Web Usability Testing: Three Cases from NYU Libraries

By Gloria Rohmann, Nadaleen Tempelman-Kluit, Mary Jean Pavelsek Gloria.Rohmann@nyu.edu; ntk2@nyu.edu; mary.pavelsek@nyu.edu

Since the mid-1990's, businesses, government, and non-profit organizations have increasingly depended on their website to serve as an electronic "front door" to products and services.

Statistics generated by website usage ("hit counts") are often cited as evidence of a website's popularity or effectiveness. These statistics can also be used to estimate how much time users spend at a particular site, whether they "buy" (or view) anything, and to provide some information on the effectiveness of the website design.

Websites are periodically redesigned to offer new services and improve user satisfaction. But before devoting the resources to design a new website or redesign an existing one, website owners, particularly those in education, increasingly rely on low cost, low-tech usability studies to test how "user-centered" the website design actually is.

What is usability?
"Usability rules the Web," says usability expert Jakob Nielsen. "Simply stated, if the customer can't find a product, then he or she will not buy it." (Nielsen, Designing Web Usability, p. 9) The same goes for websites offering information and services. Techniques for studying usability were developed by software designers and manufacturers. Representatives of likely user groups are asked to perform common tasks for which the device or software was designed, while a study team records their comments and behavior. Testing is done at various points in the development process-the first test often provides a baseline for subsequent improvements.

Methodology at NYU Libraries
NYU Libraries recently completed three usability studies. The studies were conducted entirely by Libraries staff; the only expenses were staff time and small "thank you gifts" (NYU Bookstore gift certificates) given to subjects. Web usability experts advise that reliable results can be obtained by testing as few as 10 subjects (Nielsen, Alertbox, 3/19/00). Since students and faculty at NYU were to be our subjects, we had to apply for exemption from NYU's Human Subject review from the Office of Sponsored Programs. (For more information on Human Subjects Review at NYU, see http://www.nyu.edu/osp/human.html; applications for exemptions should be submitted well in advance).

In all three studies, we started by compiling a short list of the most commonly observed user activities performed on the website, gleaned from informal polls of public service staff and the archives of the "libweb" general comments mailform and Ask-a-Librarian FAQs. In most cases we added certain "desired" activities, such as attendance at instructional classes, knowledge of discipline-specific subject pages and tutorials, and awareness of advanced reference assistance provide by subject selectors. Subjects were tested one by one; each test took about one hour. Three Library staff members took part in each session, one to ask the questions (a written copy was also provided), one to write down subject's comments and behaviors, and one to record the "path" through the website taken by each subject. Netscape's "history" file for the session was also saved and reset at the beginning of each new session. For the tests, we used the workstation in a small Library classroom with the same configuration as the public workstations in Bobst Library.

Evaluating the Libraries' Online Tutorials:
An important service of every college and university library is "user" or "bibliographic" instruction. At NYU, librarians skilled in pedagogy manage a program in which users are helped to develop research skills and information literacy. The "How to Use Bobst" section of the NYU Libraries' website (http://www.nyu.edu/library/bobst/howto.htm) provides access to a number of web-based tutorials for this purpose. How effective are these tutorials? In what format do users actually want to receive this information? These are some of the questions we wanted answered about the existing tutorials, followed by a sample of results.

  1. How linear in nature should the tutorials be? How should the navigation in the tutorials be arranged to best suit the users? Most users didn't like being locked into tutorials, but wanted a non-linear design structure
  2. Should the tutorials be primarily text-based, or should they be graphics and multimedia rich? There is too much text in most tutorials. The wording is unclear, and too much library jargon is used. Graphics are distorted and not clear.
  3. How to improve the usage of the tutorials? There is currently no motivation to use the "How to Use Bobst" section of the Libraries' website. Make sure links to tutorials are put in more relevant places on the main website. Results of this study are being incorporated into existing tutorials. New tutorials are being written based on these principles. Larger access issues will also be addressed in the redesign of the entire Libraries website (see below).
Redesigning the NYU Libraries Web
Since 1995, the NYU Libraries website (http://www.nyu.edu/library/bobst) has served as the primary access point to the Libraries' popular electronic databases and electronic journals, online catalog, interlibrary loan requests and online book renewals. It also provides links to "general information" about on-site services at Bobst and other libraries: hours, contact information, locations, classes and events, policies and procedures. In addition to a growing number of locally authored, discipline-specific subject pages and tutorials, special exhibitions complete the range of services provided by the website. Although our "hit count" looked good (over 2 million hits in the first three months of 2001), it was widely agreed that the six-year-old website design needed a substantial overhaul. We decided to begin with a usability test of the current site.

Here are some of the most consistent results. The results were used to develop principles for redesign that were incorporated in our RFP.

