2. Learning objectives
∗ Define and understand forecasting
∗ Identify the different types of forecasts
∗ Identify and discuss the various time horizons
∗ Discuss different approaches to forecasting
∗ Determine the steps in the forecasting process
∗ Describe and solve averaging techniques in
forecasting
3. Learning objectives (cont.)
∗ Solve simple moving average problems
∗ Solve exponential smoothing problems
∗ Determine what constitutes a good forecast
∗ Compare qualitative and quantitative forecasting
methods.
4. 2.1 Introduction
∗ Forecasting =
∗ Preface to planning
∗ Attempt to predict future value of a changing variable
∗ Subjective or objective
∗ Addresses:
∗ Macro-circumstances
∗ Competitiveness
∗ Market tendencies
∗ Sourcing funds required.
5. Examples of Forecasts
Management Information Services – could predict new technology, eg.
Advances in internet
Human Resources – could answer the question “will there be a need
to employ more people in the future
Goods and service design – needs of the customer in the
future
Finance – need for capital replacement, cash flows,
budgets
Operations – scheduling, inventory planning, labour
requirements, and project management
Marketing – prices for new products, promotional
plans, competition analysis
6. 2.2 The uses of forecasts
∗ Plan the system as a whole (long-range)
∗ Plan the use of the system
∗ Business forecasting (predict demand)
∗ Forecasting is never an exact science
∗ Context determines choice of method.
7. 2.3 Features common to all forecasts
∗ Assumption that past trends will be present in future
∗ No precise prediction can be made
∗ Groups more accurate
∗ Longer time horizon is less reliable.
8. 2.4 Time horizons for doing forecasts
∗ Short-range (few weeks – 12 months)
∗ More accurate
∗ Medium-range (12 months – 5 years)
∗ Long-range (5 years +)
∗ Medium and long-range: deal with organisation as a whole.
9. 2.4 Time horizons for doing forecasts
(cont.)
Time Horizon Accuracy Frequency Management Method
Level
PROCESS Long term Medium Single Top Qual or Quan
DESIGN
CAPACITY Long term Medium Single Top Qual or Quan
PLANNING
AGGREGATE Medium term High Few Middle Casual or time
PLANNING* series
SCHEDULING Short term Very high Many Lower Time series
INVENTORY M/ Short term Very high Many Lower Time series
MENT
* Capacity planning for medium term 3-18 months
10. 2.5 Requirements of an accurate forecast
∗ Use simple technique
∗ Accuracy
∗ Cost effectiveness
∗ Meaningful units
∗ Timely
∗ Reliable
∗ Should be in writing.
12. 2.7 Important situational factors to be
considered
∗ Accuracy and cost – trade off between accuracy and cost
∗ Availability of data – large pool of data and relevant
∗ Time span – the longer it is the less accurate it is
∗ Nature of the goods and services – life cycle, seasonal
variations
∗ Changes in the market - difficult for new products
∗ Use or decision factors – method used and subject should
be closely related
13. 2.8 Reasons for ineffective forecasts
∗ Failure to select applicable model
∗ Inability to recognise that forecasting must form an integral
part of the business
∗ Neglecting to monitor the accuracy of the forecast
∗ Failure to involve all of the relevant people
∗ Inability to realise that the forecast will be wrong
∗ Forecasting of incorrect items is not helpful.
14. 2.9 Approaches to forecasting
∗ Qualitative – mainly judgments of the parties involved
∗ Quantitative – calculations & statistical techniques
∗ Associative forecasting techniques – use of equations that
are descriptive of the variables used. A variable is a factor
that will influence the composition of a forecast eg. Price of
product, weather etc
15. 2.9 Approaches to forecasting
∗ Categories of forecasting techniques:
∗ Associative methods – equations that
describe the variables
∗ Judgmental forecasts – rely on subjective
judgment of an individual
∗ Time series forecasts – data is manipulated
using mathematical techniques
16. 2.9 Approaches to forecasting
∗ Qualitative approach:
∗ Relied upon when hard data not available
∗ Used when forecast required in a hurry
∗ Approaches:
∗ Consumer surveys – very expensive, validity questionable
∗ Jury of executive opinion – top level managers, long term
forecast
∗ Sales-force opinion – grassroots method, very questionable
∗ Delphi method – respondents outside company
∗ Educated guess – personal insight. Highly unreliable
∗ Historical analogy – only if a similar product exists
17. 2.9 Approaches to forecasting (cont.)
∗ Quantitative approach:
∗ Time series data – use of historical data. Assumes future can be
based on history
∗ Trends – upward or downward movement
∗ Seasonality – mostly regular
∗ Cycles – e.g. stock market indicators
∗ Irregular variations – e.g. flood. Never include in a forecast
∗ Random variations – no logical explanation
19. 2.9 Approaches to forecasting
MOVING AVERAGE IN EXCEL
PERIOD DEMAND MOVING AVERAGE
1 100
2 250
3 220
4 210 =AVERAGE(B2:B5)
5 240 =AVERAGE(B3:B6)
6 255 =AVERAGE(B4:B7)
7 245 =AVERAGE(B5:B8)
8 195 =AVERAGE(B6:B9)
20. 2.9 Approaches to forecasting
∗ Quantitative techniques:
∗ Averaging techniques – the weighted moving average
∗ Very similar to moving average technique
∗ Moving average gives equal weight to all data
∗ Weighted moving average gives different weight to each data
21. 2.9 Weighted Moving Average
Month Sales ( ‘000) 3 period MA
January 240
February 250
March 230
April 220 (0.5 x 230)+ (0.3 x 250) + (0.2 x 240) = 238
May 270 (0.5 x 220)+ (0.3 x 230) + (0.2 x 250) = 229
June 250 (0.5 x 270)+ (0.3 x 220) + (0.2 x 230) = 247
July 255 (0.5 x 250)+ (0.3 x 270) + (0.2 x 220) = 250
22. 2.9 Problems with Moving Average
Longer MA period the more smoothed. Forecast less sensitive
to real fluctuations
MA does not identify any trends in the data. Time lag +/- 2
months
Extensive records of past history must be available
Weight allocated is arbitrary – trial and error needed
23. 2.9 Approaches to forecasting
∗ Quantitative techniques:
∗ Averaging techniques – exponential smoothing
∗ Well accepted because
∗ Calculations to test accuracy are easy
∗ Technique easy to understand
∗ Accuracy high for amount of effort required
∗ Only small amounts of historical data needed
∗ Requires fewer calculations to reach the same answer as other
methods
25. 2.9 Exponential Smoothing
∗ Predicted 142000 units period 1
∗ Actual 153000 units period 1
∗ α =0.2
Demand period 2 = 142000+0.2(153000-142000) = 144200units
26. 2.9 Approaches to forecasting
∗ Associative forecasting techniques
∗ The simple linear regression metho
∗ Most widely used method
∗ Try to find a linear relationship between two variables
27. Summary
∗ Defined forecasting
∗ Defined business forecasting
∗ Common features
∗ Requirements
∗ Steps
∗ Situational factors
∗ Reasons for ineffective forecasts
∗ Approaches to forecasting.
28. For Next Week
∗ Read pages 37 -66 Operations Management
∗ Prepare 1 paragraph discussing the use of forecasts
∗ Prepare 1 paragraph discussing the reasons for
ineffective forecasts.