Manufacturing in the US, Building in Open Source… and more | E2154

26:10Right behind me, you can see the battery module and battery pack production line behind that. And you can see some of the white structure there. That's a overhead crane in the next line that's the chassis line and behind that is our motor winding line and drive unit production line so all the content is built here the batteries the electric motors the cabling systems the complete chassis the so the second part of your question i think is really what portion of this is u.s content yeah so it's about 50 right now okay which uh by automotive standards I would say is good by EV standards is really, really high.

26:49Most people have a ton of content from China. And then there's like an important corollary to that question, which is how much content do you have from China? And for us, it's under 20 % now, which is pretty exciting. How far could that number go down if you wanted it to in the next handful of years? I think it probably goes down to somewhere between five and 10%. okay it's possible to get to zero but it's sort of self-defeating the the biggest reason that we don't have such low chinese content is just because we're building stuff here yeah the battery pack in ev is about half the cost so if you are everyone else in trucking and you're buying battery packs from cattle like there you go you're at 50 chinese content right off the bat as long as you're not doing that um you know 20 is is pretty tolerable we'd like it to go down a little bit more to just continue to kind of reduce geopolitical risk.

27:43Yeah. I mean, that's, that's a material problem right now. And I think everyone has been watching the headlines has been perplexed by the back and forth. And I think with a lack of clarity, reducing your risk is just the very reasonable thing to do. So that, that absolutely tracks with me. I want to talk about growth. So clearly you have found something that people want because you had so many binding pre-orders last year. You just expanded into a hybrid EV chassis, adding even more range to your product line. So how did the company do last year? What's growth looking like this year? Maybe what's your revenue target for 2025?

28:18So right now we're building strip chassis. We started delivering those actual production units just at the start of this year. So this is our first year of commercial revenue. I hesitate to give predictions with the level of insane volatility we have right now, but something in the$100 million ballpark I think is on the table for this year for us. last year we were delivering vehicles but they were pre-production vehicles so this this segment is really defined by durability and uptime so we actually built three generations of vehicles before we started delivering any to customers and ran them through years and years of validation testing we were delivering those last year to customers to put into trials but not actually as full deliveries.

29:04Totally. From here, we're looking at primarily expanding within this segment. So we'll stay within class four through six for the next couple of years, but we'll add a cab. So right now we build strip chassis. We'll move into cab chassis, which is a broader portion of the market. We're also starting to get pretty exciting traction in our components business. Interesting. So a lot of people have realized that bringing in huge portions of your vehicle from overseas is probably bad. Buying half the vehicle from a Chinese SOE is probably bad. And so they're starting to look around and say like, hey, where can we buy compelling battery packs and drive units and HV systems?

29:46And SOE is a state-owned enterprise, kind of a public company over in China, owned by the state, or at least has a large state influence in it. Correct. Yeah. So the components business is super interesting because we were talking earlier about poverty of spirit and people not doing things and here you are doing them. So does that become, do you think in time, a material portion of the Harbinger business? We all understand the importance of a crisp, memorable, easy to spell domain name, but let's get real. The good ones, they're all taken or some poacher is holding them ransom, waiting for some crazy unrealistic payday.

30:21You don't want to use all of your startup's runway on a domain name, no. You wanna put that valuable cash back into your actual startup so you can build products and features that folks love. And there's this beautiful domain extension that everybody's starting to use. It's called.tech,.t-e-c-h. You can get a clean, crisp, super memorable name for your website right now, and you'll signal to all of your customers' investors, hey, we're a tech company, instant branding. That's why over 500 ,000 founders have collectively raised over$5 billion in investment while using a.tech domain name. So skip all the hassles and head to www.get.tech.twist or go to your favorite registrar like GoDaddy or Namecheap and grab your.tech domain today.

31:07I think it becomes material. I would expect it to always be less. I think it would always be a smaller portion than the vehicles. My expectation is that the components business can potentially be higher volume, but it's lower margin. So we want to keep that balanced. You know, if you have a buyer who's paying a ton of margin on components, that's like a temporary win because inevitably that means they're going to not have a compelling product at the end user. And so eventually they will go out of business. So effectively, you can't have too much markup on components because then the end result will be uncompetitive and then you're just selling a little bit and then they die.

31:42Okay. Exactly. That makes sense to me. Now, on the price point, you guys have an interesting, it's called the IRA risk-free guarantee, the Inflation Reduction Act. And essentially, whether or not you guys say the IRA's 45 watt credits stay or vanish, you guys will kind of make that good. And we don't need to get into the technical details, but policies have changed in the last couple of weeks as we sit here together. So I'm curious, has anything in the recent government spending bills, regulations and tax credits created a material headwind for Harbinger? Or are you guys going to be able to navigate the new kind of fiscal climate here in the United States without a lot of disruption?

32:21I think it's a little soon to say, but so far, I think we expected that the IRA credits would be a bigger tailwind. And it really wasn't. So I don't know that it's that much of a headwind. It's just kind of a return to neutral. Ever since the election, most of the customers immediately said, like, we don't think these tax credits are real. That's, you know, they are real still until September 30th. Right. $40 ,000 seems like a real amount of money to me, but it's been clear all year that customers are not really valuing those credits. So the withdrawal of those credits, I think it has a fairly low impact.

33:04Okay. Well, that's great. I was worried because I was thinking to myself, like, here's a cool company that's, you know, raising money, moving up the products, getting out into market. It would be terrible if the rug was pulled out from underneath you at kind of the last minute when you were hitting your stride and starting to get commercial traction. So I'm really glad that that's not the case. But I want to flip it around. You and I are both fans of electrification of both commercial fleets. And I'm also a fan of personal EVs, just because I think they're a great way to approach the market. What should we be doing as a nation to more quickly transition our various fleets of vehicles towards EVs, towards cleaner energy and so forth that we're not doing?

