AppLovin Corporation (APP) Earnings Call Transcript & Summary

All right. Well, yes, here, we are live. But hello, everyone. Welcome to the fourth and final day of the Morgan Stanley TMT Conference. My name is Matt Cost, I'm Morgan Stanley U.S. Internet team. I'm thrilled to be joined this morning by Adam Foroughi, Co-Founder and CEO of AppLovin. Thank you so much for being here.

Thanks for having me, Matt.

Great. So let me quickly run through the disclosures. Please note that all important disclosures, including personal holdings disclosures and Morgan Stanley disclosures appear on the MS website at morganstanley.com/researchdisclosures or at the registration desk. And with that out of the way, maybe we can jump right into it.

Cool. So I mean, I think, at least for me, probably the biggest surprise coming out of 4Q was the guidance that you gave around the ad network. I'd say, especially the $2 billion figure for 2023. I think that's somewhere around 40% growth even against a tough comp in '22 and then some very strong growth this year. So to me, that seems to say despite the pace of growth we saw in '21, there's a lot of runway still ahead. So I guess my question is what gets us there? And if you had to name the 1 or 2 initiatives that you think are going to be the most impactful to getting to that $2 billion in '23, what would those be?

Yes. So let's dive into really the buildup case around that, and we exited the year at a $1 billion software revenue run rate. And you can think about that $1 billion with a 10% adjusted contribution, 90% from our ad network, AppDiscovery. And then we obviously announced that we were closing MoPub on 1/1, integrate MoPub by 3/31, shut it down, and we build the unified MoPub plus MAX marketplace, which will be the vast majority of all in-app-traded advertising flowing through our market outside the walled gardens. And so if you build it from the marketplace, and that's how we look at it, you go -- we're starting at a $1 billion, contribute $1.50 billion for MoPub this year, so you're at $1.15 billion and grow that up this year. Well, we have AppDiscovery as a main contributor. It's grown 30% on average quarter-over-quarter for 3 sequential quarters. And so what do we need to execute on this year to get to the $1.4 billion seems very reasonable. But then if you think about the $2 billion and then the long-term prospects, what's interesting about the business model last year and then for most of this year is we don't have contribution from 2 other material parts that I'll outline right now. So MAX and MoPub, we sized that coming out of the year to target $10 billion plus in the Marketplace trading to publishers. Well, traditionally, we made money on the ad network. Now as the market moves to bidding implementation where other networks will head or bid, Facebook, Google, Unity, ironSource, et cetera, we get 5% of every dollar transaction. So if the entirety of the market was 5%, we'd have a $500 million business. Now our seat, we don't charge ourselves, but the vast majority of it will be header bid and the vast majority of it will trade that way. So that's hundreds of millions of dollars more to contribute to the business. They call it, at the end of the year, is a $1.4 billion starting point that we'll have next year that we won't have in a material part of this year. The other part is the MoPub business, which was reported at $200 million-plus contribution last year by Twitter in their earnings call. Majority of that was their ability to monetize a DSP, demand-side platform, buying inventory through their exchange that they ran. And so when they do that if $1 from Trade Desk went into the MoPub marketplace, they would take a fee off of it. In our case, it will be $0.20 on the $1. Again, we didn't have this business model in any material way prior. So we'll get this business model, and you'll get the Trade Desk, Google DoubleClick Bid Manager, Twitter Audience Platform, all these DSPs buying through our market, which is jointly much, much bigger than what MoPub was on its own getting taxed 20%. So if you look at that, that's another hundreds of millions of dollars to contribute to the software business that we didn't have before. So we end up with this marketplace that's quite fast growing, starts the year at a $10 billion run rate, going into next year and 2 businesses that are 9-figure businesses that we didn't have on an already fast-growing base. And so we just need to execute, and we should be able to achieve what we put out there. And frankly, we don't put out numbers that we're not confident in.

Great. No, that's really helpful context. I guess, on the other side of the house, I mean, you made some comments on the last call around how -- if I'm interpreting them correctly, the first-party apps are really a means to an end. And the main goal with them is to generate data to the ad network. So maybe put another way, you're not managing the business around the revenue generated by the apps. So I guess, first of all, what's your reaction to that, agree or disagree? And then secondly, what does that mean investors should expect from here, both in terms of revenue contribution, but also the level of investment you're going to put into the app?

