CoffeeTime Recommendations
Team Paper
University of Phoenix
MBA 510
MEMORANDUM
TO: Whom it may concern
FROM: Business Development (International Operations)
DATE: November 2007
SUBJECT: CoffeeTime
Multiple Regression Model
Laura Jones, CoffeeTime’s statistical expert, built a multiple regression model based on CoffeeTime’s advertising expenditures and price index. Multiple regression analysis is often used to relate a dependent variable (e.g., CoffeeTime’s weekly revenue) with several independent variables (e.g., CoffeeTime’s and Quick Brew’s advertising expenditures and CoffeeTime’s price index) (Lind, et al., 2005).
CoffeeTime has hired a media marketing firm to acquire Quick Brew’s advertising expenditures each week (UOP, 2007). CoffeeTime also monitors its price index which indicates the customer’s effective price (UOP, 2007). The price index is affected by inflation, seasonal coffee availability, and in-store promotions. Inflation and availability increase the index, while promotions decrease the index (UOP, 2007).
Based on the selection of all normal values, Laura computed the multiple R to be 0.738 and the R-square to be 0.546. When she used lagged values, she computed the multiple R to be 0.755 (slightly higher) and the R-square to be 0.570 (again, slightly higher). This shows that the lag values of the independent variables had more impact on weekly revenues than the current values. This seems logical since advertising results are not usually immediate.
To further optimize this model, Laura could have included the normal values of the current and lagged values for all three independent variables (CoffeeTime’s week advertising expenditures, CoffeeTime’s price index, and Quic ...