The Design of Everyday Things

Contents

• 01The Psychopathology of Everyday Things
• 02The Psychology of Everyday Actions
• 03Knowledge in the Head and in the World
• 04Knowing What to Do: Constraints, Discoverability, and Feedback
• 05Human Error? No, Bad Design
• 06Design Thinking
• 07Design in the World of Business

The Psychopathology of Everyday Things

When you can't figure out a door, faucet, or shower, the failure isn't yours — it's the designer's. Good design rests on two pillars: discoverability (what can I do?) and understanding (what does it mean?). Norman introduces the foundational vocabulary — affordances, signifiers, mapping, feedback, conceptual models, and constraints — that recurs throughout the book.

Affordances

An affordance is a *relationship* between an object's properties and an agent's capabilities, not a property of the object alone. A chair affords sitting *for a person*; a thumbtack affords pinning *for someone with fingers*. Affordances exist whether visible or not — the design question is whether the affordance is perceivable, which is where signifiers come in.

Signifiers

A signifier is any perceivable cue that tells a person what action is possible and where to perform it. Where an affordance defines what *can* be done, a signifier communicates what *should* be done — a flat plate says "push," a vertical handle says "pull," a worn patch on a door says "people grab here." Norman insists designers should obsess over signifiers, because they are what users actually see.

Mapping

Mapping is the correspondence between controls and what they control. Natural mappings exploit spatial analogy or cultural standards — burner knobs arranged in the same pattern as the burners, a slider that goes up to raise the volume. Good mapping eliminates the need for labels and lets users act without thinking.

Feedback

Feedback is sensory information returned to the user confirming that an action has had an effect. It must be immediate, informative, and proportionate — silence leaves the user wondering whether to retry, while excessive noise becomes a nagging backseat driver. Poor feedback turns every interaction into a guess.

Conceptual Model

A conceptual model is a simplified mental explanation of how something works, usually inferred from the device itself — its appearance, controls, and feedback. When the model matches reality, users can predict the effect of their actions; when it doesn't, they grope in the dark or invent superstitions to fill the gap.

Constraints

Constraints are limitations on possible actions, deliberately built in to guide behavior and prevent error. Physical, cultural, semantic, and logical constraints narrow the space of plausible moves so the correct one stands out — they let novel users behave correctly without instruction.

Discoverability

Discoverability is whether you can figure out what actions are possible and the current state of the device just by looking. It emerges from the combination of the five fundamental psychological concepts — affordances, signifiers, constraints, mappings, and feedback — applied appropriately. Without it, every interaction starts from confusion.

Affordance

Relationship between an object's properties and an agent's capabilities that determines what actions are possible.

Signifier

Perceivable indicator that communicates appropriate behavior or location of action.

Anti-affordance

A property that prevents an action (glass affords transparency but anti-affords passage).

Mapping

Correspondence between controls and the things they control.

Feedback

Sensory return confirming an action's effect.

Conceptual model

Simplified mental explanation of how a thing works, used to predict outcomes.

Discoverability

Whether possible actions and current state can be determined by inspection.

Norman door

A door whose physical design signals the opposite of how it should be operated — the canonical example of bad signifiers.

Multiple choice

A designer claims that "a chair has the affordance of sitting." According to Norman's revised definition, what's the most accurate correction?

• AChairs have the affordance of sitting only when they include a visible cushion as a signifier.
• BAffordance isn't a property of the chair alone — it's a *relationship* between the chair's properties and the capabilities of a particular agent.
• CSitting is a signifier, not an affordance, because it tells the user what to do.
• DAffordances apply only to digital objects; physical chairs have constraints instead.

True / False

An affordance and a signifier are two names for the same thing: a perceivable cue that tells the user what to do.

• TrueTrue
• FalseFalse

Spot the issue

A new induction cooktop arranges its four burner knobs in a single straight row along the front edge, while the burners themselves sit in a 2x2 grid on top. Users keep turning on the wrong burner. Which design principle is most directly violated?

