Disclaimer: This interview is for informational purposes only and should not be relied upon as a basis for investment decisions. In Practise is an independent publisher and all opinions expressed by guests are solely their own opinions and do not reflect the opinion of In Practise.

This interview is published anonymously at the request of the executive. 

{audio:0:00:55} What was AppLovin’s original offer to developers? 

{audio:0:01:01} It was a core app discovery and monetization platform, which to a large extent, still is the core of the business. It was a mobile ad network like many others at that point, growing very fast, but it was a way for app developers to get scale and monetize through ads.

{audio:0:01:19} Can we walk through this step by step? If I am a studio or app developer, what is the proposition of the app discovery platform to me, as a customer?

{audio:0:01:29} There are two sides, monetization and user acquisition. You can get scale within the ad network but also monetize through their ads, and it is a very competitive landscape. An app developer will never use only one provider; they work with several networks including Google, Facebook, AppLovin, ironSource and Unity, to name the biggest. I am speaking mostly from a games perspective because AppLovin has always had a games focus. They monetize and advertise with several networks simultaneously. Each network gives you additional scale which allows users to hit your ROI goals. Developers always work with Google and Facebook, the biggest providers who will give the most users. They are also fast to scale but, to get additional scale, they have to go to other channels. 

{audio:0:02:33} AppLovin was quickly becoming a third place for many back then. Their value proposition was they had scale and a functioning user acquisition platform. The algorithms were able to identify users that would hit your ROI goals back then, which has got better over the years. Today, AppLovin is definitely the third biggest channel to acquire users, especially in the games vertical, and if it is the third biggest channel, it will be relevant to monetize through them. That monetization is driven by which networks are the strongest. 

{audio:0:03:22} What defines a good network is how much advertiser demand breadth they have. Google and Facebook will always be the leaders there. Facebook have two million customers on their self-serve platform, so your advertiser breadth plays a huge role. There is an algorithm side to it, if those networks can match you with the right ad and drive the ads which convert best for you. AppLovin has got very good with their targeting capabilities over the years, using data to deliver the most high-quality users at scale, which drives higher monetization results. If an ad network has better monetization, more app developers will integrate that network because it becomes critical to monetize, which leads into the world of mediation, which we can get to later.

{audio:0:04:24} Did AppLovin become the third player simply as a result of the algorithmic side and using the data more effectively?

{audio:0:04:38} AppLovin, ultimately, had stronger tech than its direct competitors like ironSource – which was still called Supersonic – Unity, Vungle, AdColony and Chartboost. Chartboost was a big when I first started but became tiny and was acquired for 200 million by Zynga.

{audio:0:05:13} Poor tech and execution. When I first started, we looked up to Chartboost, who had a flashy office in San Francisco, and last year AppLovin went to market at 100X valuation. It's hard for anybody to say their engineers were better, but their execution and engineering was clearly better. AppLovin has always been a very bold innovator in this space. To give you an example, very early on, a big segment of mobile app install ads are in-game video ads, so you get an ad in between levels of your game. In 2014 these were called interstitial ads and were limited to static ads only; an image would pop up and you could click to remove it. AppLovin were the first to trial five second video interstitial ads, which converted significantly better than other ads. It is a better ad experience which users are more likely to engage with. 

{audio:0:06:30} AppLovin saw the conversion rates were higher and simply switched the network to video ads. That could easily annoy several clients, especially ad publishers whose app experience was suddenly interrupted by video ads that were never there. Many publishers complained about and networks are wary of negative feedback from publishers, but AppLovin went ahead anyway and, within several days, everybody sent us angry emails. Three days later, they realized they were making 45% more money and decided to run with video ads. AppLovin pushed the boundaries very early with an attitude of, if the data and the money is good, people will follow. That and their engineering acumen were critical in driving early success.

{audio:0:07:35} That got AppLovin to third position but what is their competitive advantage against a bigger scaled network, such as AdMob?

{audio:0:07:46} It is not a winner takes all market. Google and Facebook, undeniably, have a data advantage that AppLovin and others cannot compete with. Google know more about every user than everyone else does, and also have a sales team breadth nobody can compete with. They have global sales teams and all advertisers spend with Facebook and Google, which is a head start nobody can get to, but different networks have different advantages and data points they can focus on in order to drive value for advertisers. Today, advertisers use different channels to give them different types of users in different segments. 

