Multivariate methods often serve as an intelligent way to study the relatio
ns between two data sets. When the number of variables in one or both data
sets is large, which is usually the case, the correlation matrices of the d
ata sets may be singular or ill-conditioned. When this happens the weights
obtained by multivariate methods that require the inversion of the correlat
ion matrices are not unique, or highly unreliable. Here we present and appl
y a robust estimation and forecasting method that does not require us to in
vert the correlation matrices. This method, which we call robust canonical
analysis (RCA), is a straightforward extension of the simple covariance of
two variables to two data sets. As an example we use the RCA method to esti
mate the relations between a set of measures that describe how the firm man
ages its relations with its customers, and a set of variables that describe
the utility of information systems applications to the firm's operations.