A Theory of "Taste"

A few weeks ago on a bus, a friend said something that stuck with me. We were talking about some founders, and she said, as if it were obvious, that some had good taste and others didn't.

The lazy view is that she was just expressing an opinion. It's what I was raised to believe: taste is subjective. Some people like one thing, some another, and who's to say who's right? But if you take that idea seriously, it leads to conclusions that are obviously false.

If there's no such thing as good taste, there can be no such thing as a good product, or a good building, or good anything. There can only be things that are popular. It would also mean it's impossible to get better at making things. A master carpenter would be no more skilled than an apprentice. Leonardo would be no better at painting than a child.

But that's obviously false. We can see the difference between a well-made chair and a shoddy one. The gap between an expert and a novice is not an illusion. And if skill is real, then some things must be objectively better than others. Which means taste, the ability to tell the difference, must be real too.

So why does the idea of subjective taste persist? Because our reaction to things feels so personal. If we disagree about a movie, who's right? But this confuses the reaction with the thing itself. The movie is out there in the world, external to us both. And while people vary, we have a lot in common. A well-designed object is like a key made for a lock. Our brains are all variations of the same model of lock. A good design is a key that fits. A great one turns the lock with a satisfying click. And a truly brilliant design is like a master key: it opens locks you didn't even realize were there. Taste, then, isn't a matter of opinion. It's a matter of fit, and the best taste is a matter of profound fit.

But the conversation usually stops at "effectiveness," a vague and unhelpful word. It's dangerous because making things is a type of engineering, and you can't engineer a vague target. Great products are not happy accidents. That the same people can build them again and again is the proof. Which proves taste is a skill. And if it's a skill, you can deconstruct it.

To deconstruct taste, you have to start with the constraints. There seem to be three that matter most. They aren't opinions so much as observations about how the world works, the kind you notice after watching enough products succeed or fail. Ignoring them isn't a matter of aesthetics; it's a guarantee of failure.

Products are for people.

A product's value isn't inherent in its technology; it's created when a person uses it. The Segway was a marvel of engineering, a self-balancing machine that felt like the future. But it was a solution built from the technology outward, without a clear map to a pre-existing human behavior. It solved no one's problem elegantly. Contrast that with the first iPod, which started with a simple human desire: "1,000 songs in your pocket." The technology served the person, not the other way around.

Products solve problems.

A good product moves a person from a state of struggle to a state of relief. Without a real problem, a product is just a clever demo. Google Glass was a glimpse into a potential future, but for the average person, it didn't solve a problem that outweighed its high cost and social awkwardness. It was a solution in search of a problem.

All resources are finite.

A user's attention is brutally scarce. This is the fundamental constraint that forces good decisions. Ignoring it leads to products like the Bloomberg Terminal, immensely powerful for experts who have invested thousands of hours, but impenetrable to anyone else. It demonstrates that a product that makes infinite demands on a user's finite attention can only ever serve a niche.

While these axioms might seem simple, the interesting thing is to try arguing for their opposites. Go ahead, try to make the case for a product whose value from technology no one uses. Or a product that pointedly solves no problem. Or that you can make something as complicated as you want, because users have infinite attention. You'll find you're on shaky ground very quickly. These aren't opinions. They're the walls of the room we're in. And it's from these axioms that we can deduce the principles of taste, not as a matter of preference, but of logic.

The first is Empathy (E). This follows directly from the first axiom. If a product is made for a person, then it has to fit that person's world. And to make something that fits, you have to understand the shape of the space you're fitting it into. That means you need an accurate model of how the user sees things. The hard work of building that model, and then refining it over and over, is what we call Empathy. You can measure it by how little friction (μ_F) a user feels. Friction is just the distance between the user's model of the world and the product's. Your goal is to make that distance as close to zero as possible. So, E ∝ 1/μ_F.

The second is Foresight (F). This follows from Axiom 2. If you're going to solve a problem, you have to be good at choosing which problems to solve. There is more value in solving a durable problem than a fleeting one, because your solution generates value for a longer time. This is true of all created things. Some books become classics; others are bestsellers for a month and then disappear. Some movies are watched for decades, while summer blockbusters fade from memory. The difference is that the durable things solve durable problems. A classic novel is usually about some fundamental aspect of the human condition. A disposable bestseller is often about a specific celebrity or political event. We can measure this durability with a concept I'll call Value Half-Life (V_1/2), the time until a thing's core value is halved. A viral game like Flappy Bird had a half-life of months because the problem it solved, fleeting boredom, was not itself durable. A calculator or an email client has a half-life of decades because the underlying needs they serve change very slowly, if at all. Good taste in products means choosing to solve problems with a long half-life. So, F ∝ V_1/2.

