ABSTRACT This article discusses an interpretivist study of adoption of self-service based human resources management system in a company in Canada. The study is conceptually based on a conceptual and analytical framework called Information View of Organization. Both the self-service systems and the framework are new. The main finding is that a number of specific cultural and political aspects that the framework applied has illuminated can help to explain the moderate level of system adoption.
Keywords: System adoption, self service system, ERP system, information view of organization
This article presents a case study of a self-service based human resources management system (HRMS) in a Canadian Utilities Company (code name). The platform for HRMS is Enterprise Resource Planning (ERP) software by SAP™. Self-service applications require end-users to directly and totally manipulate a system with no intermediary. Many people routinely resort to self-service when using ATM machines or purchasing travel arrangements on the Web. In the organizational context, elimination of intermediaries means that data entry, querying, and other system tasks become the duty of casual users. These users can no longer count on help from specially trained power or super users. This increment in work may be perceived differently, depending on the system task and end-user group. For example, it may go unnoticed when the benefit of looking up one’s own banked hours may obscure the fact that work is needed to get to this piece of data. In contrast, an hourly worker may feel that entering his work hours into the system is an odd burden. A manager who now accesses the system for approving the reported time, rather than having a secretary perform a part of the process, may develop a similar percept. Therefore, there exists a possibility that self-service applications may not be greeted enthusiastically by new users who either never wore users’ boots (hourly workers) or those who used to be indirect/casual users (managers). In principle, self-service applications could signal a major change in work processes and organizational cultures, since all employees can become systems users. This raises the bar for the information systems (IS) function that now faces an exponential growth of the user body. The organization that adopts a self-service system is also challenged in terms of changes in work processes and relationships between occupational groups.
The discussion above charts the space of a new research problem whose importance is likely to increase: What are the key issues in adopting self-service applications in the domain of HR management? This problem references two well-established lines of IS research—ERP systems and systems adoption. Each of these has generated insights that are relevant for studying adoption of ERP-based self-service HR applications. Still, the self-service character of ERP applications, due to its novelty, has been just meagerly studied (e.g., Lapointe and Rivard, 2005; Larsen & Myers, 1999; Stein et al., 2005). In addition to studying a new research problem, this study has taken on the additional task of advancing validation of a new conceptual and analytical framework called Information View of Organization (IVO) (Travica, 2003, 2004, 2005, 2006, forthcoming).
For over a century, Canadian Utilities Company has provided the development of utility infrastructure, production, and delivery of utility services. After passing through numerous transformations of ownership and function, today the company is public. It enjoys a monopoly in a regional domestic market, while at the same time competing in the US market.
Canadian Utilities Company numbers about 5,600 employees, both white and blue collar—clerks, professionals, managers, and a whole gamut of trades that perform work inside of facilities or outdoors. The company’s operations are geographically dispersed. Moreover, the company is considered by many as a desirable employer that provides attractive salary and fringe benefits. Hence, employees tend to have longer tenures. Organizational structure draws on a bureaucratic model. The central IS function numbers about 350 employees, 20% of which work in the SAP department. The IS function is responsible for the business domain, while separate IS resources support the engineering staff.
Part of the transformation the company has been going through in recent years is centralization of the HR function. One goal is to rationalize administration, including reduction of headcount. HR specialists have been moved away from divisions into a central function, and clerical and professional staff are paired and tasked to support divisions. HR professionals provide consultation on HR matters for division management (managers and supervisors) on various issues, including salary, employment, filling positions, employee issues, labor relations issues, performance management, and training. The professionals also monitor HR processes that run across divisions. On their part, HR clerks help the professionals as well as divisions, inclusive of carrying out a help desk for HRMS.
The system studied is a mySAP™ self-service application for HR management. It contains two modules. One is for all employees to enter and query their time, expenses, travel, and professional development (this one will be labeled with SSE after the term “self-service of employees”). Another module is intended for managers, to make online approvals of employees’ reports, claims, and requests. We will code-name it with SSM (self-service of managers). Managers can also print canned reports and view professional profiles. Of course, manages are also users of SSE.
SSE and SSM are functionally connected since the output of the former provides the input for the latter. Technically, SSE is just a new user interface that is layered atop of a core R/3 SAP system. This user interface runs on the SAP portal. The older application with nearly the same functionality uses Web forms to access the R/3 system. It was introduced to clerks a few years ago, and has been adopted and used efficiently since then.
