• No results found

5 
 Case Study: Mitos – Project Desert 49


5.4
 Analysis of Challenges and Solutions 61


5.4.5
 Knowledge Distribution 76


The challenges regarding the uneven knowledge distribution can be summarized to:

• IMO does mainly work related to coding and unit testing, giving them less domain and business knowledge. The Indian resources also have poor overall picture due to lack of participation in early phases.

• Few SMEs in IMO cause IMO to depend on the SMEs in Norway. • IMO use more time than NMO to acquire functional knowledge. IMO is the lowest link in value chain

In ProjO and ProjS, IMO does mainly coding, unit testing and system testing. Several of the interviewees from these projects pinpoint their low level of functional and domain knowledge. The processes surrounding the offshoring model in these two projects, where NMO is the middleman in most of the communication between customer and IMO, cause IMO to have little access gaining more domain knowledge. An example of such a process is bug fixing: The defect is solved by onsite, offshore only code and tests it for NMO, giving them less possibility to gain more business and domain knowledge. Team lead in ProjO at IMO clearly states this point:

I don’t say we would design better solutions, but maybe sometimes we could have made better solutions, if we were involved in the problem resolving activity. By only coding a solution made by NMO, we won’t

Global Software Development

get more business understandings or better functional understanding of the solution (Ano6, Appendix B.5).

It should be noted that ProjD does not deal with this issue, as they have direct contact with the client in Norway.

In addition to express difficulties acquiring business knowledge, the quotation above also states that by not including IMO in the solution design phase, Mitos may slip on a better solution. This latter point is important to understand the next challenge, namely that of few SMEs in IMO.

Knowledge is created through social interaction between tacit and explicit knowledge (Nonaka & Takeuchi, 1998). A case study at Alcatel reports that

achievements are substantial if engineers of entirely different educations and cultures try to solve problems (Ebert & De Neve, 2001, p.68). By not

including IMO in these early phases, some innovative solutions may never be brought to life, and knowledge would not propagate on the ontological dimension of the knowledge creation spiral presented in chapter 3.3.3. Organizational knowledge creation is thus limited to some extent by not including IMO in the more knowledge intensive work assignments.

Few SMEs in IMO cause dependence to NMO

It was not possible to get numbers on the exact amount of SMEs at IMO and NMO, but according to several interviewees, there is lack of SMEs at IMO. The knowledge transfer plan (which is not presented in this report due to confidentiality) reveals 15 of 19 SMEs in ProjO are Norwegians. I found 19 SMEs on the project’s SME list, of which 15 had Norwegian names, whether they work in Norway, India or any other place is unknown. As mentioned above, when IMO does not take part in the solution design or other work higher in the value chain (requirements analysis, architecture etc.), which is work dealing with business drivers and domain knowledge, then it is not easy to develop more knowledge on these areas. Thus, making IMO having few domain or business experts.

When IMO gets a change request, we think what shall be changed; we do not think why shall this be changed (Ano6, Appendix B.5). This quotation briefly

explains the differences in the way IMO and NMO thinks, and the type of work that is assigned to each of these sites. By often thinking what shall be done, IMO develops expertise in technologies and programming languages because they will be doing routine work. At least, the work assigned to IMO will be felt as routine work if they do work that is similar for a long time. Often developers tend to prefer learning new features and new technologies,

Global Software Development

they do not want to be experts in one thing (Desouza, Awazu, & Baloh, 2006), and the interviews conducted with the developers in India supports this. Cross-site modularization of work was identified as one of the benefits of GSD, because then there will be less need for cross-site communication. Chapter 2.1 argues that diminishing the need for cross-site communication is vital, but at the same time one should be aware that too much independence causes integration difficulties of software, and less team-feeling across sites. The dependence on NMO SMEs, cause much cross-site communication, and although it may build relationships across sites, I doubt the extent of team- feeling building by such questions to SMEs. If the communication was between peers co-operating, rather than a novice asking a teacher, it may be a better chance of building a good cross-site group feeling. Thus, this dependency seems not having any specific positive effects.

Longer learning time in IMO

Ano7 (Appendix B.6) clearly states the difference in amount of time that takes to acquire functional knowledge in the different locations by saying 1,5 years

in NMO is comparable to 2-3 years in IMO. Although the numbers may be

different from the truth, it is at least clear that there is a difference in the learning time, not in IMO’s favour. The extra amount of time required in IMO can be explained by the smaller amount of exposure to functional info at IMO. NMO’s direct and frequent contact with the customer, and their capability to speak Norwegian, make it possible for them to acquire more functional knowledge in a shorter time.

Also, when the knowledge of the resources at IMO is very similar to each other, it is less likely that they learn new features from their local colleagues. Ano6 (Appendix B.5) identified this by saying everybody knows what

everybody knows here at IMO. Knowledge sharing happens best in local

groups with close working relationships because it is easier to accomplish tacit to tacit knowledge transfer, socialization, in such settings. Also, knowledge artefacts created by people you know, and who works in your team, are often the type of information preferred by individuals (Desouza, Awazu, & Baloh, 2006). Thus, two possible explanations for the longer learning time, in addition to the obvious lack of nearness and language capabilities, can be that the major part of the experts in the domain are abroad, and that individuals tend to prefer learning from their own team members and others in close working relations – which is difficult when everybody knows what everybody

knows.

Once again, the degree of shared tacit knowledge plays an important role. Physical and sociocultural distance cause the tacit knowledge shared by the

Global Software Development

Norwegian stakeholders NMO and customer on one hand, and IMO on the other hand, to be low initially. Over time domain knowledge is transferred through much externalization from Norway, and interpreted at India through internalization. This builds the tacit foundation needed at IMO for a more efficient interpretation process of the information from Norway. Thus, recognizing the need to build up the tacit foundation is crucial to decrease the learning time at IMO.

In addition to visiting each other, Mitos has developed New Joiners Kit, which is a set of documents with overview information, videos from presentations held by SMEs, screenshots and voice recordings from meetings etc. This kit is used both at NMO and IMO when the project teams get new staff members. In this way they can ensure that the new joiners get a quick introduction to the project at the same time as getting deeper knowledge through the SME- presentations, without too much effort from the existing team members. Often the size of the teams changes and increases over time, and by using the New Joiners Kit they are prepared to handle these changes in a structured way. Creating the New Joiners Kit ensures that there is a focus on knowledge transfer from early in the life of the project. The practice of New Joiners Kit has also been transferred to other projects because it was successful. Such a solution, with use of video, voice and textual presentation gives the new joiners different ways for internalization. Although, without much shared tacit knowledge it may be difficult interpreting and get a holistic view of the project, by offering multiple channels of externalized information, the internalization process may be easier.