g. marketing or hiring) it makes sense to spend money on?

Here you can find tickets for MERGE: Https://

Don’t forget to use the code UNICORN for 30% off.

Here you can setup your tech with Codesphere: 


David Heinemeier Hansson




Unicorn Bakery Whatsapp Broadcast:

Find out everything you need to know as a founder:       

Our WhatsApp Broadcast keeps you up to date with insights into the scene, news and top content.


(00:00:00) How do you resist raising venture capital? Why was it not relevant to you? To whom would you recommend raising venture capital?

(00:10:51) Would AI make it possible for me, as a non-technical founder, to build a technical company these days?

(00:16:22) What are the biggest cost traps that founders run into?

(00:23:16) How do you assess new costs like hiring or marketing?

(00:29:08) The 37signals setup: team size and product variety

(00:33:31) How did you and Jason find out what company you want to build?

(00:43:23) How much money did you save by leaving the cloud?

(00:51:41) US vs. Europe - what are the differences in living and working?

(00:57:54) What would be the reasons to build out of Europe?


 (00:00:00) Welcome to a new episode of the Unicorn Bakery. My name is Fabian Tausch and today's guest is David Heinemeier-Hansen, founder and CTO of 37 Signals. Better known for the products Basecamp, and since recently, Campfire. David is the inventor of Ruby on Rails, won the 24-hour race of Le Mans, which is very impressive, and enjoys his life as he likes because he, because 37 Signals was built as a bootstrap company for 20 plus years now. Congratulations on that. And besides selling secondaries to none other than Jeff Bezos, the company is fully in the hand of David and his co-founder Jason Fried, which is I think something we need to talk about because it's so different from most of the company stories and venture stories that we hear out there. So David, welcome to the Unicorn Bakery. Thanks for having me. Let's start with the first question that I had in mind because you bootstrapped for 20 plus years and you saw so many other companies raising lots of money and it always sounds so crazy. It sounds so cool. How did you resist raising venture and why wasn't it relevant to you? Both Jason and I were born in this world of startups at a time of great exuberance in the late 90s when the dot-com boom first really was going crazy was when WeCut started working. I worked for a number of startups in Denmark that were funded by VCs and I saw how that turned out once the dot-com bust happened and it really wasn't pretty. And I think it gave both Jason and I kind of a vaccine against that kind of business building because we had seen it firsthand what happens to companies who take a ton of money before they have revenues and how ugly it is when that goes sour. And it usually does, right? The math of venture capital is that a very small minority will end up paying for the rest. (00:01:48) But that means a very large majority ends up going bust and a large share of those go bust in really uncomfortable, terrible ways. And I saw some of that firsthand and thought like you know what I'm not interested in building that kind of business. There must be a different way to build a software business and of course there is. Most businesses in this world in the history of commerce were not built as venture capital funded businesses where they were constantly raising more money and they had to be in these extreme growth rates right from the word go in order to make it to the end. So we built a software business by bootstrapping it off a consulting business which I think is still a viable great strategy that plenty of others have followed. No strategy will guarantee you success but there are certainly differences in how bad the strategies play out when it doesn't work which again is most of the time. Most businesses will fail even the ones that are bootstrapped, even the ones that are bootstrapped off consulting revenues. And I think that's an instructive case to look at because if Basecamp for example had failed, that was our first kind of big launch, what would have happened? Nothing. We would have learned a bunch of stuff launching a product. We would still have complete control of our destiny. We would just keep working with clients for a bit longer and then maybe we'd try a second product. Maybe we'd try a third product. Maybe we'd try 10 products before we found one that really worked and were able to get us to where we wanted to go which was ultimately to convert this consulting business into a software business. And that opportunity to have as many tries as it takes without leaving you destitute is in my opinion a really appealing one. Now that doesn't mean that everyone should be starting consultancies in order to do product businesses and there's many different roads to get there but that was the road we followed and I think it is important to look at it in its contrast to the VC racing model especially when it comes to risk. There's this general sentiment that to be an entrepreneur you have to be a huge risk taker and I think it's actually wrong. I think a lot of really good business people actually hate risk. I don't like risk at all. Sometimes you take a calculated risk and you go like oh okay I've triangulated everything here and that makes sense but a lot of times you look how can I eliminate the risk. How can I make it so that we can take a bet on something without betting the company for example. And that's something we've done for over 20 years. We've taken all sorts of bets. (00:04:20) There's no guarantees in business. You don't really know what customers want until you put the product in front of them and ask for their credit card especially in software. But you can do it in a way where if it doesn't work you're still fine. It's so interesting because I think a lot of founders should ask themselves more and more the question do I want to bootstrap? Do I want to raise venture? But as so many stories and I'm guilty as well. I'm a media company with the podcast. I'm interviewing a lot of venture-backed founders and you hear more often about them than the bootstrap stories. It feels like oh if I want to be a complete entrepreneur in quotation marks I have to raise venture. And I often try to talk founders out of raising venture and I'm like hey why do you need it and do you really need it? So asking you because I know that you're good friends with a lot of venture-backed founders as well and you've seen also some stories like Shopify play out super well. Who would you recommend raising venture and who would you say hey please think about it twice? I'd say venture is a great path if you're in a winner-take-all market that needs to be captured immediately or it's gonna close. So I think the traditional examples of that are things like eBay or Facebook or Twitter, social media, anything that has substantial network effects and where those network effects are going to be super difficult to catch up to if you're not already in a leading position. Those are the kind of businesses where venture really gives you an advantage I think. Where you have to be speculative, where you see that the potential market could be enormous and we have to run incredibly fast to get there. That's the whole blitz scale thing. I don't think it matters if you're building tools. If it doesn't, if there's not a strong network effect that for example Basecamp, the tools that we built, or even hey our email service. They're both of a quality work if a company chooses to run their project management on Basecamp. It doesn't really matter if someone else uses I don't know Microsoft Project or Asana or Trello or any of the many other alternatives that are out there. It doesn't take anything away from them. Versus if you're the first one to some new social media and there's like a thousand people on it, it's not very useful. Social media networks follow Metcalfe's law where the value of the network increases exponentially with each added node. That's not true for a project