How Do Companies Make Money?

The Core Revenue Models That Power the Global Economy

Understanding how a company makes money is the first step in analyzing its long-term viability and competitive moat. While there are companies in the CorpDigest database, most fall into one of several primary revenue mechanics. These models determine not just the company's income, but also its relationship with customers and its vulnerability to market shifts.

Every major company uses one or more of these fundamental revenue mechanics. Often, the most successful firms layer multiple models — for example, selling a high-margin hardware product (Apple) while simultaneously collecting recurring service fees (iCloud) and taking a cut of third-party sales (App Store).

In this guide, we analyze real-world examples of each model, drawn from official SEC filings and annual reports. By examining the actual mechanics of money movement, we can move beyond marketing summaries and into the structural reality of the global economy in 2026.

Subscription Models: Predictable Revenue Through Recurring Payments

The subscription model generates revenue by charging customers a recurring fee — monthly, quarterly, or annually — in exchange for continued access to a product or service. What makes this model powerful is its predictability: once a customer subscribes, the company can forecast future revenue with high confidence, plan investments accordingly, and build long-term relationships rather than chasing one-time transactions.

The key metrics investors watch in subscription businesses are Monthly Recurring Revenue (MRR), Annual Recurring Revenue (ARR), churn rate (the percentage of subscribers who cancel each period), and Customer Lifetime Value (LTV). A healthy subscription business maintains churn below 5% monthly for consumer products and below 1% monthly for enterprise software. The ratio of LTV to Customer Acquisition Cost (CAC) should exceed 3:1 for the model to be sustainable.

What makes subscription revenue defensible is habit formation and switching costs. Once a user builds playlists on Spotify, curates a Netflix watch history, or stores years of files in a cloud service, the cost of leaving becomes psychological as much as financial. The product becomes embedded in daily routines, creating retention that persists even when competitors offer lower prices.

Real Example: Netflix

Netflix reported approximately $33.7 billion in revenue for fiscal year 2023, virtually all of it from subscription fees across 260 million paid memberships globally. The company operates on a tiered pricing structure ranging from ad-supported plans at $6.99 per month to premium plans at $22.99 per month. Netflix's average revenue per membership (ARM) sits around $11.70 globally, though this varies significantly by region — North American subscribers generate roughly $17 per month while Asia-Pacific members average closer to $8. The company's content spending exceeds $13 billion annually, which functions as both its primary cost and its primary moat: no competitor can easily replicate a library built over fifteen years of continuous investment.

Real Example: Spotify

Spotify generated EUR 13.2 billion in revenue in 2023, with approximately 87% coming from its 226 million Premium subscribers and the remainder from advertising on its free tier. The Premium subscription costs between $10.99 and $16.99 per month depending on the plan. Spotify's gross margin on Premium subscriptions is approximately 29%, constrained by music licensing costs that consume roughly 70% of subscription revenue. The company's churn rate for Premium subscribers is estimated at 3.9% monthly — low by consumer subscription standards but higher than enterprise SaaS products. Spotify's defensibility comes from personalized playlists, podcast exclusives, and the social graph of shared listening activity that users accumulate over years.

Marketplace and Platform Models: Earning a Cut of Every Transaction

Marketplace businesses connect buyers and sellers, taking a percentage of each transaction as revenue without owning the underlying inventory. This model creates exceptional operating leverage because revenue scales with transaction volume rather than headcount or physical assets. The platform provides trust infrastructure — payments processing, dispute resolution, reviews, and discovery — that neither party could efficiently build alone.

The critical metrics for marketplace businesses are Gross Merchandise Volume (GMV), which represents the total value of transactions flowing through the platform; take rate, which is the percentage of GMV the platform captures as revenue; and network effects strength, measured by how much each new participant increases value for existing participants. Investors distinguish carefully between GMV and actual revenue — a marketplace reporting $50 billion in GMV with a 15% take rate generates $7.5 billion in revenue, not $50 billion.

