Adam Foroughi, Applovin CEO: Surviving a 92% Drawdown, Ads as ML 1.0 & the $50B Game Ad Market
In this episode of All-In with Chamath, Jason, Sacks & Friedberg, Applovin CEO Adam Foroughi explains how his company operates as a monetization platform for mobile game developers, inserting ads within games that reach over a billion daily players. Foroughi discusses the $50 billion mobile gaming advertising market, Applovin's performance-based revenue model, and how the company's transition to deep learning algorithms significantly improved their advertising precision and competitive positioning.
The conversation covers Applovin's dramatic stock recovery following a 86% decline, which management addressed through aggressive share buybacks and cultural reinforcement. Foroughi also addresses how privacy regulations have affected advertising effectiveness, the difference between search-based and discovery-based advertising, and his perspective on how AI agents and new interfaces like AR glasses may reshape consumer shopping behavior while preserving the fundamental value of product discovery.
This is a preview of the Shortform summary of the Sep 20, 2026 episode of the All-In with Chamath, Jason, Sacks & Friedberg
Sign up for Shortform to access the whole episode summary along with additional materials like counterarguments and context.
1-Page Summary
Applovin's Model & $50b Gaming Ad Market Opportunity
Applovin operates as a monetization platform for mobile game developers, inserting ads within games that reach over a billion daily players. Co-founder Adam Foroughi explains that the company is fundamentally "an advertising company that's helping mobile game developers monetize that space." With roughly $11 billion in annual ad spend disclosed almost two years ago and 60% year-over-year growth since, Foroughi estimates current annual ad spend on their platform at around $20 billion, contributing to a total mobile gaming advertising ecosystem worth approximately $50 billion annually.
The platform's success stems from a consent-based model where adult casual gamers voluntarily watch ads in exchange for in-game rewards. Historically focused on driving engagement between games, Applovin has evolved with advances in deep learning to facilitate real-world e-commerce transactions, using machine learning to target high-intent shoppers ready to make purchases.
Applovin's revenue model is performance-based and arbitrage-driven. Advertisers pay rates below their retail prices, ensuring positive returns on ad spend. The company reports exceptional profitability with 84% EBITDA margins, according to Foroughi, demonstrating the efficiency of their automated, tech-driven approach that minimizes revenue leakage typically found in traditional advertising supply chains.
Technology Innovation: Deep Learning and Competitive Advantage
Foroughi explains that Applovin's transition from ML 1.0 regression models to ML 2.0 deep learning significantly improved their advertising algorithms' performance. This shift enhanced recommendation systems and shopping intent identification precision, with exponential growth acceleration starting after their first deep learning model launched in April. The company strategically acquired game studios to secure proprietary data for training these advanced models, as studios typically don't share data with third parties.
Applovin's mobile gaming ad specialization offers a competitive edge against larger rivals. Foroughi credits their lean structure populated by subject matter experts and access to superior mobile gaming data through studio acquisitions. This proprietary data stream and continuous algorithmic innovation create a competitive moat difficult for even the largest companies to replicate. He also highlights that cultural values of humility, intelligence, and hard work contribute to the company's innovative edge.
At the core of Applovin's operations is automation and algorithmic optimization, designed to minimize human intervention in ad arbitrage. This automated approach streamlines operations and supports large-scale revenue growth while ensuring optimal matching between advertisers and consumers.
Stock Recovery Strategy: Buybacks, Culture, Market Perception
Applovin went public in April 2021 at a $28 billion market cap, peaking at $40 billion before collapsing to just $3.8 billion by 2022—an 86% decline—despite generating $1 billion in EBITDA that year. Foroughi attributes this collapse to COVID-era IPO over-saturation and poor investor quality, with high supply from insiders eager to sell meeting low institutional demand.
The company responded with an aggressive $6 billion share repurchase program, retiring 20–25% of outstanding shares in 2.5 years. At peak values, this buyback created over $50 billion in shareholder value. Management also boosted team morale through performance-based equity incentives extended beyond typical CEO-level plans and instilling an "us versus the world" mentality that turned adversity into opportunity.
Initially, Foroughi halted investor outreach, believing engagement was pointless while the market undervalued the stock. However, after launching their deep learning model and achieving operational recovery, he resumed meetings in September 2023. The pent-up recognition played out dramatically: within a week, shares doubled from $80 to $150, and market cap soared from $28 billion to $55 billion, signaling institutional recognition of Applovin's real value.
Privacy Regulations and Their Impact on Advertising Business
Foroughi explains that Apple's privacy restrictions and EU regulations now force advertisers to target broader user cohorts rather than individuals, degrading ad relevance. Ironically, he notes, users frequently complain that ads are now less relevant and more spam-like, when research shows consumers generally prefer personalized ads that help them discover products.
Clear regulatory boundaries have ultimately helped the industry adapt. Foroughi stresses that well-defined privacy regulations enable companies to adjust their systems accordingly, and advanced algorithms like deep learning networks have proven adept at optimizing within privacy-constrained environments. Despite these restrictions, mobile gaming ads maintain high margins—Applovin continues reporting 84% margins even amid new privacy rules, with accelerating growth demonstrating the sector's resilience.
