Give the product away for free. Charge for the better version. Venture capitalist Fred Wilson named this model in a March 2006 blog post — "freemium" — and in doing so crystallised the business strategy that would define the next two decades of consumer technology. Wilson's formulation was precise: give your service away for free, acquire a lot of customers very efficiently through word of mouth and organic discovery, then offer premium-priced value-added services to your customer base. The sequence matters. Traditional business models start with the revenue model. Freemium starts with the distribution model. Revenue comes last — after the user base is built, after habits are formed, after switching costs are embedded in playlists, files, workflows, and social graphs.
The arithmetic looks catastrophic on first inspection. Typically 2–5% of free users convert to paid. That means 95–98% of your users generate zero revenue. A business school case analysis would call this a capital destruction machine — you are funding infrastructure, support, and product development for an audience that overwhelmingly refuses to pay. The analysis is correct on the inputs and wrong on the conclusion, because it treats free users as dead weight. They are not dead weight. They are the distribution layer. Every free user who shares a playlist, sends a Dropbox link, or schedules a meeting through Calendly is performing a marketing function that a paid acquisition channel could never replicate at that cost or authenticity. Freemium does not have a customer acquisition cost problem. It solved it — by making the product itself the acquisition channel.
Spotify launched in 2008 with a free, ad-supported tier offering full access to every song in its catalogue. Not a demo. Not a 30-day trial. Permanent, unlimited streaming — with ads, shuffle-only on mobile, and no offline playback. By 2024, Spotify had 615 million monthly active users, of which 239 million were paying subscribers. That is a 39% conversion rate — extraordinary by freemium standards — but it took sixteen years of habit formation to achieve. The 376 million free users were not waste. They were Spotify's distribution engine: every shared playlist, every social listening activity, every "Wrapped" campaign that flooded social media in December was a zero-cost marketing channel powered by users who paid nothing. Daniel Ek understood that the competition was not Apple Music. The competition was piracy. You cannot compete with free by charging money. You compete with free by offering a better version of free — and then offering an even better version for money.
Dropbox demonstrated that the free tier could be the product's viral loop. Drew Houston launched with 2GB of free storage in 2008 — enough to be useful, not enough to be sufficient. Every file shared with a non-Dropbox user required the recipient to create an account. Free users generated other free users, and some fraction of those users eventually needed more than 2GB. The referral programme layered virality on top of virality: invite a friend, earn 500MB of bonus storage. Dropbox grew from zero to 100 million users in five years without a traditional sales team. Slack proved the model in enterprise by offering free access for small teams, bypassing procurement departments entirely — engineers and designers installed it without asking permission, and by the time the company needed paid features, Slack was already infrastructure. LinkedIn separated the axes entirely: the social network was free, and the premium tiers sold different value to different audiences — recruiters paid for search, salespeople paid for Sales Navigator, job seekers paid for applicant insights.
The scale of the model's adoption is itself evidence of its power. By 2024, freemium had become the default business model for consumer software, developer tools, productivity apps, mobile games, and an expanding share of enterprise SaaS. Zoom, Notion, Figma, Canva, Duolingo, Discord, Grammarly, Loom — the list of billion-dollar companies built on free tiers is long enough to constitute a category. The gaming industry generates more revenue from free-to-play titles than from paid games. Fortnite, which has never charged a cent to play, earned over $26 billion in total revenue by 2023 through cosmetic purchases alone. The counterintuitive conclusion has been proven so many times it is no longer counterintuitive: giving the product away is, under the right conditions, the most profitable pricing strategy ever devised.
The model works under two conditions, and only two. First, free users must generate value that exceeds their cost to serve. This happens through network effects (LinkedIn's social graph becomes more valuable with each additional free profile), viral distribution (Dropbox's sharing turns users into acquisition channels), data generation (Spotify's listening data trains the recommendation engine that makes the product better for everyone), or content creation (YouTube, TikTok). Second, the marginal cost of serving a free user must approach zero. Software satisfies this naturally — adding one more user to a cloud platform costs fractions of a penny. Physical goods do not. A restaurant cannot give away free meals to 95% of diners and survive on the 5% who order wine. Freemium is a digital-native model because digital goods have near-zero marginal cost and infinite replicability. Violate either condition and the free tier is not a strategy — it is a subsidy.
