Table of contents

1 Introduction *

2 Theoretical Background *

3. Pharmacia and Upjohn *

4 Research Design *

5 The UniCAP 100 Project *

6 The UniCAP 1000 Project *

7 Discussion *

References *

Table 1. Communication patterns of different R&D projects *

Table 2. Perceived uncertainties of organisation A of the UniCAP 1000 project. *

Table 3. Input dependency and information gathering in organisation A of the UniCAP 1000 project. *

Table 4. Stated frequency of communication between the investigated sub-projects in organisation A of the UniCAP 1000 project. *

Table 5. Perceived uncertainties in organisation B of the UniCAP 1000 project. *

Table 6. Input dependency and information gathering in organisation B of the UniCAP 1000 project. *

Table 7. Stated frequency of communication between the investigated sub-projects in organisation B of the UniCAP 1000 project. *

Figure 1. An Information Processing Model for Managing Communication *

Figure 2. Three possible patterns of sequencing two design tasks *

Figure 3. Types of overlapping based on evolution and sensitivity *

Figure 4. The UniCAP 100 instrument *

Figure 5. The UniCAP 1000 instrument *

Figure 6. The organisation A of project activities for Unicap 1000 *

Figure 7. Sub-project interdependencies in organisation A of the UniCAP 1000 project *

Figure 8. The new organisation (organisation B) of project activities *

Figure 9. Sub-project interdependencies in organisation B of the UniCAP 1000 project *

In a world of increasing competition, the speed of change in technology and innovation has increased rapidly. Competition based on the generic strategies for stable markets, by for instance Michael Porter (1991, pp. 95-117), are no longer sufficient for sustainable profitability. The customer needs, and how they are fulfilled, constantly change because of innovative firms and new inventive technologies (see for instance D'Aveni, 1995). Profitability and success is therefore, to an increasing extent, coupled with a firm’s ability to fundamentally reinvent the basis of competition within existing industries, and to invent entirely new industries through the development of new products.

New Product Development (NPD) and the management of NPD-processes are therefore among the most crucial functions of a modern corporation (Ali, 1994, pp. 46-61; Cooper, 1994, pp. 60-76; Brown & Eisenhardt, 1995, pp. 343-378). Effective management of NPD can make a real difference. For instance, in the beginning of the 90's the three main American actors in the car industry needed 8,5 years to develop a new car model, whereas their Japanese competitors used only 4,5 years (Prahalad and Hamel, 1994, p. 179).

Communication is one of the most important factors behind NPD-success (Brown and Eisenhardt, 1995, pp. 343-378; Rothwell, 1994), and according to Tushman (1979) successful projects are characterised by intense communication, where engineers and scientists spend between 50 and 75 percent of their time communicating verbally.

The problem-solving nature of NPD-projects and the importance of information exchange in NPD-settings mean that the management of information is crucial for NPD-success. Therefore communication is likely to have a profound effect on NPD project productiveness, i.e. time, cost and quality.

The purpose of our study is to investigate:

• What are the different information needs within a NPD project?
• How is information transmitted and received by the participants within a NPD project?
• How do companies manage the information flows in a NPD project to improve time, cost and quality goals?

These questions, in conjunction with the theoretical background outlined in chapter 2, generate four propositions. These are set forth in chapter 3.

Communication is an important factor behind new product success, but it is not the only influential factor. Other factors that are found to be important are effective functional integration, planning and control procedures, market orientation and high quality production (Rothwell, 1994, pp. 33-34). Project communication is dependent on these other factors and this study will treat these other factors only in an indirect manner, i.e. how they contribute to enhanced communication within a NPD project and not how to, e.g. implement a decentralised, cross-functional project organisation.

The following chapter Theoretical Background, discusses what earlier research has found on the subject of NPD and project communication. Thereafter in chapter 3, Research Design,. we discuss the chosen methods for data collection and analysis. Then follows a short presentation of Pharmacia and Upjohn and Pharmacia and Upjohn Diagnostics AB. The study’s propositions are also included in the chapter. In the fourth chapter the UniCAP 100 case study data is presented and analysed. In chapter 5 the UniCAP 1000 case is presented and analysed. In the 6th and final chapter we discuss the study’s findings and the implications that could be drawn from the analysed data.

This chapter discusses the theoretic models that are utilised in this study. Needless to say, the chapter is not a complete investigation of what is written on the subject of communication and NPD. The content of this chapter is, rather, what we have found to be relevant for our purpose. Models referenced here concern project communication, organisation and communication and managing project activities. These models will be used to compare the empirical findings of the study with the theoretical propositions (as will be discussed in chapter 3, Research Design).

Tushman’s (1979) model on managing communication networks in projects, is built on the assumption that, due to different information requirements in high performing projects, systematically different communication networks are needed. The model is derived from an empirical study in a R&D laboratory of a large American corporation, and the findings have been synthesised with existing research on the subject (the TIMS studies).

This study focuses on the communication within NPD projects, and the Tushman model is appropriate for this purpose because it allows us to distinguish between different kinds of NPD projects. It also allows us to benchmark the actual situations in the projects studied, to a theoretical ‘best practice’. One competing, but rejected, model on organisational communication was Mendelson’s and Ravindran’s (1999) model on dynamics and performance of information-age organisations. It was rejected, in favour of the Tushman model, because it does not explicitly treat information processing requirements, but only information processing capabilities of an organisation. Many of the model’s propositions regarding information processing capabilities are, however, similar to the Tushman model.

Tushman identified three different groups, into which R&D projects can be grouped. These are research, development and technical service projects. There are three levels of communication: intra-project, intra-firm and extra-firm. These different kinds of R&D projects and levels of communication form a matrix, cf. Table 1 below, in which communication patterns, discovered in successful projects, are plotted.

A research project is a project oriented towards developing new knowledge in several technologies important to the firm. The members of a research project usually have advanced academic degrees. In a high performing research project the intra-project communication is extensive and the structure is decentralised. This means that successful research projects, when dealing with problem solving, rely more on peer decision-making than on supervisors. The connection of the project to other parts of the organisation is weak, but the communication with other R&D activities within the firm is intense. The professional communication outside the firm, with for example universities, is moderate. On the other hand, operational communication with customers and suppliers is relatively weak. Gatekeepers often handle the extra-firm connections.

Tasks in a development project are relatively less complex than are research tasks, but more complicated than technical service projects. The technology may be locally defined but not well understood. The intra-project communication pattern of successful development projects is moderate, a combination of peer and supervisory decision making. The communication is operationally oriented and is often strongly linked to the marketing and manufacturing functions of the organisation. The contact with these areas is widespread and direct. The extra organisational communication is focused on consultants, customers and suppliers. This communication is mediated by certain boundary spanning individuals, often professionally oriented project members.

In the technical service project, the core technologies are relatively well known and mature, and the problems are specific to the firm. This kind of project is more closely tied to the firm than are research projects. The members of the team generally do not have advanced degrees and they are often not professionally oriented. There is relatively little problem-solving or administrative communication within the project. Increased administrative communication can even lower the performance. There is less peer decision making and more supervisory direction. These kinds of projects should be strongly connected to organisational areas outside the laboratory. The operational extra-project communication is moderate and relies on supervisors, and is focused on suppliers and customers.

The Tushman model argues that R&D projects face different kinds of uncertainties, and if these are not properly managed, they can seriously affect NPD effectiveness. One way to manage this uncertainty, is to ensure a match of information processing requirements and capabilities between the different activities that a NPD project involves. The model has four propositions:

The first proposition (Tushman, 1979, p. 268) is that project work requirements vary in their degrees of uncertainty. Research has shown that R&D settings face three basic sources of uncertainty, which affect the information processing requirements.

Firstly – the nature of project task. Project work varies in degrees of newness to the company. It is looked upon as routine or non-routine work. The degree of newness affects the degree of work-related uncertainty facing the project. As the uncertainty increases, the amount of information processing needed to complete the project increases.

Secondly – the nature of project task interdependence. Another source of work-related uncertainty is that the activities within and outside the project could be interdependent. This interdependency affects the information processing requirements, due to the co-ordination needs between different activities.

Thirdly – the nature of project task environment. The external environment is a source of uncertainty. If the external environment changes, or is unknown, the project’s external information processing requirements increase to deal with this uncertainty. The uncertainties facing a NPD project are summarised in the upper half of Figure 1 below.

Souder et al. (1998, p. 531) have shown in a study, that perceived technical and market uncertainties affect NPD projects, and that a focus on integration, through communication between R&D/marketing and R&D/customer, reduces these environmental uncertainties. Integration promotes practises such as prototype development (reducing costly and time-consuming design changes) and market forecasts (giving the developers reasonable expectations for product success). Another study conducted by Song et al. (1997, p. 97) shows that the incorporation of customer needs affected the development of technical proficiency, three times as strongly as the skills of the project team. This points to the importance of communication capabilities within projects. Daghfous and White (1994, pp. 269-270) make a useful distinction between two different information needs. The first one is the need to reduce uncertainty, and it is described as precision. Precision means that relevant information supports expectations, and that the possible variations to these are known. The second need is to avoid negligence. Negligence means that enterprises neglect known or retrievable information, which is critical to the project.

