The Design of Everyday Things — Summary: Core Ideas and Lessons

When you push a door that should be pulled — that is not you being confused. That is the designer failing you. The Design of Everyday Things is the book that explains why, and gives you a mental toolkit to never design something confusing again.

About the Book

Don Norman is a cognitive scientist and UX pioneer who spent decades at Apple, IDEO, and university research labs studying why people struggle with ordinary objects. First published in 1988 (revised in 2013), this book is addressed to anyone who creates things other people use — designers, engineers, product managers, developers — and to anyone who has ever felt stupid using something that should have been obvious.

Ad Slot (activate after AdSense approval)

The central argument: usability problems are almost always design problems, not user problems.

Core Ideas

1. Affordances — Objects silently tell you how to use them

The Idea
An affordance is a relationship between an object and a person — it is what the object allows you to do. A chair affords sitting. A button affords pressing. A handle affords pulling. The key insight is that affordances are not painted on — they emerge from the physical properties of the object itself. A flat plate on a door affords pushing because there is nothing to grab. A round knob affords turning because your hand naturally wraps around it.

Simple Example
You walk up to a door. If it has a bar across the middle, you push without thinking. If it has a vertical handle on the right, you grab and pull. You did not read any instructions — the shape told you what to do. Now imagine that door has a pull handle on a door that must be pushed. Your hand grabs, you pull, nothing happens, you feel confused. The affordance lied. That is bad design.

🧠 Memory Anchor — The Silent Instruction
Every object is giving you instructions through its shape. A good affordance whispers the right instruction. A bad affordance shouts the wrong one. Next time you struggle with an object, ask: What is this shape telling me to do?

Good design makes the correct action feel obvious before you even think about it.

2. Signifiers — Affordances need to be visible to work

The Idea
An affordance tells you what is possible. A signifier tells you where and how to act. Norman added this concept in the 2013 revision because affordances alone are not enough — you need a visible signal pointing to the action. A touch screen affords tapping anywhere, but a button-shaped rectangle is the signifier that says tap here. Without signifiers, affordances are invisible, and invisible affordances might as well not exist.

Simple Example
Think about a door with no handle, no bar, and no label — just a flat glass panel. The glass technically affords pushing, but there is no signifier. So people push on the frame, the hinge side, the corners. Everyone looks confused. Now draw a small circle where the push point is. Suddenly everyone touches the right spot. That circle is a signifier. It costs nothing and solves everything.

🧠 Memory Anchor — The Treasure Map Rule
An affordance is the treasure. A signifier is the X on the map. Having treasure buried with no map is useless. Always mark the X.

It is not enough for the action to be possible — the design must show people where to look and what to do.

3. Mapping — Controls should mirror what they control

The Idea
Mapping is the relationship between a control and the thing it affects. When the layout of the controls matches the layout of the effects, using the object is effortless. When they do not match, every interaction requires guesswork. Norman calls a perfect match a natural mapping — and says it should be the default, not a luxury.

Simple Example
Imagine a stovetop with four burners — front-left, front-right, back-left, back-right. The controls are a row of four identical knobs in a line. Which knob controls which burner? You have no idea. You guess, turn the wrong one, and the back burner heats up when you wanted the front one. Now imagine the four knobs are arranged in the same square pattern as the burners. Suddenly you never guess wrong. Same four knobs. Completely different mapping.

🧠 Memory Anchor — The Remote Control Test
Point your TV remote at the screen. The up button should make the volume go up on screen. When controls mirror their effects, you never have to think. When they do not, you always do.

If users need a legend to figure out your controls, your mapping failed.

4. Feedback — People need to know something happened

The Idea
Feedback is the signal that confirms your action worked. When you press a button, you need to hear a click, see a light change, or feel a vibration. Without feedback, you do not know if your action registered. You press again. And again. You over-correct. Or you give up thinking the thing is broken. Feedback must be immediate, clear, and informative — it tells you what happened, not just that something happened.

Simple Example
You submit a form on a website. Nothing changes. No spinner, no message, no color change. Did it work? Did the internet cut out? Did you click the wrong button? You click submit again. Now you submitted twice. If the site had shown a simple spinning icon, you would have waited. One line of feedback prevents a cascade of errors.

