Chapters
Shownotes Transcript
Hello, acquired LPs, and welcome to another great episode of the LP show. David, this was a delightful interview with Ben from Composer.
So, so delightful. We covered so much cool stuff. The evolution of the whole Wall Street bets community. And I just love like, it's like, that's so great that like, yeah, that was like a thing. But some percentage of those people are really smart and are going to become great investors. And that's what Composer is there for. So that was such a cool part. We talked about Packy, our mutual friend, and Composer being the first not boring capital investment memo. Yeah.
Yeah. And also, Ben has like a very interesting philosophy on how to run a company and how to go about building products. So it's interesting for most listeners who don't know what Composer is. It's a no-code tool for private investors to create your own hedge fund-like strategies. And it was fascinating sort of hearing about the origin story of how he arrived at that and then going deep also on, you know, all the Wall Street vet stuff and understanding market making and dark pools. And if you want to better understand that world, it's a good episode for you.
We want to thank our longtime friend of the show, Vanta, the leading trust management platform. Vanta, of course, automates your security reviews and compliance efforts. So frameworks like SOC 2, ISO 27001, GDPR, and HIPAA compliance and monitoring, Vanta takes care of these otherwise incredibly time and resource draining efforts for your organization and makes them fast and simple.
Yep, Vanta is the perfect example of the quote that we talk about all the time here on Acquired. Jeff Bezos, his idea that a company should only focus on what actually makes your beer taste better, i.e. spend your time and resources only on what's actually going to move the needle for your product and your customers and outsource everything else that doesn't. Every company needs compliance and trust with their vendors and customers.
It plays a major role in enabling revenue because customers and partners demand it, but yet it adds zero flavor to your actual product. Vanta takes care of all of it for you. No more spreadsheets, no fragmented tools, no manual reviews to cobble together your security and compliance requirements. It is one single software pane of glass that connects to all of your services via APIs and eliminates countless hours of work.
for your organization. There are now AI capabilities to make this even more powerful, and they even integrate with over 300 external tools. Plus, they let customers build private integrations with their internal systems. And perhaps most importantly, your security reviews are now real-time instead of static, so you can monitor and share with your customers and partners to give them added confidence. So whether you're a startup or a large enterprise, and your company is ready to automate compliance and streamline security reviews like
like Vanta's 7,000 customers around the globe and go back to making your beer taste better, head on over to vanta.com slash acquired and just tell them that Ben and David sent you. And thanks to friend of the show, Christina, Vanta's CEO, all acquired listeners get $1,000 of free credit. Vanta.com slash acquired.
And lastly, as I'm sure the good folks at Composer would want us to let you know, none of this is investment advice. We may have an interest in things we discuss on the show. This is for informational and entertainment purposes only. With that, on to the interview. All right. Ben. Ben.
I love it when we have Ben's Davids on this show. Ben R., welcome to the Acquired LP Show. CEO and co-founder of Composer. It's awesome to have you here. And you have the noble honor, I think, of being the first composer
not boring portfolio company, right? That is correct, I think, right? Braintrust? No, I think that was later. Oh, not to be on the LP show. You mean the very first investment from Not Boring. Yeah, that's correct. That's correct. No, Composer, I believe, is the first investment by Packy's Not Boring fund. He was the first money into Composer as well. So first, first on both sides. He was the first money personally and then...
You were also the first Not Boring portfolio company. Amazing. Yeah. Yeah. That's probably a good place to start it. So speaking of Not Boring and our friend, Paki McCormick, before our episode here, I texted him and I was like, all right, what should we ask Ben? We have him on the show. And he said, maybe you should kick it off by asking him about doing venture in the Philippines, which I don't know a ton about your background. I only know about composers. So what's going on there?
I've done a lot of wacky things. So first off, I'll preface that I'll be totally open and at this point, not even particularly insecure about the fact that I don't have like super polished blue chip credentials to get my foot in the door. I've done a lot of like pretty wacky things. So at the time that I left for the Philippines, I was working for, frankly, a pretty janky startup in Montreal. I got this opportunity serendipitously through like a connection of a connection that was like,
well, if you have ideas, you can pitch this VC fund in the Philippines. So I had an idea that at the time didn't really make sense, but I flew all the way to Manila and
to pitch this idea to this VC fund because at that time there were no VCs you could really pitch there. So the motivation was funding to do this? It was funding actually. Yeah, it was like, you know, whatever. And it just sounded like a wacky thing. I was like, they had opportunity. They're like, why don't you come, you know, pitch that we're incubating some things. You have an interesting profile. Why don't you come pitch us? And they were interested because I was like, okay, well, like we don't normally get people from like Montreal. So why don't you come by? So I flew in and pitched them and they were like,
We like you. Your idea doesn't make a lot of sense, but why don't you just come work for us?
And I said like, okay. So I moved to the Philippines and I like lived in Manila for almost a year working in venture there. I knew it wasn't going to be a long run thing. They knew it wasn't going to be a long run thing. Frankly, like I made like very little money. It was not like what venture you associate with venture. Like, you know, it was, it was a relatively junior role, but it was like, it was so, so fun. And the interesting thing at that time, that fund was very small. It was called Kickstart Ventures.
And it was like a subsidiary of Globe Telecom, which are the big telco companies there. But it was very small at the time. It wasn't much money that they were managing. Now they're actually super successful. They have like multiple funds. They're doing super well. They're one of the biggest venture funds in Asia. And I learned a ton from the CEO of that venture fund. What's the name of it again?
Kickstart Ventures? Kickstart. Huh. Unrelated to Kickstarter. Everybody goes, that's so cool. You're working for Kickstarter. I'm like, no, no, I'm working for Kickstart Ventures. It's a VC fund. The CEO of that venture fund, Minette, I think she is the best boss I've ever had. She's like the most capable person I've ever worked for. She started really from nothing. She was like managing malls at one point and then she kept rising up, rising around. She's one of the most impressive people I've ever worked with.