33:44So what can we do to go faster? We have two goals, which are in many cases competing. We want to encourage more electrification and we want to encourage more domestic manufacturing. Those are both valid, important goals. and 45W, you kind of had to pick one or the other because it was purely encouraging electrification and it didn't have enough defensive US manufacturing because a lot of that money was just going back out the door to CATL. What we should be doing is we should go back to a similar tax credit, but we should make it much more targeted and say we want electrification, but really only if you're supporting US manufacturing.

34:25that seems incredibly common sense to me so i'm not shocked that it didn't make it through a budget reconciliation bill in the united states congress but yeah we wouldn't want any common sense just before i let you go john you guys are doing a lot in the united states it seems possible you previously worked at andrel i mean is american manufacturing coming back at a real pace here or am i just seeing a couple of highlights around the nation that make me feel excited, but might not be indicative of actual progress. I think American manufacturing is coming back, but people need to better understand what manufacturing means.

35:02Manufacturing is what's happening here behind me. We're taking parts and we're making battery modules, but this isn't a steel mill. We're not turning iron ore into steel. We're in a great position where we get to do the things that have the highest value add and the lowest health and safety risk. There seems to be a lot of government angst around like, my God, we're not making enough steel. And I just, I don't know why we want to do that. It's not a job that people want to have. It's, I mean, the government is also unhappy that we're not mining more coal, which is another kind of a death sentence of a job.

35:39So I guess we shouldn't be surprised, but, you know, the work that we're doing here, the work that we were doing in Anderle, that's super high value add work that doesn't have to involve like how manufacturing was designed in the 19 or defined in the 1920s in the children's book. Yes. Well, I think you're a vision for the future, which I can see it. I'm watching literally your staff go back and forth. No one is running. No one is on fire and everything seems to be going quite well. So it's a future that I can get behind and totally support uh john where can people find you on the great internet and also what is a job you are struggling to hire for um harbinger motors.com as well as all your common social channels linkedin twitter instagram etc we're generally hiring for a lot of positions at the moment we're doing a big expansion of our sales team so if you're someone who has experience selling trucks selling capital equipment we'd love to talk to you all right well thank you for coming on and And when you guys eventually do break into the small consumer electronic truck segment, which I'm sure you'll do down the road to make me happy, we'll have you back on.

36:45But in the meantime, John, thank you so much. And keep building here in the States. Sounds great. It's great to be on, Alex. Thanks.

36:53Hey, everybody. Welcome back to Twist. Today, I have an amazing Twist 500 interview for you with a company we have talked about on the show recently. But first, I have to rewind the clock. Often when we talk about up-and-coming startups that are out there changing the world, they're backed by well-known venture capital companies, your Andreessen Horowitz's, your Indexes, your Sequoia's. Sometimes, though, an investor steps up to the plate that we don't expect. And in the case of PostHog, they raised a large round from Stripe. Yes, the well-known payments unicorn. That caught my attention. I then dug into the company and discovered that it is one of the coolest firms out there today, one of the best startups, I think, in the world.

37:27So please join me in welcoming co-founder, Mr. James Hawkins. James, how are you doing? Hey, very well, thank you. So you guys raised$70 million from Stripe. at an announced$920 million valuation. Before we dive deep into product analytics and open source and pricing plans, how the hell did that deal come together? It all started with Twitter. So basically a while ago, I think it was in 2023, Patrick Collison liked our website and said he thought it was cool on the internet. And then things really spiraled. So here we are$70 million later. We've been very, very efficient. We're a product-led company.

38:04we'd been averaging like a two month payback period for example for an awful long time I think we had gotten kind of complacent where it's like hey we can pretty easily grow pretty fast and we didn't really know how to spend money and it kind of took us until about our fifth year of existence to really know actually we can increase our growth rate by spending more and we want to trade off a little bit of efficiency for more speed basically so yeah like change of attitude like being around the right people at the right time and having a cool website with the three ingredients I love that this is just evidence that design actually really does matter.

38:35And it totally got me. I've never read so many corporate blogs in a succession on purpose than when I was on the PostDog site. All right, let's start with what you guys do. Now, when PostDog was initially put together, product analytics was the product remit, and it had an open source core. You guys have expanded greatly in the intervening years. So tell me and everyone listening what PostDog builds today. Sure. So PostDog provides customer infrastructure. We simply aim to ship every single product that relates to customer data, today we have about 14 products that relate to mainly kind of product engineering or growth engineering use cases like product analytics obviously but like a whole list of other tools that are adjacent like session replay error tracking feature flag experimentation data warehouse this goes on and on and what we're trying to do is instead of integrating a bunch of tools together we're like no we are literally just gonna freaking provide all of them um so there's perfect first party data um and as ai has come crashing in um to then just automate everything um very long term and I'm hoping we provide business autonomy.

39:32It's kind of the end game for us. Go ahead and define business autonomy for me. We want to automate every process that happens across every major function that relates to a customer. So then product autonomy would be a subset. This is like, hey, today you tell Cursor or Windsurf or whatever, like, hey, this is the thing I would like you to build for me, please. You wait 10 minutes, it comes back, you edit it a bunch of times. But you're still deciding what to build as an engineer based on maybe what a program manager told you, based on like your own knowledge of your customers, whatever your process might be.

40:02We are trying to get to a state where we've built AI already that's like chapter data, as you kind of expect. We're starting to get into the bit where you can do things with it. Like, hey, these are actually the things you should work on. And then longer term, I suspect it'll be like, hey, you fire up cursor in the morning or whatever else. It's like, here are the 10 things that have been built overnight based on a really deep understanding of customers. It's like multi-dimensional. Like it's not just the data you get from event tracking of just the data you have from errors. It's really every color to understand the painting, basically.

40:31I actually, I think it is going to be possible to build a better product manager than a human product manager with better understanding from just how effective the models now are, are really enabling this when you have this wide range of data. So when we started the company, everyone was like, enjoy the good old days when it's two idiots and no one else to worry about. And now we're like over a hundred people. But it's getting, we've had the complete opposite experience. It's gotten much more fun as we've gotten kind of further along. And I think partly that's thanks to the changes in the tech landscape too.