Moenaert et al. (1994, p. 38) proposes that successful innovation projects are characterised by efficient uncertainty reduction in the early stages of NPD. During the initial stages, critical product specifications are to be resolved, and the integration between different functions of R&D and marketing should contribute to the reduction of many uncertainties, such as customer requirements. Empirical testing of this proposition has shown that communication from marketing to R&D significantly correlates to project success, while the reverse, communication from R&D to marketing does not. The reason for this, is that communication from marketing to R&D serves as inputs to design specifications and the like, while the reverse flow is of an informative nature and gives the technical realities of the product development.

Tushman (1979, pp. 268-269) also found that communication networks have varying capabilities to deal effectively with uncertainty and that communication patterns affect an activity’s uncertainty-reduction capabilities. He identifies two components of an efficient communication network. First the intra-project communication capabilities. Communication patterns could be described on a decentralised-to-hierarchical continuum. Decentralised communication patterns increase the opportunity for feedback, error correction, and the convergence of different points of view. Effective decentralised communication patterns were observed to be non-formalistic, and peer consultation played an important role. Decentralised communication patterns increase the ability to deal with uncertainty more effectively (having better information processing capabilities) than do hierarchical networks. The trade-offs are potentially longer response times and higher costs.

A study by Pinto and Pinto (1990, p. 208) showed that the degree of cross-functional co-operation of teams affected the type of communication. Teams with a high degree of cross-functional co-operation display higher levels of informal communication (e.g. informal discussions face-to-face or over the telephone) and informal communication is associated with obtaining project-related information, feedback, and reviewing the progress of the project. Moenaert and Souder (1990, p. 220) have observed that face-to-face contact allows for a fast and efficient transfer of communication with a high degree of understanding between the parties, and that the personal contact motivates the receiver more than does a message sent over any other channel. Face-to-face communication also has its weaknesses. The study shows that many informants perceive face-to-face communication as unstructured, transient, and fragmentary, and that it does not leave the receiver with a hard copy that could be used to justify actions.

A second feature of an efficient communication network, is what is known as extra-project communication capabilities (Tushman, 1979, pp. 268-269). No project is likely to have included in it all necessary skills for the completion of the project, thus extra-project communication capabilities are essential for the success of a NPD project. Communication across organisational boundaries, and with parties external to the company, is often difficult because project teams typically evolve norms, values, and an internal language. As boundary-spanning communication is hindered by these cultural differences, project teams often develop a boundary-spanning function/individual to communicate with important and differentiated areas within and outside the organisation. Extra-project communication deals better with uncertainty if it is directed toward areas with similar technical orientation, problem focus, or professional background, due to similarities in professional culture and language.

One method practised by many companies, that allows for communication to effectively span organisational boundaries, is to shift manpower between different functional areas (cross-functionality) e.g. put an engineer in the marketing function (Moenaert and Souder 1990, p. 221). This boundary-spanning individual can enhance the communication, since he/she readily understands the language of the transmitter/receiver of information.

Tushman (1979, p. 269-270) comes to the conclusion that the degree of uncertainty varies with project characteristics, and that projects therefore have varying information processing requirements. Due to differences in communication patterns, different projects or activities within a project have varying capabilities to deal with uncertainty. Efficient problem-solving is therefore contingent on a match between information processing capability and the information processing requirements of a project (see Figure 2, above). This also means that the communication patterns should adapt to project evolution, i.e. new communication patterns should evolve when the project moves from, e.g. the problem-solving phase to the implementation phase.

Burns and Stalker (1961) studied how organisational structure and management practices depend on the environmental conditions. They have found two different structures and labelled them as mechanic and organic. The mechanic organisation, which performs routine tasks, relies heavily on programmed behaviours, and is relatively slow in responding to the unfamiliar. Organic structures are relatively flexible and adaptive and influence is based on expertise and knowledge rather than on one’s position in the company. The mechanic structure is similar to the hierarchical communication pattern, and the organic structure is similar to the decentralised communication pattern of Tushman (1979). Burns and Stalker (1961) also argue in similarity to Tushman, that the communication in the mechanic structure will be vertical while the communication is lateral in an organisation with an organic structure.

Mintzberg (1983) studied the relation of organisational structures and the environment. What organisational structure to use is in a given situation is dependent on the characteristics of the environment. Mintzberg (1983, p. 262) identifies five basic functions of an organisation:

According to the configuration of these parts, different organisational structures can take shape. Below are the characteristics of the simple structure, professional bureaucracy and the adhocracy structures.

The simple structure has little formalisation. Authority is centralised to a single person. It could be depicted as a flat operating core, and decision making is almost centralised to a one-person strategic-apex. The simple structure should be used when the environment is simple and dynamic, and the number of employees is low. Less repetitive work and informal communication is preferred. The simple structure concentrates power to one person at the top.

The professional bureaucracy combines standardisation with decentralisation. This structure is good for organisations that require high levels of specialised expertise. The power lies in the operating core because there lies the critical skills that the organisation needs, autonomy is provided through decentralisation. The activities of the support staff is focused on serving the operating core. There is a risk of sub-unit conflict in the professional bureaucracy. This structure should be used when matched with large size, a complex and stable environment or a routine technology.

The adhocracy is characterised by a high horizontal differentiation, low vertical differentiation, low formalisation, decentralisation, and great flexibility and responsiveness. Compared to the professional bureaucracy there are few rules and regulations. A novel solution is needed for each task, standardisation and formalisation are therefore inappropriate. Because the adhocracy has little formalisation, the technostructure is almost non-existent. Since all the middle managers, the support staff and operatives are professionals, the traditional distinctions between supervisors, employee, line and staff become blurred. The result is a central pool of expert talent that can be drawn from in order to innovate, solve unique problems and perform flexible activities. Power flows to anyone in the adhocracy with expertise, regardless of his or her position. Conflicts could easily occur in the adhocracy since there are no clear boss-subordinate relationships. The technology in an adhocracy will be non-routine. It will be complex in that it will be drawn on the talents of diverse specialists. This requires co-ordination and integration of specialised and heterogeneous skills for which the adhocracy structure is to be preferred.

Eppinger et al. (1994, pp. 1-13) addresses the problem of task interdependencies of large scale engineering projects, which is also mentioned as one major source of uncertainty by Tushman (above). Activities can be grouped into one of three sequences of task relationships (see Figure 2, below). Tasks could be either dependent (serial), independent (parallel) or interdependent (coupled).

The co-ordination of dependent and independent sequences of tasks is usually straightforward. The problem lies in managing the coupled tasks, since these often require more development time, due to iterations of information transfer between involved activities.

Eppinger proposes a matrix (Design Structure Matrix). The matrix shows the interdependent tasks, and thus eases the organisation of them into integrated blocks of tasks, and thereby forming the basis for an effective transfer of information between tasks. This approach, when working, allows large projects to benefit from some of the characteristics of small integrated projects, by minimising the number of links external to the block of coupled tasks, and thus the number of dependencies (Eppinger et al, 1994, p. 1). Another effect of this is that the interface between the different coupled tasks becomes clearer.

The notion of information transfer in coupled tasks, is developed further by Krishnan et al. (1997, p. 440-443) where effective communication should enable co-ordinating concurrent development. Their model concerns the problem of lead time improvements by overlapping NPD activities, and how this affects cost and quality aspects. The model is constructed out of two basic and interrelated functions of development activities; the speed of activity evolution and the sensitivity of an activity. Evolution refers to when, in time, a preceding (upstream) activity is mature enough to allow dependent (downstream) activities to be started. A high degree of evolution means that information important to the dependent activity, such as product characteristics/specifications, is determined at an early stage. Sensitivity refers to the amount of rework a dependent activity need to undertake, in reaction to new information sent by the upstream activity. High sensitivity implies that a large amount of rework needs to be executed, in response to new upstream information.

The model is based on the idea that upstream activities could pass on information, before the completion of the upstream activity, to the downstream activity, thus enabling shorter lead times and lower costs. This, however, is contingent on the characteristics of the upstream and downstream activities. Krishnan et al. (1997, p. 447-449) construct a matrix, displaying four possible forms of activity overlapping, see Figure 3 below.

	SLOW EVOLUTION	FAST EVOLUTION
LOW SENSITIVITY	Iterative overlapping 1.	Distributive Overlapping 4.
HIGH SENSITIVITY	3. No overlapping or Divisive overlapping	2. Pre-emptive overlapping

As this model implies, there is a trade-off between time, cost and quality. In iterative overlapping, downstream rework is traded off against lead time improvements. Upstream quality is traded off against lead time in pre-emptive overlapping (ibid.). Divisive overlapping is likely to incur some quality losses, while distributive overlapping seems to be the ideal situation.

This chapter introduces the case study company and presents the two cases, UniCAP 100 and UniCAP 1000 that are analysed in the study.

