duction Smartfood, a food manufacturing company is releasing a new product in the diet category, K-Pack, which is a low-carbohydrate candy bar. Most snack products have a relatively shorth life cycle of 2 years, therefore the company wants to seize the market and launch the product as soon as possible. The marketing director has conducted a pre-test market study that predicted that the annual sales of K-Pack are 750,000 cases, using 50 cent packages of 24 in a case, with variable and fixed costs totaling 1.75 million, 3 million in advertising costs, equaling 1.55 million in profit for Smartfood. To have a better prediction a linear regression model was developed to optimize the correct marketing mix to maximize profits. Regression Model The authors of this case study use linear regression techniques to predict sales, the response variable, based on potential independent variables such as Month, Price, Advertising, Location, Store, and City. They used the p-value approach and selected the model from the full model backwards. To do this, a scatter plot matrix was used to determine any correlation between any independent variables. Using this technique, it was determined that Advertising and City were perfectly correlated, and they ultimately decided to drop city from the independent variables of consideration, because advertising correlated stronger with the dependent variable, or response variable. It was also determined that there may have been some multicollinearity