management tool. The customers we have, if we get 10x as many customers, they don't get that much value out of it. There are some network effects, there's some familiarity, people can switch between companies and oh I already know Basecamp. So it's not that there aren't benefits to it, but the benefits are just in a completely different category than if you're dealing with network effect businesses. Now sometimes it's not always obvious, like is Shopify a network effect company? I don't know. I want to ask Toby that actually. Do you think it's a network effect company? I would probably say I don't think so, but there are secondary network effects which are things like integrations and ecosystem network effects that can be important to look at even if the thing itself doesn't have strong network effects. But if you are in that tool building space, and a lot of SaaS is that, it's selling tools to individual companies where the network effects are just secondary at best. You don't need venture. If you can build a better Basecamp, you can start tomorrow and you can start getting customers. Like they show up fresh in a way they wouldn't once they've signed up, for example, for Vrax. If you build a big following on X, you're oftentimes one of the last persons to be interested in changing to another social network. I think that's why everything from Mastodon to Threads, even Threads I think is a great example. They have literally unlimited money, unlimited promotion. It's still not that easy. Oh and they have the tailwind that a ton of people hate Elon Musk like he's the Antichrist. Even with all of that, you can't just go like jump over what's already there. But when it comes to tooling and other kind of non-network effect stuff, that is possible. So to me that's the big fork in the road. Are you building a network business? Are you not? Some are in the gray areas in the middle, but others it's quite clear whether it's one or the other. I would also say and probably that sometimes falls into the network effect category as well, but it's also not as obvious in the beginning. If you're building fintech infrastructure, for example, in the European market and you know the European market quite well because you're from Denmark. If you do this in Germany, it costs like a lot of money until you can really start building something and you can't really earn it by consulting before. So there can be businesses that have a lot of costs up front, but it's not most of the companies, even when you think that. No, I think that's a good example. If there are actual barriers of entries or you're building the kind of software company, internet company that also have a large physical component. I mean Uber is a network company, for example, but let's say it wasn't. Let's say people didn't care whether it was one or the other. (00:09:41) If you'd still have to seed your market by paying drivers up front for not driving or other areas in that regard where you need a large amount of capital, yes. That's why when I share our story, I'm keen to stress that we're building software, internet software, low marginal costs, no upfront investment beyond the development itself, which by the way can be substantial. This is one of the reasons why I encourage bootstrap founders to, well not just encourage, I'd say it's actually a blocker. You have to be able to build it yourself. This is the other thing. I meet a lot of folks who have ideas about how to build a software business, but they don't know how to build software themselves. Well, you're going to need to either have a rich uncle or to have a successful exit earlier in your career or to raise venture capital because if you need to pay other people's salaries before you can even do anything, yeah, that's going to be difficult to do in a purely bootstrapped environment. So bootstrap is and probably always will remain to some extent a niche path, but a niche path that could fit tens of thousands, if not hundreds of thousands of businesses through it. Do you think, let's say I'm a non-technical founder, which I am if I would build a software company right now, I would be the non-technical one. Do you think it gets better due to all the artificial intelligence and all the stuff that I can use right now that I might be able to still build it on a low cost basis and could bootstrap it as well, even when I'm not the one building all the code? It's a good question. I think generally productivity does really matter. (00:11:16) This is one of the reasons we built most of our own software that we use to directly build our stuff, like Ruby and Rails sprung out of the needs of a bootstrap business. Then we had one technical person when we were developing Basecamp and that was me. I had to have an incredible amount of productivity available to me, so I had to have the most productive tools and in my optics that was the Ruby programming language and then I built my own tooling that was Rails and those two things combined to give us the unfair advantage, if you will, over companies who at the time were using, well I was going to say at the time, it's actually gotten worse since many of the mainstream environments require you to have very large teams, especially if you want to do native development. (00:12:00) I've been using an AI pair programmer for quite a long time now. I really like it. It is a productivity booster. I don't see it as a game changer yet. I can't tell my AI pair programmer, hey build an entire, build Basecamp. Oh no, that's quite not what I wanted. Can you change a few things? We're not at that level yet. It's entirely possible we'll get there, but it's also entirely possible that a lot of things would happen. Everyone thought it was entirely possible in 2017 that by the end of that year all cars would be driving themselves. That was not what happened. It may still happen. It may still happen that all cars will be driving themselves in maybe even the relative near future. If you've looked, obviously anyone who's observed the advances in AI, it's incredibly difficult to predict just the pace of evolution. (00:12:51) When will it get to the tipping point where it's so good that you go like, hey if I was an illustrator right now, I'd be a little worried if that was my livelihood because I'd look at the AI stuff and go like, you know what, it's actually really good at a bunch of things that fall squarely within my domain. Doesn't mean there won't be any human illustrators left, but maybe we need fewer of them. It's so difficult to predict. What is clear today though, I think that's you don't need any prediction. You can just look at what is. AI is not going to enable the non-technical founder to build an entire startup by themselves. Fair enough and sometimes good to hear for the people that are technical founders and therefore have an advantage with that. (00:13:33) Sad to hear for people like me who think maybe I had it easy way into that. Well maybe you just need to wait 18 months, right? (00:13:38) I mean, given how fast all this is going, it may entirely change, but I'd also say given how fast this is going, you could also predict another future which is that AI is just going to write everyone their own software. Just like soon it seems like everyone can get a movie of their choice made by a prompt, oh what I'd like to see tonight, can you create a 92-minute movie with these five components in it? I think it is entirely possible that AI is going to deliver software to the masses in the same way Excel spreadsheets did. Excel is perhaps the most successful software platform of all time. It has enabled final end users to create databases and all sorts of advanced modeling software and a bunch of things when they did not even need neither software developers nor another company. So it's also speculative which is why I think it's just a fun time to be in the business because there's so much uncertainty and even like what is 2026 going to look like? It's entirely possible it's not at all different from what it is right now and it's entirely possible that