• AThe cooktop lacks sufficient affordances for turning the knobs.
• BThe knobs need stronger physical constraints to prevent accidental rotation.
• CNatural mapping is broken — the spatial layout of the controls doesn't correspond to the spatial layout of what they control.
• DFeedback is missing because the burners don't glow when activated.

Multiple choice

A smart thermostat displays no indication of whether it has received a tap, accepted the new temperature, or is still loading. Users keep tapping repeatedly and overshoot the target. What is the primary design failure?

• AThe thermostat lacks a clear conceptual model of how heating works.
• BThe thermostat fails to provide immediate, informative feedback confirming the action's effect.
• CThe thermostat has too many affordances competing for attention.
• DThe thermostat has too many constraints preventing user input.

The Psychology of Everyday Actions

Action unfolds in seven stages, from goal to evaluation. Two "gulfs" — execution and evaluation — describe the gaps users must cross, and good design narrows both. Norman also introduces three levels of cognition (visceral, behavioral, reflective) and the human tendency to wrongly blame oneself for design failures.

The Seven Stages of Action

Norman models action as seven stages: form a goal, plan the action, specify an action sequence, perform it, perceive the world, interpret the perception, and compare the outcome to the goal. The first four are execution; the last three are evaluation; the goal joins them. Most design failures show up as a stage that's hard to bridge.

Gulf of Execution

The Gulf of Execution is the gap between what the user intends and what the system allows them to do. The bigger the gulf, the more effort the user spends figuring out *how* to act. Good signifiers, constraints, mappings, and a clear conceptual model all narrow it.

Gulf of Evaluation

The Gulf of Evaluation is the gap between the system's actual state and the user's ability to perceive and interpret it. Crossing it requires good feedback plus a conceptual model that lets the user make sense of the feedback. When evaluation is hard, users can't tell whether their action succeeded.

Visceral, Behavioral, Reflective Levels

Norman distinguishes three levels of processing. Visceral is fast, prewired, gut-level — the immediate reaction to how something looks. Behavioral is learned skill, mostly subconscious, governing routine action. Reflective is slow, conscious, contemplative — where meaning, judgment, and the story we tell ourselves live. Good design engages all three.

Conscious vs. Subconscious Cognition

Most thought is subconscious — fast, automatic, pattern-driven, effortless but biased. Conscious thought is slow, deliberate, logical, and expensive. Practice migrates skills from conscious to subconscious, which is why experts can act fluently while novices must think through every step.

Emotion Is Part of Cognition

Emotion is not separate from thinking — it is how we evaluate what matters. Positive affect broadens thinking and supports creativity; negative affect narrows focus and supports detail work. A frustrating design generates negative affect, which paradoxically makes users *less* able to solve the problem they're stuck on.

People as Storytellers and False Self-Blame

Humans constantly construct causal stories to explain events, and they often blame themselves ("I'm bad with tech") for what is actually a design failure. This learned helplessness obscures the designer's responsibility and reinforces itself socially, which is why Norman insists on diagnosing root causes instead.

Goal

What a person wants to achieve; the starting point of action.

Plan

General strategy for reaching a goal.

Specification

Concrete action sequence chosen to execute the plan.

Perception

Sensing the state of the world after acting.

Interpretation

Making sense of the perception in light of expectations.

Gulf of Execution

Distance between intention and the means the system provides to act.

Gulf of Evaluation

Distance between the system's state and the user's understanding of it.

Visceral level

Automatic, prewired emotional response to appearance.

Behavioral level

Learned, skilled, largely subconscious action.

Reflective level

Conscious, contemplative meaning-making.

Learned helplessness

State where repeated failure makes a task feel impossible — often induced by bad design.

Feedforward

Information given *before* an action that answers "what can I do, and how?" — the counterpart to feedback.

Multiple choice

In Norman's seven stages of action, which three stages belong to the evaluation side (rather than execution)?

• AGoal, plan, specify
• BPlan, specify, perform
• CPerceive, interpret, compare
• DGoal, perform, compare

Spot the issue

A user installs a new home security app. They want to arm the system at night but can't figure out which of the six tiles on the home screen does that — every label is jargon and there's no clear path from "I want to arm it" to a control. Which gulf is widest in this scenario?

