00:40:05 · Brian Neider
[00:40:05] S0: Part of this will be, you know, purely opinion. Part of it'll be a little bit more factual to answer that question. But I think broadly, when you think about software, if if you were to talk to a a software analyst at an investment bank or something like that and how they might organize their coverage, the world would largely be broken up into application software on one side, infrastructure software on the other side. And the way that I learned twenty years ago to think about this, and I don't think that that much has changed, On the application side, you could think of something as an application software business if it's selling or if its users, excuse me, are are like business oriented users. And so think that that non techie people are the ones primarily using it. Infrastructure software conversely, usually you have... You will have techie people typically using it. That's how you start. Now each one of those areas would break down a little bit further. So let's take application software to start. Application software typically, once you break that down one level deeper, you'll have horizontal applications and you'll have vertical applications. A vertical application would be something that's very specific to a singular end market. So we have like a business that does software for the cardiac market. Right? So patients that have defibrillators or pacemakers, that would be a a vertical market within that... Within electrophysiology. You know, there's businesses that are public, like AppFolio, which has vertical software for the real estate market. You have companies like Procore, which are public, that is vertical software within the construction market. And so that's vertical. Horizontal is more software typically for a business function across any type of accounting software or HR software or marketing software. In those cases, it's typically not all that important what the end company does. If they're doing some type of marketing function, they'll probably want some type of marketing software. If they're doing some type of accounting software, they're probably gonna want... Accounting function, they're probably gonna want some type of accounting software. And so on that side, the application side of the house, you kind of have horizontal and vertical.
00:42:15 · Brian Neider
Now the question as it relates to AI and and how AI might affect it, and this is more of an opinion in my case than than than fact, whereas the rest is sort of, I think, most people would agree on. When I think about how AI would affect it, I think that in vertical, it's harder. Vertical software is extremely mission critical typically. There's a lot of nuance and know how that only can come through significant conversations with customers, understanding of what they're trying to do, what the outcomes they're trying to get are, and specifically how they go about doing their work. And so vertical software ends up being quite particular, quite tailored to that industry. And I think that while AI will absolutely have effects within vertical software, it's slightly more difficult because the nuance and know how in those products and how they serve those niche industries are so specific that it's it's pretty hard to replicate. On the other side with horizontal software, I think that that oftentimes is looked on, and and and I would agree, as a little bit more susceptible for AI risk. And the reason is oftentimes those software platforms are more one size fits all. So as opposed to having, like, really deep understanding and workflows about how a construction firm might need to build out their project, it's more catch all across an entire functional ecosystem. And I think that those workflows are a little bit easier to sort of ascertain, understand, and in many cases, that's that's where you might have issue. So when I think about on the application side where there's issues versus where there's not, probably more issues on horizontal, less issues for AI on on vertical. There's exceptions to every rule. If you move over to the infrastructure side, and these are more, again, like for software for techie users, there's different buckets within infrastructure, but you might think about network management tools, you might think about cybersecurity tools, you might think about developer tools. And within those worlds, I think that tools around network management, observability, performance monitoring, probably in an AI world doing a little bit better actually. There's so many more devices. There's so many more agents. There's so much more data flowing through the systems.
00:44:26 · Brian Neider, CJ Gustafson
It's going to be of the utmost importance to have sources of truth around how to analyze and manage all of that. So absolutely, I feel like there's a tailwind there. Cybersecurity, same type of thing. So much more going on, so much more important to protect yourself. I think that there's gonna be tremendous demand. The place where I think that is the most susceptible for AI risk is around just traditional developer tool systems, frankly, because the way in which software development is done today is just so different than it was done even a year ago, two years ago, that a lot of the tools that were created to produce that software manufacturing engine, they're just not as applicable today. And so it's not to say again that all of them are dead. That's not the point. The point is some of the problems that they were solving are no longer even an issue anymore, not because they've solved it, just because the way in which things are done is just entirely different than the framework under which they were built, and that makes it difficult. That's sort of our framing and how we how we break out the world and how we think about the AI risks, as well as the opportunities. [00:45:34] S1: I think examples go a long way. So I wanna go through a few investments that you've made at Lead Edge because I think they show three different kinds of defensibility, if you will. You willing to play ball in this? [00:45:45] S0: I'm willing to play ball.
00:45:45 · Brian Neider, CJ Gustafson
[00:45:45] S0: I'm willing to play ball. [00:45:46] S1: Yes. First one I got for you, Rentrak. I think it's in the movie industry. So maybe you could give me the t l d r of what do they do and what makes that business hard to recreate.
00:45:55 · Brian Neider
[00:45:55] S0: Absolutely. So Rentrak previously was a business called Comscore Movies, which was owned by Comscore, who's a public company, and it's a data provider within the movie industry. They basically collect data from tens of thousands of movie theaters in regard to box office sales. They become the sole source of truth for the performance of motion pictures. And so what ends up happening is they have long term data sharing agreements with basically every movie theater chain in the in The United States as well as, you know, a lot of the independent movie theaters. And they now know what the sales of each one of the movies are on every single screen, in every single locale, in every single geography, in every single zip, and they can deliver that data. And that's ultimately purchased by a bunch of different stakeholders within the industries, most notably the studios themselves. And that data allows the studios to make better decisions around their movie releases, around marketing budgets. It allows them to do a lot of the accounting, for instance. So when you hear a a headline that says, you know, x y z movie just grossed $50,000,000 in opening weekend, that data is typically gonna be coming from Rentrak. And then on top of that, you have a situation where you might have a movie where, you know, the lead actor is being paid some percentage of the gross sales above and beyond a certain figure. Well, how do they get those numbers? How do they understand exactly what it is? Rentrak is kind of the source of truth. And so it's a really nice business, and what's hard about it is they have unique relationships with thousands and thousands of movie theaters and movie theater companies to be able to receive that data and have it flow in in a in a sort of understandable, legible way. They have unique relationships with all the studios who are willing to have them be kind of the Switzerland in the middle of everything, managing all of the data and putting it out to everyone. It's really a unique asset that's very, very difficult, if not impossible, to recreate. You know, there's a lot of ways for them to leverage AI within their business for sure, but you can't just spin up a company to do this because the amount of theaters that you need to go out to to actually pull the data in, it makes an, you know, effectively an insurmountable challenge.
00:48:16 · CJ Gustafson
[00:48:16] S1: This is great because people always in a hand wavy way are like, oh, well, AI can't replace trust. This this this is a real example of it.