the whole world has been turned upside down. I love it. I wonder how many people listened right now and were like wait a minute, I never thought about Excel being a software platform. I mean that is the long-term dream and overstated overhyped dream as I've seen it for a long time. People realize oh Excel actually is a software development platform. What if we made building web applications that simple? I've seen at least what three generations of no code tooling in my time and they've all been kind of bullshit because the reason why Excel works is it's such a bounded domain. Like it can do these things and within that it's really good. You're not going to use Excel to build Basecamp. That's a nonsensical term and that's where a bunch of the no code stuff had gone. Like oh no no no actually we could take the Excel model and we can just make it generic. (00:15:32) It could build any kind of software. Well yeah the problem is once you get into building that kind of software you actually oftentimes at least you care about all the decisions that turn out to be programming. So I think the ultimate no code tool, actually I think no code is a misnomer. (00:15:49) It's always going to be code. The question is who's going to write the code and on whom's direction. I see it as entirely possible that who's going to write the code of the future is going to be AI. They're still going to write code. It still needs to be executed on silicon but the code is going to be written potentially at the direction of someone who does not understand the code but can have an interaction with an AI that hones in on that code. But again we're back to science fiction. Science fiction we can squint and see the possibility of but science fiction nonetheless. Let's go back to a bit more of the hard facts about building a company especially bootstrapping it. Bootstrapping a company means running profitable, means avoiding cost traps while venture-backed companies can just run through them and be like hey we currently have enough money let's just not optimize for the most efficient way just run through them. Of course by now more of the venture-backed companies have to be more efficient so they could be a bit more aligned with with a bootstrap company by now but out of the last 20 plus years what would you say are the biggest cost traps that startups and founders run into? Well the one you got to start with right away is the biggest line item in almost every startup is headcount and the problem with venture capital if you will when it comes to cost is that you are rewarded nay you are required to spend all the money you raised more or less as quickly as possible. In most traditional cycles you raise a ton of money and then you have 18 months to blow it right. You spend the 18 months to get to the next growth target where you will raise money again that's how the sort of ladder goes from pre-seed, seed, series A, series B, all the way to exit. And when you have that large pressure to spend potentially tens of millions of dollars very very quickly the fastest way to spend it is to hire a lot of people because that's also how the kind of wisdom goes like if we want to grow faster we need more people and so forth. Now the problem is that software doesn't yield to those kind of economics in fact they often counter yield to that. One of the most famous tomes of software development is the mythical month by Fred Brooks and he has the wonderful quote in there that what one programmer can build in one month two programmers can build in two months. There's not a linear relationship between the number of programmers you have and how fast you can build desirable software. In fact there's often a reverse or a relationship. (00:18:25) Another way to put it I think he also mentions in that book is it takes a woman nine months to create a baby. You can't hire nine women and get one baby in one month. There's a gestation period here that I will apply to software that goes sort of similar if you have if one programmer can build a prototype in one month or in three months you can't just have three times as many program and then make it in one month it just doesn't work like that. So that's one of those fundamental incompatibilities I think between what venture-backed companies are trying to do and what software is capable of delivering. What are the inherent constraints when you deal with software? They are of this nature that arriving at first of all figuring out what kind of software we want it's a very difficult process. We've been building software for whatever 60 plus years now in its modern incarnation. No one has cracked it. No one has cracked like how do we even find out what we want and if anything the larger the piece of software the larger the project is the more likely it is to fail. The more ambitious it is the more likely it is to fail which is why so many of the venture cap or back companies fail very quickly because they're trying to build large software products very quickly. That's how they can get a moat in parts and software doesn't want to be built like that. The best complicated systems all started from simple systems. You can't start with a complicated system and have any good success rate to it. That's why so many large complicated systems when they're kicked off this is something we often see in Europe and also elsewhere large government systems that have large budgets they're always over budget they're always over time and very often they end up being cancelled. So software is a difficult species when it comes to this and the headcount is the main ingredients the main lever that startup companies have to turn on. Now that has always been true it's been true right from the word go but what is more recently true is that you can now also spend an unfathomable amount of money on your cloud and this has been one of the things I've been interested in. We've spent an unfathomable amount of money on cloud over the years and I just have different incentives. The way we run our business is Jason and I get to share what's left over when we paid all our expenses from the revenue we take in. It's a very simple model like if we take in 10 million dollars worth of revenue and we spend six million dollars on expenses we get to keep four. If I can improve those expenses so that we only spend five million dollars instead of six I get to share another million dollars with Jason. You don't have that skin in the game when you're dealing with a venture-backed company in fact you have the opposite right like the whole goal is to spend all the money you raised in 18 months and you don't get to keep any of it if you don't like the whole idea of keeping the money is only something for battle times right which we are in right now that is fair to state right two three years ago it was exactly the opposite it was the go-go days and it had been the go-go days for 12 years so you had several generations of startup founders who had just been trained that the best thing you can do when you run a startup is to spend as much money as quickly as possible all on the altar of juicing those growth rates because growth is the only thing that matters no one gives a damn whether you're profitable or not profitable how much money you're losing what your burn rate is if you can post those high growth numbers you will get to raise more money now times have changed and now more companies have found like oh actually we can't just raise fantastical sums on back of the envelope presentations like now there actually has to be revenue and there has to be a unit economics that at least seem plausible and preferably even maybe there has to be profits at some point just that you can last longer so it's interesting how these things are converging for 12 years we were in highly divergent paths bootstrap companies always had like hey if i save more money i get to keep the money venture-backed companies were always on like what do you mean save money spend all of it spend it faster spend it bigger and now we're sort of actually bootstraps stay where