A case for estimating demand for your product.

(Dataset)

 

Forecasting demand for your product is necessary to enable your firm to plan its operations (marketing, finance, and manufacturing). You have been examining the total industry demand (for all firms) since the introduction of this product. This product was introduced 22 quarters ago. Based on this data, your firm has been able to make projections and  (vaguely) determine the demand for this product.

 

Having an estimate of Total Industry Demand (TID) has been valuable for your firm in determining the overall market growth. This information has been used to determine the general direction of investments in promoting the product and improving the quality of the product. However, your firm has never been able to reliably predict the demand for its product (Firm Demand). 

 

Historically, your firm has used predictions of TID and your Market Share (MS) estimate the demand in a future quarter.

 

Firm Demand (FD) = Total Industry Demand (TID)  *  Market Share (MS)

 

For example, if your firm estimated that TID is likely to be 20000 and your MS is likely to be 25% the demand for your product will be around 5000 units. Using this estimate your firm has been able to plan production schedules, manufacturing requirements, project financial revenues, estimate costs, and make decisions on pricing, advertising budget, etc.  Due to the subjective nature of these estimates, your predictions have often been unreliable leading to high inefficiencies in managing your firm’s operations.

 

You are currently exploring the use of regression modeling for obtaining more reliable forecasts. You have explored the availability of data. It is possible for your firm to obtain quarterly data on TID and number of firms in the industry. This historical information is available from a reliable source.  Using this information you can monitor patterns (trends) in the overall demand for the industry.  You are comfortable with the idea that by extending the trend you will have a fairly good estimate of TID. 

 

The more challenging variable for prediction is your firm’s Market Share (MS). This variable is dependent on your firm’s ability to compete effectively. Competition is fairly intense and is based on pricing, promotion, and loyalty.

 

You have been exploring the availability of historical data on variables that may help you determine your firm’s MS. You would like to have data on each firm’s prices, advertising expenditures, market share, etc.  Due to the dynamic nature of your market, it is extremely difficult for you to predict the behavior of each of your competitors.

 

The company that has been providing you with the TID data, also can provide you with historical data for price, advertising, R&D and each firm’s MS.  Using this data you are now able to compute averages of price, promotion and R&D for your industry. Since there is so much uncertainty in predicting each firm’s behavior, you are more comfortable in being able to predict industry averages of price, advertising and R&D.

 

 

 

 

 

 

 

 

Keeping data availability and the limitations of your business intelligence system in mind, you are now exploring the ability to model Relative Demand.

 

Relative Demand (RD) = FD / AFD

 

Where,

FD = Your Firm’s Demand

N = Number of Firms in Industry (for this dataset N=10)

AFD = Industry Average Demand (TID/N)

 

If you are able to estimate RD, you can easily convert it into an estimate of MS.

 

MS = RD / N

 

For example, you RD is 1.2 which means that your demand is 20% higher than the average for the industry. If there are 10 firms in the industry (N=10), then your MS would be 1.2/10 or .12 (12%).  With this information you will be able to estimate the demand for your product.

 

FD = TID * MS  = TID * (RD/N)

 

So, your job is fairly clear. You have to develop models to estimate TID and RD.  The attached dataset shows the data that you have collected and organized for your analyses.

 

Since RD is relative demand, the predictor variables should also be relative to the average for the industry. Hence, you have computed PREL (relative price) and AREL (relative advertising). You will use the previous quarter’s relative demand (RD1) as a measure of Brand Loyalty.

 

PREL  = Your Price / Average Price

AREL = Your Advertising / Average Advertising

RD1 = Last Quarter’s relative Demand

 

 

You are now fairly confident that you can model the demand for your product. You will break this task into two parts: Estimate TID and Estimate RD.


OUTLINE FOR CASE ANALYSIS:

 

 

I. Describe the overall objective of this project.

 

  • Forecast demand for your product (FD)
  • Spell out the overall model you are attempting to develop (FD = TID * RD/N). Define each term.
  • Describe each component of the model and possible predictors. (TID as a function of time, average price, and average advertising; RD as a function of PREL, AREL and RD1).

 

II. Description of Variables

 

TID Variables:

 

  • Using charts and summary measures describe TID, Average Price, and Average Advertising. 
  • Use frequency histograms and line graphs (versus time or Quarter #) to visually describe these variables. 
  • Compute the mean, standard deviation, median, high value and low value for each variable.
  • Interpret the information obtained from these charts and summary measures (jointly) for each variable. 
  • Compute the correlation matrix of TID, AvgPrice, AvgAdv, and Quarter and interpret your results.

 

RD Variables:

 

  • Provide a brief (verbal) description of each variable (RD, PREL, AREL, RD1). How is each computed?  What is each variable a measure of? (for example, PREL captures relative pricing, RD1 is a measure of loyalty, etc)  
  • Using scatter plots describe relationships between each explanatory/predictor variable (PREL, AREL, and RD1) and the dependent variable RD.  Interpret these results – which of your possible predictors seem to be related to the dependent variable.

 

III. Mathematical Modeling

 

TID Model

 

  1. Perform regression analysis using Quarter (measure of time) as the independent variable to estimate TID. Describe the resulting equation (the formula) and its characteristics (R-sq, and p-value of coefficient).
  2. Perform regression analysis using Quarter, AvgPrice, and AvgAdv as predictors. Describe the final model (the formula) and its characteristics (R-sq., p-values of coefficients).
  3. Compare the models estimated above. Which one is better? What are the merits of each?

 

RD Model

 

Using PREL, AREL and RD1 as predictors develop a regression model to predict RD. Describe the final model (the formula) and its characteristics (R-sq., p-values of coefficients).

 

FD Model

 

Describe the final mathematical model to estimate the demand for you product. The model must be complete will all coefficients.