{audio:0:08:45} If I advertise on Facebook, I will be buying on the Facebook and Instagram Apps, but also on their audience network, which is smaller than AppLovin and significantly smaller than AdMob. I will buy as many users as I can, hitting the KPIs I've given the network to hit, then go to AdMob where I will buy on Google Search, the Google Play Store and YouTube. AdMob is one place Google competes on where they have data sets that make them bid on certain users across the mobile ecosystem. AppLovin also has data that will make them bid on certain users. What the competitive advantage will be I cannot say, but when an advertiser operates in this market, they get additional scale by working with all platforms.

{audio:0:10:07} How do you compare the user quality between AdMob and AppLovin?

{audio:0:10:13} It's almost the same for us but, historically, Facebook had the highest user quality. That is now limited as you cannot get endless amounts of highest quality users. The second highest amount, you get from Google but you also get from AppLovin. With many of these channels, you feed them revenue data across their system so that they learn. Revenue data is one example but you can also send them other data. In this vast mobile app ecosystem, they decide which users to bid on, based on the profiles they have built on their user bases. The quality across those channels is identical to us. Any channel who is good with absorbing revenue data will deliver high quality users, which is mainly Facebook, AppLovin and Google.

{audio:0:11:17} Do you have a KPI LTV/CAC metric as a target, feed your revenue and first-party data back into these networks and they bid on different users?

{audio:0:11:30} You add creative assets like videos and ads, and they do the rest for you. Depending on what you learn, you can then focus on the right revenue event. Ideally, you want to send an event within the first seven days of a click or install; that has to do with attribution windows. You want to send something reasonably early which show, this is a quality user; find me users like that. That is, basically, how it works across the most valuable channels.

{audio:0:12:12} A key part of the attribution efficacy of spend is the quality of targeting data. Given there are two mobile platforms in the form of Apple and Android, how is Google at an advantage by owning the platform, relative to AppLovin?

{audio:0:12:37} That is a huge advantage for them as they can dictate terms. Apple has recently proven that platforms are willing to dictate terms and their privacy changes have affected the industry. Nobody knows what Google will do, but they might adopt a light version of that. Google is an advertising company whereas Apple is not, so it won't be as draconic as what Apple did with their SKAdNetwork changes. Android has a more open ecosystem than the controlled Apple system. For example, Android allows you to download apps from other stores. Google is a data driven advertising company at its core and has a strong interest in keeping that business alive. I cannot comment on what changes they could do.

{audio:0:13:42} You spend dollars in the AppLovin network to find high value users; what is the biggest risk to AppLovin's targeting quality over the next five years?

{audio:0:13:58} AppLovin uses machine learning so the more data the better the targeting gets. That data comes from both sides; from its existing advertisers and first-party base. They refer to it as their flywheel, which makes sense, especially in light of the first-party data access, as they continue to expand their own portfolio of games through M&A. They now have many studios who produce new titles for them, feeding that first-party data machine which improves targeting, which is always two-layered. Not only do they find the person who wants to download your app, but that person has to be of high value. 

{audio:0:14:59} AppLovin has run a fascinating strategy of acquiring or building titles in all major verticals in the mobile gaming stack, from puzzles and RPG to cards and slots, which are traditionally the biggest verticals. They are in the top grossing charts for all those verticals, feeding their machine, finding the most high value players in every vertical, and making that data accessible indirectly to any advertiser. As a slots advertiser, using AppLovin is unavoidable because they have created an aura of inevitability. If you choose not to advertise with AppLovin, you will fall behind and competitors will eat you. To get additional scale you need AppLovin, because they are the only channel which has a data advantage that can compete with Facebook.

{audio:0:15:58} Is that because AppLovin have the most scale of slots players?

{audio:0:16:00} They know who the best paying slots players are, as does Facebook and Google. Google sits on the payment process so, by default, knows who the best users are, and Facebook has all sorts of other data which people have been feeding it for a decade.

{audio:0:16:22} Do Google and Facebook also both have the data?

{audio:0:16:23} An advertiser wants additional scale and for every download to hit their KPIs. They start at Facebook but, at some point, reach a limit then do the same at Google. They also work with AppLovin who will get them a significant amount of users. Advertisers work in such a competitive landscape that it is impossible not to work with them, otherwise they fall behind.

{audio:0:17:58} Going back to the privacy changes we have seen in the industry, If IDFA 2, in five years’ time, which could become much more rigorous, how will that impact targeting?