they are they never waver we get to keep the money the venture-backed companies are coming back towards a bootstrap but that's a temporary condition i think it's important to realize as soon as the go-go days start back up which usually inevitably they will they will forget all about sort of solid economics in in a near-term sense and just go like well wouldn't you rather want a small slice of a very big pie that's how it's usually sold i think the important question as a bootstrap founder or someone who wants to grow efficiently and wants to be profitable can be even when you raise venture is okay obviously you have some money left at the end of the year i'd say for for you and jason and the question is finding the optimum between how much money do are we able to like yeah have as a profit end of the year versus how much money do we invest or spend inside the organization to make it as great as possible great meaning the the best compilation of all the different sorts of things you have to do to run a successful company so when do you or how is your assessment of hey do we hire someone else don't we hire someone else do we invest in let's say marketing or all the other things like how do you assess costs and new hires for example it's a great question i think i'll start with the fundamental premise of how do you make a great company and how do you define that greatness what i found time and again is that there are lots of companies that make for a great 20 person company a great 40 person company a horrible 500 person company and startups very often fall into that that they the reason they were able to hire 500 people was because that initial small group was really effective really had some good ideas really attracted attention either from investors and usually from investors because they attracted attention from the market because they just made something good right and then they went wow yes now let's um let's hire a bunch of more people and they managed to take a company that actually was a great company and turn it into a really shitty one and it happens just time and again this is the fundamental disease of bigness and what's also interesting there is that there are lots of companies whose trajectory in terms of productivity is negative when they add people this is goes back to us we were talking about the fundamentals of software development that what one program can do in one month two programmers can do in two months 500 programmers can do in three years right the heuristic we've followed since the beginning is hire when it hurts and when it hurts is defined as it should be painful not to have this person there should be obvious things that are missing broken burdening others it's not speculative and this is what most venture capital hiring is driven on very speculative oh if we hire another 20 engineers we can probably come up with a bunch of features for them to implement and that will probably create a more attractive product which will probably generate more revenue there's a lot of probables in there and only one of them have to go sour for the whole thing not to be true and that happens just time and again that you actually end up making the product worse by adding more to it this is one of the reasons why we historically have never been really afraid of not only venture-backed companies but also large companies because the kind of software that a large company assigns 200 engineers or a venture-backed company that assigns 200 engineers to the problem will produce is just categorically different from the kind of software that five people will produce or 10 people will produce and we produce the kind of that kind of software we produce the five to ten people kind of stuff and that just ends up being very different and customers react to it very differently the number one reason people state over and over again why they love Basecamp is because it is so simple to understand that they don't need any training on it if you look at a lot of competitors that we now have for a long time Basecamp was in this odd position that it didn't attract any competition these days it has tons of competition virtually all of it it's very complicated software by this point but usually what happens is some competitor enters the market they have a really novel interesting approach perhaps and it's simple to use and people are attracted to it then they attract the funding they hire the people and you give it two or three years and it just got features coming out the ears coming out the wazoo because that's what the 200 new people did they had to do something right that's juicing I think a great example that we looked at recently was slack I don't know if anyone has gone through the the config pages of slack recently they're like there are hundreds of little check marks all representing little enterprise deals that needed that esoteric feature to be present in order to be closed and then the whole thing just grows a thousand and warts that doesn't mean that there isn't still a good product underneath all of that junk somewhere I think there is I think it's like it's a great product and that's why it had success but it's also true that now that it has that many people working on it it just expands and it grows and goes further and gets more complicated and we introduced a new chat tool called campfire which was built by one of those tiny teams right like we had for a long time we had two programmers on it and then at the end we sprinted a little bit and added a couple more but it's just it results in fundamentally different software and it therefore attracts to some extent a fundamentally different customer so hire when it hurts has to me been the way we temper ambition that actually the business should be in front of us we shouldn't be in front of the business and this is requires some mentality change usually when you think invest you think well I gotta be in front of the business I am investing in the future I often don't see that working out it's not how we run our business we we try to get the business to run ahead of us and then we have to hire to catch up with the business rather than try to drag the business to catch up with us because we've done a bunch of hiring for everybody who's not too familiar with your company how many people are you in the company in the organization and how many products are you currently building we have 68 people right now we've been swinging from high 40s to I think we got up to like low 80s at one point and we're dealing with first of all a long a large suite of heritage products as I like to call them that we've been building for 20 years we have maybe seven of those they don't see new feature development but they do see maintenance and continued service to existing customers this is one of the weird things that we do we like to not evict customers Google of course is famous for evicting customers every five seconds whenever they change their direction or a new VP is hired and they shut down some service and and so are startups um there was a good example of this recently what was it to remember Notion just bought some company and and then the company goes out and says like hey yay we got acquired and and pop the champagne on Twitter and spray it all around and then the customer's like oh cool you got acquired that sounds awesome what's gonna happen to us and they the company was like well you got six months to pack your shit and get the fuck out of here you're like wait wait wait I'm not sure that champagne was for us maybe that champagne was for you and your investors but it certainly wasn't for users so I just kind of grown allergic to that both from large companies like Google that constantly seems to do this but also from startups either because they pivot that's the positive right and doesn't mean you can't change direction sometimes it isn't right to change the direction but there's a certain glibness to it that I think amounts to a little bit of pissing in the pool right if you teach customers that they kind of can't trust any SaaS