As noted earlier in this series, information users range from novices to serious, experienced researchers. Usability means different things to different people, but in all cases the idea is to meet the users' expectations and satisfy their specific information needs.
Two international standards define usability as follows:
- "[Usability is] the extent to which a product can be used by specified users to achieve specified goals with effectiveness, efficiency and satisfaction in a specified context of use." — ISO 9241-11
- "Human-centered design is characterized by: the active involvement of users and a clear understanding of user and task requirements; an appropriate allocation of function between users and technology; the iteration of design solutions; multi-disciplinary design." — ISO 13407
This article looks at some of the important considerations for the usability of information systems, particularly those involved in searching a large information base where a large number of matches is likely.
Quite a few areas are important to the user experience, but three in particular apply to information management systems: placing users in control of the system, reducing the memory load required to effectively use the system, and providing consistency of experience. These three concepts are taken directly from the principles of cognitive science and are related to how people perceive and use complex information-gathering tools.
A basic tenet of usability is that the user is paramount; the system should present as few obstacles as possible to the achievement of the user's goals. To better assist users of information systems, the user interface should allow users to move in any search direction that appeals to them at that moment.
For example, when using a travel reservation system, the user's intent is usually to find the fastest route at the lowest cost. The system provides search tools that take the date of travel, origination, destination, and conveyance to locate a matching set of available items. However, a traveler often wants to look for other routes and compare cost, time, itinerary, carriers, and so on. Figure 1 shows the result of a search on the Cheaptickets.com Web site for travel from Denver to Seattle. The user can accept any of these matches, or change the search parameters directly from the result screen.
Figure 1. Airline flight search result
You can also put users in control of the system by allowing on-the-fly filtering and by letting users reduce the search result without forcing another search. In the Amazon.com search for a toy, shown in Figure 2, the search resulted in a set of matches. In this case, the system displays the search results with a particular default sort (such as price and shipping cost), but the user can redisplay by other attributes, such as by price alone.
Figure 2. Online item search
Another core concept for usability is reducing the amount of information the user must keep in short-term memory. Humans are limited in the amount of information they can retain, especially when that information is complex. For people who design information systems, this means providing a history of searches to which the user may refer during an information-gathering session. Moreover, it's often useful to keep the search results associated with the search terms, because doing so lets the user compare one search with another. An example of this approach is a library search for books or other materials, when the item of interest may not be available from the current library. In this case, the same search can be forwarded to another library network without the user having to remember the correct search parameters.
Many researchers follow a trail of references to other works during a literature search by using one reference to provide a set of additional references to research. For advanced users, the system should provide a way of tracking references that are found to be of interest along the way. One way to do this is to keep a grab bag of selected results that can be printed after the user has concluded the search. This is how the PubMed Entrez research system operates (see the Resources section). A selected reference is added to a private list that can be forwarded by e-mail or printed at any time during the session.
Finally, a good, easy way to reduce the memory load for users is to provide a set of defaults for the most common searches. In the case of a library, the default may be by author, because that is how many people look for specific books. For a product catalog, the default may be all products in a specific category or all items at a particular location. You want the default selections to provide the most useful result to the largest group of users.
Provide a consistent experience
The more consistent the display of information, the easier it is for users to understand the relevance of a particular search result. The results should always display in the same place, with the same layout of information, regardless of the information that is returned. You should maintain the color, font, style, and so on from the initial search, along with the result display and the means of selecting information of interest. Whenever possible, combine all these actions into a single display to limit the amount of visual change the user sees on-screen during a search.
Another consideration is to place important information higher in the display field so the user doesn't have to scroll the display to see related information. This is known in Web design as below the fold, which is the point where an average display runs out of the observable window. In many cases, there is more data to display than screen space to display it. So you should take care to group related items and place them where they can be best seen. For examples of this approach, look over the result sets from Web sites like those for travel or other reservation systems. Notice that the lowest prices are shown first. Similarly, sites like Google place the paid advertisements on top of the findings.
An information system includes two primary functional techniques: searching the information store using one or more terms, and navigating the resulting information matches. Together, these complementary functions permit the user to locate and then review the information contained in the information system.
Searching for relevant information is the heart of any information usability scheme. If users can't locate data of interest—or locate it rapidly enough—they're likely to lose patience and abandon the search. So much for all the effort that went into collecting the information in the first place! To avoid this situation, you can use an information store, which should provide an effective and flexible search mechanism.
Search engines come in many flavors and forms, but almost all use some form of term-based indexing to rapidly locate matches. They use this type of indexing because term-based search algorithms are well established and relatively easy to understand and implement. Term searches use standard-language words, usually in combination, to locate and rank cataloged information. Web-search engines such as Google and Yahoo use a sophisticated algorithm to match and rank Web pages to search terms, but they still use this standard search technique. In contrast, context-based searches are based not just on individual words but also on the context in which those words occur. However, these techniques are difficult to implement and aren't commonly used. The focus here is exclusively on term-based searches.
Term-based searches use two forms: natural language and formal nomenclature. They also use two levels of formality: basic and advanced. As described in Table 1, these two search dimensions lead to different ways of presenting search functionality to the end user. On the basic end of the spectrum is the simple search shown in Figure 3, which provides a single, faceted search based on an author's name. This search is well suited to a novice user or someone interested in browsing a basic category of information items. Figure 4 shows a somewhat more advanced search capability that allows, among other things, the ability to filter items by availability at a particular location (if you haven't guessed, this is a public library search page). On the formal side are tools for professional researchers, including PubMed Entrez, which lets investigators perform sophisticated searches on scientific or other publications. In particular, a formal language (discussed in the third article of this series) for medical headings (MeSH) can be used for highly specialized searches of medical literature.
