Common Business Models
Where the money comes from: ads, subscriptions, commissions, freemium — and how to validate each.
Where the money comes from: ads, subscriptions, take rates, upsells, freemium—and how to validate each. A business model isn't "how do you make money." It's "who's willing to pay, why they pay, and whether you earn more or less as you scale."
After reading, you should be able to answer:
- Why some products make more money as they grow while others lose more
- Where the money for a "free" product actually comes from
- Which charging model is the most realistic for something you build on your own
Free isn't free
"Free" is a pricing strategy, not the absence of charging. The money just comes from someone else. A search engine is free to use; advertisers pay for your attention. A content platform is free to watch; merchants pay for the time you spend there. A tool is free for individuals; the team plan charges for collaboration and compliance. When you design a model, ask first: who is paying for these free users? If you can't answer, "free" really is just free, and you'll fall into the trap of "lots of users, zero revenue." If you can answer, free is an acquisition tool, and its cost gets recovered on the other side. freemium
The boundary of "free" matters too: the most common split is free on the growth part and paid on the value part. Note-taking is free, export and collaboration charge; browsing is free, downloading charges. Users have to be able to understand the split—if they can't see why this part charges, free will only attract people who never pay.
Four things that decide success or failure
Same model, opposite outcomes. The difference almost always comes down to these four. First, how close is the payer: when users pay directly, the signal is cleanest—you see same-day whether the product works. With advertising, you have to please two audiences at once, and user experience and advertiser results often fight. Second, does marginal cost fall: what one more user costs decides whether the scale effect is positive or negative. Software's marginal cost approaches zero, so software businesses earn more as they grow; businesses that need heavy human service lose more as they grow. Third, is the money recurring or one-time: subscriptions recur, purchases don't, and recurring money compounds—which is why everyone tries to turn one-time transactions into subscriptions. Fourth, how elastic is the price: if you double the price, how many users stay? Seven out of ten means you're hard to replace; three out of ten means you're a commodity.
How one person collects money
Filter the charging models by "can a single person run this end to end," and not many survive. Advertising and take rates need volume—you need a crowd of users or transactions before the money starts. Platforms need two sides; one person can't feed both. The realistic ones are subscription, one-time purchase, and freemium—none of them need volume, only a small group of people willing to pay. One-time purchase has another advantage: no renewal operation needed, which suits low-frequency tool products. Subscription suits high-frequency products where value keeps flowing. If neither works, fall back to freemium—the free tier builds the base, the paid tier collects the money. The validation method is unglamorous but it works: before you build anything, find five people who match the description and ask what they currently spend—in money or time—to solve this. If the answer is zero, the model probably doesn't hold. If the answer is real money, the need has already been validated once; you only have to build something better.