Yes, I think I said it. So I probably agree with it. But just to explain it a little bit more, mobile game publishers grow their businesses 2 ways. They're either buying other content publishers. So they're aggregating M&A and inorganically growing and hoping to buy early enough to grow organically after an acquisition. We talked about how it was strategic for us for a couple of years to build up the pipeline, but it's no longer strategic in focus. So this is an organic discussion. The other way to do it is user acquisition. And we grew our games from a very small amount and even inorganically. The acquisitions have grown a lot since we bought them. Most of that growth came from user acquisition dollars spent on our own platform. We give that in the TSTV metric, which, if you took that, that was a $1.4 billion run rate coming off of Q4. I mean, our goal is to flip as many of those dollars or any dollar on our platform to third parties. Because just economically, if you think about the economic model, we bank margin from the third-party dollar right away on the dollars that we're spending. It's a long-term investment. It's an LTV-to-CAC formula. And it was an investment in data and scale, which we think we're good with now. And just if you look at the numbers, quarter-over-quarter, Q2 to Q3 to Q4, the games business was pretty flat last year. So we saw that. We still spend a lot in user acquisition, but the software business grew 30% each quarter. And what we've realized is like the games business is a subset of data. It's one of the data points we have. But we sit on top of 2 billion devices in the mobile landscape. The market in privacy-changed world has gone much more contextual. And so you're living in a contextual world where we have some just inherent advantages with our marketplace and our reach and scale, where you can't even use that data at large scale and a lot of the ad serving decisions you make. Yes, we're performing really well. So therefore, if we just put that aside for a second and say, if we just operate and add this, run it flat and then not needed to grow, the software business we're still very confident in, we end up in a really good place. And the margin structure of the business expands because our moneymaking comes from the ad tech business, not from the games business.

Got it. I guess kind of rounding out the elements of the business, you made a pretty big announcement a week or 2 ago around the acquisition of the Connected TV platform called Wurl. Obviously, this is a step into a new frontier of advertising for you guys. So what do you think that AppLovin's edge in Connected TV will be? I think it's fair to say you've been an innovative company in the mobile app marketing space. So as you work to enter CTV, what do you think that you can do better than the competition?

Yes. Just to level set, Wurl, as a business on its own today, they generate revenue by helping the content creators distribute their content to the end consumer device through channels. If an AMC, for instance, wants to show their content through an app that they own, that's not a Wurl product. But most of the users who watch AMC shows don't do it through an app. They do it through channel implementation like linear TV operates. And so Wurl is effectively the CDN and compression technology and rendering tech for these content creators. So they have a really good software business inherently there. Now when they show the show to the consumer, they control the ad space. And the share back revenue from the advertising. And so effectively, what we saw is they have an SSP. They have the same MAX product we have. They have this auction. Most dollars in Connected TV today are transacted the same way linear TV ad dollars are transacted. Brand dollars flowing through Nielsen, Nielsen reporting audience household segment reporting. And so it's exactly what mobile was when we started and saw a need for performance. And we think the market will have to go there. Now Wurl had launched a Wurl performance ad product already, where, if you think about who needs advertising in TV, if an AMC can't get their show at 5:00 p.m. watched by new users, because the only people who will watch it are the people who are used to watching that AMC show at 5, they can pay for Discovery. Now the Wurl Perform product is, okay, come run an advertisement, get a new user to watch that show, we can then track all the ads that they see in all of the minutes that they watch I showed today and any time they watch in the future, and close off the attribution for you and give you a performance model that could be scalable and get you much more Discovery. And so we saw that, and that's exactly what the AppLovin performance in-app advertising model is. We just know how to do it really well. And both their team and us saw a path to expand the market and performance and really use our technology and know-how to speed along the market from going from point A to point B.

Yes. No, that's a great point about it being kind of like an earlier stage kind of like, I think you said like mobile was maybe a decade ago. I guess, kind of fast forwarding to today, where the mobile app marketing ecosystem is changing a lot. There have, been some big changes in the past year. And you mentioned a few minutes ago about contextual advertising. Obviously, Apple introduced some privacy changes that were fairly disruptive to the ecosystem. Google was talking about doing something similar. That said, I think many would say that Apple ended up being a net beneficiary of some of those changes. So I guess the question is, would you stand by that? And then do you think that Google's changes could potentially create a similar benefit to the Apple changes?