I learned a lot from her, actually. I mean, the people there were brilliant. And the companies there were super interesting. Superstar for capital. The check sizes at that time were tiny. Things have really changed since then. Like check sizes could be like $50,000 for a sizable percentage of a company. And I really saw there. It's so wildly unfair in a sense that like where you were born, where you live affects so much. And the funny thing is I was coming from Montreal, right? And like Montreal, if you want to raise money in Montreal at that time,
The valuations, and they're still lower. There's still a discount, right? Like-
At that time, they were like a fifth of Silicon Valley, right? Like the valuation if you're raising. And it was like just local sharks who would fund you at that time. Like, you know, he owns a nightclub and a restaurant and he's going to invest in like, I don't know, an app for Montreal or whatever. And then I go to the Philippines and like the valuations are even lower, like considerably lower. And salaries are like a 10th and costs are lower there, but costs are going up everywhere in the world. So it was pretty fascinating. I think things are very different now. I should check in. What time period was this?
This would have been like six or seven years ago, right? At least or more. Yeah. Yeah. Wow. I'm sure things have changed like just so dramatically on that front, you know, with the liberalization. We'll get into how you're building composer is like completely remote. You know, you're off in the middle of nowhere right now and that's all possible now. But OK, so tell us what is composer for folks who don't know?
Absolutely. Composer is an automated trading platform that allows investors to easily build a portfolio of systematic trading strategies. So instead of struggling to implement strategies yourself with endless lines of code and endless spreadsheets, Composer breaks the strategy creation process into building blocks that can be infinitely combined using our no-code visual editor. And then once you create and invest in a strategy, Composer will automatically execute trades based on the strategy's logic.
And if you're not ready to create a strategy from scratch, you can choose from our collection of vetted pre-made strategies. And the whole kind of pitch here is Composer makes those kinds of strategies that are used by
top hedge funds as easy to access as stocks or ETFs. So by strategies, you don't mean like, you know, I think a lot of acquired listeners would probably be like, well, my strategy is I think Amazon is a generational company or XYZ or Tencent or pick whatever. I'm going to buy that and, you know, hold that forever and be long. David, don't reveal my strategy.
Well, that's the funny thing is everybody has a strategy, right? Mine just doesn't hedge anything. They're just a varying degrees of sophistication. So like technically you could do that in composer, right? Like,
What I love about something like Excel that is so compelling is that you could do really simple, almost stupid stuff with it. You can use Excel as an overpowered calculator to add three numbers. I mean, it's kind of overkill. I do that all the time. Yeah, right?
You could create a strategy that is 100% allocated to one asset, which in this case would be Amazon. And that would be your strategy. I just think Composer might be overkill for that. You could do that other ways. There's this great app called Robin Hood that we're here to tell you about. Yeah. I think defining what a strategy is in the context of Composer, it's an interesting question. And I think it's the idea that you have some sort of repeatable rule that you want to automate. Yeah.
With the operative word here being like automated. So like you don't really need automation if you want to buy 10 shares of Amazon because you just, you know, it is a one-off thing, you're done. A lot of people, if not most, actually do want to do something on an automated basis. That was the original value prop of robo-advisors was like,
you know, automated rebalancing. Exactly. Yeah. Or even like investing a part of your paycheck every month or dollar cost averaging into whatever your bucket is. That's a strategy. A strategy doesn't have to be ultra high frequency trading where you're like trying to move closer to the exchange to shave off
nanoseconds of execution time. That is a strategy too, but that's a pretty complicated one. It can be as simple as I want to rebalance 60% stocks and 40% bonds on some cadence. And that's a totally valid strategy.
But Composer obviously enables you to do considerably more sophisticated strategies than that as well. Is it a productivity tool on top of places where I already custody my money? Or do I say, thanks, Fidelity, it's been great, but I'm actually investing my money with Composer now? So our focus is on the latter. And the reason for that is that in practice, it's just such a better user experience, right?
So right now we do integrate with, we allow you to authenticate with Alpaca.
which is a third-party brokerage. And we're not like a clearing broker custody ourselves. Something that we're about to release that we're super, super excited about is that we are integrating so that it will be like funding a Composer account, right? Like we're white labeling the broker so that you go to Composer, you open an account, you transfer the money to what feels like Composer. Most of these trading apps actually have a partner that they have on the backend. But from the perspective of the customer, it will feel like...
You are just interacting with Composer. And that just leads to a much better user experience without putting any other app down. Some of the ones that I've seen that tried to sit on top of other brokers, the connection to that data is constantly breaking. It's constantly getting fried. It's just janky. It's probably also not, things are not executing that fast, right? Which, depending on your strategy, may be important. Yeah, exactly. Exactly.
I think Apple learned that a long time ago, generally speaking, when you integrate different parts instead of making them modular. Modularity sounds really cool in theory.
What we've learned, and I think product designers have learned over and over, is that when you integrate things, you just get much more seamless experiences because you have more control. So how old is Compose? A year and a half, two years old? Something like that? A little under two years. We officially incorporated end of April 2020. It's COVID, baby. You've raised a seed round from amazing investors like Paki McCormick individually and Not Pouring Capital, but also First Round Capital and plenty of other awesome folks. But you've raised a seed round. Yeah.
And you are a not technically, but from a user experience perspective, basically, you feel like a brokerage. A, it's freaking amazing that a company can stand that up in this day and age in that amount of time with that little amount of capital like that never would have happened five, 10 years ago.
That's one. Two, how the hell did you do that? What's happened is the ecosystem has evolved. And that's kind of answers the why now, where you don't have to go into this super long relationship with these legacy dinosaur banks to become a trading platform. And I think it's helpful to break down what goes into a platform like ours. And I think there's really like, if you break down our value chain, there's three main categories.
there's data, like every trading app or platform needs data, it ingests data. Then you make some decisions on top of that data. That's like the trading logic. And that data is like, hey, what's the current price of all these assets? Exactly. So like data is like, yeah, what's the market price of this asset? A lot of it's just that, frankly. That's a huge part of it. And then taking those inputs, you'd make some decisions on that. And then the output is orders. You make trades. Yeah.
And you have to send those trades at some point they have to actually clear with some bigger broker or that you set it off to work on exchange or some, you know, these apps, they send it to other sorts of pools like hedge funds or whatever to as market makers. But anyways, you have to send an order off and that order has to actually execute. What we did is we looked at like put all this this value chain on a drawing board and said, OK, data. Well, there are APIs that provide data, right? We don't have to build that from scratch.
