New metrics for evaluating preference maps

Preference maps provide a visual representation of market structure, usually depicting brand or product alternatives, product attributes, and customers in a single graphic. Using measures of consideration and attribute sets to establish criterion validity, we develop a set of metrics that can be interpreted by managers and that allow managers to evaluate maps based on their ability to accurately represent market structures for products, attributes, and consumers. Using a Monte Carlo simulation, we test the stability of the metrics for a variety of scenarios and compare them to statistical stress. Our results show that the metrics can help identify specific sources of noise and can therefore be used to interpret map fit at a more disaggregated level than stress. We apply the metrics on an empirical example and use them to develop a reweighted map for a focal product.