# Methods of Demand Forecasting – Business Economics – BBA

Aug 30, 2021

## Methods Of Demand Forecasting (For Established Products)

Broadly speaking, there are two approaches to demand forecasting- one is to obtain information about the likely purchase behavior of the buyer through collecting expert’s opinion or by conducting interviews with consumers, the other is to use past experience as a guide through a set of statistical techniques.

Both these methods rely on varying degrees of judgment. The first method is usually found suitable for short-term forecasting, the latter for long-term forecasting. There are several methods to predict the future demand. All methods can be broadly classified into two:

(A) Survey methods, (B) Statistical methods

### (A) Survey methods

The following techniques are used to conduct the survey of consumers and experts:

#### 1. Complete Enumeration Method :

Under this, the forecaster undertakes a complete survey of all consumers whose demand he intends to forecast. Once this information is collected, the sales forecasts are obtained by simply adding the probable demands of all consumers.

In this method, almost all potential users of the product are contacted and are asked about their future plan of purchasing the product in question. The quantities indicated by the consumers are added together to obtain the probable demand for the product.

For example, if only n out of m number of households in a city report the quantity (d) they are willing to purchase of a commodity, then total probable demand (D) may be calculated as

Dp = d1 + d2 + d3 + …. Dn …………. (1)

where d1, d2, d3 etc. denote demand by the individual households 1, 2, 3 etc.

#### 2. Sample Survey Method :

Under this method, only a few potential consumers and users selected from the relevant market through a sampling method are surveyed. Method of survey may be direct interview or mailed questionnaire to the sample consumers.

On the basis of the information obtained, the probable demand may be estimated. Compared to the former survey, this method is simpler, less costly, and less time-consuming but the choice of sample is very critical.

If the sample is properly chosen, then it will yield dependable results; otherwise there may be sampling error. The sampling error can decrease with every increase in sample size.

This method is generally used to estimate short-term demand from business firms, government departments and agencies, and also by the households who plan their future purchase.

Sample survey method is widely used to forecast demand. This method, however, has some limitations. The forecaster therefore should not attribute reliability to the forecast more than warranted.

Besides, sample survey method can be used to verify the demand forecast made by using quantitative or statistical methods. Although some authors suggest that this method should be used to supplement the quantitative method for forecasting rather than to replace it, this method can be gainfully used where market area is localized.

#### 3. Market Studies and Experiments :

An alternative method of collecting necessary information regarding demand is to carry out market studies and experiments in consumer’s behaviour under actual, though controlled, market conditions. This method is known in common parlance as market experiment method.

Under this method, firms first select some areas of the representative markets – three or four cities having similar features, viz., population, income levels, cultural and social background, occupational distribution, choices and preferences of consumers.

Then, they carry out market experiments by changing prices, advertisement expenditure, and other controllable variables in the demand function under the assumption that other things remain the same. The controlled variables may be changed over time either simultaneously in all the markets or in the selected markets.

After such changes are introduced in the market, the consequent changes in the demand over a period of time (a week, a fortnight, or month) are recorded. On the basis of data collected, elasticity coefficients are computed. These coefficients are then used along, with the variables of demand function to assess the demand for the product.

#### 4. Collective Opinion Method/Sales force Opinion Method / Sales Force Polling method:

Salespersons are in direct contact with the customers. Sales persons are asked about estimated sales targets in their respective sales territories in a given period of time.

Merits

• Cost effective as no additional cost is incurred on collection of data.
• Estimated figures are more reliable, as they are based on the notions of salespersons in direct contact with their customers.

Demerits

• Results may be conditioned by the bias of optimism (or pessimism) of salespersons.
• Salespersons may be unaware of the economic environment of the business and may make wrong estimates.
• This method is ideal for short term and not for long term forecasting

#### 5. Experts’ Opinion Method

i) Group Discussion: (developed by Osborn in 1953) Decisions maybe taken with the help of brainstorming sessions or by structured discussions.

ii) Delphi Technique: developed by the Rand Corporation at the beginning of the Cold War in 1950 by Olaf Helmer, Dalkey and Gordon to forecast impact of technology on warfare.

• Way of getting repeated opinion of experts without their face to face interaction.
• Consolidated opinion of experts is sent for revised views till conclusions converge on a point.

