Don't Make Me Think -- Steve Krug

Krug's book focuses on lessons supporting one major tenet of web usability and human-computer interaction: a good website (program, application, etc.) should enable users to accomplish their tasks as easily and directly as possible.

Getting sidetracked for a moment: this reminds me of a discussion of "good technical instructions" in my technical writing class. However we approached what makes instructions "good" we kept arriving at the same point -- users don't really think about them when they're good. When they're good, users don't have to think at all, which is perfectly in line with what Krug presents.

Major Points

1. Don't make the web user think: little questions on a website add up and every time a user has to think even a little bit about where he should click, there is a greater likelihood that he'll quit. Designers should aim for providing a self-evident experience on a website. However, if that is unachievable, the goal should be self-explanatory.

2. Satisficing: most people don't read an entire web page. They'll scan until they find what they think they need and click. They satisfice, or attempt to meet criteria for adequacy, rather than to identify an optimal solution. They "make the best with what they have" instead of poring over an entire website and determine the best path to the best information.

Accordingly, pages must be designed for scanning, not reading. Krug suggestions several design choices:

-creating a clear visual hierarchy
-nesting text to show connections between text, pictures, etc.
-sticking to conventions
-defining portions of the page into clear sections
-making clickable items clear
-avoiding "noise" in the design

3. Writing for the web: because users won't read everything, it serves designers well to:

-keep it brief
-get rid of half the words...then get rid of half of what's left
-instructions are often useless -- users won't read them -- make all actions self-evident or users won't stick around anyway

4. Navigation: design for browsers (clickers) and searchers (those who go for the search box immediately). Make sure navigation is persistent so users will feel comfortable knowing where they are on a website, which will increase their confidence in the website. A few components of persistent navigation:

-site ID (logo plus tagline, usually on top left of web pages)
-sections (different areas of the website, such as product lines, services, etc.)
-home link (so users can always "start over" if they feel lost)
-search option
-utilities (explain different components of a website -- how to make a purchase, contact a sales rep, etc., but aren't part of the hierarchy)

Home page will usually be an exception since the mission of the home page is different than that of the second-level, third-level, etc., pages.

5. Home page is usually more complicated since many parties will be vying for the prime real estate. In addition to the tagline and a welcome blurb (description of what the site is about, which is different from a mission statement), it must contain the following items:

-site mission
-hierarchy -- showing what the site contains
-search option
-content "teasers" and promotional offers
-shortcuts to the company's "best stuff"
-registration/log-in area, if necessary

Most importantly, the home page needs to show a user HOW to get started. Do they need to register first? Should they jump to products?

6. Page names: it's important that a user can identify where they are on the website. Clear page and section names and matching link names can help. Tabs work extremely well here because they are hard to miss and are visually appealing. A good way to test this is what Krug calls the "trunk test" -- pretending that you’ve been blindfolded and locked in a car trunk, you should be able to answer these questions about a site immediately when your blindfold is removed:

-What site is it?
-What page am I on?
-What major sections does this site have?
-Where can I go from here?
-Where am I in relation to the rest of the site?
-Where can I go to search?

7. Simple, cheap usability testing: few companies can afford a large-scale usability test, but Krug suggests that even a little bit of testing is better than no testing. He suggests that only 3-4 general web users, a private room and a computer are necessary. Show users the website and ask if they can identify what it's about, why it's valuable, how it works, and have them perform a few tasks. Designers should review the results immediately and focus on the big problems -- those are usually pretty evident right away. Take care to not break something else while "fixing" one component.

8. General usability "common courtesy" points:

-don't make users provide more info than necessary (phone, address, email -- are all of these really needed?)
-don't let fancy features get in the way of usability
-make it easy for a user to back up and try again if he navigated to the wrong place

In just a few short months, my company will launch a new website, so this was a timely read for me. So many of Krug's points made me chuckle because I can imagine running into a lot of these problems, specifically the web design arguments is lists. Everyone is a user, so everyone knows what they like and have developed hard opinions about it (I hate pulldowns vs. I love pulldowns). Rather than debating what "most people like" (because there really is no average users) we must look at what works for our site. Does this hierarchy make sense for this information presented on this page? Is this the best way to present our story? If the presentation is the best it can be, then the users will take it and run with it, even if they feel strongly one way or the other about certain components on a website.

There were several points in this book where I stopped and thought, "Well, yeah. Of COURSE you do that." Of course you test early and often. Of course you can't muddle the home page with ten competing messages. Of course you keep navigation consistent throughout. But as I prepare for the website redesign, I find that I will absolutely hit some obstacles on the way. There are several parties in my company that will want their message front and center on the home page. I'll be in the unenviable position of managing that process, but I feel better equipped after reading Krug.

I've been through web redesigns before and I've become so lost in the process that it gets difficult to step away and see it with fresh eyes. Krug's suggestion for just a few random people to test it out ahead of time seems so simple, but there are always reasons for not doing it. We're too late in the game and we don't have time. They aren't the users we're speaking to, so they won't be of any use. Additionally, pride often gets in the way. When you spend months developing a website and open yourself up to average users poking around and finding faults, your ego can get a little bruised.

I'm really looking forward to soliciting help early on this project. I'll feel much more confident with my highlighted and dog-eared copy of "Don't Make Me Think" by my side!