Why Every Place Looks the Same, and Why That's Not a Design Failure

Thoughts — things I notice while building.

Before I claim anything, look at this room for a moment. Scroll slowly.

What you just saw is a trick, but an honest one. The geometry of that room did not change. No wall moved, no plinth grew, not a single line was added — I nailed that down in the shader so I couldn't cheat myself. What was added is light. And with the light: shadow, depth, hatching, direction, weight.

At the start it was a floor plan. At the end it was a place.

Nothing happened in between — except in you.

Beauty was never in the thing

I wrote about Santayana a few weeks ago, and the sentence I still can't shake is this one: beauty is pleasure objectified. It does not sit in the object. It happens in you, and you attribute it to the object because you cannot do otherwise — the sensation sits in the same place as the seeing.

In that piece it sounded like epistemology. It is in fact a set of build instructions.

Because if beauty is a state in the beholder, then it is no longer a mystery. It is a dependent variable. Something you can trigger. Something you can — and here it gets uncomfortable — measure.

The A/B test is a beauty machine

Here is the claim I actually care about, and it is less comfortable than the usual lament.

The usual story goes: we traded beauty for optimisation. We stopped building beautiful places and started building efficient ones. Consumption ate aesthetics.

That is wrong, and wrong in an interesting way.

If beauty is measured pleasure, then the optimisation industry did not abolish it. It industrialised it. An A/B test is not a machine against aesthetics. An A/B test is a machine that does exactly what Santayana described, only with five thousand subjects and a dashboard: it measures pleasure and turns it into a property of the thing.

So we did not stop building toward beauty. We now build toward it with far better instruments than any master builder before us. And the result is, nonetheless, the same everywhere.

Why?

What falls through the measurement window

Because a measurement is always a window. And that window is narrow.

What an A/B test can capture is immediate pleasure: the first impression, the first five seconds, the click-through, the bounce. Everything that can be caught in a single moment, across enough people at once that the variance collapses.

The Austrian dramaturge Christian Mikunda spent half his working life on the other kind of pleasure. He describes experiential spaces not through taste but through dramaturgy — with tools borrowed from film: the Landmark, the eye-catcher that pulls you in at all. The gliding rhythm of circulation, the beat at which a space opens up in front of you and narrows again. The Core Attraction, the climax everything runs toward. The Concept Line, the thread that holds it together. And, this is his real subject: most of it works below the threshold of conscious attention. He calls that hypnoaesthetics.

Now put the two lists side by side.

Every single dramaturgical device loses its A/B test. Not because it is bad. Because it works across time — and every metric we have samples a single moment.

Friction always loses. Tension always loses to resolution. The detour always loses to the shortcut. A museum that walks you down a dark, far-too-long corridor before the hall opens up would fail every funnel there is — and that corridor is precisely why you still remember the hall forty years later.

We are not optimising against beauty. We are optimising for the beauty that fits inside a measurement window. The rest gets tested away — not out of malice, but because it shows up in the report as a minus.

The global optimum is beige

Now the part that explains why the result isn't merely boring but the same everywhere.

When a thousand companies optimise the same objective — conversion, dwell time, throughput — with the same instruments, on the same data, about the same audience, they do not run off in a thousand directions. They all run downhill. And downhill is the same direction everywhere.

Same gradient, same minimum.

This is where it stops being a question of taste. The Apple Store, the coworking space, the dental practice, the airport lounge and your landing page do not look alike because the same agency built them, or because designers got lazy. They look alike because they all arrived at the same local minimum. White, light oak, rounded corners, plenty of air, one plant, one LED strip. That is not a style. That is a fixed point.

And here is the cruel part: every one of those results is, taken on its own, perfectly fine. None of it is ugly. Every single one of those places is brighter, cleaner, more accessible, cheaper to run and easier to understand than whatever stood there before. The optimisation worked.

That is the problem. It worked — and the thing it couldn't see is now missing everywhere at once.

Try it on yourself

So far this is an assertion. So test it on yourself.

Below is a shopfront, and you get to make five decisions about it. They are the kind of decisions that get made every day, in real meetings, by people trying to do their job well. Next to each option is what it delivers. Pick what you would genuinely pick.

I won't give anything away here except one thing: it is not rigged. If you choose the dentils, the striped awning and the hand-painted sign five times out of five, you get exactly that facade. The funnel is open.

Only — if you didn't take it: you were free. And you decided against yourself anyway.

What can't be measured

This is where I have to be careful, because the comfortable version of this piece would now say: things used to be more beautiful. They weren't. Railway stations were cathedrals and also draughty, filthy, and unusable for anyone who couldn't walk. The magic of old places is often only magic because we no longer see the bill.

The precise claim is a different one, and it is more than enough:

Modern spaces are increasingly optimised for efficiency, consistency and measurability. Dramaturgical singularity has not been outlawed. It has simply never once appeared in a report as a plus.

That is not cultural pessimism. It is a selection pressure. And selection pressures do not require anyone to want something bad — they require one number that consistently points one way, and enough repetitions.

And what disappears under that pressure is, of all things, the part you remember.

This article was staged

One last thing, and it is the reason I built it this way at all.

This text did not inform you about staging. It staged you.

The room at the top, arriving out of nowhere before a single argument had been made — that was the Landmark. The eye-catcher that pulls you in before you've decided whether you want to stay. The scrolling that drives the drawings, your tempo, your beat, your back-and-forth — that was the gliding rhythm; I didn't play the dramaturgy at you, I handed it to you. The shopfront in the middle, the one thing you were allowed to touch — that was the Core Attraction. And the single claim running from the first line to this one was the Concept Line.

Four of Mikunda's tools, applied to a blog post. And the numbers in that shopfront, the ones that steered you, were invented. I told you so — but only after they had worked.

If this text stays with you, it won't be because of the argument. It will be because of the dark corridor I walked you down first.

That corridor is exactly what no A/B test can ever recommend.

We optimised every measurable variable. Then we wondered why nothing was memorable. Memory was never one of the variables.

Related

• The Sense of Beauty — Beauty Is Pleasure Objectified — the groundwork: why beauty was never in the thing.
• The Picture of Dorian Gray — what happens when you mistake beauty for a property.
• Why Good Explanations Fool Us — same mechanism, different field.

The four drawings in this article are not images, they are programs: procedurally rendered ink, live in the browser, without a single photograph. Scenes 1 and 4 are signed distance fields whose contours and hatching are generated in a second shader pass; scene 2 is 420 instanced polylines; scene 3 is SVG. The tremor in the lines is deliberate — it is reseeded nine times a second, not sixty, because that is how hand-drawn animation stays alive.

Christian Mikunda, "Der verbotene Ort oder Die inszenierte Verführung" (1996) and "Brand Lands, Hot Spots & Cool Spaces" (2002) — the terms Landmark, gliding rhythm, Core Attraction and Concept Line are his; I use them as tools, not as quotations, and the extension to optimisation and A/B testing is mine, not his. George Santayana, "The Sense of Beauty" (1896). The percentages in the interactive shopfront are entirely made up, and that is the point.