LaunchPod - Marty Styles === Marty: [00:00:00] We built the product called Play Finder. It's basically our product that allows you to, instead of using all the filters of the data that we have. You can just type in, show me Lamar Jackson scramble plays. And it literally should just go directly to that play for you because it already understands our filtering process. , it's giving you that quickness and that efficiency. That's really what I think the goal is of adding AI into our B2B side of how do we keep the teams. In the tool where they can get the answers they're looking for, but are we also helping their workflow? Jeff: Welcome to Launch Pod, the show from Log Rocket, where we sit down with top product and digital leaders. Today we're joined by Marty Styles, VP of Product at Pro Football Focus, the SaaS analytics platform used by every NFL team in over 225 college programs. But we're talking about their newest push where PFF is now going beyond the sidelines with a B2C play built to help millions of sports betters make winning decisions. In this episode, we discuss the playbook for taking a pro grade B2B SaaS product to the consumer market. How PFF built AI features that work for two [00:01:00] completely different audiences and the human in the loop process that keeps AI outputs accurate and trusted by NFL teams and everyday users alike. So here's our episode with Marty from PFF.​ All right, Marty, welcome to the show. This is one I am particularly interested in. Anyone who knows me knows that gambling is a fun pastime of mine. And you guys bring something I also love, which is data and insights into that world. I know you're, you're more than just a gambling side, but , that's what kind of pricked my ears up. welcome to the showman. Thanks for coming on. Marty: , thank you. Thank you, Jeff. It's great to be here with you. And, and I, myself, avid gambler here. I, I have to just say I do it a lot as well. Started off in fancy football and just naturally moved into it. It's just exciting to work for a data company in football, especially with one of the most like bet on sports. It's just every day is exciting. Jeff: So for, for everyone listening, PFF Marty, you are the VP of product over at PFF and basically started out providing insights to, I I think at this point, what every NFL team and well over 200, 225 [00:02:00] NCAA, eight football teams. Basically all went, you know what, if our data is good enough for all the top teams to use this to figure out how to be better there has been a huge growth in consumer gambling or legalized sports gambling, and there's an insights market there to help those people. Be smarter, move faster. So that's what we're gonna talk about is basically , how you guys have taken that and built that product and expanded into that market from what was a B2B background. Take away from there, like how did you guys look at that market and go, this is the world we're gonna move in. And why go B2B two B2C? Marty: Well, yeah, I think really what stood out to us is that, there's a ton of fans out there already who really are looking for sports data at all times, like outside of the game that's just really interested in a team or a player, and they want to follow that player and just like understand how they're performing week over week, however they perform historically. What's the upcoming matchup like? Who's gonna win that matchup? All of that interest around football naturally has a market already there. And when you start to include engaging products like fantasy football, it just increases the market [00:03:00] value. And so for us at PFF, we always thought, okay, well again, we grade on every play of every game for every player, and we give that information to the teams. Why not leverage that same analysis for those consumers in the market who are thinking about, maybe I'll do a fancy team for the first time, but I wanna make the right decision. When you think about sports betting as a market, it's even exponentially greater. I think it's about a $35 billion market, and it still is growing every day. Not even all the states are legalized yet. So that market is just growing exponentially, and we thought we're really set up really well for success if we bring our data to that market. Jeff: Yeah, and I, I think the interesting thing from the product side is, looking at the assessment, you have this, I'm sure the, I'm sure there's some finance and accounting people who would not like my nomenclature here, but you have this kinda like fixed cost asset. If you will, of, of the analysis you guys have done of every game, like you said, every play, every player. And it's, it's really deep my understanding is what it's like 200 data points per play. [00:04:00] There's like 80 hours , of analytics go into every game. And while some of my non-football, my friends may say it feels like a football game is 80 hours. In reality, it's not. So that's far more work going into cataloging it than it, the national game itself. You have this giant fixed cost , and a big market to all the , pro teams and college teams, but. It sounds like was some