Don't Make Me Think Revisited - Steve Krug (Book Summary)

Book cover reading don't make me think revisisted

"Don't Make Me Think Revisted: A Common Sense Approach to Web Usability" is an updated version of the classic 2000 book on web design, republished in 2014. This article is a summary of what I found to be five important points covered in Steve Krug's book. The book is a good read for anyone interested in UX or UI design, or anyone who's goals is building products that people will enjoy using.

The difference between the 2000 and the 'Revisted' 2014 version, aside from updated images and references, is the inclusion of the section on mobile devices.

Krug's #1 Rule: Don't Make Me Think
Usability Testing with just One Participant
Design for Users that Scan
The Core of Usability Testing
Design Together as a Team

Krug's #1 Rule: Don't Make Me Think

Not making the user think is Krug’s #1 rule, and given it's the book's namesake it is a consistent theme throughout. Making important things as clear as possible to the user is what we should design for.

Good Design

We should aim to for these design goals throughout the whole product:[1 p18]

Primary Goal: Everything is self-explanatory, so no thought is required to use it.
Secondary Goal: Everything that isn’t self-explanatory is self-evident, requiring minimal thought to use it.

a lady in front of a computer furrowing her eyebrows, with her chin resting in her hands.

A user with furrowed eyebrows. This is what you don't want to see in a recording of a Usability Test.Image courtesy of Andrea Piacquadio.

Design Techniques

Krug suggests some useful ways to reduce the need for a user to think in Chapters 1, 3, and 5:

Make it obvious what is interactable, so the user doesn’t have to question it, or "trial and error" through it.[1 p37]
Remove words on the screen that aren’t essential to the task at hand, reducing the noise level on the screen,[1 p49] thereby reducing user effort required when scanning.
Use the most basic wording as possible to reduce mental processing e.g. "Jobs" vs "Employment Opportunities".[1 p14]
Avoid instructions, aim for a UI that is self-explanatory.[1 p51]
Users should know exactly where they are on your website without thinking about it.[1 p17] Krug proposes a "Trunk Test" - landing on a random page in your website (as a search engine link would), and seeing if you know exactly where you are.[1 p82]
"It doesn't matter how many times I have to click, as long as each click is a mindless, unambiguous choice."[1 p43] It’s not about efficiency, or depth of navigation, it’s about having an obvious and natural design.
If you have to be slightly inconsistent to make things more clearer, then sacrifice a bit of inconsistency for clarity.[1 p33]

Usability Testing with just One Participant

Usability testing with just one user is infinitely better than no testing at all. Krug believes that even the worst test with the worst user will reveal something important to you that will improve your product.[1 p114] The willing participant doesn’t have to be someone with the product domain knowledge, as most serious usability problems can be encountered by anybody.[1 p120]

a graph showing the diminishing returns of finding fewer issues the more participants that are used

A graph of usability issues found relative to the number of participants in a test[3]

Qualitative Testing, not Quantitative

Usability tests are qualitative tests so they aren't trying to prove something, Krug believes the ideal number of participants is 3, and one of the reasons is that not all issues have to be found, just the major ones.[1 p119] Nielsen’s classic number of 5 is graphed with diminishing returns, and after 5, the results do not justify the value.[3] But even with 5, it is worth mentioning the assumptions that are made in deriving that number, such as all issues encountered are worth fixing, and that the probability of one participant encountering an issue is 31%.[2]

Although the ideal number of participants may vary, it should be obvious - but is worth emphasising, that zero participants will provide zero insights into usability issues of a product. [1 p.114][3] You can't fix issues that you're oblivious to.

Design for Users that Scan

An unfocused image of a cafe menu, with a visual hierarchy visible

We scan the menu that is in front of us at a restaurant, without committing to the details until we find the section we're interested in. A product's UI is treated exactly the same way.Image courtesy of Morena Vw.

Every user is a scanner. You were, or are, likely scanning this page yourself right now. We scan through user interfaces quickly because, among other things, we are goal oriented,[1 p22] we’re on a mission to try to achieve something, often in as little time as possible, and our patience and goodwill can be limited.[1 p166]

Even from a Web Accessibility perspective, screen reader users are scanners too.[1 p179] They only listen to the first few words of a link or line of text, before skipping forward to the next section, and often set the voice to speak at a very fast rate.[4]

We should design for scanning, in both our textual content, and the overall visuals. Loosely quoted: "If users are going to treat our designs like billboards, then design great billboards".[1 p27] Krug has a whole chapter on this, but some key takeaways are:[1 p29-33]

Create visual hierarchies through font sizing to signify significance.
Clearly define sections separating levels of information.
Remove anything that's not part of the solution.[1 p38]

The Core of Usability Testing

a photo of a hand about touch a tablet screen

When recording the video of a participant on a touch device, try to capture their hands as well for details you won't see otherwise.Image courtesy of Pixabay.

Characteristics of a Usability Test[1 c9]

The duration of the test is 1 hour.
Record a video of the Participant’s physical actions and the screen.
Ensure to record the voice of both the Participant and the Facilitator.
The Facilitator must not lead the Participant, or influence them in any way.
The Participant should be encouraged to always think out loud.

Steps to follow to run a Usability Test[1 p125]

Welcome the Participant and outline to them what to expect.
Ask the Participant questions to put them at ease.
Show them the Product home screen and ask them what they think of it.
Get the Participant to perform tasks.
After the tests are complete, probe the Participant with questions about anything that happened.
Wrap up the session by thanking them, and paying them.

Have a look at Krug's website for a list of downloadable Usability Testing templates.

Design Together as a Team

Unless you're a solo act, product creation happens in a team, where everyone's expertise is put to good use.Image courtesy of Fauxels.

A UX Designer’s role should be sharing their specialised knowledge, without telling people how things should be done.[1 p187] Sometimes people do need convincing on the value of spending time and money on UX, here are a few core tips extracted from Krug's list:

Get your boss, and your boss’ boss to watch a usability test, as it will likely bring to light something that they weren’t aware of.[1 p185]
Perform a Usability Test on a competitor’s product and share it with your team, as everyone is always interested in the competition.[1 p186]
Try to understand your work from others' perspectives to be able to better sell them on the business value of usability testing.[1 p184-187]