Section 2
How to See It
Freemium reveals itself wherever a product maintains a massive unpaying user base alongside a smaller paying cohort — and the company grows faster than paid acquisition alone would explain. The diagnostic is the ratio: if total users dwarf paying customers by 10x or more, and growth exceeds what marketing spend would produce, freemium is the engine.
Consumer Software
You're seeing Freemium when a consumer app reports hundreds of millions of users but revenue implying only a fraction are paying. Spotify's 615 million users versus 239 million subscribers. Canva's 170 million monthly users with roughly 10 million paid accounts. The free base is not waste — it is the distribution layer. The company acquires users at near-zero cost through word of mouth and organic discovery, then converts a percentage through habit formation and feature gating. The diagnostic: if removing the free tier would collapse the company's growth rate, the free tier is the product and the paid tier is the monetisation.
Enterprise SaaS
You're seeing Freemium when a B2B product spreads inside organisations without top-down procurement. Slack's free tier let individual teams adopt the tool before IT departments knew it existed. Notion offered a free personal plan that spread through startups before enterprise pricing existed. Figma gave designers free access that created organisational dependency — when every design file lived in Figma, the company had no choice but to pay for the team plan. The enterprise freemium signal is bottom-up adoption: the product enters through individual users and expands until the organisation cannot function without it. At that point, the sale is not "should we buy this?" but "how much do we pay to keep what we already depend on?"
Gaming
You're seeing Freemium when a game generates billions from a product that costs nothing to play. Fortnite earned $9.1 billion in its first two years through cosmetic purchases in a free-to-play game. League of Legends, free since 2009, generated over $1.75 billion annually at peak through champion and skin sales. The gaming freemium model is the purest expression of the concept: the core experience is complete and free, and the paid layer adds identity expression — skins, emotes, battle passes — rather than functional advantage. The 98% who never spend are the game's population. Without them, the 2% who pay would have nobody to play with.
Investing
You're seeing Freemium when a company's valuation implies future conversion rates that exceed industry benchmarks. Investors pricing a freemium business must model two variables: the total addressable user base and the long-run conversion rate. If a company has 100 million free users and investors value it at $10 billion, the implied assumption is either high conversion (10%+), high revenue per user ($100+/year), or both. The diagnostic for overvaluation: compare the implied conversion rate against the 2–5% industry benchmark. If the valuation requires 15% conversion in a market where no comparable company has exceeded 8%, the market is pricing a fantasy.
Section 3
How to Use It
Decision filter
"Before offering anything for free, answer two questions. Does the free user generate value that exceeds the cost of serving them — through network effects, viral distribution, or data contribution? And is the marginal cost of adding one more free user close to zero? If both answers are yes, freemium is viable. If either answer is no, you are subsidising users who will drain resources without converting or contributing."
As a founder
The most consequential freemium decision is where to draw the line between free and paid. Draw it too low and the free tier is useless — users bounce before forming a habit. Draw it too high and the free tier satisfies permanently — nobody converts because nobody needs to. Slack drew the line at message history: free teams lost access to older messages, which meant institutional knowledge disappeared unless they upgraded. Dropbox drew it at storage: 2GB was enough to start but not enough to sustain a working file system. Zoom drew it at meeting duration: 40-minute limits on free group calls created friction at exactly the moment teams needed the product most. Each company found the threshold where the free experience created dependency but the constraint created urgency that only the paid tier resolved. The line should make the upgrade feel like a relief, not a purchase.
As an investor
Value a freemium business on three metrics, not one. Revenue growth is necessary but insufficient. The three metrics that determine trajectory: free-to-paid conversion rate (improving or plateauing?), net revenue retention (do paid customers expand spending over time?), and free user growth rate (is the funnel still widening?). A company converting 5% of a growing free base at 120% net revenue retention is compounding on three axes simultaneously. A company converting 3% of a stagnating free base at 95% retention is decelerating on all three. The model's financial profile is nonlinear: small improvements in conversion rate, applied to a large free base, produce disproportionate revenue gains. A company with 200 million free users that moves conversion from 3% to 4% adds 2 million paying customers without acquiring a single new user.