service that they sign up for by continuously burning them with these eviction notices I think that does kind of pollute the pool for everyone else coming after you because now you want to be a startup and you want to convince someone to invest in your or to invest their time and their data in your SaaS solution and they'd be like yeah I'd like to your stuff looks cool but the last two times I did that the company ended up being acquire hired and everything was shut down in six months and it kind of sucked so so we have this notion of until the end of the internet that's the promise that we're making on our SaaS platforms then either we're going to keep operating this until the end of the internet or until the last customer turns off the lights and this has led to some sort of funky weird almost pathological cases like we have this service we created in 2005 which was a free to-do list service this was one of the first back then it was called Ajax applications like dynamic JavaScript applications and it was free again and we started in 2005 we ran it for a couple of years with new customers then we shut it off for new customers I think maybe 2009 and we've been operating it ever since and we're still operating it last I looked there were about a thousand users which is insane in itself that after we haven't developed something for like 14 years on a free service or so a thousand people use that as the main way to track their to-do list and we're like yeah I'm happy that we can keep running it for you so that's what we're trying to be something else like insert some credibility into the broader point maybe extract some of the pollution from the the pond so that there are company or customers can at least think you know what there are customers or companies who will do this that will stand by their products even if they change their minds right we've done this with uh with Basecamp our most successful products exist in three versions including a version that we have not developed with new features since 2010 that still has thousands of customers that's a multi-million dollar SaaS business with happy customers who don't want something else than the thing that they signed up for back then I kind of wish that we had more of that we're trying to introduce a little bit of that systemically with our once series of products which is installable software or at least if the company gives you the software and you are allowed to install it on your own server if they go out of business or they change their mind or whatever then they can't take the software away from you and yeah I think that's I wish there was more of that we're kind of trying to provoke the industry to at least think about it so we'll see where where that that goes I think one question that arises when I listen to you with all the noise out there of how startups work how businesses work how everybody like what's fancy what isn't how did you and Jason find what kind of company you actually want to build yeah so I think part of it was we had both had such a bad experience working for a number of startups doing that dot-com boom-bust cycle that that gave us one direction or impulse then I at least personally had a number of bosses that I thoroughly disagreed with on a large number of topics and I found that you know what a lot of things I received as established wisdom as best practices as things to do felt bullshit appeared bullshit turned out bullshit and it gave me just a very healthy skepticism over any received wisdom and gave both Jason and I I think the impulse to go back to first principles and this is one of those things that in itself has turned into a bit of a buzzword first principle thinking is that now in itself a best practice I don't know it gets a little circular at some point but this idea that we will try things without assuming whether they're going to work or they're not going to work and some of the things is not going to work obviously a lot of things don't work that you try but if you peel back enough layers and you get get deep enough down in the stack of insights that you approximate something that's like a clean slate and then you start building things up I think you have a better chance at figuring out first of all understanding why the things that work actually do work and also finding new pieces that can work well I also just think that's fun I mean both Jason and I think of the company as a product in itself and as a product it takes iteration it takes updating it takes experimentation it takes all of these approaches to to trial and error that a lot of companies don't necessarily think about their own company in that way they don't think about it's something that continuously have to be product tweaked and rewritten or refactored or what else have you from the software world metaphors that you can apply but if you do think about in that way and crucially and this is perhaps actually the key ingredient where the rest makes sense have the independence to do so then you can end up in a quite different place and I'd actually say I probably should have started with that because it's kind of like you see some masters of chef on tv and they're like doing all this cooking and then they pull out oh yeah and then I spend a little bit of time like creating this gratin and like that's the key ingredient that's why it tastes so delicious yeah you kind of kept that off camera the key ingredient here is independence you can't do a lot of things that Jason and I have done you can't experiment in the way that we've done over 20 years if you have to answer to bcs who want to see growth targets in three months it just doesn't lend itself to that kind of dynamic when you operate like that when you have these very urgent needs to produce certain numbers you're going to actually follow the playbook as far as you can right you're first going to go to the playbook the bc software company playbook try to see if you can find a chapter on whatever else you're facing and just deploy that now that's going to give you a lot of hurt thinking it may also work I mean there's a reason why you call it the playbook because it generally consists of chapters that have worked at least sometimes somewhere for someone else is it going to work for you maybe but it is a way to start and it is a way to go quickly especially if you have to spend a ton of money right away we never had that and we had so we didn't have all the money but we had all the independence and I think those are two opposing forces and you end up or you can end up in a place like we have when you have way more independence than you have resources and it's just a really interesting different place and this is more than anything why I advocate that people review their options because I just want to see interesting novel different thinking in this world and if we're all subject to the same constraints the same timelines the same playbook the same funding model we're going to get monotone similar looking types of innovation and that's just bad for entrepreneurs I think it's bad for the world we're not going to see the variety and diversity of ideas that we otherwise would have seen I think this is why a lot of our our products don't feel they don't look they don't taste the same as what comes out of the playbook based VC model it's like why we can do things like until the end of the internet it's why we can stay with the SMB market why we haven't moved up market to enterprise sales the playbook as it is for software startups usually goes you use SMB as a stepping stone to validate that you have something and then once you have something switch the whole thing over to an enterprise sales model where you charge per seat you have a large sales force and your entire business is funded by a handful of super accounts and we never had the interest in that we never hire a sales force I don't get on that well with sales as a function of a company so we routed around that and therefore ended up in this weird and again only weird in relation to the VC model weird place where