Yes. So if we look at the change, like the main change being IDFA going away on some percentage of the audience. So about prior to iOS 14.5, 75% of users had IDFA, 25% did not. They just globally chose not to share IDFA. So we lived in a contextual world on a 25. Apple gave the 75 for the whole audience the choice to share IDFA or not. We now live in a world where about 65% don't share, 35% do. So you have a starting point where you can serve an identity-based advertisement. You can serve an advertising using our first-party data intact, you can do much more relevance. When the change came out, it was a haircut to everyone's targeting capacity. So we -- as I mentioned, like we go more contextual, we can't use the first party as much. Well, the biggest loss actually was companies that depended on identity can no longer sync that identity when they didn't have the IDFA. So we took a loss and then to your point of being a beneficiary, market share opens up when identity-based advertising is no longer relevant. And so we had a gain from that, and those are just net neutral. And that's what we've signaled in the business. That's on the targeting side. And Google's view as an advertising business is to put something into the market in a couple of years to remove the ID. So they also don't want to have identity-based advertising if the user chooses not to have their data shared in that way. But they don't want to disrupt the sector in any other way. So they already have better attribution models built into the platform today than what Apple presented with Scan. And Google also has a really big advertising business. And Google also has a lot of scrutiny on that advertising business. They can't do things that are disruptive to the marketplace to benefit themselves. So it's going to end up being most likely a much better implementation for continuity in the businesses, but identity is going away. When users don't want to be served something based on the identity data, that's just not going to be a thing outside of walled gardens, where the user has consciously chosen to share identity in exchange for better products. And so in a world where identity-based advertising is gone, we believe we're set up very well with our marketplace and business offering to continue to perform at the levels that we have and possibly even better as we gain share.

Got it. Let's maybe stay for one more on the theme of kind of privacy and the platform. So I think there's been a lot of concern in the space recently around probabilistic attribution. Some people also sometimes refer to that as fingerprinting. It's a pretty complicated topic. But maybe the simplest approach would be to ask, how would AppLovin be impacted if platforms like iOS were to decide to disallow probabilistic attribution? And then how likely do you think that is to happen?

Yes. So first, there's 3 words: fingerprinting and probabilistic attribution and conversion modeling. And I'll spend a couple of minutes just diving into this because I think it's fairly complex and misunderstood. You had -- the attribution companies prior to iOS 14.5 attribute on the 25% of users that didn't share IDFA. It's on an attribution model called fingerprinting. They -- Apple came out and said, you're not allowed to fingerprint. And the attribution company you said, okay, this is probabilistic modeling. It didn't change the technology that is called it probabilistic. The self-attributing networks or walled gardens color conversion modeling. What that is, is if you're on an IP address that's a unique IP address and in a couple of hours, you install an app on your device, they can get into a pretty good guess about 70% to 75% accurate that a certain device downloaded, that a certain device made that download and then tie back all future events for the advertiser to analyze the dollars spent. The reason that exists and the reason the advertiser needs this comfort is because Apple's own attribution methodology scan is about a 1-day attribution window. So if you're an advertiser, you spend $1 million, Apple gives you a 24-hour lens into the performance. You have a lot less comfort than if you can get a 1-year view into the performance through probabilistic contribution. Now what happens if it goes away? If Apple says, we're going to obfuscate IP address. That's the only key there is. It's not a personal ISP. It also doesn't allow for anyone to go out and cross-target the user because people jump IP addresses all the time. It's not persistent. So Apple's own definition of fingerprinting is persistence, creating the ability to track a user when they choose not to let you track them. So I think one of the narratives out there is Apple is now policing their own platform. I think that's just flawed logic, they are. But this is not a persistent means of attribution. So if they get rid of it by obfuscating IP because they decide they want their own attribution model to be the de facto standard. Then what happens from there is the entire market, both self-attributing walled gardens can't conversion model, ad networks can't use probabilistic. Advertisers will have to get used to assessing their dollars spent on 24 hours for the 65% of users where they only have 24 hours of data on. 35%, they have ID-based attribution, same as it always was prior to any of these changes. So if you're an advertiser to spend $1 million from 35%. Of the $1 million, you know exactly how you yield. But you know 5% return on day 0 amounts to break even by day 365, it's pretty simple 20x multiplier on the 65%. So I think the market just goes from 100% to 75% to 35%, but the dollars have to be spent. The audience eyeballs are there and not a whole lot is going to change if Apple does IP address.