  1. Many subjects scanned for words; they would not read blocks of text. ACTION: break up essential information into chunks. Use lists, bullets and FAQs, rather than simply converting existing printed guides into web pages.
  2. Most subjects had difficulty finding articles in electronic indexes and electronic journals ACTION: Provide uniform access to electronic resources. An interactive "gateway" to such resources is planned.
  3. Navigation hierarchies got in the way of many subjects, who looked instead for search boxes where they could enter keywords. ACTION: Improve results of Site Search (Webinator) by adding meta-tags and rewrite pages to reveal key information. Use natural language; avoid jargon, such as "databases".
The website is now in process of being redesigned by an outside company. The new website will be ready for testing in Fall 2002. We will be conducting another round of usability studies at that time, before going live in Spring 2003.

Evaluating a new website: The Stern Virtual Business Library
Bobst Library also serves as the library for the Stern School of Business. Librarians and staff in Bobst identified a discrete set of users whose specific needs in the areas of business and finance were not being adequately served by the current Libraries website. The final part of a grant-supported project to create a website tailored to the needs of these users was a usability study. Users were asked a number of questions designed to show their ability to use the new website (http://www.nyu.edu/library/bobst/vbl) to find commonly requested information about business and finance.

  1. Navigation: We discovered that no matter what method was used, the desired information for each question was usually found in five minutes or less. Many subjects used the "Find" command on the tool bar and many used "Search this Site". The subjects in this test proved to be wilier and more creative searchers than we expected. Not only did they almost always find the answer, they often found it in a way we had not even considered. They found the answers their own way in spite of all our hard work to provide them with what we believed was an obvious path to the correct answer. Almost all subjects considered themselves to be advanced Internet searchers.
    ACTION: No action was taken because users always found the desired answers; they just did not find them in the ways that we had anticipated. The lesson here for information professionals is that they should stop assuming that there is a fixed set of "correct" solutions for finding information on the Web. The flexibility available with web technology has made searching for answers a more creative, forgiving and intuitive process-especially for users who are experienced web searchers.
  2. Language: Library jargon can be a problem. Most subjects did not understand what an "e-journal" or a "library catalog" is.
    ACTION: The term "journals in electronic format" was substituted for "e-journal". We are not prepared to come up with a substitution for "library catalog", as there really is no other way to describe it and the term is so entrenched. Again, web designers and information professionals must be careful not to assume that users are familiar with terms of the profession. Every effort was made while constructing this website to avoid terms of the trade and to use as much natural language as possible.
  3. Fonts: Some subjects found the type size too small for readability.
    ACTION: The font size on the website was enlarged, resulting in a much clearer, cleaner-looking site. We were surprised at the substantial difference this small detail made in the overall look and feel of the site.
  4. Color: The color intensity probably should be brightened or darkened. Used links don't contrast well with unused links. Different computers and browsers make this less or more of a problem. ACTION: The color contrast in visited links was brightened and changed to a more contrasting color.
For more information on these studies, see the NYU Libraries' Usability Website at http://www.nyu.edu/library/bobst/usability.

Books and Articles on Usability
Dickstein, Ruth, and Victoria A. Mills. "Usability Testing at the University of Arizona Library: How to let the Users in on the Design." Information Technology and Libraries 19:3 (Sept. 2000): 144-51.

Fichter, Darlene. "Head Start: Usability Testing Up Front." Online 24:1 (Jan./Feb. 2000): 79-81 Fichter, Darlene. "Testing the Website Usability Waters." Online 25:2 (Mar./Apr. 2001): 78-80.

Head, Alison J. "Hewlett-Packard: Usability Testing from the Field." Online 23:6 (Nov./Dec. 1999): 30+.

Nielsen, Jakob. Designing Web Usability. Indianapolis, IN: New Riders Publishing, 2000.

Spool, Jared M. Website Usability: A Designer's Guide. San Francisco: Morgan Kaufmann, 1999.

Usability Websites

Useit.com: Jakob Nielsen's Website. Ed. Jakob Nielsen http://www.useit.com/
This website is a jumping off spot for many excellent sites dealing with usability testing. Of particular interest to readers may be http://www.useit.com/hotlist/, a list of websites recommended by Nielsen, and Alertbox http://www.useit.com/alertbox/, Nielsen's biweekly column devoted to current issues in web usability.

Usable Web. Ed. Keith Instone http://www.usableweb.com/ Links to information on usability testing and other web usability issues.

MIT Libraries Web Advisory Group, Usability Testing: http://macfadden.mit.edu:9500/webgroup/usability/

Illustration: VBL Homepage
Gloria Rohmann is NYU Libraries' Webmaster and Head of Electronic & Media Services; Nadaleen Tempelman-Kluit is Instructional Design Librarian at NYU Libraries; Mary Jean Pavelsek is Librarian for International Business at NYU Libraries.