One fact that sets the context of the study is that HRMS was implemented without implementing a process change beforehand. Unlike a typical implementation of ERP applications, the organization studied did not conduct a formal business process reengineering (BPR) before HRMS went live. Another important characteristic is the strategic management goal of reducing clerical work gradually with help of HRMS. Currently, clerks make up about 13% of all employees. Two in three clerks do time administration for field workers, which can take up to 30% of clerks’ time. The goal of reducing administrative costs applies to SSM as well. Without self-service applications, a clerk would print out reports and other data summaries out of the R/3 system, run a preliminary check on these, and submit them to a senior for approval. SSM is expected to change this process.
With SSM, supervisors and managers are supposed to lookup data online on their own, to create reports, and to make approvals by entering checkmarks in check boxes on screens. Clerks ought to be prevented from doing what they used to do for superiors, as access to HRMS is protected by the portal password. If a manager shares the password with a clerk, the manager runs the risk of compromising the security of sensitive data he is authorized to access, including his own personal data stored in SSE. In the division of labor established with the introduction of HRMS, 700 managers and 150 clerks are placed in the user’s seat of SSM. A strategic decision was made that the clerks continue administering time and expense data on behalf of field workers (instead of these doing this through SSE), by using the older Web forms-based user interface to R/3.
Adoption of self-service ERP: Terra incognita
The ERP literature has provided valuable insights into implementation situations, conditions and effects. For example, Sarker and Lee (2000) found that adoption of an ERP system was associated with changes in the organization’s old “dysfunctional culture” characterized by a “sea of paperwork,” lack of communication between divisions, and scheduling errors. Krumbholz and Maiden (2001) found that implementation of SAP R/3 in a pharmaceutical firm did not fulfill the promises of achieving integrated data because the system was not configured uniformly across organizational units. Bulkeley (1996) discovered that a combination of unprecedented external difficulties and implementation mistakes (poor testing, insufficient user training, and a rushed rollout) was responsible for the disastrous effects of implementing in a big-bang strategy SAP modules at Fox-Meyer Drug. Markus & Tanis (2000) developed a model of the entire experience firms pass through with an ERP system. The authors have proposed a process model with four phases—from planning to routine use. In between comes a “shakedown phase” that corresponds to system adoption. In this phase, a system goes live and users strive to achieve routine use or may end up with a system shutdown. Problems may occur during a shakedown, such as the maintenance of old processes, poor software ease of use, stalled user skills, and low usage of system. A well-configured and integrated system, redesigned business processes, trained users, resources to address problems, and action to fix problems are the main methods of overcoming the problems and securing successful adoption.
Another segment of the literature on enterprise systems refers to the study of adoption of enterprise system-based self-service applications. Due to the novelty of the phenomenon, this literature is in an early stage. Stein and colleagues (2005) identified a success story of a self-service based application for pay, leave, and benefit packages in an Australian telecommunications firm. The authors reported a 40% reduction in administrative staffing, 80% reduction in management HR duties, and 50% reduction of transaction costs. In contrast, Larsen & Myers (1999) discovered serious difficulties in the adoption of a SAP accounting module in a financial firm in New Zealand, where the accounts payable process was based on self-service. The difficulties involved skipping of BRP, lack of reports, and dissatisfied users. Although the number of accountants was reduced for 68%, knowledge was lost with the fired employees, which resulted in low skill and morale. Finally, Lapointe and Rivard (2005) have found that enterprise systems may fail if the political position of a key user group is threatened. Doctors who were supposed to enter prescriptions into new electronic medical record systems torpedoed the systems in three of four hospitals in Quebec. The doctors managed to sustain their preferred manual method of dictating prescriptions to nurses, as they criticized the systems for imposing extra, non-professional work on them.
The problem of the present study also links to the topic of IS adoption. This is a well-established area of research that draws on behavioral theory and on models of innovation dissemination. The behavioral research focuses on predicting results of deliberations an individual user makes with regard to system adoption. For example, the well-known Technology Acceptance Model (Davis et al., 1989) posits that users’ perception of system usefulness and ease of use are antecedents to the user’s intention to actually use the system. In other words, the user is placed in a role of an evaluating agency that determines the adoption fate of systems. This essentially cognitive assumption is valuable for the present study. On the other hand, the behavioral theory of IS adoption has to keep developing with new types of ISes. Self-service based ERP systems in particular may present new adoption challenges that need to be reflected in new dimensions, which, in turn, could contribute to modeling of IS adoption.
