Free is not the absence of a business model. It is a business model — and in digital markets, it is often the most aggressive competitive weapon available. Chris Anderson's 2009 book "Free: The Future of a Radical Price" formalised the insight: when the marginal cost of producing and distributing a product approaches zero, giving it away becomes not just viable but strategically optimal. The economics are simple. A Google search costs Google a fraction of a cent to serve. An additional Gmail user costs almost nothing. A YouTube video streamed to one more viewer adds negligible marginal expense. When the cost of serving one more customer rounds to zero, the traditional pricing logic — charge above marginal cost to earn a margin — breaks down. The margin on zero is always zero. The question shifts from "what do we charge?" to "how do we monetise attention, data, and network effects instead?"
Google is the defining case. Larry Page and Sergey Brin gave away the world's best search engine, the world's most popular email service, the world's most used mapping application, and the world's largest video platform — all for free. Google did not charge because Google's customer was not the user. The user was the product. Google monetised the attention of two billion daily users through advertising — a business model that generated $307 billion in revenue in 2023. The search was free. The Gmail storage was free. The YouTube videos were free. The data exhaust from every search, email, and video view was not free. It was the most valuable commodity in the history of advertising, and Google collected it at industrial scale by making the front-end product irresistible and costless.
Craigslist demonstrated the destructive version of free. Newspaper classified advertising generated roughly $19.6 billion in annual revenue in 2000 in the United States alone. Craigslist made classified listings free — or nearly free — and obliterated an industry. By 2012, newspaper classified revenue had fallen to $4.6 billion, a 77% decline driven almost entirely by a website that Craig Newmark built without venture funding, without a sales team, and without charging for most of its listings. Craigslist did not build a better classified product. It built a free one. In a market where the incumbent's product was "pay $200 to list your used car," free was an insurmountable competitive advantage. Newspapers could not respond by matching the price without destroying their own revenue. They were trapped by their cost structure — a classic disruption pattern where the incumbent's business model prevents it from competing against a structurally cheaper alternative.
The catch — and it is a significant catch — is that "free" is never truly free. Someone always pays. In Google's model, advertisers pay with money. Users pay with attention and personal data. In Craigslist's model, the minimal staff and spartan design subsidise the free listings — Craig Newmark traded revenue maximisation for mission alignment. In Facebook's model, users pay with behavioural data that Facebook packages and sells to advertisers at scale. In the free-tier model of any SaaS company, the paying customers subsidise the free users through cross-subsidisation — the 5% who pay premium prices fund the infrastructure that serves the 95% who pay nothing. The economic law holds: there is no such thing as a free lunch. There is only a question of who picks up the tab, and whether they know they're doing it.
The strategic power of free lies not in the economics of the product but in the psychology of the customer. Zero is not just a low price — it is a categorically different price. Behavioural economist Dan Ariely demonstrated in experiments at MIT that the demand curve has a discontinuity at zero: dropping a price from two cents to one cent produces a modest increase in demand, but dropping from one cent to zero produces an explosion. Zero eliminates the mental transaction cost of deciding whether something is worth the price. When the price is zero, there is no evaluation to perform, no comparison to make, no risk to assess. The customer simply takes it. This psychological discontinuity means that a free product does not just compete with priced alternatives — it occupies a fundamentally different category in the customer's mind, one where adoption friction approaches zero and distribution velocity approaches maximum.
Section 2
How to See It
Free is operating as a business strategy whenever a company gives away a product or service that has obvious value — and monetises through a different mechanism than direct payment from the end user. The diagnostic: identify who is paying, what they're paying with, and why the company finds that payment more valuable than charging the user directly.
Advertising-Funded Products
You're seeing Free when a product with hundreds of millions of users charges nothing and reports billions in advertising revenue. Google Search, YouTube, Instagram, TikTok, Twitter/X, Snapchat — each gives away a consumer product that would command willingness-to-pay in a direct model and instead monetises the attention and data of the user base through advertising. The diagnostic: if the company's revenue comes overwhelmingly from advertisers rather than users, the users are not the customers — they are the inventory. The product is free because the product's purpose is to aggregate attention that can be sold. When Google's search quality declines, the strategic concern is not lost subscription revenue. It is lost advertising inventory. The free product is the pipeline. The advertising is the business.
