As many of you well know, designing a web site can be very difficult. Part of that difficulty is in creating an information architecture that represents your content organization and labeling the links to make that content easy to find.
Danielson (2003) found that the users’ perceived disorientation in navigating a Web site is related to the amount of changes in the navigation schema that the user experiences while seeking information. Danielson manipulated the level of persistent and consistent categories in the navigation and defined this manipulation as navigational volatility. He measured navigational volatility by counting the number of links that changed location from the source page to the destination page.
Danielson found that the navigational volatility was related positively to participants’ ratings of disorientation and negatively to ratings of ease of use. In other words, participants reported feeling more disoriented when the links changed from one page to the next and rated the site low in ease of use. He concluded that it is best to keep navigation links the same from page to page to help reduce the risk of users feeling disoriented and to assist them in finding information.
Information Architecture Structure
Miller and Remington (2004) looked at how information architecture and category label ambiguity impacted user performance. They manipulated the depth and width of the information architecture by having users interact with either a three-tiered or a two-tiered structure. They also manipulated the categorization by using clear labels or ambiguous labels.
They found that users performed better on sites with a deep rather than wide architecture when the site had clear labels. Participants also found items with the ambiguous labels faster in two tiered (wider architecture). This is because participants have more chances to get lose in the three tiered (narrower) architecture. The research reinforces that having a great organization scheme does not help users if the link labels are unclear.
Organizational Scheme and Labeling
Resnick and Sanchez (2004) looked how organizational schemes and labeling affected a user’s ability to find information. They created six fictitious health food store Web sites. They organized the sites by a product-centered or user-centered schema. Product-centered sites organized items by product categories whereas the user-centered design organized items according to the user’s goal. Labeling schema contained three schemas, high, medium, or low quality labels. Quality was defined based on ratings as to how well each heading represented the items in that group.
Results showed that quality of labels had a significant effect on the amount of time required to complete the task and on the userâ€™s overall performance. Users in the high quality label group committed fewer errors, and found more products than users in the medium and low quality label groups.
The results show that sites with poorly designed labels caused user to waste their attention trying to figure out what different labels meant. Labels classified as high quality were gathered from typical users. Low quality labels were gathered through benchmarking existing labels. Labels that did not meet the user’s schema result in longer search times. It is important to know what information users are expecting to find and to provide them with clear links so they don’t waste time searching for information.
Content Organization Scheme
Product based organization received higher satisfaction ratings and users committed fewer errors when compared to the task based organization. Users performed better and were more satisfied with the product-based content organization, especially in the lower quality labeling condition.
Information Organization and Older Users
Kurniawan and Zaphiris (2003) looked at information organization for older users. They were interested in looking at health information sites and ways to create better information architectures. They wanted to group and label information to make it easier for seniors to understand.
First they asked the seniors to do a card sort of all the items in one of the main categories from the site. A cluster analysis of the data resulted in the four main categories each containing 4 subgroups.
In the second experiment participants took lists of the group items and provided labels for each group. In the final experiment, the researchers showed participants the suggest label and group items and asked them to rank the label names based on their fit to the group.
Results showed that the new user-provided labels were much less formal than the original site and were easier to understand. The participants also grouped the items much differently than the original structure. The original structure had 4 branches, each main branch had two sub-categories and each sub-category contained two items. The user based structure was a lot less structured with varying numbers of items and sub-categories in each main branch.
Users also grouped items together based on their function or service provided, instead of factors such as geographic location.
These studies suggest several important considerations for designing the information architecture of a Web site. First, good link labels help user performance while bad or ambiguous labels hurt their performance. Second, additional levels of organization can be helpful for users (as long as they are appropriately labeled). No matter which content organization schema you chose, it is better to keep navigation links persistent and consistent throughout the site. Matching user’s mental models to your design will result in a more useful site.