The Ethnographer Unbounded: Considering Open Source in Corporate Environments

Introduction

We are currently working in an era of friction, as information is increasingly unowned and untethered, there are more efforts to control it, and attempts to increase stringent restrictions and laws trying to curb the right to information. Controversies regarding open access to information have recently fomented a series of protests, for example, against legislation such as PIPA/SOPA and the Research Works Act. Even professional organizations, such as the American Anthropological Association, are struggling to decide how open to make information (Boellstorff 2012; Golub 2012). These debates have led us to question how we, as practitioners of ethnography, promulgate information. Ethnographers in both the public and private sector can blog, tweet, pin, comment, bookmark, and share information across a multitude of open (available to all, at no cost) online platforms. These tools help spread ideas, promote ethnography, enhance collaboration, and further one of EPIC’s central aims to “promote public recognition of practicing ethnography as a profession.¹ We argue that in addition to recognizing the profession in the public sphere, ethnographic research itself should be promoted, as openly sharing content such as interview transcripts, analysis, commentary, applications, and resulting deliverables may also benefit practitioners and the public in surprising ways.

With this in mind, we set out to explore how “openness” of research impacts our own ethnographic work by comparing two distinct professional experiences. The first occurs within a large corporation, in which almost every stage of work is undertaken internally and confidentially. The second is an open source project, funded by a Knight Foundation grant, relying on collaboration across several public institutions. By comparing these projects, the authors explore the benefits and the tensions of working on “open” projects and seek to learn which elements of open source have beneficial applications for ethnographers working in corporate contexts. We then propose utilizing a “layering” model, along with a set of questions, that can help disentangle the openness issue. The amount of information that is shared, with whom, in which formats, at what times, are framed as questions which practitioners can ask themselves at any given point in a project to assess whether sharing information would be beneficial or not. The depth and breadth of sharing can be considered on a project-by-project basis, at multiple project stages.

We aim for openness because “closing” ethnography threatens to widen the gap between ethnographic output and a public audience; to stifle opportunities to make meaningful contributions to a wider variety of public and social issues; to stagnate interpretations of ethnographic data over time; and to limit the formats of ethnographic data representation in the public domain. All of these shortfalls splinter ethnography in the public eye, or worse, render it mysterious or meaningless. The fact is, technology and open access models are changing the way we work and communicate, and it is not a choice of whether or not to jump on board, but simply to what degree and how.

Information Societies

Living and working in an industrialized nation with widespread broadband access, we find ourselves in an “information society,” where information and its circulation are at the center and where a system of social relations are oriented toward the production, commoditization, circulation and manipulation of information (McSherry 2001). It will never be more difficult to copy digital material. Legislators, corporations, and academic organizations can oppose open information exchange, yet technology makes it inevitable. Cory Doctorow writes, “Just as the industrial economy wasn’t based on making it harder to get access to machines, the information economy won’t be based on making it harder to get access to information…the more IT we have, the easier it is to access any given piece of information” (2008, 59).

In addition to ease of sharing information, the boundaries around online privacy, in the sense of regulations, corporate policies, and individual user attitudes, are shifting. “It is no longer controversial to say that the future of cultural production will be open. The fact that open models that rely on the unpaid contributions of users are surpassing proprietary models in terms of usage and even quality, means that industries that rely on proprietary business models are feeling increasingly threatened. As Wikipedia surpasses Encyclopedia Britannica and Linux outperforms Microsoft on servers around the world, these industries are struggling to adapt” (Ford 2009, 34). Heather Ford cites Charles Leadbeater who, in a 2005 TED talk says, “The reason why despite all the efforts to cut it down, to constrain it, to hold it back – why these open models will start emerging with tremendous force, is that they multiply our productive resources. And one of the reasons they do that is that they turn users into producers; consumers into designers.”

Definitions and Models of Openness

Openness and accessibility to information is a spectrum rather than an open/closed dichotomy. As we move forward in the discussion of how open our work as practitioners of ethnography could or should be, it will be useful to clarify the range and definitions of openness and ultimately, the benefits and consequences of conducting research with various degrees of openness.

