In a lot of products the person who uses the thing, the person who chooses it, and the person who pays for it are three different people — and if you cannot say which of them is holding the money, you will build the wrong product. Not because you are careless, but because a business optimizes, consciously or not, for whoever fills the account. Intentions are what you write on the wall at the offsite. Incentives are what ship. When they disagree, incentives win every quarter, and the payer is where the incentives live.
Most product thinking collapses these three roles into one word — "the customer" — and that collapse is where the trouble starts. Split them apart and a lot of otherwise baffling product behavior resolves into simple arithmetic. The user experiences the product. The buyer makes the selection decision. The payer produces the cash. Sometimes all three are the same person, and life is good. Usually they are not, and the gaps between them are the exact shape of everything users complain about.
The three-body problem
Run it on the products you already resent, and the pattern is immediate.
Ad-funded consumer software. You are the user of the feed. You did not pay for it. The advertiser did. So the product is not built to serve you — it is built to harvest you and deliver your attention to the party that actually wired money. Every metric that matters internally is a proxy for attention captured: time in app, sessions per day, scroll depth. Your wellbeing is not in that objective function, and not because the people building it are villains. It is because the payer is the advertiser, and the product optimizes for the payer. The user being harmed and the payer being served are not a bug in the model. They are the model. "If you're not paying, you're the product" is the folk version; the sharper statement is that the product will always contort toward whoever pays, and if that is not you, it will contort away from you.
Enterprise software. The user is an employee who opens the tool forty times a day and hates it. The buyer is a VP who chose it. The payer is a procurement budget. Ask why enterprise UX is so reliably miserable and the honest answer is that usability was never the selection criterion, because the person who suffers the UX had no vote. The VP bought for reasons that have nothing to do with whether the software is pleasant: it satisfies a compliance requirement, it shifts risk off the VP's desk, it integrates with the system the last VP bought, the vendor is a "safe" name nobody gets fired for, the account exec has a relationship. Bad enterprise UX is not incompetence. It is a product correctly optimized for its buyer, who is not its user. The employee's pain is an externality the model does not price.
Healthcare. Here the split is at its most vicious, and I have watched it from inside medicine. The patient uses the drug. The physician prescribes it — the buyer, who makes the choice but consumes nothing and pays nothing. The insurer, or the state, pays. Three parties, three utility functions, wildly misaligned, and the product plus its price contorts toward the one holding the money. Drug pricing is unintelligible as a response to patient willingness-to-pay, because the patient is not the payer; it is a rational response to what a reimbursement system will clear. Pharma sales forces target physicians, not patients, because the physician is the buyer. The patient — the actual user, the person whose body is at stake — is structurally the weakest voice in the transaction, because they are the one role touching neither the decision nor the cash.
Someone always pays, and that someone writes the roadmap
The "free" product is where founders fool themselves most cheaply. Nothing is free; "free to the user" just means the payer is standing somewhere off-screen. A marketplace that is free to browse is paid for by the sellers giving up rake. A free API is often paid for by the platform that wants your app to deepen its moat. A free consumer app is paid for by advertisers, or by a future paid tier, or by investors subsidizing usage until one of the first two arrives. In every case, follow the cash to its actual source and you can predict, with unsettling accuracy, what the product will become — because it will bend, release over release, toward the interest of whoever is bending the revenue curve.
This is the load-bearing claim, so let me state it flatly: map the money to whoever actually pays, and you have predicted the product's future. Not its marketing, not its mission statement — its future. A social product that turns on advertising will get more engagement-maximizing and more hostile to your time, on a schedule, no matter what the founder believes about wellbeing, because the payer rewards exactly that. A note-taking app you pay for directly will keep serving you, because you are the payer and your renewal is the revenue. Same category of product, opposite trajectories, and the only variable that explains the divergence is who holds the purse.
The payer is also the ultimate customer selector, which is where this connects to the argument that your price selects your customers. Price sorts which payers walk in the door; the identity of the payer then dictates what you build for them once they're in. Two halves of one machine: price filters the payer, the payer filters the roadmap. A founder who sets price to select serious buyers and then quietly optimizes the product for advertisers has wired the two halves against each other, and will be confused about why the serious buyers keep leaving.
Separation in people, separation in time
I spend my working life on the seam between deciding to buy and actually paying. Kommerce runs cash-on-delivery commerce in markets where institutional trust is thin — the shopper will not prepay, because the courts and ratings and brand-recognition scaffolding a Western founder assumes simply is not there. In COD, the order and the money are held by the same person but separated by days and a delivery van, and that separation is where all the risk and all the product live: confirmation calls, address scrubbing, the entire return-and-refusal apparatus exists because the decision to buy and the production of cash are not the same event.
That generalizes past COD. The misalignment I am describing is what happens when the buy-decision and the payment are split across people; COD is what happens when they are split across time. Same disease, different axis. In both cases the ops stack, the metrics, and the roadmap organize themselves around one question — where does the cash actually come from, and what does that source demand — and in both cases founders who reason about "the customer" as a single unified agent get blindsided, because there is no single agent. There is a decision-maker and there is a cash-producer, and the product is built for whichever one you actually depend on to survive.
Draw the map before it draws itself
Here is the concrete thing to do, and it takes an afternoon.
For your product, write three columns — User, Buyer, Payer — and fill in who occupies each role. Not "our customer." The specific person or budget line. Then mark every place they diverge.
| Role | Question it answers | Whose interest it encodes |
|---|---|---|
| User | Who experiences the product? | Usability, wellbeing, daily value |
| Buyer | Who makes the selection decision? | Fit to their criteria (risk, compliance, status) |
| Payer | Who produces the cash? | Whatever the revenue actually rewards |
Where all three collapse into one person — classic self-serve SaaS, a paid consumer app — you have a rare gift: every design decision that helps the user also helps the wallet. Defend that alignment. It is worth more than most features, and it is easy to lose the day you add an ad tier or a "teams" plan bought by a manager.
Where they diverge, you have a decision you are currently making by default. When the user wants one thing and the payer rewards another — and they will conflict, repeatedly — whose interest wins? There is no universally correct answer. Ad-funded models choose the payer and are enormous. Some of the most defensible products in the world deliberately choose the user against the payer's short-term interest and charge a premium for the protection. Both are legitimate. What is not legitimate is not choosing — because an unnamed conflict does not stay unresolved. It gets resolved for you, one incremental decision at a time, always in favor of the money, until you look up and the product you have serves someone you never meant to serve.
This is also the sharper way to read the commoditize-your-complement logic: when a giant makes something free, it is not generosity and it is not even really "free." It is a payer you cannot see moving the price of your complement to zero, because a different revenue line, paid by a different party, gets fatter when they do. Free-to-the-user is a sentence with the payer deleted. Your job is to write the payer back in — for their model and for yours.
Follow the money all the way to the hand that opens the wallet. Whoever that hand belongs to is writing your roadmap right now. The only question is whether you have read it.