Applovin (Summary): The Best Case of Technology Meeting Cross-Market Demand

The following is the summary of the earnings call for AppLovin ($ AppLovin.US) for the fourth quarter of 2024. For the earnings report interpretation, please refer to “暴拉 30%！Applovin 如何暴打空头？”

I. Core Financial Information Review

1. Fourth Quarter Financial Data:

Total revenue of $1.37 billion, a year-on-year increase of 44%;

Adjusted EBITDA of $848 million, an increase of 78%, profit margin of 62%, revenue to adjusted EBITDA conversion rate of 89%;

Free cash flow of $695 million, a year-on-year increase of 105%, and a quarter-on-quarter increase of 28%.

Adjusted EBITDA to free cash flow conversion rate of 82%, cash and cash equivalents at the end of the quarter of $741 million, with 340 million shares of common stock outstanding.

Advertising business revenue of $999 million, adjusted EBITDA of $777 million, profit margin of 78%, quarter-on-quarter revenue to adjusted EBITDA conversion rate of 75%, due to the stepwise increase in data center costs, it is expected to return to normal with the GPU cost leverage effect.

Application business revenue of $373 million, a year-on-year decrease of 1%, adjusted EBITDA of $71 million, profit margin of 19%. A term sheet has been signed to plan the divestiture of the application business. The total estimated transaction amount is $900 million, which includes $500 million in cash, with the remainder representing minority equity in the merged private company. Subject to regulatory approval, the company hopes to complete this transaction in the next quarter and looks forward to seeing this business succeed under new leadership.

2. Annual Financial Data:

Total annual revenue of $4.7 billion, a year-on-year increase of 43%;

Adjusted EBITDA of $2.72 billion, an increase of 81%, profit margin of 58%;

Free cash flow of $2.1 billion, adjusted EBITDA conversion rate of 76%.

In this quarter, 1.6 million shares were repurchased and canceled at a cost of $508 million; for the entire year, 25.7 million shares were repurchased or deducted at a cost of $2.1 billion, with a weighted average price of approximately $83 per share.

3. Financial Guidance for the First Quarter of 2025:

Advertising business revenue is expected to be between $1.03 billion and $1.05 billion, adjusted EBITDA between $805 million and $825 million, with a profit margin target of 78% - 79%. Application business revenue is expected to be between $325 million and $335 million, adjusted EBITDA between $50 million and $60 million4. Overview as follows:

II. Detailed Content of the Earnings Call

2.1 Key Information from Executive Statements

1. Business Breakthrough: The fourth quarter is significant, as after the 2023 AXON upgrade, it has for the first time gained holiday shopping advertising revenue (holiday shopping advertising dollar), with non-gaming categories driving growth. The platform reaches over 1 billion mobile game users daily, with user engagement time comparable to social networks.

In the past, advertising was mainly for games, but now it attracts more types of advertisers. The fourth quarter indicates that the business model is applicable across multiple categories, and the platform's success is not limited to direct-to-consumer brands. Early pilots show that companies in various verticals can achieve conversions through the platform, with over 10 million online advertising companies globally potentially monetizing through our platform. The platform is positioned as a growth engine, shifting from a focus on games to serving the global advertising economy, where users discover new products in games, generating additional demand. There is significant demand from advertisers wanting to join the platform, and the advertising system is still under comprehensive development, currently lacking complete automation capabilities. The primary task this year is to develop automation tools.

1. Sale of Application Business: Seven years ago, we acquired game studios to train machine learning models, and now we have signed an exclusive term sheet to sell the App business, with the original App team joining a company focused on game development. Applovin will transition to a pure advertising platform, focusing on productivity, automation, and building high-impact teams. In Q4, the advertising business generated approximately $3 million in run rate adjusted EBITDA per employee, and this is expected to rise in the future.

2.2 Analyst Q&A

Q: You mentioned that different brands in various verticals have seen benefits not limited to direct-to-consumer (DTC) marketing. Does this mean that this year, as you expand the scale of non-gaming businesses and enter a self-service phase targeting a broader global audience, the goal is not just DTC marketing? Or is this merely a feasible opportunity for the future that has not yet been fully utilized?

