It started with a Wikipedia article about the Shinkansen. The bullet train. I don't remember why I was reading about Japanese bullet trains at quarter past midnight on a Wednesday, but there I was, in bed, phone propped on my chest, the house completely silent apart from the low hum of the boiler doing something it probably shouldn't. My wife was asleep. The kid had been down since half seven. And I was deep into the engineering decisions behind the nose cone of the 500 Series, which is apparently modelled on a kingfisher's beak. Biomimicry. Fascinating stuff.
By half twelve I'd moved from the Shinkansen to Japanese railway station design, then to the concept of omotenashi -- the Japanese philosophy of anticipating a guest's needs before they express them. From there to Japanese convenience stores, which I'd been reading about separately and which are, by any reasonable measure, the most perfectly designed retail spaces on earth. From there to the idea of small spaces designed with extraordinary care, which took me to a thread about Japanese capsule hotels, which took me to a design studio that specialises in micro-spaces, which took me to their portfolio, which was beautiful, and at some point I fell asleep with my phone on my face.
Three weeks later I started building a Japanese language learning app.
The path that isn't a path
If you'd asked me that Wednesday night what I was doing, I'd have said "wasting time on my phone." If you'd asked me to connect the dots between the Shinkansen's nose cone and a language learning app, I wouldn't have been able to. There was no plan. No research brief. No strategy. Just a bloke following his curiosity at an hour when he should have been asleep, clicking from one thing to the next, guided by nothing more purposeful than "that's interesting, what's next?"
But the connection was there, building itself underneath. The rabbit hole about Japanese design and attention to detail was feeding a growing obsession with Japan that had been accumulating for months -- the food, the fashion, the design philosophy, the music. And at some point that obsession crossed a threshold where it stopped being something I consumed and started being something I wanted to build around. The language app wasn't a random idea. It was the inevitable product of dozens of rabbit holes, all pointing in the same direction, none of them planned.
The method behind the mess
I've started to notice the pattern now. Every product I've built in the past year -- and there are fourteen of them, which is either impressive or unhinged depending on your perspective -- started the same way. Not with a brief. Not with market research. Not with a competitive analysis. With a rabbit hole.
CultureTerminal started because I spent a week obsessively reading about the death of RSS and the rise of algorithmic feeds. I wasn't trying to build anything. I was just annoyed that I couldn't find a good culture feed anymore, and I kept reading articles about why, and those articles led to other articles, and at some point I thought: I could just build one. So I did.
The Slack analysis I did -- pulling data from years of Slack messages to find patterns in what people share and talk about -- started because I was curious about one channel. One Slack channel I was in was consistently brilliant, and I wanted to know why. What were people sharing? When were they most active? What topics kept coming back? I started pulling the data at about nine in the evening and didn't stop until two in the morning, and by then I'd built a complete analysis tool that worked across every channel I had access to. Nobody asked for it. Nobody needed it. I just wanted to know.
Modern Retro, which is probably the project I'm proudest of, started because I fell down a rabbit hole about 1970s retail design. The colours. The typography. The packaging. Something about that era's visual language -- the warmth of it, the confidence, the way every supermarket and corner shop looked like someone had actually thought about how it should feel -- grabbed hold of me and wouldn't let go. I spent days looking at old photographs of shops, catalogues, brand identities. And then one night I thought: what if you took a modern brand and put it in that world? What would Nike look like as a 1970s corner shop? What about Apple? What about Spotify? That question became eighty-six AI-generated images, a website, a scoring system, and the thing people most associate with my name.
Curiosity isn't procrastination
There's a productivity narrative that says browsing the internet at midnight is a waste of time. Close the phone. Go to sleep. Read a book instead. And sure, sometimes that's right. Sometimes the midnight scroll is genuinely empty, a loop of the same three apps showing you nothing new, the digital equivalent of opening the fridge and closing it again.
But there's another kind of midnight browsing that looks identical from the outside but is doing something completely different inside. The rabbit hole. The follow-your-nose, click-click-click descent into a topic you didn't know you cared about. This isn't consumption. It's accumulation. Every article, every image, every thread, every Wikipedia tangent is adding to a store of reference material that your brain is filing away, connecting to other things, building patterns from.
The difference between wasting time online and doing research isn't the activity. It's the orientation. When you're doom-scrolling, you're passive. The content washes over you. When you're down a rabbit hole, you're active. You're choosing each click. You're following a thread. You're pulled forward by genuine curiosity rather than pushed along by an algorithm.
I know the difference because I do both, regularly, often in the same evening. I'll spend twenty minutes scrolling Twitter on autopilot, feeling vaguely hollow, and then something will catch my eye -- a link, a thread, a reference to something I don't know about -- and the switch flips. Suddenly I'm alert. I'm clicking. I'm opening tabs. I'm falling. And that fall, that unplanned, unstructured, completely purposeless descent into something new, is where every interesting thing I've ever built came from.
Following curiosity is the method
I used to feel guilty about this. In advertising, everything needs a rationale. A strategy. You can't spend three hours reading about Japanese convenience stores and call it work. You need a brief. You need objectives. You need to justify the time.
But now I build things. And when you build things, the rabbit hole IS the work. It's not separate from the creative process. It is the creative process. The browsing, the accumulating, the following of threads -- that's where the ideas come from. Not from brainstorms or strategy sessions or competitive audits. From midnight Wikipedia sessions about the nose cone of a bullet train.
I've stopped trying to make the process look respectable. My best ideas come from following curiosity wherever it leads, especially after nine in the evening when the kid is asleep and the house is quiet and there's nothing between me and the internet but a hot chocolate and a few hours of unstructured time. That's when the rabbit holes open up. That's when the obsessions form. That's when a Wikipedia article about a train turns into a language app that thousands of people use.
You can't plan serendipity. But you can make space for it. For me, that space is 9pm to midnight, a quiet house, and the willingness to follow a thread wherever it goes. The rabbit hole isn't the distraction from the work. The rabbit hole is the work.
