Industrial-Scale Money Laundering in the Music Industry

Jakob Wredstrøm

Could you somehow angle your web cameras a tiny bit lower?

Andrew Batey

Is that okay? More?

Jakob Wredstrøm

I mean, no, that's good. I think that's good. Yeah. So.

Andrew Batey

I lost my wide view and it put me to like a short view, so I'm not sure where I am in your view. Okay.

Morgan Hayduk

Get there.

Jakob Wredstrøm

I think it's good. I think it's good. So we're not officially recording yet, but again, so Andrew, this is your third time. Thank you for not wasting your time on the podcast.

Andrew Batey

I think you're great. So I think we're happy to be here.

Jakob Wredstrøm

I really had the I would love to go nerdy today like so I obviously always try to do my homework before I go on a podcast and I sort of got the press release what you guys did and You can explain your own words, but I was like after reading that like that was I'm not so familiar with this. So I like went into plexi like try to explain it to me very simply and I got more and more intrigued about sort of

Jakob Wredstrøm

what we're going to talk about today, like the operating system of fraud. Well, it's a lot of things. So if it's okay with you guys, I would like to get really nerdy today and take it from that angle because we've already had sort of your founder journey on. We've already started the podcast, so that's it. We already had your founder journey on, which was incredibly interesting. And then we had Andrew Yu and Konan. And this is sort of a continuation. Actually in private, I don't know why, I get to speak with lot of people guys. And in private, when people ask me like, what's the startup to look out for in the music industry? I pretty much always mention you guys, mostly because what you do is incredibly boring. And not really, like, but, but that is, that is what it takes. Like, so, so now when we're rolling, even though I wasn't sort of planning for it, guys, but for the few people who don't.

Morgan Hayduk

Yeah.

Andrew Batey

I mean, it's fair.

Jakob Wredstrøm

know you by now. is third time so you're getting regulars. Who are you guys again?

Andrew Batey

yeah, we're, beat up. We're the leaders in fraud detection and sort of an entire trust and safety suite for the music industry. and, I'm Andrew and this is my co-founder Morgan and we've been best friends for almost going on two decades, one and a half. We'll call it one and a half.

Morgan Hayduk

Yeah. It's aging us. Yeah.

Jakob Wredstrøm

It wouldn't make you older than you are.

Andrew Batey

The halves count when you're going for decades, you know?

Jakob Wredstrøm

So usually in the podcast, we have this whole journey story. And for everyone listening, could just say these guys are amazing. What they're building is incredible. So let's of skip that part with you guys sort of trying to merit yourself in this conversation. Let's just assume everyone out there, these guys are incredible, like know what they're doing and they're working with great companies. let's get to the conversation today as something new you've built that you're trying to implement. The whole... sort of whole perspective of what you're trying to build and what you're trying to address is fraud. Like if you want to have a big perspective, but then maybe I'm wrong as well, but like that's sort of my, my very simplified version of it. But now you're going into announcing something very cool and very technical. Maybe like I asked perplexity, can you simplify this to me? Like what's the most simple version of what you just sort of announced?

Morgan Hayduk

So fraud was the entry point to understanding, I think, a number of different challenges that are sort of trust or safety related in the music industry. we spent, you we've been eight years sort of down this path of fraud detection as a singular focus for the business and our. grateful for the customers that adopted us early and for all those that have come since. And as we got deeper down the path, we realized that yes, like streaming fraud on DSPs is a problem and claims fraud for collection societies is a problem and ensuring that your catalog has high integrity for labels is important. And for distributors, sort of the same thing. And so those were all related issues under this umbrella of just like fraud detection. But then, you can't help but have conversations with customers. You sit you know, in various meetings or you go to events and you hear about other sort of related problems. And I think for Andrew and I, the realization sometime over the last year was we've earned enough trust, the customers that we have and relationships across the industry, that there were things that we could be doing that we weren't yet that would solve other related problems and challenges that they necessarily did not want to build themselves, but knew they needed. And so the trust and safety operating system that we announced late last year was sort of our, you know, realization. or bringing to life that vision of can we have more than one thing for everyone that we already serve and can we find a whole bunch of new customers by bringing different products and services to market that all fit under this like big sort of pan umbrella of trust and safety. And so we can go into all the constituent parts, but the idea was really we had a unique point of view on one problem that led us to see a bunch of other related issues. And if the goal of the company is to ensure that artists get paid correctly and fairly for their work, then there were a bunch of leaky bucket problems that needed to get shored up. And we thought we were really well positioned to do that.

Jakob Wredstrøm

Okay. so when I was reading through perplexity, I was trying to understand what are the practical use cases and it, you know, search the incident and came with a few good examples that I've found to my satisfaction. But, how would you describe practical use cases? You can, you can be nerdy. We can be nerdy in the podcast.

Andrew Batey

Yeah. All right, let's start with let's think of supply chain. So let me just define for people who aren't familiar. I'm assuming everyone knows music, but just in case, if I'm an artist and I'm uploading a song, I'm going to go to a distributor. The distributor's job is to package all the metadata and get it on 100 something stores, which are streaming services store, be like a Spotify, Apple, YouTube, Amazon, whatever. So there's going be 100 of them. Audio Mac got to shut them out because they're great. But you have a bunch of these. you know, altogether and they distribute it and it goes up and then there's a bunch of streams that happen and then it goes now back the other way. So the streaming services now take the money they've made. They pay the distributors who divide it up and pay out the artists or labels directly. So there's like this supply chain. When we started, we obviously started with fraud detection on the streaming service side. And we said, why are there a bunch of accounts playing a song 63 times on a Sunday in the exact same order manipulating the streams? we would have assumed it would be artists trying to run the counts up and become more popular the same way that we would game something in a marketing context. It turned out that 80 % of the fraud we caught was financially motivated fraud because the music industry is the only one where it's performance based. And instead of paying everyone like an exact number of stream or something, they all get a percentage of that pool. So if Spotify is gonna pay out, you know, 100 % of this money. and you did one stream of 10, you get 10 % of that pool that month. So because there's a finite amount of money that's being paid and everyone's competing for the market share, the more that a fraudster can run these streams up, the more they're extracting from other people. So what that effectively means in real terms, without getting crazy too nerdy and the nuance of how pools work, because we could also spend like an entire hour of how the calculations actually work in all the spreadsheets. But let's just assume that money's being paid out. You have billions of dollars that should have been paid to real artists that are being siphoned off into all of these fraudsters accounts. And in many of these cases, the music's not really real, or I guess it's technically real music, but it could be like AI music, they're taking albums, like the gate in the past used to be how many songs can you get up because you don't want a song to be obvious. Like you want a bunch of songs that have 2000 streams so that no one notices the song with almost no streams.