🧠 Memory Anchor — Silence is Broken
In conversation, when you say something and the other person stares blankly, you repeat yourself, louder. Good design is never silent. Every action deserves an acknowledgement.

No feedback is not neutral — it is a signal that something went wrong.

5. Conceptual Models — Everyone builds a mental picture of how things work

The Idea
A conceptual model is the story you tell yourself about how something works. You do not need to understand the engineering — you build a simplified model that lets you predict what will happen when you act. Good design gives you the right model automatically. Bad design gives you nothing, so you invent a wrong model and make constant mistakes.

Simple Example
Most people believe that pressing the close door button in an elevator closes the door faster. In reality, in most modern elevators this button is legally disabled. But because the button exists and the door closes shortly after pressing it, your brain builds the model: button closes door. The model is wrong but feels correct. Now think how many wrong models you hold about apps and appliances — and how often those wrong models make you do the wrong thing.

🧠 Memory Anchor — The Three Models (DUS)
Designer model: how it actually works. User model: how you think it works. System image: what the product shows you. Good design makes D = U by making the system image crystal clear. When all three align, design works.

Your job as a designer is to give users the correct mental model through what the product looks and feels like — not through a manual.

6. The Seven Stages of Action — Every human action follows the same cycle

The Idea
Norman maps out how humans do anything. Every action — from turning on a lamp to booking a flight — moves through seven stages: form a goal, make a plan, decide the specific action, execute it, perceive what happened, interpret the result, and compare it to your goal. Design can break at any of these seven points. Understanding where in this cycle something goes wrong tells you exactly what to fix.

Simple Example
You want to turn the lights off before sleeping (goal). You decide to use the switch by the door (plan). You reach for the switch (specify). You flip it (execute). The room is still bright (perceive). You flipped the wrong switch — that one controls the fan (interpret). You have not reached your goal yet (compare). Why did you flip the wrong switch? Both switches look identical. Fix the mapping, fix the error at stage three.

🧠 Memory Anchor — GPS EPIC
Goal, Plan, Specify, Execute, Perceive, Interpret, Compare. Think GPS Epic — the route your brain takes every time it does anything. When someone fails to use your product, trace which stage broke.

Design failures happen at specific stages — diagnosing which stage breaks tells you exactly what to fix.

7. The Gulf of Execution and the Gulf of Evaluation — The two gaps that make design hard

The Idea
The Gulf of Execution is the gap between what you want to do and how to do it with the product. The Gulf of Evaluation is the gap between what happened and whether it matched what you expected. Both gulfs must be narrow for design to work. Wide gulfs mean users spend mental energy just figuring out the tool — energy that should go toward their actual goal.

Simple Example
You want to bold some text in an unfamiliar app. You look at the toolbar — fifteen unlabelled icons (Gulf of Execution: where is bold?). You click one. Nothing obvious changes (Gulf of Evaluation: did that work?). You give up. Now imagine the toolbar shows a B icon that turns blue when text is bold. Gulf of Execution: zero. Gulf of Evaluation: zero. Same app, opposite experience.

🧠 Memory Anchor — Two Rivers to Cross
Imagine standing on one bank (your goal) and the product on the other. The Gulf of Execution is the river you cross to act. The Gulf of Evaluation is the river you cross to understand what happened. Good design builds two short bridges. Bad design leaves you swimming.

Narrow both gulfs and users feel effortlessly in control; widen either one and frustration immediately follows.

8. Constraints — Good design limits wrong actions before they happen

The Idea
Constraints are deliberate limitations that prevent you from doing the wrong thing. Norman identifies four types: physical (a USB plug only fits one way), cultural (red means stop everywhere), semantic (you put a lid on a pot, not under it — context makes one option obvious), and logical (if three of four screws are in, the fourth slot is obviously for the last screw). Constraints reduce the chance of error without requiring instructions.

Simple Example
LEGO bricks are a masterclass in constraints. A 2x4 brick can only attach to other bricks in specific ways — the bumps physically prevent wrong connections. You never need to read instructions. Compare that to flat-pack furniture with identical-looking parts and no physical constraints — you bolt the shelf upside down and only notice when the drawer hits the floor.

