How Facebook built and sustains its $100 billion advertising business

Executive overview

Facebook is a media network that aggregates attention at billion-person scale and sells it to advertisers with precision targeting unavailable anywhere else. Its competitive flywheel is self-reinforcing: more users generate more data, better data enables better targeting, better targeting wins more ad budget share, and larger budgets fund more data infrastructure.

The company's 85% share of social network ad spend has been stable for five years. Its deepest structural risk is not regulation or user backlash — it is being bypassed at the platform level, as nearly happened during the desktop-to-mobile transition.

The core insight: Facebook is not a social network that happens to run ads — it is a targeted advertising platform that uses social identity as its data substrate.

The advertising flywheel and competitive moat

• 10 million active advertisers spending ~$11,000 each per year; ~1 million new advertisers added annually
• 85% share of social network ad spend, stable for at least five years
• Facebook is the third-largest CapEx spender in the S&P 500 (~$21B/year); nearest social competitor Twitter spends ~$1B — a 20x infrastructure gap
• Self-serve ad platform is the most developed in the industry; any advertiser can run a campaign without assistance
• The flywheel: scale → data → targeting quality → ad effectiveness → budget share → more data

How Facebook's ad auction works

• Facebook moved from a second-price auction (like Google) to a VCG auction in ~2013
• VCG takes every bidder's willingness to pay and allocates inventory in the optimally efficient outcome for the marketplace
• Rising CPMs are a positive signal: they indicate advertisers are finding value and competing harder for impressions
• Yield — matching the right ad to the right person at the right time — is the lever Facebook continuously optimises
• New inventory formats (Stories, Reels) expand supply; better creative and algorithm improvements drive yield higher

What advertisers actually do on the platform

• Advertisers set a budget and a target goal (e.g. customer acquisition at $300 per mattress sold)
• Targeting options: interest signals, lookalike audiences built from existing customer lists, or broad targeting
• Facebook's system takes the goal and optimises delivery toward it — not just impressions
• Winning ad creative is highly concentrated: in early-stage DTC brands, one creative often drives 60% of scale
• When a combination works, advertisers scale spend; rising demand on fixed-ish supply pushes CPMs up across the platform

What data Facebook actually uses

• Users are not individually tracked — they exist inside aggregate segments based on explicit signals (interests, demographics) and behavioural inference
• The Facebook pixel tracked users across third-party websites; the IDFA (Apple's device ID) connected behaviour across mobile apps
• Apple's ATT changes severed the IDFA link for apps that don't opt in; workarounds (email capture, first-party data) are already emerging
• The "Facebook is listening" belief is likely explained by proximity inference: two phones near each other for an extended period cause Facebook to cross-show each person's recent browsing history
• Facebook does not sell data; it uses data to place ads internally

The desktop-to-mobile transition and what it revealed

• At IPO in 2012, Facebook's ability to monetise mobile was the key investor concern
• Facebook's solution: acquire Instagram and WhatsApp to replicate its desktop footprint across the fragmented mobile app ecosystem
• Zuckerberg wrote internally (~2015) that the existential risk was being non-critical to whatever platform comes after mobile
• The company has invested heavily in Oculus and VR/AR as a hedge: even if it cannot monetise the next platform commercially, being infrastructure-critical is the goal
• COVID stress-tested platform flexibility: travel ad dollars evaporated but e-commerce dollars surged in; Facebook's revenue held while Google's was more exposed

The competitive landscape

• The "holy triangle" of digital advertising: scale, cost, and customer quality — only Google and Facebook deliver all three reliably
• Social network ad spend breakdown: Facebook >80%, LinkedIn + Twitter ~10%, Pinterest + Snap ~5%
• Streaming media (YouTube, TikTok, Netflix) is a separate category — they pay for content; Facebook does not
• Search (Google, Amazon) is intent-based and distinct from social
• DTC brands typically allocate 40–70% of paid media budget to Facebook; it is the default first channel

The ad ecosystem built on Facebook's rails

• Specialist agencies (performance marketing, creative production, influencer management)
• Analytics platforms for cross-channel attribution and mobile app tracking
• Creative studios focused on native-format UGC video (e.g. Tube Science model: pay only for used creative)
• Audience-sharing marketplaces where brands swap custom audiences
• The common thread: anything that improves matching the right message to the right person generates a business opportunity

Optionality: future revenue levers

• Instagram commerce: brands built on Instagram can now transact inside the app; reduces friction and keeps first-party data on-platform
• Facebook Marketplace: quietly displacing Craigslist with verified-identity listings; early experiments in auto and real estate
• Payments: Venmo and Cash App exist outside Facebook's ecosystem — closing that gap is described as the most obvious missing piece
• WhatsApp: 50 million business users, barely monetised; likely path is either a business services / super-app model or a funnel into the core ad platform
• VR/AR (Oculus): leading hardware; primary strategic rationale is platform infrastructure, not near-term revenue

Risks to the business

• User defection is the only reliable ad-revenue killer; it has not materialised despite years of bad press and one advertiser boycott
• Regulation that prohibits targeted advertising or data use in advertising would cause meaningful revenue impact — but politicians are also major platform users, creating a natural brake
• The next platform shift (post-mobile) is the deepest structural risk; the company is explicitly investing against it
• Trust deficit: Facebook Portal adoption is suppressed by brand distrust — the same dynamic may slow payments adoption