As a decision-maker
Use freemium as a competitive weapon, not just a pricing strategy. When your competitor charges for access, offering a free tier collapses their acquisition advantage. Zoom did this to Webex and GoToMeeting: while incumbents charged per-seat licences, Zoom's free tier let anyone host a meeting. By the time the pandemic hit in 2020, Zoom had hundreds of millions of free users who converted overnight when remote work made the 40-minute limit untenable. The free tier was not generosity. It was a strategic investment in distribution that paid off when the market shifted. Freemium is at its most powerful when the company offering it can sustain the free-tier cost longer than competitors can sustain the customer acquisition cost of competing against free.
Common misapplication: Assuming more free users automatically means more revenue. Free users who never engage, never share, and never approach the paid tier's constraints are server costs, not marketing assets. The freemium model requires active free users who either convert themselves or convert others. A million inactive accounts are worth less than ten thousand daily active free users who share, collaborate, and bump against feature gates.
A second misapplication: treating the free tier as permanent and immutable. Some companies start with generous free tiers to build a user base and then restrict them as the business matures. Evernote repeatedly tightened its free tier — limiting device syncing, note size, and offline access — as conversion plateaued. The free tier that acquires users in year one may not be the free tier that maximises revenue in year five. The line between free and paid is not a product decision. It is a strategic variable that should be recalibrated as the business scales.
Section 4
The Mechanism
Section 5
Founders & Leaders in Action
Freemium's greatest practitioners did not simply offer a free tier — they engineered the free experience to function as the company's primary distribution engine, habit-formation mechanism, and competitive moat simultaneously. The free product was not the marketing department's budget line. It was the product.
Ek built Spotify's freemium model against the unanimous opposition of the music industry. Record labels wanted paid-only access. Ek insisted that free, ad-supported streaming was the only path to converting pirates into paying customers — the real competitor was not Apple Music but illegal downloads, and you cannot beat free by charging money. The free tier gave users access to every song in Spotify's catalogue — a library worth thousands in iTunes purchases — for nothing. Ads interrupted between songs. Shuffle mode was enforced on mobile. Audio quality was capped. Offline playback was disabled. Each constraint was chosen to create a specific friction that the Premium subscription removed. The $9.99/month upgrade was not a purchase — it was a relief. No ads, unlimited skips, offline downloads, higher fidelity. Ek designed the free tier to make the paid tier feel inevitable rather than optional. By 2024, Spotify's free-to-paid conversion rate reached approximately 39% — staggeringly above the 2–5% industry average — achieved not through aggressive upselling but through sixteen years of escalating endowment. A user who builds 200 playlists over three years, who has sixteen years of listening history training the algorithm, experiences an endowment effect so powerful that $9.99/month feels trivial relative to what they would lose by leaving.
Bezos applied freemium logic to cloud infrastructure — a market that had never imagined a free tier and did not think it needed one. When AWS launched its Free Tier in 2010, the offer was audacious: 750 hours of EC2 compute, 5GB of S3 storage, and access to over 85 services, free for twelve months. Enterprise IT executives dismissed it. The free tier was not designed for enterprises. It was designed for developers, students, and startup founders — people who would learn AWS's APIs, build on AWS's infrastructure, and carry those habits into the companies they joined or built. A student learning cloud computing for free in 2012 became a startup CTO defaulting to AWS in 2015 became a VP of Engineering mandating AWS across a Fortune 500 in 2020. Each career step amplified the free-tier investment. AWS also applied freemium at the service level: Lambda launched with a million free requests per month, DynamoDB offered 25GB of free storage. Each new service's free tier lowered the experimentation barrier, and experimentation created dependency embedded in the codebase. By the time workloads exceeded the free tier, switching costs were already architectural. AWS generated over $100 billion in annual revenue by 2024, built on a foundation of developers who started for free.