we can continue to stay with an SMB market that we know love and feel part of ourselves why do you think is it that all of the products converge and what is needed to also bring this novel thinking back into the ecosystem because often or how people often start to learn things as they copy stuff then they model it and then they are at a level where they are okay now I can find my own way in here so do you think it's just a lot of founders are in the in their first cycle and therefore need to copy and then model and then come up with their novel thinking or what is what is necessary here no I actually think it's the opposite I think that it's almost like creativity most children are born highly creative I mean I have three kids of my own and when I sit down with them to draw or whatever I'm always astounded by the creativity that they seem to inherently have in them and then if you look at kids who've then been to school for 15 years a lot of them will have the creativity beaten out of them I think it's actually the same with entrepreneurs that entrepreneurs will have the creativity and the novel insights beaten out of them by the funding model by the dynamics of VC capital and then they will end up less likely to produce divergent thinking on paths that aren't compatible with the VC approach after they've been through it the first time again it does not mean you can't be creative it does not mean you can't create successful businesses it just means that like the lane is a little narrow I think what you need to create the sort of divergent outcomes is you need different constraints you need different playbooks and I think this is why we've ended up being such a outlier in some regards is because we have those different constraints and I think a lot of other other entrepreneurs if they were in our shoes and they were sitting with our kinds of constraints a different kind they would also have ended up with different kinds of businesses and we'd also have been better off for it so I think if you want if that speaks to you then you need to find a different model the model is going to produce the cookie cutter-ness in that regard it's a little bit like Hollywood so Hollywood has really started to just shrink its creativity it's because there are no creative filmmakers left in Hollywood I refuse to believe that it is because the big studios fund one thing and one thing only continuation of existing IPs because that is seen as the sure bet and there's some of that with the venture capital model too that the kind of founders and models and whatever they kind of have to fit that shape when it comes to my little corner of the world like productivity software business b2b sass it's very much the as we just talked about all right you start with your SMB stuff you you get some market validation then you use it as a stepping stone and then you immediately switch over to Percy's enterprise sales and like yeah okay that produces just a certain kind of software it just does those forces are larger and bigger than you and you can't really fight them you're going to be cutting against the grain when you do that so if you want something different you have to have different constraints there really is nothing other to say about it I think Jason and I if you had inserted us into the Silicon Valley rodeo and applied the venture capital forces on us we would have produced the same stuff I think we are less unique as individuals than the systemic approach we took is responsible for our divergent outcomes I think you summarize a lot of things that I'm sometimes unhappy with when I look at the European ecosystem for example of startups and that all feels like the same just x is y and you just solve this little problem or this little problem instead of like some doing something completely new completely different and it's interesting it feels or resonates a lot with what I thought about the last weeks and for everybody who wants to be or listen a bit more about creativity and entrepreneurship the episode before this one that I publish is with Scott Belsky currently chief strategy and chief product officer at Adobe where we talked about creativity so therefore could be could be a recommendation as well but until then we have a few more minutes that we can talk about stuff just making a small turn back to what you said earlier said earlier when we talked about cost traps we talked about the cloud or you talked about the cloud and that you left the cloud and back to hardware just for founders who can't really like grasp why that's the thing can you give us a rough number on how much money you saved in the first year out of the cloud yeah so we published an enormous amount of detail on this question exactly because I expected a fair amount of skepticism and of course there would be a lot of skepticism because cloud marketing has been exceptionally successful probably one of the most successful marketing campaigns in tech that I've witnessed since maybe like everything needs to be java from the late 90s or something the idea that the future is cloud and that everything trended towards other people owning the computers you run your software on was just such a pervasive paradigm that when it became clear to us that we had to leave I knew that we had to bring receipts there was not enough to just bring like insights it was not enough to bring ideas we had to bring receipts and those receipts had to stem from real action and the sort of overarching receipts which was we put out there was a very conservative back of the envelope calculation that said we were going to save seven million dollars over five years and we're sort of on track with that what was it last year a few months later we could already see all right well a million dollars worth of savings in terms of what if we had stayed in the cloud had ticked in so the idea that we were going to save seven million dollars over five years is a really interesting data point because there will be entrepreneurs who've been very successful raising money who'd go like let's chump change why are you even wait wasting your time trying to save seven million dollars over five years you could have spent that attention as an opportunity cost on conquering new markets gathering new growth first of all I don't think these things are in opposition I think people are vastly overestimating the difficulty that a lot of the cloud exits will entail a lot of the stuff that we have documented has been how running your or owning your own hardware actually gives you a very similar operating experience to renting hardware from hyperscalers and running it in the cloud not in all regards but in a lot of regards so it's a false dichotomy in part to say like it's about opportunity cost and second of all there will also be entrepreneurs who are just like well seven million dollars over five years I'm not I'm not spending anything like that I'm spending two hundred dollars a month on some cloud vms great if you're at the level where a you have a unvalidated business you haven't found product market fit you're not really spending any material money on cloud and you have the money to spend on it continue please don't uh start buying thousands of dollars worth of hardware for a speculative idea that might be bust in six months that's really foolish so I think that's why cloud still has an absolute place I would start on cloud probably tomorrow if I did a new venture and I didn't know whether that was gonna still be in business 12 months later or not and until my bills were higher but our bills when we did our accounting for 2022 were 3.2 million dollars in yearly cloud spent that to me is real money I guess you can get phased by all the billion dollar companies out there like I don't own a billion dollar company for me like a million dollars a year a million and a half dollars a year that I get to keep and split with Jason like that's real money either I can get to just keep that money or I can spend that money on something I'd rather have than sending large checks to AWS or Google or