The search for possible relationships between the outsourcing of logistics and logistics costs started from the most aggregate level, i.e. total logistics costs and all companies, and then each cost component and main industry were analysed separately. Table VI [Figure omitted. See Article Image.] shows the total division and the level of logistics costs in the sample (14.6 per cent of total sales). The cost components with the largest shares were transportation (4.2 per cent of sales) and inventory carrying (4.0 per cent). None of the cost means or medians is statistically significantly different between the main industries. The logistics costs were measured as a self-reported proportion of company turnover that, according to [32] Stewart (1995), is a robust base for analysis.
Logistics costs were subjected to explorative factor analysis (Appendix 3). Four of the cost components (transportation, warehousing, inventory carrying, and logistics outsourcing) formed one factor and the other two (transportation packing and indirect costs) formed the other. It thus seems that costs that are more directly linked to the operational activity of the company, and thus more concrete, behave in a similar manner as the two more abstractly defined costs.
Analysis of both main industries together and the total logistics costs gives the overall impression that the relationship between outsourcing and costs slightly resembles an inverse U-shaped curve. Logistics costs measured on the overall level as well as separately in terms of individual components, seem to be slightly lower if the company has made the decision not to outsource logistics activities at all, or to outsource more than half of them. The results follow the same pattern for the outsourcing of both materials management and value-added services, as well as of information processing. Given only single respondent in the "no outsourcers" group in the transportation-services group, the analysis just involved comparing the two remaining groups with each other. These quite systematic and visually identifiable differences turned out to be statistically insignificant ( p >0.05).
Table VII [Figure omitted. See Article Image.] shows the group-wise comparisons of the two cost factors, as well as the total costs of trading companies. It turned out that the "modest outsourcers" faced somewhat higher costs, excluding transport services, than the companies with no outsourcing at all and those with heavy outsourcing. The same applied to most of the comparisons within the manufacturing industry. However, all differences turned out to be statistically insignificant ( p >0.05).
Given the dominant role of transportation and inventory-carrying costs, each of the cost components was subjected to the same analyses separately. However, no statistically significant patterns were observed.
3.4 Logistics outsourcing and absolute logistics performance
The next phase of the analysis was to explore the possible relationships between logistics outsourcing and the level of logistics performance. The first step was to analyse the level of logistics performance with different levels of outsourcing in the two main industries combined, and the next was to compare the differences in performance between the two main industries and finally carry out within industry analyses. All analyses were done by means of ANOVA testing the means and the non-parametric Kruskal-Wallis test testing the medians.
Table VIII [Figure omitted. See Article Image.] shows the median values of logistics performance in relation to the outsourcing of materials management and value-added services. Similar analyses were carried out for all of the outsourcing groups with similar outcomes. There were no identifiable relationships in any of the outsourcing groups regardless of whether the variables were analysed in combination or separately in the two industries. The only statistically significant differences are in days of sales outstanding and in average delivery time between the two main industries. Since the outsourcing behaviours of the two main industries closely resemble one another, it is likely that these differences are caused by the different natures of the two industries, rather than outsourcing.