Platform & Marketplace
You're seeing Free when a platform charges one side of a marketplace nothing to participate, building supply or demand volume that is then monetised on the other side. Craigslist charges nothing for most listings, monetising through paid job postings and apartment listings in select cities. Zillow provides free home valuations and property data to consumers, monetising through advertising sold to real estate agents who want access to those consumers. Robinhood offers commission-free stock trading, monetising through payment for order flow from market makers. In each case, the free side of the marketplace is not a cost centre — it is the asset. The free users are the product's distribution engine, and their aggregate activity creates the value that the paying side purchases.
Software & SaaS
You're seeing Free when a software company offers a permanently free tier with meaningful functionality — not a time-limited trial but a genuinely useful product at zero cost — and monetises through conversion to paid tiers or through enterprise sales. Slack's free tier supported unlimited users with limited message history. Zoom's free tier offered 40-minute meetings with up to 100 participants. Notion's free tier provided a complete workspace for individuals. Each free tier served as the primary distribution channel — users adopted without budget approval, embedded the product in their workflow, and then hit limitations that triggered organisational purchase decisions. The free tier is not charity. It is the most efficient customer acquisition channel ever devised, converting usage into dependency and dependency into revenue.
Content & Media
You're seeing Free when a content creator or media company distributes its core product for free and monetises through adjacent revenue streams. Podcasts are free — monetised through advertising, sponsorship, and premium subscriptions. Substack writers offer free newsletters — monetised through paid tiers. YouTube creators publish free videos — monetised through AdSense, sponsorships, and merchandise. The content is free because free content maximises distribution, and distribution is the scarce resource in attention-saturated markets. A paywalled article reaches thousands. A free article reaches millions. The revenue per reader is lower. The aggregate revenue — and the brand, audience, and optionality it builds — is higher.
Section 3
How to Use It
Decision filter
"Before setting a price above zero, ask: would giving this away for free generate more value through distribution, data, network effects, or conversion than direct payment would? If the answer is yes, free is not generosity — it is the optimal pricing strategy."
As a founder
Free is your most powerful distribution weapon — if you can afford the economics. The decision to offer a free tier should be driven by two calculations. First: does usage create switching costs? If users invest time, data, and workflow into your free product, conversion to paid becomes natural because the cost of leaving exceeds the cost of paying. Slack, Dropbox, and Notion all followed this pattern — the free tier created dependency that the paid tier monetised. Second: does free usage generate a network effect? If each free user makes the product more valuable for other users, free is not a cost — it is an investment in the network's density. LinkedIn's free tier made the professional network denser, which made the paid recruiter and sales tools more valuable. Free users were not freeloaders. They were the product.
The trap is offering free without a monetisation architecture. If free users do not convert, do not generate data you can monetise, and do not create network effects that benefit paying users, you have not built a free business model. You have built a charity.
As an investor
When evaluating free-tier businesses, the critical metric is not the conversion rate from free to paid. It is the cost of serving free users relative to the lifetime value they generate through conversion, data, network effects, or virality. Spotify's free tier costs money — it pays royalties on every stream, free or paid. The free tier is justified only because a meaningful percentage of free users convert to premium ($13.99/month) and because free users attract other users through social features and playlist sharing. If the conversion economics don't work — if the cost of serving free users exceeds the revenue they eventually generate — the free tier is a subsidy, not a strategy.
Watch for the "free-tier trap": companies that grow their free user base aggressively while conversion rates decline. The vanity metric is total users. The real metric is the ratio of monetised value to serving cost across the entire user base, free and paid combined.