Pharmacia and Upjohn is a global company with 30,000 employees in more than 100 countries. Pharmacia and Upjohn’s core business is the development, manufacturing and selling of prescription pharmaceuticals. The headquarters is located in Kalamazoo, New Jersey, in the United States. (Annual Report P&U, 1998, pp. 1-4) The company’s business is divided into six areas. These are:

RX Pharma which is the company’s core business: prescription pharmaceuticals, with a product portfolio that consists of four business groups, General Therapeutics, Speciality Products, Hospital products and Diversified products (Annual Report P&U, p. 10). Consumer Health is an extension of the prescription pharmaceuticals business. It consists of consumer pharmaceutical products which can be sold over-the-counter. This function serves as a lengthening of the product life cycle of earlier prescription pharmaceuticals. Animal Health markets pharmaceuticals and feed additives for pets and livestock. Technology from R&D in pharmaceuticals for humans can be transferred and used in this area. PCS is the Pharmaceutical Commercial Services department, which supplies pharmaceuticals in bulk to third party customers (Annual Report P&U, 1997, p. 10). Nutrition The nutrition area has a portfolio of intravenous nutritional solutions, that can be used in hospitals as well as in the patient’s home (Annual Report P&U, 1997, p. 10). Diagnostics is an area aimed towards helping physicians in determining whether a patient suffers from allergy or asthma. To determine the cause of the allergy is an essential first step to treat the disease.

Pharmacia and Upjohn Diagnostics develops, produces, and markets blood test systems to support the clinical diagnosis and monitoring of allergy, asthma and autoimmune diseases. Pharmacia and Upjohn Diagnostics has 1,114 employees and an annual turnover of 204 MUSD (1998). The company is the world’s leading supplier of in vitro allergy tests. In vitro (in glass) tests are carried out in test tubes on the laboratory bench, as opposed to in vivo (in life) tests, skin tests on humans (Revealing the secrets, 1999).

The headquarter is in Uppsala, where strategic management, administration, business development, and research and development facilities are placed. The company is divided into two sections. The centre of allergy and asthma, which is located in Uppsala and the centre of autoimmune diseases located in Freiburg, Germany (Annual Report P&U, 1997, p. 10). The company strategy is to focus on areas where dominant positions give high return on investments. About 98% of the production is exported to about 60 different countries (Revealing the secrets, 1999). The allergy and asthma products stand for about 81 percent of the yearly turnover (1998). Their single most important market is Japan, followed by Germany, Italy, and the USA.

The diagnostics for allergies is specified with the help of nearly 500 specific allergens in the Pharmacia CAP Systems (Annual Report P&U, 1997, p. 10). Originally, scientists in Uppsala and Baltimore derived this system from a discovery in 1967 of an antibody, Immunoglobin E or "IgE". This is an antibody that can be specific for hundreds of different allergens. Technically, solid, biochemical-activated sponge reacts with this substance in the patient's blood. The sponge contain an antigen (allergen), the antibodies and a marker binds to the sponge during the process. Fluorescent markers that are set free measure the amount of antibody-antigen bindings. The main idea in the CAP systems is to detect small concentrations of substances in the patient's blood, which indicate a specific disease or condition. (Revealing the secrets, 1999)

UniCAP® is one of the main products of Pharmacia and Upjohn Diagnostics. UniCAP is a laboratory system, comprising in vitro tests and automated instruments. UniCAP is a new generation of the Pharmacia CAP systems. The two case studies conducted in this thesis investigate two of the automated instruments developed within the UniCAP laboratory system.

The first case is the development project of the UniCAP 100 instrument, which is an automatic batch driven (blood tests and reagents are loaded by hand) in-vitro test system. UniCAP 100 is a desktop instrument and aims to satisfy the needs of smaller laboratories. It has a capacity of 100 tests per day. The UniCAP 100 has been on the market since 1996. It is distributed to all markets. The instrument was developed in a small project that employed about 15 persons, while the research of chemical technology occupied about 35 persons in another department. The company regards the UniCAP 100 instrument as a successful product. The instrument is depicted in Figure 4.

The second case studied is the UniCAP 1000 instrument. It is under development to satisfy the needs of large laboratories, mainly in Japan. It is a completely automated in-vitro test system with a high capacity (1000 tests per day). These instruments are to be distributed and installed in the laboratories during the summer of year 2000. The instrument and the service are completely free to the customer, which pays for the allergens (the caps) and the reagents. The project occupies about 80 persons today. There are several differences between this instrument and the UniCAP 100 instrument. The UniCAP 1000 instrument uses the same kind of reagents for the different allergens, which was not possible with the UniCAP 100 instrument. The robotics is much more complex. It is completely automatic. The UniCAP 1000 instrument is depicted in Figure 5.

The Instrument and Software department, a sub-unit of the R&D organisation at the Pharmacia and Upjohn Diagnostics AB developed both instruments. The development of the instrument draws upon several different knowledge bases these are: chemistry, mechanics, software, hydraulics and electronics. The UniCAP System consists of Instrument, Software, Reagent and Pre-production, Market, Quality, Manuals and Production of Reagents. Seventy-five percent of the department resources went into the UniCAP 1000 instrument project.

This chapter cover the methodology used in answering the research questions, stated in chapter 1 (Introduction). We begin with the study’s propositions, followed by a discussion of the methods used for data collection and analysis.

Based on the theoretical framework and the research questions outlined in chapter 2, the study states four propositions concerning successful NPD project communication:

1. A high degree of uncertainty increases the communication requirements of a project.
2. Technological uncertainties are resolved through decentralised informal peer consultation and decision making, while market and interdependency uncertainties are handled through cross-functional communication via formal channels.
3. Coupling of interdependent activities enhances effective communication.
4. Effective communication enables projects to cut development time, through correct and timely transfer of critical information between development activities.

The different information needs of a NPD project are, following Tushman’s model, predicted to depend on the uncertainties facing the project. A high degree of uncertainty should thus increase the communication requirements of a project.

The second proposition concerns how information is transmitted and received within a NPD project. It is generated from Tushman’s discussion of the capabilities of communication patterns of a development project.

The third and fourth proposition treats how companies manage the information flows within a NPD project to improve time, cost and quality goals. The Eppinger model predicts that an increased coupling of interdependent activities enhances effective communication, while the Krishnan model predict that this coupling enables NPD project teams to transfer correct information in a timely fashion. These four propositions will be tested in the two NPD projects presented above.

This study will make use of primary and secondary data. The study relies mostly on primary data. Collection of primary data has mainly been achieved through interviews, but also through the study of company-internal documentation and archival records, such as project plans, projects meeting protocols and organisation charts. Secondary data used in the study consist of Annual Reports and brochures both from the Pharmacia and Upjohn Companies The reading also consisted of business literature, mainly articles on NPD and communication. The business literature was found using the DISA search engine (for books found in the Uppsala university libraries) and the LIBRIS search engine (for books found at other Universities within Sweden). The Internet was also used for searching articles on NPD and communication. The analytic approach is qualitative.

To be able to answer the questions of information needs, information transfer and how to improve effectiveness through communication within a NPD project, we must choose a research strategy that allows us to thoroughly analyse and go into deep detail in these issues. The analysis of the management of information flows is best conducted within its real-life context, and since it is not possible for us to acquire the "required control of behaviour" necessary to conduct an experiment, the case study approach was chosen. According to Yin (1994) archival analysis might be an alternative research strategy when answering "how" and "why" questions. This study attempts to measure both the formal and informal communication within a NPD project. A study of archival records is not likely to give us much useful data on the informal communication, but archival records will be used for triangulation purposes in the case of formal communication.

The thesis uses a multiple case design with a single unit of analysis. The unit of analysis is the intra-project communication pattern. The two cases of this study were selected on what Yin (1994, p. 46) refers to as a theoretical replication basis. This means that the two cases are likely to show contrasting results, and thus allow us to better observe the factors lying behind the differences in the communication patterns between the two projects. The two projects, the UniCAP 100 and UniCAP 1000 projects, developed similar products and it should therefore be easier to study the differences in the project communication patterns.

The connection between the theoretical background and its operational definitions (Eriksson & Wiedersheim-Paul, 1997, pp. 38-39) affects the internal validity of a study. Internal validity is a concern if the study is supposed to determine a causal relationship, in this study for example between uncertainties and communication. Internal validity is also at risk when making inferences, a case study usually uses inferences each time an event cannot be directly observed, for example when interviews are used. To strengthen the internal validity one may use the methods pattern-matching, explanation–building and time-series analysis. External validity concerns whether it is possible to generalise the study’s findings beyond the immediate case study. This problem is solved by the choice of case. You can generalise to other cases that are chosen on the same criteria as the case you are studying. But this requires that the findings to be tested in some other cases with the same results before they can be generalised. (Yin, 1994, pp. 35-36)

In the two cases, the collected data was analysed using a pattern-matching technique (see Yin, 1994, p. 110) in which the evidence collected is matched against the study’s propositions. If the data collected matches the propositions, this indicates an affirmative finding, given that rival explanations are eliminated. If collected data fails to match the predicted outcomes, the propositions have to be rejected.

In the first case, the UniCAP 100 project, one of the four informal sub-project managers was interviewed (by telephone). Three representatives of the UniCAP 100 project work force were also interviewed. These persons are or were also involved in the UniCAP 1000 project, which allowed them to make useful comparisons between the communication patterns of the two projects.

The second case, the UniCAP 1000 project, which is studied during two different forms for the organisation of project activities, is theoretically predicted to show a negative pattern during the first organisational solution. (i.e. it will not support the study’s propositions). The UniCAP 1000 project is also predicted to show a positive pattern (a support for the study’s propositions) during the second organisation of project activities. The second case, the UniCAP 100 project, is predicted to show a positive pattern. If these predictions are true the propositions are more strongly supported, since they were chosen on a theoretical replication basis.