Merits

• Decisions are enriched with the experience of competent experts.
• Firm need not spend time, resources in collection of data by survey.
• Very useful when product is absolutely new to all the markets

Demerits

• Experts’ may involve some amount of bias.
• With external experts, risk of loss of confidential information to rival firms.

#### 6. End Use Method or Input-Output Method:

This method is quite useful for industries which are mainly producer’s goods. In this method, the sale of the product under consideration is projected on the basis of demand survey of the industries using this product as an intermediate product, that is, the demand for the final product is the end user demand of the intermediate product used in the production of this final product.

The end user demand estimation of an intermediate product may involve many final good industries using this product at home and abroad. It helps us to understand inter-industry’ relations.

In input-output accounting two matrices used are the transaction matrix and the input co-efficient matrix. The major efforts required by this type are not in its operation but in the collection and presentation of data.

## Statistical Methods:

Basically all statistical approaches of forecasting, project historical information into the future. These are based on the assumption that future patterns tend to be extensions of past ones and that one can make useful predictions by studying the past behaviour i.e. the factors which were responsible in the past will also be operative to the same extent in future.

1) Time series analysis or trend method:   Time series forecasting uses historical figures to predict future results. For example, a restaurant may use last month’s sales figures to predict how much food it will sell the next month.   One of the drawbacks of time series forecasting is that it assumes the future will be the same (or similar) to the past. It does not address any other variables.

Statistical tool to predict future values of a variable on the basis of time series data.

Time series data are composed of:

Secular trend (T): change occurring consistently over a long time and is relatively smooth in its path.

Seasonal trend (S): seasonal variations of the data within a year

Cyclical trend (C): cyclical movement in the demand for a product that may have a tendency to recur in a few years

Random events (R): have no trend of occurrence hence they create random variation in the series. Additive Form: Y = T + S + C + R………..(1)Multiplicative Form: Y = T.S.C.R………….(2)Log Y= log T + log S + log C + log R………….(3)

Methods of Trend Projection:  It can be done using the following methods:

a. Graphical method:

• This method gives the basic tendency of a series to grow, decline and remain steady over a period of time.
• This method is useful in forecasting India’s population, demand for textiles, cement, etc. the period of time in trend analysis is always a long period.
• In this method the period is taken on X-axis and the corresponding sales values on y-axis and the points are plotted for given data on graph paper.
• Then a free hand curve passing through most of the plotted points is drawn. This curve can be used to forecast the values for future.

In Fig. 1, AB is the trend line which has been drawn as free hand curve passing through the various points representing actual sale values.

b. Moving average method: a moving average forecast uses a number of most recent historical actual data value to generate a forecast. The moving average (simple) for number of period is calculated as:

Ft = At-1+ A t-2 + ———–+ At-n

N

Ft = Forecast for the coming period

n  = No. of periods to be averaged.

At-1 = Actual occurrence in the past period

At-2, At-3 and At-n = Actual occurrence two periods ago, three periods ago etc.

This method is used to estimate the average of a demand time series and remove the effects of random fluctuations. It is most useful when demand has no trend seasonal when demand has no seasonal fluctuations.

In this method, if we use period moving average, the average demand for the n most recent time period is calculated and used as forecast for the next time period for the next period, after the demand is average is replaced with the most recent demand and the is recalculations illustration:

c. Weighted moving Average: Whereas the simple moving average gives equal weight to each component of the moving average database, a weighted moving average allows any weights to be placed on each element, providing that the sum of all weight equals one.

The formula for a weighted moving average is

Ft = W, At-1, + W2 At-2 + ———– Wn At-n

Where   W1 = Weight to be given to the actual occurrence for the period t-1

W2 = Weight to be given to the actual occurrence for the period t-1

N   = Total no. of forecast

∑n  wi= 1 (sum of all weight must equal)

i=1

For example a departmental store may find that in a four months period, the best forecast is derived by 40% of the actual sales for the most recent month, 30% of two months ago, 20% of three months ago and 10% of four months ago. If actual sale was as follows, forecast the demand in the month of 5.

