Brand positioning is frequently facilitated by the use of perceptual maps. Several approaches exist for deriving such maps. This research uses the variability inherent in customer data to build confidence regions around brands and attributes in perceptual maps. Doing so generalizes the typical descriptive approach to a truer, statistical inferential approach to mapping. The resulting visualizations clarify the interpretations regarding which brands are similar, with overlapping confidence regions, and which brands are distinct, given non-overlapping confidence ellipses. The modeling is first demonstrated on a small, synthetic dataset and then on real consumer data. The model extension is shown to be useful, and it is relatively straightforward in implementation. It is hoped that this extension to this frequently used market mapping approach should enhance interpretive precision, and therefore, lead to more accurate and successful strategic positioning decisions.