whatever you use right we could spend that money on more people we could spend that money on all sorts of interesting things and we really rather would so it comes back to that discussion about do you even care about cost if cost is not a relevant metric to you because you're on a journey of venture capital expenditures then fine don't do it spend all that VC money until you can raise more but I think what's been fascinating is that our cloud exit narrative has coincided to somewhat a degree with the general narrative that oh shit I can't raise all the money in the world anymore I can't get the crazy valuations people were getting in the fall of 2021 if I want my company my startup to survive longer than the next six months I have to extend my runway how do I extend my runway I do it by cutting my cost what are we spending on oh we're spending on that headcount oh do we need all these people probably not and second what are we spending on cloud oh shit are we spending that much money if we could somehow find a way to cut it in half like those weirdos over at 30 seconds have done maybe our runway is going to be like 30% longer 40% longer that may very well be the make and break difference between whether we survive or not so I think the whole discussion has become far more relevant for former companies and that's great because I also think this whole discussion of owning your own hardware is first of all a fundamental business lesson should you own or should you rent like that's assessment you should do with all sorts of things some should I hire someone or should I just pay a consultant if I pay a consultant yeah they're going to charge me a lot more than an employee but I can also then say okay the project is over and we part ways right um the same thing with hardware if you need a computer every day for a year then you should probably buy it right at least once you reach a certain scale if you just need it intermittently or you only need it for spikes great rent is a great option so having just some of those fundamental business discussions about how you spend your money on what at what timeline I think is instructive for all businesses and I think startups sometimes get to loop around that and then they become very good at the thing that they do which is spending money you can learn to become very good at spending money and then you somehow don't really understand how you make money how you get to create value that you can actually capture and keep that to me is actually the exciting part of a business my I'm not excited about having 10 million dollar a month a burn rate I'm excited about having profits money I can keep which comes back to again the kind of outcomes you're looking for if you are a traditional BC funded entrepreneur in most cases you will never get to keep a dollar of profit from the company you will make all of your money selling your stake of the company to someone else who then think that they probably can sell to somewhere else and then they will sell to someone else and maybe the fourth holder of that equity will realize some value in terms of profits you're very far removed from the fundamental business dynamics of making money and again lots of people have gotten very rich on that model peace be with it um if that's what you want great not what I wanted yeah fair enough um I think that's that that will let a lot of people think about that a bit more and I will also link to your articles on a bit more of the decision making and how you did it because I think you documented it quite well and I really enjoyed seeing it even when I'm not in a situation where I can save millions of dollars by leaving the cloud it was still like okay was a topic that I discussed with one of the adjust founders who sold their company for a billion at some point but they always rented because they said hey our cogs for everybody costs of goods goods sold would be too high if we would operate in the cloud because we're an infrastructure business we're not a typical SaaS business we could not even do it in the cloud we would not be a profitable company so depending on the business model it can also be an issue to to run on the cloud different story but so knowing that this can be an issue um also seeing more and more attention towards it so that people make educated decisions it's not you don't use the cloud or you like you shouldn't use the cloud it's more a is the cloud the right thing for what you're planning to do and if you make educated decisions nobody will say anything against it it's just like as you said it's very well marketed and just going for it because everybody does we we talked about this quite often in the podcast is maybe not the way to do so and David you lived in Europe Denmark for a while you you grew up there you moved to the US you moved back to Denmark um for three years during COVID and I think for a lot of listeners here it would be interesting to understand the differences between the two ecosystems and as you lived and worked from both what would you say how do Europe and the US differ it's really interesting because I moved to the US in 2005 and I was pretty young I was 25 is that right at the time and which meant that I had grown up my entire life in Denmark and that was pretty much all I'd known then I moved to the US and at first is awed by this and then lived there for 15 years and developed this sort of sense oh man I wish America could be more like Europe that was the broad strokes of my perception between the two countries I'd say and part of it was based on the fact that I mean you're growing up in a country like Denmark which is a bit of a fairytale country especially for kids it's a very open country it's very safe you can do also you see all the positives of socially democratic welfare state and then I moved to the US and I saw some of the negative effects of not having that like they're everything from homelessness to wealth inequality and so on and so forth and it's so easy to get this perception oh the US is just doing it all wrong Europe's got it all figured out Europe's got the politics figured out they got the wealth distribution figured out yeah okay but and it was the but I kind of figured out once I moved back to Europe back to Denmark for three years is that all of that niceness does happen within a context of complacency I'd probably say a complacency when it comes to legislation when it comes to business when it comes to startups that in the US is just completely different there is just such a dynamic ecosystem over here there's such a dynamic group of people who really want to for want of a more gagging cliche change the world right and we make fun of that I think especially Europeans make fun of that because you know what most people don't end up changing the world but the ones who do inside of technology a lot of them live in the United States either they're from here or they moved here because this is the place where most of it happens and we only have to look at what's going on with AI to see that the center of the universe of the AI universe is still not only just the US and but even literally Silicon Valley and I think moving back to Denmark gave me the appreciation of seeing the other side of seeing that missing seeing the risk tolerance missing seeing the energy and enthusiasm for world changing ideas missing and then seeing what we have instead I think AI again is a good example because the main thing Europe is on the map for let's put aside the fact that there is actually also interesting AI developments happening in Europe there's Mistral and there's all sorts of stuff but the main thing the headlines is Brussels the main contribution to the AI discussion is Brussels making noise about regulations and you know what it's not because I'm against regulations at all I think we dearly need some when it comes to anti-monopoly enforcement but it's not it can't be about that first the order of priorities can't be regulation this that and then somewhere down the line business formation and innovation and I think it's