40:58So just to kind of just that down, you guys started with product analytics, have expanded to essentially anything you might need for product-related work. You've added AI in now, and then it sounded like you're describing essentially some sort of AI agent backend that's going to take on tasks for people that are building products. How did you get from working with your co-founder at a financial technology company to founding this, to then building essentially AI-driven business automation? It's quite the leap. What was the initial spark? My last job, I wound up as the VP of sales the first time in my life.

41:29And I was selling compliance software to large banks for millions of dollars. I would wear suits in inappropriately hot climates. And I would be going up into skyscrapers, trying to get people to buy stuff for a couple of million dollars. And I just had this feeling that this approach we're taking feels like it won't, doesn't feel like the future. It feels like the past. And I could see vendors starting to build enterprise software that had been more focused on the end user, basically, where they go to market in enterprise. It's like, oh, 200 developers insist on using GitHub. And now the procurement team has to buy it, even though they don't really want to, for example.

42:02And so I could sort of feel this. So my co-founder was quite simply the fastest engineer I'd ever met and the best, in my opinion, kind of therefore. So this actually brings me to the open source point, because going back to the history of the company, you guys discussed how having an open source posture of building and releasing software helped you get early attention to the company. I've often heard about open source as a great business model, less as a promotional technique, but was the open source starting point critical to finding that product market fit that early? Yes. I think it was a constraint that we faced was there's already a bunch of competition in, there's a massive amount of competition for what we're building, like in product analytics i think at the time there were probably three companies doing 100 million run rate or more uh like there's like amplitude mixed panel heat have all been pretty successful to some degree or another um the answer like well we can't do cloud um that's off the table already and we kind of i am i am technical enough to be dangerous i would say so i've got like a bit of a developer background beforehand my co-founder is extremely technical um but he can do the other he's very on the commercial side as well but um we kind of felt like we really want to build this for developers and what's the ideal thing for a developer and we kind of we started talking to a bunch of more technical people about their analytics stack we discovered a lot of companies were building self-building a lot of things to keep the data inside their infrastructure so we set off on a real zig before we then zanned across actually we can do cloud if we get multi-product at the same time and so we had this very windy journey but one of the quotes we had early that has i've understood later from ali waghani who's been on board who was the person that ran the continuity fund at yc when that existed back in the day um and he said when you get bigger you can kind of see around corners and I'm like what does that mean so now I know what it means it was like actually once we have lots of users we can figure out what to build properly and so postdocs never felt like we've never kind of descended down from the hillside and gone you know what like this is the future it's going to be this this this this this this but the reason that we have a our strategy is working now is because we had the wrong strategy but we spotted some things that were working haphazardly along the way and then kind of iterated our way across to something that worked great.

44:06So it's been a very chaotic looking journey from a product perspective, I'd say. You guys decided what to build in your roadmap in part based on customer and user reactions via like upvotes essentially of certain elements of what you're planning. Have any of those ever led you astray or is that how you see around corners kind of on a day-to-day basis? Yeah, it's a love, hate. It's an input into the process and it has, I think we, sometimes it can be painful, but right now it's painful. We want to build a CRM support platform. We want to do what we're doing for the products for the entire stack of the tools for sales marketing and support teams too because it's same customers they're dealing with it's the same data um and so we're like our top feature on our roadmap is like product tours which or like we shipped an error product recently it's been really popular and now we're getting asked for like a logging product and we kind of know full well like these are getting lots of users asking for them they'll make a bunch of money they'll grow quickly and we don't really want to build them we want to build like so well if we can get a crm being adopted by developers before anyone else exists in like a yc company for example we can build comics 10 times larger um and so like okay we'll like to get from this maximum to the next one we have to walk down the dip which is what we're experimenting with and we can always go backwards and up the hill um if we don't like it um so yeah we kind of um currently we're in a mode of like lifting the fog of war where it's like okay let's just see like actually how wide we can get away with here um is it as wide as we believe it is and that's what we're validating currently so it sounds like there's not one single answer it's constantly listening to users looking at the market making decisions and then also seeing what actually generates revenue for the business and then try quadrangling between those elements yeah the constant i would say has just been quick to do these things um and just getting them out and seeing what happens um it's also a little bit based on the engineers we have like if they are excited to work on something as like if i'm super excited or like there's an engineer postdoc who's super excited we'll build it literally three or four times as quickly um and yeah it is really it's very much an art form i would say for us we tried to put frameworks in place for this um but they haven't felt optimal when it actually executed them and stuff so yeah the core thing really is like we just are very willing to embrace a little bit of chaos here um and then we'll watch intently what happens um and we're trying to reduce the cost of errors like we won't spend a year trying to ship new products so for example we don't really target very top-end enterprises hard like we have a bunch of them as customers um but our bread and butter is signing up like hundreds of thousands of companies that are a little bit smaller and that sort of means that you can target releases you can be more nimble because it's not like we have two massive enterprises breathing down our neck to achieve some roadmap item.

46:31We don't offer any roadmap updates to customers. We don't have any deadlines internally. There's no coordination, which the whole company is designed around engineering velocity, basically, which is actually the thing that really does matter to us more than anything because we kind of assume that we'll be wrong a lot. And so for a quick, we'll find good stuff and bad stuff pretty fast. That's really interesting. On the going too wide point, the first thing that struck me when I was getting to know Postdoc was just simply how many different products you guys have. It is a wide number of things and I presume it's going to go up.

46:59is there an upper maximum to the number of individual things you can have in the market and maintain well and host and support as a business and are you getting close to that point or is there a lot more room for you guys just to keep building new stuff that people ask for it's worth a million times better than you would expect um like it doesn't like you said like what are the number one like what are the most painful parts of business um managing like 14 products at once whatever it is depending on what defines the product um it's just not on the list um and these products will support tens of thousands of customers um we're just not finding it hard to do that well and we hire engineers who literally decide what to build um but they don't have product managers determining the roadmap so i think because we're willing to let go of control we hire people that we frame as product engineers internally has made it feel ultra scalable that way around there are certainly other issues with having a lot of breadth um like we have a big debate going on the moment about um when do we create a platform team for example we'd be really hesitant because worrying about creating coordination between teams or suddenly having some meetings and stuff.

