Why I Built a 3D Tower Where Every Window Is a Real Company

The idea in one sentence

One tower. Every window a company. Dark glass is for sale.

That is the whole pitch, and it is deliberately simple because the thing it describes is not. You open thepillar.io and a tower rises in front of you against a city skyline. Ride the elevator up its side and you pass floor after floor of windows. Some are dark and empty, waiting. Others glow with a real brand: a logo on the glass, a company name, a link to their site. Sign in, pick an empty pane, paste your company URL, and the tower fills that window with your brand automatically. You are now on the tower for as long as your license runs.

It is an advertising placement dressed as a place. And the reason I built it that way, rather than as yet another banner-ad network, is that a place is something people want to explore, and a banner is something people want to close.

Where the idea came from

I run a portfolio of small software products. Chrome extensions, Shopify apps, a few free web tools. Every one of them lives or dies on a familiar loop: rank in a store, get installs, convert to paid, reduce churn. It is honest work and it compounds, but it is also a grind, and every product competes for attention in a channel that is saturated with products exactly like it.

The observation that led to The Pillar was this: a good Shopify app cannot go viral. It is too specific. A merchant needs it or they do not. Nobody shares an inventory app with a friend because it delighted them. But a strange, beautiful, explorable thing on the open web can be shared, because sharing it makes the sharer look interesting. The scarce resource is not another feature. It is a reason for someone to send a link to someone else.

So I asked a different question than I usually ask. Not "what problem can I solve for a merchant," but "what could I build that a person would want to show another person." A tower where every window is a real company answers that. It is a directory turned into a landmark. It is my whole portfolio, and eventually anyone else's brand, rendered as a skyline you can fly around.

How a window actually works

Under the immersive surface, a window is a simple, honest thing. It is a licensed advertising pane for a set period. Not property. Not a token. Not a security. You are not buying a piece of a building, because there is no building; there is a rendered tower and a license to display your brand on one of its panes for as long as you pay. I say this plainly on the site and in the terms, because the fastest way to turn a fun idea into a liability is to let anyone believe they are buying an asset.

The mechanics:

• The tower has exactly 1,920 windows. Eighty floors, four faces, six panes per face. That number is fixed, which makes the inventory finite and the higher floors genuinely scarcer.
• Around the tower, the surrounding city buildings carry billboards, a cheaper entry tier of ad space out in the streets.
• You sign in, pick a pane, and pay. The window lights up with your brand. A top-floor corner window costs more than a low-floor one, because prestige on the tower works the way prestige in a real skyline does: height and corners command a premium.
• Prices run from a few dollars a month for a city billboard up to a few hundred for a penthouse corner. Every price is shown before you commit, and it is exactly what you are charged.

When a window is taken, it gets its own shareable page. Share that link and it unfurls as a card showing the brand on the glass and where it sits on the tower. A crawler with no JavaScript reads the card and the structured data. A human clicking through gets flown straight to that window on the live tower. Both audiences are served by the same page, which matters for a project whose entire premise is that people will want to share it.

Leading by example: my own brands go up first

An empty tower is a sad tower, and a marketplace with no listings is a ghost town nobody wants to be the first to join. So the tower does not launch empty. My own products occupy a founder's row of windows on the top floors of the entrance face, each with its real logo on the glass and a link to its site. The same brands run as full billboards on the city buildings, shown as their actual social share cards so the skyline reads like real advertising from a distance.

This is not filler. It is the honest demonstration of what a window looks like when a real company takes one. When you fly up the tower and see a lit penthouse row, you are seeing exactly what your own window would look like. Lead by example, then invite others in. A marketplace that will not use its own product is telling you something, and I did not want to tell anyone that.

The part nobody warns you about: making it fast

A 3D city in a browser is easy to make and hard to make usable. The first working version ran at nine frames per second on the kind of laptop a lot of people actually own. That is not a minor blemish. That is broken. A beautiful thing that stutters is worse than a plain thing that works, because it promises delight and then withholds it.

