The book ends by changing the question

For most of the book, Norman asks us to understand the systems we have designed.

By the end, we should be able to

1. Compare top-down and bottom-up routes to change.
2. Explain why warnings and technology are necessary but insufficient.
3. Apply incremental, humanity-centered action to an HCI problem.
4. Connect action to the book’s four themes: artificial, meaningful, sustainable, and humanity centered.

The scale of the problem exceeds personal choice

• Individuals can change their own behavior.
• Planetary problems also involve infrastructure, regulation, industry, and entrenched interests.
• Large changes create immediate costs, political conflict, and resistance.

Top-down action changes the rules

• Governments establish policy, law, incentives, and enforcement.
• Large organizations can redirect money and institutional priorities.
• Common standards can distribute the cost of change across an industry.
• Bureaucracy, political turnover, and tightly coupled megaprojects can slow progress.

Seat belts show both the power and delay of policy

• Seat belts appeared in limited use in the 1930s.
• Installation in new US vehicles became mandatory in 1966.
• Usage requirements spread across the states only decades later.

Bottom-up action builds pressure

• Individuals organize around shared concerns.
• Small groups make problems visible and demonstrate alternatives.
• Coalitions convert dispersed concern into social, market, and political pressure.
• Local action can spread without waiting for a single global plan.

Top-down and bottom-up are not opponents

Discussion: map the leverage

In pairs, identify:

1. One action a design team can take now.
2. One action that requires organizational authority.
3. One action that requires policy or an industry standard.
4. One group that must help define what “better” means.

Coalition-building turns fragments into force

• Many groups already pursue compatible goals.
• Working alone duplicates effort and limits visibility.
• Shared methods, results, and failures help others move faster.
• Partnerships amplify influence without requiring one centralized megaproject.

Transition creates real local harm

• A sustainable long-term change may eliminate present-day jobs.
• New requirements can raise costs for firms and consumers.
• Generic calls for “retraining” may ignore workers’ existing abilities and identities.
• A credible transition must address livelihoods, agency, and fairness.

Warning everyone is only the first step

• Recognition must precede coordinated action.
• Scientific evidence does not automatically become policy.
• Repeated warnings can coexist with decades of delay.
• Institutions may agree publicly and fail to implement privately.

Technology solves only part of a sociotechnical problem

Human behavior is infrastructure

• Policy is enacted—or blocked—through human institutions.
• People judge distant benefits against immediate losses.
• Trust, identity, emotion, and perceived fairness shape adoption.
• A technically sound intervention can fail when its social conditions are ignored.

Complex change needs many kinds of expertise

• Technological
• Economic
• Political and legal
• Organizational and commercial
• Social and behavioral
• Design

Wicked problems improve; they do not resolve neatly

• The problem cannot be defined once and for all.
• Causes and consequences cross institutional boundaries.
• Each intervention changes the situation being addressed.
• No single solution satisfies every stakeholder or value.

Start with a minimum viable outcome

1. Find a tractable cause whose improvement would matter.
2. Make a small intervention with the affected community.
3. Observe consequences, including unintended ones.
4. Preserve what works; revise or stop what does not.
5. Use the learning to make the next step easier.

Loosely coupled action can scale without becoming rigid

• Many local groups pursue related outcomes.
• Each adapts to its own context.
• Groups share methods, evidence, and failures.
• Coordination makes the collective effort visible to decision makers.

Application: design an action, not just an interface

Design a minimum viable outcome:

• What improvement should workers experience?
• Who must participate in defining it?
• What is the smallest meaningful change to try?
• What evidence would tell you whom it helped—or harmed?
• What organizational or policy change might follow?

Artificial means changeable

• Objects, institutions, metrics, customs, and technologies have been designed over time.
• Path dependence makes them feel natural and difficult to question.
• Their history constrains the present—but does not make the present inevitable.

Meaningful means measuring what matters

• A single number can conceal a multidimensional reality.
• Cost, productivity, and efficiency do not exhaust human value.
• Qualitative evidence, stories, and dashboards can reveal different dimensions.
• Measures should guide attention without erasing lived experience.

Sustainable means designing out waste

• Treat waste and pollution as design failures, not inevitable by-products.
• Prioritize repair, reuse, regeneration, and return.
• Reduce the need for energy and material before optimizing supply.
• Attack causes in production systems, not only symptoms at disposal.

Humanity centered means designing with communities

• Expand the boundary from an individual user to humanity, other living things, and the environment.
• Replace prescriptive expertise with facilitation and partnership.
• Respect local knowledge, disagreement, and self-determination.
• Judge success over longer time horizons and wider systems.

Action joins the four themes

Exit reflection

Choose an interactive system you know.

1. What harmful condition does it treat as inevitable?
2. What outcome would be meaningful to the people most affected?
3. What is one minimum viable outcome you could pursue?
4. Which bottom-up and top-down actors would have to work together?