My friend says that cross-country skiing is a good sport because it keeps you humble. Every time you start to get too pleased with yourself, you manage to slip and fall down. I can now say the same for usability testing: doing it is an excellent way to remind yourself that you have not created the pinnacle of user interface design. Luckily, I was pretty sure of that to start out with.
Anyway, yesterday I visited the two other branches of the county library to do some usability testing of the new website I’ve been working on. I was about to sit down and put together my findings, but it occurred to me that it might be helpful to someone down the line some day if I wrote a little about the whole process.
The book on usability that everyone recommended to me is Don’t Make Me Think: A Common Sense Approach to Web Usability by Steve Krug. While under normal circumstances I would be off put by that title (it is, after all, one of my goals to make people think), when it comes to web design, it’s right on. I don’t want people to have to think about how to find the information: I want them to think about the information.
Krug talks a lot about how to design websites and how not to, and at the end of the book he outlines how to do “simple” usability testing. His “simple” involves a separate room, a video camera, several people to run the test, and $50 each for the people who take it. All of that is above and beyond what we can manage here, but Steve Lawson told me they’d done usability testing without the camera and with cookies or something instead of $50–so I figured I could also do it without the separate room and the multiple people. I did get treats for the testers from the Meeteetse Chocolatier.
Since I’d returned the Krug book many months ago, I poked around to see if I could find some more good advice on the web, and I came up with the University of Virginia library’s Web Usability page, which had some good instructions and lots of sample questions used for testing various parts of their websites.
Of course, I did a lot of things wrong: I designed the website, and I wrote the questions and did both the interviewing and observing. These are all big no-nos, but you do what you can, and I think I learned a lot even with this small, highly imperfect system.
I asked the other branch managers to find me two volunteer testers each. I said they didn’t need to be computer experts–just average people who might use the library or the website. I planned for an hour with each person, which I figured was going to be way too much time and was–each tester took perhaps 20-30 minutes total.
We did the tests sitting at one of the public computers in each library. The tester sat at the computer, and I sat a little to the side so I could see the screen. I explained the idea–I would ask them questions or give them little tasks to try to do–and emphasized that the test was a test of the web design, not of them. For each question, I recorded whether they were able to answer it and made some notes about the process they followed to get there. I tried to keep track of how much time they took on each question, but that got too hard to do along with writing comments and watching, so I gave up. Here are some of the results:
I started by asking, “Is the library open on Thursday nights?” Half the respondents saw the library hours in the sidebar right away; the other half clicked on the tab for their branch and scrolled from there. I was glad I’d decided to put the library hours in both places, so things began well, but they went downhill rapidly.
Almost everyone mistook the catalog search box in the sidebar for a search box for the site. If they couldn’t find whatever I had asked them about (how long they could check out a movie; what kind of ID they needed to get a library card; what programs are available for teens), they’d try typing something into the search box. That’s perfectly logical–in most cases, a search box will search the site its on. In this case, though, it searches the library catalog. As you may guess, when you type in “library card id” as a catalog search, you don’t get many helpful results.
A few people figured out they were in the catalog and not in the site when that happened, but most didn’t really distinguish. Almost everyone tried to use the “Go Back” button in the catalog to get back to the website, which of course doesn’t work.
Another result of people thinking that the catalog search was a site search was that when it came to asking them where they’d go to look up a book or a movie, many of them were stuck. They wanted the catalog–many of them even said, “I’m looking for WYLDCAT,” the name of our statewide library catalog. They didn’t see anything that said “WYLDCAT,” or any of the familiar graphics, and they didn’t think of the search box as having anything to do with the catalog. So they went all over the place–they’d try clicking the tab for their library, or clicking on the Books category, or clicking on one of the New Books posts. One of them said, when I asked how they’d find a book, “I’d go to one of those computers,” pointing to the computers reserved for catalog searching.
The question I thought would be the hardest–”Can you look for information about your genealogy?” turned out to be the one everyone got right on the first try. Everyone clicked on the Research tab, and then scrolled (or used the anchor tags) till they got to the genealogy databases.
The other thing that turned out to be really puzzling was the contact form. I meant to get rid of the “website” box before I did the testing, since I figured that would be confusing, but as it turned out, the whole thing was confusing. Many people were just baffled. . . “But how can I put in an e-mail address? I don’t know any e-mail address to send it to?” At the debriefing, after we finished, I explain that it was an e-mail form, and the E-mail box was for their e-mail address. Though some said that making it say “your e-mail address” would help, I got the feeling that the contact form really isn’t the way to go for this project.
So, here are some changes I’m going to make:
- Add the WYLDCAT logo above the library catalog search box, and revise it to say “Search for books, movies, etc.” or something like that
- Add a “search this site” search box
- Redo the contact page and give multiple ways to contact each library: address, telephone number, and at least one e-mail address (There is a generic address for each library, and some e-mail addresses are on the old website. I need to find out who is willing to be contacted and who isn’t. That might have been a smart thing to figure out earlier on, eh?)
- Put a few more things like phone numbers and important details in bold type
- Move the information from the Friends and Foundation tabs somewhere else (probably to the sidebar) to make room for an FAQ tab up at the top–though I think I may just call it “Help,” since I’m not sure how many people will translate “FAQ” to “Frequently Asked Questions.”
- Relabel “Flickr Pictures” to something like “Pictures From Your Library,” and maybe add a note explaining that they’re hosted by Flickr. Flickr as a site was not familiar to anyone, and a few people thought maybe it was connected in some way to the movie question.
- Get rid of the News Categories bit in the sidebar. It seemed to confuse people to have a tab that said Cody and a link that said “Cody Library.” Unless you’re familiar with blogs, it’s not readily apparent that this is effectively a blog.
- I’m not totally sure about the whole front-page-as-blog thing. . . I think people may expect something more static. I could make a front page and have the blog in a separate tab–or maybe just have one blog post at the top that stays the same. I haen’t made up my mind yet.
Doing usability testing–even in the small, podunk kind of way that I did it–was hugely helpful. It seems obvious now that people might confuse a catalog search box with a regular search box, but I never would have guessed that on my own. When I’ve made some changes to the website, I may try it out a few more people and do a final tweaking, and then, I think, we’ll be ready to go live.