product to build to go to the B2C market, but you could still leverage the thing that brought you to the table. Right? So it's that, that very straightforward, two box, this is the upper right box of what you wanna go into market was Marty: and I think for us, we recognize the fact that. We are not the actual sports book. We are not the actual fantasy platform. But more and more consumers are looking to make data-driven decisions and are looking at paying subscriptions for those decisions. So I almost think of it as think about , your financials. Yourself. Maybe you're managing your wallet and you wanna make some good decisions. You might reach out to Fidelity and get yourself, you know, a financial planner. You might get someone to help you figure out what are the right investments [00:05:00] to make. And just like Bloomberg has its own just advice and things like that, we consider ourselves that as well. You're managing your, your bankroll, you're managing your bets. You wanna make sure that more money's coming in, then more money's leaving. So we're a really good person to have, besides you, to keep that thought. And there's subscriptions out there right now that are priced pretty well in the market and we thought, okay, we can match that same price when it came to football. Jeff: Then, I mean the, Just the validation of the number of teams that contract with you. I don't think it's very far off. I was gonna make the analogy and then you kinda did it for me of almost , the Bloomberg terminal , of sports betting or fantasy football. Which is, let's be honest, just non-monetary gambling. Marty: It's, it's. Jeff: Cool. So maybe, first of all, , what are the kind of core pieces here that go into, or come out of rather that, that 80 hours of analysis per game, like I assume it's not just freeform analytics. , there's actual insights to be had here. Marty: Yes, yes. So I, I won't get too far into the details 'cause it's eight hours, it's a long time. But just imagine you [00:06:00] have someone who's watching the game and they're taking in first the basic information like the jersey numbers, the players' names who are on the field. Which positions are on the field. Obviously you've got your quarterback, your offensive line, but , are we at a three receivers set up? Are we at a two receiver set up that basic information is first charted? Then we think about the player participation, like what are they actually individually doing on the play? So is the wide receiver blocking on this play, or is the wide receiver going out for a post route? Or is the offensive linemen pulling for the running back or are they staying put and actually protecting the quarterback? That information happens. And then the final stage is our analysis. We're thinking about, okay, let's start to grade each player, right? So, okay, let's say you, Jeff, you're the receiver and you're going up against that cornerback. Let's say you miss your block, we're gonna have to give you, maybe a minus grade for that block being missed. However, maybe the next play. You hit that post route and you catch the ball and , it's obviously past the 10 yards and now we got our first down. We're gonna grade you appropriately for that [00:07:00] success. And so those 80 hours is Ted going through what's happening for every player, grading every player, and then obviously that analysis that goes along with it, right? Like success rate and that. So there's a lot that goes into it. I really appreciate our data collectors here at the company because they provide us in product the ability to create these beautiful. Experiences to help. To help the. Jeff: Right. And that, if I understand, that's part of what differentiates the product and, and part of kinda why this has been successful is this end of , it's one thing to kinda collect analytics from a aggregate level, stats, all that, but this is trained human, experts actually analyzing these things and providing that next layer. And so for someone to come in and try to surpass what you guys are doing, it'd be a huge investment just to, just to get near it. Marty: Huge investment and, and I think that's why I was impressed with PFF when I first heard about what they were doing, and that's why I joined the company is just the, the level of just class and just [00:08:00] professionalism this team takes to do that work. It is unmatchable. Like there's not many companies out there that can say they do what we do. And again, that's why we've been able to have all 32 teams use our product and we have over 225 NCAA teams using our product because they understand how much effort is going into collect this data and the amount of analysis that comes off of. Jeff: So I guess now, now think about the B2C product, right? Like, you guys have clearly delivered a ton of value to the super mega, ultra pro level to have every single pro team using the tool and a huge number of, of, college teams. And, we can all be honest. The top college tier teams, their budgets are probably not far off of the pro teams. On the B2C side, though, I mean even pretty intense data nerds are gonna have a little