The Inmates Are Running the Asylum: Why High-Tech Products Drive Us Crazy and How to Restore the Sanity
Source: https://www.informit.com/store/inmates-are-running-the-asylum-why-high-tech-products-9780672326141 ↗
Alan Cooper invented Visual Basic's interaction model and then spent the rest of his career arguing that engineers should not design the products they build.
This book introduced personas as a design method — not the diluted marketing personas used today, but richly researched archetypes grounded in behavioral patterns.
Cooper's core claim is that software is hostile to normal people because the people who build it optimize for their own mental models, and that design must be an independent discipline with authority in the development process.
The book was published in 1999, when this was a genuinely controversial position in the industry, and Cooper fought the cultural war for design's seat at the table when almost no one else was fighting it.
The organizational argument — that interaction design must happen before engineering begins — remains uncomfortable and largely unimplemented in most product teams.
Central argument
Cooper argues that software products are systematically hostile to ordinary users because engineers design them to reflect their own mental models and edge-case thinking rather than the goals of actual users. His core thesis is that this is not a skill gap but a structural problem: when the people who build a product also design it, they will inevitably optimize for implementation logic over human behavior. The solution he proposes is to establish interaction design as an independent discipline with authority that precedes engineering — and to base that design on personas, behavioral archetypes derived from research, which force teams to reason about specific human goals rather than abstract feature sets.
Critique
Cooper's argument rests on a clean separation between design and engineering that assumes these are distinct cognitive modes best kept in separate roles and sequential phases — but this organizational model struggles to account for the tight feedback loops that make modern iterative development effective. The book treats engineering involvement in design almost exclusively as a contaminant, which risks dismissing the legitimate domain knowledge engineers bring to understanding what is technically feasible or where clever constraints can drive genuinely better user experiences. There is also a tension Cooper never fully resolves: personas are only as good as the research behind them, yet the book says relatively little about the conditions under which persona construction goes wrong or gets captured by organizational bias.
Why it matters for product
For a CPO, the book's sharpest edge is its organizational argument: if interaction design does not happen before engineering begins, the product's behavior will be determined by implementation decisions made under pressure, not by deliberate design choices — a dynamic that is still the default in most roadmap-driven teams. Cooper's framing helps diagnose a common failure mode where discovery and design are nominally present in the process but lack the authority to stop or reshape engineering work that has already started. The persona method, understood in its original form as a behavioral archetype rather than a demographic profile, also offers a useful corrective to product strategies built around market segments rather than the specific jobs and frustrations of identifiable human beings.