Table 1. Search matrix
| Natural language | Formal language | |
|---|---|---|
| Basic search | Public library | PubMed journal search |
| Advanced search | Google advanced | MeSH (medical term) search |
Figure 3. Library book basic search
Figure 4. Library book advanced search
Searches can also be broadened or narrowed based on additional terms, or the possible matches can be narrowed (for example, by using an exact string match like "hot-house flowers" where only that phrase is found). If the information organization uses a thesaurus (discussed in article three of this series) to describe information categories, then a search can return broader or narrower matches based on the term hierarchy.
The key to a successful search is to match the sophistication and complexity of the search interface to the expected user groups. The most successful sites provide a broad range of search capabilities to allow all groups to find the most appropriate level.
Depending on the level of search-term sophistication and the size of the underlying data store, the return set of matches for any particular query may be so large as to be unusable. Consider a standard Google Web search, where thousands or millions of matches may be found among the uncountable number of Web pages on the Internet. If this set were presented to the user without some form of navigation and ranking, the chances of finding useful information would be slim. This is why high-quality search engines provide a sophisticated ranking and browsing capability.
Information ranking is a popular area of research, due to the explosive growth of information databases and information accessibility. There are many ways to rank the findings of any search, but the most common are based on one or more of the following:
- Number of occurrences of search terms in the information
- Proximity of terms within an information item
- Occurrence in thesaurus or index associated with the information collection
This approach to ranking allows for a simple first-pass sorting of information into relevant groups. These techniques are often successful, especially when used in combination with careful choices of search terms.
Regardless of the ranking and scoring technique, it's likely that the user will need to browse through the search-return set to review each choice. There is a trade-off between showing a useful information summary and overloading the user with detail. One useful technique is to show the matched terms with a few of the surrounding sentences. This usually provides enough context for the user to determine whether the match is useful or should be ignored.
Navigation models that are based on an understanding of the user's intention work best. For example, consider these five different information areas, which are frequently used with search engines:
- Inventory. Users want to know how many of something are available and where they're located.
- Product catalog. Users want to know what is available and how much it costs.
- Reservation system. Users want to know how many reservations are available and when.
- Encyclopedia. Users are interested in general reference.
- Professional journal search. Users are involved in detailed research.
In each case, the navigation model associated with the information returned from a search is different. In the inventory case, users are looking for a summary of all items that match the search query, with the associated location and count. Someone looking at a product catalog is interested in what products are available and is probably looking to purchase one or more items. A reservation system should provide an organized list of results, such as airline reservations that can be sorted by lowest cost or shortest route. For the final two items, the user's intent is the same—to locate and understand information about a particular topic—but the degree is different. A researcher is much more likely to be a savvy user who will understand a complex display of search results, as opposed to an encyclopedia user, who may be a young child.
One final point on information navigation involves the idea of bread crumbs. This technique is often used in Web site navigation to permit the user to back-browse to earlier pages or to understand where in a page hierarchy the current view is located. Similarly, a user who is browsing a particular search result may be interested in following multiple lines at the same time, may wish to back up to an earlier search result and take a different path, or may wish to save the search path for future use. In these cases, information-system usability is directly influenced by a user's ability to track along a search trail.
During the development of an information-system interface, the development team's decisions regarding critical points of interest may have profound effects on final usability. When you're investigating users' needs, the users should provide guidance about the kinds of searches they expect and how the results will be used. For example, novices will have different searches and uses for the results than advanced users. If the system is complex, then a prototype may be required to further guide the team in selecting appropriate technologies and presentation techniques. Finally, you should evaluate the system's effectiveness by directly measuring the user community's satisfaction in finding information of interest.
The needs of any information-system user community are often difficult to discover on the first try. Users typically discover new ways of utilizing the system as they have the opportunity to work and explore. So, when you're developing a new system (or significantly enhancing an existing one), it's a good idea to "build a little, deliver a little" to a select set of knowledgeable end users who can help you discover good approaches as well as ones that are less useful.
A time-honored way to understand complex systems is to build a small, lightweight prototype that illustrates key features of a larger system. For information usability, such a prototype may involve investigating a search algorithm, experimenting with various navigation models, or providing users with a chance to comment on the visual presentation. The earlier an end user can see and work with the evolving system, the sooner they can help direct the eventual solution.
You can validate your approach and audience expectations by creating one or more of the following:
- Click-model. A user interface with only stubbed data or static screens
- Canned demo. A limited demonstration of the system's capabilities
- Limited release. A beta release to a friendly audience
After a system is in use, it's a good idea to periodically revisit the usability question with users. Over time, a user community may become disenchanted with a once useful system as newer technologies are introduced. You can visit with key users, conduct usability workshops, or distribute surveys to discover areas where the system can be improved.
- Visit the
Usability Professionals' Association
Web site to learn more about usability,
usability testing, and general best practices.
- See
The Art of
Rapid Prototyping
by Scott Isensee et al for more information on the
development of prototypes.
- The
NCBI
Entrez Web sites offers more information about the PubMed Entrez
medical journal search engine and the MeSH term thesaurus.
- Browse the
technology bookstore for books on these and other technical
topics.
Benjamin A. Lieberman serves as the Principal Architect for BioLogic Software Consulting. Dr. Lieberman provides consulting and training services on a wide variety of software development topics, including requirements analysis, software analysis and design, configuration management, and development process improvement. Dr. Lieberman is also an accomplished professional writer with a book (The Art of Software Modeling, Auerbach Publishing, 2006) and numerous software-related articles to his credit. Dr. Lieberman holds a doctorate degree in Biophysics and Genetics from the University of Colorado, Health Sciences Center, Denver, Colorado.