bit of trouble operating at the capacity of, of, a full analytics team with a NFL team. How did you start to look at, okay, we have this product and this, this rich data, but what needs to be done differently now that we are serving, a giant number of individuals versus, a tight market , of very [00:09:00] professional people. Marty: I like to think of my product team as a chef when you think about it, right? Like you've got all these raw ingredients, you've got the grades, the raw data that we've collected just like the matchup analysis, but then all of those things separated you might not be able to actually action upon, right? Like if you're a sports better, and I just give you one piece of that data, you might just say. Well, what else do you have? This isn't enough for me to make a decision. Yeah, they have a great matchup, but now I don't know how he's performed in the last five games. I don't know, if there's any injury I should be aware of that's gonna affect this game, that information like consolidated, can actually inform your bet a lot better. And so I like to think of my team as like that chef cooking up a stew and trying to make a more like suitable dish for you out of those raw ingredients, right? And so that's what I see the PFF plus subscription as, right? Which is not only are we going to give you that mashup analysis that has historical performance, but we're gonna wrap it together and tell you, okay, the probability that this bet [00:10:00] covers is much more higher because of data 0.1, 2, 3, 4, and five, all being great. versus there's maybe two out of five data points that we have on this player, and this isn't really a good bet to think about. and even if it's not a good bet, that might make you think, let me take the under instead of taking the over. And so that inverse is there. We have very complex data and the teams really want to see that complexity. But when you think about the consumer, you really, from a product perspective and a design perspective, the job at hand is how do you simplify it down to what's most important for the customer to understand? Jeff: Right. And so I guess this kind of key insights feature, which is take this data and how do you kinda synthesize it with LLMs and create, like you said, that kind of head-to-head matchup, what's gonna happen , and what are the insights there? So when you're thinking of that piece, who are the people using that and what's, how did you kinda determine what they actually needed there 'cause like you could have casual bets, just betting on a random game all the way to people. Trying to think about, betting across 12 games in a, in a week [00:11:00] or someone trying to model their entire fantasy team against three other teams or something. Marty: You know what was interesting is we ran a lot of surveys the prior season before Key Insights came out, and one of the things that was like the resounding point was, Hey, PFF, you have all this great data. But it can be at times be overwhelming. And I really need a simplified way of understanding what's the key takeaway you want me to understand from this player? And what really resonated with me was the fact that I heard that from both the most advanced like sharp betters who are more like building their own models all the way down to the casuals who had just met my brand for the first time. And that was really the premise. building key insights was thinking about. Okay, to your point, how can I let you understand the most you can how can I give you the best key takeaway about CD lamb's upcoming matchup without giving you report one, report two, tool A, tool B. And that's really, I think, the best part about the introduction. Of AI into this product development market, which is you can now leverage AI and these LLMs to [00:12:00] help you simplify the most complex data using natural language. And so when we started to do product testing with Key Insights, it took some iteration. Because we wanted the LLM to keep our tone, how we talk around football and understanding the matchups and not just spit out just stats. We wanted it to follow that natural narrative of, like, think about it, if you're listening to maybe Chris Collinsworth, our owner on Sunday Night Football, he does a really good job of like summarizing what he's seen on the field for anyone at home to understand. We wanted that same kind of narrative to come through when you read our key insight of like, oh, this feels like I'm just listening to Chris. that to me felt like we had hit the point, like after iteration, after issue, where anyone could hear and feel that message come through. Jeff: , that gets into an interesting part about LMS in general in AI, is how did you match that tone that you were looking for? Because one thing to have the data and synthesize data, and like you said, it can spit out facts and figures that's some level kind of quantitative modeling [00:13:00] that the computers , are typically good at. But how did you get to a point where it was delivering , in the way you wanted tonally , Marty: for us it was probably more of the, the copy text, like