Defensibility in marketplace models comes from network effects: more sellers attract more buyers, which attracts more sellers, creating a flywheel that becomes increasingly difficult for competitors to replicate. The risk is disintermediation — buyers and sellers connecting directly to avoid platform fees once they establish trust through the initial marketplace interaction.

Real Example: Airbnb

Airbnb reported $9.9 billion in revenue for 2023 on approximately $73 billion in Gross Booking Value (GBV). This implies an effective take rate of roughly 13.5%, split between a service fee charged to guests (typically 14% of the booking subtotal) and a host service fee (typically 3%). The company facilitated over 448 million nights and experiences booked during the year. Airbnb's operating margin reached 16% in 2023, demonstrating the operating leverage inherent in asset-light marketplace models — the company owns zero real estate yet earns billions from housing transactions. Its moat is the accumulated trust layer: 7 million active listings, millions of verified reviews, and a brand synonymous with alternative accommodation that took fifteen years to build.

Real Example: Amazon Marketplace

Amazon's third-party seller services generated $140.1 billion in 2023, representing roughly 24% of Amazon's total net sales. Over 60% of units sold on Amazon come from third-party sellers rather than Amazon's own retail inventory. Amazon charges sellers a referral fee (typically 8-15% of the sale price depending on category), plus Fulfillment by Amazon (FBA) fees for storage and shipping. The effective take rate on third-party sales is estimated between 35-50% when all fees are included. Amazon's marketplace defensibility stems from its 200 million Prime members who default to Amazon for purchases, creating demand-side scale that no competing marketplace can match.

Advertising Models: Monetizing Attention at Scale

Advertising-driven businesses offer their core product for free to users and generate revenue by selling access to that user base to third-party advertisers. The product is attention; the customer is the advertiser, not the user. This model produces extraordinarily high gross margins (often 70-85%) because the marginal cost of serving one additional ad is near zero, but it creates dependency on advertiser spending cycles and regulatory tolerance of data collection.

Key metrics in advertising businesses include Cost Per Mille (CPM) — the price per thousand ad impressions; Cost Per Click (CPC) — the price each time a user clicks an ad; Average Revenue Per User (ARPU); and ad load, which measures how many ads are shown per session. Investors also track the ratio of user engagement time to ad revenue, which reveals monetization efficiency. A platform with high engagement but low ARPU has pricing power it has not yet exercised.

The defensibility of advertising models rests on data moats — proprietary datasets about user behavior that enable precise ad targeting. A platform that knows what you search for, who you communicate with, and what content you engage with can deliver ads so relevant that advertisers pay premium prices. This creates a compounding advantage: more users generate more data, which improves targeting, which attracts more advertiser spending, which funds better products, which attracts more users.

Real Example: Alphabet (Google)

Alphabet generated $307.4 billion in total revenue for 2023, with $237.9 billion (77%) coming from Google advertising across Search, YouTube, and the Google Network. Google Search alone produced $175 billion — the single largest advertising revenue stream in history. The company serves over 8.5 billion searches per day, each one an explicit signal of user intent that advertisers bid on through the Google Ads auction system. Google's average CPC varies by industry but typically ranges from $1 to $4 for Search ads. The company's data moat is arguably the deepest in technology: twenty-five years of search history, Gmail content signals, Maps location data, YouTube viewing patterns, and Android device telemetry create a targeting capability no competitor can replicate.

Real Example: Meta Platforms

Meta Platforms reported $134.9 billion in revenue for 2023, with approximately 97% derived from advertising across Facebook, Instagram, Messenger, and WhatsApp. The company reaches 3.98 billion monthly active people across its family of apps. Meta's ARPU varies dramatically by geography: North American users generate approximately $68 per quarter while Asia-Pacific users generate roughly $5.50. Meta's advertising engine processes over 12 million ad variations simultaneously, using machine learning to match ads to users based on behavioral signals. The company's moat is its social graph — the map of human relationships that makes its targeting uniquely effective for brand awareness and direct-response advertising alike.