Andrew Batey

If you have a song that suddenly charts, you have a problem. You have a big magnifying glass. If you're the fraudster, you do not want this. You want to live in the middle to long tail. And so you need content. And it used to be content was really hard to get. You'd have to hack Dropbox, hack, you know, a streaming platform that has user generated uploads that doesn't fingerprint. You'd have to go and digitize albums from Africa and Latin America and Asia from like the 70s that no one's paying attention to. There were a lot of things you had to do. to sort of just get songs as the prerequisite to creating your fraud. AI changed that game. And so now all of a sudden it's very easy to upload tracks. So on our side, we're seeing the streaming service end, looking at all the fraud, but they're not necessarily sure who's being paid out because they're paying the distributors. Then we go down to the distributor side in our next sort of journey and we say, hey, distributors, who are you paying out? Why aren't you doing customer due diligence on these people? Why are you paying out terrorist organizations and criminal fraud orgs or click farms from Russia? How is it possible that we are paying the wrong people? So then you start getting into customer due diligence, which is basically like KYC and AML, like know your customer, anti-money laundering. You start really focusing on that. And Morgan and I are looking at this entire industry going, why is this all separated? Why are they doing AI music detection in a different place than where you're uploading it and checking for it. Why are you not checking who you're paying out? All of these things are interconnected because ultimately they're all prerequisites of the fraudster getting their money. And so if our goal is to stop fraud and we've stopped it from being paid out from the DSP side in terms of don't pay this, but there's massive amounts of types of iodophors that have occurred. And someone once brought it up like, it's like a water balloon. When you squeeze one side, it just bubbles out the other. Like you don't really fully, it's not you close the water balloon. We've squeezed here at the DSP side. And then you started seeing it go in collection societies with CMOs and sort of basically all of the rights holders and claims. You start seeing it with them uploading massive amounts of other types of music, you know, account takeovers. Like when you start shutting down bots, they start then taking over your account instead of creating their own. So you start in like accidentally.

Andrew Batey

driving towards these other parts of the industry. So when I think of trust and safety, you asked me to go nerdy, so I'm talking and rambling, but I'm like, hey, we started with fraud detection. And then we realized when we removed the first big boulder out of the stream, all the other little ones that we had to fix too. So then you go to the next level and you start pulling that rock out and that rock out and that rock out. And at some point Morgan and I said, We have like 30 rocks that we've removed. Like it should probably just be an operating system for everybody because at some point you end up with all these services that you're building for everybody and that just one cohesive system makes sense because at this point we are effectively the trust and safety platform for most of the music industry.

Jakob Wredstrøm

I think that's interesting. When you build something so novel and you don't really know what happens when you start scratching the surface, and guess that's sort what you figured out. Are you to some degree saying that your first approach with fraud, of course, had positive impact, had a lot of knowledge and learning, but these fraudsters are just moving around as much as they can, and you need to do something holistic about it. Is that what you're basically saying?

Morgan Hayduk

Yeah, I think there's two things happening at once. There's the evolution of as you get... As an industry develops more resilience at cracking down on the obvious forms of fraud, the basic sort of schemes, the fraudsters adapt. So you can go back and look at the history of any online or recently online industry that has high dollar values attached to it, whether it's the early days of Google AdWords or read the PayPal Founders book. There's a chapter at the beginning of all of those stories and it's pretty damn close to the start about when they first realized fraud was a problem. And so for music, you the initial discovery was 30 second or 31 second streams by obvious bots. Once those start getting caught, the fraudsters evolve their tactics. make more, you know, interestingly, creatively programmed bot farms that behave more like people. They do things like a count takeover. so like Andrew squeezing the balloon example is exactly that, right? Like in one way you're, you know, driving innovation by catching the obvious stuff. The second part is I think as fraudsters start to get to know an industry better, they too start to see all of the play where money is sort of hiding. And so in the early days, I would guess if you went back to like 2014-15, you would see primarily fraud focused on extracting recorded royalties, recorded music royalties, where song was played on a DSP, DSP pays the distributor, distributor pays the fraudster. That was sort of the chain. If you spend long enough in any industry, you start to figure it out. Some clever fraudster along the way went, hey, there's this other like 20-ish percent of music income that comes from publishing. How do we get that? we just go to the society and we claim the work there and then they send us the check. Cool. Let's do that. And then at the same time, you have other ones going, hey, these societies, they, you they don't necessarily match every work. They don't know exactly who should get paid for all of them. What if we just went and programmatically claimed millions or tens of millions of dollars worth of ownership interests in copyrights in hopes that some of them don't have a match and we get the check in the mail? Like it's an oversimplification, but that's sort of the idea. So I think what you see is like a combination of innovation and intelligence and learning and

Morgan Hayduk

and adapting all happening at the bad guy level. And so for us, sit, you know, sitting above that, looking at it going, well, if they're adapting, we need to as well. And so where can we be helpful? And so the first things we sort of discussed were, you know, working with collection societies, working with distributors, but then it became more, you know, there are other vectors, there's identity. Do we know who these people are? Do the distributors, do the collection societies really have a detailed understanding of who their customers are? And if they don't, why and should they? And then the other challenge obviously being AI music detection, which is fairly recent, but evolving so quickly that we again looked at it and said, we have to have something in this space. mean, this is, you know, however you slice it, this is one of the exploding vectors in the sort of arms race against fraud. And we need to be able to be, you know, a provider to our customers of an AI detection solution that at least helps them, you know, at the front end of the sort of copyright content, whatever ingestion stack. know whether something is AI music or human. There will be lots of business model iterations that come out of who's a good actor, who's participating in the licensed sort AI music economy. And I think they will be treated differently by distributors and by streaming services. They also probably aren't going to be the problem from our perspective. The problem is that large language models are open source. In a lot of cases, you can go and create a large language model that makes music that no one knows about.

