Using Usability Testing to Determine "Related Links" in An Online Brokerage Web Site
By Ioannis Vasmatzidis, Eliot Jablonka, and Hsin Eu
In content-rich web sites, success is defined by the user's ability to locate relevant content quickly and efficiently. To achieve such success, information architects and usability specialists direct their efforts towards developing intuitive and well-organized navigation structures to support user content search tasks. However, the rich content of the site may lend itself to more than a single navigation scheme that might satisfy the navigation demands of the site. In such cases, to decide on the best approach, usability tests are conducted. Observing users navigating through the site provides the critical information needed to finalize the information structure of the site. We used this technique to redesign the navigation of the www.harrisdirect.com web site, an online brokerage web site. This site is very rich in content. In addition to online trading, the site allows users access to a multitude of news, investment research, market research and educational content resources. The inherent inter-relationships of this content enabled us to further improve on the final navigation design by introducing related links in a manner that was derived from the usability test.
The Concept of Related Links
Related links enhance the navigability of web sites in two ways. First, they provide access to content that is related to the current page. In the context of a financial services web site, if a user is reading a news story on IBM, a related link could provide access to a research report on IBM. In many cases, such links are known as "See also" and "See more" links. Second, they allow users to take action on information that is being reviewed. To use the same financial services web site example, if a user is reading a research report on IBM, related links would be those that allow users to buy or sell this stock.
Using Usability Tests to Decide on Related Links
The main goal of our usability test was to optimize the hierarchical structure of the site content. Related links were introduced in the site only after we felt that the overall site hierarchy satisfied the way users intended to search for content in the site.
During the test we asked users to locate information in the site such as research reports on a particular company, or financial planning tools. We then used this information to finalize the grouping and information structure of the site. After the structure was finalized, we revisited the results of the study and looked for less frequent patterns of user behavior that could not be best covered by the hierarchical navigation. By patterns, we mean pages that were visited by users in an effort to locate the content required by the usability test tasks. Such behavior was not very common - otherwise it would have influenced the overall information grouping - but was deemed by the usability professionals who conducted the study, and the business domain experts as reasonable ways of navigating the web site. As a result, we developed the following rules of thumb for developing related links:
1. A related link should be added to a page if a user consistently visited that page while taking an alternative path (not supported by the optimal structure) of navigation.
2. If a related link allowed a user to navigate from page A to page B, we added a similar related link to allow users to navigate from page B to page A. We call this the "reciprocal property" in the design of related links.
To illustrate with an example, we asked users to access educational information on tax-related issues. In the prototype tested, this information resided in the "Education" category under the "Tax" section. However, more than two test participants navigated to the "Tax Resources" section of the "Advice & Planning" category. As a result, we added the "Tax" education related link to the "Advice & Planning" category under "Tax Resources" and the "Tax Resources" related link to the "Education" category under "Tax" (see Figure 1).
Figure 1. Related links were introduced in the site only after we felt that the overall site hierarchy
satisfied the way users intended to search for content in the site.
In this article, we advocate the use of usability testing as a technique for deriving related links. This technique is especially useful in content-rich web sites, where predicting the path users will follow to complete a task or meet a goal, is not always obvious. Usability tests can provide critical insight as to how users navigate within the information space of the site. Related links can then be designed to accommodate all navigational patterns observed in the test. Finalizing the related links of a site needs to be done in a way that reflects realistic navigational behavior. Usability engineers should be careful not to design a related link that accommodates behavior resulting from poor usability testing methods (for example, a behavior resulting from a poorly phrased usability test task).
Pershing Technology Group
Pershing Technology Group
Pershing Technology Group