And then trading, well, brokers are actually becoming more and more commoditized. That was part of the sort of market analysis we did. I said, okay, well, all these brokers are now offering commission-free trading. So clearly this is maturing to the point that it's moving towards commoditization. So it's no longer that differentiated of a technology. And at that time, there still weren't that many good APIs though. So it's not that mature. It's not the equivalent of like that there was Plaid for like a brokerage platform. Yeah.
No, not really. Not as seamless as Plaid. Definitely not. It was still heavier than that. Although there are new entrants that are really changing that. So someone like an Alpaca, we work with them. Alpaca is an API. It's a modern REST API where you make API calls with instructions to that API.
to buy or sell some quantity of shares, of securities. There's also embed financial is launched recently. So there's these new entrants that are coming in that are trying to be sort of like the stripe of trading. A lot of companies don't have to handle
the sort of incidental complexity and all the crap that goes into billing or payments because of Stripe. So they can really focus and specialize on where they add value instead of this undifferentiated heavy lifting. What we saw, and just like you saw with cloud too, AWS and GCP, all these sorts of primitives, all these building blocks were starting to form in the industry that made it a lot easier to build a company like ours.
Having that ecosystem readiness is so essential to something like this. It's all about timing. This stuff is all timing. And we saw that opportunity. It was still early though, right? It was like at that time, it was still early. Like it was not... Alpaca at that time, I think had just raised like a series A. These were not mature companies, but I could see that...
Being Canadian here for a second, you want to go to where the puck is going to be. That's what Wayne Gretzky said. I probably just, yeah. Right? So you want to go to where the puck is going to be. And we saw where the puck was going to be. And that's what we saw. But in those three components, the trading logic or the strategy layer, that nobody's really done well at all, actually. Right.
Like that, you still have to like, I don't know, do it in Excel or do it in Python or something. Or there's like B2B companies, right? No one's made this for individuals. There's plenty of people that will like do it for like software for financial advisors or for hedge funds or, but the hedge funds are typically writing something proprietary. Yeah, it's mostly proprietary. Even that software, my guess, like the stuff I've seen is really either like kind of janky or it's like old, you know, it's old school stuff.
B2B legacy software. I don't know of any well-designed, you know, web-first software
modern, consumer-grade UX for this problem. I was solving a problem that I had, and so I looked around and I was like, okay, there's nothing. How'd you pick a segment and a business model for this? It sounds like you could add value for anyone who's investing, and then the question is, okay, who do you go after and how do you monetize it? So in terms of our target customer, the main pain point that brings our most fervent early adopters is
is that they had a similar problem as I had, which is they have like these, they're trying to cobble together some Excel spreadsheets. They're packing together Python or R scripts or whatever, and just failing. And to be clear, not because they're stupid. A lot of our early adopters aren't engineers.
a lot of people work in tech and are technical. And I think that will only grow, like the number of technical people is growing. And even then, it was just not worth the pain and misery that went into this, right? Like, even if you're a programmer, it's not that fun to just deal with lots and lots and lots of incidental complexity and dev environments. Some people like that stuff, but most don't, right? Like you ask a lot of programmers, like what part of your job you don't like? And it's like, I don't like DevOps, or I don't like doing crap that
isn't actually, I'm just redoing something that just has to get done, but it's not intellectually stimulating. A lot of the actual dev work that would go into, say, sitting at the automated trading system is very laborious and not particularly rewarding because it's just like crap you have to do. And then obviously a lot of our customers aren't engineers. So like then what about all the people that can't code, which is still a lot of people or people that like know a little code. Like I think there's a huge market that's kind of ignored and it's going to keep growing. It's going to keep growing of people that are not
professional software engineers, but they might write some custom functions in Excel, and they're a little geeky and analytical. We get a lot of those people too. I mean, BI is a gigantic emerging, still rapidly growing field, right? Everyone is learning SQL at all parts of the organization. Exactly. Exactly. I mean, of course, this hurts my ego a bit, but I'm kind of one of those people, right? I'm not a professional software engineer. I worked in data science, but
And I can code, but most professional software engineers do not want me shipping production code, right? So there's a lot of people like that. Well, it's the exact same dynamic as a Webflow, right? Lots of people who use Webflow...
can make a website on their own are fully capable of coding a website, but like you wouldn't really want to do that. Right. Like, you know, and there are also lots of people who use Webflow who are not at all capable of coding a website. And that's a perfect analogy. That's I mean, that's that's exactly it. And we were really inspired by those tools. We were inspired by Webflow. We were inspired by Figma.
We were inspired by Squarespace and all those kinds of tools that give those powers to people who either can't code, know some code, or even people who can code, but it's just a terrible use of their time to reinvent the wheel. I'm pretty sure we're talking about like April 2020, sort of beginning of COVID. This is pre-GameStop and WallStreetBets going supernova, right? Absolutely. And at that time,
It was absolutely a very niche market when you talk about systematic trading. There was weird internet forums. Again, they were growing at a fast rate. And so a lot of times I look at things and it's all about, I look at like, okay, don't look at the intercept, look at the slope, right? Kind of thing. And the slope was like, eh, okay.
And it's crazy. If COVID at that time taught us anything, it's when things are growing exponentially. Well, that's the other thing. Look at not just the slope. Maybe there's look at the second derivative, right? And if that's positive, then you have acceleration. Then, well, if COVID taught us anything, it's that, you know, don't underestimate exponential growth. These are small communities, but they were growing really fast. Yeah, when things are already obviously huge, those things tend to be saturated. It's already, you know, the secret's out.
And here I kind of saw a secret. I was like, whoa, this is going to be something because it's growing really fast. There is something here. And I looked at like subreddit stats. I looked at like, what are the fastest growing subreddits? One of the fastest growing subreddits two years ago was
was our algo trading, which now has like 1.4 million members. When I was starting Composer, it had around 100,000 in that range. Where can you go to get analytics on subreddit growth? There's like subreddit stats or something.com. If you just Google it, subreddit stats, you can look up subreddits and it'll give you the stats. That feels like a great place to mine for startup ideas. Exactly. Everybody who's looking for an idea, every MBA is going to go now. Every VC associate out there is now.
Is now going to want to research which subreddits are going. But it's... The other thing, though, was, frankly, in, like, all the Wall Street bets stuff... So, Wall Street bets started, like, popping off way before the GameStop thing. It was popping off because people were, like... Not only not investing in it, but it was a terrible idea. They were...