Jakob Wredstrøm

Mm-hmm.

Morgan Hayduk

it can still generate tens of thousands of tracks for you if you're trying to scale a fraud scheme. So we need a reliable way to detect that music at the front end before it gets onto a service and be able to tag it accordingly. And then if we catch it as part of a fraud scheme, great. But even if we don't, maybe that'll have some sort of second order consequence for the royalty model and how the owners of that music gets paid. And that's important too. those are like, yeah, walking down the path, you can kind of go choose your own adventure here on which of the sort of various products we've recently created. you want to get into, like all of them are an offshoot of seeing the evolution of fraud from our perspective and trying to stay one step ahead and offer something that our customers can go to one place and get everything they need as opposed to like, you know, we got to get a KYC vendor and we got to get an AI music detection vendor and we still need the fraud guys. then, you know, now all of sudden you've got a vendor soup sandwich that you just don't really want to manage.

Jakob Wredstrøm

Hmm. And we talked a bit about this on the podcast with it with Khan, but I want to sort of maybe paint the picture a bit clearer. Maybe this is a loaded question, but I think in the past.

Andrew Batey

Hmm.

Jakob Wredstrøm

When I thought about string of fraud, I might have just like some association to the, similar to a call center somewhere, in, know, and someone scamming, but this sounds a tiny bit more like, you know, quite professional, quite intelligent, organized crime. Like what's the persona at scale that we're seeing here?

Andrew Batey

I mean, think even within fraud, there's different motivations, right? So if you're organized crime syndicate, you're probably money laundering more than anything. You're probably trying to move money through the system, clean it, and then get it into another location digitally. So you're cross-bordering money. I don't know if anyone knows how heavy it is to ship dollars, but it's quite expensive to take US dollars and ship them from Columbia. to doha. another approach, for example, is to turn that into cryptocurrency, pay streaming farms with crypto using the you turn into crypto with ATMs. They have ATMs everywhere, especially in America. You convert that to crypto, you pay the streaming farm operators off. It gets a bunch of streams on songs that you tell it to go play. And then that money conveniently gets delivered to you through your distributor or through your mechanisms to the entity you set up in Doha or Canada or Hong Kong. or Cypress wherever you want. it becomes a large money laundering operation. And because the music industry is unprotected and because you're talking about billions of dollars, it's very conceivable to move a few hundred million dollars this way without anyone noticing. And so if your sort of options are bury the money in the backyard or like you can only invest in so many car washes and laundry mats or whatever, like this is just another avenue. you to move this money because you're sitting on massive amounts of it. And in a sort of warlord type scenario where people in that jurisdiction will do what you say because you've effectively gained control, a lot of the people that are even KYCing themselves, like KYC facial check, that's already easy to sort of fake. But even if it wasn't, you could probably get a bunch of willing people to go ahead and put their face on there and do the checks with their passport. It becomes really a matter of like a count mouse game of sort of trying to combat all these people. And I guess my point is like, there's different motivations inside the fraud. There's that side. There's the side where I just want to steal pennies from everybody and them not notice. So I'm just running like massive amounts of these types of fraudulent streams. And like they might log in as you and or I'll log in as Morgan and play a song 10 times. I'm a victim.

Andrew Batey

You know, my account, won't say which platform it was. It showed me my number one artist I've never heard of before. But somehow this last year I played them over 2000 times and I've never heard of this artist or this band, you know? And so it happens. The thing is people don't guard their streaming service accounts like they do their banks. You want it to be easy. You probably choose an easier password than you would normally on anything else. Like you're paying a set fee. you don't realize the implications or sort of the sort of cost to the artist when that stuff happens. It's not directly obvious to the average consumer. And so I think that's there's all these types of frauds that are occurring and the motivation of fraudsters is quite has a wide range. And I think that's part of our struggle is always figuring out what is their motivation, what is their pathway, how are we going to try and combat them, how are we going to beat this? Even when I look at AI music detection. Two years ago, most of the AI slop was obvious. We would listen to it and go, yeah, that's an AI track. It's getting so good these days. It's very difficult to discern the difference, which is why we needed an internal tool. Because what we found often was that when we saw a lot of fraud, it often was tied to an AI generated track. And so we're not looking necessarily for AI and calling it fraud. But when we have 47 signals that this is fraudulent, And then it also is AI music. It's another confirming vector that just says, yeah, we were correct in the other 47 different flags that we flagged. So I think, again, all of those things become sort of critical overall. And the fraudsters have different motivations depending on what they're trying to accomplish. And music has by and large been like this massive unprotected bucket. And I laugh now because I use the water balloon analogy, which

Jakob Wredstrøm

Hmm.

Andrew Batey

If you think a little bigger, maybe the water balloon's giant and the music industry is just this part. And Morgan always talked about pushing them back to PayPal. But like if we just squeeze hard enough on music and just shove them back over to the other parts, that's really what we're trying to do. I don't think that the water balloon pops. I think it's more just like, let's just make the music part so difficult that they're just over there Venmo scamming people or using Zelle or something, you know?

Jakob Wredstrøm

Okay, well that's interesting. Just for the listeners, I'm usually sort of an early adopter of many things, but then I'm also very slow as a person on other things. And it's only really when I started building technology myself, like two years ago, where I figured out, I really shouldn't be using my password everywhere else, like the same password. Because when you know these systems, like how easy it is to crack these things, it's crazy. And how- of the internet is. But guys, one question I've asked myself sort of following your journey is like, how are you guys able to do this? Like it's a big claim and you know, if I didn't know you guys I would try to call bullshit on you but like I'm gonna give you the benefit of the doubt like I've known you for a while now so I'm gonna trust your word for it and everyone around you. So. But why are you qualified to do this? Like that doesn't sound easy.