As a decision-maker
Recognise when a competitor's free offering is a strategic weapon aimed at your revenue model. When Google launched Google Docs for free, it was not competing with Microsoft Office on features. It was making the category free — a move designed to commoditise Office's core value proposition and shift the competitive battleground to ecosystem lock-in, where Google had advantages through Gmail, Drive, and Chrome. Microsoft's response — Office 365 with collaboration features — was a defensive reframing, not a feature upgrade. The strategic question when a competitor goes free is not "how do we match the price?" It is "what value do we provide that survives the price going to zero?" If your entire value proposition is the product itself, free competitors will destroy you. If your value proposition includes integration, support, data, and ecosystem — elements that cannot be replicated for free — you survive.
Common misapplication: Assuming free works in every market. Free requires near-zero marginal costs to be sustainable. Digital products — software, content, search, social media — have marginal costs approaching zero, making free viable. Physical products do not. A restaurant cannot give away free meals and monetise through advertising (though some have tried). A hardware company cannot give away devices at zero margin indefinitely (though Amazon's Kindle strategy came close, subsidising hardware to monetise content). Free is a digital-native strategy. Attempting it in high-marginal-cost businesses is a path to bankruptcy.
Second misapplication: Treating free as permanent. Free is often a land-grab strategy that funds distribution during the growth phase and then transitions to paid models as the market matures. Uber subsidised rides below cost for years — effectively making urban transportation "free" relative to its true cost — to build network density. The subsidies ended once the network reached critical mass. Netflix offered a free trial month that drove millions of subscriptions. The trial disappeared once the brand no longer needed it. Free is a phase in many business models, not an endpoint.
Section 4
The Mechanism
Section 5
Founders & Leaders in Action
The leaders who weaponised free most effectively share a common insight: free is not a pricing decision. It is a distribution decision. They recognised that in markets where attention is scarce and switching costs are low, the company that removes the adoption barrier entirely — that reduces the price to zero — captures distribution at a velocity that no priced competitor can match. The monetisation comes later, through different mechanisms, but only if the distribution was achieved first.
Sergey BrinCo-founder & President of Technology, Google/Alphabet, 1998–2019
Brin and Larry Page built the most profitable free product in history. Google Search was free from day one — not because Brin and Page were indifferent to revenue, but because they understood that search was the attention aggregator and advertising was the monetisation layer. The insight was structural: charge for search, and you limit your user base to those willing to pay. Make search free, and you aggregate the world's attention — attention that is worth more per query to advertisers than any individual user would pay per search.
The economics validated the strategy at extraordinary scale. By 2023, Google processed approximately 8.5 billion searches per day. Charging even $0.01 per search would have generated $31 billion annually — but it would have reduced search volume by an order of magnitude, as free alternatives captured cost-sensitive users. Instead, Google monetised through advertising at approximately $0.04 per search — four times the hypothetical per-search price — because the zero price point maximised the audience that advertisers would pay to reach. Free was not a sacrifice. It was the mechanism that maximised total revenue by maximising the denominator (searches) rather than charging per unit.
Brin extended the logic across Google's product portfolio. Gmail launched in 2004 with 1GB of free storage — 500 times what Hotmail and Yahoo Mail offered — because email users generated search data, consumed advertising, and deepened their integration with Google's ecosystem. Google Maps launched free because location data enhanced advertising targeting. YouTube was acquired for $1.65 billion in 2006 and kept free because video attention was monetisable through pre-roll ads, display ads, and eventually YouTube Premium subscriptions. Each free product was not a cost centre but a data pipeline that fed the advertising engine. Brin's strategic clarity was absolute: the more Google gave away, the more data it collected, the more targeted its advertising became, and the more advertisers paid per impression. Free was the input. Advertising revenue was the output. The relationship was not inversely proportional — it was multiplicative.
