A Web on the Wind: The Structure of Invisible Work

A Special issue of Computer-Supported Cooperative Work

Bonnie A. Nardi and Yrjö Engeström, guest editors.
Forthcoming, August, 1998


Bonnie A. Nardi

What is work? Efforts to design and introduce new technologies and to restructure the workplace (which often involve new technology) will be most successful when built on a firm foundation of knowledge about how work actually gets accomplished.

Much work is visible. It yields to being being mapped, flowcharted, quantified, measured. When planning for restructuring or new technology, visible work is the focus of attention. It is the only work that is seen, so efforts to restructure center on how visible work can be manipulated, redrawn, reorganized, automated or supported with new technology. But a growing body of empirical evidence demonstrates that there is more to work than is captured in flow charts and conventional metrics.

This special issue documents four kinds of invisible work: (1) work done in invisible places, such as the highly skilled behind-the-scenes work of reference librarians, (2) work defined as routine or manual that actually requires considerable problem solving and knowledge, such as the work of telephone operators, (3) work done by invisible people such as domestics, and (4) informal work processes that are not part of anybody's job description but which are crucial for the collective functioning of the workplace, such as regular but open-ended meetings without a specific agenda, informal conversations, gossip, humor, storytelling.

Understanding the nature and structure of invisible work is crucial to designing and managing organizations. When organizations are restructured and work is reorganized, invisible but valuable work is often eliminated. No one recognizes that it is being done, or that it is of value, so the time and personnel it requires are not allotted in new plans. Recently at Apple Computer, for example, the Apple Library, considered one of the best, if not the best industrial library in the world, was closed down during budget cutting. One famous high level executive was shown the library in September, 1997, in a last-ditch effort to persuade him of its importance before the budget ax was about to fall. His remark to the head of the library was, "The engineers should already know all this, shouldn't they?" Another executive said the library was unnecessary because it simply encouraged employees to waste time getting clever quotes for their slide presentations. The invisibility of the work of the librarians led (in part) to the Library's demise. In other settings, librarians are targets for elimination because it is believed that they can be replaced by software. For example, some government officials do not understand the importance of what librarians do, and are willing to cut library funds because they believe that "digital libraries" and computer-based services can successfully replace human librarians.

Invisible work also occurs in the case of workers' input to continuous improvement and development of practices, products, and technologies. Such work has tacitly become an increasingly important layer of "second order work" on top of regular work duties.

Sometimes new work processes are introduced that undermine good but invisible work being done by employees. For example, the work of telephone operators is often defined by telephone companies as rote and routine, amenable to being tightly scripted on a second by second basis (saving the phone company money). Upon closer scrutiny, it is found that telephone operators frequently solve problems for customers in a resourceful, proactive manner (e.g., "My heat has been turned off. Who do I call?") Attempts to constrain and reduce interactions with customers result in poorer customer service and less satisfaction with phone company service, leaving companies vulnerable, over the long run, to competition. In another arena, many banks that have cut back customer service in an effort to reduce costs, are now being openly challenged by "boutique" banks that emphasize personal service and human contact.

We call this issue A Web on the Wind after the work of the Russian physiologist Nikolai Aleksandrovich Bernshtein (1896-1966). Bernshtein proposed that the rudiments of complex physiological phenomena were like a web on the wind: highly structured, but floating and difficult to detect unless one looks carefully. A gossamer web on the wind is visible if you train your eye on it, though it floats past undetected when you are unseeing. The web has a complex, even beautiful structure, and is both as strong and delicate as a finely spun web. Exploring that gossamer--both its strength and delicacy--is the objective of this issue.

The contributors to this issue come from varied backgrounds including psychology, anthropology, economics, design, artificial intelligence, sociology, and education. Theoretical backgrounds are diverse, including ethnocriticism, activity theory, Weberian perspectives, grounded theory and social criticism. As the papers in the issue will show, invisible work takes many guises: as tacit and contextual knowledge, as informal social networks, as expertise acquired by old hands, as long-term teamwork. These things disappear in bloodless diagrams showing the putative advantages of "matrix" organizations, in blithe proposals for "intelligent" software agents.

