اصول ناب، یادگیری، و کار دانش: شواهدی از یک ارائه دهنده خدمات نرم افزار

In this paper, we examine the applicability of lean production to knowledge work by investigating the implementation of a lean production system at an Indian software services firm. We first discuss specific aspects of knowledge work—task uncertainty, process invisibility, and architectural ambiguity—that call into question the relevance of lean production in this setting. Then, combining a detailed case study and empirical analysis, we find that lean software projects perform better than non-lean software projects at the company for most performance outcomes. We document the influence of the lean initiative on internal processes and examine how the techniques affect learning by improving both problem identification and problem resolution. Finally, we extend the lean production framework by highlighting the need to (1) identify problems early in the process and (2) keep problems and solutions together in time, space, and person.

Lean principles, exemplified by the Toyota Production System (TPS), continue to greatly interest the operations community. Many credit Toyota's sustained success to their persistent and pervasive application of these ideas to manufacturing and management systems (Hino, 2006 and Liker, 2004). This thinking has motivated many manufacturing companies to imitate, either wholesale or in part, lean principles in their improvement programs. While lean production has led to improved performance in many cases (Li et al., 2005 and Shah and Ward, 2007), failed implementations are common, and as Shah and Ward (2007, p. 785) note, there is significant “confusion and inconsistency” in how lean production works and how it is best implemented. In recent years, organizations have sought to apply lean production to knowledge work (e.g., Poppendieck and Poppendieck., 2003 and Schutta, 2005). While almost all work consists of manipulating both physical goods and information, work referred to as “knowledge work” primarily involves the use of information (Drucker, 1999). The utility and impact of lean production in non-manufacturing contexts remain points of contention, leaving many managers to wonder if they are merely applying inappropriate and faddish ideas while others argue that lean principles have universal applicability (c.f. Sousa and Voss, 2001). In this paper, we ask two related questions: (1) Do principles of lean production apply to knowledge work? (2) How can we extend the existing framework of lean production to a new context that differs substantially from that in which lean was developed? To answer these questions, we report our observations and analysis of the application of lean production at Wipro Technologies, a large Indian firm competing in the global software services industry (i.e., custom software development). In the following section we examine lean production in the context of knowledge work and identify this paper's contribution to the literature, before discussing the principles of lean production as identified in manufacturing in Section 3. Section 4 details our case study research design while Section 5 uses quantitative data from Wipro to examine the performance of lean software development projects as compared to non-lean software development projects. Projects at Wipro are the primary way that work is delivered to customers. A non-lean project is executed in a traditional manner, while a lean project is delivered using lean principles. Section 6 then qualitatively examines how Wipro's lean initiative changed the way that the firm operated. In Section 7 we discuss extensions to the lean production model while Section 8 offers concluding remarks.

In this paper, we see that implementation of a lean production system in knowledge work is possible and that it changes how the organization learns through hypothesis-driven problem solving, streamlined communications, simplified process architectures and, to a lesser degree, specified tasks. In its attempt to implement a lean production system, we see that core processes were altered, resulting in improved operational performance. Like any study, ours has limitations, so one should be cautious in applying its results. In particular, our study examines implementation of lean production in knowledge work, investigating the experience of one company. It is possible that our observations will not generalize to other settings. Also, while the interim results at Wipro are promising, the implementation has far to go to deliver fully on its promise. This limitation is a necessary but unattractive consequence of the detail and lack of recall bias that the real-time nature of our study permits. Finally, question arises as to whether Wipro is actually doing “lean” production. Our study does not rely on the epistemological concern of whether Wipro's approach is truly lean, however. Since no definition of “lean” in software is accepted, we rely on the fact that Wipro was consciously trying to create a lean system for software services. Their ideas were inspired by lean thinking, in any event, so we are able to learn from their attempted mapping. This paper makes several contributions to the existing knowledge base. First, we identify a significant challenge to using ideas from lean production in a knowledge-based industry: lack of repetition. Second, our empirical examination suggests that manufacturing-based principles are applicable to knowledge work. Third, we use descriptive analysis to examine how the challenges identified above, were overcome. Our qualitative work illustrates how lean production improves both problem identification and problem resolution within knowledge work. Finally, we extend the framework of lean production, offering two propositions for problem solving. We hope that, together with other recent work (e.g., Shah et al., 2008), our observations of details during implementation provide the beginnings of a roadmap for other knowledge-based industries seeking to apply the same ideas (c.f. Boyer et al., 2005). Such details are the most important (and most often underemphasized) part of any lean initiative, far outstripping the import of a strategic mandate that ‘we are doing lean’. While implementation of its lean production system is far from complete, Wipro's new approach may offer a way to manage uncertain and complex projects through a high-assurance, iterative model. If the greatest value from lean production comes when complementary practices are implemented as a system (Hino, 2006 and MacDuffie, 1995), then this process may take some time as the company finds the right contextually dependent set of practices to eventually constitute the “Wipro Production System.” While proceeding by analogy to physical processes has led Wipro's lean initiative to this point, they will need to use knowledge gained from their context to proceed further. As the company's earlier public, external approach to process management left Wipro open to imitation, its switch to lean techniques may prove beneficial if more difficult for competitors to emulate. This may, in turn, create opportunity for an ongoing operations-based competitive advantage.