situational optimizing. We had to basically look at, and this was the benefit of our business. We have a broadcast side of our business where we actually sell our data directly to some of the broadcasters, Chris included. And so we have tons of just historical. Scripts on how they read the plays, how they speak to the plays. We have our own content. Rogers here at the company and they write these great articles around football and what they see. And so it was really us just taking bits and pieces of that with the metrics themselves and saying, Hey, LLM, this is how we would say this. And instead of it just spitting out those stats, we go and iterate and iterate and iterate on just the sound and the tone that it would use. , but I think that's, again, the power of our business. We already have a lot of that content at the ready to leverage and so , it wasn't hard at all for us to continue to iterate on that. Jeff: It is funny. I think that's one thing I've heard [00:14:00] across teams has been. This piece about training and iteration. Content wise, like we, we use it for some content stuff here. And similarly it's, it outputs and, and you have this giant kind of. Base of, the work product. You say, this is good, this is what good looks like. And then you start to go through, well, you did this part well try to do this differently. And then, do it again. And well, hey, you got a little bit worse here, but this is better and here's why this is better. And just keep going and going and going until you get Then I guess there is , the element of hallucinations happen in LMS periodically it's a risk from a data perspective. How did you guys find to prevent someone might ask a question , that's a little outta left field. And typically a problem with LMS has been, they'll answer it no matter what. What's the safeguards , or do you find a methodology there to ensure that when people get a little more crazy with their questions, they don't start to get just wacky answers? Marty: So , to be clear, we don't actually sell the chat bot experience like all the the LLMs that we use are for summaries. But with that being said, internally, we get the reply back from the L LM and we'd say, okay, that's not how we would say that. [00:15:00] And so it's been a really interesting, like shift of our business model where we're leveraging more of our content writers. As more QA now as well for , these natural summaries we're creating. So before you see anything on the PFF site, it's already been QA and editorialized by our team, I really think that shows how well we've done the transition a little bit. With this new AI and it coming in and leveraging our own people to say, okay, how can we be more of the validation for what's being said? Right? Because the last thing we want is for, a consumer to see a key summary or that key insight and make a decision. And we realize in the back end that, oh wait, we got the matchup wrong. Or wait, the. Jeff: a bad idea. Don't make that. Marty: no, we don't want that at all because like, again, this is a subscription model, so there's a lot of QA that happens Jeff: Yeah. Marty: before you see. Jeff: That model I have heard across so many verticals now of, it is generated by AI , and the process is sped up. But at some level there's some [00:16:00] form of human check and that goes all the way to medical, where doctors can have conversations transcribed and everything summarized, but it won't actually prescribe the medicine. It'll, it'll remind the doctor, Hey, you. You said you wanted to prescribe this, do you want to do that? But they have to check it off and go through a process to ensure that they do that to architectural, project companies. A large ERP. We had a guy from Del Tech on who talked about the same thing is in those kind of architectural processes and project management. It's still going to do some human in the loop validation before , it executes anything because there's just a high cost to gain that wrong. Similarly. On our end where we do digital experience analytics in this case we would, sit on your app and if people were running into frustration or problems getting these key insights or using it , in whatever intended way, we'd be able to flag that automatically to you and explain what the issue was and why it was impactful. And the whole value prop there was the AI watches the sessions and tells you what's important, but. Very quickly. If you give even a small number of false positives, you erode trust incredibly fast. Similarly [00:17:00] here, if you give a couple bad bets, people are gonna lose faith real fast. Marty: Right. we definitely think that type of technology would be useful. I mean, when it comes to our product, we want to continuously be monitoring how our consumers are reacting to it, what their pain points are. A technology like that would just help the product team here just react a lot faster. So I, I definitely could see the benefit. Jeff: You guys also have the prop be tool, right? So how is that different? And then kinda how did you kinda weigh in, we need this key insights functionality, but we also need