I'll talk about for a moment, best embodied by the speed of learning. It used to take us two weeks to get a campaign to a statistical significant data set where we could really confidently grow user acquisition for an advertiser. Now we can get to that significance within a day. We're also predicting impressions across the entire ecosystem at a rate of about a billion a second at this point. And so you think about how profound that is for a business like ours that came from linear regression. It gives us a much more effective prediction engine than we had before. Now, in terms of executing against the roadmap, in Q2, we had what was an innovation that I think is representative of the type of breakthrough that we could see throughout the year. I think the pace of innovation would suggest that we'll see more. what it looked like for us in this past quarter was the feature set that helped us underwrite the value of a particular impression was enriched. So if you think about the variables that we're able to look at to make a decision, that expanded. So we get to see more of a full story of the user's journey. And that certainly acts like a force multiplier in terms of our predictive power. We were able to roll it out across all of our user acquisition models in Q2, and so those gains are now embedded in the system. While the release is episodic in nature, the gains are durable, and that is a window into the benefits of Cortex. I think we're very excited about what the next few years hold, and there are some very well-known peers and businesses that are chronicling their journey through neural nets, or what we call Cortex, and it's very exciting for us. Now, the second question about AppsFlyer specifically and mobile measurement, as you mentioned, I'm happy to address that one directly. AppsFlyer has been a really clear partner to us for well over a decade as a mobile measurement partner, and the discussions that they've had about their investments have been well well known to us as as you mentioned they just entered into a minority non-exclusive investment with a cohort of investors across the space our access and our relationship with apps flyer has not changed we're very supportive of them and we continue to work closely together just a little bit more of a window into our logic on our decision our our balance sheet and our cash generation, as Tarek has articulated, they give us flexibility and we keep a very high bar. So what we're looking at is strategic fit, expected returns, valuation, and alternative uses of cash, whether we act. And for us, as we just talked through with Cortex, our first priority is funding organic growth while maintaining that strong balance sheet. We've delevered meaningfully this year. And when we look at a capability, we certainly have to map it against the capabilities that we're building against organically. And so M&A is not in our model, not required for our growth, but we maintain a deep connectedness to AppsFlyer and the broader ecosystem as it relates to strategic items.

Thank you. Our next question comes from the line of Matt Swanson of RBC. Your line is open, Matt.

Thank you so much for taking my question. Tarek, it's kind of unique that you guys give annual guidance based on your peer set. Could you just talk about the difference in philosophies between quarterly and annual guidance given the timing of model improvements and how that can be variable?

Yeah, so I think at a high level, we would say that it's the same philosophy for both. You know, it's the realistic and prudent approach to setting expectations. That being said, I think as you think about our visibility into model breakthroughs, the closer in those are, the more we can specifically identify those. And so the model, as we've guided you right now, does not include any model breakthroughs. But if we saw one of those coming in the near term, we would include that in our guidance. But that's really the main difference is that sort of timeframe and visibility into the breakthroughs. The underlying philosophy is the same.

Thank you. Our next question comes from the line of Steven Zhu of UBS. Your line is open, Steven.

Hi, this is Ishabesh putting in for Steven Zhu. Thanks so much for taking our question. We wanted to ask, I mean, of course, it's difficult to predict where the next lift for models will be coming from, but can you talk to us a little bit about Cortex in its current form and work that you're doing to drive continuous improvement?

That's a great question. That's something that we think about all the time. That's the day-to-day innovations within Cortex. I think just taking a look at Q3 and beyond, we have a very active model development pipeline. And as we just discussed with Ross, a runway that we have available to us as we look to propagate and reinforce our data advantage in this market is very exciting. So when we look at the future, I think the empirical evidence is probably the best indicator of where this business can go. 11 consecutive quarters of growth with Cortex at our sales, and we've been accelerating the pace of innovation. As Tarek mentioned, the breakthroughs don't arrive on a schedule. And so when you look at the model updates, creative changes, or any sort of improvements that we have to the business, they don't follow that specific reporting calendar. And so those are above and beyond what we expect from the business on a quarter-to-quarter basis. And so when we look at where the breakthroughs can come from, just examples could be innovation in terms of the windows of training. you hear the topic of progressive training a lot in the market this is just one example training on more recent data in a more intelligent way with more features available to you but there are so many different pockets that are exciting to us and so we will guarantee you that we'll discuss specific advances once they're deployed and measurable but the outlook really reflects the capabilities that we have right now and the customer demand that's visible today although the the pattern of breakthroughs has been quite reliable

sounds good thank you thank you our next question comes from the line of Alec Brandolo of Wells Fargo your line is open Alec yeah hey thanks so much I appreciate the question could you maybe help us understand how you guys are thinking about the health of the mobile gaming market. Obviously, the leading network had a challenging second quarter, and I think some of the game publishers, I would say 1Q and 2Q bookings, were a little bit light. So just any update on kind of how your conversations and your client conversations are progressing there would be helpful.