Open Access (OA) generally refers to literature that is “available online to be read for free by anyone, anytime, anywhere – as long as they have Internet access” (Crawford 2011). According to Peter Suber (2012), Director of the Harvard Open Access Project, “OA removes price barriers (subscriptions, licensing fees, pay-per-view fees) and permission barriers (most copyright and licensing restrictions).” The moral argument for Open Access claims it removes barriers to information, rendering knowledge accessible to rich and poor alike. It lays the “foundation for uniting humanity in a common intellectual conversation and quest for knowledge” (Crawford 2011, 9). Daniel Miller (2012) calls for the development of Open Access for anthropological books and journals and critiques the way we have “ceded control of dissemination to inappropriate commercial concerns.” HAU: Journal of Ethnographic Theory and Journal of Business Anthropology are two relatively new examples of open access social science journals. Many blogs about ethnography would also fall under Open Access, including Ethnography Matters and Savage Minds, for example.

Open Source also promotes free redistribution of work, but in addition, promotes access to an end product’s design and implementation instructions. Most commonly used to describe free access and distribution of computer software, open source projects are usually peer-produced and maintained by collaboration. Open source projects always include the source code or “blueprints” and documentation at no cost to the public². Creators give “customers control over the technologies they use, instead of enabling the vendors to control their customers through restricting access to the code behind the technologies” (Young 1999, x). Neither open access nor open source models translate directly to opening ethnography in corporate environments. However, we will use them as starting points to imagine openness within corporations and possibilities for various forms of open corporate ethnography. The source material of ethnography, for example, could include raw data (audio, video, written transcripts and translations), analyses in the form of domains, themes, or trends, background context and goals of the work, and applications or implementation of the work. It might also include final reports, presentations and opportunities for feedback.

This concept of openly sharing work is not completely foreign in corporate contexts. Ethnographers have been using participatory design or co-creation, and more recently crowdsourcing, to engage stakeholders, including employees, partners, customers, citizens, or end users etc. to participate directly in the creation of new products and services offered by a company to help ensure the product meets needs, is usable, and adds value (Howe 2006; McCracken 2009; Prahalad & Ramaswamy 2004; Schuler & Namioka 1993).

To dissuade thinking in absolutes of “open” and “closed,” we propose a layering model that shows degrees of openness and helps us consider which data to make accessible and to whom in projects. There are at least three sets of layers to consider. One, layers of data; what kinds of information are being shared. The shallow layer is sharing little ethnographic work. This might mean only distributing final reports or presentations. A layer out might include analyses and themes that provide a deeper context for the project. Further still, is sharing more work such as description of methodology, screeners, etc., then the last layer would be sharing all work including certain raw data, such as transcriptions, audio/video recordings, and interview instruments. Two, layers of sharing; who will have access. “Open” does not always have to mean full public access. For example, it could include a white paper on findings that is shared with other departments in your company, or sharing analyses with other practitioners through an online group. Three, layers of time; when is it appropriate to share information, from before a project starts, towards the end, or after its completed. It is through the interplay of these layers of data, participants, and time that corporate ethnographers may find emerging opportunities to open up their research in beneficial ways. This model allows us to consider options of openness on a project by project basis.

FIGURE 1. Layering model that shows examples of what kinds of ethnographic information can be shared, at different stages of a project, and with whom.

Case Study 1: A Corporate Ethnography Experience

Dornadic speaks from her experience doing “consumer-centric research” in-house for the research arm of a large property-casualty insurance company from May 2010 to February 2012. All stages of all research activities were proprietary and executed under confidentiality constraints. Dornadic’s experience is not unique. Ethnographers employed by corporations are often bounded by legal obligation to not disclose research data or resulting analyses and conclusions. “Some limitations on access, topics, and documentation are common in all types of ethnographic work…in corporate work…these kinds of constraints are bureaucratically embedded. They are also driven by competitive concerns about intellectual property, privacy, and image” (Jordan 2009, 102).

The steps to undertake an ethnographic study in Dornadic’s experience might unfold as follows. Initial scope of study and goals are either determined independently, meaning within the research center, or with internal partners from various departments across the organization. For example, an internal client from the Marketing department might approach our team asking for feedback on a new web-based auto insurance quote concept with Baby Boomers as their target customer segment. Recruitment of participants is done internally, usually initiated through our network of contacts. We release minimal information to recruit, sufficient to receive informed consent from research participants. Before starting the study, we have research participants sign a non-disclosure agreement (NDA). Even when we reveal little or nothing about the insurance company’s process or products, when all information is one-directional, meaning information comes solely from participants to the corporation, we still have them sign paperwork first. This sends several messages: 1. The participant’s knowledge and time are so valuable to the company we do not want them telling anyone else what they share. 2. They can trust us to protect their privacy. We do not want external competitors to know about the study, and we will not reveal any of their personal information 3. We can take legal action if they betray the rights of the company; we have a certain legal power and authority over them.

