Toward Industrialization of Ethnography

Level1 (Initial): Traditional ethnographic research

Traditional ethnographies are assessed as being level 1, which is “Initial”. Sato has commented, “As it has been pointed out for a long time, it is almost impossible to create a ‘cookbook’ type of manual for fieldwork, which is because nothing is as far from the intrinsic nature of on-site investigation than strict application of some formulae or rules” (Sato 2002). The essence of ethnography is in the depiction of culture and there may be a great variety of target sites. For that reason, traditional ethnography combines different investigation methods best suited to the situations at the sites. For example, Lancaster University, which has made use of ethnography to design systems for air traffic, says that there is no well-established investigation method:

“There is no one method of ethnographic analysis. The approach we have adopted involves various practices, alerting concepts, and semi-technical terms, which we can indicate but have no space to defend in detail. The fieldworkers immersed themselves in the work by spending several months observing activities on and around the sites, talking to staff, and talking with them so that they could develop an understanding of what the controllers do.” (Hughes, 1992)

Level 2 (Repeatable): Repeatable ethnography

Spradley et al. (1979) has structured the process of ethnographic interviews and defined elements that are important in interviewing. Beyer et al. (Beyer 1997) has defined a method of gaining an understanding of the user’s context for system design and has defined procedures for understanding the context such as the artifact model and sequence model.

These techniques are intended for navigating the process of ethnography and can be seen as indicators of the procedure of ethnography that has been conducted at a personal level.

Case examples have been reported in which ethnography was used to determine specifications in the fields of system development and product development. For example, Bentley et al. (Bentley 1992) have reported on case examples where ethnography has been used in the design of air traffic control systems, in which they say that it is difficult to put successful cases to wider applications:

“We have had the advantage that the people involved from both sociology, and software engineering were willing to be flexible and willing to recognize that the other discipline’s viewpoint may be valid, irrespective of how alien it might be. In this respect, we believe our collaboration has been successful but there have been difficulties which we believe are likely to arise in other collaborative projects of this nature.”

This shows that success can only be achieved under certain conditions, which leads to the conclusion that the standard process of ethnography has not been defined.

Level 3 (Defined): A case study

In this section, activities in Fujitsu are described, and it is assessed whether Fujitsu satisfies level 3 or not.

Fujitsu is a Japanese based system integration company and that has made use of fieldwork not as a tool for research but as a tool for qualitatively analyzing customers’ work practices and problems that supplement quantitative customer analysis in the conventional marketing and product design. Ikeya et al. (Ikeya 2007), an anthropologist who conducted fieldwork training for Fujitsu, has commented: “As they accumulate their experiences, they customize fieldwork into a more systematic method in order to cope with constraints they have on the overall schedule. They have developed a quite effective way of doing team ethnography.”

In the activity called “engineering of ethnography,” Kishimoto et al. (Kishimoto 2009) standardized operation procedures and outputs to implement them at more than 50 customer sites and gave training to more than 400 employees for the purpose of conducting fieldwork in a few weeks, which is generally executed over a few months to a few years in the academic world, to business.

Kishimoto and his teams collected a series of field notes and findings made by a trained team and they counted the volume of the notes and the number of findings day by day. Based on this empirical study, the number of days for shadowing has been established. The team also developed a technique and guidelines for standardizing the quality of activities of ethnography in addition to the quantitative aspect.

It is defined that several viewpoints such as time sequence, space and human relations are necessary to obtain a large amount of qualitative information in a short time – as shown in Figure 2.

FIGURE 2. A meta-model of work context

To assist with exhaustively covering these viewpoints for interviewing, several types of worksheets are provided (Ishigaki 2009). Quantitative issues have been standardized by designing in advance which worksheet to use at a given time or situation, which also allows for the definition of the types of worksheets to use and the extent of information obtained by interviewing, like shown in Figure 3.

FIGURE 3 – Worksheet examples

In the study’s events, a trained team was charged with the negotiations for a system renewal and the front-end of requirements definition for a branch office at a financial institution. They conducted interviews and a one-day observation at a financial institution to analyze the workers’ invisible human awareness of issues and their intrinsic motivation. The findings were shared in a session with executives, general managers of each branch office, and with directors and chiefs of the department of clerical work. The results were praised higher than the team’s original expectations and they succeeded in the system development negotiation without lapsing into price competition.

