User-centered design is a term used to describe the process by which products and services are researched, tested, and developed while keeping the end-user's needs and desires constantly in mind. In this broad definition, "design" can refer to work in the fields of Human Computer Interaction (HCI), industrial design, engineering, and computer science. All of these disciplines are engaged on some level with creating products and services for people. In order to better understand people and their needs, practitioners in these fields started using social science techniques and methods in the 1980s (although there are examples of using social science to influence design going back to the 1920s and 30s) [1]. Increasingly, corporations are hiring social scientists to work with designers and technologists to inform and influence the design process. Instead of starting by creating a thing, a machine, a widget, and then hoping it will work when used by people, the idea behind user-centered design is to start with people, their needs, and the context in which a device, object, or service will be used. Ethnographic methods, including video documentation, are an extremely useful way of bringing "users," i.e., real people, into the early stages of the design process as examples of current practice, inspiration for future practice, and guides into the context in which a product or service might some day be applied.

This paper will not focus on all the numerous ways video can be used in a user-centered design process, but will concentrate on why video is an effective tool for communicating findings and research insights to colleagues during the early stages of that process.

While documentary footage is an effective tool for communicating insights and inspiring design ideas, it is usually one of several tools used and usually does not stand alone. Documentary-style video, in other words, video footage that is shot as an interview or observation unfolds, without a script or direction, needs interpretation to be most effective. Visual representations are particularly vulnerable to misinterpretation if the viewer is given no guidance or context for what he or she sees. It is up to the ethnographic researchers to interpret the footage, analyze the practices of the people being studied, synthesize that data, draw conclusions, and convey what they have learned to their corporate colleagues.

Field notes are another way of documenting field work. Researcher's written field notes comprise observations he or she felt were interesting or significant at the time. It is impractical to write down everything that happens during an ethnographic visit. Notes are jotted down in the moment, and a more complete account is written at a later time at the office or hotel. For that reason it could be erroneously argued that videotaping an ethnographic interview results in a more "truthful" or more "accurate" account than simply relying on the researcher's notes and recollections. While conducting an interview, researchers are often asked by study participants why the visit is being videotaped. A common reply is, "So I don't have to rely on my notes to remember you, and what we talk about today." The video serves as a record.

A videotape does provide a detailed record of an encounter. It can be transcribed, and if the audio recording of the visit was done well, the result will be a very thorough record of everything that was said (near the microphone) during the interview. The videotape and the transcript become a version of the truth.

What this notion of truth and reality doesn't take into account, however, are the numerous decisions that have to be made during, and the prevailing circumstances of, the filmmaking process. How a shot is framed, where the camera is placed, what questions are being asked, who is being heard, and who is not, and whose perspective is being shown: all of these can dramatically influence a videotaped account of field work. The researchers' impressions of a particular family, for instance, may be significantly altered by the presence of some family members, and absence of others, or the necessity to focus the camera on one part of a family while other members cook dinner, go to work, nurse a sick relation in another room, etc. From a practical standpoint, this is because research teams are generally small, and a videographer (or videographers) must choose where to be, where to point the camera, and whom to follow when the action changes. But no matter how many cameras, researchers, and willing participants are present, it is not possible to capture on videotape every aspect of a life, a home, a workplace or, of course, a culture. Additionally, how the final product is edited, scored, subtitled, and presented also play a large role in determining the impact the footage will have, and the insights it will impart. Even a film that is many-hours long is subject to interpretation by both the filmmaker and the audience. Frederck Wiseman, a documentary filmmaker known for creating long films devoid of narration, interviews or commentary, has said:

I don't manipulate the events, but the editing is highly manipulative and the shooting is highly manipulative, not in the sense that people do things differently from what they will ordinarily do, but the way that people are shot. First of all what you choose to shoot, the way you shoot it, the way you edit it and the way you structure it [2].

In a user-centered design research setting, many hours of videotape are reduced to a few key minutes in the interest of saving colleagues' time, capturing short attention spans, and communicating that which is most compelling, insightful, and germane. Because of this brevity, video ethnographers are sometimes asked by colleagues, "How do we know what really happened when we only see footage after it has been edited?" It is a legitimate question. How does a fellow researcher know whether the video being presented accurately reflects reality? How does he or she know how to "read" the images? When does video need to be framed and contextualized for the viewer? Why can or can't it stand on its own? How does the specific use of video for ethnography and user-centered design research help determine how it should be viewed and by whom? Should different audiences be given different levels of contextualization and complexity? What is appropriate to present in an active, workshop setting, and what works in a relatively passive presentation?

Filmmaking and ethnography are both interpretive by nature. Both are framed and constructed by one person (or several people) to convey a particular point of view and both are interpretations of people and events. As Nafus and Anderson [3] have pointed out, a common pitfall of using visuals for ethnography in industry is that pictures (or videotape) are thought to "show the real," and are often considered by a corporate audience not to require interpretation or analysis.

While visuals can be problematic, they can also provide richness and depth to field data, and when properly framed and contextualized for an audience those pitfalls can, hopefully, be avoided or diminished. In this paper I do not address the need to move on from the "real period" that Nafus and Anderson assert, and will not cover the innovative applications of documentary filmmaking discussed by Raijmakers, et al., as new methods of using video in user research [4].

Rather, in this paper I demonstrate the power of using video for ethnographic research in an industry setting, and I examine some of the complex issues video representation can elicit. There is no single "reality," but through interpretation and analysis, researchers can create a version of what they witnessed in the field that speaks to a theme or gives insight that is relevant, inspirational, and useful. The result is a representation of reality that is both grounded in what the ethnographers experienced and instructive for colleagues who were not there. This version of "reality" can be powerful, meaningful, and highly effective in the design research process.