Supplier search in industrial clusters: Sheffield metal working in the 1990s

Industrial clusters can be characterised by high levels of personal interaction between the owners/managers of firms. It has been argued that within industrial clusters community and firm tend to merge. One result of the notion of a cluster as a community is that it might be expected that personal interaction between members of the community will have an important influence on intra-cluster trading patterns. Whereas there appears to be a number of anecdotal stories of the impact of the 'personal' upon the 'economic' within clusters we do not know whether such relationships are uniquely part of cluster behaviour or whether they are found more widely within the industrial system. The paper reports the result of a new interview survey of seventy small metal working firms in the Sheffield metal working cluster in the UK. Although dealing with a traditional industrial sector the analysis is focused upon contemporary business patterns. It explores the ways in which owner/managers of small metal working firms search for new suppliers. In the empirical analysis the search process is conceptualised as being characterised by two stages: the identification of potential new suppliers and the selection of a specific new supplier. The research is in undertaken in two parts. It first measures the role of the personal networks of the owner/managers of small firms in the identification and selection of suppliers. Second, the research examines whether personal factors are more important in the identification and selection of within cluster suppliers than in the identification and selection of suppliers based outside the cluster. It is shown that, overall, personal networks are of major significance in the identification of suppliers and that information received from third parties are more important than direct contacts between the owner/manager and the potential supplier. However, in the selection decision, price and availability are dominant considerations and personal factors such as trust and reputation of only minor significance. It was not possible to identify a cluster effect in disaggregation of the data to separate out the relationships with cluster suppliers and relationships with suppliers based outside the cluster. There was no evidence that personal factors play a more important role in the establishment of within cluster links. In sum, personal networks are important in the identification of within cluster links but they are equally important in the establishment of links outside the cluster. This suggests that the importance of personal interaction within clusters has been overplayed.