The six components represent clusters of ﬁrms with similar distributions of service values across cities. Thus far, I have focused upon component loadings because they relate components to ﬁrms, as just illustrated. However, loadings are not the only results that are derived from a principal components analysis. Just as each ﬁrm (variable) displays a pattern of service values across cities, so also do the clusters of ﬁrms that are components (composite variable). In general, this can be thought of as patterns of ‘composite service values’ across cities, and these are computed as component scores. The latter are derived for each city across all components from the service values of ﬁrms and their loadings on each component.