Because of the extreme division of labor in postindustrial society, work is, in a sense, always invisible to everyone but its own practitioners (and even they are not always aware of their own special expertise and how it functions). The question is: when do we need to make the invisible visible? There are many answers depending on the situation. Visibility and invisibility are neither good nor bad in themselves. There may be costs to revealing or concealing expertise and work. As the papers discuss, visibility and invisibility are not monolithic quantities; they are relative to various perspectives within an organization. As designers of CSCW systems, it will be in everyone's best interest if we get a better grip on these perspectives as we attempt to create useful systems that meet human needs.

The issue opens with Leigh Star and Anselm Strauss's paper, "Layers of Silence, Arenas of Voice: The Ecology of Visible and Invisible Work." Star and Strauss ask the fundamental question: "...what exactly is work, and to whom it might (or should) be visible or invisible." Their paper accomplishes many tasks, including a focused review of the literature and a discussion of three kinds of invisibility: creating a non-person (e.g., the domestic worker), disembedding background work (e.g., the work of nurses whose activities disappear into the background) and abstracting and manipulation of indicators (people and work are both rendered invisible as they dissolve into a set of formal or quantitative indicators). Star and Strauss note that the idea of invisible work is not new: the social critic Ivan Illich described and discussed what he called "shadow work" almost twenty years ago. That we are all-even highly paid, highly skilled workers-subject to misinterpretations of the work we do is underlined by Star and Strauss's observation that physicians are increasingly concerned with the control exerted by HMO's which "effectively bar[s] them from practicing proper medical care." Further, Star and Strauss observe that technologies such as "affective computing, smart badges, and other forms of linking presence, thought and emotion in advanced CSCW applications," may allow for monitoring and quantitatively assessing even the most creative work done by the top tier of workers.

Michael Muller's paper, "Invisible Work of Telephone Operators: An Ethnocritical Analysis," relates the results of an in-depth study of telephone operators at US West, a large telephone company. Muller uncovered the extensive knowledge work done by telephone operators as they helped customers work through ill-formed requests for directory assistance. Muller forcefully makes the point that this knowledge work had been missed by both management because of their preconceptions about what operators were doing, and by researchers whose research methods effectively prohibited them from seeing what was happening as operators worked with customers . Muller notes that research paradigms lead to invisibility because they often achieve depth at the expense of breadth; for example, in the case of GOMs modeling and other related techniques. These methods are useful up to a point, but they are of no help in uncovering non-repetitive, non-routine, conceptual work. Most crucially, such techniques "do not let the researcher know that something has been left out." Muller discusses how the notion of the "canonical call" rendered the variability of actual calls invisible and led to a poor design for a partially automated directory assistance system. Using the CARD method, a participatory design method with simple visual tools that can be very effective in guided sessions presented by a skilled facilitator, Muller and his colleagues redesigned a directory assistance system to re-pace the dialog between machine and human operators, changing the allocation of time and content between human and machine segments of the dialog. The new system was measurably better than the first prototype which was not informed by their research. Measures included customer acceptance, operator acceptance and time spent on each directory assistance inquiry, i.e., work-time savings-the bottom line. Management was enthusiastic about the research and acted upon the findings.

Yrjö Engeström's contribution, "Expansive Visibilization of Work: An Activity Theory-Theoretical Perspective," also points out the problem of theoretical and practical representations and how they can render work invisible. He notes that such widely used paradigms as business process reengineering confine discussion to "describing how a given process would operate with all the external and internal performance measures optimized." A corrective to this approach is provided by ethnographies of work that stress the situated nature of action. They do not go far enough however, believes Engeström, because they do not ask, "what is driving people in their work," nor do they consider "how transformations in the collective organization of work are accomplished." Engeström is concerned with what he calls "developmental turmoils"-how an organization can change and grow as contradictions and problems in the organization are confronted. Engeström reports on a successful intervention in a Finnish hospital. Using methods developed in the developmental work research tradition in Finland (and other European countries), he and his colleagues intervened in the reorganization of a large children's hospital. As in Muller's work, simple visual tools based on his theoretical orientation provided a point of discussion with those whose lives would be affected by the organizational changes and redesign of work and technology.

