Adam Foroughi, Applovin CEO: Surviving a 92% Drawdown, Ads as ML 1.0 & the $50B Game Ad Market

Business summary (AppLovin)

• What AppLovin does: An advertising platform that helps mobile game developers monetize via ads/recommendations inside games. It drives users from one game experience to the next and increasingly enables shopper-intent / e-commerce-like outcomes using ML.
• Market framing: Mobile casual gaming has scaled dramatically:
  • 1B+ people/day play mobile casual games.
  • Company disclosed $11B/year ad spend on its platform (nearly two years prior to the interview), growing to an implied ~$20B today.
  • Estimated total annual ad spend in the mobile gaming ecosystem: ~$50B/year.

Strategy & operating model

• Core GTM / product strategy: Build performance advertising by creating “discovery” moments—showing users something relevant enough to trigger fast transaction/engagement.
• ML playbook evolution (explicit “waves”):
  • ML 1.0: Regression-based models aimed at driving users to the next game experience (intent inferred from ad/reward engagement).
  • ML 2.0: Transition to deep learning and algorithmic optimization, shifting toward shopper-like behavior prediction (bigger economic opportunity beyond “next game”).
• Performance-based economics:
  • Sells advertising to advertisers with revenue tied to scaling (advertiser performance improves as the model improves).
  • Described as transaction-level economics: advertiser “covers the cost of the consumer immediately,” enabling scalable margins.

Key frameworks / playbooks mentioned

• “Us against the world” culture playbook during public-market collapse (internal morale + aggressive response).
• Performance stock plan (applied broadly, not only to the CEO) to align leadership with recovery upside.
• Learning-velocity operating philosophy: stay lean, be focused, and move faster than incumbents (explicitly framed as how to compete with Meta/Google).
• Moat via ML + differentiated data: advantage comes from complex models and data access, compounding at scale.

Metrics & KPIs (explicit numbers)

Market / demand

• Mobile casual games reach: >1B users/day
• Company-reported platform ad spend: $11B/year (disclosed nearly two years before the interview)
• Implied current platform ad spend: ~$20B
• Total mobile gaming ad market estimate: ~$50B/year
• Company growth: ~60% YoY (roughly, “since then”)

Public market / financial outcomes

• IPO timeline: went public April 2021 (co-IPO)
• Market cap at IPO: ~$28B (initial post-IPO figure)
• EBITDA at IPO: ~$600M
• Peak market cap: about $40B
• 2022 market cap trough: about $3.8B (stock “down literally every day” that year)
• 2022 EBITDA: ~$1B
• Drawdown mentioned: ~92% peak-to-trough
• Post-recovery stock movement (illustrative):
  • Stock recovered to ~$80 in roughly Sep 2023 timeframe.
  • Then ~within “a week” after an investor-facing period: ~$80 → $150
  • Market cap: ~$28B → $55B in that window (as described)

Capital allocation / shareholder return

• Aggressive buyback program:
  • Total repurchased: ~$6B in stock
  • Shares retired: ~20–25% of shares outstanding
  • Buyback peak value (for that $6B at one moment): over $50B

Unit economics / profitability

• EBITDA margin cited: ~84% (“Ebidon margins… number one in the market”)

Case examples / concrete actions

• Responding to extreme public-market skepticism (operational + comms):
  • CEO stopped talking to investors (“waste of time”) during the collapse.
  • Redirected strategy into cash generation + buybacks to become “our best investor.”
  • Later resumed investor engagement when performance improved and buyback capacity reduced.
• Data strategy via acquisitions (then divestment):
  • Bought game studios initially as a data play to train early deep learning models.
  • After model success and when third parties wanted in, AppLovin divested studios because game studios generally don’t want to share data externally.
• Investor/market psychology management:
  • Acknowledged leadership hardship: team members received calls asking about wellbeing.
  • Used compensation alignment (performance stock plan) + “we’re against the world” narrative to retain execution focus.

Leadership & organizational tactics

• Culture under existential stress: treat market hostility as an operating constraint; maintain execution focus and belief in the product.
• Alignment: broadened performance equity across key roles to mirror CEO-level incentive logic.
• Talent model: distributed engineering footprint:
  • Engineering offices in Palo Alto, Beijing, Singapore.
• Management philosophy: hire/empower subject matter experts; leaders aim to be among the “dumber” people in rooms to sustain learning.

Competitive advantage logic (how they “win” vs Meta/Google)

• Lean + fast iteration: compete with giants by focusing tightly on one domain (mobile gaming → transactions/intent).
• ML + differentiated data creates a compounding moat: model complexity and adoption at scale makes imitation difficult.
• Performance economics reinforce adoption: advertisers scale because outcomes improve with model performance.

Advertising + AI/agents (high-level, execution-focused)

• Bottom-funnel vs discovery advertising:
  • Search/LM closes the loop when intent is known—this resembles Google search competition.
  • AppLovin’s differentiation is when intent is unknown: ads act as recommendations that create discovery, enabling incremental economic value.
• Agentic commerce: expects some shoppers will use agents, but AppLovin believes its core audience prefers window-shopping + transaction experience, not fully automated agent-driven optimization (agents don’t replace the dopamine/experience of discovery for mainstream users).

Practical recommendations implied by the discussion

• Build a compounding ML loop where better models directly improve advertiser ROI (tie product quality to customer economics).
• Use performance-based incentive structures across leadership to preserve motivation during external shocks.
• Treat data strategy as a product requirement, not a byproduct: if training data is scarce due to partner reluctance, acquire/control it—then divest when the core model is established.
• Stay lean to move faster than incumbents in tightly defined markets.

Presenters / sources

• Adam Foroughi (Adam Foroughi / “Adam”): CEO, AppLovin
• Interviewers / other speakers: Multiple hosts are referenced in the subtitles (“Adam, thanks…”, “Please welcome…”, “Jason asked you…”), but names are not clearly provided in the transcript excerpt.