Licensing and Intellectual Property Models: Monetizing Software and Patents

Licensing businesses generate revenue by granting third parties the right to use proprietary technology, software, or intellectual property under contractual terms. This model produces exceptionally high margins because the marginal cost of licensing an additional copy of software is effectively zero — the development cost is fixed and amortized across all customers. The transition from perpetual licenses (one-time payment) to subscription-based SaaS (recurring payment) has been the defining structural shift in enterprise technology over the past decade.

Investors evaluate licensing businesses through metrics like Annual Contract Value (ACV), Net Revenue Retention (NRR) — which measures whether existing customers spend more or less over time — and the mix between on-premise licenses and cloud subscriptions. A company with NRR above 120% is growing its revenue base even without acquiring new customers, which is the hallmark of a best-in-class enterprise software business.

Defensibility in licensing models comes from switching costs and ecosystem lock-in. Once an enterprise deploys Oracle databases across hundreds of applications, or builds its entire workflow on Microsoft Office, the cost of migrating to a competitor is measured in years and millions of dollars. This creates pricing power that persists even when technically superior alternatives exist.

Real Example: Microsoft

Microsoft reported $211.9 billion in revenue for fiscal year 2023, with its Intelligent Cloud segment (including Azure) generating $96.8 billion and Productivity and Business Processes (including Office 365 and LinkedIn) generating $69.3 billion. The company's transition from perpetual Windows and Office licenses to Microsoft 365 subscriptions and Azure cloud consumption has transformed its revenue quality — over 50% of revenue is now recurring. Microsoft's commercial remaining performance obligation (contracted future revenue) exceeded $213 billion, providing extraordinary revenue visibility. The company's moat is ecosystem breadth: Windows, Office, Azure, Teams, GitHub, LinkedIn, and Xbox create an interconnected product surface that competitors can challenge in individual categories but cannot replicate as a system.

Real Example: Oracle

Oracle generated $52.9 billion in revenue for fiscal year 2024, with cloud services and license support accounting for $39.4 billion (74% of total revenue). Oracle's database technology runs mission-critical workloads for the majority of Fortune 500 companies, creating switching costs so high that customers routinely renew contracts despite aggressive pricing. The company's cloud infrastructure (OCI) revenue grew 49% year-over-year as Oracle transitions its installed base from on-premise licenses to cloud subscriptions. Oracle's remaining performance obligations exceeded $80 billion, reflecting multi-year enterprise commitments that provide revenue certainty regardless of macroeconomic conditions.

Direct Product Sales: Hardware Margins and Ecosystem Lock-In

Direct product sales generate revenue through one-time transactions where customers purchase physical goods or digital products outright. This is the oldest and most intuitive business model, but it carries structural challenges: revenue is lumpy (dependent on product launch cycles), margins are constrained by manufacturing and logistics costs, and customer relationships reset after each purchase. The most successful product companies mitigate these weaknesses by building ecosystems that generate recurring revenue alongside hardware sales.

Key metrics for product businesses include gross margin (revenue minus cost of goods sold), inventory turnover, average selling price (ASP), and services attach rate — the percentage of hardware customers who also subscribe to companion services. Investors increasingly value the services layer because it converts one-time buyers into recurring revenue streams with higher margins than the hardware itself.

Defensibility in product models comes from brand premium, ecosystem lock-in, and vertical integration. A company that controls both hardware and software (like Apple) can optimize the user experience in ways that component assemblers cannot, justifying premium pricing that produces margins far above industry averages.

Real Example: Apple

Apple reported $383.3 billion in revenue for fiscal year 2023, with iPhone generating $200.6 billion (52% of total), Services contributing $85.2 billion (22%), Mac at $29.4 billion, iPad at $28.3 billion, and Wearables at $39.8 billion. Apple's hardware gross margin is approximately 36%, while its Services segment (App Store, iCloud, Apple Music, Apple TV+, AppleCare) operates at roughly 71% gross margin. The company's installed base exceeds 2.2 billion active devices, each one a potential services customer. Apple's services attach rate continues to climb — the average iPhone user now pays for 2.1 Apple services, up from 1.4 three years ago. The ecosystem moat is formidable: once a customer owns an iPhone, Apple Watch, AirPods, and MacBook, switching to Android means replacing an entire integrated system, not just a single device.