{audio:0:18:16} The deprecation of IDFA in Apple has been rigorous enough for everybody. I am not the greatest expert on Apple's privacy rule changes, but if the user doesn't give consent to IDFA tracking, that means you cannot track that revenue data and share it with third parties. That makes it more difficult to target users, especially in verticals with deep monetization. With RPG games, few players monetize, and it takes them longer than 24 hours to spend. Using data to understand user behavior is critical, and if you cannot send a call back to the networks to let them know when a revenue event happens, you are flying blind. AppLovin, Unity and ironSource moved to probabilistic installs, which doesn't rely on IDFA to track users. They use probabilistic models to assume which installs came from where.

{audio:0:19:54} Is that a risk in the longer term when you're built on top of two platforms?

{audio:0:20:05} It is a great risk to this business and I cannot see a way for AppLovin to integrate more vertically. There are still players in both the Apple and Google playgrounds and AppLovin has surprised me before, but I cannot think of any way to de-risk that.

{audio:0:20:30} They do rely on this data but they also have first-party data which informs a lot of the machine learning, so is that their strategy?

{audio:0:20:39} That is their strategy and has driven success and given them a competitive advantage over other networks. It is very difficult for the other major ad networks to catch up on, especially as multiples for gaming companies have gone up so much. AppLovin acquired a great portfolio of gaming studios early on, churning out great content. That will be very hard for an ironSource to catch up with.

{audio:0:21:18} How would you compare the first-party data of AppLovin versus AdMob?

{audio:0:21:24} Google's data is better because they have payment data, but not on iOS, where they use whichever touch points they have on you in the Google ecosystem. Apple doesn't have data on AdMob users but they don't have an ad network. Apple has an ad business but so far they don’t show ads outside their own store.

{audio:0:21:50} Does the fact that AppLovin is vertically integrated, with first-party content, give them a competitive advantage over ironSource, Unity and others?

{audio:0:22:14} AppLovin acquired MoPub earlier this year, they are now the largest mediation player in the in-app ads market, and are known as the king of the bid. Everybody mediates their ad inventory but they use one mediator that decides who gets to fill your ad request. AppLovin invested heavily in that market by first acquiring and then building out their MAX platform. Building mediation tools around its own content business gives them the best solution. MoPub was the biggest mediation company operated by Twitter, followed by MAX from AppLovin, then Google, ironSource and Fyber, which is another smaller player in this market. Those five players own the entire market, with MoPub and MAX being the biggest. 

{audio:0:24:14} AppLovin's tool will be better than everybody else so, over time, people will switch over to whoever generates the highest yield. Some publishers would be more wary of switching to a mediation company who is a direct competitor through its gaming business, but if their competitors use a solution which generates up to 20% higher yield on their ad inventory, they will have to switch otherwise they will fall behind. MoPub was the only one who competed with AppLovin's MAX, and now that AppLovin acquired MoPub, which was an unloved child in the Twitter family, they have effectively become the largest mediation player in the market. 

{audio:0:25:17} Today the majority of ads go through the AppLovin pipes on Google Android and Apple iOS, giving AppLovin view on every bid, giving them perfect visibility. Combine that visibility with their data advantage, it becomes difficult for new entrants to the market to compete, but will also allow them to scale their business significantly in the next few years.

{audio:0:25:54} Would you only use one mediation player?

{audio:0:26:02} It gets too complicated otherwise, because you still work with the same networks below that mediation network. The networks are the same on every stack so they would be competing with each other, which wouldn't work. You would choose one and stick with them.

{audio:0:26:30} You buy more traffic and users via all networks because you want to get incremental demand but then you only monetize your own inventory via one mediation platform which, hopefully, generates the highest CPM?

{audio:0:26:44} People do split tests and give ad requests to multiple networks to see who performs better. Everybody needs demand from all networks for successful mediation.

{audio:0:27:08} That is complicated as you are competing with Google who is also your partner?

{audio:0:27:15} There is an interesting push and pull in the mediation business. Ideally, you own the inventory and only show your ads, but the yield generated to publishers will decrease. You have to create the most competitive landscape in order to stay competitive as a mediator. You have to push Google and Facebook to ensure they do as best as they can, in order to generate the highest yield. AppLovin built its mediation through the lens of its own content business. They built a tool which made their own games the highest yielding ones, which is what made MAX a superior product. It is a weird situation where you have to ensure your competitor show less ads, but it also encourages ad network operators to be competitive, and pushes operators to continuously innovate in order to drive results.

{audio:0:28:50} As a developer today, why did you choose AppLovin mediation versus AdMob?

{audio:0:28:54} We used AdMob but our average earn per user went up after moving to MAX.

