|
|
Karen Holtzblatt
Hugh Beyer
InContext Enterprises, Inc.
249 Ayer Rd. , Suite 301
Harvard, MA 01451
telephone: (508) 772-0001
email: karen@acm.org, beyer@acm.org
Field data gathering techniques such as Contextual Inquiry enable a design team to gather the detailed data they need. These techniques produce enormous amounts of information on how the customers of a system work. This creates a new problem-how to represent all this detail in a coherent, comprehensible form, which can be a suitable basis for design. An affinity diagram effectively shows the scope of the customer problem, but is less effective at capturing and coherently representing the details of how people work. Design teams need a way to organize this detail so they can use it within their own development process.
In this tutorial we present the latest methods for representing detailed information about work practice and using these representations to drive system design. These methods have been adopted over the last few years by major product development and information systems organizations. We show how to represent the work of individual users in models, how to generalize these to describe a whole market or department, and how to use these to drive innovative design. We present the process by which we build and use the models and practice key steps. We show how these methods fit into the the overall design process, and summarize Contextual Design, which gathers field data and uses it to drive design through a well-defined series of steps.
The tutorial is appropriate for those who have used field techniques, especially Contextual Inquiry, and would like to put more structure on the process of using field data.
analysis methods, design techniques, customer-centered design, ethnography, usability engineering, methodology, team design, domain analysis, work modeling, software engineering, task analysis, user models, user studies work analysis
Systems and products are built to help people work better. They cannot be built well without understanding how people work. Techniques such as Contextual Inquiry gather the necessary data, but producing a good system requires that the data and its use be incorporated into a coherent design process. Such a design process would lead a team from data about specific users to a design addressing the needs of an entire market or department, providing ways to represent the design and iterate it with users. So good system design starts with a deep understanding of how people work.
But understanding work is hard: there is no discipline of understanding how people work, the concepts, distinctions, and issues of work practice are not general knowledge, and we have no language for describing work practice. Without a language, it is hard to communicate work practice to others. To remedy this deficiency, we have developed work models, drawings that incorporate important distinctions about work. These models show the roles people play in the organization and how they communicate; the social and emotional context in which work happens; the sequence of actions which accomplish work; the details of the physical site and work place in which work happens; and the artifacts which support work and capture work results.
In this section of the tutorial, we introduce the problem of representing customer work for design. We discuss the need for a langauge in which to talk about key work distinctions. We review some existing modeling techniques and discuss the distinctions they are designed to reveal. We introduce the work models we have developed and show how they offer multiple perspectives on work practice.
Modeling customer work is a natrual extension of Contextual Inquiry. We show how to capture work models in an interpretation session simultaneously with capturing affinity notes, using a video of a working team to illustrate the process. We have found that the work models shorten interpretation sessions significantly, because they structure and simplify the discussion.
We describe each type of work model in turn and give participants practice building each type in pairs from a simplified transcript.
Products and systems are built for sale to a market or use by a department; they are not built for individual users. But we gather data from individual users-how do we represent what these users tell us about all users? Without a well-defined way to generalize from specific users, we appear to be designing from anecdotal evidence. Work model consolidation is such a well-defined process, resulting in a small set (5-7) of work models which characterize the work structure and basic work strategies across all customers. These models can be shown to account or fail to account for the work practice of any individual user.
In this section, we describe and demonstrate the consolidation process for each type of work model. We show how each consolidated model reveals issues to which a design must respond. Participants work in pairs to read and derive issues from consolidated models of each type.
Ultimately, system design is the invention of the system's response to a user problem. Without adequate customer data this invention is ungrounded-it is not driven from deep knowledge of how people organize their work, and cannot be developed in its details to support customers' work well. Without a coherent understanding of work, design tends to degenerate into lists of features that do not consider the system as a whole, or that depend on the designers keeping the whole system in their heads. Work models capture precisely the detail necessary to ensure invention is grounded in customer work, provides a coherent solution to a whole work problem, and is developed to support actual work practice.
In this section, we describe the grounded brainstorm process by which we do this. Participants use the issues they identified from the consolidated models to brainstorm design alternatives, evaluate them, and synthesize the best of each alternative into a new design.
We conclude the tutorial with a brief discussion of how to use consolidated models to work out the details of the new design. We offer some alternatives for driving the design process to completion, including Contextual Design.
A clearly defined, customer-centered design process guiding a team from initial data gathering to system design is possible today. In this tutorial we lead participants through the key transition in that process: making customer work practice real, generalizing to a market or department, and inventing a system response which remains true to the data. We show how the material in this tutorial links to field research techniques such as Contextual Inquiry and to the later design process which formalizes and elaborates the initial design.
H. Beyer, "Calling Down the Lightning," in IEEE Software. September 1994, Vol 11 No 5, p. 106
B. Curtis, M. I. Kellner, and J. Over, "Process Modeling," Communications of the ACM, September 92, V 35, No. 9.
P. Ehn, Work-Oriented Design of Computer Artifacts. Gummessons, Falkoping, Sweden 1988, international distribution by Almqvist & Wiksell International, also Coronet Books, Philadelphia, PA.
K. Holtzblatt and S. Jones, "Contextual Inquiry: A Participatory Technique for System Design," Participatory Design: Principles and Practice. Aki Namioka and Doug Schuler (Eds.), Hillsdale, NJ: Lawrence Earlbaum Pub. 1993.
K. Holtzblatt and H. Beyer, "Making Customer- Centered Design Work for Teams," Communications of the ACM, October, 1993.
K. Holtzblatt and H. Beyer, "Representing work for the Purpose of Design," in Representations of Work, HICSS Monograph (Hawaii International Conference on System Sciences), January 1994. Lucy Suchman, Editor.
K. Holtzblatt and H. Beyer, "Contextual Design: Principles and Practice," Field Methods for Software and Systems Design. D. Wixon and J. Ramey (Eds.), John Wiley & Sons, Inc., NY, NY, (forthcoming).
K. Holtzblatt, "If We're a Team, Why Don't We Act Like One?", in interactions, July 1994, Vol. 1 No. 3, p. 17
M. Kyng, "Making Representations Work," in Representations of Work, HICSS Monograph (Hawaii International Conference on System Sciences), January 1994. Lucy Suchman, Editor.
P. Sachs, "Transforming Work: The Role of Learning in Organizational Change," in Representations of Work, HICSS Monograph (Hawaii International Conference on System Sciences), January 1994. Lucy Suchman, Editor.
L. Suchman, ed. Communications of the ACM issue on 'Representations of Work', September 1995.