Ek built Spotify on a bet that free music streaming — legal, high-quality, and ad-supported — could convert an industry of pirates into an industry of paying subscribers. The music industry in 2006 was being destroyed by free: Napster, LimeWire, and BitTorrent had made music effectively free through piracy, and the industry's revenue had fallen from $23.3 billion in 1999 to $15.9 billion by 2006. Ek's insight was that people were not stealing music because they were cheap. They were stealing it because the legal alternative — paying $0.99 per track on iTunes — was worse than the illegal alternative. Free, instant, unlimited music with no friction beat paid, per-track, DRM-restricted music on every dimension except legality. Ek's solution was to make the legal product competitive with piracy by matching its price: free.
Spotify's free tier offered unlimited streaming supported by advertisements — legal, high-quality, and zero-cost to the user. The free tier was not a trial period. It was a permanent product designed to compete with piracy by offering the same value proposition (free unlimited music) with better reliability and no legal risk. The monetisation came from two channels: advertising revenue on the free tier, and conversion to Spotify Premium, which offered ad-free listening, offline downloads, and higher audio quality for $9.99/month (now $11.99).
The conversion economics were the bet. Ek wagered that enough free users would convert to premium to make the model work — and that the remaining free users would generate enough advertising revenue to cover the royalty costs of serving them. By 2024, Spotify had 615 million total users, of which 239 million were premium subscribers — a conversion rate of approximately 39%. That conversion rate, combined with the advertising revenue from free users, made the economics work: premium subscribers generated the bulk of Spotify's $15.7 billion in 2024 revenue, while the free tier served as the world's most efficient customer acquisition funnel. The free tier's cost — royalties paid on ad-supported streams — was the customer acquisition cost. The premium tier's revenue was the payoff. Ek had turned the music industry's existential crisis (free piracy) into a business model (free with a premium upsell) by accepting that free was the new default and building monetisation around it rather than against it.
Section 6
Visual Explanation
The diagram maps the logic of free from top to bottom. When marginal costs approach zero (digital products), the optimal price is zero — because zero eliminates all adoption friction and maximises the three strategic assets that free products generate: distribution (user volume), data (behavioural information), and network effects (each user increasing value for others). The four models below show how free is monetised in practice: through advertising (selling the attention of free users), freemium (converting free users to paid), cross-subsidisation (giving away one product to sell a complement), and marketplace dynamics (charging one side to access the other). The closing line captures the essential insight: someone always pays. The strategic question is not whether to charge but who to charge and what currency to accept — money, attention, data, or network contribution.
Section 7
Connected Models
Free operates at the intersection of pricing strategy, distribution theory, and platform economics. Its power comes not from the absence of revenue but from the strategic decision to monetise through mechanisms other than direct payment — mechanisms that often generate more total value than charging the end user would. The connections below map how free enables network effects, how it relates to adjacent pricing strategies, and where its logic creates tension with attention economics.
Reinforces
Network Effects
Free is the most efficient fuel for network effects. A social network that charges $5/month will never reach the density where network effects become self-sustaining — the price barrier filters out too many potential users. The same network offered for free reaches critical mass faster because zero removes the adoption barrier entirely. Facebook, LinkedIn, WhatsApp, and Twitter all used free to build networks so dense that the network effect itself became the competitive moat. Once a billion people are on the platform, no paid competitor can match the network's value — even if the paid competitor's product is superior. Free sacrificed per-user revenue to maximise the denominator. The network effect turned that denominator into an asset that no amount of money could replicate. The reinforcement is direct: free accelerates network growth, and network growth makes the free product more valuable, attracting more users, which strengthens the network further.
Reinforces
Platform Business Model
Free is the default pricing strategy for platforms because platforms derive value from the interactions between two or more user groups — and charging one side inhibits the growth that makes the platform valuable to the other side. Google charges advertisers because advertisers pay to reach users. Charging users would reduce the audience that advertisers pay to access. Uber subsidised rider fares (making rides effectively cheaper-than-free through venture-funded promotions) to build the rider demand that attracted drivers. The platform business model requires critical mass on at least one side, and free — or below-free subsidies — is the fastest path to that critical mass. Every successful platform in the last two decades has used free on at least one side of the marketplace as the distribution mechanism.