{audio:0:29:05} Google has a reputation of not being the most publisher-friendly, in that they optimize for their own yield, whereas in the MAX setup, we optimized for our yield.

{audio:0:29:22} Is the take rate different?

{audio:0:29:24} Google's take rate was significantly higher than AppLovin's when we used their mediation. Admittedly, this was several years ago and header bidding wasn't around to the scale it is now. Back then, Google was the only bidder in your inventory, but that has changed to where Google has created a real time bidding environment with Facebook and other channels competing next to mediated networks. In MAX, you use a blend of real time bidding networks and so-called mediated networks that you call at fix price points.

{audio:0:30:14} How are we seeing the shift from waterfall to real time bidding?

{audio:0:30:28} What I know from MAX is that a request for an ad goes out and the mediator first aggregates all the real time bids for said request. The mediator asks all the bidders what is the highest they are willing to pay to get that request.

{audio:0:30:47} Would those be other networks as well, such as Google and Facebook?

{audio:0:30:51} In MAX's case, Google is not a real time bidder, but Facebook, AppLovin, Mintegral, Vungle, InMobi, AdColony and Yahoo all bid.

{audio:0:31:06} Why isn't Google a real time bidder?

{audio:0:31:09} Google is not a real time bidder in any place other than Unity, because there are commercial agreements between these companies. When you become a real time bidder, you help somebody significantly because you add more demand bid to every request.

{audio:0:31:42} How does the waterfall system work versus real time?

{audio:0:31:46} A request goes out and the mediator asks real time bidders who wants to bid. That highest bid gets compared to the waterfall setup. You will call Google or ironSource three times in your waterfall. You would call all those networks at various price points. You would ask Google if they are willing to pay $50 CPM for this request, and if they say no, you ask ironSource if they are willing to pay $45 CPM for this request. As long as those are still higher than the highest bid, then it is decided if it goes to the bidder or the waterfall network.

{audio:0:32:46} How is that changing CPMs?

{audio:0:32:48} The more demand you have for each request that goes out, the higher your CPM will be, because it's a real time auction on every request.

{audio:0:33:07} Are you seeing that materially increasing over the past year?

{audio:0:33:12} Yes, we've definitely seen CPMs go up over the last few years, across the board.

{audio:0:33:23} Is this kind of like a hybrid between waterfall and real time?

{audio:0:33:27} Exactly; it is a hybrid setup.

{audio:0:33:29} Is it ever going to go full real time?

{audio:0:33:31} Everybody would hope so but it's up to the networks to switch. Facebook has been open and switched over to bidding everywhere. Google has been more difficult because networks compete with each other as they use different mediation solutions. A bidder using my solution makes me more competitive, so why would I want to help you?

{audio:0:34:01} Why is Facebook willing to play ball in real time bidding and Google is not?

{audio:0:34:10} Facebook are a competitive platform but are not a mediator, so they want to have access to every bid possible because that is better for their business. Those who want an open market are happy to bid everywhere.

{audio:0:34:56} How do you compare MAX to ironSource, in mediation specifically?

{audio:0:35:00} AppLovin's MAX out competes ironSource on a revenue per user.

{audio:0:35:08} What are the main reasons for that? Does it simply come back to scale?

{audio:0:35:12} AppLovin's mediation is a better mediation because it was built with the publisher in mind being a mediation for its own scaled large content business. One of AppLovin's biggest games is called Wordscapes by PeopleFun, and it is the biggest word game in the market. Word is an ad-revenue driven vertical and if you want the biggest word game in the market, you need the highest yielding mediation solution. That's an example of how AppLovin's MAX beats others.

{audio:0:36:04} They effectively build it in-house and simply monetize it externally?

{audio:0:36:10} Testing and building on their own products and content has worked well for them.

{audio:0:36:16} It's interesting how Google's AdMob mediation has fallen behind AppLovin?

{audio:0:36:26} I don't know many publishers who use AdMob. Everybody would be happy to test them but do they want to give their data to a competitor – which arguably AppLovin is – in the gaming vertical? That puts a lot of people off, but people prefer using AppLovin over Google.

{audio:0:36:52} Is that because of the yield and the increased CPM you get?

{audio:0:36:55} There is data access and other tools that come with a good mediation solution, which helps publishers better understand their data, and that is the icing on the cake.

{audio:0:37:09} How sticky is the mediation for you? How difficult is it to move to AdMob?