47:59Yeah, it's not always easy, I would say. But it does feel to me like we could, like our goal is like 50 plus products minimum, potentially a couple of different, depending on how we get on. But I think it's going to be very possible to do that. And there is Preston, like AWS, for example, has hundreds of products. Oh, absolutely. They are a little bit bigger than us. But even like, we actually looked at Y Combinator, We're like, well, they have 250, whatever, like roughly 250 companies in a batch with only 500 people shipping them or whatever. And there's 250 totally distinct businesses being managed and optimized to a certain extent.

48:31And so we're looking at kind of how YC is structured initially and like why are people so productive in the batch? And then we're also looking at AWS as like a very low stage example of this. So we think AI is massively accelerating this too, which is like needless. We already thought this was happening. Like open source software is more available. It's going quicker to develop stuff anyway. So we already felt like software was going to get kind of commoditized. And I think AI is putting the front of the gas here and we're like, well, like rather than hiding from this stuff by going into enterprise, we should just drive that change because it's better for users.

49:01We're circling something interesting here, which is with very small teams, nimble engineering and a lack of centralized road mapping, you can move very quickly and build stuff that a lot of people want to use. I'm actually familiar with several other companies in the product analytics space. I've been talking to Amplitude since before they went public. So I have a little bit of insight into them and their market and their growth and so forth. I'm just kind of blown away by how much you've managed to barge into markets that are filled with, you know, incumbent companies, often public with lots of money and time and history behind them.

49:31And you're able to grab to grab market share. What makes that possible? Is it that their products are just too enterprise focused or is it you're just engineering velocity that allows you to show up to so many different parties and walk out with a punch bowl? I think a standard playbook that I think is getting outdated now is you build a product, you get a little bit of traction, and you build like a 10k a month business, you go raise like your seed round, you do a series A, and then you're like, man, and then you start getting serious. You're like, okay, I need to keep up my rate of revenue growth.

49:57How do I do this? And I think the easiest, like most intuitively possible way of doing that is, I need to add more salespeople and go up market because the order values will jump like 10x or something. And we're kind of more competent at this stage, you can handle these enterprises and whatever. we built and people companies have all and fast forward and it's like you know after 20 years of sass and like vc coming in here and it's like cool there's now what i would turn a like fragile mess of tools in enterprise and so i think our opportunity is like well we think it is harder but we won't grow as fast if we like we could right now just go hey we're just got a whole bunch of sales people that are very enterprise focused we're going really upmarket we're not going to invest in going wider we'll just invest in depth of each product and we'll take that little toolkit we've already built upmarket and we probably get to 100 million quicker if we did that.

50:38But I don't think we'll continue to compound. And so we've just sort of been willing to like, yeah, we're going to lean into engineering velocity above everything else. We'll do some really counterintuitive things there. But it's kind of almost a constraint of going, well, if we went on market, we're just going to lose. Like we're never going to beat Datadog today at enterprise sales or something. Kind of a dated paradigm, really. Yeah, and I'm like, engineers exist first. Like they exist when there's two of you starting the startup from scratch. So if we just go with engineers, we're upstream of every single piece of software that company will ever use.

51:04So if we can just offer it all, then they'll buy all of it if we're the best vendor. And then we keep doing a good job of looking after people. So there's a few factors that I think that have meant that we're successful. If I'm being harsh on myself, it's because we can't invent something. We've struggled to be, I haven't been getting up a product to invent something totally box fresh that has zero competition or whatever. Like maybe that would have been better if I could do it, but I haven't been able to. We're now getting to the point where we are starting to innovate and it's proving a lot of fun, but I've had the chance to now hang out with lots of people and learn stuff and spend time with my co-founders doing the same thing.

51:30And now we're getting a little bit more competent at like actual, what I would say, actual innovation. that's kind of why it's a bunch of things at once but the main one is like we've sort of been forced into being different everyone else because the market is so saturated and which has worked out really well like i think conceptually it's easier to do a better job of something already exists than it is to completely invent something from scratch and so we're merely doing about what we think is a better job of something that already is out there so you mentioned getting to 100 million i presume you mean arr by that point and you said that if you had gone perhaps upmarket originally like the usual sass playbook you might have gotten to that milestone faster but you will get there you also said that you're compounding which i think shows that if you have a a wide base early as those customers grow they'll grow with you which makes a lot of sense to me but you also do some things that that do decrease your revenue which is an interesting thing that i don't often see uh one thing you do you have side hustle insurance so if someone blows through usage caps um you don't take their entire house you guys have done actual price cuts which harkens back to the aws days of your back when they actually cut prices if people were kind of think back that far um so i guess my question is how do you balance between what i very politely call extreme customer friendliness and the need as a business to keep the growth coming because you do have backers and you do have people that are expecting a certain return from you the core one i mean the core tenant of our finance that i think is service really well and our core principle is um don't need to make ensure that we don't need to fundraise basically which means that we can take a long-term view on stuff so we don't have to rely on we don't have the dependency on like oh god like if revenue dips and have to fundraise at the same time we have to explain it to everyone and we'll lose momentum it'll be hard to do that um but when you don't really care because you're like well i just kind of fundamentally we're running the business efficient enough that we can just get profitability whatever we can take move make moves like this because we think they have a lot of upside if we screw them up it's like oh it's like you always just knocked us back a couple of months it still doesn't matter like we've just got to back ourselves that we'll be able to get it back um yeah just not feeling desperate is and like and that's partly why we've raised is um there's a huge psychological effect of um having the ability to take lots of long-term views on things and including like you know it comes out and also like the marketing channels you pick for example um and all the ones that are like really that work better in the short run like paid ads or something are worse for you in the long run because they're really saturated so um yeah i think it's you just can't be burning a huge amount you have to fundamentally pretty efficient business to get away with moves like this so that you can run yourselves in a good financial spot and you have to be able to fundraise in some way, shape or form.