They were buying like naked puts on spy, like huge amounts of them, like with leverage and shit and doing crazy shit. Because they were like, okay, there's going to be like a huge correction because of COVID, which is a pretty easy call at the time. Which seemed very logical. I mean, we were hitting circuit breakers every day. Yeah. So they were buying huge amounts of like out of the money puts. And I was following it. And it was really funny. It was like, it was, I mean, maybe I just have an infantile sense. I was following it just because it was funny.
But the thing that I saw was that on Wall Street, you saw a lot of these eccentric people there that were becoming more sophisticated. You were watching them learn. I think Wall Street is overly impressed with themselves. And all industries are. It's not just Wall Street. Look, I don't want to say that tech is so enlightened and Wall Street's stupid. Yeah.
people in tech also like to think they're smarter than they actually are. But Wall Street really does it. So they like, you know, you use lots of jargon and acronyms to make everything more complicated and now you can't do this. And what Wall Street bets kind of said is like, no, we can, like we understand what you're doing. And that's really like when I see like the GameStop thing, I think the point that a lot of people missed in that, a lot of pundits missed is they're like, oh, these people are idiots. Some of them were, some of them were really smart. Like even Matt Levine had to be like,
Actually, they were right and I was wrong. They figured out how to orchestrate a gamma squeeze. I thought the market makers were smarter than them. They were smarter than the market makers. That's pretty wild. They outsmarted market makers who had a really imbalanced order flow and managed to do all kinds of crazy stuff. So I mean-
That's pretty wild to me. They could understand. It was like they were using Robin Hood, but they were talking about all the Greeks and they actually were teaching each other and they're being really vulgar. So I think people dismiss them. But the point here is that they were much more sophisticated than people gave credit to just because of the irreverent tone of the conversation. But they weren't stupid. Some of them were, not all of them. Some were actually pretty bright.
The thesis there was that there's like a much more sophisticated segment than we maybe give credit for that's out there.
There's a natural tee up here to like, okay, so did that segment start using Composer once you launched it to the world? And what's the journey been like? But before that, I'm interested in this idea of the Wall Street betsification of finance. And I'm curious, what like second and third order impacts has it had on the markets as a whole that we have all these people not only trading on platforms like Robinhood, but like starting to use leverage for the first time? And how
How has market making changed? And in particular, like you mentioned, some of these apps use pools, like different pools in order to match the supply to demand. And it maybe doesn't make it all the way to the market or to the exchange. What weird stuff has happened in the last couple of years in the stock market that is the fallout of this? Well, a lot of trading activity happens now off exchange, right?
And the reason for that, it's not as ominous or nefarious as I think sometimes people make it out to be. It has, at least in the short term, made transaction costs less for consumers. So just as like a primer, you know, if you execute an order on exchange, on an official stock exchange, a customer that does a market order might pay what's called the NBBO, which is like the best bidder offer.
And that's essentially that they're required to give you the best spread that's available on any public exchange for that security so that you can't get screwed. And what apps like Robinhood and other trading apps have realized is like, okay, well, what we could do instead is actually get better execution
by selling to like a big market-making hedge fund and sell that order flow to these hedge funds. And they will actually share some of the spread back with us and we'll share some of that spread with our customers. So they get essentially a revenue share of this tighter spread. It works for the hedge fund because they're still making money.
Because if you're sending a bunch of retail order flow to one of these hedge funds... That's the hedge fund as the counterparty for the trade? Exactly. Yeah, exactly. So they'll take all that order flow and they're not taking any directional position. They're just saying, okay, we'll make a market in this. And they're going to assume that, frankly, all that retail order flow isn't informed. When I say uninformed, it just means that
They don't have any knowledge about where prices are going to move at a really short timescale, which is true for basically any retail trader. And therefore, they don't have to worry that if they fill it, that they're going to get screwed. So it's actually desirable that that order flow is what they would call uninformed. And so they can basically pretty reliably know that if it's retail, that they're going to make that spread, which reduces risk, which means they can actually offer a narrower spread
on those orders. So the bid-ask spread actually narrows as a result because there's less risk to that hedge fund. So they like all that retail order flow. And what they do is because there's a narrower spread, then what some of these apps will do is they actually will give better spreads to the customer. Again, they legally cannot give a worse spread than the NBBO.
It's not so nefarious. It's like it's not so toxic for the consumer where it becomes sort of, it can lead to some skewed incentives because some of these apps, they make money the larger the spread is. So they might steer a customer to more illiquid or assets with larger spreads because then they'll make more money. So what's showing up on your home screen or your push notifications or something? Exactly. So it can screw up incentives. But if you know, certainly if you know what you're doing,
I don't think it's a bad thing because what this PFOF does, it's a revenue model, right? So the hedge fund shares, it's a rev share agreement, shares something that's spread with like Robinhood or whatever. And then Robinhood or whatever app doesn't have to charge any commissions, right?
So then you're giving commission-free trading to your customers. I mean, if it's lowering costs and you're not like a high-frequency trader, I don't see why this would be bad for customers. I think it's just very... It is quite complicated. Like it's a mouthful that I just explained, right? And so I think when things are confusing, people's first reaction is to be like, it must be like evil. Most of the activity that's happening on...
you know, a consumer platform like Robinhood or the like isn't strategies based on micro timescale, you know, optimization. Exactly. By definition, it's not. You know, if a high frequency trading hedge fund makes like
a little extra spread on my order on Apple, like that's meaningless to me, right? Like as a retail investor, that's like, I want to buy like five shares of Apple and hold them for a while. And not only that, you probably would have paid more if you traded on exchange anyways. Now, if you want to get really technical and I want to go down that whole rabbit hole, there is a game theoretic argument that if you
And I don't want to throw a bone to regulators, but if you put everything on exchange, you could argue that possibly spreads would narrow because the existence of these dark pools then increases the spread on exchange. And that's why...
you get better order execution off exchange. There's an argument. But it's not like Citadel knows that Apple's going to go down over the next month and I don't know that and they're screwing me. That's not like quite what's happening at all. No, that's not their model at all. They're not taking a directional position at all. They're just hoping that all the orders balance out. It's completely market neutral. There's no directional position on that kind of business. It's interesting. The dark pools thing is...
