Key Ideas from Don't Make Me Think

Steve Krug wrote Don't Make Me Think in 2000 and it still shows up on every recommended reading list for anyone who builds things people have to use. That tells you something. The web has changed dramatically since the first edition — mobile, SPAs, design systems — but the underlying observation hasn't moved an inch: people will not work harder than they have to. If your interface makes them think, they'll leave.

The book is thin, direct, and full of common sense that turns out not to be common practice. Here's what's in it.

1. The Core Principle

The title is the whole thesis. Every unnecessary decision you force on a user — What does this do? Is this clickable? Am I in the right place? — spends a little of their goodwill and patience. Those costs add up fast. Your job as a designer or developer is to eliminate those moments before they happen.

Krug's standard for a self-evident interface: a typical user should be able to look at a page and immediately understand what it is and how to use it, without having to think about it. If you find yourself needing to explain your UI, the UI is the problem.

2. Users Scan, Not Read

This is the most important thing to understand about how people actually use web pages. They don't read them. They scan for the thing that catches their eye and looks close enough to what they want, click it, and bail if it doesn't work out.

Krug identifies three ways users actually behave that designers routinely ignore:

They satisfice. Users don't look for the best option — they take the first reasonable one. Optimal decision-making on a web page is rare. People are in a hurry, or distracted, or just want to be done.
They muddle through. Most users never figure out how your site actually works. They find something that kind of works and stick with it. If your navigation is confusing but they find what they need anyway, they'll use the same confusing path next time.
They don't read instructions. People skip onboarding text, skip tooltips, skip the helper copy you spent an afternoon writing. Design for the person who is ignoring everything except the button they're about to click.

3. Design for Scanning

Once you accept that users are scanning, the implications for design follow naturally. Krug's advice is to treat every page like a billboard — something a person passes at speed and needs to register instantly.

Practically, this means:

Clear visual hierarchy. Important things should look important. Headlines, calls to action, and navigation should stand out, not compete.
Short chunks. Break content into small, digestible pieces. Long paragraphs disappear into noise.
Obvious clickables. Links and buttons should look like links and buttons. Don't make users hover to find out what's interactive.
Eliminate noise. Every element on a page that doesn't help the user accomplish a task is working against you. Not neutral — actively against you.

4. Conventions Are Your Friends

Web conventions exist because millions of people have developed habits from using the same patterns across thousands of sites. The shopping cart icon in the top right. The logo linking to the home page. The search bar at the top. Breadcrumbs at the top of a content page. These things are conventions because they work.

Krug's point is blunt: don't break conventions for the sake of originality. Every time you deviate from something users expect, you make them think. That's a cost. Creativity should go into the content, the product, the experience — not into reinventing navigation because you thought it would look interesting.

5. Omit Needless Words

Krug borrows directly from Strunk and White here and applies it to UI copy. Most web pages have far more words than they need, and most of those words are not helping anyone.

He calls out happy talk specifically — the self-congratulatory, vague filler text that exists on nearly every homepage. Things like "Welcome to our website, where you'll find everything you need." Nobody reads it. It's noise. Cut it.

Every word competes for attention. Fewer, better words win.

6. Navigation Is Wayfinding

Navigation isn't decoration. It's the system that tells users where they are, what they can do, and where they can go. Krug argues that every page needs to answer a few questions at a glance:

Where am I?
What's here?
Where can I go from here?

The page title, the breadcrumb, the active state in the nav — these are all answers to those questions. When they're missing or ambiguous, users feel lost, and lost users leave.

Persistent navigation (the same structure on every page) is how you give users a consistent mental model. If it changes between sections, they have to reorient every time they move.

7. The Home Page Has the Hardest Job

The home page has to accomplish more than any other page on the site: establish who you are, what you do, where to start, and what the most important paths are — all without overwhelming someone who arrived knowing nothing about you.

Krug's advice is to be brutally clear. What is this? What can I do here? Most home pages fail not because they lack information, but because they try to say too many things at once and the most important thing gets lost.

8. The Goodwill Reservoir

Krug introduces a useful mental model: users arrive at your site with a reservoir of goodwill. Every time something annoys them — a pop-up at the wrong moment, a form that clears when they make a mistake, a link that opens an unexpected PDF, hiding information they expect to find easily — they spend some of it. When the reservoir is empty, they're gone.

Things that drain it fast:

Hiding your phone number or contact information
Making users register before they can see what they're registering for
Designing for the average user when you know your user has a specific need
Making errors feel like the user's fault

Things that restore it: obvious navigation, fast load times, useful error messages, not wasting their time.

9. Usability Testing Doesn't Have to Be Fancy

Krug's most practical chapter is about testing. His argument: testing one person early is better than testing fifty people later. You don't need a lab, a research team, or a large budget. You need a real person, a prototype, and someone watching.

The goal isn't statistical significance — it's surfacing the obvious problems before they ship. Most usability issues are obvious once you watch someone encounter them. You don't need a lot of data to know that your main CTA is getting missed.

Test early, test cheaply, test often. The feedback from real users doing real tasks is worth more than any internal design review.

10. Accessibility Is Good Design

In the book's later editions, Krug includes a chapter on accessibility that makes a point worth repeating: building for accessibility doesn't create a separate experience for users with disabilities. It makes the site better for everyone.

Alt text helps screen readers and also shows when images fail to load. Good color contrast helps users in bright sunlight or on low-quality screens. Keyboard navigation helps power users and anyone who'd rather not use a mouse. The accommodations designed for the few end up serving the many.

The thing that makes Don't Make Me Think last is that it's not really about the web — it's about understanding that people using your product are busy, distracted, and unwilling to work harder than they have to. That was true in 2000 and it's true now. Design that respects that reality will always win over design that ignores it.