53:56I treat finance as kind of like a, our revenue growth is almost like a constraint on the business where it's like, we need to go pretty quick, but like we will sacrifice it for more. Like, again, mentally it's like, well, if we cut your pricing today, you'll keep using us. You'll probably use us more, you'll retain better. You'll probably tell your friends more often about us. And then eventually you'll buy another 20, like in my head, I'm like, you're gonna buy 20 more products or we'll have 20 more to sell you. So it's like, it doesn't matter. You're like at 2 % of your total spend with us over the next 10 years, for example.

54:22course. Like we should really optimize for that. And that's how we're thinking about these types of changes. I would have said that they don't make our revenue look good in the short run, but they have like that change increased our revenue growth rate, which was so counterintuitive, but basically retention went up, usage went up and growth went up massively, which offset the difference. You can be cut pricing for a whole existing customer base for two biggest products, for example, by like an average of 60 % or something. What was the total hit to your revenue from that decision in the absolute short term, like day two, what was the...

54:53I don't know what it was, runway wise, but yeah, it was our two biggest products and it was 60 % average decrease in spend for all of them. I think it like almost halved the, it like halved the average order value of customers signing up, but we tripled the number of customers signing up. That's, that's, that's what I was going to ask, what was the impact of it, you know, day three? Yeah, it was skewed towards customers which had a lower, that higher margin, because it's towards the bottom end of our revenue scale, which actually means the revenue impact was like unnoticeable sorry the margin impact was unnoticeable in our numbers so we make a lot of margin on the bottom end of our pricing scale and like much less further up it was after like a long meeting where we were debating doing some ridiculously complicated like let's do a reverse trial where if you put your credit card in and this and then this and this and this then you downgrade them but in this i'm like man this feels like peak midwit meme if you've seen that where this pricing thing is so freaking it's like we're just trying to get around the fact we should be making it cheaper if we want to win and i think the other thing is we concluded was because we're efficient, we grow inbound, we don't have to pay for a sales team, we have super low cap payback periods, like I think we can actually sustain, you know, pricing battle quite well, because of this.

55:55And so we can kind of afford it in a way that like, if we were very dependent on high overhead to acquire the next customer and so on, we don't fundamentally higher operating costs, we can't cross sell it as easily if you just have one product. So we kind of felt well from first principles, we would win if we did a pricing war. And so we're happy to like, undercutting our competition. Yeah. Okay. So last question for me is this. I thought the core trick of PostHog was going to be open source because I'm a big advocate believer in open source software companies in general. But listening to you talk about how the company has approached the market and learned and grown, it seems like your lack of burn has provided the most flexibility for PostHog to choose its own path.

56:38So is that right? And then also, do you think that's a viable strategy that other startups should emulate. Yeah, I think that has been an incredibly important part of our journey. I think it's that. And the other one that's worked has been the whole reason we're multi-products, we found this alternate path. We built a successful, not venture scale, we were building a reasonably successful growing business off the back of self-hosting, where we were just chanting self-hosting customers, we're making money off this thing. It just doesn't feel like the business that we could be, doesn't feel like the best business we could possibly build.

57:06It's all right, we're growing fine, but we're spending we're doing so much support we don't think this is a great experience from a lot of our customers who are like because we have to provide lots of support they have to deal with lots of support requests all the way around as well and if and then like and some customers were self-hosting for no reason and we're like man if we just like why don't you toast it for them lo and behold we have now got a cloud product fast forward you're like wait the cloud customers are way happier than the self-hosted ones and yet loads of people are using it and like i don't know why they're signing up our cloud products it's just like kind of worse than our competitors ones but it was like well it's because it's a different audience and like because our branding and like we got away with like a lot of branding and developer just by being really good at handling developers and knowing how to run a company for them meant that even though the products were like very similar or else was even worse at that stage and we just got enough traction that then we could lean in and the other ingredient that was very important was the whole reason we're multi-product is because we let our engineers decide what to build and one of our engineers randomly built the next not randomly like he had reasons for it but he decided to build session replay as a second product and it was just popular immediately.

58:06Users had kind of asked for it in our repo and then went, oh, hang on, I can see how this works out. Now, if we stay a little bit more down market, we can go much wider, much quicker than everyone else. Because we don't have to pay for sales team, all our money we can just put into engineering so we can have a much bigger engineering team than a normal company at this stage. So actually we do have quite a lot of engineering bandwidth because we're not losing money elsewhere. So yeah, having kind of like low financial desperation or whatever has made it super easy to do cool stuff like this. and it also like now our marketing works like that too like if the we can just do like really weird brand marketing stuff that sometimes like you mentioned you put something on twitter and you get more exposure than if you'd spent half when we also do the billboards but like you get more views than if you spent like half a million bucks on billboards for example in one tweet that you spent 25 seconds to write um or like we have a newsletter we're building up it's got 75 000 subscribers um that's like it's new it's not been going on for very long but it's like really climbing quickly and we're like well this is not going to help us in this year or next year like but in five years this might be this is probably the most important thing we could possibly do in marketing if it works and we can sort of see that we're trending into that sort of direction so we're just continuing to invest in it and now when we hire candidates and stuff it's like cool we have inbound only recruitment because people already know who we are they're like learning from us um we've managed to get that thought leadership because we're just willing to invest in stuff that is going to suck on a spreadsheet for a long time basically so yeah that's been very important because it's made us the key is to like maintain urgency when you don't have like crippling fear of death coming.

59:32But yeah, like, you have something that's all much better than I'm about to repeat back to you, basically. No, you're all good. I really appreciate it. And if folks want to learn more, it's posthog.com. And James, where can they find you online? Probably Twitter's the best place. James 406. I want to get into why that's your username because we don't have time. But thank you so much for coming on, James. And when you hit product 20, come back on the show and tell me what the next four or five were. Cool, I will do. Thanks so much. Thanks, James.

