The new structured covariance probit demand model helps overcome challenges in predicting consumer choices by allowing the covariance to change based on market characteristics. This model requires fewer parameters and is better at predicting consumer behavior than the traditional logit model. The researchers used a Bayesian estimation approach and incorporated a flexible method to model demand heterogeneity. Empirical analysis showed that the new probit model accurately captures realistic consumer switching behavior, while the logit model is too restrictive.