🧠 Memory Anchor — P-C-S-L (Please Constrain Sensibly, Logically)
Physical, Cultural, Semantic, Logical — the four constraint types. Before finalising any design, ask: which of these can I use to make the wrong action impossible?

The best error message is the one never shown — because the design made the error impossible.

9. Knowledge in the World vs Knowledge in the Head — Stop making people memorise things

The Idea
Knowledge in the head is what you have memorised — your PIN, keyboard shortcuts, app gestures. Knowledge in the world is information built into the environment — the H on a hot tap, the raised dot on the F key, the colour coding on a cable. Good design externalises knowledge — it puts the information where you need it so you do not have to memorise it.

Simple Example
A gas stove with burner labels on each knob versus a stove where you must remember which knob is which. The labelled stove lets you cook safely on day one. The unlabelled one takes weeks of practice to feel confident. One uses the world. The other taxes your memory.

🧠 Memory Anchor — The Cheat Sheet Rule
If users need a cheat sheet to use your product regularly, you have put too much knowledge in their heads and not enough in the world. Every sticky note on a monitor is a design failure wearing a Post-it costume.

Design that requires memorisation punishes new users and fatigues experienced ones — put the knowledge where the action is.

10. Human Error Is Really System Error — Stop blaming users

The Idea
When someone presses the wrong button, deletes the wrong file, or uses something backwards, the instinct is to say user error. Norman argues this is almost always wrong. When the same mistake happens repeatedly across many different users, the cause is the design, not the humans. Norman categorises errors as slips (you knew what to do but did the wrong thing automatically) and mistakes (you had the wrong plan from the start). Both are designable problems.

Simple Example
In aviation, near-identical incidents are logged in shared databases and the cockpit layout is redesigned. In hospitals, the same errors happen silently and the surgeon is blamed. Aviation crashes dropped dramatically. Hospital errors did not. The difference is whether you blame the person or redesign the system.

🧠 Memory Anchor — Slips vs Mistakes
Slip: Right goal, wrong action. (You meant to save but pressed delete.) Fix by making the wrong action harder. Mistake: Wrong goal from the start. (You thought delete would move it to trash.) Fix by improving the conceptual model. Different error, different fix.

If the same mistake happens repeatedly, the design is the problem — redesign the system before you retrain the human.

11. Design for Error — Assume something will go wrong and make it recoverable

The Idea
No matter how good the design, humans will make errors. Design for error means: assume errors will happen, make them hard to cause, easy to detect, and easy to reverse. Undo is a perfect example — it does not prevent the error, but it makes recovery effortless. Confirmation dialogs for irreversible actions, clear error messages that explain what happened and how to fix it — these are features, not afterthoughts.

Simple Example
Gmail's Undo Send gives you a 5-second window to cancel an email after sending. It has saved millions of sent-to-the-wrong-person disasters. Google did not make sending emails harder — they made the error recoverable. Compare that to a bank transfer app with no confirmation step and no cancel option. One design plans for error. The other punishes it.

🧠 Memory Anchor — The Safety Net Rule
Every trapeze artist works above a safety net — not because they plan to fall, but because falls happen regardless. Design your safety net before your users perform. The net does not make the act worse; it makes the whole show possible.

Designing for error is not pessimism — it is the most optimistic thing you can do, because it trusts that people will keep trying after a mistake.

12. The Paradox of Technology — More features create more confusion

The Idea
As technology improves, designers add more features. More features mean more controls. More controls mean more complexity. More complexity means more errors and frustration. Norman calls this the paradox of technology — the advances meant to help us end up overwhelming us. The solution is not fewer features, but better design that hides complexity until it is needed (progressive disclosure) and makes defaults intelligent.

Simple Example
The average microwave has a clock, timer, power levels, defrost, sensor cooking, and twelve preset buttons. Most people use three things: the number pad, start, and add-30-seconds. The other fifteen features sit unused. A well-designed microwave shows the three common things prominently and hides the rest behind a single more options menu. Same features. Radically simpler experience.

🧠 Memory Anchor — The Iceberg Principle
Show 10% of features on the surface — the ones 90% of people need 90% of the time. Hide the other 90% below the waterline. Users who need advanced features will dive. Users who do not will never know the iceberg is there.