A: When building the platform, we strategically chose the DTC business area because we understood that if the technology is effective in this area, it will also be effective in others. Currently, we have privately allowed clients from other categories to go live and found that the advertising model works in these areas as well. Seeing clients from various categories succeed on the platform gives us confidence for this year. As we launch more tools and move into a more automated and open state, we hope to attract a large group of advertisers with annual revenues exceeding $10 million globally. We will steadily advance in the future, focusing on utilizing tools, artificial intelligence, and automation technology to provide solutions and successful experiences for new advertisers while maintaining the company's cultural valuesQ: Regarding CTV advertising, can you talk about the changes at Wurl, especially compared to our last discussion? What changes are prompting more attention to this supply opportunity now?

A: Actually, we haven't really focused on CTV advertising yet. The acquisition of Wurl is aimed at leveraging its connections with media companies to access some online resources for business development, such as MAX bidding. Our core business and most of our revenue come from the demand-side platform (DSP), where advertisers find us, and we help them place ads. However, we need supply, and Wurl brings a lot of supply. Last year, the platform could only display game ads, but showing game ads on full-screen TVs, which require users to download games on their mobile phones, was not effective. With the addition of consumer advertising, DTC commercial advertising, and a wide range of advertisers in fintech, automotive, and other categories to the platform, expanding creativity to large screens could be very appealing. However, there are challenges such as attribution issues and a lack of clear calls to action. If successful, it will open up a new large-scale performance advertising channel, so we will start working on this this year.

Q: What changes have you seen this quarter in model improvements, and how significant is their impact on growth?

A: Growth factors can be broken down into several aspects. First, the model continues to learn, and we expect to benefit from this every quarter. Currently, there is no sign of slowdown, contributing to this quarter's growth. Second, there are gradual and incremental improvements to the existing model. Although unpredictable and not huge enhancements, some incremental effects have been realized this quarter. Third, the launch of new models like ChatGPT 4 and 4o has brought significant incremental improvements. Additionally, there are seasonal factors in the fourth quarter related to shopping and shopping behavior, and people are spending more time on mobile devices this quarter, which may increase the likelihood of mobile game consumption and transactions. These factors together drive growth.

Q: Regarding the e-commerce business, do you still believe it will make a substantial contribution to revenue by 2025? What is the incremental growth potential of the e-commerce business, and how does it provide a framework for the model in 2025?

A: We remain confident that the e-commerce business will make a substantial contribution to revenue by 2025, but it is difficult to predict when and to what extent growth will occur. We will provide more details when we see growth happening.

First, the traditional business (game and app advertising) itself has a 20% annual growth rate. In addition, the two factors of model improvement and platform openness attracting more customers may bring significant upside potential. In terms of model improvement, AI development and research are still in the early stages, and improvements in the model will impact revenue;

Second, regarding platform openness, the e-commerce business is currently in a pilot phase, and as we attract a large number of customers in the future, the business will continue to grow strongly. We view it as part of the overall business, better monetizing over 1 billion active users without focusing on the revenue of each category.

Q: How many people are currently participating in the e-commerce pilot project?

A: We do not disclose specific numbers. From industry research and Twitter discussions, it is clear that the number of participants is definitely not in the dozens or thousands. The EBITDA metrics for each employee indicate the level of team streamlining. Due to the limitations of the core team, the entire advertising business has about 1,000 people, including all business segments like Adjust, Wurl, and AppLovin, while the e-commerce marketing team has only about 20 peopleFrom an operational perspective, the team cannot manually onboard a large number of advertisers in the short term.

Q: The self-service mentioned earlier will truly open up the situation and onboard more clients, right?

