|
|
|
Home What we do Storytelling for User Experience Articles and downloads About us |
|
On Beyond Help: by Whitney Quesenbery It is well accepted that understanding the users and a thorough analysis of their goals and tasks is a prerequisite for usability. To produce a document, online information or knowledge base that is truly usable, the designer and writer must also consider different user approaches to the information in order to create it in a form that meets those needs. The underlying technology must also be considered, as it affects the presentation of the information as well as the functionality available to users. To meet user needs for useful online information, all of these elements must be factored into the design – and technical communicators must master the skills necessary to make the right choices. We are in the midst of a sea change. More information, and more types of information, are available online. Users are becoming increasingly aware that they need more efficient ways to find the information they need. Interface and information designers, developers and technical communicators must rise to the challenge. To do this, we must employ the techniques of user-centered design to understand user’s tasks and goals. We must insist on usability testing to evaluate the success of the interfaces. Most importantly, we must find ways to design the user interaction in a way which matches the user’s needs, which is easily learned, and which is efficient, effective and engaging to use – in other words, we must design for usability. The interface design is more than a pretty face "smeared like peanut butter" (Lewis & Reiman 1994) onto difficult-to-use software. Usability must be built in through the combination of visual design, navigation (or interaction) design and information design that shape the user experience. The starting point for usability is understanding the people who will use the software. It is a basic principle of user-centered process that the design grows out of this analysis. The interface must match the user’s model of the task and help them accomplish their goals while speaking in their own language (Cognetics and UseIt web sites). Even popular modern software development methodologies such as the Unified Modeling Language (Booch 1999) start with a technique, such as use cases, for analyzing the interaction between the user and the system. If we know all of this, why is so much software and information still so difficult to use, still modeled around the data or system requirements? One answer, of course, is that until the design is truly user-centered, it will not be usable. Another is that even when users are considered, this analysis must go beyond simple tasks or workflow to an understanding of the context and the work styles that users bring to their tasks. Consider these two scenarios:
What makes these two experiences so different? In both cases, the goal is to find a present, the people shop in the same store, and they purchase the same book. However, the process each follows to reach her goal is very different, in ways that simple task analysis does not illuminate. In the ‘bricks and mortar’ store of these shopping scenarios, attractive displays facilitate browsing, the organization of goods into departments creates hierarchical categorization, and directories act as search agents. In other words, the store environment presents an interface rich enough to serve many needs. To be effective, the interfaces for online information systems must be as rich and flexible as the physical environments they replace. They must not only supply a direct path to reach the users’ goals, but must be able to accommodate different approaches to the task. This means that the interface design must not only organize the content for easy access, but must incorporate the right combination of technologies and interaction techniques to allow the user to work in their own style.
USER STYLES IN APPROACHING INFORMATION In common parlance, all user strategies for exploring an information space are referred to as "browsing." This is a convenient simplification, but does not provide enough guidance in how to design an interface to assist the user in meeting their goals. Early hypertext research looked at how people used hypermedia knowledge bases to try to distinguish browsing patterns. One study (Canter et al, 1985) observed users navigating through a hypertext database. The researchers looked at elements including the number of nodes traversed, the characteristics of the path and the ratio of individual nodes visited to total visits. They identified five search strategies, which they called scanning, browsing, searching, exploring and wandering. Of these, only searching was primarily goal-oriented. A later study (McAleese 1989) examined these strategies in comparison to two interface design models – global aerial view and screen-by-screen view. McAleese found a correlation between type of interface design and predominant search strategy. This suggests that design choices can be used to create an interface which is helpful or one which obstructs the user in their work. The original five search strategies are valuable in supplying a descriptive name for observed behavior. However, they do not offer a way to link information from user and task analysis to a design strategy. Nor do they provide guidance on the interaction design required to support these behaviors. An early version of this work (Cognetics 1995) formulated a list of five approaches and matched each with a goals statement from the user’s point of view:
