Forecasting

Product/Operations Management

Chapter 4.0 – Forecasting

Forecast – A statement about the future.

Forecasting – the science of predicting states of the future.

Two uses of a forecast

1.      To help managers plan the system –planning the system involves the long-range plans about the types of products and services to offer. What facilities and equipment to have, where to locate, etc.

2.       To help managers plan the use of the system – this refers to short range and intermediate range, which involves tasks such as planning inventory & work force level, planning purchasing and production, budgeting, scheduling.

Forecasting pertains to more than predicting demand. Forecasts are also used to predict profits, revenue costs, productivity changes, prices, and availability of energy, and raw materials, interest rates, movements of key economic indicators (GNP, inflation, government borrowings) and prices of stocks and bonds.

In spite of use of computers and other sophisticated mathematical models, forecasting is not an exact science. No single technique works at all times.

Generally, the responsibility for preparing forecasts lies with marketing and sales rather than operations. But, operations people are often called on to make forecasts and help others prepare the forecasts.

Features common to all forecasts

1.      Forecasting techniques generally assumes that the same underlying causal system that existed in the past will continue to exist in the future.

2.      Forecasts are rarely perfect. Actual results usually differ from predicted values. No one can precisely predict the values. Allowances should be made for inaccuracies.

3.      Forecasts for group of items tend to be more accurate than for individual items because forecasting errors among items in a group usually have a cancelling effect or offsetting effect.

4.      Forecast accuracy decreases as the time period covered by the forecast (time horizon) increases. Short- range forecasts must contend with fewer uncertainties than larger- range forecasts, so they tend to be more accurate. 

Steps in the Forecasting Process

1.      Determine the purpose

2.      Establish a time horizon

3.      Select a forecasting technique

4.      Gather and analyze the appropriate data

5.      Prepare the forecast

6.      Monitor the forecast   

Two General Approaches to Forecasting:

1.       Qualitative – It consists mainly of subjective inputs, often defy precise numerical descriptions. These have personal biases and include soft informations based on personal opinions, hunches, and other human factors.  These are judgmental forecasts.

A.      Collective Opinion Method – Collecting opinions of individuals in a organization regarding the future demand for a given product. These persons are assumed to possess knowledge of future behavior and may Staff of Marketing and Sales Department. Sometimes, outside consultants are also asked.

B.      Delphi Technique -  This calls for a questionnaire being prepared and distributed to the participants for their initial inputs and after which, the filled-up questionnaires are returned to an organized committee which then complies the report

Then the committee develops and distributes a second questionnaire which summarizes and highlights the results of the initial questionnaire. The process is repeated until a consensus is achieved.

2.       Quantitative – This involves either the extension of historical (time series) data or the development of associative models. This method generally assumes that the same underlying causal system that existed in the past will continue to exist in the future. Causal demand patterns can be classified as either of the following demand patterns:

a.       Trend – Long term movement in data ( population shifts, income & cultural changes)

b.      Seasonal – Short term regular variations related to factors such as weather, holidays, vacations.

c.        Cycle – wavelike variation lasting more than one year. Often related to a variety of economic, political and even agricultural situations.

d.      Irregular Variations – caused by unusual circumstances not relative of typical variations. Such as weather conditions, strikes, or a major change in a product/service.

e.      Constant Demand – more or less permanent circumstances. 

f.        Random Variations – residual variations that remain after all other behaviors have been accounted for.

Time Series Analysis

Averaging – A technique to smooth out some of the fluctuations in a time series because  

 individual highs and lows in the data offset each other when they are combined into an      average.

Three Averaging Techniques:

1.       Naive Forecasts – The simplest forecasting technique. It is the forecast for any period equals the previous period’s actual value.  For example, if demand forecast last week was 50 units, the forecast for upcoming week is 50 units.

2.       Moving Averages – A technique that averages a number of recent actual values, updated as new values become available. This technique uses a number of the most recent actual data in generating a forecast. It can be computed using the following equation:

                                                                                            ______

                                                                                                   n

                       where:
                                   i = ‘age’of the data ( i = 1,2,3…)

                                  n =  number of periods in the moving average

                   Ai = Actual value with age i

                  MA = Forecast

Example 1 Compute a three-period moving average forecast given demand for shopping carts for the last five periods.

Period	Age	Demand
1 2 3 4 5	5 4 3 2 1	42 40 43 40 41

Solution:

                   MA3 = 43 + 40 + 41          =  41.33

                                      3

If actual demand in period 6 turns out to be 39, the moving average for period 7 would be:

                   MA3 = 40 + 41 + 39           = 40.00

                                       3

Note that in a moving average, as each new actual value becomes available, the forecast is updated by adding the newest value and dropping the oldest and then re computing the average. Consequently, the forecast “moves” by reflecting only the most recent values.

Weighted Average – is similar to a moving average, except that it assigns more weight to the most recent values in a time series.

 For instance, the most recent value may be assigned a weight of 0.40, the next most recent 0.30, the next 0.20 & the next of 0.10. Note that the weights sum to 1.00 and the heaviest weights are assigned to the most recent values.

Example  2:  

Period	Age	Demand
1 2 3 4 5	5 4 3 2 1	42 40 43 40 41

a.       Compute a weighted average forecast using weights of .40 for the most recent value; .30 for the next most recent; .20 for the next; .10 for the next.

b.      If actual demand for period 6 is 39; forecast the demand for period 7 using the same weights as in part a.

Solution:  a) Forecast = .40 (41) + .30 (40) + .20 (43) + .10 (40) = 41.0

                  b) Forecast = .40 (39) + .30 (41) + .20 (40) + .10 (43) = 40.2

Note: If four weights are used, only the four most recent values are assigned weights.

3.       Exponential Smoothing – weighted averaging method based on previous forecast plus a percentage of the cast forecast error. This is one of the most widely used techniques in forecasting.

α – smoothing constant

 Error ( Actual – Forecast)

Example 3: Suppose the previous forecast was 42 units, actual demand was 40 units and α = 10; the new forecast is:

        Ft = 42 + .10 (40-42) = 41.80

Then, if the actual demand turns out to be 43, the next forecast would be:

        Ft = 41.80 + .10 (43 -41.80) = 41.92

A number of different approaches can be used to obtain a starting forecast, such as the average of the first several periods; a subjective estimate; or the first actual value as the forecast for period 2 (i.e. the naïve approach. For simplicity, the naïve approach is used in this module.

Example 4: Use exponential smoothing to develop a series of forecasts for the following data, and compute (Actual-Forecast) = Error, for each period.

a.       Use a smoothing factor of 0.10

b.      Use a smoothing factor of 0.40

c.       Plot the actual data and both sets for forecasts on a single graph.

Period (t)	Actual Demand
1	42
2	40
3	43
4	40
5	41
6	39
7	46
8	44
9	45
10	38
11	40
12

Solution:

A & B 

                                                                             A                                                  B                                                                            

Period (t)	Actual Demand	α  = .10 Forecast    Error	α  = .40 Forecast     Error
1	42	-                       -	-                    -
2	40	42               -2	42                -2
3	43	41.8             1.2	41.2              1.8
4	40	41.92         -1.92	41.92          -1.92
5	41	41.73         -0.73	41.15          -0.15
6	39	41.66         -2.66	41.09          -2.09
7	46	41.39          4.61	40.25            5.75
8	44	41.85          2.15	42.55            1.45
9	45	42.07          2.93	43.13            1.87
10	38	42.35         -4.35	43.88           -5.88
11	40	41.92         -1.92	41.53           -1.53
12		41.73	40.92

Solution: C

DEMAND

                                                                 PERIOD