All analysis and report writing is done internally within our team. We immediately share information with internal partners if the work had been initially requested by them. This usually takes the form of a PowerPoint presentation and conference call to review findings and recommendations. Anonymity is granted to research participants, however, it matters little as we, the researchers, are the only ones who ever see the raw data transcripts and rough analysis. We rarely share research findings with participants unless it is to help clarify our findings or to initiate a following stage of research. If subsequent brainstorming or workshops are desired, we translate research findings into design fodder, usually in the form of research patterns (bullet points), quotes, and images, and continue an iterative design process of brainstorming and testing concepts with new research participants. Work for internal clients is often reactionary and attempts to answer specific questions. It is unlikely that we remine data or that the data can continue to grow or be used in future projects. If the research is initiated within our group, we continue the design process independently until the research is “ready” to be released to other groups within the company. We are reluctant to share information that was not yet finished or to share underdeveloped ideas in fear that other groups would misappropriate or reject it. Our release of data across the organization is extremely controlled. Employees need to be granted access into our department’s folders to see our unfinished work, and rarely, in my experience working at the company, does information extend beyond the corporation. Even an initiative from senior management to strive for Open Innovation (see Chesbrough 2011) resulted in careful information management. External partners have to sign NDAs and are given information on a “need-to-know” basis.

Case Study 2: An Open Ethnography Experience

Conand managed a collaborative project between the local television station CreaTV San Jose and the San Jose State University Anthropology department, publicly funded by the John and James K. Knight Foundation between 2009 and 2011. The purpose was to uncover models of civic engagement being explored and utilized in Silicon Valley by individuals whom the team referred to as “SparkPlugs,” or leaders in their respective communities who create change. One of the SparkPlugs, for example, is the head of De-Bug, a media, community organizing and entrepreneurial collective in San Jose. By setting out to create a multimedia resource, freely available in the public domain, about Silicon Valley community engagement, the team also experimented with alternative, digital media-based forms of anthropological data representation. The ultimate goal in this was to reach a wider, more diverse audience who could then use that resource to support action.

From the onset, it was decided that efforts would be made to ensure the project was as open as possible, meaning collaborative in process, and accessible and sharable in outcome. Conand’s team also wanted to share as much of the raw data as was possible and relevant. Participants were not given anonymity, volunteering to participate with the knowledge that their experiences would ultimately be shared in an open access forum. The project team conducted semi-structured interviews with all participating SparkPlugs regarding their use of technology, organizational forms, information exchange, and networks used in their respective engagement work. The analysis of these would help inform a more structured instrument to utilize in later filmed interviews. After conducting the subsequent film interviews, all footage was edited down to an eight minute vignette per SparkPlug, based on analytical categories developed by the team. A public film screening and discussion was held in San Jose to showcase the project, and a website was developed where the project would be housed. The website includes: the finished film vignettes, description of how the project was developed, project team member profiles, bios and video interviews from regional experts to provide context, analytical themes, interview transcriptions, and a brief survey in order to continue collecting public feedback.

The film screening and website were key modes of information distribution. The screening and discussion about Silicon Valley SparkPlugs took place at a local theater shortly after the completion of the project and included participatory activities. In this way, the project team members, the SparkPlugs themselves, as well as the audience experienced the project outcome together, publicly. The event included a brief question and answer session with the seven individuals who were featured in the vignettes acting as the panel, fielding questions from audience members. Finally, it closed with a group activity to engage the audience. The research team had prepared questions on large format posters regarding civic engagement that the audience could respond to by brainstorming ideas on Post-Its. While this seems very low-tech, it created a very tangible and visual process of sharing information in such a large group. Quickly, the front of the room was covered in colorful Post-Its as a very visual sign that ideas had been generated. The concepts and suggestions that came from this activity provided insight for a second phase of the project. Ideally, the project and the event goal was to incite action, but at the very least, for people to leave that event talking about what they had seen, heard, and thought about, connecting with each other. It served the dual purpose of making people feel good (even if they did nothing with the information after that screening) and helped the researchers move into the next phase of their project by using the output of the event as data.

The website, the main channel of communication about the project, was developed in a way that allows for layers of the project to be uncovered, depending on an individual’s level of interest in the project. For example, an interested community member could visit the website, watch one or more vignettes, and then be done with it. Another might find the vignettes interesting enough to want to watch the contextual videos, or read more about the participants. Beyond that, there may be fellow ethnographers who find the project relevant to their own work, and would like to look through the transcriptions or theme analysis for deeper investigation. To summarize, this model of openness allows for various levels of engagement and discovery by multiple interested parties, from the layman to the seasoned researcher, who can use or share the research as they see fit.

