All organizations operate in an atmosphere of uncertainty but decisions must be made today that affect the future of the organization.
Demand
Forecasting
All
organizations operate in an atmosphere of uncertainty but decisions must be
made today that affect the future of the organization. There are various ways
of making forecasts that rely on logical methods of manipulating the data that
have been generated by historical events. A forecast is a prediction or
estimation of a future situation, under given conditions. Demand forecast will
help the manager to take the following decisions effectively.
The steps
to be followed:
1. Identification of objectives
2. Nature of product and market
3. Determinants of demand
4. Analysis of factors
5. Choice of technology
6. Testing the accuracy
Criteria
to choose a method of forecasting are:
1. Accuracy
2. Plausibility
3. Durability
4. Flexibility
5. Availability
The following are needed for
demand forecasting:
1. Appropriate production scheduling
2. Suitable purchase policy
3. Appropriate price policy
4. Setting realistic sales targets for salesmen
5. Forecasting financial requirements
6. Business planning
7. Financial planning
8. Planning man-power requirements
To select the appropriate forecasting technique, the manager/forecaster
must be able to accomplish the following:
1. Define
the nature of the forecasting problem
2. Explain
the nature of the data under investigation
3.
Describe the capabilities and
limitations of potentially useful forecasting techniques.
4.
Develop some predetermined
criteria on which the selection decision can be made. Demand Forecasting Methods:
1. Survey of
buyers’ intension 2. Delphi
method 3. Expert
opinion 4. Collective
opinion 5. Naïve
model 6. Smoothing
techniques 7. Time
series / trend projection 8. Controlled
experiments 9. Judgmental
approachTime Series / Trend Projection
The linear trend is the most
commonly used method of time series analysis. The following are various trend
projections used under various circumstances.
Linear Trend Equation:
Y = a + b
X Y =
demand X = time
period a,b constant values representing intercept and slope of the line. To
calculate Y for any value of X we have to solve the following equations,(i)
and (ii). We can derive the values
of ‘a’ and ‘b’ through solving these equations and by substituting the same in
the above given linear trend equation we can forecast demand for ‘X’ time
period.
Estimate the sales for 2012, 2015 and fit a linear regression equation
and draw a trend line.
Techniques that should be used
when forecasting stationary series
(the demand patterns influencing the series are relatively stable) include
naïve method, simple average method, moving average, and autoregressive moving
average (ARMA) and Box-Jenkins method. When forecasting trend series then, moving averages,
simple regression, growth curves, exponential models and autoregressive
integrated moving average (ARIMA) models and Box-Jenkins methods can be used. For seasonal series census X-12, winter’s exponential smoothing,
multiple regression and ARIMA models can be used. When forecasting cyclical series econometric models,
economic indicators, multiple regression and ARIMA models can be used. The major forecasting techniques
are: naïve, simple average, moving averages, exponential smoothing, linear
exponential smoothing, quadratic exponential smoothing, seasonal exponential
smoothing, adaptive filtering, simple regression, multiple regression,
classical decomposition, exponential trend models, S-curve fitting, Compertz
models, growth curves, census X-12, Box-Jenkins, leading indicators,
econometric models and time series multiple regression may be used. The causal forecasting models
(simple, multiple regression analysis) will be useful to decide the production,
personnel hiring, and facility planning in the short run. In Time series
forecasting models like decomposition is suitable to decide the new plant,
equipment planning. Moving average and exponential smoothing is used for
operations such as inventory, scheduling and pricing decisions. The
autoregressive models, Box-Jenkins techniques are used to forecast price,
inventory, production, stock and sales related decisions. Neural network method
is for forecasting applications in development phase of the organization. Apart from the above mentioned statistical methods
the survey methods are also commonly
used. They are: 1. Complete Enumeration Method: the
survey covers all the potential
consumers in the market and an interview is conducted to find out the probable
demand. The sum of all gives the total demand for the industry. If the number
of customers is too many this method cannot be used.
2. Sample Survey Method: the
complete enumeration is not possible
always. The forecaster can go in for sample survey method. In this method, only
few (a sample) customers are selected from the total and interviewed and then
the average demand is estimated. 3. Expert’s Opinion: the experienced people from the
same field or from marketing agents
can also be taken into consideration for collecting information about the
future demand. The above discussed qualitative and quantitative
methods are commonly used to forecast the future demand and based on this
information firms will take production decision.
Tags : Managerial Economics - Demand Analysis
Last 30 days 1345 views