Fixing it was the hardest month of the project, and the details are their own long story, but the shape of it is worth telling because it is the same lesson every performance problem teaches. The scene was issuing nearly six thousand drawing instructions to the graphics card every frame, most of them from traffic, pedestrians, and billboards that had each been built as hundreds of separate objects. The fix was to batch them: draw an entire fleet of cars, or an entire crowd, in a single instruction instead of hundreds. Drawing instructions dropped by more than eighty percent. The frame rate nearly tripled on the weak hardware and became effortless on anything modern.

The point for anyone building anything ambitious on the web: the exciting part is the idea, but the part that decides whether anyone stays is whether it runs. I could have shipped the pretty version and watched people bounce off the lag. Instead I spent the unglamorous weeks making it real. That is usually where projects like this quietly die, and it is exactly where they earn the right to exist.

Why height and corners cost more

The pricing is not flat, and the reason is worth explaining because it is the same instinct that governs real real estate, translated into a rendered tower. A window on floor 3 costs a few dollars. A corner window on floor 80 costs a few hundred. Nothing about the pixels is more expensive to render up there. The difference is prestige, and prestige on a skyline is a real thing you can price.

Height reads as status. When someone flies around the tower, the top floors are what they see first and remember longest, the way a person on a street looks up at the crown of a building, not its lobby. Corners are scarcer than mid-face panes and they catch two views at once, so they carry a premium the way a corner office does. I did not invent this. I borrowed it from how people already feel about tall buildings, and the pricing simply reflects a feeling the visitor already has before they read a single number.

There is a design lesson buried in that. When you price something novel, do not invent a value system from scratch and hope people adopt it. Find a value system people already carry in their heads and map your thing onto it. Everyone already knows that the penthouse costs more. I just had to not fight that intuition.

The sharing mechanics, and why they got so much attention

If the entire growth plan is "people share it," then the moment someone shares a link is the most important moment in the whole product, and it has to work perfectly. That turned out to be harder than the 3D.

Every window and billboard has its own page. When you paste that link into a social app, two completely different visitors arrive. The first is a crawler, a bot with no JavaScript that reads the page to build a preview card. The second is a human who clicks the card. The crawler needs a static, server-rendered card and clean structured data describing the brand, the slot, and its place on the tower. The human needs to be flown straight to that window on the live, interactive tower. The same URL has to satisfy both, or the share either previews as a blank box or dumps the human onto a dead page.

Getting this right meant the share page renders a real card that a bot can read, and then, only for a human with a browser, quietly redirects into the live tower pointed at that exact window. I also had to fix a subtle bug where shared windows were all claiming to be on the same floor because the code fell back to a hardcoded value when it could not parse a billboard's ID. Little things, but they are the difference between a link that spreads and a link that embarrasses you.

The lesson generalizes to any site that expects to be shared: your share preview is a product surface, not an afterthought. It is often the very first thing a new person sees of you, rendered by a machine that does not run your JavaScript. If that machine gets a blank card, your best distribution channel is quietly broken and you will not even see it in your analytics.

It is a game, too

Somewhere in the building, the tower stopped being just a marketplace and became a place you could mess around in. That was not the original plan, but it was the right accident. If people are going to explore, give them things to do.

So now you can leave the elevator and walk the streets, drive a car, fly a helicopter or a plane, or glide over the city in a wingsuit. You can walk in the front door and climb all eighty floors on foot through the tower's interior, each floor an open room ringed by the same ad panes you see from outside. The city around you is alive: traffic on the roads, pedestrians on the sidewalks, animals, weather that changes the whole mood of the scene. At night, a UFO occasionally drifts in and abducts someone off the street, and sometimes they shoot back. You can tap a pedestrian and they will say something.

None of this sells more windows directly. All of it makes the tower a place worth returning to, and a place worth returning to is a place worth sharing, and sharing is the entire growth strategy. The game is not a distraction from the marketplace. The game is what makes the marketplace visible.

Building the living city was its own set of decisions. The crowd started as boxes with heads, which looked exactly as bad as it sounds, and became little jointed people with a real walk cycle: legs swinging from the hip, arms counter-swinging, all of it running efficiently enough that I could put hundreds of them on screen at once. The traffic learned to fill the lanes and avoid piling up. The pedestrians became things you could tap and talk to, each with a small seeded personality so the same walker always says the same kind of thing, and a few of them will actually mention a real company that currently holds a window, pulled live from the tower's occupancy. The UFO show is a fifteen-second cutscene that only runs at night, drifting in, beaming someone up, occasionally taking a hit from the ground and fleeing, sometimes dropping the person safely and sometimes carrying them off.