Got it. No, that's super helpful. I guess kind of thinking about go-to-market and addressing those customers. I think really since the Adjust acquisition last year, you've talked about the opportunity to leverage the sales force to reach out to customers. And I think that's something that prior to that, you didn't have a sales force at AppLovin historically. So with MoPub now, you could argue the opportunities even larger to leverage the sales force. So where are you in the process of integrating those capabilities? And how big of a needle-mover can that be from here?

Yes. So MoPub and Adjust came together pretty quickly. So we're in the midst of putting 2 successful teams, unifying them within ours. Some of the salespeople will sell their traditional product just attributions run separately. Some will port over to sell AppLovin's offering. Now what's interesting about this year being transformative with the MoPub integration is we put the spec number out there and like that is a benchmark for salespeople to go bring in clients, but we're going to bring in some specs that will throw all the math off. The MoPub specs, a Twitter Audience Platform reaches thousands of advertisers. But to us, that's one. It could become a mega spec paying us a lot. So it throws all the numbers off. But Trade Desk is 1 to 1,000. So same sort of deal where we're going to create the connection to these really big pipes. And then our sales people are going to sell through them, but we'll report the one. And so it's interesting because we'll unify the sales forces, we'll have efficiency created, but it may not be as obvious in the numbers this year while we bring on these mega specs that we haven't had traditionally. But these are huge dollars. They enable the omnichannel spenders and a lot of the nongaming dollars to flow through come through a DoubleClick Bid Manager at Trade Desk. Now that we'll have that in our platform, we'll be able to sell through those benefit from those dollars, tax dollars in our marketplace. And then a couple of years from now, we'll be talking about P times Q. We will be going through an evolutionary period this year, though, that we'll land it in a more stable place longer term, but won't be in the current year.

Got it. I like that term mega specs. I can't wait to add that to the model. So I think that to a great extent, and you were talking about header bidding before, the ad market that you're operating in is auction-based. There are some examples in online advertising in the past, whereas auction-based ad markets become more competitive, you can see pressure develop on that network take rates over time. So I'm wondering, have you observed any trend in take rates as you've grown the company? And if they're fairly stable, do you expect that to remain the case?

Yes. It's interesting because we are the market, right? We launched MAX to facilitate a movement to header bidding faster than where it was going, and that was really the impetus around it. Us and Facebook were the 2 earliest adopters of real-time bidding. And we have a couple of beliefs. One is we don't optimize to take rate. We'll take $0.01 out of $1 because it's still a penny we otherwise wouldn't have had. We'll take $0.80 of it if we have really good data, and we can be accurate, and we make $0.80 on the other dollar. We report only the net. So our take rate is not an impactful metric that we optimize to in any sort of way. But the other belief is, as you take the market to real time, you create more efficiency of market. You bring more competition online, you create more discovery, more efficiency of market. As that marketplace gets more efficient, the publishers utilizing it create a world where they get more dollars into their pocket, but they don't pocket it. The go reinvest it in UA. The flywheel or funnel in mobile gaming and mobile app is the more LTV, the more dollars you can spend because LTV-to-CAC formula widens. So therefore, let's go raise the bids and now we're more competitive. So for us, like it's not about take rates or competition. The real growth driver is the MAX market growing. And the way the MAX market grows, as we bring it together and get it to the $10 billion plus are: one, more efficiency yields the publisher more so they can reinvest more. More competition creates more discovery because there's more platforms to buy on so the publishers get more eyeballs as the pie grows. The entire market is going to grow as the MAX marketplace grows. And the third one that no one talks about is the mobile gaming TAM that's widely recognized is the IAP market TAM. The $100 billion-plus IAP market is saturated. It's not growing anymore. Why is it not growing? IDFA headwinds made it really hard for developers to target a purchaser for an IAP game. 5% of users buy, 95% don't. IDFA and identity was required for marketing at scale. Well, that's gone. So these developers now have to adapt to much more mainstream games, and they have to monetize the 95%. That's going to happen with advertising. That comes right back to our marketplace. That will be a multiyear not only shift in development pipeline and how they think about going from more niche IAP-type titles to much more mainstream casual-type titles. But that's also going to lead to more eyeballs getting monetized with advertising, and that will create more tailwinds. So that MAX market grows. We're going to grow the entire market of advertising is going to grow. That will fuel more discovery in casual games, and that's really the funnel that we're excited about.