Future of Advertising in AI World With Agents and New Interfaces
Foroughi describes how AI agents will streamline purchasing for some consumers, automating routine purchases like supplement subscriptions. However, he points out that most mass market consumers value the shopping experience itself—browsing, comparing goods, and the emotional satisfaction from discovery—more than the financial optimization agents provide for lower-value purchases.
He distinguishes between search-based ads, where consumers already know what they want, and discovery-based advertising, which presents products consumers didn't know existed. Discovery-based ads create economic growth by generating new purchases and trials that wouldn't otherwise occur, providing emotional satisfaction through the thrill of discovery while expanding consumer spending.
Looking toward new computing paradigms like AR glasses, Chamath Palihapitiya raises the prospect of new touchpoints for product recommendations. Foroughi believes that while technologically advanced consumers may quickly adopt agent-driven commerce, most shoppers will remain motivated by the experience and joy of product discovery. Interface shifts may change how consumers encounter ads, but the fundamental value of discovery-based advertising will persist.
1-Page Summary
Additional Materials
Clarifications
- A monetization platform integrates advertising technology into mobile games to generate revenue for developers. It manages the placement, timing, and targeting of ads shown to players, often rewarding them with in-game benefits for watching. This system balances user experience with maximizing ad revenue by ensuring ads are relevant and engaging. It also provides analytics and optimization tools to improve ad performance and earnings over time.
- A performance-based revenue model means Applovin earns money only when ads achieve specific results, like clicks or purchases. Arbitrage-driven refers to buying ad space at lower prices and selling it at higher rates, profiting from the price difference. This model aligns Applovin’s incentives with advertisers’ success, reducing risk for advertisers. It also leverages Applovin’s technology to optimize ad placements for maximum return.
- EBITDA stands for Earnings Before Interest, Taxes, Depreciation, and Amortization, measuring a company's operating profitability. EBITDA margin is the percentage of revenue remaining after these operating expenses are deducted. An 84% margin means the company keeps 84 cents of every dollar earned before non-operating costs, indicating very high efficiency and low operating costs. Such a margin is exceptional because most companies have much lower margins due to higher expenses.
- ML 1.0 regression models use simple mathematical equations to predict outcomes based on input features, often assuming linear relationships. ML 2.0 deep learning employs multi-layered neural networks that can model complex, non-linear patterns by learning hierarchical feature representations. Deep learning excels at handling large, unstructured data like images or text, improving prediction accuracy. This shift enables more precise targeting and personalization in advertising algorithms.
- Proprietary data from game studios includes detailed user behavior, in-game actions, and engagement patterns unique to their games. This data helps machine learning models identify player preferences and predict which ads or offers will be most effective. By training on this exclusive data, models improve targeting accuracy beyond what generic data allows. Access to such data creates a competitive advantage because it is not available to other advertisers or platforms.
- Ad arbitrage in advertising involves buying ad space at lower prices and reselling it at higher rates by targeting audiences more likely to engage, generating profit from the price difference. Algorithmic optimization uses data-driven algorithms to continuously adjust ad placements, bids, and targeting to maximize campaign performance and return on investment. Together, they enable platforms like Applovin to efficiently match ads with users who have high purchase intent, increasing advertiser ROI while maintaining profitability. This automation reduces manual intervention, speeds decision-making, and minimizes wasted ad spend.
- The stock price collapse despite strong EBITDA often occurs when market sentiment and investor confidence decline sharply. High insider selling can flood the market with shares, reducing demand and driving prices down. Additionally, IPO over-saturation can lead to investor fatigue and lower valuations. Strong earnings alone may not offset negative perceptions or macroeconomic factors affecting stock prices.
- A $6 billion share repurchase program means the company buys back its own shares from the market using $6 billion in cash. This reduces the number of shares available, increasing earnings per share and often boosting the stock price. It signals confidence from management that the stock is undervalued. Repurchases can also improve financial ratios and return value to shareholders without paying dividends.
- Apple's privacy restrictions, like App Tracking Transparency, require apps to get user permission before tracking their data across other apps and websites. EU regulations, such as GDPR, limit how companies collect and use personal data, enforcing strict consent and transparency rules. These rules reduce advertisers' ability to target individual users precisely, forcing broader audience targeting. Consequently, ad effectiveness can decline, as ads become less personalized and relevant.
- Search-based ads target consumers actively looking for a specific product or service, matching their explicit intent. Discovery-based advertising introduces consumers to new or unexpected products they were not actively seeking. This type of advertising relies on curiosity and emotional engagement to drive purchases. It expands market demand by creating interest beyond existing consumer needs.
- AI agents are software programs that use artificial intelligence to perform tasks on behalf of users, such as making purchases automatically. They analyze user preferences, past behavior, and product data to select and buy items without manual input. These agents can handle routine or repetitive purchases efficiently, saving time and effort. Their effectiveness depends on accurate data and user trust in automation.
- Interface shifts refer to changes in how users interact with technology, such as moving from screens to augmented reality (AR) glasses. AR glasses overlay digital information onto the real world, creating new, immersive ways for users to receive ads. This changes advertising by enabling context-aware, seamless product recommendations integrated into daily life. Such shifts require advertisers to design experiences that fit naturally into these new interaction modes.