this other one. Like where, where did you draw the line of this is functionality we need to build Marty: Oh man. So I'm, I'm really excited about this product. It just actually released this week So it's in the PFF app now this week. Consumers can go in, try it out. We basically, the product. Is a player prop tool product. So with looking at the props that are in the current markets, so for instance, maybe passing yards, we're able to tell you for Lamar Jackson, his upcoming matchup, it's a great matchup because he's got a good offensive line. His wide receivers are gonna have a good matchup versus those cornerbacks in that defensive [00:18:00] back group as well as like from a scrambling perspective, that's always an opportunity for him to find ways to get more yardage. So. We thought, okay, how do we just marry everything we have on that player, including what his PFF information is, but also what the market is telling us. So if he's at maybe a minus one 20 in odds right now for his passing yards, we're able to look at that odd based on our PFF data and historical data and tell you, is there an edge there for you? Like, is this priced in the right way? And it may be , in the sense there's an edge there. Where potentially it's undervalued. And if you were to place this bet, you're actually gonna make more upside than you might have thought. And because of our model, we can show you that. So last year when we did the key insights, we quickly saw how much users were coming back to the tool weekly, and then through our surveys we were recognizing, they were like, Hey, can you put this with the actual prop? Like can you tell me like this key takeaway? Is it good for the over or the under? And so it just naturally was a good starting place to get to where we are now with the player prop tool, [00:19:00] which was really the orchestration of everything we have available in that player. But like I said, we're a chef. We're picking the things that we feel are most imperative for you to know, not just every data point we have over 200. We don't want to give you everything and overwhelm you. So it's been a really cool transition. Jeff: , I think that's something to, to pick out real quick is this idea of, watch behavior, understand how people are using it, but also, contextually with the rise of things like draftking and, and FanDuel. Like those are. Typically their big use cases tend to be as much around prop bets as maybe sportsbook in Vegas. Were much more focused on the games themselves. I'm curious kinda on that front when we talked earlier, you had mentioned one of the big focuses, from a persona standpoint is intermediate betters, like not the super, super expert not the super entry level beginner, but how did you determine, this middle ground or mid-level person? Just doing semi-regular wagers was the way to focus. Marty: , it goes back to my example with Bloomberg. There's that person who trust is now getting started on their financial journey [00:20:00] and they're probably just. Asking friends and they're probably doing a little bit of information gathering, but they haven't yet subscribed for more of like a data-driven secondary tool and that would be that intermediate who's doing that more data-driven. They're looking for a product to actually help them on their journey. When you get to the advanced level, if you're sticking with that Bloomberg example, you probably have already gone as far as like, I've got my planner, but I've also got my own modeling. Of how my stocks are gonna go. And we have that same like persona within the betting industry. You've got what we call sharp bets or more advanced betters who are at home building their own models, leveraging AI and advanced modeling to figure out, okay, where's the edge in these pricing? Where can I get some value? Also, here's all this great football data I've probably scraped and gathered from different sources and putting that into their model. And so I would say we felt like, okay. The good spot to start with our data is probably that intermediate. Let's be that secondary screen for someone who's at like DraftKings or FanDuel. And right before they press the button, they're [00:21:00] like, I want to verify. I want to make sure I'm making the right decision. And the casual may just say, you know what? Oh well, I'm gonna go for it. My gut tells me yes. And that more advanced player is saying, what does my personal model tell me? And so there's a lot of value in that middle ground, but it also gives us space. To actually start attacking the other two, the bookends, Right. If we do this product right, hopefully the casuals will come in, take a look at it and say, this is why I need to be a little bit more data-driven. I'm gonna give this a shot. But then it also gives us the ability, maybe we go deeper, maybe we go more into arbitrage, betting more into these different modeling types of betting use cases where the advanced players may want to be with us. So we think this middle layer gives us an opportunity to attack both sides of the market too. Jeff: Have you guys looked at all about the idea of I know