Päivittäistavarakaupan dynamiikka ja ostoskäyttäytyminen pääkaupunkiseudulla

The purpose of the study is to describe the structural development of the convenience good trade in Helsinki metropolitan area during years 1990 -1999. The second aim is to locate the housing areas of diversified households based on differentiated shopping behaviour and to develop a dynamic shopping model based on accurate geographic information. The data of the model were collected and the parameters were defined in 1990 according to 664 convenience good outlets and 897 interviewed households. A strong shift towards bigger retail outlet units took place during the research period. The number of over 1000 m2 units grew almost 2,5 times and the outlets concentrated in big (over 1500 m2) shopping centers, which doubled their market share during the period. The total number of retail units concentrated in big shopping centers by 27 %. In shopping centers with floor area less than 1500 m2 the total number fell to half but the size of single units grew by 16 % in both shopping center types. The number of big shopping centers doubled as well and they were distributed more evenly over the research area according to the purchasing power and accessibility. Two thirds of the interviewed households had to find a new outlet because of shutdown. The length of shopping trips increased in general and the establishing of new outlets provided a shortened trip only for few households. The theoretical total length of trips to big shopping centers more than doubled, but the length of average single trip grew by 8 %. The larger the territorial unit the less there are differences between households. By means of principal component analysis it was possible to obtain 8 different types of areally differentiated household categories based on grid data. The grid data is suitable for targeted areal marketing, but it was not possible to define distinct area profiles based on all principal component score categories. The spheres of influence of retail units were defined by potential models, cost-distance models and empirical evidence. The dynamic shopping behaviour model gives patronage estimates to every single outlet and a distance dependent coefficient to shop to every customer. The floor area and the attraction coefficient together can explain 88 % of the total variance of the turnover. By identifying the outlets of different business chains it is still possible to increase the explained variance noticeably. The variables of empirical data could only give hints about the differentiated shopping behaviour of sample households