- "Institutional recognition" means that large, professional investors like mutual funds and pension funds acknowledge a company's true value and start buying its stock. Their participation often boosts stock prices due to the large volume of shares they purchase. This recognition signals confidence in the company's prospects and can attract more investors. It contrasts with retail investors, who are individual, non-professional traders.
Counterarguments
- While Applovin claims its consent-based ad model is voluntary, some critics argue that in-game rewards can create a form of coercion, especially if game progression is significantly slowed without watching ads.
- The high profitability (84% EBITDA margins) may not be sustainable long-term as competition increases, privacy regulations tighten further, or user tolerance for ads declines.
- The acquisition of game studios for proprietary data could raise concerns about market consolidation and reduced competition in the mobile gaming ecosystem.
- Although Applovin emphasizes the value of discovery-based advertising, some users find such ads intrusive or disruptive to their gaming experience.
- The assertion that consumers generally prefer personalized ads is contested; some research indicates a significant portion of users are uncomfortable with data collection and targeted advertising, regardless of ad relevance.
- The focus on automation and minimizing human intervention may lead to less oversight and accountability, potentially increasing the risk of algorithmic bias or unintended negative outcomes.
- While Applovin attributes its stock price collapse to external market factors, some investors might argue that internal factors such as business model risks or overvaluation also played a role.
- The claim that privacy regulations have not impacted profitability may not hold if future regulations become more stringent or if enforcement increases.
- The strategy of large-scale share buybacks can be criticized for prioritizing short-term shareholder value over long-term investment in innovation or employee welfare.
- The narrative that most consumers value the shopping experience over convenience may not apply universally, as trends show increasing adoption of automated and agent-driven commerce, especially among younger demographics.
Get access to the context and additional materials
Applovin's Model & $50b Gaming Ad Market Opportunity
Applovin operates as a monetization platform for mobile game developers, enabling ads within games that collectively reach over a billion daily players. The company taps into a massive and fast-growing market, facilitating consent-based advertising that has redefined mobile gaming's economic landscape.
Applovin Is a Monetization Platform for Mobile Game Developers, Inserting Ads Within Games That Reach Over one Billion Daily Players
Applovin’s core business is helping mobile game developers monetize their user base through advertising. According to co-founder Adam Foroughi, “what we are ultimately is an advertising company that’s helping mobile game developers monetize that space.” The platform reaches over a billion people daily, primarily adults and heads of households, all playing mobile casual games. These games represent a vast, engaged audience.
The company disclosed roughly $11 billion in annual ad spend on its platform almost two years ago and has since grown about 60% year over year. That puts current annual ad spend on their platform at an estimated $20 billion. Considering other ad platforms operating in the same space, Foroughi estimates that the entire mobile gaming ecosystem now supports roughly $50 billion in annual advertising.
Adult casual gamers are the key audience. They often choose to watch ads voluntarily in exchange for in-game rewards, creating a consent-based, intent-driven model which allows Applovin to capture high levels of user engagement while respecting user choice.
Evolved From Gaming Rides to Deep Learning For Predicting Shopping Behavior and E-Commerce Transactions
Historically, Applovin used its platform to drive engagement from one game to another, sparking user intent within the gaming ecosystem itself. But the company has evolved alongside advances in deep learning. Foroughi explains that deep learning models are now so robust that Applovin can use its massive audience to drive real-world transactions, introducing shopping behavior and e-commerce opportunities directly through game ads.
Their machine learning capabilities help target high-intent shoppers, identifying when a user is ready to make a purchase. This shift—from internal game promotion to facilitating e-commerce transactions—has significantly boosted investor enthusiasm and accelerated Applovin's growth trajectory, allowing the company to tap into much larger economic sectors.
Applovin's Revenue Model: Performance-Based and Arbitrage-Driven System Based On Transactions and Return On Ad Spend
Applovin’s revenue is rooted in a performance-based, arbitrage model. Advertisers, such as cosmetics brands selling lipstick, use Applovin’s platform to reach consumers. Foroughi describes the process: ...
Here’s what you’ll find in our full summary
Applovin's Model & $50b Gaming Ad Market Opportunity
Additional Materials
Clarifications
- A monetization platform integrates with mobile games to enable developers to earn revenue, primarily by displaying ads or offering in-app purchases. It manages ad placement, targeting, and payment processing, allowing developers to focus on game creation. These platforms use data to optimize which ads are shown to maximize user engagement and revenue. They also ensure ads comply with user consent and privacy regulations.
- Consent-based, intent-driven advertising means users willingly choose to watch ads, often in exchange for rewards, showing clear interest. This approach respects user privacy and avoids intrusive ads by requiring permission before engagement. It targets users who have demonstrated intent, increasing the likelihood of meaningful interactions. This model improves ad effectiveness and user experience by aligning ads with user preferences and actions.
- "Annual ad spend" refers to the total amount advertisers pay to place ads on Applovin's platform each year. This figure indicates the platform's market size and advertiser demand but is not the same as Applovin's revenue. Applovin earns a portion of this spend, typically a fee or percentage, based on ad performance and transactions. Higher ad spend generally means more revenue potential for the platform.