• AThe Gulf of Evaluation, because the user can't interpret what state the system ended up in.
• BThe Gulf of Execution, because the user can't translate their intent into an action the system accepts.
• CThe Gulf of Evaluation, because the home screen isn't giving feedback fast enough after each tile tap.
• DThe Gulf of Execution, because the user lacks the procedural skill that practice would have built.

Multiple choice

After tapping "send" in a chat app, a user sees no spinner, no checkmark, and no error — just the message sitting there. They can't tell whether it was sent, is sending, or failed. Which gulf is the design primarily failing to bridge?

• AThe Gulf of Execution — the user couldn't figure out how to act
• BThe Gulf of Evaluation — the user can't perceive or interpret the system's resulting state
• CThe Gulf of Execution — the send button lacks a clear affordance
• DThe Gulf of Evaluation — the user lacks the conceptual model needed to imagine network behavior

True / False

When a user repeatedly fails to operate a confusing coffee maker and concludes "I'm just bad with technology," Norman would say their self-assessment is essentially correct — most everyday failures are user error.

• TrueTrue
• FalseFalse

Knowledge in the Head and in the World

People act accurately despite imprecise memory because they combine what they remember with what the world shows them. Memory is unreliable for arbitrary details, so good design externalizes knowledge through visible cues and natural mappings, letting recognition substitute for recall.

Knowledge in the World vs. Knowledge in the Head

Knowledge in the world is perceivable from the environment — labels, layouts, signifiers. Knowledge in the head is what you've memorized. Behavior emerges from their combination, so a designer can dramatically reduce learning burden by putting more required knowledge into the world.

Precise Behavior from Imprecise Knowledge

People act precisely even with vague internal models because four factors compensate: information is in the world, great precision is rarely needed, physical constraints rule out wrong actions, and cultural conventions fill in the rest. You don't need to remember every detail of a coin to spot it in your change.

Declarative vs. Procedural Knowledge

Declarative knowledge is knowledge of facts and rules — easy to articulate, easy to teach in words ("stop at red lights"). Procedural knowledge is knowledge of how to perform a skill — largely subconscious, hard to verbalize, best taught by demonstration and practice (returning a tennis serve, riding a bike).

Working Memory Has a Tiny Capacity

Short-term (working) memory holds only a handful of items — Norman uses about five as a working number — and is easily disrupted by interruption. Designs that require users to remember more than that across a sequence of steps will fail in the wild, no matter how clearly each step is labeled.

Meaningful Material Is Easier to Remember

Arbitrary information (random PINs, isolated facts) is brutal to memorize and quickly forgotten; meaningful information that fits an existing schema is far easier to retain. Skilled performers chunk arbitrary material into meaningful units to beat working-memory limits.

Natural Mapping

A natural mapping is a layout where the spatial arrangement of controls mirrors the spatial arrangement of what they control. Stove knobs arranged in the same pattern as their burners is the classic example — the mapping is so obvious no label is needed.

Reminders Combine World and Head

Effective reminders carry both a signal (something needs doing) and a message (what specifically). Putting reminders in the world — a sticky note, a calendar alert, an object placed where you'll trip over it — offloads the cognitive work onto the environment.

Tradeoff Between the Two Knowledges

Knowledge in the world is easy to use and learn but clutters the environment and depends on it being present. Knowledge in the head is fast and unobtrusive but requires learning and is constantly being forgotten. Designers must consciously decide which to favor for each task.

Knowledge in the world

Externally available information embedded in signifiers, labels, and perceptible structure.

Knowledge in the head

Information stored in human memory — facts, rules, skills.

Declarative knowledge

Knowledge of facts and rules; verbalizable and transmissible.

Procedural knowledge

Knowledge of how to do something; tacit, acquired through practice.

Working memory (STM)

Limited-capacity transient store holding currently-used information.

Long-term memory (LTM)

Vast, durable but reconstructive memory store for experience.

Natural mapping

Use of spatial or cultural analogies between controls and effects.

Mnemonic

Memory aid that imposes structure on arbitrary material.