The present study is conceptually based on Information View of Organization (IVO) (Travica, 2003, 2004, 2005, 2006, forthcoming). This is a new conceptual and analytical framework that focuses attention to aspects that sit at the intersection of IS research, organization theory, and cognitive theory (more in the next section). The broad-based and flexible character of IVO made it a suitable theoretical foundation for investigating the variant of ERP systems. Advancing the testing of the IVO framework was the second research goal of this study.
Information View of Organization
This study is conceptually based on IVO that was introduced by Travica (2003). This is a conceptual and analytical framework that draws on the IS, organization, and cognitive theory. For more details on IVO and its research applications, the interested reader should consult the above-cited works. A brief description suitable to the present study will suffice here. As depicted in Figure 1, IVO places information artifacts and IT at the nexus of the organization. The concept of information used is broad (data, meaning, knowledge, and wisdom), with cognitive underpinnings, and is different from the usual concept implied in the data-information dichotomy. IT is also conceived more broadly, as any means of creating and manipulating information. Information and IT are tied in a yin-yang relationship. Information gives purpose to organizational IT, while being molded by it; IT makes a difference in information (medium is a message), while being in service of information. These premises are motivated by the goal of IVO to bridge gaps between research traditions that chart the space of the IS field.
Figure 1. The Information View of Organization Framework
Figure 1 further shows that the first orbit around information and IT consists of well-established themes in IS research, including IS adoption (acceptance). The next orbit consists of aspects that represent an intersection of information, IT and IS aspects with well-established themes in organization theory. IVO gives an IS theory slant to “views of organization,” while IS theory gets firmer anchoring in organization theory. The intersection aspects include infoculture and infopolitics. Infoculture refers to beliefs, behaviors and artifacts that are related to information and IT. Infopolitics refers to power, agendas, and fight/flight behaviors related to information and IT. These concepts have a background in both IS theory and organization theory (cf. Travica, 2005). The adjacent orbit exhibits organizational aspects that correspond to IVO aspects (e.g., organizational culture, corresponds to infoculture). The outermost orbit in the IVO framework refers to organizational goals and issues (things demanding attention, challenges, and problems). These provide the raison d' être of organizational ISes, information, and IT.
It is assumed that the aspects in the IVO framework can participate in different relationships. Apart from the obvious relationship between infopolitics and organizational politics (the former is a segment of the latter), infopolitics could be related to system adoption. For example, established authority over access to data may be challenged by a new IS. This scenario may motivate a group on the losing side to fight the new system (cf. Markus, 1983). Before engaging in infopolitical fight, the group members evaluate the IS based on their infopolitical and political agendas. Thus, the IVO aspect of IS Evaluation comes into play (Figure 1). Furthermore, the evaluators can target IT built into the system as well as data coming from the system and the meaning these data conjure up in the mind. These two may not necessarily be congruent. Furthermore, IS adoption typically implies that users adjust to new work processes, an aspect of which is infoprocess (information flows that accompany task flows), and so on. Since IVO is a framework and a flexible one, these relationships are theoretical possibilities to be studied empirically rather than preconceptions to be imposed on research design.
This study is based on interpretivist epistemology. Basic postulates of case study methodlogy were followed (Lee, 1989; Yin, 2003), and some elements of action research were implemented (Baskerville & Myers, 2004). The case captures a period of adoption process of HRMS from October 2005 to April 2006. Elements of action research have materialized through communication between the author and few key actors in the company in the course of the study. Preliminary study results that were communicated to the key informant during the study process could have influenced some actors and events in system maintenance.
Data were colleted in a triangulated manner. Intensive interviewing was the main method, combined with the author’s hands-on experience of the system, observation of systems use during some interviews, and document analysis (the company’s surveys of HRMS users’ satisfaction, HRMS plans and management reports, and broader corporate documentation). Interviewed were 33 persons in total, including 4 focus groups, which yielded 22 hours of interviews. Interviews were tape recorded with consent of the respondents, and notes were taken during each interview. With the exception of one focus group held via videoconferencing, interviews were face-to-face. Data collection and sampling started from an internal informant, and unfolded through snowballing and intentional sampling. The resulting sample exhibits variability in terms of professional background, hierarchy, and geographical location.