- A performance-based revenue model means Applovin earns money only when ads lead to measurable actions, like purchases. Arbitrage-driven refers to buying ad space at a lower cost and selling it at a higher price, profiting from the difference. This model incentivizes efficiency and accuracy in targeting to maximize returns. It aligns Applovin’s earnings with advertisers’ success, reducing risk for both parties.
- Applovin’s deep learning models analyze vast amounts of user data, such as in-game behavior and ad interactions, to identify patterns linked to purchasing intent. These models use neural networks to detect subtle signals indicating when a user is likely to buy a product. By continuously learning from new data, the system improves its accuracy in predicting shopping behavior over time. This enables Applovin to target ads to users at moments when they are most receptive to making a purchase.
- Advertisers pay Applovin a fee lower than the product's retail price but higher than the product's profit margin, ensuring the ad cost is covered by the sale profit. This means the brand still makes money after paying for the ad because the product's cost of goods sold is less than the retail price. Applovin's platform targets users likely to buy, increasing conversion rates and reducing wasted ad spend. This efficient targeting allows advertisers to scale budgets while maintaining profitability.
- Revenue leakage in advertising supply chains refers to the loss of potential income due to inefficiencies, fraud, or hidden fees between advertisers and publishers. It occurs when a portion of the advertiser's budget fails to reach the intended audience or is absorbed by intermediaries. This reduces the overall effectiveness and profitability of ad campaigns. Minimizing leakage improves transparency and maximizes returns for advertisers.
- EBITDA margins measure a company's operating profitability by showing earnings b ...
Counterarguments
- While Applovin claims its advertising model is "consent-based" and "intent-driven," some critics argue that the incentive structure (offering in-game rewards for ad views) may pressure users into watching ads, blurring the line of true consent.
- The focus on adult casual gamers does not preclude exposure to younger audiences, as many mobile games are played by minors, raising concerns about ad targeting and child privacy.
- High reported EBITDA margins (84%) may reflect aggressive cost-cutting or accounting practices, and may not be sustainable as competition increases or regulatory scrutiny intensifies.
- The rapid growth in ad spend on the platform could be partly driven by increased ad load or frequency, which may negatively impact user experience and long-term engagement.
- The arbitrage-driven revenue model depends on continued advertiser willingness to pay for conversions; if return on ad spend declines, advertisers may reduce budgets, impacting Applovin’s growth.
- Heavy reliance on machine learning and automation introduces risks related to algorithmic bias, lack of transparency, and potential errors in targeting or attribution.
- The shift toward e-commerce and real-world transactions v ...
Get access to the context and additional materials
Technology Innovation: Deep Learning and Competitive Advantage
Applovin's Switch From Ml 1.0 Regression Models to Ml 2.0 Deep Learning Unlocked Exponential Growth
Adam Foroughi explains that Applovin's transition from ML 1.0 regression models to ML 2.0 deep learning profoundly improved the performance of their advertising algorithms. By moving away from traditional statistical methods and embracing deep learning, Applovin significantly enhanced its recommendation systems and the precision with which it identifies shopping intent. Foroughi notes that the better the deep learning model works, the greater the advertiser's returns, driving more revenue and rapid company growth. The exponential acceleration in growth started after launching their first deep learning model in April.
Foroughi also highlights the strategic acquisition of game studios as a means to secure proprietary data needed for training these advanced models since studios typically do not share data with third parties. By acquiring studios, Applovin could seed their deep learning models with valuable training data, leading to rapid market success.
The continuous evolution in AI intertwines the progress of deep learning used in advertising recommendations with the development of large language models. Foroughi points out that recommendation algorithms and language models share similar research trajectories; advancements in one often influence the other. Much of the innovation and research that fuels today's AI breakthroughs derives from advertising technology foundations. The economic value of these models is instantly quantifiable—accurate ad predictions immediately translate into revenue and performance metrics, underscoring the significance of deep learning in advertising.
Mobile Gaming Ad Specialization Offers Competitive Edge Against Larger Rivals
Applovin's focused expertise in mobile gaming ads enables them to compete—and often outpace—larger, more resource-rich rivals. Foroughi credits this specialization to a lean structure populated by subject matter experts in mobile gaming and e-commerce psychology. Their deep knowledge allows for rapid decision-making and adaptability, which is a critical advantage over slower-moving, larger companies.
Another critical advantage stems from Applovin's access to superior mobile gaming data, facilitated by their acquisition of game studios. This constant stream of proprietary data and ongoing algorithmic innovation forms a competitive moat that is challenging for even the largest companies to replicate. Foroughi emphasizes that while resources matter, the complexity and continuous improvement of the underlying technology, combined with unique data, make it extremely difficult for competitors to catch up once a company reaches scale.
Foroughi also highlights that cultural values of humility, i ...