This is only something I learned recently and kind of blew my mind that the reason it's dark is because supply and demand are not being matched by the exchange. They're happening off exchange in a different pool. So if you're subscribed to exchange data, you're actually not going to see that that transaction is even taking place. That like the amount of data that's like visible to anyone who's looking at exchanges is not
less than it used to be because of this new business model. Yeah, that's absolutely right. Like getting a holistic picture of all like order flow is really, really hard right now for that exact reason. It's fascinating. I totally buy that game theoretic argument though, that you get much narrower spreads if we were all forced to actually go to the single central market maker for any given asset. Yeah, it's sort of like a prisoner's dilemma type thing.
Although I'm probably out of my depth now. I'm not an economist. So maybe let's rewind a little bit. You were the reason that, you know, we're joking about Paki and not pouring and all this. You worked with Paki at Breather. I did for many years. For many years. For several years. You were head of data science and then head of product at Breather? Yeah. The trajectory is I actually started, I just started as a rank and file data scientist and I was
promoted a few times and then I ended up as VP product running product and data science. And then we were in this crazy like office of the CEO situation where we were co-CEOs of Breather in this wild transition period.
So it was a, it was an adventure for sure. You know, you mentioned that when you started Composure, it was like you were having this problem. Like you would imagine hearing this episode so far as like, oh, at some point in time, Ben, like, you know, was a quant, you know, data scientist at a hedge fund, right? Like, and that's how you came into this. That's not what happened. No. And I don't think that's an accident. And I learned increasingly why. Yeah.
Because people who would come from that background probably wouldn't have the empathy, frankly, for the problems involved. Not that they're low empathy in general or bad people. Product thinking is different than quant thinking. And I think that by nature of my roles, I happen to see the value of both. And I'm not saying one is at all superior to the other. It's just a very different way of approaching problems, right? Like a quant hedge fund mindset is,
And I actually remember I was interviewing for a role at one point at like D.E. Shaw or a hedge fund like that. And they were trying to bring me in for something related to like incubating something in their technology. So I said, why are you doing this? You have all these people with better credentials than me. Why are you interviewing me for this role? I said, I just found me and I'm kind of selling myself out of the role. And they said, because like these people, they're like the world's most preeminent expert on this like one tiny thing, this one like very narrow thing that they're optimizing. Like it's so mature at this point that
They did a postdoc on this thing and they want to like focus on that thing. And that's it. They're not going to go like rebuild a whole new product or way of approaching something from scratch. That's a lot of it. I will say that I've also been trading investing since I was a teenager. I've had this as an interest. And then I've worked on a lot of problems that involve data. And that's a huge component of this as well. But then also there's a product aspect. So there's a certain amount of synthesis here.
That was required. As opposed to say, like, optimizing... Again, like, this is a product company. It's not...
How can I create an adaptive order algorithm that shaves nanoseconds off of like some trade or something? It's not that. And to the extent that there's an optimization function to create the best product here, you don't even really, when you're starting the company, know what you're optimizing for. I mean, I guess it's like enterprise value or something, like some very long... It's a wicked problem. It's a wicked problem, not a tame problem. It's a very wicked problem. And it requires...
divergent thinking rather than convergent thinking in the beginning. You have to do this generative process of generating a lot of different ideas and then connecting them.
And that's not comfortable for a lot of people. I mean, when I tell you this sentence, there should be some kind of product that lets you compose different building blocks to execute trading strategies. Like, that's not a product spec. There's like a zillion questions that stem from that. And it's like, okay, who's using it? And like, how much should we presume that they already know? And like, what's the atomic...
unit of a building block and like all this stuff just creates a like a wicked problem when you start everything's very high level like inherently you're like starting with like the it's funny because a lot of people talk about how ideas are a dime a dozen execution is everything and for the most of the life cycle of a company that is absolutely the case
But I do think in the very beginning, like the idea does matter because it's, that is the execution. In other words, like it's very hard to even disentangle the two in the beginning. So like, like you're saying, you have to take this idea that's kind of a high level problem statement and then keep breaking it down and going to the next level of abstraction. Cause that's like a very abstract problem. You have to keep breaking it down and resolving it to a lower level of abstraction until you're actually building a product. And that process is, is difficult. Yeah.
So I don't even know if it's ideation or it's execution, but it's definitely a big part of the initial creation of a startup.
And it's one of the hardest parts. And it's one of the most mysterious. People have tried to scientize it and systematize it and mostly failed, in my opinion. You're talking to a startup studio founder here and the other Ben G on the line here. But no, but that makes it a good problem. Because people have failed to do it, I think that's what makes it an interesting thing to try to create a studio for. It's funny. I can tell you, we have not gotten better at...
having better ideas. We still, six and a half years in, kill 90% of the ideas that we actively work on. Despite having lots of systems in process to work through these ideas as fast as possible, we've gotten more efficient at killing ones when we think they're going to end up bad, but it doesn't mean that your raw material input gets higher quality.
Yeah, I think one of the most fascinating things there, for example, is like, why is the distribution of startup outcomes largely unchanged? Like, why would that be a natural law? Is it like, is it a natural law? Like, is it just natural law that 90% of very early ideas fail? And then like, there's a sort of distribution that like, as you go through stages, but it's like been remarkably stable.
And, you know, even the top tier VCs, what separates them from like third tier VCs or second tier VCs is simply that the winners, it's just the payoff. It's not the actual probability distribution. It's the payoff function. It's that the ones that hit just hit harder.
Oh, that's like our two-part series on Andreessen. It was like, you know, that was the whole thesis. And everybody was like, show me the money, show me the money, show me the money. And then they're like, Coinbase DPO, you know, like, there's the money. You know, they made $11 billion. People were talking shit about A16Z. It's like, oh, the IRR isn't very good. Oh, yeah, that 2015. Yeah. Yeah, it's like, you don't know that yet. But then it's really interesting. It's so hit driven. But yeah, like their distribution of outcomes is probably not that different from any other firm. It's just that...
Their, you know, head of the power law netted them $11 billion. Like, that's the difference. Yeah, you get one event that returns the whole fund and that's the game. And it's a little frustrating at a level of a startup that you don't have that sort of diversification.
