Roadmap to the Lecture:Fundamental Types of Forecasts, Finer Classification of Forecasts

<< INTRODUCTION TO PRODUCTION AND OPERATIONS MANAGEMENT:Demand Management

Time Series Forecasts:Techniques for Averaging, Simple Moving Average Solution >>

Production and Operations Management MGT613

Lesson 08

Roadmap to the Lecture

·Discuss the requirements of a good forecast.

·Steps in making a forecast.

·Fundamental types of forecast.

·Finer classification of forecast

·Discuss characteristics of Judgmental Forecasts.

·Delphi Method.

·Time Series Analysis.

·Naïve Forecast.

Requirements of a Good Forecast

·Timely.

The forecast should be timely. Indicating that forecasting horizon should provide enough time

to implement possible changes. Capacity cannot be expanded instantly it requires some time to plan,

coordinate and increase the required resources.

·Reliable. Forecasts should be reliable meaning that it should work consistently. A forecast that is

partially correct will succeed at sometime and sometime fail making the end users question the purpose

and intent of forecasting.

·Accuracy. Forecasts should be accurate. In fact it should carry the degree of accuracy, so the users are

aware of the limitations of the forecast. This will also help the end users to plan for possible errors and

provide a basis for comparing the forecast with other alternative forecasts.

·Meaningful Forecast should be expressed in meaningful units. Financial Planners will use Rupees to

show how much capital would be required; Mechanical Project Schedulers would require Forecasts to

carry the type of machines and crafts of technicians required.

·Written/Documented. The forecasts should be presented in writing. A documented forecast always

provides a chance to measure the variance between estimate and actual result at a later stage.

·Simple to understand and use meaning that Forecasts should not be dependant upon usage of

sophisticated computer techniques or task specific highly qualified technical personnel. A failure or

limitation on the part of this can lead to an incorrect decision and less acceptance amongst end users

Steps in the Forecasting Process

·Determine the purpose of the forecast meaning what is the purpose and when will it be required.

This will provide the level of detail for resources required man, machine, time and capital.

·Establish a time horizon. We already know that as time increases the accuracy of the Forecast

decreases

·Select a forecasting technique whether qualitative or quantitative

·Gather and analyze the appropriate data. It goes without saying that before a forecast can be

delivered data is required. The closer the real life data more realistic would be the forecast. This may be

the time when you would like to identify the important assumptions and suppositions.

·Prepare the forecast.

·Monitor the forecast. A forecast has to be closely monitored to determine whether it is fulfilling its

basic purpose. This helps in re-examining the method, assumptions and validity of the data and

preparing a revised forecast.

Fundamental Types of Forecasts

·Qualitative Techniques which use subjective inputs and no numerical data. It relies solely on soft

information like human factors, personal opinion, hunches. Thus Qualitative Forecasts are often biased

and tilted towards what the management wants to predict.

Production and Operations Management MGT613

·Quantitative Forecast involves the extension of the historical data. It sometimes make use of

forecasting technique that uses explanatory variables to predict future demands. Quantitative techniques

are favored where quality attributes cant be quantified.

·In reality both need to be used together to develop a judicious and realistic forecast.

Finer Classification of Forecasts

·Judgmental - uses subjective inputs meaning that a judgmental forecast rely on analysis of subjective

inputs obtained from various sources, such as consumer surveys, the sales staff, managers and

executives, and panels of experts. These insights are not available publicly.

·Time series - uses historical data assuming the future will be like the past and depend on developing

relationships between variables that can be expressed to predict future values. Some time series forecast

try to smoothen out random variations in historical data. There are some time series forecast which

identify specific patterns and then may even extrapolate those patterns into the future.

· Associative models - uses explanatory variables to predict the future for example demand for a small

car may be dependant upon increase in price of petrol or CNG. The analysis in this case would employ a

mathematical model that would relate the predicted variable with the predictor variable or variables.

Judgmental Forecasts Characteristics

·Judgmental Forecasts rely solely on judgment and opinion to make forecasts.

·In the absence of enough time, it is easy to use qualitative type of forecast.

·In case of changing external environment economic and political conditions, organizations may use

judgmental forecasts.

·When introducing new products, services, new features, new packaging, judgmental forecasts are used

in preference over quantitative techniques.

Judgmental Forecasts

·Executive opinions normally consist of a group of senior level managers from different interfaces,

used for long range planning and new product development. Advantage being the collective pool of

information from all divisions and departments, disadvantage being that one person will dominate other

interfaces, which can lead to erroneous forecasts.

·Sales force opinions have the advantage of being in direct contact with customers. The sales force can

detect the customers' change of plan, However it suffers from the fact that it can not differentiate

between what the customer can do and will do. Current data of sales can often lead to over pessimistic

and overly optimistic forecasts, which then results in incorrect sales projections.

·Consumer

surveys are based on sample taken from potential customers. These type of surveys require

skill to develop, administer and interpret the results. Often fall victim of the consumers irrational

behavior of buying.

·Outside opinion which is a mix of consumer and potential customers. This kind of opinion is now a

days readily available through internet, telephonic surveys and newspapers. Its biggest limitation is a

fixed format which often fails to quantify the exact demand forecast.

·Delphi

method: Managers and staff complete a series of questionnaires, each developed from the

previous one, to achieve a consensus forecast. Commonly used for Technological forecasting, when to

introduce a new technology. It's a long term one time activity and has the same issues like expert

opinion type of judgmental forecast.

Time Series Analysis

·Time series forecasting models try to predict the future based on past data

·We as Managers can pick models based on:

1. Time horizon to forecast

2. Data availability

3. Accuracy required

4. Size of forecasting budget

5. Availability of qualified personnel

Production and Operations Management MGT613

Naïve Forecasts

·Simple to use

·Virtually no cost

·Quick and easy to prepare

·Data analysis is nonexistent

·Easily understandable

Drawbacks

·Cannot provide high accuracy

·Can be a standard for accuracy

Table of Contents: