Analysis and interpretation of data

Data analysis is central in building theory from case studies. In this study the analysis together with the data collection formed an iterative process, moving among the data, the literature and the emerging theory. The within-case analysis of the information obtained through the interviews was guided by the themes of the interview guidelines. The qualitative analysis followed the instructions by Miles and Huberman (1994, 10 - 12) through the phases of data reduction, data display and conclusion drawing. The collected data was managed and gradually condensed into tables and matrices. By focusing on the success dimensions, the data was transformed through selection and summarising to larger patterns. This brought out fresh views, which helped in reaching the conclusions. The literature was consulted again to refine the findings proposed by the field studies. A brief summary of the case comparisons based on the semi-structured interviews is introduced in Appendix 5.

The answers of the structured interviews (questionnaire) have been categorised to factors followed by the introduced framework of the success dimensions. Each factor was given a quality rating in order to describe the status of the factor. The quality rating is based on the mean values and the unanimity in the interviewees’

assessments. The mean value has been counted to each factor based on the interviewee’s assessments (Likert scale 1 – 7), called later as an informant’s mean value. The company mean value is summarised based on the interviewees’

assessments from this case company. The quality rating of the factor is based on the rate of the company mean value and on the unanimity in the informants’

assessments as illustrated in Table 6.5. For example to have the quality rating of very good, the company mean value needs to be more than 6.0 and all the individual informants of that case company need to have the mean value at least as high as 5.0. A summary of the case companies’ and informants’ mean values based on the structured interview (questionnaire) have been introduced in Appendix 6, as well as the quality ratings.

Table 6.5 Coding principles of categorisation

Quality rating Company mean value Every informants mean value

Excellent > 6.5 AND > 6.0

Very good > 6.0 AND > 5.0

Good > 5.0 AND > 4.5

Satisfactory > 4.5 AND > 4.0

Neutral > 4.0 AND > 3.0

Fair > 3.0 AND > 2.0

Poor > 2.0

Very poor < 2.0

This study focuses on the success dimensions and success factors of business development projects implemented in SMEs. For that reason, it is important to clearly figure out the factors perceived as very successful by most of the informants. In order to distinguish these factors from others, the scale of the categorisation of the quality ratings is eight despite the fact that the assessment was done by using the scale from 1 to 7. For the same reason, one of the coding principles is unanimity in the informants’ answers, but that principle is also meant to prevent overrated answers from biasing the results.

The structured interview includes a set of questions based on the Project Implementation Profile (PIP) analysis tool. It is a tool developed by Pinto and Slevin (1987) to help project managers in their everyday work. PIP is a questionnaire booklet expected to be filled in by the project team members. It is based on the ten critical project implementation success factors introduced in Chapter 3.4. The questionnaire includes five statements per each of ten success factors (Pinto &

Prescott 1990, 323 – 325). The person who is answering the questionnaire expresses his agreement or disagreement with each statement using the scale from 1 to 7, where one means strong disagreement and seven means strong agreement.

In the end, the sum of the scores of each factor is calculated separately to form a value for the factor. According to the PIP systematic the value of each factor is compared to a reference table, which is based on a database of 408 studied projects (Slevin 1989, 329). If the value of the factor is below 50 %, the situation is considered as critical and some actions concerning the factor are recommended.

The example value 50 % means that within the reference projects 50 % of the persons have ranked her or his factor higher in their own project.

In the present study PIP was used as a quantitative analysis tool to support the interviews and the analysis. The main advantage of this analysis tool is that it illustrates the status of the strategic and tactical success factors of the projects.

According to Pinto and Slevin (1987), the PIP-tool is practicable in different phases of the project. In this research it was used as a tool to assess the project success afterwards. Some of the statements concerning two critical factors, client consultation and client acceptance, were difficult to answer. Due to the deficient answers, these two factors were dropped from the analysis. Instead, the customer perspective was analysed connected to the project success inside the category of the project’s impact on the customer.

The business performance was viewed on the basis of both objective financial measures and subjective perceptual measures. The data of the objective measures was collected from secondary sources (annual accounts, publicly available from The National Board of Patents and Registration). The data of the subjective measures was collected directly from the case companies. According to Venkatraman and Ramanujam (1986, 806), this approach is appropriate when a broader conceptualisation of the business performance is needed for addressing a specific research question, but the data of the financial performance may not be forthcoming from the primary sources due to reasons of confidentiality and sensitivity. Ward, Leong and Boyer (1994, 345) have observed that the informants appear more reluctant to disclose objective information than perceptual information. The informants of the present study were aware of the fact that the objective data would be collected from independent, secondary sources.

Further, in this kind of situation, Venkatraman and Ramanujam (1986, 811) recommend to address the dimensionality issue both theoretically and empirically.

The model of the project success is modified from the one introduced by Shenhar et al. (2001). It consists of categories important for project performance measurement and business performance measurement. In addition to this, the 12 measures validated by Gupta and Govindarajan (1984, 34) were chosen to belong to the body of the measures reflecting the business performance. The factors of the project efficiency follow the Iron Triangle of the project management introduced by Atkinson (1999), measuring cost, time and quality. The informants were asked to add more measures to the list, which they considered as important for the implemented business development project.

In some cases the content of the subjective measures are defined according to an operational definition, but the measurement is done as perceptual, called as quasi-perceptual measures. According to Ketokivi and Schroeder (2004, 261), the measurement instrument that uses quasi-perceptual data may be less affected as it defines the content of the measurement exactly. In this study, for instance, asking what has been the average sales growth of an individual new product category, defines the content of the sales growth exactly, but leaves the measurement to the informant’s discretion, as well as the option not to answer truthfully.

The number of informants was three per case company. During the coding phase one of the coding principles was unanimity of the informants’ answers. Further, the informants’ answers of the business performance were based on perceptual and quasi-perceptual measures, and the answers were compared with objective financial data. The researcher considered the perceived project performance as plausible if all the three informants gave similar answers and the objective data supported this united answer.

Finally, the qualitative and the quantitative analyses were documented as a process. Feedback from the informants was used. The feedback was acquired at least from one key informant from each case company when the first drafts were available. The details about the informant feedback are introduced in Appendix 2.

In this study both qualitative and quantitative evidence has been used, forming a dynamic balance for the analysing process (cf. Dey 1998; Eisenhardt 1989; Stake 1995). The researcher’s role has been an independent outsider.