Reinforces
Distribution
Section 8
One Key Quote
"Information wants to be free. Information also wants to be expensive. That tension will not go away."
— Stewart Brand, Hackers Conference (1984)
Brand's statement is the most concise articulation of the paradox that defines the economics of free. Information — and by extension, any digital product whose marginal reproduction cost approaches zero — "wants to be free" in the sense that the economics of copying push the price toward zero. A song, a search result, a software application can be reproduced and distributed at negligible cost. The marginal economics pull the price downward, toward free, because any price above marginal cost creates an arbitrage opportunity for a competitor willing to charge less.
But information also "wants to be expensive" because creating it is not free. Google invested billions in building the search algorithm. Spotify pays billions in music royalties. Netflix spends $17 billion annually on content. The creation cost is enormous. The reproduction cost is trivial. The business model puzzle is bridging the gap: how do you fund expensive creation when the market price of reproduction is zero?
The resolution for the companies that have solved this puzzle is indirect monetisation. Google funds expensive search infrastructure through advertising. Spotify funds expensive royalty payments through premium subscriptions and ads. Netflix funds expensive content through subscriptions. In each case, the product distributed to the user is priced at or near free (Google Search is free, Spotify's free tier is free, Netflix's per-hour cost is pennies). The creation cost is recovered not through direct payment for the product but through monetisation of the attention, data, or subscriber base that the free (or near-free) product aggregates. Brand's tension "will not go away" because both forces are real: the economics of digital reproduction push prices toward zero, while the economics of digital creation demand revenue to sustain investment. The companies that thrive are the ones that embrace both sides of the tension — using free distribution to aggregate value and indirect monetisation to capture it.
Section 9
Analyst's Take
Faster Than Normal — Editorial View
Free is the most misunderstood pricing strategy in technology. Most people treat it as an absence — "they don't charge" — when it is actually a deliberate architectural decision with specific economic logic. The question is never "why is it free?" The question is "who is paying, what are they paying with, and is the exchange sustainable?"
The pattern I track most closely: free as a competitive weapon against incumbents with high-margin business models. Craigslist destroyed $15 billion in newspaper classified revenue not by building a better product but by making the product free. Google Docs threatened Microsoft Office not by matching features but by making collaboration free. Android took 72% of the global smartphone market not by being the best mobile OS but by being the free one that hardware manufacturers could adopt without licensing fees. In each case, free was not a pricing decision — it was a strategic attack on the incumbent's revenue model. The incumbent could not respond by matching the price without destroying its own economics. Free is the ultimate competitive asymmetry: the attacker prices at zero and monetises elsewhere, while the defender's entire business model depends on the price staying above zero.
The most dangerous version of free is "free for now." Venture-subsidised free — where the product is funded by investor capital rather than sustainable economics — is a land grab that works only if the company can transition to monetisation before the capital runs out. Uber's rider subsidies, DoorDash's delivery promotions, and WeWork's below-market rent all followed this pattern: free (or below-cost) to build market share, then a planned transition to profitable pricing. The pattern works when the subsidised product creates switching costs, network effects, or habits that persist after the price increases. It fails when the product is a commodity and customers leave when the subsidy ends — which is why Uber's ride-hailing business still struggles with profitability while its payments and advertising businesses grow.
The AI era is testing free's limits in new ways. OpenAI launched ChatGPT with a free tier in November 2022 and reached 100 million users in two months — the fastest consumer product adoption in history. The free tier worked as a distribution strategy. But the cost structure is different from traditional free digital products: each ChatGPT query costs OpenAI meaningfully more than a Google search costs Google, because large language model inference requires expensive GPU compute. The marginal cost is not approaching zero — it is approaching cents per query at massive scale. OpenAI's economics depend on either reducing inference costs dramatically or converting free users to the $20/month Plus tier at rates that justify the infrastructure investment. If inference costs decline on the same trajectory as storage and bandwidth costs ( for compute), free AI will follow the Google playbook. If they don't, AI may be the first major technology category where free proves unsustainable at scale.