Here’s what you’ll find in our full summary
Technology Innovation: Deep Learning and Competitive Advantage
Additional Materials
Clarifications
- ML 1.0 regression models use simple mathematical formulas to find relationships between variables, often assuming linear patterns. ML 2.0 deep learning employs neural networks with many layers to automatically learn complex patterns from large amounts of data. Deep learning can model non-linear, intricate relationships that traditional regression cannot capture. This allows for more accurate predictions and better handling of diverse, unstructured data like images or text.
- "Shopping intent" refers to the likelihood that a user is interested in purchasing a product or service. Advertising algorithms analyze user behavior, such as search queries, clicks, and browsing patterns, to predict this intent. Identifying shopping intent helps target ads to users who are more likely to convert into buyers. This increases ad effectiveness and advertiser return on investment.
- Deep learning models improve recommendation systems by learning complex patterns from large amounts of data, capturing subtle user preferences and behaviors. They use multiple layers of artificial neurons to automatically extract relevant features without manual input. This enables more accurate predictions of what users want, even in diverse or changing contexts. As a result, recommendations become more personalized and effective at driving user engagement.
- Ad arbitrage is the practice of buying advertising space at a low cost and then reselling it at a higher price to generate profit. Minimizing human intervention reduces errors, speeds up decision-making, and allows algorithms to optimize ad placements and pricing in real-time. This automation improves efficiency and scalability, enabling the platform to handle large volumes of transactions without costly manual oversight. It also helps prevent value leakage by ensuring precise matching and pricing based on data-driven predictions.
- Acquiring game studios gives Applovin exclusive access to detailed user behavior and engagement data that is not publicly available. This proprietary data is crucial for training deep learning models to better predict player preferences and ad effectiveness. Without such data, models rely on less accurate, generic datasets, limiting performance. Owning the data source creates a competitive advantage by enabling more precise and personalized advertising algorithms.
- Advertising recommendation algorithms and large language models both rely on deep learning techniques to analyze vast amounts of data and identify patterns. They use similar neural network architectures, such as transformers, to process and predict user behavior or language sequences. Advances in one area, like improved model training or architecture, often benefit the other by providing new methods or insights. Both aim to make accurate predictions—whether about user preferences or language context—enhancing personalization and relevance.
- Value leakage in advertising technology refers to the loss of potential revenue or efficiency due to inefficiencies in the ad delivery process. It can occur when ads are shown to uninterested audiences, when pricing is suboptimal, or when there is fraud or wasted impressions. Minimizing value leakage ensures that advertisers get the best return on investment and publishers maximize their earnings. Effective algorithms reduce this leakage by improving targeting, pricing, and campaign optimization.
- Predictive analytics uses historical data and machine learning models to forecast future behaviors and outcomes. For campaign pricing, it estimates the value of ad placements by predicting how likely users are to engage or convert, enabling dynamic price adjustments. For advertiser-consumer matching, it predicts which users are most likely to respond positively to specific ads, improving targeting accuracy. This data-driven approach maximizes return on investment by optimizing both co ...
Counterarguments
- The effectiveness of deep learning models is highly dependent on the quality and diversity of training data; proprietary data from acquired studios may introduce biases or limit generalizability to broader markets.
- Strategic acquisitions for data access can raise concerns about market consolidation and reduced competition, potentially stifling innovation in the industry.
- While automation and algorithmic optimization can increase efficiency, they may also reduce transparency and accountability in ad targeting and pricing decisions.
- The immediate economic value of deep learning in advertising does not account for potential long-term risks, such as user privacy concerns or regulatory challenges related to data usage.
- Specialization in mobile gaming ads may limit Applovin’s ability to diversify into other advertising verticals, potentially exposing the company to market volatility within the gaming sector.
- Relying heavily on proprietary data and closed ecosystems could hinder co ...
Get access to the context and additional materials
Stock Recovery Strategy: Buybacks, Culture, Market Perception
Applovin's April 2021 IPO at a $28 Billion Valuation Was Followed by a 2022 Market Downturn That Reduced the Company's Market Cap to $3.8 Billion, an 86% Decline Despite Generating $1 Billion in EBITDA
In April 2021, Applovin went public at a $28 billion market cap, based on strong fundamentals including $600 million of EBITDA, and at one point peaked as high as $40 billion. However, by 2022 the stock suffered a relentless slide, bottoming out at just a $3.8 billion market cap—an 86% decline—even though the company delivered $1 billion in EBITDA that year. Adam Foroughi, Applovin’s CEO, attributes much of this collapse to the over-saturation of COVID-era IPOs, which resulted in poor investor quality. Many blue chip firms didn’t perform rigorous research, and much of the supply on the public markets came from private equity, ex-co-founders, and other insiders eager to sell. With demand low and supply suddenly high, Applovin’s valuation quickly collapsed. Foroughi, drawing on his finance background, recognized a huge gap between the company’s market performance and its intrinsic financial health—an environment ripe for strategic opportunity.
Response to Stock Collapse: Aggressive Share Repurchase Transforms Pessimism into Long-Term Wealth
Foroughi and his team turned crisis into opportunity by initiating a massive stock buyback. He reasoned that, since Applovin was generating substantial free cash flow, it could “become our best investor.” The company embarked on a $6 billion repurchase program, retiring 20–25% of its outstanding shares in just 2.5 years. At peak values, this buyback created over $50 billion in shareholder value. This aggressive buyback required deep commitment to company fundamentals and the courage to ignore negative short-term market sentiment. As Applovin’s core advertising technology and business performance rebounded, this approach proved prescient—the company was able to turn a period of deep market pessimism into a platform for long-term wealth creation.