one of those lessons that I had to move back to Europe to truly appreciate I had to move back to Europe to appreciate the fact that the niceness of European societies as it were and as it is also under threat and also under stress in some parts comes with some downsides that most societies just have inherent trade-offs that there are no societies who've just figured it all out and just can excel on all the metrics at the same time and I think Europe to some extent is arrogant about it in ways that doesn't only not behoove it but doesn't benefit that we in Europe and I say that as a proud Dane continue to be a proud Dane do ride on the coattails of the United States when it comes to technology and don't take our own digital sovereignty really seriously I mean the main contribution to the web in the last I don't know how many years from the EU has been the fucking cookie banner I mean that's just a blight an absolute terrible mistake on the fabric of the web that is harming people around the world in just constant annoyance and sapped attention for no good reason now there is a beautiful ethics behind that but those ethics did not translate into actionable usable user improving legislation and I think this is where I'm starting to lose faith that the EU even has any capacity to to fix this and I think if this is where if you zoom out and you look at the overall economics of the picture you see you go back to the beginning of the turn of the millennium 2000 and you chart sort of economic development US versus Europe Europe is just continuing to fall behind again it's not all like economic progress is not the be all end all of everything I would say plenty of Danes do not envy their American counterparts in wherever they are in the socio-economic hierarchy and want to swap places but neither is it true the other way around and that's okay can we have can we let America be America to some extent and then also let Europe be Europe and then like we'll at the edges try to learn from each other without saying like well America needs to become Europe where Europe needs to become America yeah it's become more nuanced and I've gotten a greater appreciation for the American model and what it gives even as it's so obvious what it takes yeah thanks for sharing the perspectives on both sides if you would still mend the argument for a founder who's currently thinking do I built in Europe or do I built in in the US what would be the reasons to build in Europe I say if you are bootstrapping and you can internalize the American mindset in terms of startups and world perspective Europe is a wonderful place to build I often and perhaps more so than when I just lived there would say like you know what there was nothing that required us to be in the United States to build base camp but if you're looking to raise serious money go the venture capital route I do think there's a distinct advantage I also think there's a distinct advantage if you're in certain sub fields there are these clusters of talent and so forth that if you're looking to very quickly build a large business do help now I'm mostly on the side that like why do you want to do that um does it have to be really large that it has to be really quick what did you build a great business that only hired when it needed and it needed that much to get started do you need to be in the US for that no I don't I don't think so so I think it is a bit of a mixed bag I mean more than anything though I think it's a an adoption of a perspective and it's an adoption of an ambition and I was surprised to see that during my recent few years in Europe it was less like there was just something inherent and maybe it's because all the individual European markets are so small compared to the US and it's so difficult to attack Europe as a integrated market despite the fact that this is literally why we have the EU is to have a common internal market but there are so many little barriers one example is I've invested in a number of companies and one of the companies is a system that does time registering for employees and it needs to integrate with various systems in the United States you can integrate with literally like four providers and you will have covered 150 million employees in Europe it's all per country they all have their little different labor laws and they all have their little different registrations it's very difficult to roll out to all of Europe so it's very difficult to build I usually don't like that word but I'm using nonetheless scale because I do think there is a sense of scale that matters in terms of escape velocity can you get to a sustainable business for a lot of software companies I mean this is specifically looking at Denmark and that is one of the smallest markets in Europe it's I mean just under six million people you can look at other markets that are bigger and the problem is slightly different but it's just it's too small like the kind of size of businesses you can build and the amount of penetration you need in the market to sustain a business like it's so high in the United States the market is just incredibly vast it's hugely receptive to new technology to increase productivity in ways that European businesses just generally aren't if you're building software for example and you just have to sell to non-software businesses you're going to face a lot more resistance in adoption of technology generally speaking than you will find in the United States so not only is the market smaller in terms of literal number of companies and potential buyers it's also smaller in the percentage of those buyers who are willing to hear you out and this is why when I deal with the Danish companies that I've been investing in we talk about the American market as like the second or the third market yes you should bootstrap on your home market that just makes things easier you're still at a slight disadvantage to the companies who bootstrap on the American market that already starts with the biggest software market in the world right off the bat but you need to get there sooner rather than later now maybe their arguments are saying that actually in some ways in some niches the fact that Europe is such a fragmented and impenetrable market gives you some protectionism yes true maybe kind of but it also then shields you from the international competition that'll get you to be the best right I mean if you can hide away in some obscure market and I say that saying like Germany is an obscure market or France is an obscure market sounds weird but it is compared to the size of the American market I see that maybe you're not under the same stress that will arrive at the kind of software that'll conquer the world again the soft there's all software have to conquer the world no no it doesn't but would Europe be better off if it had more software and servers and internet players at the scale of Spotify clearly we would clearly it's a problem for Europe that Amazon Microsoft Apple Nvidia all of it is in the United States and uses China as its manufacturing base like what's left for Europe just to write stupid cookie banner laws man I hope not I hope there's a broader ambition there and I know there is I met plenty of European entrepreneurs who want something different want something better the scale is just very small and I think you have to accept that you gotta accept starting out that Europe really is 20 years behind how do you catch up do you catch up doing what you've been doing so far I don't know man it doesn't seem like it's working I would love to end on this note because then founders can think about what this means for them and their business so therefore David thank you so much for sharing all your thoughts experiences etc and I will link to your blog LinkedIn etc because then everybody can read and dig deeper into what you're talking about there and what you're sharing linking to your products etc because I really enjoyed it it was a great conversation thank you so much for being a part of the unicorn bakery thanks for having me on this is great