a lot of friend groups there's always the kinda one or two who are. I, I'd say far more serious about it than I am. They're not quite, , the top etch a lot of people who I know who, you run into periodically where they're just very serious. People make the most, but like, if they were to tell me, 'cause I'm, I'm, [00:22:00] I like it, but I'm not the most diligent about sports bet. It's more just like, Hey, they'll make the game more fun. And I tend to like the, gambling end of it. But if one of them were to tell me, Hey, this gives me a real edge, I'd probably go check it out. How much Is that a part of the strategy here? Marty: A lot of it is from a product marketing standpoint. Like when our, one of our slogans right now, or the slogan we've gone with for the 25 Cs is making winning decisions. How can we help you make a winning decision? And it goes to your point, which is if that friend of yours says, I just made a great call, like I found this edge and I make more money on this bet than I thought. To your point, you're probably gonna hear that and say, how'd you do that? I want to track that same process for my own betting needs. We're trying to follow that same context when product marketing our new product, like the player prop tool. We're hopeful that with our marketing strategy, we can get in front of consumers who are looking for that edge, who are looking to get more out of the experience. And we also think that's gonna drive. Retention is gonna drive repeat usage because once we get that hook and you feel good about that first bet, you'll probably [00:23:00] potentially come back and want to try it again. And I think that's just gonna help us from that engagement perspective as the season goes on. Jeff: Right. I mean there, there's so, so many natural like viral things in here where people always wanna brag about wins that they got and people want to tell their friend, people always wanna look smart, right? So how do you, how do you kinda leverage the, , the downside though is if someone tells me about it and I try it and I lose money, I mean, be like, what the hell man? That's a bad tip you gave me there. Marty: Exactly. Exactly. And I would say that's probably something when we went into this betting market, we were really conscious about, we were like, we don't really want to come into this market and say here's our pick. We want to come into the market and say, here's our analysis that goes with the prop that you're looking at, and still leave the decision with you. But now that you've got that information, it's more informed. Even if that bet doesn't play out the way you like, I'm hoping that the process like that engagement of like you went and got the data, you felt confident, that's the process that's we're hoping is the Jeff: Yeah. Marty: team, per se, to come to have you come back. Jeff: Have you guys seen what that kind of like aha moment is? Because I know right? If [00:24:00] you look at like Twitter, early on it was, you followed X number of people, I think, and that was actually a huge part. Wasn't, it wasn't even like posting, it was just you followed X number of people were consuming or right. Facebook, I think very, very early on had some metric around connections that you generated. If you had like three friends on the platform. You were like so much more expected to last and have a long time what does that look like here? Because I, there's gotta be some number of, maybe one successful bet or one bet , or one insight, but like at some number it probably becomes validated. This is, this is a part of my habit now of how I do this. Marty: So for us, the interesting part though is we're not really much aware. We're not of what your outcome is . And so with that, what I have to look at as a key KPI is what is your repeat usage like? , the key insights is just a page of insights. So what I'm looking for is how many times in the session are you coming back to me? What's your week over week? , repeat users look like. If I see that teeing up for you, this gives me and the marketing team an opportunity to reach out to you and say. How are things going? Like here's a little bit more information that might help you out on your journey, or have you checked out this [00:25:00] additional product? So for us, we can't really get too in the weeds to think about how many bets is it going to take for that user to come back because I can't control that. But what I can control is. Did you enjoy the experience you had in front of you and making that more of an engaging one for you? But I am interested hopefully this season with the introduction of the player prop tool to understand, hey, what did that handle, what did those bet outcomes look like? And we'll run surveys to try to understand that too. It makes for an interesting problem to solve for sure. Jeff: When you start to get more into the obtuse. , or hard to track value props. I guess like the value prop is very clear, actually not obtuse at all. , it's, people can make better decisions and hopefully, if you outlay X amount of dollars with this information, you can make 20% more than without