The analysis of the UniCAP 1000 project also includes a time-series analysis (see Yin, 1994, pp. 113-119) where the organisation, and its effect on the communication patterns of the project, is analysed at two points in time. The concern here is to identify how the organisational forms have affected the communication patterns, and to trace these factors at the two time points. It is important here to distinguish between the changes that have come as a natural answer to project evolution, i.e. functions to prepare for production and changes that have been implemented to enhance the effectiveness of project work (and thus likely to affect the communication pattern of these activities).

The interviewees from the UniCAP 100 and UniCAP 1000’s organisation were chosen by the authors with the help of Nina Brunk working at the Instrument and Software Department at Pharmacia and Upjohn Diagnostics. All project managers were asked for an appointment but not all had the time.

Interviews were conducted with the former project manager (2 hours), the present project manager (1 hour) and the project co-ordinator (1,5 hours) of the UniCAP 1000 project. In the UniCAP 1000 project, one hour interviews were conducted with 3 of the sub-project managers, representing four out of the seven sub-projects of the former UniCAP 1000 organisation (later referred to as organisation A). The interview with the sub-project manager of the CAP and Sample Modules were conducted by telephone. Of the new, present, organisation (later referred to as organisation B) of project activities we interviewed seven out of ten sub-project managers (1 hour per interview). Two of these interviews were conducted through E-mail (giving short numerical answers). One interview was also conducted with the staff function of the UniCAP 1000 project. The interviews were conducted at Pharmacia and Upjohn Diagnostics. All the interviews except for the telephone interviews and the E-mail interviews were taped and transcripts were made to strengthen the reliability. The compiled data from each interview was sent to the interviewee for comments. At least two persons from the thesis group were present at each interview.

The project managers were interviewed to allow for us to get the necessary overview of the project activities, their interdependencies and the uncertainties facing the project. The interviewees (project managers and sub-project managers) from UniCAP 1000 project covered most parts of the two organisations of project activities. This allowed us to get a better picture of the intra-project communication, the communication between the different sub-projects and also to compare the different statements with each other.

In the UniCAP 1000 case some sub-project managers have a role that is twofold. Firstly they, of course, manage their part of the project, and secondly they take part in the actual production, which makes them part of both the work force and the project management. The sub-project managers will thus, in some cases, come to represent two levels of the project organisation.

The interview protocol was revised two times, once during an instructor review, and once by the authors, following a preparatory interview with the former project manager of the UniCAP 1000 project. The second revision was substantially leading to two separate protocols, one more focused and structured, aimed at the project leaders, and one more open-ended, aimed at the project co-ordinator. The present project manager was interviewed using a revised ‘sub-project manager protocol’. The reason for this is that he recently entered upon his office (about one month ago). The interviews were conducted in Swedish, allowing the interviewees to express themselves more freely. See Appendix 1 and Appendix 2 respectively for the interview protocols in English.

This study investigates the intra-project communication of two NPD projects. The focus of our study is on the problem-solving communication, i.e. the transfer of information regarding project co-ordination, technical specifications, et c., between different activities within NPD projects. This study labels effective communication as problem-solving communication.

Effective communication occurs when the information requirements of a party are met (as proposed by the Tushman model, above) allowing the party to fulfil its obligations. In this sense our study measures effective communication in three instances. Firstly: are the different parties interdependent, i.e. do they need to co-ordinate their activities in some way? Secondly: if interdependent, do they communicate with each other? Thirdly: is the communication timely and correct? We are thereby assuming the content of communication to be project related and of a problem solving nature.

The interviewees were asked questions on uncertainty and interdependence (to measure the project and more specific the sub-projects communication requirements), communication (which parts of the project they communicated with and to what extent), and the correctness of information (if they received faulty information and if information was timely). See the interview protocols in Appendix 1 and Appendix 2.

The project consisted of one project manager and four sub-project managers, heading the four different sub-projects. These were Process (responsible for the functioning of the chemical process in the instrument), Software, Electronics and Mechanics. The organisation was not an actual project organisation it existed informally within the main organisation. The project members worked full time on the project. Two to four people worked on each sub-project.

The technological uncertainty was high due the new chemical process the instrument utilised. Except for the chemical process no other technology used in the development of the instrument were totally new to the company.

The market uncertainty was stated to be in between medium and high. The reason for this uncertainty was that the market function put up customer specifications, which were a mix of both wishes and demands. The contacts with the marketing function were mainly through the weekly project meetings, but the attendance from the marketing function was varying. This lead to that the project team had to ask the marketing function directly whether it was possible to deviate from the customer specifications. This was often possible since the marketing function wanted an instrument on the market as soon as possible.

The communication uncertainties were perceived to be low. Technical problems were discussed informally around a prototype instrument. At these meetings the whole project team was present, which gave an efficient diffusion of information.

The greatest interdependencies existed between Process and Software, while Electronics and Mechanics had more of a supporting function for the instrument. The dependencies for input were easily resolved through direct contact and it was easy to get the needed information. The informal communication was extensive, about 90 percent of the total communication. Formal project documentation was not extensive, there were few specifications and test results. Project documentation was written afterwards to improve the traceability of instrument changes. The reason for this was that a focus on actual production would bring the instrument faster to the market. Otherwise formal communication consisted of the weekly project meetings. At these meetings decisions were decided upon formally.

We classify the project as a development project. There were few rules and regulations. This might be because the UniCAP had communication patterns that have some of the characteristics of a research project (according to Tushman’s classification of NPD-projects).

A development project is characterised by moderate intra-project communication and a combination of peer and supervisory decision making. The structure of communication should be in between decentralised and centralised. In a research project the amount of intra project communication should be intense and decentralised (see section 2.2.1, pp. 3-5 above). The Process and Instrument Software sub-projects of the UniCAP 100 project had intense contacts due to the strong interdependencies. The intense communication reduced the interdependency uncertainties between these two parties. Electronics and Mechanics had, as stated above, a more supporting function and also a less intensive information exchange.

Communication was informal and all members of the project continuously met in the same room for problem solving. This allowed for intense communication. Although the project had informal project managers decentralised peer-decision making dominated. Formal decisions were taken afterwards to document project progress.

The newness of the chemical process was high (it was new to the world). The newness of other technologies used in the development of the instrument was low. The technological uncertainty was high because of the complexity of integrating the chemical process into the instrument. The project task interdependencies of the UniCAP 100 project were high. Giving a high degree of overall project uncertainty. These uncertainties were lowered by effective cross-functional communication. The external environment was uncertain. The customer requirements, which were a mix of demands and wishes from the marketing function made the market uncertainties higher. This uncertainty was also increased by insufficient communication with the marketing function.

The staff was assigned to work on the project based on their previous experience. The project was simply a group of people with specialised knowledge that had been organised to develop the product. According to Mintzberg’s classification the UniCAP 100 project was similar to the Adhocracy (section 2.2, pp. 9-10). The UniCAP 100 project had, in essence, a cross-functional project organisation. The project team was small, consisting of about 15 people. All project members attended the informal meetings, which took place around a prototype instrument, to discuss project problem solving and evolution. Ninety percent of all communication was informal. The communication network could therefore be said to be fully connected, i.e. all sub-projects were communicating with each other.

This chapter presents the interviews made with representatives of the former organisation, referred to as organisation A, and the present organisation, referred to as organisation B, of the UniCAP 1000 project.

In the beginning the project was divided into seven modules named: the System Functions Module (responsible for the physical location of different components and modules within the chassis, some support functions and external design), Electronics and Power, Process Module (contains the chemical process which performs the actual tests, CAP module (the cap contains the allergens), a Sample (Test Tube) Module and finally Instrument Software. The organisation also included an Instrument Verification Function. The organisation is depicted in Figure 6.

These modules are, except for the Verification function, the same as the physical parts of the instrument. The reason for this approach was to make the instrument more distinct, so that it would be easier to describe the interfaces. The subprojects consisted mainly of people of the same professional background. There were no real cross-functional teams. The intended plan was that the subprojects would work on their problems separately and then put everything together. Two of the sub-projects (CAP and Sample Module) were outsourced to a supplier (Partnertech).

The interviewees were asked to estimate, on a scale ranging from 1 to 7, where 1 is low and 7 is high, their perceived uncertainties regarding technology, market, time and communication. These uncertainties were measured as their general perceived degree of uncertainty during their project tenure. Their answers are presented in the text and in Table 2, below. In the table a line indicates that no answer was given and a word (low, medium and high) indicates that the interviewee could not give a numerical answer.

The Electronics & Power, Process, Instrument Verification modules were not investigated.

The interviewees were further asked to estimate the newness of the technology to the company/sub-project team, their input dependency (to what extent they were dependent on inputs form the other sub-projects), timely information (how often information arrived as planned) and to what extent they planned or acted to incorporate anticipated information needs. Table 3, below, presents the data, which is commented upon in the following sections.

The Electronics & Power, Process, Instrument Verification modules were not investigated.

There were no uncertainties with the underlying technology. The greatest technical uncertainty was the complexity, the size of the project and the many interrelationships. The market uncertainty was low. The customer requirements were known. The communication between projects with the same professional background was good. There was some communication problems between persons with different professional backgrounds.