Section 6
Visual Explanation
The funnel narrows from the full free user base at the top through engaged users to the paying cohort at the bottom — typically 2–5%, though Spotify's sixteen-year maturation pushed it to 39%. The three boxes below the funnel show why the unpaying majority is not dead weight: they serve as the distribution engine through sharing and referrals, create network effects that make the product more valuable for everyone, and generate data that improves recommendations, search, and personalisation. The two conditions at the bottom are the prerequisites: freemium works only when free users generate more value than they cost to serve, and when the marginal cost of one additional free user approaches zero. Violate either condition and the free tier becomes a subsidy, not a strategy.
Section 7
Connected Models
Freemium sits at the intersection of pricing strategy, distribution theory, and behavioural psychology. Its power comes not from the pricing itself — giving things away is easy — but from the compounding interactions between free distribution, habit formation, network density, and the structural advantages that a massive user base creates over time.
Reinforces
Network Effects
Freemium and network effects form the most powerful growth loop in consumer technology. The free tier maximises the user base, and a larger user base increases the value of the product for every user — which attracts more free users, which strengthens the network further. LinkedIn's free tier built the professional graph to over 1 billion members. That graph is what makes Premium, Sales Navigator, and Recruiter valuable — without free-tier network density, the paid products would have no audience to monetise. Slack's free tier created network effects within organisations: once half the team was on Slack, the other half had no practical choice. The free tier does not just acquire users. It makes the product exponentially more valuable with each user added.
Reinforces
Distribution
Freemium is a distribution strategy disguised as a pricing model. The free tier eliminates the primary barrier to adoption — cost — and transforms every user into a potential distribution node. Dropbox's referral programme layered viral distribution on top of the free tier: invite a friend, earn extra storage. Calendly's free tier turned every meeting invitation into a product advertisement — every recipient saw the Calendly branding and became a potential user. The insight: freemium companies do not have separate product and distribution strategies. The free product is the distribution strategy. The paid tier is the business model. The two operate on different layers and reinforce each other continuously.
Reinforces
[Hook](/mental-models/hook)
Nir Eyal's Hook Model — trigger, action, variable reward, investment — describes the loop that creates habitual product use. Freemium companies weaponise this loop because the free tier is optimised entirely for engagement, not revenue. Spotify's free tier delivers the trigger (a playlist notification), the action (press play), the variable reward (discovering a new song), and the investment (saving it to a playlist that increases switching costs). The longer the user stays in the free tier, the deeper the habit and the higher the eventual conversion probability. The free tier is not a loss-making charity. It is the habit-formation engine that makes the paid upgrade feel inevitable rather than optional.
Section 8
One Key Quote
"Give your service away for free, possibly ad supported but maybe not, acquire a lot of customers very efficiently through word of mouth, referral networks, organic search marketing, etc., then offer premium priced value added services or an enhanced version of your service to your customer base."
— Fred Wilson, AVC blog, 'My Favourite Business Model' (March 2006)
Wilson published this sentence in a blog post that would become one of the most influential business model descriptions of the internet era. The power is in the sequence: give away, acquire, then offer. Not build, price, sell. The traditional model starts with the revenue question. Freemium starts with the distribution question — because Wilson recognised that the most expensive line on most startups' income statements, customer acquisition cost, could be reduced to near zero by making the product itself the acquisition channel.
The deliberate ambiguity of "possibly ad supported but maybe not" was a crucial hedge. Freemium is not the same as ad-supported free. Ad models monetise every user through attention. Freemium monetises a fraction through features. The incentive structures diverge entirely: ad-supported products want you scrolling endlessly. Freemium products want you reaching the limit of the free tier and deciding the upgrade is worth it. One optimises for time spent. The other optimises for value delivered. Wilson left both doors open, but the companies that built the most durable freemium businesses — Spotify, Slack, Dropbox — chose the feature-gated path over the ad-supported one, because feature gates align the business with the user rather than against them.
