It started with a stupid question. What if Supreme had existed in 1974? Not in an ironic way. Not as a joke. But genuinely -- what would a Supreme retail store have looked like on a high street in the mid-seventies? The signage, the window display, the shopfront. That specific combination of a modern brand's identity with the grain and warmth and slightly naff charm of seventies retail design.
I don't remember where the question came from. I think I was scrolling through vintage shop photography on some corner of the internet, the kind of rabbit hole I fall into regularly, and I saw a picture of a record shop from about 1973. Orange carpet, wood panelling, those chunky plastic letters on the fascia. Beautiful in a way that modern retail has completely forgotten. And I thought: what would it look like if you put a brand I know -- a brand everyone knows -- into that world?
That was it. That was the entire brief. No market research. No competitive analysis. No user personas. No business model. Just a question that wouldn't leave me alone.
The first image
I used fal.ai to generate the first image. I'd been playing with AI image generation for other projects, but this was different. The prompt needed to be specific in a way that most AI image prompts aren't. I didn't just want "a 1970s shop." I wanted a 1970s shop that felt real. That had the right film grain, the right colour temperature, the right slightly faded quality. I wanted it to look like someone had found an old photograph in a drawer.
The first few attempts were terrible. Too clean, too modern, too obviously AI-generated. The brand logos looked wrong. The proportions were off. The seventies aesthetic was there in a costume-party kind of way -- surface level, no depth. I kept adjusting the prompts, pushing for more specificity. The wood panelling. The fluorescent lighting. The hand-painted price cards in the window. The slight tilt of the camera, like someone took it casually on their way past.
And then one came out right. I don't remember which brand it was -- it might have been Nike -- but the image had that thing. That click. It looked like a photograph, not a generation. It looked like it had existed. And I sat there staring at it thinking: I need to do this for every brand I can think of.
The escalation
This is the part that's hard to explain to people who don't build things. The escalation from one image to ninety-six brands wasn't a business decision. It wasn't strategic. It was compulsive. Each brand I generated opened up three more that I wanted to try. Supreme led to Nike led to Patagonia led to Apple led to -- absurdly, brilliantly -- IKEA and Peloton and Oatly. Brands that had no business existing in 1974, which made the images more interesting, not less.
I built a scoring system. Five factors, twenty points each, giving every brand a score out of a hundred based on how absurd it would be for them to have existed in the seventies. Peloton scores high because a connected exercise bike in 1974 is genuinely laughable. Levi's scores low because they actually did exist then. The absurdity score became part of the project's identity -- it gave structure to something that could have been shapeless.
I added categories. Technology brands. Fashion brands. Food and drink. Sports. Each category brought its own visual challenges. A Tesla shopfront in 1974 needs to look different from an Aesop shopfront in 1974. The lighting, the neighbourhood, the type of high street -- all of it has to feel right for the brand while also feeling right for the era. That tension is where the best images live.
Why this is my proudest project
I've built fourteen products. Some of them are genuinely useful. Trove helps me understand my own taste. CultureTerminal aggregates culture news. Little London helps parents find things to do. These are practical products that solve real problems. Modern Retro solves no problem at all. It's not useful. It's not practical. It doesn't make anyone's life easier or more productive.
And it's the one I'm most proud of. By a distance.
Because Modern Retro is pure taste. It exists because I thought it should exist. Every decision -- which brands to include, what the aesthetic should be, how the site should feel -- came from conviction rather than data. There was no audience to validate against. No metrics to optimise for. No brief to satisfy. Just a vision of what this thing should be and the stubbornness to keep pushing until it matched what was in my head.
In advertising, every piece of work is filtered through research, through client feedback, through the opinions of people who are paid to have opinions. The work gets smoothed and sanded until the sharp edges are gone. Modern Retro has all its sharp edges intact because nobody was there to sand them off. The absurdity of Peloton in 1974 is a sharp edge. The scoring system is a sharp edge. The whole concept is a sharp edge. And that's exactly why people respond to it.
Building from taste, not demand
The conventional wisdom is that you should build things people want. Find a problem, validate the demand, then build the solution. I understand this advice. In fifteen years of advertising strategy, I've given versions of it to hundreds of clients. Start with the audience. Understand the need. Build from insight.
Modern Retro is the argument against that wisdom. Or rather, it's the argument that there's another kind of building -- building from taste, from obsession, from a question that won't let go -- that produces things the demand-driven approach never would. Nobody was searching for "modern brands reimagined as 1970s retail stores." There's no keyword data that would have led anyone to this concept. No focus group would have surfaced it. It came from a specific person with specific taste browsing a specific corner of the internet at a specific moment.
That's not replicable. That's not scalable. That's not a framework you can teach in a workshop. And I think that's exactly why it works. In a world where everyone has access to the same tools and the same data, the things that stand out are the things that came from somewhere personal. From taste. From obsession. From a question that bothered one person enough to spend a week of evenings answering it.
The question that matters
Nobody asked for Modern Retro. That's the point. The best things I've built are the ones nobody asked for. The ones that came from curiosity rather than demand, from taste rather than research, from a late-night question rather than a morning standup. If I'd waited for someone to ask, it wouldn't exist. If I'd validated the idea first, I would have talked myself out of it. The only reason ninety-six brands have 1970s shopfronts on the internet is because I didn't ask permission and I didn't ask the market. I just started generating images one Tuesday night and couldn't stop.
That's the case for building from taste. Not instead of building from demand -- there's room for both. But alongside it. Make the useful things. Solve the real problems. And then, when it's late and the house is quiet and nobody's watching, make the thing that nobody asked for. That's usually the one that matters most.