At the module testing, when the modules were put together, the instrument did not work as it should. The assembled instrument clarified the interdependencies and the interfaces that had up to this point been unclear. This had an positive effect upon project communication. It increased the intra sub-project communication.

Partnertech had earlier worked with P&U Diagnostics in the UniCAP 100 project. The company had the required knowledge for these kinds of projects and the technical uncertainty was low. The market uncertainty was high, the requirements from the marketing function of UniCAP System could change suddenly and the work had to be redone. Because of being an external supplier, Partnertech was able to have a higher degree of documentation than the other subprojects - this lowered their time uncertainties. Partnertech do not think that the subproject’s communication was suffering from their location in Åtvidaberg.

The project managers for the CAP module and the Test Tube module spent two days in Uppsala and three days in Åtvidaberg each week. At the beginning of the project the interfaces were clear, but later on they "slipped away and became like a snake pit". The project was not well described and documented. Getting the information that was needed required a lot of personal commitment. With some persons the communication was good but there were problems with others, due to resistance against outsourcing. The hardest group to get information from was the marketing group. The communication with them was mostly informal.

The technological uncertainty was perceived to be average. The technical aspects of the tasks were relatively well known to the System Functions sub-project, apart from the cold-storage technology. Uncertainty was also increased by the complexity of the instrument compared to the UniCAP 100 project. The customer specifications were varying and sometimes becoming blurred. This gave an average market uncertainty. The sub-project group knew for sure that they could not clear the deadlines, but was unsure if it would be moved forward. There was a lot of information communicated in the project, but it was not certain whether it got through to the intended receivers. There was no response to information sent to the management of the company. To improve certainty, they tried to have a lot of immediate contact concerning construction and features, with the marketing function. The technical aspects of the development were solved internally in the sub-project.

The sub-project is responsible for building the chassis in which to assemble the different modules of the instrument, they were therefore a receiver of information. The sub-project was highly dependent on inputs from other sub-projects to be able to fit the different components together in the instrument. They had to search for the information they needed. There were also problems to get technical information before decisions were taken, and thereby the ability to influence the decisions. Informal communication was used to solve technical problems, and formal to inform oneself on the project status. There were continuous informal contacts between sub-projects within the project and a lot of information exchange. The formal information was not always enough to act upon and people were sometimes reluctant to give away information. They wanted to finish their work before communicating anything. The project manager thinks that with more strict rules guarding the information exchange the project might have been finished on time.

The technique was well known and therefore the technological uncertainty was perceived to be low. The market uncertainties were medium to low. They got good response from the marketing function when they contacted it. The time uncertainties were high. The sub-project was often behind schedule and they had to employ more personnel (the sub-project was reinforced several times). According to ISW this had to do with that the department is not used to work and plan such large projects. The communication uncertainty was average. There was much guessing on information and sometimes the collected information did not fit.

The input dependency was high. The ISW sub-project needed input from all other sub-projects. The information received was seldom timely. It was planned that the other sub-projects should deliver specifications to the ISW sub-project, but this did not work. The second step was that the ISW sub-project gained responsibility for the specifications, but the actual content, what the software should perform, should be communicated to them. This worked badly to. To meet these communication uncertainties, different organisational functions were developed. These functions consisted of sub-projects important to the ISW sub-project. This meant that the ISW sub-project was more and more involved with the other sub-projects in deciding their specifications.

This section presents data on sub-project interdependencies, the communication networks between the sub-projects and the structure of communication.

The System Functions sub-project needs input from Electronics and Power, Process Module, CAP Module and the Test Tube Module to be sure that the different components fit in the instrument (see Figure 7, below). Electronics and Power have connections with System Functions but also the Process Module. The CAP and Sample Modules have connections with System Functions, Electronics and Power and the Process Module. Instrument Software is dependent upon input from Electronics and Power, Process Module and the CAP and Test Tube Modules. The ellipse circling the Test Tube and CAP sub-projects indicates that Partnertech externally supplied these to the project.

In the figure the double-headed arrows indicate that there exists interdependence between the sub-projects. A single-headed arrow indicates information dependence on the behalf of the receiver.

The System Functions group communicated with all other sub-projects (see Table 4, below). The stated frequency of this communication was very often. The communication often concerned ‘soft integration’, which means that the different modules and components are test assembled in a CAD model. If the module/component fit (within the chassis) and does not collide with other parts within the instrument it is all right to change the dimensions of the part. This could also mean that, in order to make room, parts within the instrument would have to be rearranged.

The CAP and Sample Modules also communicated with every other sub-project. The stated frequency of this communication was now and then. The communication concerned specifications, choice of components and time schedules. The Instrument Software sub-project communicates with all parts of the project. The communication was mainly about specifications. No frequency measure was stated, but the ISW communicated with other sub-projects many times a day to solve problems related to project implementation. The communication of Electronics and Power, Process Module and Instrument Verification were not investigated.

All interviewees stated that very little information was delivered to them, rather they had to actively search for the information they needed.

Much of the communication within the project could be described as informal. Informal communication between project members with different background was common and the informal communication between the different sub-projects was stated to be very common. The communication was decentralised and informal communication channels, such as unplanned meeting or direct face-to-face discussions, were used to solve technical problems. One problem associated with the use of informal communication channels was that the verbal information sometimes was misinterpreted and thereby faulty information was, not intentionally, transmitted. Two of the sub-projects were located in Åtvidaberg, which lead to that the informal communication between these the other sub-projects were low, since the sub-project manager was in Uppsala two days a week.

The main channel for formal communication was the weekly project meeting. At these meetings the project activities were co-ordinated, i.e. planning and scheduling issues were treated. The conditions for communicating information during the project meetings were perceived to be good. Otherwise the formal project documentation was perceived to be insufficient. Effects of this were that the interfaces became blurred and that responsibilities in the ‘boarder land’ became unclear. The two sub-projects that were outsourced stated that they benefited from having better formal documentation routines than the rest of the project, they had higher demands since they were external suppliers.

During the fall 1998 the project was reorganised. Management felt that the project needed a reorganisation since it was behind schedule. The project needed enhanced communication between interdependent functions and to formalise some of the information flows.

The role of the project manager has been changed in the new organisation. It is more about managing the project staff than the technical aspects of the project. He defines goals and activities, look to resources and see to it that the staff is comfortable. In the old organisation the project manager had the responsibility over both staff and technical issues and due to the project size didn’t manage to carry out all necessary activities. New functions have been added in the new project organisation (see Figure 8). Integration and Verification is made more distinct. The technical leadership lies in a separate function, the System Architect Group which had the responsibilities for System design/description; the Instrument Change Control Board (ICCB) which handles changes and documents them; Technical Reviews that shall check the overall construction; the Issue Management which collects and documents known problems within the project, so that problem becomes officially known to all parts of the project. Project planning and Administration was not new but is reinforced. The function shall structure and supervise the documentation that has earlier been neglected. It has responsibility over what documents that should be prepared and how these should be available. The Quality function is also reinforced.

All functions have a separate sub-project manager. Some of the sub-projects are multidisciplinary. These are Process Module, Pipeting, Sample and Reagents Handling and Hydraulics. In these there are people with competencies of software, chemistry, mechanics, and electronics.

The interviewees were asked to estimate, on a scale ranging from 1 to 7, where 1 is low and 7 is high, their, at present, perceived uncertainties regarding technology, market, time and communication. Their answers are presented in the text and in Table 5 below. In the table a line indicates that no answer was given and a word (low, medium and high) indicates that the interviewee did not give a numerical answer.

The interviewees were further asked to estimate the newness of the technology to the company/sub-project team, their input dependency (to what extent they were dependent on inputs form the other sub-projects), timely information (how often information arrived as planned) and to what extent they planned or acted to incorporate anticipated information needs. Table 6, below, presents the data, which is commented upon in the following sections.

Communication uncertainty is increased in the project by sub-projects not being aware of the information needs of the other activities within the project. This might be explained with that they do not have enough time to understand each other’s role. Time uncertainties are also high. It is very easy to lose focus on the critical tasks of the project. Uncertainties of market demands are high.

Communication with the marketing function could be improved. A market representative from the market department of Pharmacia and Upjohn Diagnostics is invited to the weekly instrument development project meetings, but his attendance is low. An ongoing dialog with the market could improve productiveness by resolving issues such as whether it is possible to reduce the functionality of the instrument if that would decrease development time.

In the new organisation uncertainties have decreased because they focus on the planning of project activities and the interdependencies among the different sub-projects are better known. Despite this they sometimes still have problems with the integration of the product. The System Architect Group is created just for this task of planning and thereby to decrease some of the information uncertainties.

The Technological uncertainty is perceived to be medium. The technology is fairly well known. The market uncertainty is also medium. The time constraints were medium to high while the communication uncertainties is high. The input dependencies were also high.

Technological and market uncertainties were medium to high, while time uncertainties were perceived to be high. The communication uncertainties facing the sub-project were low. Some new technologies were utilised and the input dependency was high.

In the hydraulics sub-project the technological uncertainties is medium, some chemicals in the system are undefined and these could possibly have a negative effect on the system. The time uncertainties are low. The sub-project will be able to incorporate changes in reaction to new demands from the other sub-projects. The communication uncertainty is medium and concerns the communication of how to build a consciousness about other project activities and project work.