The third is Simplicity (S). If a user's attention is finite (Axiom 3), then every piece of complexity you add is a tax on that scarce resource. Therefore, a product that respects this constraint must be economical with its demands. The ruthless curation of the unnecessary isn't an aesthetic preference; it's the only logical response to the user's cognitive limits. Simplicity can be measured by the Utility-to-Complexity Ratio (U/C_x). So, S ∝ U/C_x.

Finally, there is Craftsmanship (C). This pillar is a consequence of the first two axioms working together. A product that is buggy or unreliable fails to reliably solve the problem (violating Axiom 2) and in doing so, breaks trust with the person using it (violating Axiom 1). An unreliable solution is not a solution at all. Therefore, the skill of high-fidelity execution is required. The enemy of craftsmanship is defect density (D_d). Great craftsmanship minimizes it. So, C ∝ 1/D_d.

These pillars are not independent. The relationship feels multiplicative, because a zero in any one area seems to cause total failure.

We can try to model this relationship with a formula. But we have to be careful here. A formula can create an illusion of precision, and the variables we're talking about (Friction, Half-Life, Complexity) are not things you can measure with a ruler. The formula's real power isn't in spitting out a number; it's in forcing you to ask the right questions. Its purpose is to structure your thinking, not to replace it.

The most robust part of the model is the multiplicative relationship:

This just says that all four matter, and that a failure in one can't be compensated for by brilliance in the others. This much seems undeniably true.

If we want to speculate further, we can try to express how the components of these pillars might relate to one another. One way to model it would be like this:

This version of the model is more specific, and therefore more fragile. But it's useful because it makes a claim about tradeoffs. It suggests that you want to maximize the things that create lasting value and utility, while minimizing the things that create friction, complexity, and decay. It's a mathematical argument for a specific kind of focused, user-centric minimalism.

This raises a question, though. If taste can be deconstructed, why is it so rare?

Because having taste isn't about mastering four separate skills. It's about winning a single, unified battle against the path of least resistance. It's a battle against entropy. The universe doesn't want a chair to be well-made; it wants the wood to rot. Society doesn't want a product to be simple; it wants to add one more feature. Our own brains don't want to do the hard work of empathy; they want to build what's easy.

Empathy is the hard work of winning the fight against your own ego when it's easier to build what you want. Simplicity is the courage to say no when it's easier to say yes. Foresight is the patience to work on durable problems when the world rewards chasing trends. And craftsmanship is the discipline to polish every surface when it's easier to ship something that's "good enough."

So taste is rare for the same reason discipline is rare. It means choosing the harder path at every step. This is what people with taste do. Their intuition runs the numbers on that equation instinctively, not because they're geniuses, but because the discipline has become part of them. They've internalized the battle. The result looks like magic to the rest of us. But it's not. This internalization is what makes someone an artist, or more simply, a person with taste.

Notes

1. One reason so many products lack taste is that large organizations are terrible at this kind of thinking. In a big company, adding a feature is often a political victory for someone, regardless of its impact on the product's overall taste score. Startups have an advantage here, because they are small enough that a single person, or a small group, can hold the entire model of the product in their heads and defend it.
2. Here is a practical example of how the formula can be used as a sketch. Imagine we have a simple note-taking app, and someone suggests adding a "Kanban View." It sounds professional, and other apps have it. The temptation is to say yes. But let's run a quick gut check.
   We don't need precise numbers, just rough estimates. Let's say adding this feature increases overall Utility (U) by a factor of 1.5 for a small group of power users. But it also doubles the number of interactive elements, so Complexity (C_x) goes up by 2x. For the majority of users who just want to write a note, there's a new button they have to see and ignore, so Friction (μF) increases slightly, say by 1.2x. And this new view is a whole new surface for bugs, so Defect Density (D_d), the potential for things to go wrong, might increase by 1.8x.
   The numerator in our equation changes by 1.5. The denominator changes by (1.2 * 2.0 * 1.8), which is about 4.3. The taste score changes by a factor of 1.5 / 4.3, which is about 0.35. We've made the product a lot worse.
   I haven't given rigorous definitions for these variables, because you don't need them. The formula isn't a calculator; it's a way to structure your thinking. What we call a 'gut check' is really a subconscious weighing of tradeoffs. This model just makes that process conscious. It's useful for big questions, like whether to add a feature, but its real value is in training your intuition for the countless small decisions that good work consists of: the wording of a button, the curve of a corner, the timing of an animation. It forces you to see the hidden costs of every choice, and that's its real purpose.
3. While this essay discusses taste as a discipline for making things, the Polish poet Zbigniew Herbert explored it as a form of moral resistance. In his poem "The Power of Taste," he argues that refusing to accept the ugliness and cheap rhetoric of a totalitarian regime wasn't a matter of great courage, but "fundamentally it was a matter of taste." For Herbert, taste was a form of conscience, an instinctive recoil from the false and the ugly. It's a powerful reminder that the same faculty we use to judge the curve of a chair can be used to judge the shape of a society. Both are, in their own way, a battle against entropy.