Morgan Hayduk

I mean, look, like on the, why are you qualified? I'm definitely not qualified. I don't know, Andrew, on a technical level, I don't think either of us are qualified to do this. We've been very fortunate that from day one, we've hired really, really good people who are infinitely smarter than we are and pointed them at thorny problems that they've had a knack for solving. I mean, if you look at our team right now, not to go fully down the path of like the feel good part of building up a business, but like. I think 11 of the first 12 people we hired were now about 50 people, but 11 of the first 12 still work with us. It says something about their willingness to believe in the problem, believe in their ability to solve it, and sort of run down this path that is not clearly illuminated the whole way. We've been really fortunate to be just surrounded by awesome, bright, technical folks who have looked at these problems and gone, one, yep, these are real, two, technology can help solve them, and three, we're the people to build them. Andrew and I are the fortunate beneficiaries of both an industry that was receptive and a team that was willing. I think more than anything. And yeah, like I, I don't know if there's like a better answer to it than that. You can just do things, right? Like that's the, I know that's like a Twitter meme at this point. But you know, someone else, I love this quote. Someone said, you know, the whole world is just a museum of other people's passion projects. Right? Like we just decided this was a thing we wanted to do because we saw a need for it. And fast forward eight years, like, yeah, this is the museum entry for Andrew and Morgan right now. And a whole bunch of other people, our co-founder, Poria, and the rest of the team of our passion project, which is making fraud less prevalent in music and making artists' lives better. that's, I get off my soapbox now, but like that, the question, it's funny, because it's like, of course, who's qualified to do much of anything, honestly? When you're starting something that's brand new, you just have to sort of have conviction and then hopefully attract enough talented people that you can actually go do the thing.

Andrew Batey

think one other thing to add onto that too is that Morgan and I were passionate and props truly to Morgan because there was a point where, you know, if you go back and probably listen to our first podcast with you, we talked about our journey and talking about how we were in audit technology and then we switched into fraud detection. And in 2019, I did, we did a red team thing at the office and I was like, Hey, I really believe fraud is potentially a problem here, but no one really believed in the industry. It was as big as it was going to be. And so we didn't really go down that path at first. Like we, read team some stuff and sort of, it's funny, you go back now and all the vectors of attack are ones we now see regularly and hit on. our ideas were spot on in terms of how it was being hacked. the size of which we weren't, we weren't sure of, but I guess my point is props to Morgan on that because there was a real moment where we were like, are we going to continue doing auditing Or are we going to solve the problem? Cause if I tell you how many times a song was played, but then tell you that it shouldn't count, what was the point of knowing how many times it was played? Like if I tell you 100 million times it was played, but 40 million of those we should remove, then auditing audits for streams doesn't matter. and I really wasn't sure that I was in it to do like another startup, like start over. Morgan was like truly passionate about it. And I think that's where like being friends for 15 years, like really helps because I'm like, well, if you're going to ride, I'm going to die. So like, let's, let's ride or die together. Like, let's go. And like, we just, we just went that direction and said, it's almost like.

Morgan Hayduk

you

Andrew Batey

You have this like some cost of all the stuff you've built and you're like, you know, throw it out. Let's just start over. And I think that the piece that we're had that other people didn't have the qualification question is like our team was largely from another company we had built. And we could people we knew that had we'd built a machine learning product before. So if anything, it was kind of interesting because we were going back to old hat, like one, I'd been a hacker in the early days. So think it became easier for me to like red team products and think through how their minds might work. Two, we had a team that built really big, sophisticated machine learning and analytics products. So when you start going into ML and then eventually AI components, you start like you're already on that sort of trajectory. You have the skill sets or the tools to be able to accomplish that. That's just not how people were directed. Just like Morgan is former lobbyist, but he doesn't mean he doesn't know how to lobby. Like he's now a tech founder for like the last eight, 10 years, you know, it's like, doesn't mean he doesn't know how to be a lobbyist. just like. chose not to do that for right now. And so again, all these machine learning guys and machine learning guys we had, we had all the right tools in the right place. I think also made it helpful for us to just make that determination to go. And so you say like, why us? I think we are the most skilled team. I've had the benefit of working in lots of startups my entire career, and I've never seen an engineering team as qualified or as good as this one. We have multiple PhDs on staff that do crazy stuff. We have literal rocket scientists on staff. like PhD in like aerospace, like it's insane. When you ask yourself who built the AI music detection model for us, like who was the leading driver? Like obviously it was a whole team, but the guy who like actually ran through the door for us internally to help build it is a rocket scientist. Like it is, we have the most insane people who love music and want to defend music.

Morgan Hayduk

It's true. He's gonna love that shout out.

Andrew Batey

And I think that is like what makes this company truly unique is we have this amazing set of skill sets, but a passion to drive us forward. And I think that's made all of these kind of like, you know, maybe challenges seem more like opportunities.

Jakob Wredstrøm

There's a lot to dissect of what you just said, but if you just want to follow a couple of media when it comes to big companies acquire other big companies engineers that's in this space, that would tell something about how attractive in general such a position is. How do you retain such talent? Obviously some capital is needed, but I'm pretty sure you can't compete with Metta or companies like that try to be competitive at that level. What's your strategy?

Morgan Hayduk

I think, I mean, setting aside the like, you need to know you have a stable place that you enjoy working. to come to every day surrounded by good people working on challenging things. mean, I think all of those things are sort of table stakes. have to pay people, you know, fairly and ensure that they feel like they can, you know, have upside in the company, make sure that everyone in our case, everyone has an ownership interest in the company. And that was intentional from day one. I think the biggest thing is like, look, if you're an engineer thinking about whether you want to work at a startup or whether you want to work at a big company, it's if you're having the question, if you're asking yourself the question, you probably want to go work at the big company. Like people who want to work at startups, who want to work at startups, and they want to do it for tangible and intangible reasons. Some people want the upside. Some people want the, you know, the equity lotto ticket that they're signing up for when they get that first, you know, vesting schedule. But the other thing they want is autonomy and like flexibility and freedom to own something and not just, not just in an economic sense, but actually like own the deliverable. I'm not going to pick on any company in particular, but I imagine that there are large teams of very well compensated people who decide on what shade of button should be for six months at a time on one single product within one single widget of one single giant company. And if that is what gets you going, amazing, do it, please. All power to you. Like all work is good work. But if you want to come and like the example of the data scientist on our team who basically started AI music detection because he saw the need even before we did, you don't get to do that in big companies. You got to run through so many levels of corporate bureaucracy and sign off and agreement and alignment and every other buzzword for just meetings after meetings that you just sort of squeeze some of the, I don't know, the joy out of