Reminder

External cue carrying both a signal that something is due and the message of what.

Cultural framing

Shared schemas that let members of a culture interpret situations and predict behavior.

Multiple choice

A kitchen designer wants users to operate a four-burner stove correctly the first time, without reading labels. Which approach best leverages knowledge in the world?

• APrint large, clearly-labeled stickers next to each knob explaining which burner it controls
• BArrange the four knobs in the same square pattern as the four burners so position alone reveals the mapping
• CShip a laminated quick-reference card that owners can keep in a drawer
• DTrain users with a short tutorial video the first time they turn on the appliance

True / False

Because skilled experts can chunk information, they have a larger working memory capacity than novices.

• TrueTrue
• FalseFalse

Spot the issue

A mobile banking app forces users to memorize an eight-digit transaction code displayed on one screen, then re-enter it on a different screen reached after three taps. Customer support is flooded with complaints that people "keep typing the wrong number." What's the underlying design problem?

• AThe font on the code screen is not large enough for older users to read
• BThe app exceeds working memory capacity and offers no way to carry the knowledge in the world across the intervening steps
• CThe transaction code is not a meaningful number, so users are unmotivated to remember it
• DUsers have not received enough training on how to use the multi-step flow

Multiple choice

Which scenario best illustrates the tradeoff between knowledge in the world and knowledge in the head?

• AA pilot memorizes emergency procedures because looking them up mid-emergency would be too slow, even though a checklist exists on paper
• BA toddler learns to walk by repeated practice rather than by being shown a diagram
• CA driver uses GPS turn-by-turn directions instead of a paper map
• DA student writes down a phone number rather than trying to remember it

Knowing What to Do: Constraints, Discoverability, and Feedback

When faced with an unfamiliar object, people work out what to do by exploiting constraints — physical, cultural, semantic, and logical. Norman shows how designers deploy these constraints, plus forcing functions, to make correct behavior obvious and incorrect behavior impossible.

Physical Constraints

Physical constraints are properties of the physical world that make wrong actions impossible — a large peg won't fit a small hole, a USB-A connector only goes in one way. They need no learning and can't be violated, which makes them the strongest constraint when available.

Cultural Constraints

Cultural constraints are conventions shared by a group: red means stop, scroll direction, which side of the road to drive on. They are powerful within a culture but vary across cultures, so global designs must respect them carefully or violate them on purpose.

Semantic Constraints

Semantic constraints come from the meaning of the situation — a motorcycle rider's helmet must face forward because that's where the eyes are. Once the meaning is clear, most configurations are ruled out without anyone teaching the rule.

Logical Constraints

Logical constraints are natural relations of inference: if you assemble a thing and have one piece left with one slot left, the piece goes in the slot. Natural mapping is a logical-constraint workhorse — the leftmost control belongs to the leftmost effect.

Forcing Functions

A forcing function is a strong constraint that blocks the next step until a prior step is correctly completed. ATMs that return your card before dispensing cash, microwaves that won't run with the door open — these prevent whole categories of error by design, not by reminder.

Interlocks, Lock-Ins, Lock-Outs

Three named forcing functions. An interlock orders operations (car must be in park before the key can come out). A lock-in keeps an operation engaged so it won't be inadvertently stopped (the "save your work?" dialog). A lock-out prevents entry into a dangerous state (stairwells that stop at the ground floor so fleeing people don't run into the basement).

Conventions and Standardization

Conventions are cultural constraints that reduce ambiguity over time; standards codify them so they need only be learned once (clock faces, traffic signals, QWERTY). Standardization is the design solution of last resort — used when no good natural mapping exists.

Discoverability Through Combined Constraints

Discoverability emerges when all four constraints, signifiers, affordances, mappings, and feedback line up. A well-designed object lets a first-time user behave correctly without instruction — and when this happens, it feels obvious, even though it took real work to engineer.

Physical constraint

Real-world limitation from physical properties (size, shape, fit).

Cultural constraint

Convention shared by a cultural group that limits appropriate actions.

Semantic constraint

Limitation imposed by the meaning of the situation.

Logical constraint

Constraint based on natural reasoning, such as one-to-one correspondence.