1:00:02Hey, everybody. Welcome back to Twist. This is Alex, and we have another Twist 500 interview for you today. This one comes with a little bit of history. I first met Alpaca back when they were a small company raising their first funding rounds. And now here we are a half decade later, the company is scaling and growing, recently raised more money, and has a really interesting perspective and take on building fintech in 2025. So please join me in welcoming you to the show. It's Yoshi from Alpaca. Yoshi, how you doing, man? Great. It's been great. Like, it's been a long time, Alex. It's great to be here.

1:00:33It's, I ask that because it's 4.30 in the morning in Kyoto, where you are, and Yoshi got up way before dawn to come record with us today, and we're very thankful for that. It's Saturday where he is, so Yoshi gets 10 founder points. All right. So it's been a minute since we've talked, so I'm out of date, Yoshi. Why don't you just tell me and the folks listening what Alpaca does today? Yes. Alpaca is a brokerage infrastructure company and then also a FINRA SEC registered self-clearing broker dealer. You know, it sounds pretty boring, but what we do is that we build embedded brokerage product that a lot of fintech applications and then broker dealers and banks all over the world can embed and add investing and wealth services to their end customers.

1:01:19So the way that I've always thought about this is if you have an application that has a financial bent to it and you want to expand your product offering, you go over to Alpaca, plug into your API or sign a contract, and then you can offer equities trading, options trading, and a number of other wealth-related services to your product as kind of a white-labeled solution, Yoshi. Is that fair? Yes, that is correct. You know, we take care of all the mid-office, back-office, all the compliance requirements on our side, and also the reporting as well, and everything comes through the API. So the developers and businesses don't have to worry about backend stuff.

1:01:56It's interesting because when we used to talk about APIs, kind of back when Twilio was blowing up, we would talk about abstracting complexity away and then offering a single developer hook that provides a complex service that's very easy. and it's, I think Alpacas are just a really good example about how that still works because people still want to use API delivered products. It's not all AI today, you know, there is still money to be made, but that's actually where I wanted to start because I was going back through my coverage of your guys' I think 2020 and 2021 era funding rounds and you and I were going back and forth on how Alpaca made money.

1:02:31And one thing you told me, if memory serves at the time, Yoshi, was that you guys weren't charging much directly for the service, but we're instead splitting payment for order flow revenues generated by end user trading. So I'm just curious if you could give me an update on how the company makes money today and if you're still doing that revenue split on order flow payments. Sure. So at the end of the day, our business that we do, even though we say this is a brokerage API or brokerage infrastructure, at the end of the day, type of the business is the securities clearing and securities custody business.

1:03:06And since we are membership of DTCC, this is the really core semi-governmental entity that actually has all the custody of the securities. So we are membership member of that, which means that we take care of all the end customers, cash and securities on our book on the custody. The business model becomes nothing trivial. It has been something that existed for a long time. So all of those clearing farms makes money of the transactions interest as the kind of main things. And the inside of the transaction revenue, there is, you know, as you mentioned, payment border flow, specifically for the United States stocks, equities trading and options trading.

1:03:49There are certain market maker model that connects to the bunch of the exchanges that exist in the United States. And then they create a model that pays back some of the rebates that was not rebates, the money that's been built between the bid and offer price to be offered back to the broker dealers like ourselves, which we can split to the partner broker dealers that we work with. All right. Now, one thing I did notice that you guys have somewhat recently become a fully self-clearing broker dealer. I kind of know what that means. I'm pretty familiar with market making, but it might help just to hear it from you about why it's important that Alpaca is now a self-clearing broker dealer.

1:04:29And also, if you could just tell folks listening who aren't as familiar what a market maker is and why they matter. Yes. When someone buys, for example, Nvidia stocks, the things happens as the trade is executed, which trade execution means that we're basically deciding the price when this person A buys the Nvidia stock at what price. And then after that price is decided, actual exchange of the Nvidia stocks with the cash that person pays, that process is basically called trade clearing. So in terms of the first one, trade clearing, like companies like us who are behind the scene, there are multiple layers of the businesses actually that's built up on this financial industry system, because maybe while you're interacting with your app or businesses may not have the full integration into the deepest core of the booking records, which is the DTCC.

1:05:30The finance services is like, it's always relying on each other. So you outsource these licenses, you outsource these functions. It's really connected everywhere. So with that said, us being at the very much one-stop shop, having all the licenses and membership, it reduces the fees that we have to charge because we're not outsourcing anything to anyone. Does it also allow for tighter spreads when you offer a price for a particular equity? So in terms of the tighter spreads, it becomes more of the, yes, exactly, definitely. Because when we are pricing back to our partner brokers, who is going to basically charge back to the end customers, the fee that we are getting charged from those systems become the, the, the, basically the lowest in the industry, because if you own everything in terms of the licenses and then also the systems, and also this is, this is an interesting part.

1:06:26Most of the financial services company, like most of them, they outsource the system because they are not tech companies. So they don't build stuff, which means it creates a lot of overhead, the unit expense, which we don't have. So we can offer the tightest price, basically, logically speaking. And just to clear up an acronym, the DTCC is the Depository Trust and Clearing Corporation. That is correct. Yes. And you said you were a member of it. Does that mean you guys have like an equity stake or is it a partnership? How does that work? Yeah. So it's not like an equity or partnership, but there's so much like an application process that you have to file.

1:07:06So the DTCC basically allows us to be a member. Same as like becoming a bank or becoming a broker dealer. You have to prove and you have to show that you have all the compliance ready. You have all the risk system ready. You have all the right people ready and you have enough capital ready. So the DTCC welcomes you as the direct member that's connected to the DTCC system. All right. So what I loved about Alpaca at the time, back in the day, was that everyone wanted to have a broader suite of financial services. Once they got customers in, all these fintech companies wanted to offer more stuff to them because the more stuff they could offer, the more money they could make made perfect sense to me.