Design of interviews varied from open-ended to semi-structured, with the exception of the structured interview conducted with the CEO. The level of structuring for the interview depended on the stage of investigation, professional background, and hierarchical position of a respondent. Interviews were more open ended at the start of the project since the focus was on learning about HRMS, organizational context, and issues associated with the adoption of HRMS. More open interviews were also conducted in initial steps of investigating a new issue. An interview would start by asking a respondent to describe his/her main tasks and the role of HRMS in them. Then, probes would be used in order to learn about details of information and IT aspects involved in the tasks, and to elicit comparisons between the present and past. The respondent would also be prompted to speak of his/her feelings regarding the ups and downs of the new way of working and the system. If feasible, the respondent would be encouraged to use of HRMS in a typical or a troubling task. From the IVO perspective, these methods served as an inquiry into infoprocess, IT, information, and infoculture aspects. As interview notes and transcripts were coded iteratively during data collection, the researcher’s learning about “issues” was advancing with time. Once a coding category acquired initial contours, interviewing would become more focused. The goal from that point on was to drill down into details of the emerging concepts that reflected the adoption situation.
The level of interview structuring also depended on the professional background and hierarchical position of respondents. For example, clerks would be questioned more on usability issues because, as super users of both the old and the new system, they could make detailed evaluations of user interface and comparisons with the older Web-based HR application. In addition to providing rich IT-related data, clerks also proved to be sharp evaluators of process/infoprocess efficiency. In contrast, interviews with managers would probe on IT aspects and, then, move onto aspects of infoculture and infoeconomics if the respondents would demonstrate less familiarity with technological details (which had happened more often than not). So for example, the managers would be invited to assess their workload now and before, to comment on the idea of making the use of HRMS a standard part of managers’ jobs, and to estimate a value of the new online approval processes. IS management was additionally asked about their technology-related views. As for respondents at the executive level, they proved to be most informative on issues fitting the rubric of infopolitics. Probing them on infoprocess and infoculture from a stance of change management turned to be, contrary to expectations, a less resourceful path. Finally, professionals provided accounts corresponding to their particular areas of expertise—HR and IT. Their stories helped learn about infoprocesses, infoculture, infopolitics, and infoeconomics.
Coding was conducted during data collection, and included coding interview notes and transcripts, constant comparison of the codes, writing memos on code definitions, diagramming relationships between codes, and taking a note of emerging research questions. For the most part of the research process, coding was kept at a basic level by using common sense vocabulary and resisting to aggregate categories into broader ones. Aggregation started when all the data were collected. The final step in coding was to match higher-level categories to IVO concepts (see the example below). Many of these data analysis methods are used in different venues of qualitative research, such as ethnography (Orlikowski, 1992; Suchman, 1987; Tedlock, 2000) and grounded theory (Charmaz, 2006; Glasser & Strauss, 1967). The intertwined activities of data analysis and data collection paved the way for “data to speak” (Glasser & Strauss, ibid.). The data collected created research questions and ideas that moved the investigation to the next step. One should not assume, however, that a perfect linearity reigned. Indeed, the entire research process had many hiatuses, loops, and parallel lines of investigation. This is an example of the data collection/coding process.
A seasoned expert that had worked for years as a trainer for and tester of ERP applications uttered a remarkable statement: “We don’t finish off something before we get to another technology change.” Indeed, this person used the underlying idea as a leit motive making seven occurrences in his interview. Here is another example:
“We have added the idea that we are going to a paperless system, but we are printing off hundreds of employee’s notifications of what they are paid. Plus we are printing for the supervisors to tell them what we paid the employees. So, I don’t think we finished that job off very well at all.”
This lead prompted to start exploring how other interviewees felt about well roundness of applications and about keeping up with ongoing changes. It turned out that the opinion of the ERP tester/instruction had an apparent following. As one respondent put it: “I just get comfortable with one thing, and then all of a sudden we are into something different. Now it is this [SSE].” In a later round of coding, these instances were grouped into the category labeled “Systems Development Pace.” Upon constructing other categories in the same way, comparisons-driven analysis sorted the systems pace category into a larger conceptual “box” of cultural conditions for system adoption. In order to facilitate acceptance of HRMS, it is important to know about users’ assumptions and values regarding a desirable pace of developing systems. If a number of them feel as if the systems development has been rushed, important details neglected and never refined, and that the frequency of rollouts surpasses their coping capabilities, then it is plausible that their adoption behavior will be affected accordingly. Since these assumptions and values concern IT and associated information, they belong to the rubric of infoculture in the IVO framework. This assignment was the very last step in coding. It is common to interpretivist methodologies, while deviating from the grounded theory approach that inspired the preceding coding process.