While the focus was on openness, it is essential to understand that there were activities undertaken to manage content. For example, it was decided early in the project that the interview transcripts would be shared on the website, and therefore it was necessary to review the transcriptions and edit them in order to ensure they were appropriate for the website as an open source, public document. The process included editing grammar and punctuation (for example, there are no commas in speech, the researchers decided where to add them), editing out repetition, asides, “hems and haws,” unnecessary interviewer dialogue, as well as mentions of embarrassing or harmful information. Names of other people (besides the participants) and organizations, and places or words in other languages were changed, translated, or clarified. Since these individuals trusted Conand’s team and participated in the project knowing there was no anonymity, it was important to the team to ensure they were presented in the best light.

What Corporations Can Gain by “Opening Up”

Corporate ethnographers can enhance their research by conducting open source ethnography. In particular, we saw open sharing of ethnography as an opportunity for ongoing learning, new partnerships and ways of working, and new forms of data representation. The open source nature of Conand’s project increases the potential for continuous learning and multiple feedback loops. As the interview transcripts and video footage “live” online, there is endless possibility for how it can be expanded and appropriated in the future. The SparkPlugs website includes a description of the project, its intentions and goals, the researchers’ backgrounds, and the data collection and analysis process along with the data, so that their methods can be repeated and their data be remined for new purposes. The work maintains rigor since its context and background information are included with the raw data and results. Although the SparkPlugs project has ended, there is no expiration date on the data. Dornadic, often frustrated by the short shelf life of her ethnographic work in corporations, can imagine this as one possible solution to the “one-off project” problem in which a quick turn-around project is undertaken to answer specific questions and then is forgotten shortly thereafter. Open information can be shared internally across the organization and can be reexamined when new questions arise or new projects commence. A site could be created where information is collected, donated, or borrowed. It could be referred to in future projects, workshops, or planning meetings. The same effect can occur if information is shared more broadly, with external partners, for example, or with the public. We need not release every design spec and conclusion to obtain interesting or useful information or feedback.

When researchers share ownership of their data it opens up new ways of working and new partnerships, be it with colleagues, research participants, organizations, or the public. Conand created a network of organizations that became involved in the SparkPlugs project. Individuals from these organizations were brought together for her project and she could rely on them, as they could on each other, for help or information in the future. We can imagine corporate collaborations occurring across organizational departments where they may not have occurred before. We can imagine joint corporate projects, where multiple organizations benefit from sharing data and learning from each other. We could imagine groups that started outside the corporation being brought in as partners to share ideas and information. This has the potential to influence organizational processes. When we are not limited by proprietary and confidentiality boundaries, there is greater potential for learning.

Creating new partnerships, both inside and outside the corporation, is beneficial because it infuses projects with more brain power, garnering more ideas, and possibly a larger variety of ideas. During the SparkPlugs film screening and brainstorm session, Conand was able to get a theater-full of people working on the topics and problems that came from the research. The information collected from this session helped inform the next stage of the project. Rather than the usual process of filing away data and rarely sharing back with participants, let alone the public, discussions and brainstorming with participants led to insights that would not have otherwise been gleaned. More inclusive models of brainstorming allow for feedback loops, more points of evaluation of a project or process along the way, and a more holistic planning process in general.

Including research participants and the public in the process had an empowering effect. The SparkPlug research participants enjoyed sharing their stories publicly, feeling a renewed sense of purpose in their work, pride, and validation that their work is important enough to be documented on film and shared online. If the research were proprietary and confidential they would not have had the opportunities to discuss and share with each other or in public with a large audience. In the corporate realm, Dornadic’s research participants know very little about how their information will be used later. They are less involved overall, have less at stake or claim to the project, and it is more difficult to build rapport and ongoing relationships.

The SparkPlugs project is an example of bringing ethnography into the public realm. Ethnography here takes a very accessible and familiar form – short online videos. As online forums provide the space for data to be accessed, they also allow for different forms of data representation to be utilized. Thinking beyond the familiar forms of essays and white papers can increase the opportunities for a larger audience. “There are clearly genres within anthropology where we have failed to properly convey or disseminate our results because we have ignored the potential to use sound, animation, and other expressive genres that can be placed online to complement written texts” (Miller 2012, 389). The film vignettes in Conand’s work are an ethnographic investigation of community engagement work, but are condensed into bite-sized chunks of visual data that are accessible to a general audience. Because the participants knew from the beginning that all information would be filmed and publicly available, sharing the video online was never contested. If practitioners are serious about the desire to promote public recognition of ethnographic practice and its applications, then the utility of the online forum, and the accessibility of different forms of representation, say video or graphic, should be recognized and more widely utilized. Within a corporate setting, experimentation with data representation can occur based on project type, participant involvement, and audience, be it internal or external to the organization.