Sure, Alex. We track this market really closely, and so I can speak to gaming and other verticals directly. We see a really healthy market. Our gaming business grew alongside all other verticals. year-over-year in Q2, and gaming definitely deserves its due here. When you look at the market longitudinally, gaming has been a great innovator of the app economy, free-to-play, in-app advertising, hybrid monetization, live operations. Those models were really pioneered in gaming and are now the foundation of the business across the sector where a lot of verticals are leveraging those best practices to get their app businesses started. And so gaming continues to be foundational to our platform. We benefit twice from it in that when you look at how we work with games, it's as demand where gaming advertisers scale with performance and as supply where gaming publishers monetize engagement and benefited from episodic moments such as the World Cup, for instance, as publishers. Now, I think when you reference the third-party data trackers, There's a mix of data that's available, and so the hybrid modernization or off-store purchases are not necessarily available, and so the headline data understates the market's health. There's really no single gaming trend that I would say is moving in one direction. Just like if you look at the sort of macro of the app economy, each and every subvertical has their pattern. That's in part informed by some of the innovation of the businesses themselves. And gaming has that same property. But as a business, our opportunity is driven by the overall market. And when you look at the growth of the app economy, the data that we subscribe to shows that the CAGR of the market is about 11% overall with 14% ascribed to verticals outside gaming. And I think when you look at our playbook, it will go to where innovation is asymmetrically concentrating. And so I think gaming will remain a key vertical, but it's great for us to have both, and that's a benefit of our business model as we look forward to the back half of the year.

Perfect. Thank you. Thank you. Our next question comes from the line of Ralph Shackert. of William Blair. Your line is open. Ralph.

Good afternoon. Thanks for taking the question. Jeremy, maybe you can just give us a sense of the order of magnitude of the model breakthrough that you saw in the quarter and perhaps how that would compare to other quarters and maybe comparable periods. And then I know you talked about existing customers driving most of the growth, but just any sense on new customer additions in the quarter as well. Thank you.

Sure, Ralph. I'll take that first one, and then I'll pass it to Tarek to take the second question on the customer KPIs. As it relates to the model breakthrough, the benefit that we are accruing from our most recent breakthrough is actually sees the entire user acquisition side of our business. And so I think just to give a little bit more color, the feature set that we used to help us underwrite the quality of a given type of impression or type of ad request was broadened and so for us like if you think about just the the amount of variables that we're using to underwrite with precision that expanded and you see that act as a force multiplier through the business moving forward that's a sustained advantage for us I think that the the variables that we consider when we look at impact are the scope that the release hits across the business and the time that it affects the print right as we think about our quarterly guide etc or in year and for us this benefit is going to accrue and benefit our business through through the year and beyond and so you can see that in the raise to the model I hope that that provides some some color on on the benefits of the release. Tarek, you want to take the question on the KPIs?

Yeah, so if you look at the growth in our business between existing and new customers, the majority of our growth was driven by existing customers in Q2. That reflects increased customer spend as performance improves through ongoing product enhancements, driving higher revenue on the core advertising platform. New customer acquisition is also important, but generally the minority of growth. Recent cohorts also scale faster than old ones as Cortex improves. You'll see in the 10Q that comes out tomorrow, that's one of the KPIs we'll release, 58% of our growth came from expansion with existing customers versus 42% from new customers acquired over the last 12 months.

Great. Thanks, Jeremy. Thanks, Tarek.

Thank you. Our next question comes from the line of Deepak, from Hatsivana, of Cancer Fitzgerald. Your line is open, Deepak.