Morgan Hayduk

of the freedom to build and create. And so I think for us, it's like making sure that people feel empowered to do that and have good ideas and want to bring them forward and actually want to go and execute on them. Because it's one thing to say, like, I've got a great idea and now I know I'm going to have the next. You know, two, three, four, five, six quarters, trying to advocate for my idea to have a chance to see it get built. It's another thing for, you know, with Andrew and I, for someone to come to us and say, Hey, like I've been doing this for a little while off the corner of my desk and it's working. We should probably turn this into a full fledged product, which is basically what happened with AI music detection. you don't get to do that in big companies. And so think that's the trade-off. think for a lot of, for a lot of, you know, engineers and otherwise it's, you know, there's economic benefits both ways. Certainly disproportionate probably upside in startups but more predictable predictability and stability and big companies but like the trade-off on autonomy is you know One has it and I think in a lot of ways the other just probably doesn't in most circumstances

Andrew Batey

think also building on that, Morgan and I and Boria were very intentional about company culture early on. One of our very first meetings was an offsite. We had this idea of a company and then we had this idea of what our culture was going to be from the start. We actually sat down and I was the guy. I openly talked trash about company culture for a good decade of my career and really didn't believe in it. And then I ended up in a company with 700 people. that we grew from nothing where the company culture was absolutely toxic and there was zero way to get out of it other than to sell the company and hope somebody else could fix it. And it was like the worst possible outcome I could imagine for what was my baby that I didn't want to associate with anymore. And so at the very beginning, we all sat down and said, what do we want this company to look like? What are our core values? And even how we do performance reviews, it's not how many tickets you've closed or how much code you've written or whatever. It's how much continuing education have you done? Do you have a bias for action? Like, are you actually building the thing off the side of your desk because you have ownership of it and you want us to, you want to see if it goes forward? Do you have good judgment? Like we've built our entire company culture around what we want our company to be. And I think where we see that is people are not leaving, but also when they do leave shout out Sina, he leaves and you know, like two months later comes back and says this was the most amazing place I've ever worked and I don't, I don't, I would like to come back.

Morgan Hayduk

Ha!

Andrew Batey

please. You know, and so like, it's all of these sort of artifacts that are left behind. And you ask how like we were intentional, not just about what the company looked like and what we wanted. But and like even our performance reviews mirroring to those things like that's how you're sort of graded is on how you're carrying the culture and all these aspects. But also, I think you build camaraderie or your team because of trust and trust is built through vulnerability. And so we've spent a lot of time building moments of vulnerability within our team, whether that's through our team events, through our sort of bios that we do. We have this thing at the end of the year, at the end of the week called Cheers, where they're completely anonymized and they're not political. Like if you give someone props for something, no one knows where it came from or what. And we've completely built systems around it. So it's automated. There's chat bots that the sort of prep everything, get it going, gets delivered directly to your inbox. We all read them together. We've built these moments of vulnerability. that don't take lots of time but have created connectivity in a way that is very structured but not onerous. And I think that intentional thought at the beginning has really carried us forward as a company. And I think stands, you know, stands up to any company culture anywhere because the culture isn't catered lunches and team events. Like we have those too, but no one cares about those. They care about doing the work with their best friends next to them. And how do you create best friends? It's these moments of vulnerability and trust that we've sort of. fostered and built and I think that was very intentional and as much as all the other stuff's great too like I'm really proud of that component because we actually did that day one like we had a day one sit down we're like what do we actually want here as a company

Jakob Wredstrøm

So you've been a part of building something before that you don't want to be associated to anymore for many different reasons. I won't go into that, but it sounds like there's some previous learnings. Does that come from an intellectual place or an intuitive place when you make these decisions upfront and you continue to withhold those?

Andrew Batey

think they're intellectual because we like very systemically set them up on purpose and like defined what they were. think intuition would imply that we were doing them kind of on the fly and just had the instinct to accomplish them. think we were very specific with that, at least in this iteration of the company from Morgan and me and Poria being together doing this. Do you disagree Morgan? kind of feel like that's

Morgan Hayduk

No, no. was just thinking, it's a bit of, I think it's a bit of both. mean, I think like, especially this last round of sort of products that we're launching, like that came very much from a place of feeling pulled in a direction after enough customer conversations and industry conversations to just know that there was some there. Some of it's a gut. mean, I think, you know, early on that the thought about, you know, fraud detection versus the auditing product, you know,

Andrew Batey

It has to be intellectual.

Morgan Hayduk

we were very fortunate to have one partner who gave us enough data that we could make a somewhat analytical and empirical decision about, but it still wasn't representative of the whole of the industry. had to extrapolate a little bit based on what we thought and felt. So I think it's, you know, the short answer is like, yes, bit of column A, bit of column B, but we've really gone into this last round of thinking through like, what are the things that the industry really needs? And, you know, and then two questions sort of came from that one, like what does the industry need to, can we build it? And in the case of a number of the products, we said yes, we can build them. In the case of one, we actually went out and found a partner who'd built just the most elegant and incredible customer due diligence, which is like KYC on steroids platform. And we sort of were fortunate to be introduced to them and sort of say, look guys, like. Maybe there's a world where if we set our mind to doing this nonstop full time for some period of years, we could get close to a place where we're doing a reasonable facsimile of what you've built. you guys have done it. This is the most elegant solution that I've ever seen for KYC. And you happen to have been forged in a crucible of another very challenging, high touch, high value market. For them, it was fine art. And so I think that was definitely a combination intellectual exercise where it was, what can we What do we have the capacity to do internally? What are the right sort of skills that we have already within the company? And which products can we build as a result of those? And then if there's still a gap, how do we go find the best partners in the world to help us execute on filling the last gap, which in this case was customer diligence.

Jakob Wredstrøm

You know, I know you guys have been in music for a while, especially you, Andrew, and you're probably expecting there could be a few hiccups in the way from now. Like you have a great proposition. You found the problem. You've built the product. What do you expect could go wrong while trying to implement this into the world?