It has been pointed out that fieldwork by engineers without fieldwork experience tends to result in analysis from an etic² point of view such as pointed out by Obata et al.(Obata 2007): “They evaluated work practices based on their own observations and perspectives rather than describing problems from the workers’ point of view. We reminded them repeatedly about the guidelines and procedures for an ethnographic approach”. So, guidelines to help lead to emic³ results, as depicted in Figure 4, were developed

FIGURE 4 – A process in the guideline (Obata 2007)

Aim is proposed by Yagi et al. (2009) to help researchers gain an understanding right down to the inner deep level of customers. Aim is composed of a phase of asking customers about past positive experiences and a phase of discussing, based on the result of the previous phase, the source of energy for action and values of the customers. “The most distinct features of this model are the ability to extract “future vision and direction and past positive experiences,” The focus is on positive feelings—values, features, source of motivation that envelop the everyday duties of workplace users and the products and services that they use.(Yagi 2009)

FIGURE 5 – Assessment of the case

Figure 5 is a summary of an assessment of the case. As the figure shows, Fujitsu’s standard processes are in place and they are trying to establish consistency of process performance across the organization. Through this assessment case study, Fujitsu can be assessed as level 2 which is “Repeatable” and they are on the way to level 3 which is “ Defined.” To be assessed as level 3, Fujitsu needs to establish consistency of process performance across the organization and needs some degree of improvement over time. Measurement is required as level 4. Although Fujitsu’s method contains steps or concepts for improving the quality, apparently it does not extend to the assessment of quality.

Toward Level 4: Measured

A number of researchers have pointed out qualitative issues with fieldwork executed by engineers. Obata et al. (Obata 2007) mention, in addition to the tendency toward being etic, that the description of episodes tends to be insufficient and the observation on the relation between an episode and recognition often contains a jump: “However, the descriptions of episodes were sometimes insufficient. Members of the inspection team who were outside the project teams had difficulty understanding what was going on in the projects. Professional ethnographers sometimes pointed out that the relationship of the episodes to the discussions was unclear.”

Jordan et al. (2008) point out that fieldwork conducted by engineers lacks a theoretical framework and does not go beyond stopgap solutions: “They were missing the kind of theoretical framework that would allow them to go beyond proposing “band-aid” solutions and instead consider the implications of their local observations for the company as a whole.”

This indicates that there is more than one viewpoint in assessing quality, to say the least. One is validity, or whether the validity of the analysis is sufficient as a baseline. Assessment of whether or not a proposition objectively deduced from facts is clearly distinguished from the assertion of an observer is required. The second is ethnocentric findings. The core of ethnography is to obtain the viewpoint of residents, and how much has been found from an emic point of view can be a measure of assessment. The third is business effectiveness. There is apparently a discussion of whether or not the results of analysis are able to support the deduction of essential improvement plans in business.

Rejon et al. (2009) propose a framework of metrics and index for ethnography, which are believed to provide measures of business effectiveness. “Although there is no proven formula for ethnographic value recognition, based on our projects, we have defined a method called SEED (Savings, Efficiency, Envisioning and Differentiation) that identifies several key factors for tracing how ethnography affects business and what contributions to business it makes.”

Toward Industrialized Ethnography

Fujitsu’s case is considered a successful endeavor where a large company successfully brought ethnography into the business context. They are successfully continuing to manage ethnography research teams like software development projects. These teams are required to make their output measurable by the request from their internal customers; however they are yet to have good indicators.

If organizations using ethno-methodology are assessed as higher than level 2 with CMMI maturity level, their customers will be able to expect consistent process with consistent results. The customers will be able to see how much the research has proceeded, what the current middle output is, how good the output is when comparing with their initial expectations to judge to if the project should continue or not, for example. The detailed process will be modified to adjust for the field where researchers study and there will be no same result. In software industries the situation is the same and there are no identical projects and no identical software. However CMMI is working effectively in software industries because it fits requirements of customers.

Software development is industrialized with CMMI and other lots of efforts. In the software development industry, only some special software such as encrypt systems or the core part of operating systems is developed by highly educated programmers. Most of the software is developed by programmers who have received a few months of training. In the architecture world, some customers ask for famous professionals to design their house, but most of the people are satisfied with houses designed by anonymous designers.