Some new technology is used by the hydraulics sub-project and a couple of the technical solutions have disadvantages that are hard to deal with. The input dependency is high, since hydraulics have interfaces to almost all other sub-projects. Information is seldom timely and for the most part the sub-project actively collects the needed information.

Some of the technology had to be innovated and some existing technology had to be modified. Viability tests of the developed/used technology reduced the technological uncertainties. The technical problems increased time uncertainty. The uncertainty considering customer demands were high. Several times they thought they knew market demands but were proven to be wrong. If communication with the market had been better the subproject would have been enabled to develop a more "correct product in time". Many times specifications were developed that were not demanded by the customer; the instrument has in this aspect too much functionality.

The mechanics sub-project perceived the technological uncertainty to be low because of it being well known and mature. The market uncertainties were perceived to be average because customer expectations might vary on instrument size and appearance, these were however not looked upon as a threat to project success. The largest uncertainty was space. The instrument is crowded internally, so changes in one of the functions of the instrument are likely to induce rearrangements of other functions/components. This is time consuming and explains why the time uncertainty is perceived to be average, when it would otherwise have been low. The communication uncertainty is high, because while one sub-project could make changes without notifying them, they could have allowed another sub-project to implement changes, affecting the same space of the chassis, this causes a risk of collision of components within the instrument.

No new technology is used in the mechanics sub-project. The sub-project is highly dependent on input from the other sub-projects to get information needed to build the chassis and to effectively plan for spacing within the instrument. They often receive timely information. The sub-project often plans or acts ahead to incorporate information necessary for their work, to learn what is going on and to learn about likely changes in the different functions of the instrument.

The perceived technological uncertainty is low in the instrument software (ISW) sub-project since the technology is well known to the sub-project members. The market uncertainty is medium. The customer requirements are fairly well known. The time uncertainty is also medium. The sub-project has enough time to finish its work in time. The communication uncertainty is high because the input often does not come in the right form and/or the quality of information is not good enough to act upon. Information that should be in written form, i.e. through specifications, is instead communicated verbally.

Some new technology is utilised by the ISW sub-project since the development in software technology is very fast. The sub-project is highly dependent on input from the other functional areas. The ISW is described as a mirror image of the instrument. All functionality performed by the instrument has an ‘mirror image’ in the programme-code. Therefore they are involved in almost every other area of the project. The input is often late which leads to that planning ahead to incorporate needed information is a necessity. This activity is very common for the ISW sub-project.

The integration and verification sub-project faces a high technological uncertainty due to problems with getting the different functions of the instrument to work properly together. The market uncertainty is medium. The time uncertainties are high, because of the technological uncertainties mentioned above. The sub-project measures the communication uncertainty as medium. They know what type of information they need. The main uncertainty lies in the aspect of timely inputs.

The technology behind the integration and verification is in many aspects new to the sub-project team. These kinds of tests have not been conducted before at the department. The input dependency is very high because integration and verification stands as receiver of deliverables all sub-systems of the instrument, except for the process module, which is tested separately. The sub-project perceives that needed information arrives late and they often act to incorporate needed information.

This section presents data on sub-project interdependencies, the communication networks between the sub-projects and the structure of communication.

There are very strong interdependencies between the different sub-projects (see Figure 9, below). In the figure the double-headed arrows indicate that there exists interdependence between the sub-projects. A single-headed arrow indicates information dependence on the behalf of the receiver. Process Development (responsible for getting the chemical process to work on the instrument) and Process Module (responsible for the electronics, mechanics and software parts for the process part of the instrument) are tightly connected (indicated with a thick line). Sample and Reagent, Pipeting, Process Module and Process Development are dependent upon one another.

Hydraulics has connections to all other sub-projects except the Operator Software (OpS). Electronics and Power got some connection with mechanics, which in turn affect the hydraulics. There is a weak dependency with Sample and Reagent as well as Pipeting (weak dependencies are not depicted in the figure).

Electronics and Power and Process Module are strongly dependent on one another (indicated by a thick line). The dependency between Electronics and Power and Process Development are not that strong. Instrument Software (ISW) got connections with every other sub-project. There is a weak connection between ISW and Mechanics (not depicted in the figure). There is an interdependency relationship between ISW and OpS. The OpS got weak connections with Integration and Verification (I&V) and Process Development, which uses the OpS to test-drive the instrument. I&V which need deliverables from all sub-projects to be able to integrate the different modules and to verify functionality. Mechanics are most involved with the Process Module, but dependent on information from the other sub-projects to assemble the instrument. The inter-dependencies between Pipeting, Process Module, Sample and Reagent and OpS were not studied.

Process Development (see Table 7, below) communicates very often with the Process Module and ISW sub-projects. They seldom communicate with Sample and Reagent, Pipeting, OpS and Hydraulics. The sub-project communicates now and then with the Integration and Verification group. Process Module communicates often with Process Development, now and then with electronics and power and integration and verification. The sub-project communicates seldom with pipeting, sample and reagent handling and hydraulics. The Hydraulics sub-project communicates very often with Process Module and Sample and Reagents. The sub-project communicates seldom with Mechanics. They communicate often with ISW and Electronics and Power.

Electronics and Power communicate often with Process Module, Hydraulics, Mechanics and ISW. They communicate now and then with Integration and Verification and Sample and Reagents. The group seldom communicates with Process Development, Pipeting and OpS. Instrument Software communicates always with Process Module and very often with Process Development, Electronics and Power, OpS and Hydraulics. The sub-project communicate often with Pipeting and Sample and Reagents, and almost constantly with Integration and Verification. Integration and Verification communicates with all sub-projects. No frequency Figure was stated but the sub-project communicates several times per day with other sub-projects to solve problems related to project implementation. Mechanics communicate (at present) almost constantly with Hydraulics and very often with Process Module and Electronics and Power.

Just as in the organisation A of project activities all interviewees stated that very little information was delivered to them, they still had to actively search for the information they needed.

The informal communication between project members with different professional background is stated to be very common within both the cross-functional and functional sub-projects. The informal communication between different sub-projects was stated to be common. The informal communication channel most frequently used, was the unplanned meeting.

The formal communication consists of the weekly project meetings, the distribution of project and sub-project meeting protocols (through E-mail) and project documentation (reporting of instrument imperfections and planned changes, managed by the System Architecture Board). Some sub-projects stated that much of the planning and scheduling still were informal and that the project still was suffering from too little formal communication.

A research project is oriented towards developing new knowledge in several technologies that are important to the company. As the UniCAP 1000 project was aimed towards the automating of known technologies, not to develop new ones, it cannot be defined as a research project. As the technologies for automating on the other hand is not well understood, and the automation requires a modification of the testing technology, it is not a technical service project. We will therefore define the project as a development project.

In organisation A of UniCAP 1000 project there was continuous informal contacts between subprojects and a lot of information exchange. It was not certain whether information transmitted reached the intended receivers. To get the needed information one often had to take the initiative and search for it. This sometimes created problems in influencing specifications before they were formally determined, to check whether the different specifications were consistent with each other and to be able to influence the design.

According to Tushman’s model of communication, successful development projects should have moderate intra-project communication. The informal intra-project communication was intense and it seems that the information exchange pattern in this project more resembles a research project where information exchange is intensive and problem solving rely more on peer decision making than on managers. In a successful development project the intra project communication is moderate and the decision making is a combination of peer decision making and supervisory decision making.

Contacts with the main organisation that we encountered were either with the managers of the line organisation of the Instrument and Software department or with the marketing function. The contacts with the line managers and the marketing function were either informal or through the project meetings. The contacts with the marketing functions were made by direct contact. These contacts were perceived to be important for the project, because the information from marketing forms the basis of the specifications.

According to Tushman extra organisational contacts in development projects are mediated by boundary spanning individuals and are focused on customers and suppliers. The communication with the partner was intense and made by direct contact. Specific individuals also mediated the contacts as the partner had well functioning communication with some persons in the project and problems communicating with others depending on personality. According to Tushman the communication with suppliers is usually weak in development projects, but the role of the supplier in this case was twofold. They first and foremost brought technical expertise into the project and took part in the development of the instrument, but they also supplied the physical components and constructions they developed. They were thus both consultants and suppliers. The contacts with the customers were mediated by the marketing function. The customer involvement can be said to be low.

The informal communication is still extensive between the sub-projects. Most interviewees say that communication between subprojects is nearly as high as communication inside the subproject. The sub-projects still have to actively seek the information they need. We would like to interpret this as an effect of too little formal communication. Formal communication could structure the information exchange and thus reduce the amount of informal communication. The new organisation has implemented procedures; supervised by the System Architect Group, for a formal communication of design changes, but these protocols are not always used. Some project managers also think there is too much information in the system so that information is hard to find it.

The communication with the marketing function seems to be weaker compared to organisation A. All interviewees complain of uncertain contacts with the marketing function. Some sub-project managers stated that; if the communication with the marketing function had been better they would have been able to develop a more correct product in time.

Extra firm communication with the customers is made through the Market Function, but the customer requirements are not mirrored in the product specifications, e.g. the product specifications contain more automation than the customers need. The communication with the partner is not as high now as the Cap Module and Sample Modules are finished. The project still has some resources borrowed permanently by the project and contacts have been established to enable adapting the modules when needed. The operational Extra-firm contacts can therefore be said to be moderate and professional contacts to be weak as in a pattern for a development project.

