The two-stage, hypothesis-based GEXSO study intends to provide praxis-related recommendations for decisions with global and international focus. Both studies cover expert interviews to identify relevant influencing factors and hypotheses regarding their correlation. Subsequently, influencing factors and cause-effect relationships are examined in a quantitative and questionnaire-based primary data study in companies within the machinery and plant engineering sector and the automotive supplier industry. The survey took place from September 2011 until June 2012.
Overall, experts from 45 companies were interviewed on key issues of the GEXSO study. 43 of the 45 interviews were taken into account during the evaluation. All participants of the study have long-standing experience in supply chain positions within their respective industry sectors, e.g. as supply chain managers, purchasing managers, logistics managers etc. In cases where the expert position did not already imply an international focus, we posed international questions regarding supply chain management within the scope of projects. In order to maximize the knowledge gained, and not to influence the assessments of the experts, we used an exploratory approach (supported by an interview guideline) with open questions. Afterwards, documented results of the interviews were presented to the participants of the study for review and approval. Then, we coded the answers and summarized them with the help of a cluster analysis.
For a better evaluation of suspected relations by GEXSO, the questionnaire was presented to experts from different companies within the machinery and plant engineering sector, and the automotive supplier industry. In order to determine suitable and competent participants, contact persons were researched thoroughly and then contacted individually over the phone or online. Sources of information for the research were customer data, membership lists of unions and associations, profile pages on the internet, and articles and publications on the topic supply chain management. The questionnaire was distributed via several channels of communication (online, as electronic data, via post). Furthermore, we offered structured personal and phone interviews to the participants. Approximately 80 per cent of the experts were interviewed personally or over the phone. With a response rate of 12 per cent, we compiled 45 fully completed questionnaires for evaluation. Unfortunately, we were not able to include two companies in the final evaluation due to their business model. In addition to previously determined influence factors we also inquired about company characteristics, key figures, and strategies for success. Over all, we identified 110 indicators and almost 55,000 data points.
Methods of the study
In addition to a qualitative cluster analysis of the expert interviews, the results of the questionnaire were analyzed employing quantitative, statistical methods. For measurement models, we exclusively used reflexive indicators that could be categorized into influencing factors, and only incorporated indicators with very large factor loading values (larger than 0.6) were used. A very large number of indicators had factor loading values larger than 0.8. Therefore, we were able to measure influence factors previously determined in the expert interviews with a high degree of clarity. In order to scale appropriate indicators, we generally used a 6-level Likert scale: neutral values between ‘not applicable’ (=1) and ‘fully applicable’ (=6) were excluded by providing only an even number of options. Although we were not able to test expert interviews in a holistic causal model, we examined them employing a correlation analysis. For evaluation of the moderating factors such as complexity and internationalization, and other company characteristics, we used nominal, interval and absolute scales. These results are used to provide company clusters and typifications. We also utilized them for the index calculation of the globalization, complexity and the GEXSO index. The applied models are based on existing models that are recognized in literature and already applied in other contexts. For instance, the term ‘globalization degree’ is based on a model by Christopher Bartlett and Sumantra Goshal and measured with the aid of specific company characteristics. For evaluation of supply chain risks, the questionnaire also contained a short explorative section with one open question concerning the top five risks. Analogous to the evaluation process of the expert interviews, answers were coded and classified into supply risks, process risks, management risks, demand risks and environmental risks according to a risk scheme that is based on the cause model used by the BVL and the SCOR model. Risks were rated, i.e. the first risk listed was rated fivefold, the second fourfold, etc.