There's a special kind of violence that happens in focus group facilities — the kind with the one-way mirrors and the stale biscuits and the moderator who keeps saying "there are no wrong answers."
Ideas go in sharp. They come out smooth.
I've watched it happen repeatedly. A creative team spends weeks crafting something bold — something that makes them slightly uncomfortable, which is usually a good sign. It gets presented to eight strangers who've been paid ₹ 4,000 to give their opinions. Within twenty minutes, the edges are gone.
"I don't quite get it." "Could you make it clearer?" "My mother wouldn't understand this."
And so the team goes back. They clarify. They simplify. They make it more obvious, more palatable, more... nothing.
The problem isn't the people in the room — they're doing exactly what we've asked them to do. The problem is the belief that consensus equals quality.
Great ideas don't test well. They can't. Because great ideas require context, repeated exposure, cultural momentum. You can't test "Just Do It" by asking eight people in Varsova if they understand it.
Consider the Sony Walkman. When Sony tested it, people hated it. "Why would I want to listen to music alone?" The research said kill it. Akio Morita San ignored the research. The Walkman went on to define a generation.
Or the Herman Miller Aeron chair. In focus groups throughout the 1990s, people called it ugly, uncomfortable, too expensive. The mesh looked cheap. The design looked alien. Every tested metric said don't launch it. Herman Miller launched it anyway. It became the most successful office chair in history, generating over $2 billion in sales.
Or think about the original Volkswagen "Think Small" campaign — arguably one of the greatest advertising campaigns ever made. It positioned a small, ugly, foreign car in an American market obsessed with big, beautiful, domestic vehicles. We may only imagine what a focus group would have done to that?
I'm not suggesting research is useless. Research is excellent at telling you what's wrong with what already exists. It's terrible at validating what should exist next.
So what's the alternative? Test concepts, not executions. Test problems, not solutions. And when we've made something that makes us nervous — something that doesn't feel like everything else — maybe trust that feeling instead of asking eight strangers to make it safer.