The use of modeling and statistics for the design and development of pricing strategy is prevalent in academia as well as the industry. One of the more commonly used tools by researchers and managers alike for the estimation of linear demand models is the ordinary least squares (OLS) regression. Unfortunately, a majority of data sets to which such models are applied suffer from nonstationarity ‐ that is, the dependence of a variable on its prior values ‐ thereby violating the assumptions of a basic (naïve) regression model. Estimates of variables under these conditions are known commonly to be inflated and inaccurate. While this problem is well‐known and can be corrected for among statisticians and econometricians, a simple and effective tool has not yet been designed for managers ‐ the actual users of such models. Studies some of the problems encountered when using a naïve model and proposes a simple method to check for nonstationarity and redesign the model to account for the same. Using scanner data on soup, shows that the redesigned model predicts better, fits better and offers more meaningful results. Finally, looks at the implications of estimating such models for pricing strategies and issues. Surface response analysis shows how a manager can use such models for conducting insightful studies on price sensitivity.