Section 9
Analyst's Take
Faster Than Normal — Editorial View
The free tier is not a marketing expense. It is the product's primary distribution mechanism. The companies that treat the free tier as a cost centre — something to be minimised, restricted, and eventually eliminated — fundamentally misunderstand the model. Spotify's free tier is not a promotional giveaway. It is the reason Spotify has 615 million users instead of 50 million. Every free user who shares a playlist or triggers a social listening notification is performing a marketing function that paid acquisition could never replicate at that scale or authenticity. The free tier's cost is infrastructure. Its value is distribution. The ratio between those two numbers is what separates freemium brilliance from freemium bankruptcy.
The conversion rate obsession is a trap. Most freemium analyses fixate on free-to-paid conversion as the single health metric. A 3% conversion rate on a base growing 50% annually is more valuable than 10% conversion on a base growing 5%. The free tier's growth rate is the leading indicator. The conversion rate is lagging. Optimise the leading indicator first — build a free tier so good that users cannot help recommending it — and the conversion rate will compound on top of an expanding base. The companies that chase conversion at the expense of free-tier growth shrink the funnel they need to expand.
The AI-era freemium challenge is real: inference costs break the zero-marginal-cost assumption. Freemium works when adding a free user costs fractions of a penny. AI products face meaningful per-query costs — every free user running GPT-4-class inferences costs real money. The AI freemium winners will be the companies that find the cost threshold where free users generate network value — data, feedback, word-of-mouth — that exceeds the inference cost of serving them. The losers will be the ones offering expensive AI features for free without a credible path to conversion or a structural reason why free users improve the product for paid users.
The pattern I watch most closely: freemium compression. As freemium becomes the default model in every software category, free tiers become more generous and paid tiers harder to differentiate. When Notion, Coda, and Slite all offer robust free tiers, the free tier ceases to be a competitive advantage and becomes table stakes. The next phase will not be about who has the best free tier — it will be about who converts free users most efficiently and retains paid users most durably. Distribution is commoditised. Conversion and retention are the new moats.
One structural truth that most founders learn too late: the best freemium businesses are not the ones that convert the highest percentage of free users. They are the ones where the free tier generates so much ambient value — network density, data, content, distribution — that the company would keep it even if the conversion rate dropped to zero. Spotify would keep the free tier even at 0% conversion because ad revenue from 376 million listeners is a business in itself. LinkedIn would keep the free tier at 0% conversion because the professional graph is the foundation of every paid product. When your free tier is independently defensible — when it creates value beyond conversion — you have built a moat that competitors cannot replicate by offering a better paid product.
Section 10
Test Yourself
Freemium looks deceptively simple — free plus paid. The complexity is in the execution: where to draw the line, which users to optimise for, and whether the unit economics actually work when you run the numbers honestly. The scenarios below test whether you can distinguish freemium done well from freemium done badly — and whether you can identify when the model applies versus when it is being forced onto a business that would be better served charging from day one.
The most common analytical error is treating all free users as equivalent. They are not. A free user who shares, invites, and approaches the feature gate is a conversion candidate and a distribution asset. A free user who signs up and never returns is a line item on your server bill. The scenarios require you to distinguish between the two.
Is this freemium model working?
Scenario 1
A project management SaaS offers a free tier (up to 10 users, unlimited projects) and a paid tier at $12/user/month (unlimited users, advanced reporting, integrations). After two years: 3 million free accounts, 45,000 paid accounts. Free tier grows 8% month-over-month. Paid tier grows 3% month-over-month. The CFO argues the free tier should be restricted to force conversions.
Scenario 2
A cloud storage startup offers 15GB free, $3/month for 200GB. Marginal cost per free user: $0.02/month. After three years: 80 million free users, 2.4 million paid subscribers. Free users share files with non-users, driving 60% of new signups organically. A competitor launches with 50GB free.
Scenario 3
A fitness app offers free bodyweight workouts and charges $15/month for personalised plans, nutrition tracking, and coach access. After eighteen months: 5 million downloads, 800,000 monthly active users, 12,000 paid subscribers. Free content is consumed passively — users watch videos but do not create profiles, track progress, or engage with social features.
Section 11
Top Resources
The freemium literature spans venture capital strategy, behavioural economics, and SaaS growth frameworks. Start with Wilson's original naming and Anderson's theoretical foundation, then move to the practitioner literature from the companies that proved the model works at billion-user scale.