One honest confession from that stretch, because it is a mistake worth sharing: at one point I had the whole living-city update built and was confused why none of it appeared on the live site. The answer was embarrassing and instructive. The work sat on a branch I had never merged. Built is not the same as live. A finished feature on an unmerged branch shows the world exactly nothing, and I had been staring at the old version wondering where my aliens went. Every builder learns this lesson, usually more than once, and the tower taught me my turn.

Honesty as a design constraint

A project like this attracts a specific kind of trouble if you are careless. The moment "own a window" starts to sound like "own an asset that will appreciate," you have wandered into territory that is both dishonest and legally radioactive. I designed against that from the start.

The language is deliberate and repeated verbatim everywhere: a window is a licensed ad pane, not real property and not a security. The pricing is transparent and matches the charge to the cent. There is no resale market implying appreciation, no promise that a window will be worth more later, no mechanic that rewards speculation over use. If something on my side fails during a purchase, the buyer is automatically refunded and told about it. The whole thing is built to be exactly what it says it is: a fun, finite, honest place to put your brand on some glass.

That constraint is not a limitation on the idea. It is what lets the idea exist without becoming a problem. The most creative projects still have to be truthful, and truthful turns out to be a feature, because people can feel when a thing is being straight with them.

Finite on purpose

There are exactly 1,920 windows and there will only ever be 1,920 windows. I could have made the tower taller when demand justified it, or spun up a second tower, or let the inventory grow to meet whatever interest showed up. I chose not to, and the choice matters.

Infinite inventory is worthless inventory. If anyone can always get a window, no window is special, and the whole premise of prestige on the top floors collapses. A skyline is compelling precisely because there is a limited amount of it and a limited number of great spots on it. The scarcity is not a growth ceiling I regret. It is the feature that makes a window mean something. A penthouse corner is valuable because there are only four of them per floor and only a handful of floors near the top, and no amount of demand will mint a fifth.

This runs against the instinct most software carries, which is to scale everything infinitely because marginal cost is near zero. But value and cost are different things. It costs me nothing to render a thousandth tower. It would cost the idea everything. Some things are worth more because there is less of them, and a designer who understands that will make a stronger product than one who reflexively scales inventory to infinity because the server can handle it.

What building this taught me about my other work

I came back to my Shopify apps and free tools with a few things I did not have before. The first is that the share preview is a real surface, and I audited every product to make sure a bot with no JavaScript gets a proper card, because I had watched how much the tower's growth depended on that one detail. The second is that finite and honest are not constraints to route around, they are things people can feel, and a product that is straight with its user earns a trust that a slicker, vaguer one never does. The third, and the one I underrated most, is that attention is upstream of everything. The best app in a category still needs a reason for someone to look, and a landmark can buy that look in a way that a better feature list cannot.

What it is really for

Let me be honest about the business case, because I am honest about everything else. The Pillar is not going to replace my Shopify apps as a revenue engine. Windows sell for a few dollars to a few hundred, and the tower is finite. As a pure ad business, the math is modest.

Its real job is different. It is the most shareable thing in my portfolio, which makes it the top of a funnel that every other product sits inside. Someone flies around a strange 3D tower, notices the founder's row of real products lit up on the top floors, clicks through to one of them, and discovers an inventory app or a margin tool they actually needed. The tower is a landmark that happens to be a directory. Attention is the scarce thing, and a landmark earns attention that a directory never could.

And more simply than that: it is the kind of thing I wanted to see exist, so I built it. Not every project needs a spreadsheet to justify it. Some of them earn their keep by being the reason someone remembers your name.

Go see it

The best way to understand The Pillar is to open it. Fly around the tower, ride the elevator, walk the streets, find the lit windows and see whose brand is on the glass. If you run a company and want your logo up there, pick an empty pane and light it. If you just want to see a strange, fast, honest experiment in what a browser can do, that is reason enough.

It is live at thepillar.io. It is one of the products I build and run at Extensions Market, alongside the Shopify apps and free tools that solve the more boring problems. The Pillar solves a more interesting one: how do you get someone to look in the first place.