Browse Browsers are explorers. They enter an information space without a clearly defined goal except to understand what is available in this space. Browsers are well understood in the physical shopping culture. They are the people who walk the aisles of a store looking for something to catch their eye. Even without a definite shopping goal, they might make a purchase if a sufficiently tantalizing item is displayed in their path. Or, they may have a goal which is not well-formed and are using browsing activity as a precursor to a more directed approach. Readers of print books can also browse easily, flipping through the pages to find a passage that catches their eye, or the right section of the text. Online interfaces can have difficulty replicating the rapidity and ease of browsing in a well-designed physical space because of the limited resolution and depth of the computer interface. Visualization techniques that attempt to create an abstract representation of a large data set are one solution. An example of this approach is the Market Map on the Smart Money web site (Smart Money 2000), which represents 500 popular stocks as a grid of rectangles. Each rectangle represents one stock and uses color and shape to indicate the size of the company and its recent performance. Once a user has learned to read the map, there is a lot of information packed into a single screen. However, the very abstraction which makes these interfaces work well on-screen has also prevented their general acceptance, as each must be learned individually. Browsing interfaces that present selections in thumbnail images, with links to detailed pages, are more widely used. Another approach to browsing is to provide many links to allow the user to follow lines of interest in a free-form way. Both in-line links embedded within the text and lists of "see also" links to related information serve this purpose. The goals of an interface design for browsing are:
Find Find and Query activities both use what are conventionally known as search interfaces. In both cases, users indicate criteria for the items they wish to locate. The difference is in whether the user already has a well-formed picture of the item they are looking for. A Finder can identify the target and only needs help in locating it within the information space, while a Query approach provides a general description of the desired item without a specific target or knowledge of whether such an item exists. In a shopping interface, these users know the name, model or other key identifiers for the product and can provide this information immediately. An example of a Find activity in documentation is a user who is looking for the syntax of a specific command or the correct procedure for a specific function. Simple examples of Find can be found in interfaces that supply a way to enter a product number such as a book’s ISBN code, manufacturer product name, or software command to go directly to the matching item. Typically, these interfaces skip the presentation of a list of options when only one item is found. Find interfaces can also be integrated into the interface so they operate transparently. In many programming environments, any command or reserved word can be highlighted and a command given to go directly to the syntax documentation. Designs to support finders can include constructed indexes and interfaces that allow the user to enter a product description. The more directly the user can reach the desired item, the more successful the interface is.
The goals of an interface to support a Find approach are:
Query In contrast to Find, people who use Query strategies can provide a general description of the items they hope to locate, but do not have a specific item in mind. Their mental image is well enough formed that they can provide several criteria at once, rather than needing to use a structured approach. The interface should not assume that the user’s objective is unclear because the user cannot identify the target items directly. Query strategies are a typical research approach as users look for items that meet their needs. The interfaces for Query can be difficult to use when they mimic the underlying search mechanism. Users have a notoriously difficult time formulating their queries with the precision demanded by these engines and can be baffled by the visual complexity with which they are often presented. Good query interfaces support the user in this task by presenting examples of valid entries, hiding syntax requirements, and allowing entry to be done in any order. Some advanced interfaces for query combine a visual display of the results of the query with the query itself. This combination can be very effective in helping users fine-tune the search and see the range of information available.