Kristina Westerberg's paper, "Collaborative Networks among Female Middle Managers in a Hierarchical Organization," uncovered the way a formal, visible hierarchical organization of home health care workers co-exists with informal, invisible horizontal social networks. Both are crucial to the work of the health care managers in Sweden. The informal network is used to solve problems and make judgments that escape the boundaries of the formal hierarchy. The informal networks are not persistent-they are built and rebuilt as needed, providing a flexibility and elasticity to the health care system. Westerberg makes clear that she does not privilege one system over another: both serve important functions in making sure that Swedish citizens receive high quality home health care. She observes that a new computer system planned for the workers served only the visible hierarchical structure. Formal functions fulfilled by computer applications such as form filling and statistical analysis were welcomed by the health care managers and made the work easier for everyone. Westerberg notes, however, that the invisible horizontal network could also be served by enhanced communications supported by computers-but this never came up for discussion. Westerberg observes that making the work of the horizontal networks visible will have to be undertaken if computer support is to be forthcoming. So here is a case where "visibilization"-to use Engeström's term-is crucial for further improvement in the organization's use of technology.

Libby Bishop utilizes a macroeconomic viewpoint in her paper, "Visible and Invisible Work: The Emerging Post-Industrial Employment Relation." Like Westerberg, she documents the co-existing formal and informal arrangements by which work gets done. Using Weber's classic distinction between the instrumental and status contracts of work, she discusses the tensions and contradictions emerging in today's workplace. The instrumental contract is the visible, agreed-upon performance or result promised by worker and employer. But by its side is always the much less visible status contract that affords identity, status and a sense of connection to others in the enterprise and to the enterprise itself. As Bishop says, "The elements of work address not only what I make...but who I am." These relations exist in tension, especially in times of economic change. For example, with the current shift to temporary work, outsourcing, and contracting, the status contract comes under fire. It is less easy to establish an enduring identity as a member of a particular enterprise. When we are all "consultants" our status is less visible to ourselves and others. Whatever we might say against hierarchy, it gives people a clear sense of their relations to others and an unambiguous view of the ladder they must climb if they wish to enhance their status. Technology enters the picture because the flexibility of moving workers around between companies, locations and functions is possible in part because of advanced communications technologies. Bishop calls this new fluidity the "exile from the office." It includes diverse manifestations such as the need for workers to find new jobs as companies downsize, the implementation of telecommuting, and office redesign schemes like hoteling in which workers have no permanent personal space. Stripped of space and enduring connections, we become what Bishop calls "worker widgets." Bishop reports several cases studies that show that invisible informal communication work-a crucial part of the status contract- is often severely disrupted as companies try to rationalize all work functions. The status contract, being diffuse and lacking explicitness-for good reason-is damaged and then puts pressure on the instrumental contract to become even more explicit. When that happens, awkwardnesses and distortions in communication occur as people are forced to be too explicit; e.g., in the case of the use of software systems such as the Coordinator or the need to follow highly scripted communication protocols. Sometimes we need to have the freedom to not be accountable! With small deceptions and fuzzinesses, we can smooth feelings, rearrange schedules to accommodate pressing needs, sneak off when we need to sneak off. If everything is blazingly explicit, these fine-grained adjustments to the work day become impossible, resulting in stress and tension. Bishop provides an important reminder that "visibility is not always advantageous." Bishop's work is part of the tradition of "interaction research" in which practical outcomes are targeted. She offers a case study of an office redesign that she participated in where a better design emerged when workers had a say in how high partitions between cubicles should be, and were allowed to decide on hoteling vs. dedicated space. Workers chose to accommodate a new team member with dedicated space instead of hoteling by decreasing their own square footage. Their interest was in protecting the informal communication of the status contract (though they probably hadn't read Weber!).