{audio:0:37:17} It is difficult but not impossible. If an app developer wanted to switch, they would have to rebuild several systems which is not ideal. It would be a multi-week transition, especially when you rely heavily on that kind of data, so I would say medium sticky.

{audio:0:37:49} How important do you think MAX mediation is to AppLovin?

{audio:0:37:55} It is an integral piece to AppLovin's business. With the on boarding of MoPub to MAX, AppLovin will have visibility on the vast majority of bids in the in-app ads market. That removes all supply constraints from an ad network point of view and allows them to compete on everything. The stronger their ad network becomes, the more demand they have, and the more they can execute on building out their demand side. 

{audio:0:38:36} There is a circular trend because it is a fast-growing market and the pie is growing. They now have a big chunk of the pie, and the better they can execute on their demand and their first-party data injection, the more the business will grow. After they acquired Adjust, they now have a global sales team, and many non-gaming integrations which they never had before, which is a big advantage Google has. Google advertisers come from all verticals, but bit by bit, AppLovin is adding more demand, and has the ability to bid on every request, so the focus is driving and increasing demand to drive the business forward.

{audio:0:39:37} Why did AppLovin buy MoPub when Twitter couldn't make it work?

{audio:0:39:41} MoPub had a reputation of not having many development resources focused on it at Twitter. Twitter's network take rate on share of voice in MoPub was in the low single digits, which speaks for the strength of its ad product. They are not known to have a strong performance advertising product, whereas AppLovin is a performance advertising company at its core, and has a take rate many multiples higher than Twitter's. Adding that inventory to AppLovin will generate a lot more cash for the owner than Twitter has.

{audio:0:40:53} Is that because the ROAS will be so much better?

{audio:0:40:55} Because they will have a higher share of voice with all the MoPub publishers.

{audio:0:41:13} They have purchased a huge amount of gaming and non-gaming inventory and will plug that into their mediation machine to offer real time bidding?

{audio:0:41:32} At the same time, driving first-party data from their own content to get access to every bid. AppLovin said they will process 15 billion in ad revenue through their pipes this year alone. In a growing market, access is key. Having access to every bid on every banner or video ad shown inside any app, will only fortify AppLovin. For advertisers to get to those users most effectively, they will go to AppLovin, which in turn makes the mediation system stronger and makes their yield higher, which makes them even more inevitable as a mediation partner for any future publishers. The flywheel works on both sides so they will simply scoop up more supply, making themselves more interesting to advertisers. If they want to advertise and get to users outside of social, You Tube and search, they have to go through AppLovin.

{audio:0:42:59} Can they see every bid on every ad in an app for gaming and non-gaming?

{audio:0:43:08} In everybody who uses them, yes.

{audio:0:43:09} What are the biggest barriers for AppLovin to grow in non-gaming apps?

{audio:0:43:17} That piece requires a lot of work because non-gaming works differently from gaming. Historically, AppLovin hasn't competed there and that is where the DSP part becomes interesting. The MoPub acquisition added a huge DSP footprint integration to the MAX stack. AppLovin started their AppLovin Exchange, which is where all its DSPs are housed. DSPs play a very important role in the world of digital advertising, and the Trade Desk is a famous DSP. 

{audio:0:44:23} AppLovin is plugged in to all these DSPs who are a big driver in the native inventory market. In the mediation game, the DSP pays a fee to the mediator for access to the bid or for each impression they serve in a revenue share model. AppLovin are onboarding DSPs to give them demand for the non-gaming native inventory. Inevitably, they will gear up their own advertiser base to compete on non-gaming, something they haven't done in the past. There is definitely a lot of growth wind for AppLovin there.

{audio:0:45:13} Because they now have MoPub's DSPs and their own to integrate?

{audio:0:45:24} There are two sides to it. If you acquired this inventory and show these native ads on any mobile app that shows ads, in order to compete in those, you need to have a big DSP footprint, which will make your entire network more competitive. That, in turn, pushes AppLovin to focus on its own advertiser demand to compete on native inventory.

{audio:0:46:03} Briefly, on the economics of the core app ad network for AppLovin, how does the rev share typically work? A developer allocates an ad spend, has a target and pays a certain CPM; what is AppLovin's take?

AppLovin: King of the Bid

Former Director at AppLovin

inpractise.com/articles/applovin-king-of-the-bid

AppLovin: King of the Bid(January 28, 2022)

Former Director at AppLovin

Interview Transcript

Related Content

AppLovin: Building a Mobile Ad Network

Former Director at AppLovin

Subscribe now

Related Content

AppLovin: Building a Mobile Ad Network