Forecasting and Types of demand variations

Forecasting can be termed as the prediction of the future sales or demand of a particular product.

It is a projection based upon past data and the art of human judgment. The survival of any organization depends upon how well they can project demand in the future.

Types of demand variations

Trend variation (T):

It shows the long-term upward or downward movement in the demand pattern of a particular product.




Seasonal (S):

It shows the short-term regular variation related to a particular time of a day or day of a week.




It shows the long term wave like demand variation normally for more than a year.


Irregular variations ( I ):

These variations are caused due to unusual circumstances which are not reflective of the normal behavior. These may be due to the government policy change, strike, price of hike, shut down and many other.



Forecast must be divided in to two types they are

  • Qualitative
  • Quantitative


It is also known as subjective. The qualitative forecast must be consists of judgmental


It is also known as objective forecast. The quantitative must be divided into two types they are time services and casual or econometric


The judgmental must depend upon the art of the human judgment, to know how well a human being can predict the demand of a product in future. These methods does not require past data or sales figures.

The judgmental forecast must be divided into four types they are

  • Opinion survey
  • Market trail
  • Market research
  • Delphi technique

 Opinion survey:

In this method, opinions are collected from the customer or retailer and distributor regarding the demand pattern of a particular product. The details about various factors which influence the demand are related to get the forecast.


Market trail:

It is pre launched activity and the product is introduced amongst the limited population, which is used to project the demand from the bigger population. It is applied for low-cost consumables like cosmetics, cool drinks, chocolates items etc.

Delphi Technique:

In this method, a panel of experts are asked sequential question in which the response to one question leads to the next question. It is a step –by- step procedure in which information available to some expert and the final forecast is obtained by the common opinion of all the experts.

Time series:

In the time series method, the demand of a product is given in the time order. In this method  past data are arranged in the chronological order as dependent variable, and the time as independent variable based upon the past date. One needs to project the future demand of the product.

The time series are divided into four types. They are

  • Past average
  • Moving average
  • Weighted moving average
  • Exponential smoothing

Past Average:

In this method forecast is given by the average or mean of the actual demand data for the previous period.

Simple moving Average

It is also known as rolling average. This method uses the past data and calculates the rolling average for a constant period. Fresh average is computed at the end of each period by adding the actual demand data for the most recent period and deleting the data for older period. The method of changing the data from period to period it is known as moving average method.

Weighted Moving Average:

In this method the summation of the weights should be equal to one. This method is similar to moving average method the only difference. This method gives the highest weight and the weight assigned in such a manner that summation of all weights is always equal to one.

Exponential smoothing:

In this method the user requires only the current demand data for the recent period and the forecasted value for the current period to give the next forecast. This method gives weight to all the previous data but the weights assigned are in exponentially decreasing order. The most recent data is gives the highest weight and the weight assigned to the older data decreases exponentially.

Responsiveness and stability


Responsiveness indicates that the forecast have a fluctuating or swigging pattern. It is preferred for new product and for that particular number of period it is kept small.


Stability means that the forecast pattern is flat smooth or has fewer fluctuations. It is preferred for old existing product and for that particular number of period it is kept large.

Casual forecast:

It is also known as economic forecast. In this method forecasted tries to establish cause and effect relation between the demand of a product and any other variable on which demand is dependent. The objective is to establish a relation such that changes in one variable become useful for the prediction of other.

Correlation Analysis:

Correlation coefficient tells the degree of closeness between the two variables and it is an indicator of the extent to which knowledge of one variable becomes useful for the prediction of other. The correlation coefficient between the two variables x and y is given by



Where  \vec x  and \vec y  are the average value of individual x and y value.

Linear Regression Analysis:

It is a mathematical technique of obtaining the line of best fit, between the demand of a product and any other variable on which demand is dependent. In regression analysis the relationship between the dependent variable y and some independent variable x can be represented by a straight line.



Y= dependent variable

x= independent variable

Forecast Errors

When the forecast error is studied for long duration, it turns out to be helpful to find the particular pattern or trend. This may regulate our future production. Commonly used methods to find the forecast are.

  • Mean absolute deviation (MAD)
  • Mean forecast error (MFE)
  • Mean square Error (MSE)
  • Mean absolute percentage error (MAPE)

Mean absolute Deviation: (MAD)

Mean absolute deviation indicates the average magnitude of error in every period without taking sign of error into consideration.


Mean forecast Error: (MFE)

It is also known as bias. It measures the forecast error with regards to direction and shows many tendency of over (or) under forecast. Positive bias is indicates under estimated forecasting and negative bias is indicated over estimated forecasting.




Mean square error: (MSE)

Mean square error is used to compute standard deviation for forecast error which is utilized to plot control chart for forecast error.



Mean Absolute percentage (MAPE)

Mean absolute percentage is used to find the average of percentage error compared to actual demand and it is used to put demand in perspective, as there is difference between the 10 out of 100 and 10 out of 1000.

  • It helps in determining the volume of production and the production rate
  • It forms the basis for production budget, material budget and labour budget
  • It is essential for production design and development
  • It suggest the need for plant expansion
  • It helps in determining extent of marketing, advertising and distribution required.

Click Here To Know More About: Sequencing and Rules of sequencing