Hey, guys. Congrats on the strong quarter, and thanks for taking the question. Jeremy, maybe one more on Kotex. Just curious, you noted self-learning as the driver of outperformance in 2Q. Recursive self-learning is an area where AI labs are making good progress and continue to see model gains delivered by the models themselves. How are you thinking about setting up recursive self-learning abilities at Kotex, and is that something that you have kind of tapped into? And then second one, Liftoff obviously has been a well-known company for a while now, but the IPO brings certainly new level of visibility and competitive advantages. Can you talk about some of the benefits you're seeing maybe in hiring side or go-to-market, perhaps business development now as a public company? Thank you so much.

Sure, Deepak. I'll take that first one, then I'll let Tara take the second question. As we look at the contribution of Cortex to our business, you dug into one of the two ways in which that Cortex will evolve. So one is self-learning, and a very prominent term, as you mentioned, is recursive learning. I think for simplicity, self-learning helps ascribe the value of a model that is updating with the most recent data on its own. And the second is breakthroughs, which could be modifications to the architecture or the feature set that is more conducive to a release work that our team is doing. Now, as you mentioned, there's this steady baseline of self-learning that accrues in this business, and it's a quite powerful economic application of AI in that our recommendation engine gets smarter through its pours into the market, if you will, and with each and every update, the model is using fresher information that is benefiting from more recent trends. Now, when you think about how we could benefit from reinforcement learning or from progressive training, from self-learning, these are all different shades of the model's architecture working to our benefit, and it comes part and parcel with the business model today. The improvements that we roll out sit above and beyond that self-learning. Now, I think that as advances are available to our business as it relates to the enhancements of the self-learning engine itself, you certainly will hear about that. But I would consider that as a breakthrough that sits on top of the self-learning that we're modeling against today. Tariq, do you want to take the second question on the move to going public and how that's working out for our messaging network?

Yeah, absolutely. So being a public company has a number of benefits that we anticipate will come over time. Two that I can tell you have already started to manifest. One is the hiring. It is actually a much easier value proposition when you're in the hiring process and you're talking to a recruit to be able to communicate the value of public company stock. It's an easier thing to quantify. It's something we're already seeing some benefits from on the recruiting side. The other is, I'd say, on the business development side. We are an instrumental part of the customers that we work with, and we're working with companies that want to make sure, you know, they have a credible counterparty that they're engaging with. Being a public company definitely adds a level of credibility that helps make those sales engagements even easier. Those are just two simple examples in the near term. I think there will be more as time progresses.

Thank you. Our next question comes from the line of Benjamin Black of Deutsche Bank. Please go ahead, Benjamin.

Great. Thank you for taking my question. So, you know, roughly half your revenue is still within gaming. So I guess the question is, you know, how do you think that mix will evolve over the next two to three years? And it would also be good to hear, you know, which emerging sort of app categories you think are most attractive within a time span. And then secondly, your incremental margin steps up into, I think, the low 80% range. So could you perhaps speak to the philosophy around reinvestments into the business and how should we think about the revenue flow through to profitability over the medium term? Thank you very much.

Yeah, thanks for the question about the market. I think it gives me an opportunity to talk a little bit about how the mix is evolving. I think that it's quite fascinating how our mix evolves intra-quarter and across quarters truly as a reflection of the overall consumption patterns in the market. And we don't target a specific mix, but we really benefit from and appreciate having this diversified portfolio of advertisers and publishers across all these different verticals, inclusive of gaming, as you mentioned. Now, overall in the market, non-gaming sector as a whole, it's a collection of verticals, has been growing more quickly than gaming, and we've certainly benefited from that innovation. Now, I think it's also interesting to dig a layer deeper. We benefit at Liftoff, we benefit twice from the innovation that's happening in this market. The first layer, and we talked about this with Deepak a moment ago, We are benefiting from the advancements in AI as they manifest in our business. The application through Cortex, the enhancements to our prediction quality, our ability to find more valuable consumers for our advertisers. But we also benefit in that our advertisers are applying the newest technology in AI to enhance their funnels, to enhance the quality of their applications. And we saw that on display in Q2 with a lot of the apps that were playing a role in the World Cup across verticals were built in the last few years and built with the newest technology and conversion funnels that leveraged the latest in AI. And so we end up being a beneficiary, not just on a proprietary basis, but through the categories usage of the technology as well. And I think that's a good lead-in to what verticals could emerge in the coming quarters. The fact is we will be there for any given vertical, given that we have such a panoramic view of the consumer from working across all these verticals, the next vertical that emerges will be some sort of derivative or hybrid of a few that we know, that consumer experience. and we will be in an advantage position to work with that new vertical. What type of consumer experience will it be? It's fascinating to watch as various native AI apps are built, as we see different kinds of consumer applications and environments emerge. But I can tell you with confidence that we intend to play a very big role in whatever new vertical does emerge and whatever that zeitgeist may be. I'll pass it over to Tarek to take the second body of questions.