The goals of a query interface are:
Structured A structured approach often relies on a defined hierarchy to assist the user in narrowing the search domain through successive choices. This approach can be as simple as a multi-level table of contents or a dynamically-constructed context based on the user’s previous selections. This approach is frequently used in interfaces for both shopping and help. In these applications, the user is presented with a series of choices from a small selection of options, avoiding the need to present a large, heterogeneous list of items. This technique works well when the hierarchy is unambiguous and the user can make clear choices. Structured interfaces can fail when there are items that might logically belong to several branches; if they can only be found by following one path, they are effectively hidden from a user with a mental model of the information that is different from that of the designer. Failures also occur when choices are presented in language that is not familiar to the user. In this situation, the user may misinterpret the options and follow an unproductive path . Structured interfaces can also be used as teaching tools or as a way to guide users through an unfamiliar domain. A complex problem can be broken into a series of simpler choices, for example. The screen real estate gained by presenting only part of the information set can be filled with user assistance, offering explanations of the choice being made at each step. Another important use for structured interfaces is to allow choices to be presented dynamically, where each successive set of options is determined by the choices made in the previous step. A variant of this is to use a structured approach to build on a query interface, narrowing the results domain by suggesting additional search terms until the set of candidate items is reduced to a number manageable for browsing. The goals of a structured interface are:
Guided A guided interface offers few navigation options in presenting a single narrative sequence. Any interface that presents an authored order for the information is guided. This includes pages in an online book chapter, a multimedia movie or a series of web pages that simply link from one to another in order. A guided interface can be superimposed over a browsing interface by adding controls that allow the user to move through the material in sequence. The "browse" buttons in WinHelp work in this way, as do ordered tables of contents and "next page" buttons in some web sites. Or, a guided interface can be embedded in a larger system as a way to present a constrained task; for example, the ordering systems in many e-commerce sites, which present a series of forms that allow the user to specify items to purchase, shipping address, gift wrap and payment information separately. These interfaces are often limited in scope and short in duration, presenting small granular ‘chunks’ of information. Once the user has identified an information chunk or a task (typically through one of the other approaches) and confirmed their interest in it, they are more willing to follow a sequence as authored with confidence. The goals of a guided interface are:
Users are rarely limited to one approach Typically, a user will engage in more than one of these approaches during a single session. For example, browsing activity can be transformed to a find approach when the user sees something that triggers an idea. A query that returns too many items can be successful if converted to a structured approach. And, depending on the kind of content material, the end of any search might be a short guided session to allow the user to absorb the information as presented. It is important that the interface be able to handle these changes of style, rather than assuming that one is pursued to the exclusion of others. An online shopping scenario provides one example of how the user’s approach is transformed during the course of a single experience.
In this scenario, browsing, structured, query and guided were all used for different aspects of the experience. It is important not only that all of these approaches are supported in the interface, but that the user was able to move from one to the other seamlessly. Too often, the structured table of contents can only be reached from the main menu or search functions do not start from the user’s current context. It may be easier to design and program in discrete modules, but this separation of functions does not enhance the user’s experience. TECHNOLOGIES FOR PRESENTING INFORMATION A dynamic structured interface requires technology to determine and present new choices. In looking at user styles, the success of the interface depends in part on the capabilities of the underlying technology. For example, both find and query rely on a search engine, browse and structured on hyperlinks or other techniques for connecting related information. To design an information system or user interface that meets user’s needs, the author must ensure that the technology will support the appropriate strategies for approaching information. Authoring technologies can be categorized as:
Multimedia Multimedia programs integrate graphics, text, audio, video and animation into an interactive presentation. They are often used for presentations, kiosks, self-running demos and training. Their advantages center on the presentational control they offer the author and the ability to use movie-like combinations of media. Typical multimedia tools include Macromedia Director and Flash, RealVideo and other video editing tools and some computer-based training tools such as Macromedia Authorware. Multimedia is strongest in a guided approach, but can support browsing with appropriately indexed navigation controls. Digital paper Digital paper is especially seductive because it offers the possibility of a display which is identical to a print-based publication. In that sense, it is a transitional medium filling a gap between our long experience in designing for paper and the newer medium of the screen. As online information becomes the rule rather than the exception, the value of looking "just like paper" will diminish. An advantage, however, is digital paper’s strong cross-platform support and the ability to deliver high fidelity print-on-demand. Although there were several early competitors, the Adobe Acrobat format has emerged as the primary program in this arena. On its own, digital paper is a guided interface, but with the addition of a table of contents, bookmarks or other links it can support a structured or browse interface as well. An integrated search engine can add limited support for query. Online books Before the explosion of the World Wide Web, online books could be defined as a type of stand-alone program that produced a publication for online delivery. A few of these