1:07:46Time passes. The COVID era, what I call the savings and investing boom, you know, kind of like the early meme stock era and Robinhood blowing up and so forth. We're now pretty far past that. So I'm curious if you could just give me a bit of an overview of the last couple of years of adding new customers from you guys and how they're adding end users. Just what's it been like in the last two or three years? Yeah, so surprisingly, and thanks to our teams working really hard, we've been more than doubling revenue last several years. I always do admire what Vlad and Bejir did in the Robin Hood. You know, obviously it's always being talked about and, you know, paid attention to, but, you know, they have been always kind of making a way, creating new things, like, you know, like, you know, really focused on like what comes next.

1:08:38I think like, you know, there is definitely the maturity that's happening for the B2C fintech applications. You know, maturity means that adding a little more service verticals. And then I think like winners are more becoming winning players and to becoming more of the enterprises. So having becoming more invincible in terms of the market. So that's Yoshi. Can I, can I jump in? So are you saying that because the upstart companies from a few years ago are becoming the incumbents? Yes. It's a good time for Alpaca to be out there helping to power a lot of these companies because essentially your customers are maturing and becoming the new leaders.

1:09:15Yeah, definitely. And then like, you know, we, I think like the reason why I wanted to bring up the example of the Robinhood is that there is always a continuous innovation on the B2C products. And then like, you know, the players who are good continue to add new things to continue to grow. So that's the kind of the proof point that I see from the Robinhood. And the second thing is that by looking at the Robinhood, there are like many other following Robinhood types of applications, not only in the United States, but also outside the United States. That's our maybe few years behind or five years behind or seven years behind, which means that the growth curve is happening in a very different diversified way in terms of the revenue, assets under custody growth.

1:09:58And that's our built on top of us. So like, you know, we see that growth like very diversified. So like that's why we see like very constant growth of our revenue and assets under custody, regardless of the despite of the situations that we see. So Robinhood does very well for itself. I just pulled up its stock chart. It's currently worth$64 billion, which is more than anthropic, shout out. And people around the world have seen this and gone, oh, we need to build the Robinhood for Malaysia and the Robinhood for Nigeria, et cetera. And they're a little bit behind the initial Robinhood curve, but if the curve still goes up, you guys at Alpaca, if you're serving those people, do very well.

1:10:34Okay, is part of the demand for Alpaca services the fact that people really want access to US equities in particular? Yes. Um, so 65 % of the market cap, uh, of the, uh, whole, uh, whole equity markets in the world is United States, uh, U S stock market. Yeah. And then, you know, obviously I don't have exact stats, but like, you know, computers people use in terms of that Mac, Mac or windows or Google, or even open AI or whatever. I think it has like much, much higher penetration, which means that like an 8.5 billion people on this earth probably has thought about the United States listed companies the most.

1:11:12So I think that is the reason why we started from offering the U.S. stocks. And then that has been our core asset classes. But last year, you guys added quite a lot. I was going back through your 2024 recap and also looking at your original blog posts. Things like options, high yield savings accounts, IRAs, fractional shares. The thing that I'm struggling with is what's left to add? I mean, do you remember when robo-advising was a big deal? Everyone's talking about Wealthfront and the other one that I now forget. That's now table stakes. So, Yoshi, how much more stuff is there for Alpaca to build that's like net new in a feature sense?

1:11:48Yeah, yeah. So, I think like, you know, it probably goes back to, again, you know, how the Robinhood is adding stuff as well, right? And then a lot of companies in the world, even the fintech companies, but also big companies who are launching a bunch of digital banks and digital arms are looking at what the Robinhood is doing. Because, like, you know, they are best at growing the customers, best growing the businesses. So, again, it's not only about the stocks and options. There's the saving accounts products and even fixed income products. There's money market sweep. And then, of course, there's a prediction markets and then investment advisor products.

1:12:23And this is really showing the shift of the money from just speculation and trading into the more the long term saving and wealth management. So I think it's now coming back to the core thesis of the money is coming from the boomers to the new generation and how we're supposed to cover more of the cash. And then that's, again, happening everywhere in the world. But there's always three years, five years, seven years, 10 years behind. So we just need to keep building those infrastructure for them to make sure that they can do what they need to do, like Robinhood has been doing in the United States.

1:12:57So essentially, Robinhood is not just the millennial brokerage of choice. It is an indication of new generations of humans demanding different services than their parents, effectively back to the boomers point. That sounds pretty bullish for Alpaca because it means that the companies you're serving that are offering your equities and other trading services to their customers have a lot of growth ahead of them. So I guess the question then becomes amongst your more mature customer set, people you've had for a couple of years that are not new, are they still growing very quickly or as quickly as your your more recent customer additions?

1:13:32Yes, yes, definitely. Like, you know, the growth curve of the existing customers are higher. And then, you know, those guys, those guys like. So we work with 220 applications in the 40 different countries. like, you know, the benefits of the big banks having the new offering, which, you know, we serve the infrastructure for as well. Definitely the first period of the growth is really high because you can cross sell it. I think that's another thing that we have been seeing is that, you know, we now work a lot more enterprise customers, including legacy banks and then legacy incumbents broker dealers because like they're also fighting against competing against new fintech players and they have to like you know make sure that they have to grow against the new competitors that's coming up so we're serving both segments i was kind of working my way on the feature point to ask about stable coins because i feel like everyone's been been yapping about them yeah and people want to have access to them they're not a trading thing as much as equities they're not really an investment that you hold forever, like money market or bonds or so forth.

1:14:40But do stable coins come into the alpaca world at all? I'm just not sure if they have a fit, but I just wanted to ask. Definitely, like, you know, stable coins do come into the play because in terms of the money movements, you know, we have to deal with the traditional rails of like moving money around. For example, a customer in Saudi Arabia buys, you know, the Microsoft stocks and then he comes to the United States because in order to purchase the stocks, the money has to move. So there's a bunch of the correspondent banks and all those payment rails. So when that stable coins happen, definitely like, you know, the money movements faster, even though we have to on ramp it or ramp it.