A: Self-service and automation tools will be very helpful. In a limited pilot, a few clients have generated considerable revenue for this category. As we gradually open up, we believe it will have a significant impact on client businesses, benefit partner publishers, increase the diversity of mobile game advertisers, and avoid the constant appearance of competitor ads in game publishers' advertising, opening the door for more supply to go online, especially with the future onboarding of the MAX platform.

Q: With the launch of e-commerce solutions, have you seen reactions from existing competitors? Are you prepared for this? After a full rollout, what do you think the industry will look like, and how will it differ from today?

A: We haven't paid much attention to the competition. The platform is not taking the same advertising revenue from others. For example, a mattress manufacturer may gain some business from advertising on social media, but after joining our platform, they can acquire new transactions from new customers who may not respond to their ads on social media. They will not reduce their advertising spend on social media but will increase their investment on our platform to achieve incremental sales. This means we are expanding the economy, allowing advertisers to achieve measurable profits based on performance, expanding their business without harming competitors.

Q: Regarding e-commerce and new vertical market promotions, has the strategy evolved since the pilot, especially during the holiday period? What key insights or shifts have occurred? How do you view market promotion? What are the priority areas for demand generation in 2025?

A: We have been focusing on the mid-sized DTC (direct-to-consumer) market, with annual transaction volumes between $10 million and $250 million. These companies act quickly, have short legal processes, and do not require large teams for onboarding. The companies participating in the pilot are of moderate size, and due to social media noise, feedback loops, and initial positive results from customers, many clients have proactively approached us, with a long queue of clients waiting to onboard the platform. The team has been busy during the holiday period, but aside from the increase in the number of clients, there have been no other changes in strategy until we launch tools that allow clients to onboard themselves, achieving the same results as the current management support.

Q: Regarding category-specific optimization issues, can AXON independently address these adjacent non-DTC areas, start driving conversions, and adjust the model on its own, or does the team need to make specific category adjustments? How much product work is needed to expand into these non-DTC categories?

A: Currently, we do not have the resources to build custom models for every vertical. The goal is to build more complex models that are smart enough to figure out how to market any product on their own. The reality is that once companies from a wide range of categories onboard, the model will be able to function effectively.

Q: So currently, demand is high, and the challenge we face is just market promotion, or making advertisers aware that we can provide demand?

A: Yes. The challenge lies in launching tools to onboard more companies while ensuring that advertisers can place ads in a profitable, incremental, and measurable way. As long as each advertiser relationship on the platform can achieve such results, there will be positive word-of-mouth, naturally bringing us clientsQ: What is the situation regarding how to meet additional e-commerce demand, and is this different from inventory that converts better for games?

A: This is the same inventory, covering full-screen video ads for games played by over 1 billion active users daily, without differentiated inventory access; it is a unified bidding process. Whether mattress ads or game ads are selected depends on model decisions, and the inventory situation is the same for all categories for end users and end suppliers.

Q: What early feedback has been received regarding the D28 (28 days) advertising ROAS model in gaming? Will it drive additional spending this year?

A: These models have been launched for several quarters. Communication with game developers indicates that the models are generally very effective for everyone, especially the longer-term models, which are the most expressive models provided by the platform, enhancing mobile clients' advertising spending on the platform that can yield effective returns over the past few quarters.

Q: Investors are concerned about the take rate between gaming and e-commerce. We seem not to charge a fixed rate like other ad tech platforms. As more shifts to e-commerce, how should we think about the overall business composition? Does it affect the commission rate?

A: As the monetization of purchasable inventory improves, the take rate will naturally rise. However, the platform has never optimized the commission rate; the system will purchase users deemed monetizable regardless of price. It is important to pay the highest price the market can offer; the better the performance, the stronger the business, as the commission rate can expand.

Q: With e-commerce now accounting for a larger share of the business, what about seasonal issues? How should we view seasonality in terms of revenue growth and cost structure in the future?

A: In terms of costs, it mainly involves data center costs and salaries, driven by usage rates. With revenue growth, it is expected that about 10% of the annual increase will come from rising data center costs, so cost fluctuations are not significant. In terms of e-commerce, as it becomes a larger part of the business, it will be affected by e-commerce seasonality, such as slight increases in revenue during Black Friday and holiday periods.

