In any company I've worked at, the best products never came from a Gantt chart. They didn't emerge from a quarterly planning session or a neatly prioritised backlog. The best products, the ones people actually remember, came from someone who couldn't stop thinking about an idea. Someone who was supposed to be doing something else entirely, but kept circling back to a concept that wouldn't leave them alone.
I've spent fifteen years in advertising and strategy. I've sat through hundreds of workshops designed to generate ideas on demand. Some of them produced good work. But the ideas that truly stuck, the ones that shipped fast and resonated hard, always had a different energy. They weren't assigned. They were compulsive.
That compulsion is something worth paying attention to. Not as a quirk. As a signal.
Obsession as a product filter
Most ideas are forgettable. That's fine. That's how it should be. You encounter dozens of interesting concepts in any given week. Most of them are interesting for about ten minutes. They catch your attention, you turn them over briefly, and then they dissolve. The ones that matter are the ones that refuse to dissolve. They keep returning. They start developing texture and detail without you consciously working on them. They start growing limbs.
This is not a productivity hack or a creative framework. It's a filter. The ideas that survive a full day of neglect, of competing priorities, of every reasonable excuse to forget them, are the ones worth building. If you can walk away from an idea and it stays behind, it was just a thought. If the idea follows you, it's something more.
I've now shipped over twenty-four products, all built with Claude Code, all driven by this same filter. Every single one started as a thought that wouldn't go away. Not one of them originated in a sprint planning meeting.
How Modern Retro happened
I was browsing old photographs of 1970s high streets. Just looking. Not working. And a question appeared: what would a modern brand look like in that setting? What would Nike look like as a cramped retail shop in 1974, with orange signage and brown trim and slightly wonky hand-painted lettering?
That question would not go away. Within hours I was mentally cataloguing which brands would be the most visually interesting, what the right aesthetic parameters would be, how the scoring system should work. I hadn't opened a laptop. I hadn't written a thing. But the product was already taking shape.
Within a week, Modern Retro was live with ninety-six brands, a five-factor scoring system, and a visual identity so specific that people assumed a design studio had built it. The reality was simpler: the idea had grabbed me so completely that building it felt less like a choice and more like a release of pressure. That's what real creative obsession feels like. Not discipline. Inevitability.
Modern Retro has since become my most obsessively iterated project. Not because I planned it that way, but because the idea keeps generating new questions. What about this brand? What about this era? What about a different country? When the concept itself is generative, the obsession sustains itself.
How First Out grew
First Out started as the simplest possible product. One question: which Tube exit should I use? It was a small annoyance I experienced constantly, and the existing tools for solving it were clunky or non-existent. So I built a straightforward exit finder.
Then the obsession deepened. Every time I used the app, I noticed something else it could do. Step-free routes. Interchange navigation. Live disruption data. Accessibility features. Buggy-friendly exits for parents. Each addition came from the same place: a persistent thought that wouldn't let go until I'd built it.
First Out now has over ninety-five features, a PWA, and a native iOS wrapper. It went from a single-purpose tool to a comprehensive London transport companion. None of that growth was planned. All of it was obsessive. Each feature was a thought that followed me until I shipped it.
Why this matters for how we build
The standard product development process is designed to be rational. Research, validate, prioritise, plan, build, test, ship. That process has its place. It prevents waste. It manages resources. It keeps teams aligned.
But it also systematically filters out the most interesting ideas. The strange ones. The ones that don't fit neatly into an existing category or target a well-defined user need. The ideas that someone can't stop thinking about, even though they can't yet articulate why.
Those are exactly the ideas that create new markets, new aesthetics, new categories. They're the ideas that feel obvious in retrospect but couldn't have been predicted in advance. And they almost always come from obsession, not process.
I'm not arguing against process. I'm arguing that the best processes make room for obsession. They recognise that someone who can't stop thinking about an idea is showing you something valuable. That persistence isn't a distraction from real work. It often is the real work, trying to surface.
Following the signal
The tools we have now make obsession more productive than ever. With AI-assisted development, the gap between "I can't stop thinking about this" and "it's live" has collapsed from months to hours. That changes the calculus entirely. The cost of following an obsessive idea used to be enormous. Now it's almost nothing. The cost of ignoring one, of letting a genuinely compelling concept fade because it wasn't on the roadmap, is the only real risk left.
I've built twenty-four products in the time it used to take to get a single deck approved. Not all of them are world-changing. But every single one came from following an idea that wouldn't let go. And the best ones, the ones that people actually use and talk about, were the ones where the obsession ran deepest.
Pay attention to the ideas that won't leave you alone. They're not distractions. They're the most honest signal you have about what's worth building next.