Section 10
Test Yourself
Free is present in more business models than most people recognise, and its economics are more nuanced than "give it away and figure out revenue later." The scenarios below test whether you can identify when free is a strategic weapon versus when it is a subsidy without a plan, and whether you can trace the actual payment from the person who bears the cost.
Is Free operating as a viable strategy here?
Scenario 1
A B2B analytics startup launches a free tier offering dashboards for up to 3 users with 30-day data retention. Paid plans start at $49/month for unlimited users and 12-month retention. After 18 months, the startup has 50,000 free users and 2,100 paid accounts. The cost to serve free users is $8/month per account. The average paid account generates $127/month in revenue.
Scenario 2
A food delivery startup offers free delivery on all orders (no minimum) for its first two years of operation, funded by $150M in venture capital. Customer acquisition is rapid — 3 million active users. When the startup introduces a $4.99 delivery fee in year three, 40% of users stop ordering within 60 days. The remaining users' order frequency drops by 25%.
Scenario 3
A professional networking platform is free for all users. It monetises through three channels: premium subscriptions ($29.99/month for recruiters and sales professionals), advertising (targeted ads based on professional profile data), and enterprise hiring tools ($8,000+/year per corporate seat). Of its 900 million users, approximately 6% pay for premium features. Total annual revenue exceeds $15 billion.
Section 11
Top Resources
The literature on free spans economics, behavioural psychology, media theory, and platform strategy. The strongest foundation combines Anderson's taxonomy of free business models with Ariely's research on the psychology of zero pricing and the platform economics literature that explains how free creates network effects and multi-sided marketplace dynamics.
The definitive treatment of free as a business model. Anderson, then editor-in-chief of Wired, identifies four models of free (cross-subsidisation, advertising, freemium, and gift economy) and traces the economic logic from King Gillette's razors to Google's search engine. The book's central argument — that digital economics have made free the natural price for information goods — remains the foundational text for understanding why the most valuable companies in the world give their core products away. The chapters on the psychology of free and the history of cross-subsidisation are particularly strong.
Ariely's chapter on the zero price effect is the most important piece of research for understanding why free is categorically different from cheap. His experiments demonstrating that demand discontinuously spikes at zero — not just because the price is low but because zero eliminates the mental transaction cost of evaluating whether the purchase is worth it — provide the behavioural foundation for every free-tier strategy in technology. The book contextualises the zero price effect within a broader framework of irrational economic behaviour.
The most rigorous treatment of platform economics and multi-sided market dynamics — the structural foundation that makes free viable for platforms. The book explains why platforms subsidise one side (free for users) to build the network that the other side (advertisers, recruiters, merchants) pays to access. The analytical framework for deciding which side to subsidise, how to design the free offering, and when to transition to monetisation is essential for any founder considering a free-tier strategy.
Wu traces the history of businesses that aggregate attention through free or subsidised products and monetise through advertising — from penny newspapers in the 1830s to Google and Facebook. The book provides the historical context that Anderson's "Free" lacks: the recognition that free-as-business-model is not a digital invention but a pattern that recurs whenever a new medium makes distribution cheap enough to give away content and sell the audience. The chapters on the attention-extraction arms race and its social costs are particularly relevant in the post-social-media era.
Written before Google, Facebook, or smartphones, Shapiro and Varian's analysis of information economics is remarkably prescient. Their treatment of versioning (offer a free basic version, charge for premium features), lock-in (free creates switching costs that enable future monetisation), and network effects (free maximises the network that creates value) anticipated the business models that would come to define the internet economy. Varian later became Google's chief economist — the book reads like the theoretical blueprint for the company he would help build.
Free — When marginal cost approaches zero, free becomes the optimal price for maximising distribution. Monetisation shifts from the user to advertisers, data, network effects, or premium conversion.
