Question

Notex Manufacturing makes various batteries used in mobile devices. The company has a major customer so batteries are shipped in bulk to this customer. The company also distributes these batteries to retail stores as replacement parts. The batteries are packaged individually to retail stores. In all, the company makes about 15 different batteries. Currently, the company does not use any forecasting to predict the demand for the batteries. Instead, it has employed rule of thumb to decide about the volume of the production. This has caused some issues for the company including stock-out for some types of batteries or overstocking for some others. The other problem is an increase in the price of raw materials though the company believes it is a temporary condition. Due to complaints from suppliers and also customers the company has decided to introduce a systematic approach toward forecasting.

Therefore, the company has decided to forecast two most important products. The following table (see next page) shows the data on product demand for the two products from order records for the previous 19 months.

Question

Which forecasting method/s do you suggest for the two products? Briefly explain why? Forecast for the next month for each product. (Use MAD for measuring error).

Month	Jan	Feb	Mar	Apr	May	Jun	Jul	Aug	Sep	Oct	Nov	Dec	Jan	Feb	Mar	Apr	May	Jun	Jul
Product 1	33	37	38	40	42	47	43	49	51	55	62	68	69	70	65	55	50	51	49
Product 2	33	34	35	35	36	36	37	39	81*	38	40	38	39	40	39	40	41	42	41

Answer 1

Since you have the data available for the past 19 months, I would suggest a linear regression method for calculating the forecast for the next month for both the products. This method considers previous trends, season wise usage, and irregularity in the data while forecasting for the future.

While using this method I am able to identify the demand for next month as per the following.

Product 1: 63

Product 2: 43