Q: What are the driving factors for e-commerce's quarter-over-quarter growth in the first quarter? Should we consider it more as existing advertisers increasing spending or new incremental advertisers driving it?

A: The e-commerce situation has not yet been segmented. In traditional advertising, the first quarter typically sees a decline due to two fewer days in the fourth quarter and seasonal factors, and the e-commerce category will be similar. Even without adding new advertisers, this category may still expand relative to the fourth quarter. Overall, the business is expected to see a 4% quarter-over-quarter growth in the first quarter, despite having two fewer days, because there is confidence that even considering seasonality or fourth-quarter growth, the first quarter will continue to grow.

Q: How willing are long-term e-commerce advertisers compared to new advertisers to increase their daily spending limits?

A: The situation is unstable because the e-commerce category just entered in the fourth quarter. Many e-commerce companies invest heavily during Black Friday and holiday periods, and inventory needs to be replenished during Christmas, leading to spikes and drops in spending. The platform operates in performance marketing; as they replenish inventory entering the first quarter, spending will restart. The observed fluctuations are due to the nature of the business; seeing good results on the platform means that over time, after adjusting for seasonal factors, there will be a sustained growth in revenue lines, not only because success on the platform allows businesses to grow and invest more in marketing funds but also because they are still in the early stages in this fieldThe e-commerce model is several years behind the gaming model and still has a long way to go; the platform will continue to improve.

Q: I am quite interested in your point that the technology you mentioned can be applied to different types of advertisers. There are different categories within e-commerce, and you have launched Axon Pixel, which e-commerce merchants are already using on their websites to assist with attribution. Is this sufficient to reach all different types of e-commerce advertisers? Has the technology at least been fully deployed in terms of attribution to reach 10 million advertisers globally? Given the efficiency of artificial intelligence, will it be used for in-app installations in non-gaming categories? Will this become a growth area by 2025?

A: We currently have relevant technology to help advertisers achieve effective attribution, but I cannot say that the technology is sufficiently mature in all aspects of the business; we are continuously adding tools to achieve more accurate attribution and collaborate with advertisers. Nevertheless, it is effective in the website and e-commerce categories as well as in the broader non-gaming network categories, and any customer with a website can automate ad placements after accessing the platform. However, for gaming advertisers, non-gaming app advertisers, and merchants with both websites and applications, more work needs to be done to get them online, which is what we are working on this year. It is expected that as we move into the second half of this year, next year, and the longer-term future, we will touch on new advertising categories, but a turning point will not occur in 2025.

Q: Since taking over the human resources function last quarter, what changes have been made in recruitment and streamlining operations, or what should we pay attention to?

A: The focus is on simplifying teams and processes and truly concentrating. We announced matters related to the sale of the App business, and the company currently faces significant opportunities and a long-term development blueprint. When such opportunities arise, it is necessary for everyone to focus on them. The company is profitable but has been announcing layoffs, which we do not want to continue doing. After restructuring human resources, we will cut areas that are not fully aligned with the strategy to focus on current opportunities, and after achieving this, there will be more exciting moments.

Q: Non-gaming helps supply; do you expect to gain market share in the intermediary aspect, and how will additional supply help the gaming business over time?

A: Our market share is already quite high; let's not talk about market share but rather about inventory. We have been acquiring new game publishers to run game ads, and we are currently a leader in monetization. There is growth in the gaming supply business, but there is not a huge space in the short term. Many consumers of in-app purchase applications spend a lot of time on games but do not run ads because they can effectively monetize user time through in-app purchases. For example, King generated considerable additional revenue by launching non-gaming ads when it went public before being acquired. Most ads in the traditional mobile gaming intermediary market target games that compete with the publishers' games. By introducing more non-gaming category advertisers and selling MAX products and DSP platforms to major or exclusive game publishers that monetize primarily through in-app purchases, introducing this supply will create another increment.