Forcing function

Constraint that prevents the next step until the current one is correctly completed.

Interlock

Forcing function that imposes order on operations.

Lock-in

Forcing function that keeps an operation engaged.

Lock-out

Forcing function that prevents entry into a dangerous space or state.

Convention

Cultural constraint that has become widely accepted, often invisible until violated.

Standard

Codified convention fixed across products and contexts.

Multiple choice

A USB-A connector only fits its port one way around, so a user physically cannot insert it incorrectly. Which kind of constraint is doing the work here?

• ACultural constraint
• BSemantic constraint
• CPhysical constraint
• DLogical constraint

Spot the issue

A startup designs a smart oven for a global market. The control panel uses a green LED to mean "heating element on" and red to mean "ready / safe to open." European testers report they keep getting burned because they assumed the red light meant "stop, don't touch." What constraint did the design get wrong?

• AIt violated a physical constraint — the lights should physically block opening the oven
• BIt violated a cultural constraint — red is conventionally read as "stop / danger," not "ready"
• CIt violated a logical constraint — there should be one LED per heating element, not two
• DIt violated a semantic constraint — the LEDs are too small to see from across the kitchen

Multiple choice

A microwave oven is wired so that pressing Start does nothing while the door is open — the machine simply refuses to run. What kind of design intervention is this, in Norman's vocabulary?

• AA lock-in that keeps the heating cycle engaged once started
• BAn interlock that orders operations so a hazardous step (radiation on) can't occur before a safe precondition (door closed)
• CA semantic constraint that relies on users understanding microwave physics
• DA natural mapping between the door state and the timer dial

True / False

A forcing function is essentially a polite reminder: it nudges the user to complete a step before moving on, but a determined user can always proceed anyway.

• TrueTrue
• FalseFalse

Human Error? No, Bad Design

What we call "human error" is almost always the predictable result of poor design colliding with normal human cognition. Norman draws on James Reason's taxonomy of slips and mistakes, the Swiss cheese model, and root-cause analysis to argue for resilience engineering and a no-blame approach to failure.

Slips vs. Mistakes

The two fundamental error categories. A slip is when the goal is right but the execution goes wrong — you intend one thing and your hand does another. A mistake is when the goal or plan itself is wrong — the execution may be flawless, but it accomplishes the wrong thing.

Two Kinds of Slips

Action-based slips: the wrong action is performed (pouring juice into your coffee cup instead of milk). Memory-lapse slips: a step is skipped or repeated because memory fails (leaving the original in the copier, forgetting to turn the stove off). Both are subconscious failures of skilled behavior.

Three Kinds of Mistakes

Rule-based mistakes diagnose the situation correctly but apply the wrong rule. Knowledge-based mistakes happen in unfamiliar situations where the actor lacks the knowledge to choose well. Memory-lapse mistakes occur when a goal or plan stage gets skipped entirely. The fix differs for each.

The Swiss Cheese Model

James Reason's model: organizational defenses are layered like slices of Swiss cheese, each with holes (weaknesses). An accident occurs only when the holes in all the slices momentarily align, letting a hazard pass straight through. Adding imperfect layers dramatically reduces risk because the holes rarely align.

Root Cause Analysis and the Five Whys

Root cause analysis is the discipline of refusing to stop at the proximate cause and digging until you reach the systemic source. The Five Whys (Toyota Production System) operationalizes this — keep asking "why?" until you reach a design or organizational cause rather than a person to blame.

Resilience Engineering

Rather than trying to prevent every possible error, resilience engineering designs systems to monitor for trouble, respond to disruption, and recover gracefully when failures do happen. The goal is robustness against inevitable error, not the fantasy of eliminating error.

Designing for Error

Assume errors will happen and build defenses: provide good feedback, make actions reversible (undo), confirm destructive operations, add forcing functions, run sensibility checks on inputs, and write error messages that help users recover rather than scold them for failing.

The Blame Culture Problem

Blaming the sharp-end operator discourages reporting, loses information about systemic weakness, and lets bad design persist. Aviation's no-blame reporting culture treats every error as data about how to redesign the system — and is why air travel is one of the safest things humans do.

