Invest with AI
00:27:51 · Kirk McKeown
[00:27:51] S0: You know, it's I'm I'm I'm I haven't been as close to it um since I started CarbonArc for a bunch of reasons, but you know, sort of uh the the net the natural is, you know, sort of what especially with the proliferation of card data, you know, sort of the number of folks that were trading names for the data rather than for the business increased pretty pretty meaningfully, um, which meant they were only there for the event and not necessarily for the turnaround of a business or what have you. And and and and I think I think one of the things that became most apparent is especially with like factor frameworks as they emerged as well, it was like the rise of factors and then the rise of data created this like really weird weird volatility cycle in a lot of names where mistakes or misses were way outsized, uh, you know, sort of uh impacted to the and then it would blow up people's portfolios and create all sorts of uh dislocations that just frankly, um, you know, sort of weren't fundamental. They were technical, but technical still hurts and technical still puts you in a really weird place. And so it just, you know, it it's gotten to the point where I think, you know, sort of one of the one of the big parts of the data business is either increasing conviction so that you can buy more on those dislocations, or um, you know, sort of competing, you know, in by figuring out what the data is going to uh you know, predicting the credit card data, being in front of those types of data assets to have more visibility into what's going on, right? Um it's those, you know, that's largely consumer, but I would say soft consumer and food. Um, you know, sort of uh, you know, um like you know, a lot of the big ticket stuff uh isn't well covered, um, largely because uh buying a mattress or uh buying something that is sort of big ticket is is not as episodic as going to Chipotle. Um so it's not as strong a card name. Um and then, you know, soft uh what I would call soft tech, you know what I mean? Um and online, uh internet and online, right? Um your hard tech is still, you know, sort of uh, you know, a more uh hotly contested data space. Um people are trying to find it. It's it's very hard. Um you can see it. Um but you know, sort of, and then industrials is actually really hard to build because it's like building a cottage industry in every vertical, right?
00:30:00 · Kirk McKeown, Brett Caughran
Um it's very hard to find horizontal data assets in in industrials. [00:30:04] S1: Yeah. Yeah, and sort of like you wrap it all together. I was talking to a friend who runs a consumer pad at a multi and he spends about three million dollars a year on data. And I just asked him, like, is that how does that show up in PL? He's like, he doesn't. It's just table stakes. I need to sort of understand why these stocks are priced at these levels. And I'm like, that's quite a big, that's quite a big nut to just under understand expectations. Um but sort of walk me through that dynamic, sort of how alternative data has become more beta, has become this now casting dynamic. And then I'd sort of love to understand how that evolution led to the sort of conviction to to launch CarbonArc.
00:30:46 · Kirk McKeown
[00:30:46] S0: Yeah. So, you know, it was one of those things where like when you started to look at, you know, how data was bought and sold um in the markets in the mid-teens, um, you had long lead times for legal and compliance and contracts, um, and then portability, usability, and flexibility of the data was also really hard. You needed a huge team, like we talked about, and you also sort of needed a balance sheet. And so that kept prices artificially high, right? And so it kept it up in the ether so that the people could buy were buying, and the people could sell, couldn't sell down markets, they had to keep the prices up to validate why they were selling the data in the first place and maximize their opportunity because they only had 500 qualified buyers, right? Uh the the the total customer base, potential customer base was really small. Um the view we took was um if you could smash down the cost of insight, move the insight instead of the asset, insight being an aggregation or what have you, a cohort, um, you could expand your TAM meaningfully to folks that just wanted to buy single tickers or um you know single zip codes, you know, um, and increase density and velocity of consumption um and and sort of uh build into that. You know, those ideas, you know, are are you know sort of a have aged well. Um you know, sort of uh, you know, we were early in the way we were thinking about it. Um but you know, sort of fundamentally the whole my whole the whole core thesis is when you when you when if you were pitching a name to a PM, you never wanted to go in and pitch Chipotle because the credit card data is better. You wanted to have three, four, or five inputs to sort of say credit card data is better, foot traffic's better, especially in the Northeast, where they're they've a they have a denser denser store store store base. Um we're seeing sort of that they increased ad spending throughout the quarter, and the reviews on Reddit are positive, right? Now you have surface area and a portfolio of inputs that give you sort of higher conviction into an event, i.e., you know, sort of in sort of how they're talking about sort of how they're gonna talk about it on the call and what they're gonna guide to, you feel good that there's good momentum going into the call. And so that was sort of that's always been my view, but like to do that costs millions of dollars.