That's a great description of the funnel as you put it. I guess kind of thinking to a natural extension of what you're doing on that side. I think you've talked in the past a lot about growing your exposure to more customers outside of gaming for the ad network. So where are you for nongaming advertising today versus where you'd like to be? And then when you think about the data model and contextual techniques that you guys are using, how do you feel that those will lend themselves to advertising outside of the gaming room.

Yes. So MoPub's business was vast majority nongaming, Twitter audience platform, Twitter's business itself, Trade Desk Google DoubleClick Bid Manager. These are pipes that non-gaming customers that are omnichannel utilize for their buys. So in order to get nongaming dollars to flow through our market, we have to hook up the pipes. And so that's happening right now. That's active. And so as we integrate MoPub successfully, we get these customer relationships. You're going to have a ton of nongaming demand come through our marketplace. The nongaming demand that we're going to sell through to does not depend on our data. It depends on their own omnichannel data, and it's going to go through these other pipes, not AppDiscovery and that's where we take a different tax. We just take the dollar that's transacted transparent pricing model. And that's one where we go. MoPub, a couple of hundred million dollar a year business, flow that through to our marketplace, it's going to be materially larger and faster growing than MoPub was. That becomes a really big business over time and a good part of ours and taps right into that non-gaming demand.

Got it. So I guess as we came into this year, especially following some of the disruption that we saw last year, I think people were curious to see if companies, as you said, targeted or behavioral advertising kind of that share came up for grabs last year. So people were curious to see the companies that may have been shared donors or growing slower last year, would they get much more competitive in '22. So I guess, are you seeing any changes in the competitive environment as we head into the year? Or are the ad dollars that sort of came into contextual last year proving to be sticky?

Competitive landscape is pretty constant at this point. We've got a good lens into it with the marketplace. Identity-based advertising, obviously, is much less efficient on iOS. If you can't do it, it's not there. Now I think just tying back to what I said a couple of minutes ago, we want competition and more discovery in the market, like the IP games market would be growing if there was identity-based advertising at scale. So it's good if there's competition in the market. Now competition goes out, then we just gained share. We're a big player in the market at this point. The best comp to us has been Unity Operate. That's the most similar business to what we have our growth rate has been much higher in the last few quarters. Our scale is now much bigger than Unity Operate as well. And so -- and then the dollars per advertiser spend on our platform is multiples bigger on the averages because we just have 1/3 of the advertisers, therefore, a bigger business must have 3, 4x share of wallet. So if you look at all those numbers, we're doing really well in the software business. We have to go sell our product out, but we've never been traditionally worried about competition coming in because we know the pie grows with competition. It pushes dollars out. The publisher spend more, everyone's benefiting from them.

Got it. I think that on the quarter, one thing that stood out was the EBITDA margin guidance, which came out in I think the high 20s percent. And Harold actually went into some really helpful incremental detail on that on the Wurl call a week or 2 ago about the puts and takes. But from your perspective, maybe first talk about what initiatives are baked in for your guidance this year on the OpEx side? And then secondly, what are the potential sources of upside into '23 if those initiatives are starting this year and then continuing?

Yes. So I think -- so new initiatives aren't factored into our revenue guidance. We also -- you did a great job of writing up on Harold's comments on the world call, super accurate. We do once a year guidance, and we only update it if we're not going to hit it. And for ahead, we're not updating it. And so we reserve dollars just in case we need to invest around new initiatives growing. We haven't factored those into revenue. We reserve the cost, but we don't spend the dollars, then it will just flow through. Now next year, we didn't give an EBITDA guidance. But if software goes from 1 to 2 A net revenue reported business. If all that growth comes from the ad tech business and its net revenue reported business, there's not a lot of cost against that $600 million incremental. So our margins will continue to widen as that software business continues to execute as it does, if we're not investing heavily into new initiatives. Now if we're investing heavily into new initiatives, 2 is not going to be 2 anymore. And so we'll talk about that when we get there.