Slip

Error where intent is right but execution deviates; almost always a subconscious failure.

Mistake

Error where the conscious goal or plan itself is wrong.

Capture slip

Slip where a frequent action "captures" and replaces a less common intended one.

Description-similarity slip

Slip caused when the description of the target is ambiguous enough to match a similar wrong target.

Mode error

Error from performing an action correct in one mode while the device is in another.

Postcompletion error

Memory-lapse slip where the main goal is completed but a cleanup step is forgotten.

Sharp end vs. blunt end

The operator at the moment of failure vs. the organization whose prior decisions shaped the conditions.

Forcing function

Constraint that blocks the next step until a prior condition is met.

Swiss cheese model

Reason's metaphor for layered defenses with imperfect holes that occasionally align.

Multiple choice

A nurse administers a dangerous overdose because two unrelated medication vials have nearly identical labels and sit next to each other on the shelf. Using Norman's taxonomy, this is best classified as which kind of error?

• AA knowledge-based mistake — the nurse lacked training on the correct medication.
• BA rule-based mistake — the nurse applied the wrong rule to a correctly diagnosed situation.
• CA description-similarity slip — the user's mental description of the target was ambiguous enough to match a similar wrong target.
• DA postcompletion error — a cleanup step was forgotten after the main goal was achieved.

True / False

According to the Swiss cheese model, the right way to prevent accidents is to find and remove the single defensive layer where the hole appeared.

• TrueTrue
• FalseFalse

Spot the issue

After a pilot lands with the gear up, an airline issues a memo reminding pilots to "double-check the landing gear before touchdown" and considers the incident closed. What's wrong with this response?

• AThe memo should have gone to maintenance crews instead of pilots.
• BIt stops at the proximate cause (the pilot) instead of digging for the systemic conditions that let the slip reach the runway — no root cause analysis was done.
• CIt blames the blunt end of the organization instead of the sharp-end operator.
• DIt adds a forcing function when a cultural constraint would have been sufficient.

Multiple choice

A product team adds a confirmation dialog for the "Delete Account" button, an undo window for sent messages, and rewrites error text to suggest what to try next. Which design philosophy does this best illustrate?

• AForcing function design — every dialog blocks the next step until prior conditions are met.
• BDesigning for error — assuming errors will happen and building reversibility, confirmation, and helpful recovery into the system.
• CKnowledge-in-the-head design — moving information out of the world and into trained user memory.
• DActivity-centered design — focusing on the activity rather than the user.

Design Thinking

Design thinking starts by questioning whether you're solving the right problem. Norman lays out human-centered design (HCD) — iterative, observational, prototype-driven — and the double-diamond model that captures its rhythm of diverging to explore and converging to commit.

Solving the Right Problem

The biggest failure in design is solving the wrong problem — taking the brief at face value instead of digging until the real need surfaces. Norman insists the stated requirement is rarely the actual problem; reframing the question is usually where the value lives.

Human-Centered Design

Human-centered design (HCD) starts with a deep understanding of the people who will use the product — their needs, capabilities, behaviors, and context. It is fundamentally iterative: solutions are tested with real users and refined, not specified up front and built once.

The Double-Diamond Model

Two successive diamonds. The first diverges to explore the problem space broadly, then converges on a well-defined problem statement. The second diverges to explore many solutions, then converges on the chosen one. Each diamond captures the rhythm of expansion and contraction that good design requires.

The Four Activities of HCD

HCD cycles through four activities. Observation: study users in their real context. Idea generation (ideation): produce many candidate solutions without premature criticism. Prototyping: build cheap, fast representations of the leading ideas. Testing: put prototypes in front of real users and learn from their behavior. Repeat until it converges.

Activity-Centered vs. Human-Centered

When a product is used by many different people across cultures, designing for an individual user becomes impossible. Activity-centered design focuses on the activity itself instead — cars are designed around driving, not around any particular driver — and trusts that people will adapt to fit a tool that fits its activity.