Q: How do you view the amount of spending growth that the current mobile gaming environment or ecosystem can absorb? Can inventory value double or even triple? Do you see this as a limiting factor for future growth?A: I do not believe this is a limiting factor. Over the past year, the supply has remained basically unchanged, while the company has grown by 80% - 90%, and both supply and demand are still in the early stages. On the demand side, we have just entered some areas that can better monetize inventory, and the current monetization level is lower than that of some major social platforms. Although the ad units are high-impact full-screen videos, we have not yet brought in all types of advertisers globally. In the future, as we bring in more advertisers and increase demand density, the algorithms and data footprints will become stronger. Even if the supply remains unchanged, the same daily active users will be better monetized. The path of demand expansion driving growth is still long, and supply is just a bonus; it will not have a significant impact on growth like demand expansion.

Q: What are your thoughts on moving Audience+ from the pilot phase to self-service? What have you learned from the pilot that will help improve or change the solution before moving to self-service? When do you expect it to achieve self-service in 2025?

A: We will launch self-service at the appropriate time. The key is to have tools that automate every step of the process, including AI agents that help advertisers go live and start using the platform, making them feel like they are interacting with a person while actually interacting with a robot and being able to start smoothly. Another key point is that when opening a large platform, we must build a lot of tools and content review controls to prevent fraud. Just like Facebook and Google, if a large platform does not build tools to prevent fraud, there will be a lot of fraudulent activities flooding in after the platform is opened. We will take this seriously and cautiously. Currently, the growth rate is fast, and we are controlling the pace, developing all the tools in the background. Once we are ready, we will launch, and it will have a significant impact, with effects accumulating over quarters, years, and even more than a decade.

Q: Is the application sales situation phased, gradually phasing out some studios, or will all 10 studios stop operating at once after the transaction is completed, with application revenue no longer existing? When will the transaction be completed, and how can we understand the timing dynamics of this year's application revenue modeling?

A: We will sell the entire application business, removing it from the profit and loss statement and the balance sheet all at once. The transaction is expected to be completed in the second quarter, but it may be affected by regulatory approvals, so the timing may vary slightly. The goal is to complete it in the second quarter.

Q: How much of the personalized advertising experience plan can be achieved through what has already been built, rather than requiring additional investment to achieve personalization goals? What should we focus on externally to understand how this plan might be implemented and more widely promoted in the business in the coming years?

A: Currently, it is all research and development. Taking "Candy Crush" as an example, right now, humans upload 20 ads, and AI decides to display one of them, which is not very personalized. The ad formats and video types that users see are exactly the same as those seen by other users who view the campaign. With generative AI and large language models, we can run top creative ideas through models and create many variations. Facebook and Google are also talking about generative AI-driven ads. The platform is researching this and believes it will have a significant impact on driving consumer responses to ad units, as there is a full-screen video experience, and often there are HTML ads or even mini-games. There is still a lot of work to be done to personalize this experience, and we need to build models that can automatically complete these tasks and operate within GPU constraints. This is one of the key focuses for this yearQ: Do social ads used by e-commerce brands in mobile games provide an opportunity for e-commerce and other verticals to create ads that resemble mobile game ads?

A: Possibly. Hopefully, it's not decided by humans, but by AI and machines. Once we enter the realm of machine creation, it can generate more variants than humans, whether it's gamified ads for e-commerce products or things that are currently unimaginable; theoretically, machines can do it all, and advertisers and consumers will benefit from all the new ad variants produced by the system.

Q: What feedback have you received from game companies when entering e-commerce and other verticals?

A: Game companies see better performance on the platform every quarter and are pleased that this is a catalyst and necessity for their growth. Publishers are delighted to see the shift towards e-commerce, as the worst thing for a game publisher is to monetize their games using competitors' ads, while the platform allows MAX publishers to see these ads running. Some publishers are shifting to non-gaming categories, and we are excited about this trend.

The risk disclosure and statement of this article: Dolphin Investment Research Disclaimer and General Disclosure