Management Boosted Team Morale and Retention During the Crisis Through Performance-Based Equity Incentives and an "Us Versus the World" Mentality That Turned Adversity Into Opportunity
Faced with demoralized employees and widespread financial loss among staff, Foroughi shifted focus inward. The management team extended performance stock plans—typically only available to CEOs—across key personnel. They spoke openly to staff, acknowledging that many felt they’d lost significant personal gains (“you thought you had a house and now you don’t”), but promising that those who stuck it out for the comeback could share in massive upside. Foroughi instilled an “us versus the world” mentality: as investors abandoned the company, Applovin would bet on itself and rely on collective effort for recovery. This approach helped retain top talent and align their interests with the recovery, turning crisis into the nucleus of a focused, motivated culture.
Strategic Silence on Investor Relations During Crash Shifted to Aggressive Communication Amid Recovery
Post-Launch of Deep Learning Model, CEO Foroughi's Investor Outreach Boos ...
Here’s what you’ll find in our full summary
Stock Recovery Strategy: Buybacks, Culture, Market Perception
Additional Materials
Clarifications
- EBITDA stands for Earnings Before Interest, Taxes, Depreciation, and Amortization. It measures a company's operating performance by showing profit from core business activities, excluding financial and accounting decisions. Investors use EBITDA to compare profitability across companies and industries by focusing on operational efficiency. It helps assess cash flow potential and the ability to service debt.
- Market capitalization, or market cap, is the total value of a company's outstanding shares of stock. It is calculated by multiplying the current stock price by the total number of shares available. Market cap reflects the market's perception of a company's overall value and size. It helps investors compare companies and assess investment risk.
- An IPO is when a private company offers its shares to the public for the first time to raise capital. It involves regulatory approval, setting an initial share price, and listing on a stock exchange. This process allows public investors to buy ownership stakes and provides liquidity for early investors. IPOs often attract media attention and can significantly impact a company’s valuation.
- During the COVID-19 pandemic, many companies rushed to go public, creating an oversupply of new stocks. This flood diluted investor attention and capital, making it harder for individual IPOs to attract quality, long-term investors. Many investors bought without thorough research, leading to volatile and often inflated valuations. When market sentiment shifted, these stocks faced sharp declines as demand dropped and insiders sold shares.
- Private equity firms and insiders often hold large shares before a company goes public. When they sell these shares after an IPO, it increases the number of shares available on the market. This sudden increase in supply can lower the stock price if demand does not keep up. Their selling can signal a lack of confidence, further discouraging new investors.
- A share repurchase program is when a company buys back its own shares from the market, reducing the total number of shares outstanding. This often increases the value of remaining shares by boosting earnings per share (EPS) and signaling confidence in the company’s prospects. It can also improve financial ratios and return excess cash to shareholders without paying dividends. Buybacks may lead to higher stock prices as demand for shares rises and supply decreases.
- Regression models use simple mathematical equations to predict outcomes based on input variables, often assuming linear relationships. Deep learning models use layered neural networks to automatically learn complex patterns and interactions from large datasets. In advertising, deep learning can better capture user behavior and context, improving ad targeting and performance. This leads to more accurate predictions and optimized ad delivery compared to traditional regression approaches.
- The transition from ML 1.0 to ML 2.0 signifies moving from basic regression models to advanced deep learning algorithms, which can analyze complex patterns in data more effectively. This upgrade improves prediction accuracy and decision-making in advertising technology, leading to better targeting and higher revenue. Enhanced machine learning models enable the platform to optimize ad placements dynamically, increasing user engagement and advertiser ROI. Consequently, this technological leap drives stronger business growth and competitive advantage.
- Investor relations shape how investors perceive a company's value and future prospects. Clear, consistent communication builds trust and reduces uncertainty, encouraging investment. Lack of communication can lead to misinformation, fear, and undervaluation. Positive engagement attracts institutional investors, boosting demand and stock price.
- Performance-based equity incentives are stock or options granted to employees that vest only if specific company or individual p ...
Counterarguments
- While the buyback program is credited with creating shareholder value at peak prices, buybacks primarily benefit remaining shareholders only if the company’s fundamentals and future prospects justify higher valuations; otherwise, they can be seen as financial engineering rather than true value creation.
- The attribution of the stock collapse solely to market oversaturation and poor investor quality overlooks the possibility that Applovin’s business model, sector risks, or growth prospects may have also contributed to investor skepticism and the valuation decline.
- The narrative emphasizes management’s foresight and discipline, but does not address whether alternative uses of cash (such as investment in R&D, acquisitions, or debt reduction) might have provided greater long-term value than buybacks.
- Extending performance-based equity incentives to more employees can help retention, but it also dilutes existing shareholders and may not fully compensate for the magnitude of financial losses experienced by staff during the downturn.
