Question

You have just been hired by WidgetCo; a long trusted manufacturer of Widgets. Unfortunately, the company has been suffering from very high inventory levels as compared to others in the highly competitive Widget industry. Your boss thinks there might be a problem in the forecasting and wants you to figure it out.

You have selected a representative family of items and asked for the complete 2 years history of forecasts and actuals. Sadly, they only have the last twelve months, as shown in the table below.

Year	Month	Actual demand (units)	Forecast (units)
2019	March	2938	2913
2019	April	3120	3208
2019	May	3234	3377
2019	June	3417	3514
2019	July	3485	3665
2019	August	3724	3813
2019	September	3827	3906
2019	October	3786	4040
2019	November	4024	4133
2019	December	4096	4219
2020	January	4170	4392
2020	February	4476	4447

You ask to meet with the current demand planner for this family of items, and he tells you that they do next period forecasting – that is, the forecast for June is made at the end of May. He uses a forecasting method of his own design.

The forecast for March and April are far from what you have expected. In this sense, you want to use a more advanced forecasting method: linear regression. Using this method, what is your estimate for the demand in March and April 2020?

Answer 1

y^ = 2863.1667 + 127.4231 x

for march and April 2020, t = 13 and 14

t	y^
13	4519.667
14	4647.09