Morgan Hayduk

I mean, for me, the biggest one is I think for everyone that we speak to about all of these different problems and the solutions that we have that pretty much everyone comes at it from the place of like, we ultimately want the best for artists. We ultimately are in this business. We're on the business side of music because we are trying to do the right thing by the creative community and by artists and songwriters and everyone else. The sort of big existential risk is like when rubber meets road and someone attaches a dollar sign to a thing that is beneficial, in the long run for the artist community? Does it fit within the tight budget of that quarter or that year's spending plan? And are you actually willing to invest in these things that matter? And so far we've been very fortunate, I think, in finding that the answer is yes. But there's a world where, when you talk about the incentives to do the right thing...

Andrew Batey

and

Morgan Hayduk

you've got to find a financial motivation and a principled motivation because oftentimes the principled one is insufficient. So I think that's the sort of existential risk for us is that, you know, the principles that we all agree we're trying to uphold, if we can't find an attendant financial benefit to the customers for doing the right thing, we may not be able to get the right thing across the line, even if it is entirely, you know, geared towards something that we all say we care deeply about.

Jakob Wredstrøm

This is not a question I usually ask, but you guys are grown up men. If you were to name, okay, if you were to degree alternatives to right now, like, know, why would someone choose someone else for something else instead of Europe right now? Who are you? I wouldn't call it competing with, but who are you fighting for market share in this sort of new opportunity?

Andrew Batey

My wife would disagree, but you know.

Andrew Batey

I think we have an unfair advantage because we discovered a problem so early and nascent and then helped define what the problem was. And in doing so became the solution for a lot of the industry. So I would say, I think that in terms of another company, that's not necessarily where we compete. I think the perception could be that, but I think that we are clearly the market leader. And I think we... We don't take that with a grain of salt. Like we deliver with our customers. Like we build stuff. If our customers ask for something, we don't like nickel and dime them and add like an amendment. Like we build the thing that they need and then we figure it out later because they're satisfied and they love working with us. And I think that is why customers stick with us. That's why customers tell other customers to work with us. And that's why we have the market share we have. I think where we actually compete is with internal teams who are trying to build like a version of this and, and they're It sometimes takes them a while to realize, wow, this might be tens of millions of dollars and a lot of time in an area that we aren't necessarily best suited to solve. Just like we didn't want to rebuild a KYC AML product when there was something that was so amazing that wasn't in music that should be in music, we'd rather just partner. It's really no different. So I think what we often see with our customers is that there's a smaller team internal that still works with us and we become an additive tool for their team to sort of like amplify all the things they're seeing or they would like to do but without them necessarily having to rebuild it from the ground up so where we often compete is like a build versus buy scenario not a like us versus another competitor scenario and I think that's often that's actually the part that takes a long time to get through I think.

Morgan Hayduk

Now, I guess the one other thing to add there though is that is definitely true on the fraud detection side, whether it's streaming fraud, claims fraud, account takeover defense, all of that work. AI music detection is a much more crowded bar. There are competitors in that space, big and small, they're competing at a whole bunch of, in a very similar way. As of right now,

Andrew Batey

Fair.

Morgan Hayduk

The jury is still out on attribution. So I think anyone who claims that they can do precise attribution of what actual underlying copyrights in the training set resulted in the output. I don't buy it. I just don't think that's a technical problem that's been solved yet. So on the detection side, I think there actually are competing options out there. for us, we have a couple of things that we're doing that I don't want to give the game away because they're actually like, I think they are the most defensible reasons why you would choose us over another. competitor in the market other than just on price. But there are a couple of things we've done structurally and how we're actually deploying that product that make a huge difference to our buyers and to their partners for what service they ultimately use. The other thing we haven't really touched on this today and we may not get all the way there, but that's okay. I'll just say it briefly. The other piece of the trust and safety operating system that we built is a recommendation engine. Now we had a streaming service that has tens and tens of millions of monthly active users. come to us and we sort of built this hand in glove with them. We're not the first AI-powered content recommendation tool on the market. In fact, there are many. We do think, to Andrew's words of having an unfair advantage, we think we've gotten really good at filtering out that which should not be influencing an algorithm. Our whole fraud experience has taught us how to find negative outliers and make sure that they don't make it into sales reports. The same logic applies to the recommendation system, right? In the case of music companies that work with us, we actually take our fraud reporting directly out of the recommendation system inputs. So you know you're not getting a biased output from an AI-powered recommendation system that has artifacts of fraud creeping into what gets recommended to a user. But we also just think that generalizes quite well to finding... suspicious behavior or problematic behavior in large data sets that you don't want influencing recommendations. So that's a product we're going to take across music, I hope, but also well outside of music. We're having conversations in film and television and publishing, know, publishing both in like the, not in the music sense, but in the, you know, sort of editorial and news sense. There are lots of interesting places to take content recommendations, but that's also a very competitive space. We just feel really strongly that the product that we've built is...

Morgan Hayduk

It's pretty exceptional. I've been floored by watching that team over the last six months build this from nothing. And again, going back to the question earlier about what gives us the right? It just so happens that the data scientist who leads recommendation systems for us did his PhD thesis on recommendation systems. And so when we had the opportunity to go, we might have a customer for this. Do you think you can build it? He was like, yeah, of course. Like this is. been training my whole life for this moment, put me in coach, and so we did it. And so yeah, anyway, that's another just example of a place where we're gonna have some competition, but I think we also feel pretty good about our ability to compete.

Jakob Wredstrøm

Amazing.

Andrew Batey

Summaries, we have a lot of PhDs here.

Morgan Hayduk

Yeah

Jakob Wredstrøm

Well, it's great. well, know, time is running out, but again, I'll use the opportunity to ask a few scattered questions just to sort of stimulate my own curiosity and take advantage of this. But I'm here to learn and hopefully the audience is here as well. Andrew, know that you've, of course also you've logged in, but Andrew, you've raised a lot of capital in your career, but you've also sit on the other side of the table. And one of the big things that lot of startups with this space meet is the reluctance of the size, opportunity, and music. Not because you need to necessarily come directly into that, but if you were today to raise more capital, let's say you want to raise some more capital, what would be your simplified argument as of why this would be worth investing in?

Andrew Batey

business?

Jakob Wredstrøm

your business in particular, but if you can dissect it into sort of the opportunity and how that can actually be worth in like return on investment.

