The Human-Centered Design Process

Definition

Human-centered design (HCD) is a procedure for ensuring a product actually meets human needs, run through four iterated activities: observation, idea generation, prototyping, and testing. Norman frames the whole process using the British Design Council's 2005 "double-diamond" model: two separate diverge-converge cycles, one to find the right problem (expand broadly, then converge on a problem statement) and a second, independent one to find the right solution (expand broadly again, then converge on a proposal) — rather than jumping from a given problem statement straight to a single solution.

In the Book

Chapter 6 opens by insisting designers treat "the original problem as a suggestion, not as a final statement" — echoing the Five Whys technique from Chapter 5 — and resist converging on a solution before they've determined the real underlying problem. The double-diamond (Figure 6.1) splits into "discover" and "define" (diverging then converging on the problem) followed by "develop" and "deliver" (diverging then converging on the solution). Norman then walks through the four HCD activities in order: observation, done through "applied ethnography" — watching prospective users in their actual homes, commutes, and workplaces rather than relying on self-report, because people are unreliable narrators of their own needs; idea generation; prototyping, kept cheap and fast specifically to enable testing early; and testing with real people, feeding back into another round of observation. He acknowledges the process looks chaotic to outsiders — a product manager who assigned a scoped problem finds designers instead traveling the world and generating impractical half-formed ideas — but argues this apparent randomness follows disciplined method, and that the practical fix is to give teams freedom within firm deadline and budget constraints, since "there is nothing like a firm deadline to get creative minds to reach convergence."

Why It Matters

Splitting "solve the problem" into a distinct problem-finding phase and a distinct solution-finding phase blocks the most common failure mode in design and strategy work generally: converging immediately on the first plausible solution to a problem statement that was itself never interrogated. The iterate-through-four-stages structure also generalizes past product design to any situation requiring understanding before intervention — policy, organizational change, research — where premature convergence produces a well-executed answer to the wrong question.