Observation and Applied Ethnography

Designers must watch potential users in the actual environment where the product will be used, because what people *say* they do diverges sharply from what they actually do. Applied ethnography is a shorter, focused version of academic ethnography, used to ground design in real behavior.

Ideation Rules

Generate many ideas, including wild or apparently bad ones. Don't criticize during generation. Question everything, especially the obvious. The cardinal rule is to defer judgment — quantity produces the variety from which a good solution can later be selected.

Rapid Prototyping and Testing With Five Users

Prototypes should be quick, cheap, and disposable — sketches, paper mockups, storyboards, Wizard-of-Oz simulations. Test with a *small* number of representative users, iterated repeatedly, rather than one big evaluation at the end; Norman cites Jakob Nielsen's finding that five users uncover roughly 85% of usability problems.

Design thinking

Approach emphasizing understanding the real problem before generating solutions, then iterating rapidly.

Human-centered design (HCD)

Process placing people's needs, capabilities, and behavior first.

Activity-centered design

Approach focused on the activity to be supported when users vary too widely.

Double-diamond model

Four-phase representation (Discover, Define, Develop, Deliver) of two diverge-converge cycles.

Ideation

Divergent-thinking phase of producing many candidate solutions without evaluation.

Prototype

Cheap, fast representation built to be tested and discarded.

Wizard of Oz technique

Prototyping method where a hidden human stands in for not-yet-built technology.

Applied ethnography

Field observation method adapted from anthropology, conducted in days or weeks.

Iteration

Repeated observe-ideate-prototype-test cycle central to HCD.

Spot the issue

A hospital executive tells a design team: "Our nurses keep forgetting to log patient vitals on the new tablet — design us a more aggressive reminder system." The team ships louder alerts and the problem gets worse. What did the team fail to do?

• AThey skipped ideation and went straight to prototyping.
• BThey didn't apply enough cultural constraints to the reminder UI.
• CThey accepted the brief at face value instead of solving the right problem — observing the nurses might reveal that the tablet is unreachable during procedures, not that nurses need reminding.
• DThey used activity-centered design when they should have used human-centered design.

Multiple choice

A team kicks off a project by spending two weeks interviewing users and shadowing them in the field, then narrows their notes to a single problem statement. They then sketch dozens of competing solutions before picking one to prototype. Which model best describes this rhythm?

• AThe seven stages of action
• BThe Swiss cheese model
• CThe double-diamond model — diverging then converging on the problem, then diverging then converging on the solution
• DThe Gulf of Execution

True / False

To get statistically meaningful usability data, designers should test each prototype with at least fifty users before iterating.

• TrueTrue
• FalseFalse

Multiple choice

A team designs a public-transit ticket kiosk that must serve tourists, commuters, elderly riders, children, and visually impaired users — people whose needs vary wildly. According to Norman, what is the most appropriate design approach?

• AStick strictly to human-centered design and create a separate persona for every user segment.
• BUse activity-centered design — design around the activity of buying transit tickets, and trust that diverse users will adapt to a tool that fits the activity well.
• CSkip observation and rely on engineering specs, since users this varied can't be studied meaningfully.
• DUse the Wizard of Oz technique throughout production to dodge the variation problem.

Design in the World of Business

Real products are shaped by competition, marketing, cost, and legacy compatibility — pressures that produce featuritis and Norman doors. Norman distinguishes incremental from radical innovation, examines the marketing-engineering divide, and closes with the designer's moral obligations in a world of smart, networked products.

Featuritis (Creeping Featurism)

In competitive markets, every product feels pressure to match or exceed rivals' feature lists, producing featuritis — the disease of ever-expanding functionality that swells products with rarely-used capabilities, complicates the interface, raises cost, and degrades usability. Features are easy to add and excruciatingly hard to remove once customers depend on them.

The Norman Door

A door whose physical design tells you to do the opposite of what's actually required, or gives no signal at all about whether to push or pull. The Norman door is the canonical example of a discoverability and signifier failure — and the term has entered popular usage as the shorthand for design that blames the user.