Andrew Batey

Our business is easy. So these are some summary level stats for you. 75 % of music companies are acquired for under $15 million. The ones that are the other 25 % are typically user-generated, or sorry, user-focused. So think like streaming services in some contexts like AI music where it can apply to everybody in terms of people creating it. And the one other carve out is... If you can win a market and be systemic or a core need for people. When Morgan and I have Emporia first set out with this, we decided what we wanted to be. And I think it's not just unique to me. Morgan and I went to business school together. think in terms of like large scale, we were not interested in building a small company. We were interested in doing something that was positive, have impact, but building a large billion dollar outcome. In fact, before we started Morgan and I actually did customer calls, sat with lots of different bankers. We looked at the sort of like graveyard of other music companies that existed and talked to those founders and figured out where did they all fail? What pathway was wrong? And we set out, we set out with a very clear end point. We weren't sure we had a pretty good direction, like which way we were going to go. Obviously with fraud detection and stuff, it's moved. But the point is like we knew where we were going. We knew where we were starting and we sort of charted a path based on all of these external outcomes. So my argument to everybody is When I'm looking at a company or when even we've decided to invest our own time into this company, it wasn't a question of like, is this going to be investable? It's like we weren't going to waste our time if it wasn't. And I would say the reason I think BeatDap is in fact the investable company is that we will be a multi-billion dollar company because of all of these tools and aggregates of all the things we've done and the market share we've earned and the way that we treat our customers and the way that our customers appreciate the job we've done and we've fulfilled. That is why we are going to be a large scale company. And that's inevitably why someone would want to invest in our company. I think where most music specific startups fail is that they're thinking in terms of like a service business or something that's not going to ever be a multi-billion dollar outcome. And that's fine. Like you can have a great business. Like I think of like publishing admin, think of like rights admin in general, think of distribution. Like you can be a really great

Andrew Batey

profitable company and services are a great way to do that. People pay you for a service, you accomplish the task, you do those things. You can be a scalable service business, but it's hard to build a sort of investor case when you're so reliant on a service or you can't be a platform that scales. So I guess my point is that most music companies just don't fit the bill of that type of investment. And most music companies should really be geared for angel investors or strategic investors. Morgan and I knew from the very beginning that we did not want to take money from any strategic because we knew we needed to be neutral. If we were going to be the sort of like the, you know, the truth, the ground truth for everybody, we couldn't have even a perception of thumb on the scale. So we've never taken money from a DSP, a label, any strategic music investor. That was the much harder path to choose at the beginning because in the very early days, people who understand music are the ones trying to give you money. So we had to like really intentionally design what we wanted to be, how we were going to get there and then sort of not take the early off ramps that exist for a lot of music companies really trusting in that direction. And so I think that is the reason why we are investable. That's the reason why we've largely been successful. It also could have been a reason why we would have failed in truth, but we chose to go for that big. like this is what we want to do, this big, hairy, audacious goal, and that's where we're going and we know where we're going. And not sort of like the sort of path where we created a smaller business for ourselves.

Jakob Wredstrøm

and

Andrew Batey

Maybe that was too direct, but that's kind of how I failed to feel about it. I lied up about that, obviously, because I really believe in our company. And I also believe that Morgan and I wouldn't waste our time if we didn't believe it was going to be something great.

Jakob Wredstrøm

No, I don't know. No, I'm just trying to hear.

Jakob Wredstrøm

I follow the question to that is how big a part of that argument is based on you as a storyteller than the actual, you know, maths behind it.

Andrew Batey

Let me just see some of Morgan's models. There's math on math on math. Sometimes I'm calling Morgan like, how does this equation work? It's crazy. But his assumptions, I mean, the beauty of doing this with your best friend or a good friend, we went to grad school together. We were in a case traveling team together. Like we have looked at lots of different problems together and constructed models and stress tested each other. And, you know, we can scream at each other and still have family dinner. Like there, is like a,

Jakob Wredstrøm

I'm

Morgan Hayduk

you

Andrew Batey

You it's not that it shouldn't be that unique. That's how companies I think should be built. it is, you know, there is math there, guess. Morgan, please, I'm stealing words from you. You've done an incredible job with all the math.

Morgan Hayduk

No, I'm just enjoying the praise. I'm just going to sit here and revel in the bask in the afterglow. Let's keep going.

Andrew Batey

No, he builds incredible even early on. Like when you're looking at market sizing, like he'll do bottoms up, tops down, like we'll look at both different directions. So if this is the plan that we plan on, this is what we're charging our customers. How many total customers are there? What percentage do we think can actually win? What's the timeline to win that? What is that ladder up to from grounds up in terms of potential budgeting? If you go top down, this is the problem they say they have. This is how much damaging it is. If you take a percentage of market share and move this way, do those two things line up. They both line up. Check, check like. We are looking constantly at all of the math as we go. It's not like that is not there, but I think the best companies are part art, part science. And we have sort of the quant and the qualitative side together that sort of drive our direction. And I think Morgan and I have always been sort of the yin and yang. You know, I'm the one running off the cliff and he's the one making sure we have a parachute. You know, and I think that, but sometimes like in the case of fraud detection, I was ready to take my ball and go home and Morgan still wanted to play. You know, like it's, it's like a, just like a really like tight, you know, this is probably the wrong way. It's more like this. We're just like, you know, like we're, we're and Poria too, like he's so brilliant. He's the most gifted engineer I've ever had the privilege of working with. And, and he's, he, he ca I mean, I remember being in the road with him. We were fundraising in the early days and he had a broken laptop. So he like brought this giant, like keyboard with him and put it on top and like keyboard from the, from the hotel room. It was like the funniest setup. And I mean, I would never take any of that grind back. Like all of us sleeping in crappy hotel rooms, Morgan will never forgive me for the comfort in, anyway, there's one place I think Morgan lost some really nice sunglasses and probably had bed bugs. Like, you know, at the end of the day, All those like really make this whole experience so much more.

Jakob Wredstrøm

I have a weird second last question. I'm not going to make it the last question just because it's a bit of a downer, at least trying to understand it. So if both of you individually would mention one reason for you guys not succeeding, what would that be?

Andrew Batey

Okay.