it. But like you said, , it's hard to track that outcome even though it's easy to understand what it is. Although maybe, like you said, the prop heading makes it easier where you can start to see, they looked at these five, how do those five play out and that kind of thing. I'm, I'm curious if we can maybe regress a step. And think about the tech behind kinda where you guys have gone and your use of AI internally and how you're using that to better present the [00:26:00] data. There's the conclusion. One can draw that a lot of the kind of backend analysis work that you guys are doing, , that 80 hours per game, some of that could be potentially replaced at some point by, AI vision models or any of that kinda stuff. How do you think about that? Marty: Honestly, I think it's really innovative and it's the right thinking to have. We as a business obviously want to make that process even more efficient. So we're exploring things like computer vision, like more vision modeling. And really what I'm excited about with it is not just the efficiencies, but the new analysis that can come from it, right? And so I'm really looking for a world where now this creates even more data points for us to collect. Think about athleticism, think about speed that's happening on the field. That opens the door a lot faster to gather those different types of data points. And then from us on the data science side and product, we can now go off and create even more digital products from it. And so I like to think of it as an introduction of what new things can PFF provide, not just consumers, but now the teams too, right? Because they'll be interested in that data [00:27:00] too. Jeff: It's an interesting thing to look at because if I came up with it and you guys think about it, there's probably other people do you think there's like unique insights that the human level brings on that, . Marty: I think there is. I think there still is , the person who's looking at the screen, who knows the player, I think they have the ability to just know a little bit more about like, okay, this is what Lamar brings to the game. And this is why he chose to scramble versus not. Or they can take analysis to think about. Alright, we've got Aaron Jones' running back. He tends to want to do more inside than outside run. Not saying he does, but the person watching has that moment. Now to think more about the why. The computer vision takes more of the job of, oh, okay, what's happening? I'm just gonna go ahead and start collecting that. So I'm really excited for the time. It gives the human to think more about the context of the game, instead of just every finite detail. We can train the modeling to do that for us more. . Jeff: That's one area, kinda to your point that I've seen be an interesting kind of combo , of almost bionic humans sure, almost anyone can gather all that data. You can just throw a vision model at [00:28:00] something and money, and that's all it takes. But having the expertise to know. What are , the three questions you really need to ask in this situation or what's , the bullet point to bring out? That seems to be where a lot of this differentiation in any tool that's using AI like that or could be disrupted by ai, like that is going to come out Marty: I mean, it also opens up the door for, similarities and analysis of like, okay, I'm watching this film right now, but who plays similar to Lamar in a sense? Is it Jayden? Is he running a similar scheme just like he is? Oh, AI already recognizes it, and here's some additional plays you should look at. me as a product manager with the product we have, which is ultimate that we sell to the teams that whole product is really just the premise is , getting you directly to the film, having you start with our reports, our filtering, and get you down to the film that you're looking for. But once you get in that film, as a product person, I really want you to stay engaged. I'm like, how do I get you to watch another piece of film? And so if I can leverage AI as a sense of like. What's similar here are they running a scheme just like the Giants would, are they running a different front just like the [00:29:00] Saints would? Okay, we've already identified that for you. Here's another key insight. Why don't you check this video out? If I can keep that session time longer. That's a really good key indicator that that product's working. It's also another reason why we built the product called Play Finder. It's basically our product that allows you to, instead of using all the filters of the data that we have. To get to that film, you can just type in, show me Lamar Jackson scramble plays. And it literally should just go directly to that play for you because it already understands our filtering process. , it's giving you that quickness and that efficiency. But that's really what I think the goal is of adding AI into our B2B side of our business is how do we keep the teams. In the tool where they can get the answers they're looking for, but are we also helping their workflow? Like the weekend's over and now you're working on the scouting, the next team. I