Incremental vs. Radical Innovation

Incremental innovation is slow, continuous refinement of an existing product through small improvements — by far the most common form, and the source of nearly all real progress. Radical innovation starts over with a wholly new concept or technology — rare, risky, and frequently failing, but occasionally creating an entirely new product category.

Why Radical Innovation Usually Fails

Most attempts at radical innovation fail. Norman cites figures suggesting roughly 90% of radical new-product ideas don't succeed, and the rare winners often take decades to reach acceptance. Companies that bet entirely on radical innovation tend to be outperformed by those that combine occasional radical bets with disciplined incremental improvement.

The Marketing-Engineering Divide

Marketers focus on what makes customers buy — often a long, comparable feature list and a competitive price. Engineers and designers focus on what makes the product work well in daily use. These are different problems, often pulling in opposite directions, and reconciling them is a central business-design challenge.

Legacy and Standardization Lock-In

Products must remain compatible with existing infrastructure, conventions, and learned user skills — the QWERTY keyboard is Norman's classic example of a suboptimal but locked-in standard. Standards survive not because they're good but because the coordination cost of switching is prohibitive; designers must accept and work within them.

Smart, Connected Products

Everyday objects are becoming intelligent, networked, and continuously updated. This raises new design challenges — privacy, autonomy, machine-human communication, and the temptation to add features simply because connectivity allows them — and it makes good HCD *more* important, not less.

The Designer's Moral Obligation

Designers' decisions shape the everyday experience of millions of people, so the responsibility is ethical. Good design serves human needs, accommodates human limits, and respects human dignity; bad design — whether from laziness, business pressure, or featuritis — makes daily life harder and shifts the blame onto the user.

Featuritis (creeping featurism)

Relentless tendency to add features that bloat the product, complicate the interface, and degrade usability.

Norman door

Door whose physical design provides incorrect or absent signifiers about how to operate it.

Incremental innovation

Slow, continuous small-step improvement of an existing product.

Radical innovation

Breakthrough creating a new product category; rare, high-risk, mostly failing.

Legacy problem

Constraint that new designs remain compatible with installed infrastructure and learned skills.

QWERTY effect

Phenomenon of a widely adopted standard persisting despite being inferior.

Marketing requirements vs. usage requirements

Distinction between criteria driving purchase and criteria determining daily-use quality.

Smart product

Everyday object augmented with sensing, computation, and connectivity.

Designer's moral obligation

Ethical duty arising from the scale at which design decisions shape lives.

Spot the issue

A consumer-electronics company benchmarks its new smart speaker against three competitors and concludes that to win shelf placement it must match every feature on every rival's box — voice training, multi-room sync, sleep timer, smart-home hub, language packs, recipe mode, and a karaoke mode. The shipped product is hard to set up and users complain it's confusing. Which pattern is this?

• AA failure of activity-centered design — they should have studied karaoke users.
• BFeaturitis — competitive pressure produced a bloated feature list that complicated the interface and degraded usability.
• CThe QWERTY effect — they were locked into a legacy standard.
• DA radical innovation failure — they tried to ship a wholly new product category at once.

Multiple choice

According to Norman, where does *most* real progress in product design actually come from?

• ARadical innovation — bold leaps that create entirely new product categories.
• BMarketing-driven feature races that force engineering to keep pace.
• CIncremental innovation — slow, continuous refinement of an existing product through many small improvements.
• DStandardization committees that fix conventions across the industry.

True / False

The reason the QWERTY keyboard layout is still standard is that decades of usability research have confirmed it is the most efficient arrangement for typing English.

• TrueTrue
• FalseFalse

Spot the issue

A startup pitches its appliance with a stunning, technically novel touch interface that has no precedent in any kitchen product. Despite glowing reviews and a few enthusiastic early adopters, sales collapse within eighteen months and the company folds. Using Norman's framing, what's the most likely diagnosis?

• AThe team failed at incremental innovation by ignoring competitors' feature lists.
• BThey attempted a radical innovation — which fails roughly 90% of the time and, even when it succeeds, often takes decades to reach mainstream acceptance.
• CThey violated a forcing function by letting users open the appliance mid-cycle.
• DThey confused activity-centered with human-centered design.