B2B SaaS: Land-and-Expand Revenue Growth

Business-to-business Software as a Service (B2B SaaS) companies sell cloud-hosted software tools to other businesses on a subscription basis. What distinguishes B2B SaaS from consumer subscriptions is the land-and-expand motion: a sales team closes an initial contract with one department or use case, then systematically expands usage across the organization over subsequent years. This creates revenue growth from existing customers that can exceed new customer acquisition in mature SaaS businesses.

The defining metrics for B2B SaaS are Annual Recurring Revenue (ARR), Net Revenue Retention (NRR), and the Rule of 40 (growth rate plus profit margin should exceed 40%). NRR above 130% means the company grows its existing customer revenue by 30% annually before counting any new logos — a powerful compounding engine. Investors also track the CAC payback period (months to recover customer acquisition cost) and the magic number (net new ARR divided by sales and marketing spend).

B2B SaaS defensibility comes from workflow integration and organizational dependency. When a company's sales team runs entirely on Salesforce, its customer data, automation rules, reporting dashboards, and institutional knowledge all live within that platform. Switching CRMs means retraining hundreds of employees, migrating years of data, and rebuilding custom integrations — a project so disruptive that most companies simply renew their contracts regardless of price increases.

Real Example: Salesforce

Salesforce reported $34.9 billion in revenue for fiscal year 2024, growing 11% year-over-year. The company's remaining performance obligation (contracted future revenue) exceeded $56 billion, providing multi-year revenue visibility. Salesforce's Net Revenue Retention rate consistently exceeds 120%, meaning existing customers expand their spending by at least 20% annually through additional seats, modules, and platform upgrades. The company serves over 150,000 customers ranging from small businesses to enterprises like Toyota, Amazon Web Services, and the U.S. federal government. Salesforce's land-and-expand motion is visible in its product portfolio: a customer might start with Sales Cloud ($25-$300 per user per month), then add Service Cloud, Marketing Cloud, Commerce Cloud, and the Einstein AI platform — each expansion multiplying the annual contract value without requiring a new sales cycle.

Freemium: Converting Free Users Into Paying Customers

The freemium model offers a functional product at no cost to attract a large user base, then converts a percentage of those free users into paying customers through premium features, enhanced capabilities, or removal of limitations. This model combines the viral distribution advantages of free products with the revenue quality of subscriptions. The strategic challenge is calibrating the free tier — too generous and users never upgrade; too restrictive and they never adopt the product in the first place.

Key metrics for freemium businesses include the free-to-paid conversion rate (typically 2-5% for consumer products, 10-25% for B2B tools), the time-to-conversion (how long free users take to upgrade), and the viral coefficient (how many new users each existing user brings through organic sharing). Investors evaluate whether the free tier functions as an efficient acquisition channel — if the cost of serving free users is lower than traditional marketing spend per converted customer, the model is working.

Freemium defensibility relies on network effects within the free tier and the accumulation of user-generated data that makes the product more valuable over time. A free user who has built hundreds of playlists, connected with friends, and trained the recommendation algorithm has created switching costs for themselves — even though they have never paid a dollar.

Real Example: Spotify (Freemium Conversion)

Spotify operates one of the most studied freemium funnels in technology. Of its 602 million monthly active users in 2023, approximately 226 million (37.5%) were paying Premium subscribers — an exceptionally high conversion rate for a consumer freemium product. The free tier serves as both an acquisition channel and an advertising revenue stream, generating EUR 1.7 billion annually. Free users experience audio ads between songs, limited skips, and no offline playback — friction points carefully designed to demonstrate Premium's value without making the free product unusable. Spotify's data shows that the average free user converts to Premium within 7 months of registration, and once converted, maintains their subscription for an average of 4.2 years. The company's Discover Weekly and Release Radar algorithms improve with listening history, creating personalization that free users would lose by switching to a competitor — a subtle but effective retention mechanism that operates identically on both tiers.