- The decision to halt investor relations during the downturn may have prolonged negative sentiment and delayed recovery, as proactive communicat ...
Get access to the context and additional materials
Privacy Regulations and Their Impact on Advertising Business
Apple's Privacy Rules and EU Regulations Limit Ad Targeting Precision, Requiring Broader User Cohorts and Less Granular Relevance
Adam Foroughi explains that previous advertising models on iOS allowed for precise user targeting, enabling advertisers to deliver highly relevant ads to individuals. With Apple's recent privacy restrictions and EU regulatory changes, advertisers are now forced to target grouped audiences or broader user cohorts, which significantly degrades ad relevance and increases wasted ad impressions. Foroughi notes the irony that users, despite privacy concerns, frequently complain after these changes that the ads they see are less relevant and appear more like spam than before. Research supports that consumers generally prefer personalized ads that help them discover products and provide value—such as relevant in-game ads during mobile gaming—over irrelevant or generic advertisements.
Regulatory Clarity Helps Companies Align Systems With Requirements, Letting Algorithms Optimize Within Defined Boundaries Instead of Ambiguous Environments
Foroughi stresses the importance of clear and well-defined privacy regulations. He notes that clear rules enable technology companies to adjust their systems accordingly and avoid operating in uncertainty. Since the implementation of explicit privacy regulations, there has been a stabilizing effect: technology companies have adapted, deploying compliant systems and leaning on advanced tools such as deep learning networks. These algorithms are now adept at adapting to privacy-constrained environments, using cohort-based targeting to ...
Here’s what you’ll find in our full summary
Privacy Regulations and Their Impact on Advertising Business
Additional Materials
Clarifications
- Cohort-based targeting groups users by shared characteristics or behaviors instead of identifying individuals. It protects privacy by anonymizing data within these groups, preventing tracking of single users. Individual user targeting uses personal data to deliver ads tailored to a specific person’s preferences. Cohort targeting reduces precision but complies with stricter privacy rules.
- Apple's privacy restrictions primarily involve the App Tracking Transparency (ATT) framework, which requires apps to get user permission before tracking their data across other apps and websites. This limits advertisers' ability to collect detailed user data for precise ad targeting. Without tracking, advertisers must rely on aggregated data or broader audience segments, reducing ad personalization. Consequently, ad effectiveness and relevance decline as targeting becomes less specific.
- The EU regulations refer primarily to the General Data Protection Regulation (GDPR) and the ePrivacy Directive, which restrict how personal data can be collected and used for advertising. These laws require advertisers to obtain explicit user consent before tracking or profiling individuals for targeted ads. They also limit the use of cookies and other tracking technologies, pushing advertisers to use aggregated data instead of individual-level data. This reduces the precision of ad targeting and increases reliance on broader audience segments.
- Broader user cohorts group many individuals with diverse interests together, reducing the ability to tailor ads to specific preferences. This lack of precision means ads are less likely to match individual needs, lowering engagement. As a result, more users see irrelevant ads, increasing wasted impressions. Precise targeting minimizes waste by focusing ads on users most likely to respond.
- Deep learning networks analyze large datasets to identify patterns and predict user behavior without needing individual-level data. They enable advertisers to optimize ad targeting by grouping users into cohorts with similar interests. These models adapt to privacy constraints by focusing on aggregated data rather than personal identifiers. This approach maintains ad effectiveness while respecting user privacy.
- Mobile gaming ads often use in-app placements that integrate seamlessly with gameplay, enhancing user engagement. They rely heavily on real-time bidding and user behavior data to optimize ad delivery within the game environment. Unlike many other ads, they can leverage interactive formats like rewarded videos, which incentivize users to watch ads for in-game benefits. This interactivity and context-specific targeting help maintain high ad effectiveness despite privacy restrictions.
- An "84% margin" refers to the profit percentage a company retains from its advertising re ...
Counterarguments
- While some users complain about less relevant ads, many others value increased privacy and reduced tracking, even at the cost of less personalized advertising.
- The claim that consumers generally prefer personalized ads is context-dependent; surveys and studies have shown significant segments of users are uncomfortable with data collection and targeted advertising.
- Broader cohort-based targeting can reduce the risk of discriminatory or manipulative ad practices that were possible with hyper-targeted ads.
- Privacy regulations are designed to protect fundamental rights and prevent misuse of personal data, which can outweigh the benefits of ad relevance for many stakeholders.
- High profit margins in mobile gaming ads may reflect market dynamics or cost structures unrelated to privacy regulations, and do not necessarily indicate that privacy rules have no negative impact on smaller advertisers or new entrants.
- The resilience of the mobile gaming ad sector does not guarantee similar outcomes in ot ...
Get access to the context and additional materials
Future of Advertising in Ai World With Agents and New Interfaces
Ai Agents Streamlining Purchases and Enhancing Consumer Commerce
Adam Foroughi describes how AI agents will increasingly streamline the purchasing process. He gives the example of putting his supplement subscription into an agent that could automatically optimize and deliver it monthly. This automation allows consumers to delegate routine, repeat purchases to autonomous agents that make optimal decisions on their behalf.