Free is the ultimate distribution hack. In a market where customer acquisition costs for SaaS companies average $200–$500 per customer through paid channels, a free tier that converts at 5% produces an effective acquisition cost of zero for the free users and a blended cost far below paid-channel alternatives. Dropbox's referral programme — give free storage to both the referrer and the referred — turned every user into a distribution channel, growing the product from 100,000 to 4 million users in 15 months without paid advertising. Free turns the product itself into the distribution mechanism: users adopt because there's no barrier, and they share because there's no cost. When distribution is the primary constraint on growth — which it is for most digital products — free is the most effective solution.
Leads-to
[Freemium](/mental-models/freemium)
Freemium is the most common monetisation architecture built on top of free. The free tier serves as distribution and acquisition. The premium tier serves as monetisation. The relationship is sequential: free comes first, freemium comes after. The strategic challenge is designing the boundary between free and paid such that the free tier is useful enough to drive adoption but limited enough to drive conversion. Slack limited free-tier message history. Zoom limited free-tier meeting duration. Spotify limited free-tier audio quality and added advertisements. Each limitation was calibrated to let users experience enough value to become dependent and then feel enough friction to justify paying. Freemium does not work without a compelling free tier — and a free tier without a freemium model is a cost centre, not a business.
Leads-to
Loss Leader Strategy
Free is the extreme case of the loss leader strategy — pricing a product below cost (in this case, at zero) to drive traffic, engagement, or purchase of a complementary product. Amazon's Kindle was sold at or near cost because the monetisation was content purchases. Gillette gave away razors because the monetisation was blade refills. Game consoles have been sold at a loss for decades because the monetisation is game licensing. Free extends the loss leader logic to its endpoint: the "lead" product costs zero, maximising the customer base that the complementary product monetises. The risk is identical to the traditional loss leader: if the complementary monetisation fails to materialise, the loss leader is just a loss.
Tension
[Attention](/mental-models/attention)
Free creates tension with the economics of attention because every free product competes for a finite resource — the user's time and focus. When search is free, email is free, social media is free, video is free, news is free, and messaging is free, the scarce resource is not access to information but the attention required to process it. Free products proliferate because the marginal cost of production is zero, but the marginal cost of consumption — the user's attention — is decidedly not zero. The tension: free maximises production of digital goods while attention constrains consumption. The result is an attention economy where the strategic competition is not "can I build a product people will pay for?" but "can I build a product that captures and holds attention against an infinite supply of free alternatives?" The companies that win in this tension — Google, TikTok, Netflix — are the ones that monetise not just the product but the attention itself.
My operating rule for evaluating free businesses: follow the subsidy. Every free product is subsidised by someone — advertisers, premium users, investors, or the users themselves (through data). The sustainability of the free model depends entirely on whether the subsidy is self-reinforcing or depleting. Google's advertising subsidy is self-reinforcing: more users generate more ad inventory, which generates more revenue, which funds better search, which attracts more users. A venture-subsidised delivery service is depleting: each free delivery burns investor capital without generating a self-reinforcing revenue loop. The businesses that make free work permanently are the ones where the subsidy creates a flywheel. The ones that fail are the ones where the subsidy is a straight line to a cliff.
The ethical dimension deserves more attention than it receives. When the product is free and the user is the product, the incentive structure tilts toward extracting maximum value from user attention and data — often at the expense of user wellbeing. Facebook's algorithmic feed optimises for engagement because engagement drives advertising revenue. The feed is "free." The cost is measured in attention, mental health, and political polarisation. The companies that will define the next era of free will be the ones that find monetisation models that align the company's incentives with the user's interests — not the ones that extract the most value from the least informed participants in the transaction.
Scenario 4
An AI image generation startup offers unlimited free image generation for all users. Each image costs the startup $0.03 in GPU compute. The startup has 10 million monthly active users generating an average of 50 images per month. Monthly compute cost: $15 million. Monthly revenue from advertising and API licensing: $2 million. The startup has 8 months of runway remaining.