Why Corporations Should be Cautious

As optimistic as we are about open ethnography, we understand the limitations and risk associated with it. Recruiting and obtaining consent from research participants, for example, may be more difficult in corporate contexts. Research participants may already be suspicious of the motives of a corporation and hesitant to share information when it remains confidential and private. They may not want to be publicly linked to the corporation, which would make it more challenging to recruit. However, for those participants who are uncomfortable being associated with a particular company or have hesitations to share personal information, there are other strategies to employ. Pseudonym strategies, which are common in anthropology, allow for contributions to be made and shared while retaining privacy. Participant anonymity and an open source agenda are not necessarily mutually exclusive.

Ethnographers must understand that open source ethnography increases risk for the corporation and the participants. Sharing the “source material” such as raw data and project descriptions is arguably the most dangerous proposition. Information may be misappropriated, abused, or linked to the organization’s strategic plan, dulling its competitive edge. Corporate ethnographers already find themselves in an ethical tug-of-war where, “The practitioner must negotiate an intricate balance between the interests of the clients who commissioned the work and those of the community being studied” (Kedia & van Willigen 2005, 16). The ethnographer must be especially sensitive to privacy issues, “realizing that the potential for abusing the researcher-informant relationship is high” (Kedia & van Willigen 2005, 18). Research goals, uses of the information, and inability to control information must be clearly discussed and understood by the participant.

Even expanding the reach of ethnographic research in-house, trying to extend sharing across the organization can have drawbacks. Grant McCracken (2009, 11-12) describes a “death by committee” scenario in which there are too many internal processes that will kill an idea before it fully materializes. He gives the example of Geoffrey Frost who joined Motorola in 1999, who recognized the Razr as a promising product and therefore kept it secret to the company itself. He concealed it from the corporation, reaching out to only a select few, claiming the phone was a “brand builder” and would not make money. He did this to protect the product from the organization’s own internal processes that might have led to the delay or cancellation of the product. Once released in 2004, the phone did well. In 2005, it sold 20 million units and by end of 2006 sold 50 million.

Stepping Out in the Open

Perhaps, at this point, you are wondering if you should try to open up, or make freely available, your own ethnographic work, or to build an open source ethnography project from scratch. A critical (and perhaps less daunting) first step in considering openness in corporate contexts would be finding beneficial opportunities to disseminate ethnographic work more widely across the organization, to internal employees. Despite McCracken’s example, distributing work across organizations has the potential to spark new collaboration and innovations (Kanter 2006; Chesbrough 2011). At each stage of each project, one can consider whether it is beneficial to share work openly and with whom, other professionals, colleagues, research participants, or the public.

Questions to Consider Open Ethnography

We propose a set of questions which can help clarify which elements of a project might benefit by being open so that more projects can be open from the beginning. These questions can be repeated over the length of a project as goals and information-gathering activities change. We argue that increasing accessibility to ethnographic work has the potential to enhance research, project outcomes, partnerships, and make ethnography more relevant in the public sphere. Accessibility and usability of ethnographic data (format, style, description of methodologies etc.) will impact how ethnographic data are used.

Q1: What do we hope to gain by opening up?

• To establish new networks of partners?
• More inclusive and diverse brainstorming process?

Q2: How sensitive is the topic area?

• Are people willing to share on this topic?
• Could it potentially cause harm?
• Are there legal limitations to sharing? (e.g. research in a medical clinic)

Q3: Who should be involved?

• Internally within the organization?
• Externally, across organizations?
• Truly open source, co-created and maintained by anyone?
• When and to what degree in the process?

Q4: What materials should be shared?

• Who might benefit from each piece of information?
• Is there potential for misappropriation?

Q5: When in the process should sharing occur?

• Are there benefits to sharing earlier in the process?
• Does it make more sense to wait and share more developed results?

Q6: Who has access?

• Is it limited to similar organizations and professional colleagues?
• Could it include clients, customers, and the general public?
• Are they passive receivers of information or active participants? What is their role?

Q7: What formats are there for sharing data?

• What is the most appropriate format for the audience/participants in question?
• What will have the most impact? What will be most accessible?

Q8: How do current organizational processes limit or enhance opportunities for sharing?

• How could sharing change organizational processes?