The gap in the literature is operational: most books explain why freemium works but few explain how to set the line between free and paid. The resources below cover both the theory and the case studies needed to apply the model to specific products and markets.
The blog post that named freemium and defined the model's structure. Wilson's formulation — give away free, acquire efficiently, offer premium to the base — remains the clearest articulation of the strategy twenty years later. Read this before anything else on freemium because every subsequent treatment builds on Wilson's foundation, and most do not improve on his concision.
Anderson's book-length treatment of why digital goods gravitate toward zero price and how businesses profit from giving things away. The economic argument — near-zero marginal cost makes free the inevitable distribution price — is the theoretical infrastructure for understanding why freemium works in software and fails in physical goods. The cross-subsidy framework (free core subsidised by paid premium) is the mechanism Wilson named but Anderson explained.
Eyal's Hook Model explains the behavioural loop — trigger, action, variable reward, investment — that makes freemium conversion work. The book is not explicitly about freemium, but the habit cycle it describes is the engine that converts free users to paid users. Understanding how products create compulsive engagement explains why Spotify converts at 39% while most apps stall at 3%.
Ariely's research on the zero-price effect — the irrational overvaluation of free goods — is the psychological bedrock of freemium. His experiments showing that consumers choose a free inferior option over a cheap superior option explain why the free tier drives adoption at rates that even very low-priced tiers cannot match. The chapter on free versus cheap is essential for anyone designing the boundary between tiers.
The definitive account of how Spotify built the most successful freemium business in consumer technology. Details Ek's negotiations with labels over the free tier, the internal debates about conversion optimisation, and the sixteen-year arc from piracy competitor to 239 million paying subscribers. Essential for understanding freemium as a multi-year strategic commitment rather than a pricing experiment.
Freemium — The funnel economics of giving the product away and converting a fraction to paid. The model works when free users serve as the distribution channel or cost nearly nothing to serve.
Tension
[Churn](/mental-models/churn)
Freemium's strength — zero barriers to entry — is also its structural weakness: zero barriers to exit. A user who paid nothing to join has nothing anchoring them if a competitor offers a better free experience. Free-tier churn rates routinely exceed 50% annually in consumer apps, compared to 5–10% for paid SaaS products. The paradox: the free tier acquires users at massive scale but retains them at low rates. The resolution is conversion velocity — moving free users to paid before novelty fades, or creating investment (playlists, files, professional networks) that makes the free tier progressively more expensive to abandon. Companies that solve free-tier retention win. Companies that treat the free tier as an endlessly leaky top-of-funnel lose the compounding advantage that makes freemium work.
Leads-to
Loss Leader Strategy
Freemium is the digital evolution of the loss leader. Grocery stores sell milk below cost to draw shoppers through the door, where they purchase higher-margin items. The free tier is digital milk — offered at a loss to bring users into an ecosystem where some will purchase the premium product. Amazon's Kindle devices have historically been sold at or near cost because the real margin comes from ebook and subscription purchases. The logic is identical: absorb the loss upfront, capture the lifetime value downstream. The distinction is that freemium's loss leader scales infinitely because digital distribution has no shelf space and no per-unit cost floor.
Leads-to
Free
Chris Anderson's concept of "Free" — that the internet economy gravitates toward zero-price products — is the theoretical foundation that freemium operationalises. Anderson argued that when marginal costs approach zero, free becomes the most efficient price because it maximises distribution and creates opportunities for adjacent monetisation. Freemium is the business model that translates Anderson's insight into an executable strategy: give the core product away following the "Free" logic, and capture revenue from the minority who value enhancements enough to pay. The shift from Anderson's "everything will be free" to Wilson's "free plus premium" was the move from economic theory to operating playbook.
Scenario 4
An AI writing tool offers 5,000 words/month free and $20/month for unlimited usage. Inference cost: $0.003 per query. The company has 2 million free users averaging 3,000 words/month and 150,000 paid users. The free tier drives 40% of paid conversions through word-of-mouth. The remaining 60% come from paid marketing.