Got it. maybe obviously, I think the 3 initiatives you mentioned were CTV, like carrier solutions and NFTs and blockchain. Obviously, you've had a lot to say about CTV in the remaining time. But maybe on those other 2 how big do you think those opportunities could be for AppLovin and how focused are you on executing on them in the near term?

Yes. We're entrepreneurs. So with that, like the focus is 10 15-person teams trying to build a new business. And if it builds, you start scaling it from there. And really, the carrier OEM market and marketing solutions, that's a well-defined space, not a huge TAM. It's just a lucrative extension of our core marketing platform. But the blockchain space, blockchain NFTs especially in gaming, is not a well-defined thing yet. It can be really big. We all know this Web 3.0 stuff that's hyped, but there hasn't been a great use case yet out there that we can all look at and say, like, that's the future. Now we also know that people were spending $100 billion a year on an IP buying air. There's no value to that digital good that I'm buying today. But if you're buying a car, if you're buying something that might depreciate or appreciate based on its utility and based on its rarity, then that could actually expand the economy of the free-to-play market and have this layer of NFTs and blockchain on top. And so our view there is with MAX marketplace, we're directly plugged into and deeply integrated with most of the app developers in the market of any sort of scale. If we can expand their monetization opportunity, they can invest more in the user acquisition, that benefits us greatly. If we can do it with the marketplace on top of the marketplace, that's a really good business bet. And so that's why we're building.

Got it. So the vision I know it's something you're experimenting with. It's a vision of sort of enabling the sale or trading of NFTs basically as a form of [ monetization ].

For our developers.

Got it. Now it's really interesting. I guess kind of following up on new initiatives, maybe we can talk a little bit about your M&A philosophy. At this point, do you feel like you have most of the capabilities that you need in-house with Wurl now and on the CTV front? And if not, how are you philosophically thinking about buy versus build?

Any new initiative, we'll always take that buy-versus-build thought process. And so with CTV is already mature enough as a market, even though it's really early stage when it comes to monetization. So we felt like buy was the better approach, obviously, right behind the earnings call, how we executed on it. Blockchain NFTs, nothing exists. So there's nothing to buy. Everything is priced on future value, but we feel like we can build. And then the carrier OEM space, there's a couple of players, they are competitors of ours on different products. We already have the base technology. So extending it out seems a lot more cost effective than being acquisitive. So we think we have a lot of the key components and then we're very bullish on our own business. That's why we just launched a share buyback. So traditionally, we would have kept our capital and have it available for strategic M&A, but we're bullish on the business that we have in front of us and at the prices we're at today in the market, we're buying our own shares. So we'll make a decision on how to deploy capital depending on what we see in front of us is the best way to return dollars to our shareholders.

That's great to hear. I think in the final minute or 2, maybe we can just follow up on MoPub. And I think you touched on a fair bit of this. But you're in the process of integrating, and I think you've mentioned a 90% retention rate from MoPub customers as you've worked to transition them to MAX. I guess talk about the thought process behind transitioning MoPub fully into MAX as opposed to kind of running them in parallel, and then the opportunities that you see that opening up?

Yes. It shuts down in 21 days or -- so it's all going to be transitioned. We've got 2 types of customers: one, nonpaying the publisher, but that's the value. And so the publisher has to come over to create the supply and then the demand naturally comes over. And so this quarter is all about bringing the publishers online. That's the 90% plus. The demand side will just be 100% retained. It just has to integrate over time. So you're getting all of the demand side actively integrating in parallel. That's just with some of these companies will be slower than the publishers because the thing goes offline, the publishers won't have business left. So they have to be cut over. Whereas the demand side will come over, some now, some have already come over and some will come over, over the next few months. And so once that happens, fully intact, it makes it really nice because from an integration perspective of technologies, we have no tech debt. We already had the MAX platform. We had 3 months to build the MoPub feature set that we were missing on top of it, which we did, and then that helped us retain the audience.

Great. I think we're right up against time, but thank you so much again for being here.

Thanks for having me.