However, Foroughi points out that most mass market consumers are different from early adopters who embrace the latest agent-based technologies. The typical consumer values the experience of shopping itself—they enjoy browsing, comparing goods, and engaging in the transaction process. Even if an agent could inform them later that it could have saved them 20% on a $50 purchase, most consumers value the emotional satisfaction and [restricted term] reward from shopping, discovery, and comparison more than the financial optimization provided by agents for lower-value purchases.
Discovery-Based Ads Create More Economic Growth Than Search-Based Ads By Unveiling Unknown Products
Foroughi distinguishes between two models of advertising. One is bottom-of-funnel, based on search or use of language models, where the consumer already knows what they want—like searching for dress shoes on Google or using a large language model to complete a purchase. These search or assistant-driven channels only channel existing demand and do not materially expand consumer spending or overall economic activity. The transaction was bound to happen with or without the ad, serving to guide but not fundamentally change the outcome.
In contrast, discovery-based advertising operates by presenting consumers with products they did not know existed or did not know they wanted. Such ads, as seen in the business model of platforms like Meta (Facebook), are not based on previously declared intent. Instead, they create “discovery moments”—fun, engaging experiences that drive [restricted term] hits and build anticipation for receiving new purchases. Discovery-based ads expand consumer spending and create economic growth by generating new purchases and trials that would not otherwise occur.
Foroughi emphasizes that this type of advertising not only provides emotional satisfaction to consumers—with the thrill of discovery, comparison, and the shopping journey—but also powers economic expansion by unveiling novel products and opportunities.
Consumer Interfaces Like Ar Glasses and New Computing Paradigms Will Create New Advertising Contexts, Though Relevant Product ...
Here’s what you’ll find in our full summary
Future of Advertising in Ai World With Agents and New Interfaces
Additional Materials
Clarifications
- AI agents are software programs that use artificial intelligence to perform tasks autonomously on behalf of users. They can analyze user preferences, past behavior, and external data to make purchasing decisions without manual input. These agents continuously learn and optimize choices to save time and money for consumers. They act like personal assistants that handle routine shopping tasks automatically.
- Search-based advertising targets consumers who already have a specific product or need in mind, showing ads that match their explicit intent. Discovery-based advertising reaches consumers without prior intent, introducing them to new or unexpected products. This approach relies on engaging content that sparks curiosity and emotional interest, often leading to impulse or exploratory purchases. Discovery ads expand market demand by creating new desires rather than just fulfilling existing ones.
- "Bottom-of-funnel" advertising targets consumers who are close to making a purchase decision. It focuses on converting existing interest into a sale rather than generating new interest. This stage follows awareness and consideration phases in the marketing funnel. Ads here often include calls to action like "buy now" or "sign up today."
- Platforms like Meta use algorithms to analyze user behavior, interests, and social connections to show ads for products users have not explicitly searched for. These ads appear in users' feeds as personalized suggestions, often blending seamlessly with regular content. The goal is to spark curiosity and encourage impulse purchases by introducing novel items. This method leverages vast data to create highly targeted, engaging discovery experiences.
- [restricted term] is a neurotransmitter in the brain that plays a key role in reward and pleasure. When consumers discover new products or enjoy shopping, [restricted term] is released, creating feelings of satisfaction and motivation. Advertisers leverage this by designing ads and experiences that trigger [restricted term], encouraging engagement and repeat behavior. This neurological response helps explain why consumers value the emotional experience of shopping beyond just saving money.
- New computing paradigms refer to emerging ways people interact with technology beyond traditional screens, such as voice assistants, virtual reality, or brain-computer interfaces. Consumer interfaces are the devices or methods through which users access digital content, like smartphones, smartwatches, or AR (augmented reality) glasses. AR glasses overlay digital information onto the real world, allowing users to see and interact with virtual objects while ...
Counterarguments
- The assertion that most mass market consumers prefer the emotional satisfaction of shopping over financial optimization may not hold true for all demographics, especially in times of economic hardship or among price-sensitive consumers who prioritize savings.
- The claim that search-based advertising does not significantly increase consumer spending overlooks the fact that effective search ads can upsell, cross-sell, or introduce higher-value alternatives, potentially increasing transaction sizes and overall spending.
- Discovery-based advertising can also lead to overconsumption, impulse buying, and increased consumer debt, raising questions about its net positive impact on economic well-being.
- The idea that discovery-based ads always provide emotional satisfaction ignores the growing consumer fatigue and annoyance with excessive or irrelevant advertising, which can diminish user experience.
- The persistence of discovery-based advertising’s value may be challenged by increasing consumer adoption of ad blockers, privacy tools, and regulations that limit data-driven targeting.
- The assumption that new interfaces like AR glasses will simply ...
Get access to the context and additional materials
1-Page Summary- Applovin's Model & $50b Gaming Ad Market Opportunity
- Technology Innovation: Deep Learning and Competitive Advantage
- Stock Recovery Strategy: Buybacks, Culture, Market Perception
- Privacy Regulations and Their Impact on Advertising Business
- Future of Advertising in Ai World With Agents and New Interfaces
Create Summaries for anything on the web
Download the Shortform Chrome extension for your browser