In organisation A, the technological newness was perceived to be low or average by all interviewees. It was the perceived complexity of the task that heightened the technological uncertainties in the eyes of the project manager. To the other interviewees the perceived technological uncertainties had approximately the same levels as the newness of the technology (see Table 2 and 3 above). According to Tushman the degree of newness affects the uncertainties facing the project and the amount of information processing needed is then also heightened. In this organisation information exchange should thus be relatively low. This was the case up until the moment when the instrument was put together and tested. The instrument did not work as it should. At this moment the interdependencies became apparent and the communication became more extensive. It seems that the project managers had made too low estimates of the uncertainties facing the project.

The technological uncertainty in the project has increased and according to Tushman high uncertainties increases the information processing requirements. The formal communication has increased and new staff functions have been created to administrate this communication. Planning for information is now marginally lower. The communication uncertainties are still perceived to be high. The amount of timely information is medium. This was also the case in the former organisation.

The nature of project task interdependencies are high throughout the project, most managers are dependent on information from all or many other subprojects. The input dependencies are high. (Figure 7 and 9) These uncertainties, however, did not change with the reorganisation of the project. It is the complexity of the interrelationships that make the co-ordination of information harder. Many project managers, especially in the new organisation, say that information that should be delivered has instead to be gathered. It seems like the sub-projects are not aware of how dependent the other sub-projects are upon their information. The question of how information/knowledge important to the project were diffused to the different activities, the project managers answered that they communicated this kind of information through project meetings. It seems that this is insufficient as most sub-projects are out searching for information themselves.

The project manager of organisation A perceived the market uncertainties to be low. The sub-project managers perceived it to be low to average. After the reorganisation the market uncertainty is perceived to be a little higher than it was earlier. The input dependencies are still perceived to be high, and most interviewees say that informal information exchange is high. The external environment has changed. The target markets of the UniCAP 1000 instrument were initially medium to large laboratories, in the United States, Europe and the Far East. The focus is now on large laboratories in Japan.

According to Tushman a project's external processing requirements increase with changes in the environment. Focusing on fewer markets should decrease the environmental uncertainty. But many project managers, in the new organisation, perceived the market uncertainties to be higher. This is due to the fact that the change of target markets was not communicated to the project team through changed specifications (regarding the customer requirements).

In the new project the Market uncertainties are perceived to be average to high. The contact with the Marketing function is perceived to be the main reason. Studies by Song et al (1997) and Moenaert et al. (1994) promotes a strong coupling between marketing and development. In the UniCAP 1000 project the marketing function often did not participate in the project meetings. A higher attendance should allow for better communication between the market and the project team.

Moenart et al. proposes that successful projects are characterised by efficient uncertainty reduction in the early stages of NPD. This has not happened. The uncertainties are even perceived to be higher in the new organisation. Many sub-project managers said that the lack of clear specification has increased the uncertainties.

The organisation A of the UniCAP 1000 project was similar to Mintzberg’s simple structure with a dominating strategic apex (section 2.2, pp. 9-10). The communication was less formalised and was supposed to be decentralised.

The UniCAP 1000 project is characterised by high interdependencies among the different activities of the project. Due to the authority and good technical knowledge of the former project leader, he took much of the responsibility for the different activities. This made much of the communication centralised. The increase in size of the project made it difficult for the project manager to have both the technical and managerial responsibilities and thus made the organisation structure insufficient.

The new organisation is similar to Mintzberg’s professional bureaucracy with a powerful operating core (section 2.2, pp. 9-10). The System Architect Group formalises the interdependencies among the different subproject activities enabling integration.

Through the System Architect Group they have achieved an amount of formalisation which is necessary for the successful integration of the different project activities. The power of the project manager is weakened in favour of the sub-projects. The project manager of this organisation has less technical knowledge than the former project manager. Technical decisions are thus decentralised to the sub-project mangers. This should give the new project manager time to focus on extra-firm communication requirements such as customer demands.

The different sub-projects were very dependent upon information from each other, which is mirrored by the fully connected communication network. All the studied sub-projects communicated with all the other sub-projects. The capabilities of the communication network were however insufficient. The reason for this seems to be that to little use were made of formal communication channels. Many sub-projects stated that they lacked formal specifications and documentation on e.g. customer requirements. The insufficient formal documentation made the interfaces between different components/modules blurred. This was compensated by intense informal communication, which also handled sub-project planning and scheduling. Verbal informal communication was not perceived to be the best channel to communicate technical specifications, since it is often fragmentary and non-exact.

The communication network of the new organisation is not fully connected. The project evolution rather than lack of communication could explain this. The mechanics sub-project, for instance, does not need to communicate extensively at this point in time with e.g. Pipeting. In most cases the sub-projects have stated the same amount of communication between each other, but discrepancies exists. In the case of process module which tend to communicate less to other sub-projects than the reverse. This indicates that process module have information that is important to other sub-projects.

The interdependencies of the sub-projects are still very high at this level. The capabilities of the communication network have improved and more use is made of the formal communication channels. More attention is paid to project documentation and the structure of the project meetings. Protocols of these meetings are also distributed to the whole project. Some sub-project mangers still perceive the formal communication to be insufficient, while others are concerned about too much paperwork. All interviewed sub-project managers did think that project communication was moving in the right direction.

This chapter will discuss the findings of the study and determine whether the propositions (section 4.1, pp. 17-18) were supported or not. The chapter will also discuss the answers to the research questions stated in the introduction.

Three questions were posed in the first chapter:

• What are the different information needs within a NPD project?
• How is information transmitted and received by the participants within a NPD project?
• How do companies manage the information flows in a NPD project to improve time, cost and quality goals?

The answers to these questions are as follows:

The information needs are dependent upon the uncertainties facing the NPD project. Activities within a NPD project need information that reduces these uncertainties. In the studied NPD projects the different sub-projects needed information on design specifications to allow for co-ordination and integration of project work and on customer requirement to know what features the product should have and also why certain features should be there. The NPD projects also needed information on time constraints and project status, i.e. they needed to know what they should deliver and at what time it should be delivered.

In the studied projects the information was mainly transmitted through informal communication channels. The most common way to communicate formally was through the weekly project meetings. The cases showed that too little formal communication could not be fully compensated with more extensive informal communication.

The case studies show an evolution in the project management. The UniCAP 100 case was managed more like a research project than a development project. This could be explained the Pharmacia and Upjohn Diagnostics origin as a researched based company. The organisation A of the UniCAP 1000 project was based on the experiences of the UniCAP 100 project. They saw that larger, more complex, project needed a more formal organisation structure but also more formal communication.

Many sub-projects had problems in gathering needed information and stated that this was due to that the other sub-projects did not want to give away any information since their tasks were not finished. This shows the problem of whether activities should give away information that later has to be redefined or if they should keep the others waiting. Project management needs to chose policy and make it clear to the project members, since it implies a trade-off between time, cost and quality goals of the project.

In the UniCAP 100 project the activities were, in effect, organised as coupled activities (section 2.3, p. 11) and it seems that they successfully co-ordinated the information flows within the project. In organisation A of the UniCAP 1000 project the activities were organised to be executed in parallel with clearly specified interfaces. Time constraints made that the formal communication and thereby the specifications were neglected. This caused difficulties in coupling the different activities and came to affect development time and indirectly the costs of the project. In organisation B of the UniCAP 1000 project an increased cross-functional coupling of interdependent activities enhanced the co-ordination of information. The increased coupling within and between the sub-projects seems to allow for a more effective communication, which should improve time, cost and quality aspects of the development project.

Another aspect of not receiving needed information is whether the activity can or can not give away information at that specific time, which raises the question of how activities should be sequenced. Whether they should be sequential, parallel or coupled. This, however, demands an exact time planning of project activities. The project had no thorough time planning.

The research questions and the theoretical background generated four propositions (section 4.1 pp. 17-18) these are discussed below.

1. A high degree of uncertainty increases the communication requirements of a project.

A comparison of the UniCAP 100 and 1000 cases shows that the degree of uncertainty facing the project affects the communication requirements. A high degree of uncertainty increases the communication requirements. This proposition is supported across both cases.

The uncertainty not only puts demands on the amount of communication needed, but also on the type or structure (informal/formal) of communication. Highly complex and uncertain development projects demand rigorous formal communication (through project documentation), while smaller and less complex projects can manage almost without any formal communication.

Further, the UniCAP 1000 case implies that too little formal communication increase the demands on the informal communication. The UniCAP 1000 case also implies that both formal and informal channels must treat the same information.

2. Technological uncertainties are resolved through decentralised informal peer consultation and decision making, while market and interdependency uncertainties are handled through cross-functional communication via formal channels.

In the UniCAP 100 case the technological uncertainties were resolved through informal and decentralised peer consultation. The uncertainties concerning the integration of the instrument involved informal decentralised cross-functional consultation. The interdependency uncertainties were resolved through cross-functional communication but not via formal channels. The contacts with the marketing function were planned to go via the project meetings (a formal channel), but due to low attendance questions were often posed directly to the marketing function.