So then it becomes a question of can you recreate that, you know, internally? While still having conviction, obviously, you can't just throw spaghetti at the wall. So obviously, because you don't have infinite resources, so you're faced with this abstract problem that every time you sort of put a stake in the ground, you end up with five more questions. So like, what were some early...
guardrails that you put in place where you're like, okay, we know for sure we're doing this, not that. And how did you sort of get from the broad idea to what it is specifically in code that runs and designs that look nice today? So I think you need to have...
high-level set of guiding principles. You need values and you need guiding principles. And this becomes almost just like companies have values. And we had these very early before we even incorporated. We said, okay, we need to codify our values in a cultural sense, in an HR sense. I think you need product values. So you need attributes that are evergreen. Otherwise, you don't have a distinctive
product culture. So there was three that I came up with pretty early. They're gorgeous, accurate, and fast. So those are three examples of attributes that were going to guide this. And then just more to the point that this was going to be a product that while solving something quite technical and serving a relatively technical audience was going to be guided by design principles and by usability. Because almost no one else in the space
of like these sort of prosumer trading tools. They're just like, some of them are actually quite successful businesses. They're not sexy, but they actually are, some of them are actually extremely successful businesses, but they're not well-designed and they don't put usability first. And so one of the easy constraints was this high level constraint, which was like, okay, we have to balance flexibility with usability. And, you know, we have to be opinionated in our design.
So that we steer people to use the product in a way that it just functions and works better. So that means that you can't do what a lot of failed no-code tools do, which is you say, okay, I want this to be like a totally full-functioning programming language that's like Turing complete and does all these things and does everything at the expense of it being at all usable.
I mean, that continues to be like one of the hardest. It's so hard. I wrote this article with packing one of the things about Excel. And one of the things we really talk about this because we were like, we've been talking about this for a long time. It's like, why is Excel such an enduring piece of software? Why is it so successful? If we had to like have a bullet point thesis, like it's the only thing that's usable and flexible.
And that's actually incredibly rare. Like how many things can you think that are like have total flexibility and are extremely usable for regular people? Like that's such a insanely difficult challenge. So just that alone focuses you a lot. If you say it has to be usable, but still flexible enough to hit these XYZ use cases, that's an incredibly focusing product objective. What was the journey from, you know, April, 2020, you're starting this,
there's a product that works and that people can use. It took a while to actually build the MVP, not surprisingly. Like we're doing something that isn't just like a riff on something that already exists. So when you're doing something category creating like this, it was pretty hard, right? Like we had to do a lot of mock-ups and prototypes and
And we spent months just doing user research on prototypes and proof of concepts, creating like videos, Figma prototypes. We did just like hundreds of hours of qualitative research of user interviews, just like, does this make sense? Does this make sense?
Because again, getting something this powerful usable is so hard. So one of the biggest challenges is like, it's almost like it's a cognitive science problem more than anything. It's like, do you understand this? You know, like if they validated that, it doesn't make sense to ship a whole bunch of code. So we spent like months in that stage. It took almost a year to really have something that was actually ready to be used by real humans. And then it really like evolved quickly from there.
Once you have that initial foundation and then you can just keep iterating. But to build an initial foundation was, it was hard. It was definitely really hard. And so you've got this like usability thing you're testing, but you also have to test like, will people make money trading with these strategies? Because you started with like a finite set of like your own trading strategies, right? Yeah, that's right. I think it really helps that all of these are back tested. So like we really, and this is part of being opinionated and actually, you know,
Kyle, who writes content for us, who's fantastic. And actually, he was at Vanguard for a while and joined us. At least we have some people who know what they're doing. He wrote this piece today on the value of backtesting. And backtesting isn't perfect, obviously. First off, they always give you that asterisk that says pass returns are not predicted or the
predictive of future returns. And that's true. There's something very pernicious in backtesting, which is overfitting, which is just that you can try-- you throw enough spaghetti at the wall, you're going to find something that works. You can cherry pick data that makes any strategy look good on a backtest.
It's not perfect, but I have found personally, even if we're going to, it is so much better than the alternative, which is just making shit up as you go. Like how can, how can the latter be superior? Well, even just the backtesting, like I got to imagine that's like a pretty freaking killer feature of composer. Like if I come up with a strategy before composer and I want to backtest it, what the hell am I going to do? Like, I can't do that. It's incredibly difficult. There's a free tool that a
But it's very basic. It only uses monthly data. So it's very coarse. So like the most recent data it has will be like 45 days or whatever on average delayed. It doesn't feel like a modern analytics tool in that sense. But it is impressive data.
was the most impressive free tool that I was using at that time. But yeah, backtesting is super hard. It's super, super hard. It's computationally difficult and intense. You have to get all that data clean, all that historical data and have it somewhere. Backtesting is really hard. And then creating an accurate simulation of what actually happened, that in itself is really hard. I was listening to Jim Simmons at Rentech, which is by far the most successful hedge fund of all time. There's a great book, The Man Who Solved the Market. And he was being interviewed and he's like,
Yeah, our strategy is people would probably be surprised they're not as complicated as you probably think. He's like, the one thing that I'm just like not even worried about anybody competing with us on is just like accurate data and accurate backtesting. He's like, we spent years just like accurately estimating transaction costs and
how orders are filled and what things actually look like had they traded. He said that question alone is so hard that nobody, he thinks nobody except for them actually does it well. And that's their advantage. And his whole thing though is that the future is like the past. Like Rentac, the reason it's the most successful is like we're just really, really good at actually figuring out what happened in the past. Most people fail because they didn't actually simulate things properly. You can actually truly simulate it, then like, no, actually a lot of these things are pretty predictive of the future. That's their whole like...
raison d'etre is that, is just like really accurate backtesting.
So, clearly, you can do some pretty incredible things with that if you just do it really well, if you know what you're doing. Every time I read a book, there's a new concept I'm obsessed with. And after I read The Principles, I got obsessed with backtesting. And it's like, once you see it, you can't unsee it. Did you try and implement backtesting in other places in the company, like in your decision-making process or in your hiring or in your... Do you apply it more broadly than just for the trading strategies? Yeah.
Yeah, maybe I don't call it backtesting. But yes, yes, I think we are genuinely a more data driven company. We like to have a record or provenance trail of decisions of events. I wouldn't say we're totally systematized. We're not like Bridgewater. My guess is that's mostly PR. I don't rate people after interactions.
