Statistics Case Study

OUTLINE

Case Background

Lenovo is a US$34 billion personal technology company and the world's largest PC vendor. They have more than 33,000 employees in more than 60 countries serving customers in more than 160 countries. A global Fortune 500 company, Lenovo has headquarters is in Beijing, China and Morrisville, North Carolina, U.S.

Goal

You have been asked by CFO of Lenovo to develop the most accurate method for forecasting Lenovo sales on quarterly basis. You have been provided with sales figures of the past 6 years for each quarter that are publicly available on investor relations page of Lenovo website.

Body of the Analysis

1. Plot time series and comment on its pattern.

2. Develop time series decomposition. Start by using four quarters moving average and centered moving average to develop indexes for each quarter. Then deseasonalize revenue and develop linear and quadratic estimated regression equations.

3. Forecast sales for all past years and also the next four quarters using both regression equations

4. Plot forecasted sales using both methods with the actual sales on a graph and comment on it.

5. Calculate mean square error of linear and quadratic regression models and comment on the results.

Alternative Solutions

1. Instead of decomposing time series, use actual sales that are not seasonally adjusted and develop additional data using dummy variables to reflect seasonability of each quarter.

2. Develop estimated quadratic regression equation with dummy variables.

3. Use quadratic regression model equation with dummy variables to forecast past sales and sales for the next four quarters.

4. Plot forecasted sales on the same graph and compare it with the actual sales and results from time decomposition methods.

5. Calculate mean square error using dummy variables method and compare results with MSE of time decomposition models. Comment on the results.

Implementation plan

Develop plan for Lenovo to implement new forecasting method that you chose as the most accurate and discuss implementation through the entire corporation and its business units.

Recommendations

Recommend the most accurate method of forecasting for Lenovo based on your analysis. Comment on benefits but also mention shortcomings.

David Borovka | Milcent Taruwinga BUSN5011 Case Study – Forecasting for Lenovo