Financial Services: Transaction Fees and Payment Volume

Financial services companies generate revenue from the movement, storage, lending, and insurance of money. Unlike technology companies that sell products or attention, financial services firms monetize the infrastructure layer of economic activity itself. Every time money moves — between a buyer and seller, between a borrower and lender, across borders, or through time via insurance — a financial services company earns a fee. This creates revenue streams that scale with economic activity rather than with any single product or customer relationship.

The critical metrics for payment networks and financial services companies include Total Payment Volume (TPV), net revenue per transaction, net interest margin (for lending businesses), and the number of transactions processed. Investors distinguish between payment networks (which earn fees on volume without taking credit risk) and banks (which earn interest by lending depositor funds). Payment networks like Visa and Mastercard operate at 60-70% operating margins because they bear no credit risk — they simply process transactions and collect fees from issuing and acquiring banks.

Defensibility in financial services comes from regulatory moats, network ubiquity, and trust. Building a payment network requires regulatory approval in every jurisdiction, integration with thousands of banks, acceptance at millions of merchants, and the trust of billions of consumers. These barriers compound over decades, making incumbent payment networks among the most durable businesses in the global economy.

Real Example: Visa

Visa reported $32.7 billion in net revenue for fiscal year 2023, processing $14.8 trillion in total payment volume across 4.3 billion Visa cards worldwide. Visa's effective take rate is approximately 0.22% of total payment volume — a fraction of a percent that, applied to trillions of dollars, produces enormous revenue. The company operates at a 67% operating margin because it functions as a technology network rather than a lender: Visa does not extend credit, hold deposits, or bear default risk. Its revenue comes from service fees (based on payment volume), data processing fees (per-transaction charges), and international transaction fees (cross-border surcharges). Visa's moat is acceptance ubiquity — the network is accepted at over 100 million merchant locations in 200+ countries, creating a standard that no new entrant can replicate without decades of infrastructure investment.

Real Example: PayPal

PayPal processed $1.53 trillion in Total Payment Volume during 2023, generating $29.8 billion in net revenue — an effective take rate of approximately 1.95%, significantly higher than card networks because PayPal provides both the network and the processing layer. The company serves 426 million active accounts across 200 markets. PayPal's transaction revenue comes from merchant fees (typically 2.99% plus $0.49 per transaction for standard processing) and currency conversion spreads on cross-border payments. The company's Venmo subsidiary processed $270 billion in payment volume, monetized through merchant acceptance fees and instant transfer charges. PayPal's defensibility stems from its two-sided network: merchants accept PayPal because consumers have accounts, and consumers maintain accounts because merchants accept PayPal — a classic network effect that strengthens with each new participant on either side.

How to Evaluate Any Business Model

Regardless of which revenue model a company employs, analysts evaluate business quality through a consistent framework. First, revenue durability: how likely is the revenue to persist next year without additional sales effort? Subscription and licensing models score highest here; advertising and product sales score lowest. Second, margin structure: what percentage of each revenue dollar converts to profit after all costs? Software businesses typically operate at 70-80% gross margins while hardware businesses operate at 30-40%. Third, scalability: can the company grow revenue without proportionally growing costs? Platform and marketplace models scale most efficiently because each additional transaction requires minimal incremental infrastructure.

The most valuable companies in the world typically combine multiple revenue models in ways that reinforce each other. Apple sells hardware (product model) that locks users into services (subscription model) distributed through a marketplace (platform model). Google gives away search and email (freemium model) to collect data that powers advertising (attention model) while selling cloud infrastructure (B2B SaaS model). Microsoft licenses software (licensing model) delivered as subscriptions (SaaS model) to enterprises that expand usage over time (land-and-expand model). Understanding these layered architectures is the key to reading corporate financial statements with genuine comprehension rather than surface-level familiarity.