And if they rated after me, I mean, I would probably, you know, destroy my ego. So I think we're all too delicate and fragile to have the true rent tech social currency score or whatever. But like we have an amazing analytics lead, for example, like really early in our company history. So like we have more tracking, for example, set up than a typical startup or stage because we like to track things and see how they perform and how things. Yeah. So yes, I think it has influenced. It has absolutely influenced us.
how we make decisions, how we analyze things. David, I remember you did this in kind of a crude way when you were at Madrona and you were trying to answer the question of like, what makes for the most successful founders of the biggest companies of all time? And it was coarse, you know, it was like a... Well, you know what I used? Excel. Yeah, yeah, yeah. Like I have the spreadsheet, but it's like, okay, like let's look at the backgrounds of all the people who started $100 billion companies and let's look at, you know, data
different markers and private company data like if you're trying to analyze startups is so much harder because you don't have the ticker you know you don't have the the second by second tape well it's also all private right right most of that data is completely private yeah exactly to your point ben are like this is so hard i did that analysis when theranos was worth nine billion dollars
So like that was a big data point in the analysis, right? Like, well, obviously that shouldn't have been worth $9 billion, you know. There was pretty material non-public information in that case, right? So yeah, one thing we do is we do money ball hiring. I'm not even sure if our staff fully knows that, but yes, our hiring is very structured in money ball. Like it's all, it's highly, highly analytical. And we didn't actually completely start that way. You know, it's funny how many times you have to be
kind of hit on the head with this. You have to keep relearning the same thing over and over and like, oh no, this definitely applies in this domain as well. But yes, like we learned the hard way actually that like, no, you need to be as analytical as possible. Moneyball. Tell us more. Some of it's a trade secret. I can't, I can't tell you everything because it is, because this, you know, that hiring like obviously is like the heart, like the last year, maybe if the markets crash, it'll flip. But like, because there was so much money sloshing around, fundraising was easy the last year or so, the last nine months at least. And then hiring is like brutal, right? Because-
It's just the other side. It's like any market, right? It's obtaining capital is easy in this market and deploying it in a profitable way. So some of it I can't share, but yeah, like, well, okay. So maybe not the specifics of the hiring process, but you know, you guys are fully remote of which many companies are now, but like,
Tell us about what that means for you because you could take it to a pretty extreme extreme. We are completely distributed 16 people all in different cities. They're not necessarily cities split between Canada and the US. And I started the company. I was in Nicaragua for most of the company's history. So I moved back to Canada a few months ago.
Because the situation there got a little dicey. And the funny thing is, like, I'll be the first to acknowledge that remote and being that distributed is not perfect. Like, there are many, many issues and downsides to it. But I think if you just acknowledge them head on, that you can sort of mitigate a lot of those issues. But, you know, life is a series of trade-offs.
How do you communicate? Is it Slack? Is it email? Is it synchronous? Is it asynchronous? Like what? Yeah. So I have a very controversial take that actually. So even though we're remote, I think that if you're remote, like async is actually a terrible idea. Like fully async is like terrible. I know, I know the GitLab loves it. Other people love it. A lot of people, a lot of, a lot of, especially a lot of engineers love it.
So like if anything, you know, I'm the type of person who would like async, but I just like have finally had to accept that. I don't understand. I don't think like a consumer product company can be like fully async. And the reason why is like, yes, async is, is less disruptive, but like you, you need synchronous communication. You need a lot of synchronous communication if you're remote, because you already, if you're physically disengaged,
not in proximity, that already affects trust. It already affects the bandwidth of communication so much. Things get lost in contact and translation. They get confused and muddled. And there's been so much research on that. So why would you make it harder? Right? So the big thing, and we're always adapting and learning here is just like, even recently realized like just, you know, the biggest feature that we take advantage of is like slack huddles, just like hop in a huddle. Okay. If there's any confusion, just hop in a huddle.
And I think there was like this weird insecurity that a lot of remote companies had. And it's the classic, like right now everything is so polarized. Like you have to be one extreme or another. It's like, if we're remote, then we have to be like GitLab and do everything asynchronously. And you know, the only way we communicate is through like comments on like a GitHub, you know, PR or something or a Jira ticket. It's like, okay. Or, or you're like in office and you spend all day in meetings. It's like, well, isn't there like, like, you know, can we be a little more thoughtful? Yeah.
about this. So I think with everything with this stuff, you just kind of, you know, you do it. You have to take reverse principles, like say, okay, what makes sense for us? What doesn't make sense? Tell us a little bit about where Composer is today. We launched in open beta the end of last year. And the really cool thing is we have a lot of like super fervent, you know, super users that are out there spreading the Composer gospel, which is really, really cool to see.
especially as someone who was a early, you know, built something as a customer of my own product to see people who are even more fanatical than myself is a pretty fun and pretty exciting thing too. Where do those people hang out? Do you have your own communities or is it on Reddit or? It's a variety of places. A lot of it is Reddit. That's a lot of it. Some of it's like Bogle heads. You see people on different online communities in these sort of like
niches. And I'm sure there are also some in discords and things like that, that we can't even, they're harder to discover. Yeah. Dark pools. Exactly. Exactly. So that's, that's really, really cool. I mean, the other really exciting thing, like I mentioned, is just like, it's just the team. Like we have this, we have this really, really killer team. Not, that was a very hard thing to build. So like I said, we're at 16 today.
That's pretty wild because that team is now extending, you know, my original vision in ways that I never thought of. And they come from really different disciplines now. Like it's very, very interdisciplinary team. And so they can kind of cross pollinate ideas and work on things in a way that, you know, really does create like this collective brain trust that, yeah, is a lot, a whole lot more powerful than anything I could have come up with.