1:15:16But there's definitely the future there, in my opinion. So we do definitely follow very, very closely what's happening there to make sure that we can provide the best service for our partners. Would you consider building your own alpaca USD coin just for your own internal settlements? Because to me, thinking about the example of someone purchasing Microsoft stock in Saudi Arabia, that is far away geographically. It's far away in currency terms. It's not close. And so if I was you, and this is maybe ignorant, I would think having my own in-house stable coin would allow me to effectively take money from my right hand and give it to my left and skip all those borders and other sticky issues.

1:15:55Is that how it would work? Or am I oversimplifying to the point of stupidity? I think there are multiple takes that how the future will be in terms of stable coins. I think some people say that it's going to be predominantly two or few stable coins. I wouldn't say even use the word of stable coins. It's going to be with a stable dollar, in my opinion, as a concept. And some people think about building the infrastructure so that so many people can build a bunch of the stable US dollar, like One Money by Brian. I think there are multiple futures possibilities. But what I think is that both of them are actually right.

1:16:34I kind of think about it more like ETF, right? So if you see Bitcoin ETFs, there are a bunch of them. And then in order to create better experiences, we need to lower the entry hurdle. So we need to create the infrastructure for that EDF, those stable coins to be built, made easily. So I think both things have to be happening at the same time. But for Alpaca side, we haven't really decided in a way. We're definitely looking at it, what is the future? But our focus and priority is to solve the problem. So if the existing environment allows us to solve the problem faster, cheaper, quicker, we'll take that route.

1:17:11If it doesn't, we have to build our own. So it really just comes down to how good is the offering out there and could we do better internally? Okay. Well, when you do make that decision, let me know. Cause I want to ask you that question again, once you've made it, because I think it'll be illustrative for other, well, I'm just thinking about other founders who might be thinking along similar lines and you guys are, I mean, you're a series C company, you're pretty mature, you're well-established. And so I think there's probably a good lesson there for when you make that choice. All right. On that theme of talking to other founders.

1:17:41When we were talking back in 21, FinTech was the hottest thing. We were joking before the show that it was, you know, the AI of its time. And then there was, you know, Zerp ended, everyone freaked out. Suddenly startups had to be profitable. It was all very strange for a while. And now with Circle, eToro, you know, Chime, FinTech feels very hot again. So you've gone through a period of time in which your company was out of fashion. What changed operationally for you because fintech lost its luster? And any tips for founders out there who might be building today in a market that is considered less sexy than AI?

1:18:16I think market cycle always happens. The concept that I always like to think about is the Bridgewater's Ray Dalio continues to talk about short-term debt cycle and the long-term debt cycle. So I want to always like, you know, the kind of step back to see what's really happening in the world and the bird eye view so that I don't freak out, even though I'm freaking out every moment about revenue and growth and client and all the things. So my team members probably say, what are you talking about, Yoshi? But I think like, you know, for definitely like there's always a cycle. And I think what really important is that we have to, in my opinion, react as quickly as possible to what's happening in the world.

1:18:55because it's impossible to predict the futures. But what we can do is that we can react as quick as possible to what's happening to beat out our emotional barriers or whatever that stops you from acting. So I think that's one key component, I think. And the second component is that these things actually allow us to grow and mature us extremely in a good way to be very strong and very organized of how we run the company. So I always think about those two things as the benefits going through the multiple market cycles. So essentially, because you had to go through a period of time in which investors were no longer showing up in a car with just briefcases full of$1 ,000 bills, you had to become a more mature company, probably think about cash a bit more seriously.

1:19:41And so now that, you know, Fintech's back at it, I mean, your Series C was, I know, announced this year, actually happened last year, but with these IPOs, I presume you'll have more investor interest and so forth. Now you are much more stable and just healthy company. Like, is it my read of what you're saying? Yes, yes, definitely. Like, I think in addition to the healthy, it's more predictable as a business and then more visible and transparent as the business. And I think that's what matters, in my opinion, because if there's a visibility and transparency of what we are doing, we can we can like we have many choices.

1:20:16What levers that you pull to expect certain outcomes? And now it's a very nuanced and balanced situation. And I think that's really the role of the CEO that we have to navigate which direction, which nuances that we have to be taking at that moment. Yeah, you have to steer the ship and not worry about the little waves, but keep an eye on the bigger waves. All right. Oh, one last question, Yoshi, before I let you go and either go back to bed or have a very early breakfast. We haven't seen these FinTech IPOs and it's gotten me very excited because I love an S1. And I would put Alpaca in the not going public this year, but maybe the back half of 26 or something, because, you know, you've been doubling for a long enough time now.

1:20:56You're getting big. Yeah. But at the same time, there's also been a kind of an uptick in M &A. So I'm just curious from your chair, are more people reaching out, kind of testing the waters to see if Alpaca wants to exit early? What's it like out there for a founder of a mature and healthy FinTech startup? Yes, definitely like a lot of inquiries and the conversation that, you know, we receive. uh and then i think um it's one thing is that fintech getting the traction but also uh multiple uh industries of the fintech are uh getting traction and getting more capital also in the in the u.s of course like you know robin is adding more products and uh you know becoming even mature and like in the brand name and and i don't think that's only the case only in the united states like if we look at the europe there's like you know of course revolut and monzo and 26 and Klarna and, you know, all those guys are also becoming even more winners and adding new things.

1:21:49So that means that those companies are even growing faster. So there's more cash and opportunities they're looking for. So in addition to the fintech as a general, as an industry, but also like each of the key players are also becoming bigger. So I think like they're also like looking at the opportunities, how they can also grow faster. So I think like this trend will continue. and you know, of course, like I'm friends, I try to be friends with everyone to try to understand what's happening in the world so I'd love to be talking with as many people as possible always. All right, well, you heard it here first, no buying alpaca until they go public.

1:22:23I want to see the gosh darn documents. Yoshi, thank you so much. Let me know when you're done with the stable coin choice and also I want to talk to you again at the end of the year just to go over the growth rate and the metrics to see how it all wraps up because I suspect it's going to be quite good. Thank you very much and it's alpaca.markets. Yep. Yes, that is Christalapaka.markets. All right. Thank you very much, man. Talk to you soon. Thank you, Alex.