The Bishop, Muller and Engeström papers utilize different theoretical frameworks to delve into issues of work and change, with an eye to practical, measurable outcomes. The success of their efforts underscores the fact that theory does matter (or, rather, it can matter when handled wisely). These studies of invisible work are not merely academic exercises; they have had profound practical implications. In her paper, " `It's Just a Matter of Common Sense': Ethnography as Invisible Work," Diana Forsythe turns the analytical lens on ethnographers-those who have made significant contributions to uncovering everyone else's invisible work. Forsythe notes with irony that now that ethnographers have convinced researchers and corporations of the value of ethnographic work in technology design, they face a new and unexpected problem: appropriation of their methods. Ethnography does look easy. It's just talking to people, right?! Three day training classes are offered in ethnography-how hard could it be? Forsythe carefully deconstructs the skills an ethnographer actually brings to his or her work, and makes clear that doing ethnography properly is not learned in five easy steps. Forsythe includes some engaging personal anecdotes from her years working in medical settings, including the time she was called by a physician a "walking tape recorder." The physician saw only that she was making copious notes and did not understand the analytical rigor with which she was sifting through (in real time) the myriad sensory data coming at her in the rich medical encounters she observed. Her intellectual work was invisible. Forsythe discusses how the invisibility of ethnographic work can lead to a situation in which the ethnographer's field notes and interview transcripts are considered raw material that anyone can appropriate and analyze. (This is a surprisingly common situation that I have experienced myself.) One physician suggested posting Forsythe's interview transcripts on the Web! Forsythe argues that field notes and interviews are intellectual capital or intellectual property that belong to the ethnographer. Not just anyone can produce a careful interview or make the detailed nuanced observations that an ethnographer can. Drawing on Star's concept of "deletion," Forsythe observes how certain kinds of activities are "deleted," or simply not considered salient, from various kinds of accounts. In essence, we get into the habit of ignoring that which is not considered interesting-even with respect to our own work . For example, Forsythe reports from her studies of artificial intelligence researchers how technical people describe the work they do in terms of programming or system design, but consistently delete social activities such as meetings and other kinds of social interaction. The work would not get done without these interactions, but they are deleted by the researchers as inferior to "the real work." Forsythe discusses deletion in people's perceptions of ethnographic work.

It seemed appropriate to end the issue with a cautionary tale, and Kate Ehrlich and Debra Cash complete the issue with "The Invisible World of Intermediaries: A Cautionary Tale." They pinpoint a particularly salient issue for design in today's world: the danger of eliminating human intermediaries such as librarians, editors, various kinds of agents, and customer support representatives. They observe that the work of intermediaries can be invisible from several points of view: that of the customer, the organization (like the health care workers' horizontal networks in Westerberg's study), and the intermediaries' managers (like the telephone operators Muller reported on). Erlich and Cash remark on the pervasive rhetoric about how technology can replace people, noting that Bill Gates, for example, argues in his book The Road Ahead, that consumers will be blissfully united with producers, eliminating what Gates refers to as the meddling "friction" of intermediaries such as sales people, stock brokers, real estate brokers, investment advisors, insurance agents, travel agents. Erlich and Cash point to another endangered class: those who "find, filter, sort and interpret existing information," including journalists, reviewers, librarians, editors and customer support representatives. They note that the critical invisibility here is that the user's task is conceptualized as simply connecting with raw information, when what users actually often need is a way to make sense of the information. Using data from their extensive studies of librarians and customer support representatives, they document the valuable services provided by human intermediaries that are not made unnecessary by direct end-user access. Ehrlich and Cash suggest ways to fruitfully bring together technology and people to better serve human needs, rather than eliminating humans from the equation.

The time that elapsed during the development of this issue engendered both joy and great sadness. The joy was in working with an extremely thoughtful group of writers and scholars. The sadness was the occurrence of the deaths of two of our authors. After a long and extraordinarily productive life, Anselm Strauss, the co-author of "Layers of Silence, Arenas of Voice," died of a heart attack. Diana Forsythe, author of "It's Just a Matter of Common Sense," died tragically in a drowning accident in Alaska in July, 1997. One of the last things she did before her death was drop off on my front porch a Mac disk with the final version of her paper. Both Anselm and Diana are missed very much for their scholarly contributions and their friendship. This issue is dedicated to them.

In memory of Diana and Anselm, we hope this issue will enable people to see work differently, to talk about work differently, and to act differently when confronting the realities of work and the people who do it. We want the invisible web on the wind to become a part of our everyday perception, to heighten awareness of what it really takes to get work done. We hope to affect the way productivity is defined and measured, how managers allocate resources, how scholars theorize about work, and how we can take account of both visible and invisible work as we design CSCW systems.


I am grateful to the authors for their creativity, insight and the care with which they composed their papers. It was an enriching experience to receive the fruits of their empirical labors and theoretical analyses. Because of Diana Forsythe's participation in the issue, she and I had a ready-made excuse for getting together, and I was lucky to experience several very special lunches with her before her untimely death. My co-editor Yrjö Engeström provided perspective and the big picture at times when I was drowning in the details of editing. Most enjoyably, the authors were kind enough to acknowledge from time to time (in that incomparable medium, email) the invisible work of their editor. Speaking of invisible, our anonymous reviewers made cogent suggestions and materially improved the papers. Kjeld Schmidt was a supportive, patient and helpful journal editor.

Last updated May 17, 1998

back to ... Bonnie Nardi