want you to get that scouting report done a lot faster so the players have it in their hands, that's one of the key pieces I'm hoping to solve on that side of the business. Jeff: It's striking me as funny how similar some of the goals of [00:30:00] our product teams are across both our companies. Where, I as I talked to you, we do digital experience analytics , and historically, we started from providing session replay. The difference being yours is video of people , in football games. Ours is a video of users, of people using your application. But it's the same idea, right? It's how do you query? I wanna know where people struggling with this, funnel or how are people struggling? Putting together a system of prop bets to analyze. But it's the same idea of how do we surface. That kinda situation without you having to dig, how do we make it so you don't have to watch, you can watch more situational video of exactly what you wanna see, not have to scrub through, a two hour game to find one play. Marty: I also think it brings up a really good point about where we see the industry growing with ai. Like I'm probably the type of person who would say, I want to first find the core problem, and what is that a problem to be solved? And then figure out, can AI help me enable this problem to be solved a lot faster? And I think the good product teams of the industry. Are following that same philosophy, which isn't just splatter AI everywhere, but [00:31:00] is am I still solving the core problem that the consumer has and then leveraging AI to do that at PFF, our vision for, 2030 is to be a more AI driven analytics company. And so we're taking a really keen lens across our portfolio and saying, where are the right places to start introducing AI in the film process? Like you're saying, that's a really good place to do it because again, it's the speed of get me to the next video where I can see the next insight, And that's where that can help. Jeff: It's how do you surface, in a world where there is near infinite amount of video you can watch. How do you. Proactively surface. Here's what you need to see. Here's the important bit that's going to help you in whatever your endeavor is. It's cool to see is a lot of just the future of what lays ahead of you guys in a industry like this where. It is an industry has moved very quickly , or is in the process moving very quickly from a world where there was all this kinda like by gut betters, right? , and anyone who had a system was advanced and now it's more and more, and this just enables that further push. I mean, I can envision a world where it be [00:32:00] ludicrous not to make a bet without this kind of insight. Marty: That's the, that's the world I'm trying to live. Right, right. That is world. That's exactly what our mission is, is to make more informed betters. Right. Not just gut feeling, because I, I'll tell you that, like, that feeling of losing a bet, I, I personally don't like, I hate that feeling. So like, the more data I have, the more confident I feel. Trust me, I'm gonna ride that bet till hits. I just think we are hoping to change the market with what. Jeff: Yeah. One last thing that I always think about when I thought about this very early on just came to my head is I've known so many people that don't love sports, but you're in social situations where you need to talk about it. One last application for, you want, you want consumer applications. All these ar glasses are coming out. Just throw up something on one of those, AR glasses so that as you're having a conversation, you can just sound like you know what you're talking about. Marty: Right, right. Did you know that? He's like, the success rate is 50% for outside zone, like John Robinson. He's gonna get it in the end zone. Like I do want to still be at the bar. Being able to hear people spit that PFF knowledge, that still is the core of that consumer [00:33:00] business. It's just giving our data analysis to everyone to use, Jeff: I would buy that tool too. I mean, that'd be great. Like being able to sound like the most knowledgeable person there. That sounds fantastic. Marty: Right? And everybody's just like, well, no, no, we gotta invite Jeff. He knows the game. You gotta ring him. Jeff: Oh, I gotta remember my glasses shoot. Marty: Right. Jeff: no, I'm excited to see , where this ends up in a couple years. 'cause I think there's just, there's so much runway here. It's a cool market. You guys got an interesting take on it. It's been a pleasure having you on Mario. Thank you so much for coming on. This was super fun, super enlightening. If people wanna follow up and talk to you about ai sports betting, football, any of that kinda stuff, is LinkedIn the best place or is there somewhere better to reach out to Marty: Best place is LinkedIn. You can also follow me on Twitter and the PFF team on Twitter. We talk about football all the time. Jeff, it's been a pleasure to be with you today. I hope you win some best this season. And we'll be talking for sure. Jeff: I'll definitely let you know. Thanks for coming on. Have a good rest of your day. Marty: Thank you.