The proposition that technological uncertainties are resolved through decentralised informal peer consultation and decision making finds support in the UniCAP 100 case. The latter part of the proposition was not supported. This could be explained by the informal characteristics of this particular project. The behaviour of the marketing function can also explain the deviation from the expected results

The proposition that technological uncertainties are resolved through decentralised informal peer consultation holds true in both organisations of project activities of the UniCAP 1000 case, but the decision making was not informal and not decentralised in the first organisation. In the new organisation of the UniCAP 1000 project the decision making is more decentralised but it is not informal. The second part of the proposition, that market and interdependency uncertainties are handled through cross-functional communication via formal channels is not supported in the first organisation, but it finds support in the new organisation of project activities. The contacts with the marketing function were planned to be formal in the UniCAP 1000 project, but it has become informal due to the low project meeting attendance of the marketing function.

3. Coupling of interdependent activities enhances effective communication.

In the UniCAP 100 project the project team was divided into four competencies (process, software, electronics and mechanics), but the size of the project team and the way project work were conducted lead to a strong coupling of project activities. It allowed for an efficient transfer of communication. The problem solving communication also benefited from the team working and discussing around a prototype of the instrument. It made the communication more concrete.

During the organisation A of project activities in the UniCAP 1000 case the intended plan was that the subprojects would work on their problems separately, through clear specifications and well-defined interfaces, and then put everything together. When the modules were tested the instrument did not work properly. This is explained by an insufficient coupling of the interdependent activities, but also by the insufficient formal communication. One example to overcome this the ISW sub-projects development of cross-functional functions, informal organisational functions within the project. These informal functions allowed for a greater coupling of interdependent activities.

The new organisation of project activities in the UniCAP 1000 case has several sub-projects that are cross functional. Many interviewees perceived that the communication within the project had improved as a result of organising interdependent activities within a single sub-project. All sub-projects are more closely monitored by the System Architect Group and ‘forced’ into formal information exchange. This was also made to couple interdependent tasks.

The two cases support that coupling of interdependent activities enhances effective communication.

 

 

4. Effective communication enables projects to cut development time, through correct and timely transfer of critical information between development activities.

According to our definition (Chapter 3.3.1, p. 22) effective communication occurs when interdependent activities communicate with each other and when the information transferred is timely and correct.

The communication within the UniCAP 100 project was characterised by a fast transfer and iteration of information between the different sub-projects. The transferred information was instant (when asked). Whether the communicated information was correct or not builds on the assumption that the different sub-projects had well defined deliverables to the other project activities. The communication in the UniCAP 100 case was more an ongoing problem-solving dialogue. The UniCAP 100 project was finished somewhat behind schedule.

In the organisation A of the UniCAP 1000 project the communication between the sub-projects was intense. The communication was seldom timely and the interviewed sub-project leaders stated that the lack of clear specification sometimes made verbally communicated ‘specifications’ incorrect.

In the new, second form of project activities the intra-project communication is intense. The formal communication has improved with the help of the System Architect Group and the Project Planning and Administration staff functions. This allows for a more correct communication, but the communication is not timely.

The proposition finds weak support in the UniCAP 100 case. The picture is less clear in the UniCAP 1000 project. It seems that the reorganisation has improved the quality of communication, but it is hard to conclude whether this has speeded up the development of the instrument or not.

In short we encountered these problems with project communication:

• The communication with the marketing function is insufficient.
• There were problems with product specifications, the formal communication seems insufficient.
• Sub-project members seem not to know what information other sub-projects need.
• All sub-project managers stated that they had to pull information from the other sub-projects rather than getting the required information, which probably caused the high amount of informal communication.

Suggestions:

Following the study of the two development project we have some suggestions regarding project communication.

• Create a close formalised connection, at a low project level, between the marketing function and the project organisation.

• Specifications should be determined early in the project. Changes that affect the features of the product should be clearly communicated through changes in the specifications.

• Clarify the interdependencies between project activities at an early point in time. There are several tools for this such as PERT diagrams (Haynes, 1997, pp. 36-38) or Eppinger’s (1994) Design Structure Matrix (DSM). Gantt charts, which software such as MS Project is based on, is better suited for smaller projects. PERT diagrams are not as detailed as the DSM and give, according to us, a better graphical representation of the time interdependencies (Haynes 1997, p. 32).

Subjects for future research are to study the interactions within and between sub-projects: how project members communicate with each other. Another idea is to study projects at a higher level: how project managers communicate with top-management and the rest of the company and how this affects project success.

Finally we would like to thank Nina Brunk at Pharmacia and Upjohn Diagnostics for helping us setting up the interviews and showing great interest in the thesis. We would also all project members who gave us time for the interviews at a short notice.

Ali, B. 1994. Pioneering Versus Incremental Innovation: Review and Research Propositions. Journal of Product Innovation Management. 11: 46-61

Brown, S. and Eisenhardt, K. M. 1995. Product Development: Past Research, Present Findings, and Future Directions. Academy of Management Review, 20(2): 343-378

Burns, T. and Stalker, G.M. 1961. The Management of Innovation. Tavistock, London, UK

Cooper, R.G. 1994. New products. The Factors that drive Success. International Marketing Review, 11(1): 60-76

Daghfous, A and White, G. R. 1994. Information and Innovation: a Comprehensive Representation. Research Policy, (23)3: 267-280

D’Aveni, R. A., Gunther, R. and Harrigan, K. R. 1995. Hyper-competitive Rivalries. Free Press, New York

Eppinger, S. D., Whitney, D. E., Smith, R. P. and Gebala, D. A. 1994. A Model-Based Method for Organizing Tasks in Product Development. Research in Engineering Design,

6: 1-13

Eriksson, L. T. and Wiedersheim-Paul, F. 1997. Att Utreda Forska och Rapportera. Liber Ekonomi, Malmö, Sweden.

Hamel, G. and Prahalad, C. K. 1997. Att konkurrera för framtiden. ISL Förlag, Göteborg, Sweden.

Haynes, M. E. 1997. Project Management – How to Manage a Project Through the Crucial Stages From Idea to Implementation. Kogan Page, London, UK.

Krishnan, V., Eppinger, S. D. and Whitney, D. E. 1997. A Model-Based Framework to Overlap Product Development Activities. Management Science. April (43)4: 437-451

Mendelson, H. and Pillai, R. R. 1999. Information Age Organizations, Dynamics and Performance. Journal Of Economic Behavior and Organization, (38)3: 253-281

Mintzberg, H. 1983. Structures in Fives: Designing Effective Organizations. Prentice Hall, Englewood Cliffs, NJ.

Moenaert, R. K., Souder, W. E., De Meyer, A., Deschoolmeester, D. 1994. R&D-Marketing Integration Mechanisms, Communication Flows, and Innovation Success. Journal of Product Innovation Management, 11: 31-45

Pinto, M. B. and Pinto, J. K. 1990. Project Team Communication and Cross-Functional Cooperation in New Program Development. Journal of Product Innovation Management, 7: 200-212

Porter, M. E. 1991. Toward a Dynamic Theory of Strategy. Strategic Management Journal, 12:95-117

Revealing the secrets, Allergy, Asthma and Autoimmune Diseases. Pharmacia and Upjohn Diagnostics AB, 1999

Rothwell, R. 1994. Industrial innovation: Success, strategy, trends. In Dodgson, M. and Rothwell, R. The Handbook of Industrial Innovation. Edward Elgar, Hants, U.K

Song, X. M., Souder, Wm. E. and Dyer, B. 1997. A Casual Model of the Impact of Skills, Synergy, and Design Sensitivity on New Product Performance. Journal of Product Innovation Management, March, 14: 88-102

Souder, W. E., Sherman, D. and Davis-Cooper, R. 1998. Environmental Uncertainty, Organzational Integration and New Product Development Effectiveness: a Test of Contingency Theory. Journal of Product Innovation Management, November (16)6: 520-534

Tushman, M. L. 1979. Managing Communication Network in R&D Laboratories in Tushman, M. L., and Moore, W. L. eds. 1988. Readings in the Management of Innovations. 2^nd ed. Harper Collins, New York.

Wiedershem-Paul, F.

Yin. R. K. 1994. Case Study Research – Design and Methods. Sage Publications, Newsbury Park London.

The UniCAP 100 Project:

Sub-project manager of the Process sub-project, 1999-05-23, by telephone

Employee System Functions sub-project, 1999-05-12

Employee System Functions sub-project, 1999-05-11

Employee Instrument Software sub-project, 1999-05-12

The UniCAP 1000 Project:

The former project manager, 1999-04-26

Sub-project manager System Functions sub-project, 1999-05-12

Sub-project manager CAP and Sample Modules, 1999-05-19, by telephone

Sub-project manager Instrument Software, 1999-05-12

The new project manager, 1999-05-12

Manager System Architect Group, 1999-05-20

Sub-project manager Process Development sub-project, 1999-05-26, by e-mail

Sub-project manager Process Module sub-project, 1999-06-05, by e-mail

Sub-project manager Hydraulics sub-project, 1999-05-17

Sub-project manager Electronics & Power, 1999-05-20

Sub-project manager Mechanics sub-project, 1999-05-11

Sub-project manager Instrument Software (ISW), 1999-05-18

Sub-project manager Integration & Verification sub-project, 1999-05-12

Tillbaka