programs still exist to serve the need of publishers and software developers to deliver books and documentation in electronic form. But even on the Web, it is still important to distinguish coherent bodies of information, with a ‘beginning, middle and end’ presented online. These sites often have a normal reading order, even if hypertext techniques can be used to provide direct access to specific content. In addition to hypermedia authoring tools, many word processors or desktop publishing programs now offer a way to convert content directly into an online format. Online books are the most varied, but most support a structured interface through a table of contents as well as a narrative guided sequence. With good navigation and authored hyperlinks, online books can support browsing well. Their support for both find and query depend on the search engine available, though they do not tend to have enough fielded criteria for a very powerful search. Contextual Help The world of contextual help – general assistance or detailed instruction for specific functions in a software program – may not be visible to the general user as a special category of program, but it has its own authoring tools and technology. These programs are designed to produce online help in the form of details on demand, step-by-step instruction, identification of screen objects and other forms of on-screen assistance. Their biggest disadvantage has been that these programs operate as a separately, co-existing but often not well integrated with the software they support. New technologies to support embedding contextual help in the interface eliminate some of the usability pitfalls of these programs. The primary approach for contextual help is find – direct access to specific information. They tend to have a structured interface when accessed without context, and may have limited support for browse, and query. Reference databases Databases are collections of documents which can be accessed through search-and-retrieval techniques. These programs typically store not only the text of the document but also searchable fields which supply categorization, identification and keywords as well as access to full-text search. Many programs which supply full text search without the fielded criteria, but they rarely create as effective interfaces for users since they support neither Find nor Query well. Reference databases’ strength is in their ability to manage very large information spaces containing a wide variety of documents, but they rarely offer strong native presentation or display tools. Text-oriented databases may not handle multimedia or graphical information well. Typical document database tools include Dataware and Lotus Notes as well as document management systems like Documentum, although standard databases such as Oracle or M/SQL can also be used. A database’s primary approach is query, with find a close second. They may support browse through hyperlinks or special visualization techniques. Choosing the Right Technology Until recently, these programs tended to be proprietary and isolated. Few of them worked together, so the choice of an authoring tool also dictated the capabilities of the interface, enforcing its own point of view through a limited feature set. In the five years since this material was first presented (Quesenbery 1995, 1997), the Web has exploded and authoring tools have added more features to accommodate the demands of both designers and users. This categorization remains useful however, for two reasons. First, it provides guidance in selecting the type of presentation technology for specific information or interfaces. Second, programs remain true to their roots, no matter how many new features are added. It is still important to consider a program’s strengths in assessing its ability to meet user needs. Here’s an example. When the editors at Primary Source Media first designed the American Journey CD-ROM series, they had to make a choice between a hypertext program with nominal search capabilities and a database program with poor text display and linking features. In making the decision, they considered both their goals for the discs (a collection of historical reference material packaged with background essays for high school students) and how they envisioned the students using the disc. They decided that browsing and structured navigation would be the most important approaches and that an attractive book-like display was more important than a complex boolean search capability. Once stated in that way, the choice of the hypertext product over the search engine was obvious. Including approaches to information as a decision factor made it possible to select the correct technology to support users needs. MEETING USER NEEDS It’s an exciting time for technical communication. More information is going online, and that information is more critical to both business and personal lives. In the work place, everything from human resource information to procedure manuals are on the corporate intranet. In their private lives, people expect to hop on the web to find not only things to buy, but resources for health, financial, hobby and community information, and they hope to find this information easily. If this information is not presented in ways that make it easy to locate and read, users will seek elsewhere. Meeting this expectation will require new skills from all information designers, including technical communicators. User-centered approaches from user and task analysis to presentation design are all important, but the designer (and writer) must also consider different user approaches to the information in order to meet user needs. The underlying technology must also be considered, as it affects the style of as well as the functionality available to users. To meet user requirements for usable online information, all of these elements must be factored into the design – and technical communicators must master the skills necessary to make the right choices. REFERENCES
This paper was published in the the STC journal, Technical Communication, Volume 48, No.2, p.182-188 May 2001 It was awarded the 2001 Frank R. Smith Outstanding Journal Article. Adobe Acrobat, Dataware, Documentum, Lotus Notes, Macromedia Authorware, Macromedia Director, Macromedia Flash, M/SQL, Oracle, Real Video are all trademarks of their respective owners. The URL for this article is: http://www.wqusability.com/articles/on-beyond-help.html Whitney Quesenbery works on user experience and usability with a passion for clear communication. She is the co-author of Storytelling for User Experience from Rosenfeld Media. Before she was seduced by a little beige computer, Whitney was a theatrical lighting designer. The lessons from the theatre stay with her in creating user experiences. She can be reached at www.WQusability.com |