You were the first not boring portfolio company. You and Baki wrote Excel Never Dies together. There's the not boring memo on Composer. How much of the interest in Composer, either on the user side or on the hiring side, or frankly, on the capital raising side, do you look back and you can attribute to that? Oh, huge. I mean, I won't even hold any punches there. I think
Without a doubt, especially relative to the check side. I mean, like the highest value add was having Paki, having that. And that was luck, right? Like when he wrote the memo, I think he had like 7,000 subscribers and he was not a known entity. I mean, we have funny stories there, but when we were doing that, like some people were even like, why is this guy? I mean, is he going to add value? You know, like people ask questions like that. And now those, some of those same people are
hitting me up like, please, can you just make one itch? I just want to talk to Paget for 10 minutes and can you squeeze him into this? You know, it's wild how fast things change in this world now. Like that was 18 months ago. I mean, just even saying that, that's crazy. That was 18 months ago. That delta is insane. So some of that was luck, but his platform, that kind of thing is the future. Those kinds of
of newsletters and thought leadership as a huge part of our distribution strategy going forward. And he informed a lot of that. And he inspired us to think about Composer as not just like an app, not just trading, not just strategies, not just a no-code tool, but it's content, right? Like each of these strategies is content and really rethinking what content even means. Like these strategies are content you can write around them and then you can share them
And that's part of the reason, like I said, we have someone who like writes full time for us internally. Like all those moves were inspired by just watching like Paki's meteoric rise, right? Like that was hugely, hugely inspiring to us. And then I'm biased here, but he's also like actually, I can honestly say he's a fundamentally very good person, which helps a lot. He's someone, he is, he's a good person. Like that helps a lot.
That means a lot. I can definitely second that. As someone who's very interested in this evolving media business model, do you use the not boring investment memo as collateral for recruiting? Do you find yourself sending that to people? Oh, yeah. Especially early on. Now it's a little stale, but absolutely in the beginning. Yeah, absolutely. We would share it with investors. We'd share it with people. It was very valuable.
And then also he created that like diagram that showed us as like, I still use certain sort of metaphors and things that he created in that and sort of images like calling a composer an entropy wrangler or that, that, that, that image that I think really resonated with a lot of people where it was like, okay, he created this like,
little diagram that showed, you know, Robin Hood on one side and Roe advisors, another side and composers in the middle. Like that was a packet creation. And that was just like, he has a way of taking things that are kind of complicated and distilling them in a way that's just very relatable. And we've kind of remixed some of those elements into our own collateral for sure.
Your story is bringing up for me memories of being in business school and beyond, but of Andy Ratcliffe and his framework that he always talked about of that, like to be a successful startup, you need a secret or an innovation. You need an unfair advantage on the product side, but also on the distribution side. And you got to get both of those. How does this guy have so much distribution? He didn't spend any money on it. Well, it's better to be lucky than smart, right? Yeah.
Right. But like, it's cool. Like, you know, you had this like secret on the product side of like, Hey, there's all these people out there. They're hanging out in reddits that like some of them are idiots, but a lot of them are not. And some of them who are idiots are going to learn to be not. And they're
and they're going to want tools. That's like a product secret, right? Absolutely. And then you get like this, you know, free, amazing, like a tiny, you know, angel check into your company leads to, you know, something that can completely change the distribution equation for you. Oh, totally. You know, people argue about what's more important than product or distribution. It's like, well, it's both. I was in a retreat with some CEOs and one of them said, yeah, it's like asking like what's more important, a brain or a heart? Yeah. Yeah.
You can't live without either. So you better have both. And both matter a whole lot. And there are situations where like a product can be so...
can have such perfect product market fit that like the answer on distribution is well it spread virally because people just loved it deeply and shared it with their friends that's an answer too but that is not an answer that you can scientifically count on in the process of crafting your startup yeah i mean elad gill had some good things on that and i think the argument there is
That most of the things that look like that wasn't actually the case. Like Facebook, all the sort of classic examples, it wasn't actually the case. Like Facebook did all these aggressive tactics for distribution. Airbnb did all these aggressive hacks for distribution. Google paid billions, billions. Google paid billions for distribution and doing like- Still does every year. Really like-
gangster shit back in the day. They were doing like, they were distributing like toolbars through like, you know, adware packages. Like it was super aggressive. They were paying Dell to install a cool, and it was by far the best search engine. So, you know, and I say this as a, as a, someone who, who is much more from a product and data back, not a marketing person, but like, it's something you have to learn. It's like, no, like everybody, everybody has to do distribution.
Well, Ben, this has been awesome. There's a lot of investors that listen to the show. Anything they should be thinking about or that's exciting that you want to share with potential future investors? I'm not too worried about fundraising.
Uh, so I'm shocked, shocked that you're not worried about that. No, it's a well-deserved. I'm not going to do the, uh, used car salesman thing. All right. Well, what about who should try out composer and where should they go to play with the product? Any serious investor that wants to earn better risk adjusted returns and wants to stop flying by the seat of their pants and
wants to get rid of their messy Excel sheets, they should go to www.composer.trade. Sign up today. You can play around with it. You can start, you don't even have to transfer money to start playing with it. And yeah, do that today. I love it. Oh, one final question I meant to ask earlier. I love the name. Where did Composer come from? Are you first or are the like Web3 composability people first?
I think we were before that. Web3 wasn't a term when we came up with Composer. I think Packy made composability a Web3 term because of you. They swagger jacked us. No, I'm just kidding. The Composer thing, it goes on a couple levels. One is that I love music. And originally, I didn't have the talent for it, but I wanted to be a musician back in the day or a composer or something. So there's that sort of personal aspect of it. And I do think that there are all these parallels between great software and great music composition.
And Rich Hickey's actually talked about that a lot, who's the creator of Clojure, incidentally. I didn't even know that until-- but we actually use Clojure. And that's a total coincidence. So that's really, really cool. The other is a little bit more direct and a little more practical, which is just that Composer was designed from the very, very beginning. And you're asked in the beginning, OK, what was a guiding principle that narrowed down this design space? And we wanted things to be composable.
One of the hardest things when you create strategies is like, okay, I need a container. I need to neatly encapsulate the strategy so that I can create a strategy that is composed of other strategies. And it sounds abstract, but it's not. We all do this when we invest because you use an ETF. An ETF is an abstraction, really, over a whole bunch of underlying assets. And there's a great book called Trillions, you should read it, that talks a lot about this. But we take ETFs for granted. It was a tremendous innovation.
And it really was like an abstraction that allows you to treat this giant basket of stocks as an individual stock. And we said, wait, could we do that with a sort of logic? So not just assets, but actually the logic that operates on those assets. And then could you compose those strategies into sort of like Russian doll sort of structure? That's the origin story of that name. Love it. Love it. Well, all right, LPs.
This has been awesome. Ben, thank you for joining us. Thank you for having me. With that, LPs, we'll see you next time. We'll see you next time. Cheers.