Choosing weight: Experience and hit and trial methods. As a general rule, the most recent past is the most important indicator of future. Therefore it should get higher weighted and if data are seasonal, more weight can be assigned to the months if season.

d. Least square method

This is one of the best method to determine trend. In most cases, we try to fit a straight line to the given data. The line is known as ‘Line of best fit’ as we try to minimise the sum of the squares of deviation between the observed and the fitted values of the data. The basic assumption here is that the relationship between the various factors remains unchanged in future period also.

• Estimates coefficients of a linear function

Y=a+bX where a =intercept and b =slope

• The normal equations: ΣY=na + bΣX

ΣXY= aΣX+ bΣX2

Once the coefficients of the trend equation are estimated, we can easily project the trend for future periods.

Let Y denote the demand and X the period for a certain commodity. Then the linear relationship between Y and X is given by

Y = a + bX ……………………………………… (3)

the nature of the relationship is determined by the values of a and b. The values of a and b can be estimated with the help of the past information about Y and X. If x arid y denote the deviations of X and Y from their respective means, then the least square estimates of a and b are given by a and b where n is the number of observations.

Illustration: The sales figure(000s) for a firm over a period of seven years are given below.(1) Fit a Trend Line  by the method of least square,(2) Based on the trend Project the sales for the year 2010.

Ans: Trend Line:-      S = 7.14 + .71t

Projected the sales for the year 2010:-     Rs.11.42(000s)

e. Exponential Smoothing: In the previous methods of forecasting (simple and weighted moving average) the major drawback is the need to continually carry a large amount of historical data. As each new piece of data is added in these methods, the new forecast is calculated. In many applications, the most recent occurrence are more indicative of future than those in the more distant past. If this premise is valid- than the importance of data diminishes as the past becomes more distant, then exponential smoothing is the most logical and earliest method to use.

In exponential smoothing method, only three pieces of data are needed to forecast the future, the most recent forecast, the actual demand that occurred for that forecast period and a smoothing constant.

The smoothing constants determine the level of smoothing. The value of is determined both by the nature of the product and the managers sense of what constitute a good response rate. For example if a firm produces a standard item relating stable demand, the reaction rate  to differences b/w actual and forecast demand would tend to be small. However, if the firm is expecting growth, it would be desirable to have a high reaction rate, may be 15 to 30% points, to give greater importance. The equation for a single exponential smoothing forecast is

Ft = Ft-1 + α (At-1 – Ft-1)

Where              Ft = The exponentially smoothed forecast for period t.

Ft-1 = The exponentially smoothed forecast made for prior period.

At-1 = Actual demand in the prior period.

α = The desired response rate or smoothing constant.

This equation states that the new forecast is equal to the old forecast plus a portion of error (the difference b/w the previous forecast and what actually occurred).

Illustration: Assume that the long run demand for a product under study is relating stable and a smoothing constant α is considered 0.05. Assume that last month forecast (Ft-1) was 1,050 units. If 1,000 actually were demanded, rather than 1,050, the forecast for this month would be

Ft = Ft-1+ α (At-1 – Ft-1)

= 1,050 + 0.05 (1,000-1,050)

= 1,050 + 0.05 (-50)

= 1,047.5 units

2) Regression and Correlation: These methods combine economic theory and statistical technique of estimation. Under these methods the relationship between the sales (dependent variable) and other variables (independent variables such as price of related goods, income, advertisement etc.) is ascertained. Such relationship established on the basis of past data may be used to analyse the future trend. The regression and correlation analysis is also called the econometric model building.

3) Barometric Techniques

• Barometric Technique alerts businesses to changes in the overall economic conditions.
• Helps in predicting future trends on the basis of index of relevant economic indicators especially when the past data do not show a clear tendency of movement in a particular direction.

Indicators may be

• Leading indicators: economic series that typically go up or down ahead of other series e.g. birth rate of children is the leading series for demand of seats in schools.
• Coincident indicators : move up or down simultaneously with the level of economic activities e.g. national income series is coincident with series of employment in economy.
• Lagging series: which moves with economic series after a timelag e.g. industrial wages is lagging series of price index for industrial workers.

4) Simultaneous Equations Method

It is based on the fact that in any economic decision every variable influences every other variable in an economic environment.

• Incorporates mutual dependence among variables.
• It is a simultaneous and two way relationships,
• A typical simultaneous equation model may comprise of:
• Endogenous variables: included in the model as dependent variables
• Exogenous variables: given from outside the model