00:32:51 · Kirk McKeown, Brett Caughran
Yeah. And so, like, how do you do that for hundreds of dollars? Right. And so that's why we started CarbonArc, um, which was because like everybody wants to, everybody still wants to make a decision, trade a stock, whatever they want to do. Um, you know, we just wanted to make it so that anybody could do it, and then you could start pushing it down market into, you know, sort of uh corporates and other folks that are trying to figure out a lot of the same problems that Wall Street are from the other side. [00:33:18] S1: To to some it makes a ton of sense and like this is how the ecosystem should operate. Also, if you have 500 core buyers who have to spend 300K for a data set on the potential that they use that three or four times a year, that's actually very good for the data incumbents. It's almost sort of a supercharge dynamic of the gym business model where you know 80% of your people never show up or once once once or twice. Can you just talk talk to us about the the recommercialization frictions, particularly in the really high caliber data sets? Like how do you get those, like if those data sets all played ball with you, that's actually leads to dilution of their revenue model.
00:34:00 · Kirk McKeown
[00:34:00] S0: Yeah, I know. So in the the the short answer is uh, you know, we we we we have done um we haven't sort of um done deals with all of the number ones, right? Like the the the nice thing is the number ones are uh people everybody knows, right? Um and uh and we've done deals with a few of them. Um we haven't done deals with all of them. Um but like one thing we've managed is we've managed uh, you know, sort of uh like a class uh commercial list where they have they they maintain you know pull position in certain spaces, right? So like, you know, it's it's it's gone into how we develop product, right? So like um we work with uh credit card partner and the product we deliver is a differentiated product versus what they deliver, um and it's fits in a seam of the market um that is you know sort of uh meaningful as a as a as a as a strong second source or solid first source um based on hit rate, size, and and price, right? It's uh you know, it's sort of we work with Facteus on uh as a partner, they're they're they're great guys, they're the you know, sort of incredible partners, very commercial. Um and we and you know, we created a panel of 10 million cards um that we sort of curate differently than they do and um and and and sort of uh we're at a price point that's that's very different in the market. Um, but we we we're very active with our partners and hey, we're in here, um, you know, want to make sure we're good, you know, sort of it all of it's really about being a counterparty and just making sure that like you do what you say and say what you do. And, you know, like a lot, like a lot of the space has been super opaque and just creating transparency, everybody's trying to get to the right place. It's a trillion dollar opportunity, and there's more room for everybody to win. Um, you know, the monopolists that sort of uh, you know, sort of try to block out the space and block out the sunlight. Um, you know, sort of the problem that they have is that, you know, sort of there's more data coming to market every day.
00:35:50 · Brett Caughran, Kirk McKeown
[00:35:50] S1: Yeah. [00:35:50] S0: Um CapEx isn't growing as fast as sources, and there's gonna be downward pressure on their names. And in card specifically, one of the reasons people were paying the prices they were paying was for stability. And what we saw over the last six months is that it's not as stable as it once was from a uh feedstock supply side perspective. Um, not to similar to what happened with change in 24. Can you can you say more about that? There was some there were some panel outages at different providers that created some instability in the credit card market. [00:36:19] S1: Um I'm not gonna get into specifics, but uh, you know, sort of that's the bank, that's the banks make a decision not to sort of distribute those panels. That's exactly right. [00:36:27] S0: And so like that, you know, but like if you pay high prices, you expect stability, right? And like, you know, it's it's why, you know, there was a re there was a re-rating in claims data after change, right? So like you know, claims data used to be more expensive than it is now. Part of that's because of the stability framework, right? People don't want to pay as much if they're not sure it's gonna be around tomorrow. [00:36:46] S1: What um are these sort of uh are are the motes of these number ones purely commercial, or does A like do AI agents allow like the the sort of number two and number three to sort of close like get get more parity on the tier one data sets?