Every feature you add is a tax on every user — make sure the tax is worth it and hide it for those who do not need it.

13. Discoverability — You should be able to figure out any product just by looking at it

The Idea
Discoverability is the degree to which a product reveals what it can do and how to do it — without a manual, tutorial, or help article. If users cannot figure out the basic functions of your product by exploring it for two minutes, discoverability has failed. Discoverability comes from combining good affordances, clear signifiers, correct mappings, and immediate feedback — all of Norman's principles working together.

Simple Example
Plug in a new USB keyboard you have never used. Within 60 seconds, you are typing. The keys are labelled. The layout matches every keyboard you have seen. Pressing a key gives tactile and auditory feedback. Nothing required explanation. Now try to use a smart home device with no labels, a single multi-function button, and no display. You spend 20 minutes reading a PDF. One product is discoverable. The keyboard is decades older but infinitely more usable.

🧠 Memory Anchor — The Two-Minute Tourist
Imagine handing your product to a tourist who speaks no English and giving them two minutes to complete the core task. If they fail, your discoverability failed. Good design works across language barriers because it uses shape, position, and feedback — not words.

If users need onboarding to use basic features, the design owes them an apology.

14. Human-Centred Design — Design starts with understanding people, not technology

The Idea
Human-centred design (HCD) is Norman's proposed process: start by deeply understanding the people who will use the product — their goals, their context, their mistakes, their mental models. Then design, prototype, test with real users, and redesign based on what you observe. Repeat until it works. HCD explicitly rejects the approach of building something technically impressive and hoping people figure it out. Technology serves humans, not the other way around.

Simple Example
A hospital redesigned its patient check-in process. Engineers built a sleek touchscreen kiosk with 12 fields. Patients — many elderly, many anxious, some in pain — took 8 minutes and frequently needed staff help. A human-centred redesign team spent two days watching patients check in. They cut the kiosk to 3 fields. Average check-in dropped to 90 seconds. The technology got simpler by understanding the human better.

🧠 Memory Anchor — Observe, Do Not Assume
The most dangerous sentence in design is: Our users will obviously do X. No. Watch real people use real things in real situations. What users say they do and what they actually do are almost never the same. Your assumptions are the problem. Observation is the fix.

You cannot design for people you have not watched — assumptions are not research.

How the Ideas Connect

These ideas are not a checklist — they are a system. Affordances and signifiers answer what can I do. Mapping and constraints answer where and how. Feedback and conceptual models answer did it work and do I understand why. The seven stages of action give you a diagnostic tool to pinpoint exactly where the design breaks. The gulfs of execution and evaluation measure the distance between a user and success on either side of an action.

Human-centred design is the process that ties everything together — you observe real users, see which principles are violated, fix them, and test again. All the principles exist to serve one goal: closing the gap between what a person wants to do and what the product allows them to do, effortlessly and without error.

Who Should Read This

A developer who has ever shipped a feature users complained was confusing — this book will show you exactly why and give you the vocabulary to fix it.
A product manager who needs to push back on feature bloat with a principled argument — Norman gives you the framework and the language.
Anyone who has ever felt stupid using a door, a microwave, or an app — you were not stupid; the design was broken, and this book proves it.

Who Can Skip This

If you are a senior UX designer who already lives by these principles daily, the 2013 edition adds little you do not know. The core ideas are foundational but not new to someone with 5+ years of design practice.

The One Thing to Do Today

Pick one physical object you use every day — a door, a tap, a light switch, an app on your phone. Spend three minutes breaking it down: What are its affordances? Are the signifiers clear? Does the mapping make sense? Does it give feedback? Where does it fail? You do not need to fix anything today — just practice the diagnostic habit. Once you can see design failures clearly in everyday objects, you will never be able to unsee them in the things you build. That is exactly the point.

Rating and Verdict

Readability: 4/5 — Clear and full of stories, though some sections repeat the same point from different angles.
Actionability: 5/5 — Every principle is immediately applicable to anything you build or evaluate.
Originality: 5/5 — Published in 1988, it invented the vocabulary of usability that the entire industry now uses.
Verdict: This summary gives you the full mental model — but the original book is worth reading once for its case studies and the texture of how Norman thinks. If you design or build anything other people use, it belongs on your shelf.