Morgan Hayduk

I mean, I think it's... The squishy answer is to say that it's what I said earlier, which is that people actually don't believe in as much in what we think are the principled reasons to do this. I think that's probably the squishy safe answer. I think that the real answer is almost every problem is an interpersonal problem. And if you're going to blow up a company, at this stage, it's probably not going to be because you didn't carry the one in a model. Or you, you know, miss sized a market or you misunderstood a customer need. It's probably going to be because you, you know, set fire to your own tent is how I would think about it. And so like, that's the existential risk of it's, it's mitigated by, know, like Andrew said on the call, like both him and I go back a long way. And at this point with Korea, we all go back a long way because you've been doing this together and Andrew, we're going to back further than that. So you hope that having that like, you know, well of interpersonal relationship and trust, mitigates some of the risk of, you know, some interpersonal thing blowing up a business. But I think that is why, you know, more than anything companies fall apart is because, you know, it's not a product issue. It's not a technical issue. It's probably not even a funding or fundraising or a revenue issue. It's an interpersonal problem between one or more of, you know, the founders or going back to the culture point, it's an interpersonal problem between members of the team that you inadvertently fostered by having a shit culture. And so I think that's our, like always persistently the risk that's sitting there right under the surface for any business. It's true in ours, it's true in others. You know, we're still like fickle, fallible human beings at the end of every day. You know, we want to feel all of our, you know, hierarchy of needs have been fulfilled. And if something is out of sorts and we can't find a way to work through it constructively.

Morgan Hayduk

Then we'll probably work through it destructively and that will have a horrible outcome for the business.

Jakob Wredstrøm

Andrew, can you find something different than that? Do want to make a statement?

Andrew Batey

I think that's well said. Have no other statement. Look, I think Morgan's so eloquent. I love hearing him sometimes. Like he is very, very correct. Think this is a, every business has a potential people problem, you know? And to some extent our job is to try and make sure we don't fall into that trap. I think we've done our best to do that. I think we've continued to do the best we can to do that.

Morgan Hayduk

You

Jakob Wredstrøm

Mm.

Jakob Wredstrøm

Mm.

Jakob Wredstrøm

Yeah, makes sense.

Andrew Batey

It never goes away, right? It's like, look, like, was it Bezos that said every company is just trying to survive? Like, he just, you see, survive long enough. Like that's, that's effectively what every company is trying to do. Yeah. Whether it's now or 300 years from now.

Jakob Wredstrøm

The more positive question. Well, it's not really your question. It's just sort of curiosity. Like, I think sometimes in contrast to sort of the, typical understanding of the American, well, Canadian, founder and sort of this sort of, you know, going for the unicorn, association, it's more like when, when will you.

Morgan Hayduk

Alright, what's your feel-good question?

Andrew Batey

Yeah

Andrew Batey

Thanks.

Jakob Wredstrøm

Feel fulfilled in your pursuit. I guess that's sort of my question. Like do you imagine there's a certain? Tangible point where you believe you have succeeded with beat-down

Morgan Hayduk

You should close this one out Andrew because I feel like you're going to have a better answer here. But I will tell you when I was in 10th grade playing basketball, we had a sports psychologist, one of the guys on my team, his dad was the sports psychologist for Team Canada. And so he was like just kind of coming back after the Nash Olympics where Steve Nash played on the Canadian men's basketball team. And so we were like, we're a nothing team from a little town in British Columbia, 10,000 people. But we happened to have Canada's best sports psychologist as our team sports psychologist. And his name was Dr. David Cox and if there's no chance he's listening to this, but if he is, what's up David? But he used to just drill into us process, not outcome, right? Like you just do things, how you do anything is how you do everything. You do your absolute best in what you're doing. You run a good process and the outcome will take care of itself. I don't know when the end point is. Like the outcome is very hard to predict. I do know that as long as we and me been like, I speak for myself here, but I think for everyone also on the team, like the. The goal is to do the right thing by the company and by our customers every single day. And if we keep doing that, the outcome takes care of itself. If you start trying to fixate on a defined point in time whereby Q3, 2029, we are going to do X, not that you don't have goals, but like to be as specific as to sort of, you know, borrow some millennial language, like manifest the outcome that you want. I just, it just never, it's never worked for me. I've always sort of believed that you just, you know, follow the process, do the process as best you can, the results take care of themselves.

Jakob Wredstrøm

Did you have a?

Andrew Batey

Feel the exact same way. I know I love I love I do things because I want to do them, not because the outcome I think success is the consequence. You know, that financial like the sort of revenue outcome or the billion dollar is a consequence of sort of the product or the sort of the process you've done like as Morgan pointed out. So I do think says I love putting puzzles together, like and to the extent that this continues to be one giant puzzle we're putting together. It fulfills me in a way that isn't relevant or tied necessarily to the outcome on the end. I just know that I want to be solving big problems, and my brain wants to be solving really big, complex things. And I think that to the extent that we can keep doing that, I'm excited to be here. I admittedly thought we would have solved this puzzle four or five years ago, but it's still exciting. And the puzzle changes, and the sandbox we're playing in changes. It's just fun and exciting constantly. And I think that's part of why you do these things. I cannot imagine, as Morgan said so perfectly earlier, deciding on a color on a widget for some big company for the next six to 12 months. I just couldn't do it at all.

Jakob Wredstrøm

Make sense. Appreciate the time, guys. Let's make a of a date for a fourth time in like a year's time. We'll see where you guys up to. I appreciate you getting back. It's always incredibly interesting learning from you and listening to your thoughts. And it's very. For me to just open up and listen, know, not to pass any judgment to anyone, but like sometimes you gotta be careful with who you heed advice from. This episode was not one of them for me, so I'm just listening and learning. Thank you guys.

Andrew Batey

Appreciate it. Thank you. Do we get some kind of ring or something after 10 years? What do we do? We get like a gift? A letterman jacket like what do we get? What's the?

Morgan Hayduk

Thank Yeah, where's the jacket?

Jakob Wredstrøm

I don't know. Maybe. I'll figure it out.

Andrew Batey

Thanks for having us. Cheers.

Morgan Hayduk

Thanks, Jacob.

Jakob Wredstrøm

Here we go.