“Make something people love” is the most repeated advice in product. In a large number of markets, it is also incomplete, in a way that quietly limits good companies — because it hides a question. Which people? Buying a product can involve three different ones: the user who works with the product, the budget-holder who pays for the product, and the decision-maker who chooses the product. When they are the same person, as in most consumer markets, the advice is perfect — the person who loves the product is the person who buys it. When they are three different people, it falls apart.

The most powerful person in the purchase often isn’t the user

I build companies and products for a living, and the markets I work in are typically the second kind. In practice they are messier than three neat boxes. A real purchase looks more like a RACI chart — people responsible, accountable, consulted, informed — and the user is frequently the one merely informed: told what they are going to be given. More awkward still, the most influential role often belongs to someone who is none of the three. A recommender who shapes the choice without using the product, paying for it, or owning whether it works. Call this person the mediator — the analyst, the procurement lead, the IT gatekeeper, the specialist whose recommendation the decision-maker signs off almost by reflex. They are usually the most powerful party in the purchase, and the least exposed to the end result.

That gap — power without exposure — can create a big problem. “Nobody ever got fired for buying IBM” was never a statement about IBM; it is a statement about the mediator. Their downside is asymmetric: a safe choice that underperforms costs them little, while a bold choice that fails costs them their standing. So they optimise for defensibility, coverage, and the absence of blame — and product quality drifts toward whatever the mediator rewards. The user’s experience is frequently not near the top of that list.

There is a deeper layer beneath even this. It is not only that these parties want different things; the structure of the market — regulation, procurement rules, funding mechanics, professional gatekeeping — can push the decision away from merit altogether. You can win the user, the budget-holder and the decision-maker and still lose to the incumbent the rules quietly favour, or to the box your product doesn’t tick. The structure decides what “good” is even allowed to mean.

Two markets with the same shape

The clearest place to watch this is enterprise software — the market most readers will have lived inside. The people who choose it are usually a procurement or IT function whose dominant concerns are security, compliance, and covering every conceivable requirement. Each concern is legitimate. Together they select for the product that ticks the most boxes rather than the one that does the daily job best — and a tool built to satisfy every possible use case tends to be mediocre at the common ones. The people who live in that software all day had little say in choosing it, and their frustration shows up nowhere the decision-maker has to look. It isn’t that anyone ignored the user out of malice. The user simply wasn’t what the choice was graded on.

Healthcare shows the same shape in a cleaner form. In the UK a GP prescribes, the patient takes the medicine, and the NHS pays — three roles, three different people, almost no overlap. It is instructive because the two forces pull in opposite directions. Clinical accountability is high: the prescriber answers for the outcome and usually sees the patient again, so efficacy is fiercely protected. But the patient’s price sensitivity is near zero, so cost discipline has to be imposed from somewhere else entirely. (Academics who studied a more mundane version of this — the American textbook market, where the professor assigns and the student is forced to buy — found it behaves just like the market for prescription drugs.) The structure decides what counts as good, and it is never simply the user’s preference or convenience.

The way in

None of this is fixed, and I’ve spent years working out the way through. Great products do win these markets — but they win differently, and slower. The route is to manufacture the one thing the structure withholds: user power. A product good enough that people adopt it without being told, talk about it, and start asking for it by name can build demand from the bottom until the mediator has to follow.

Slack moved through companies this way for years before most IT functions formally chose it; by the time procurement arrived, the decision had effectively been made on the floor. Figma spread desk to desk among designers until it became the standard the organisation bought rather than the one it selected. This is a real strategy, but a patient and expensive one. It means running a consumer-style adoption curve inside an institutional market, and carrying the cost of winning users who, for a long time, are not the ones signing the cheque. The companies that try it without the patience or the capital to outlast the mediator are the ones that quietly run out of road. Those that succeed have a moat that’s difficult for others to overcome.

The pattern underneath

There’s a pattern underneath all this. The quality of a product — as the person using it actually experiences it — tracks two things: how much power that user held in the decision to buy it, and how directly the mediator answers for getting that decision wrong. Where both are high, you get products people love, because loving them is how they were chosen. Where both are low — the user has no say and the mediator never has to look at the consequences — you get products people tolerate, and a market that rewards the tolerating. Most of these markets sit closer to the second than their participants like to admit.

I have watched teams with first-rate, commercially sharp product instincts — the right instinct, to sit close to the user and build the best possible thing — still find themselves pushing against this dynamic. For a long time I read it as something a team could execute its way out of. It isn’t. The misalignment is in the structure of the market, not the calibre of the people working in it.

The questions I ask before entering a market

So the opinion the structure points to is an uncomfortable one. The instinct to build the best possible product is necessary, but on its own it is not the optimum strategy. In these markets you have to decide — consciously — which of these parties you are really building for, and most teams never make the choice at all. They build for the user, because that is what good people do, and then they are puzzled when the market rewards someone who built for the buyer. Caring about the user is not the same as choosing them.

When I look at a market now, three questions tell me most of what I need to know. Who actually holds the pen in this purchase — and have I ever watched them make the decision, or only assumed I knew how they make it? Does the person recommending the product carry any cost when it turns out badly, or do they simply move on to the next one? And if the user has little power today, do I have the patience and the capital to build it for them from the bottom — or am I quietly depending on a decision-maker whose incentives I have just spent this piece describing? The answers won’t tell you whether to enter a market. They tell you which game you are actually playing in it.

It is the same pattern that surfaces wherever the person who uses the product and feels the outcome isn’t the person who holds the decision; I’ve written elsewhere about what it does to public money. In markets it is quieter — it shows up as the distance between the products we are told to build and the ones that actually get bought. The advice to make something people love was never wrong. It just never said which people.