Yeah, I think the questions about our incremental margins and how we plan to deploy those, I think you're exactly right that there's a high degree of flow through on our incremental revenue and the low 80% range is exactly right. And it's reflective of our strong operating leverage of the business that we're able to pass through such a high flow through. In terms of how we would invest that money, we talked about the capital allocation priorities on the prepared remarks. I think our first priority is always going to be reinvesting in the business. We are a capital-light model, so we don't need a ton of CapEx here, but we do want to invest appropriately in R&D, new vertical expansion, and other strategic priorities. Our second priority would be maintaining a reasonable leverage position. We target to be below three times on a net leverage basis. We've achieved that level already, but we want to maintain that level of leverage. And then I think third would be the return of capital. As leverage normalizes, we expect a return of capital to shareholders would be something that would be an increasingly attractive option. We don't have anything new to announce on that front today, but that's something that will certainly be top of mind for us.

And then as Jeremy mentioned, opportunities M&A is something that's always on out there that we could look at so you know lots of interesting things we can do but that's gives you a sense of where we've where we've deployed the incremental margin very helpful and thanks very much thank you our next question comes from the line of Bernie McTiernan of Needham and Company please go ahead Bernie great thanks for taking questions to two for me, if I could, first, Tarek, the 68-42 split that you mentioned earlier in terms of existing customers versus new customers, that's actually higher than I would have thought for new customers. So is there any impact from the World Cup on that number? Is that the typical kind of like roughly 70-30 we should be anticipating?

And then secondly, you guys just benefit from this large event with the World Cup any network reflect their flywheel that that's coming from that that's you know driving growth in future periods happy to take the first part of that so yeah it was the 58% from existing customers 42 from from new customers that that is you're correct that is a little bit more from new customers than we've seen in the previous quarter and that was affected by the World Cup we did see the World Cup in the prediction markets influence that. I think what I'd say is we would expect the majority to be from existing customers on a go forward basis, but there could be some fluctuations as you see changes in the app economy and different parts of that evolving. It's actually really an exciting part of our business is that we're able to service a new vertical like prediction markets and hit some of those customers and participate in that. Let Jeremy answer the second part.

Yeah, thanks Tarek. Great question, Bernie. I know you're an expert on this topic we did see demand particularly across prediction markets however it was really fascinating to observe the adjacent verticals engage in in that moment and as you mentioned and I'll get into what what is to come or what transpired after that's flowing into Q3 so we saw the OSBs engage we also saw live score apps engage on both the advertiser side but also if you think about the publisher side we are in a market that has so much space on the on the publisher side already but we can still see that there are users that are engaging with these live scoring apps for example that are creating more liquidity pockets for advertisers to access so it's really a two-sided benefit. Now, what happens next? Well, now we need to engage those high intent users or those apps need to engage those high intent users and effectively expand their DAU base in the moments to follow that flow into the NFL season. And so for our business, our job is to provide valuable users. And so as we look into Q3, you certainly are working off an adjusted base of users that raises the floor for our overall business. And that increase in engagement and advertiser activity, that enriches the signals that are flowing through the platform. And so if you think about how does Cortex benefit, well, that supports the broader Cortex self-learning flywheel there's a lot of data that comes through in a moment like the World Cup just an example we get to benefit from all of that information across the entire business as a consumer you may be interested in a prediction markets apps but you're also going to be booking a hotel room or hailing a ride or engaging in your in your banking app or playing a game and so this is how this is an example of the window into the consumer in a moment of engagement benefits that flywheel in the quarters to come I think explicitly across the sports category but certainly those same users might be engaging in back-to-school activity through e-commerce etc and we're excited about what what that can mean for our business moving forward that's great thank you both gentlemen Chairman, as we are out of time, this does conclude today's conference call.

Thank you for participating. You may now disconnect.