MODULE 2: Demand Forecasting FORECASTING

 Art and science of predicting a future

event
TOPICS:  Underlying basis of all business
decisions
-WHAT IS FORECASTING?
- Production
 Forecasting Time Horizon - Inventory
 Influence of Product Life Cycle - Personnel
 Types of Forecast - Facilities

-STRATEGIC IMPORTANCE OF FORECASTING FORECASTING TIME HORIZONS

-FORECASTING APPROACHES  SHORT-RANGE FORECAST

- Up to 1 year, generally less than 3
 Qualitative Forecasting Models months
 Quantitative Forecasting Models - Purchasing, job scheduling workforce
levels. Job assignments, production
levels
-TIME SERIES FORECASTING
- tend to be more accurate than longer-
 Decomposition of a Time Series term forecast
 Naïve Approach
 Moving Averages  MEDIUM-RANGE FORECAST
 Exponential Smoothing - 3 months to 3 years
 Exponential Smoothing with Trend - Sales and production, planning,
Adjustment budgeting
 Trend Projections
 Seasonal Variations in Data  LONG-RANGE FORECAST
 Cyclical Variations in Data - 3+ years
- New products planning facility
location, research and development
-ASSOCIATIVE FORECASTING METHODS: - deal with more comprehensive issues
REGRESSION AND CORRELATION ANALYSIS
INFLUENCE OF PRODUCT LIFE CYCLE
 Using Regression Analysis for
Forecasting  INTRODUCTION-GROWTH-MATURITY-
 Standard Error Estimate DECLINE
 Correlation Coefficients for Regression  Introduction and growth require longer
Lines forecasts than maturity and decline
 Multiple Regression Analysis  As product passes through life cycle,
forecasts are useful in projecting;
staffing levels, inventory levels, factory
-MONITORING AND CONTROLLING FORECAST capacity
 Adaptive Smoothing
 Focus Forecasting

- FORECASTING IN THE SERVICE SECTOR

  Jury of Executive Opinion- Pool
opinions of high-level experts,
TYPES OF FORECAST sometimes augment by statistical
 ECONOMIC FORECAST- addresses models
business cycle- inflation rate, money  Delphi Model- Panel of experts, queried
supply, housing starts, etc. iteratively , 3 types of Participants:
Decision Makers, Staff, Respondents
 TECHNOLOGICAL FORECASTS- predicts  Sales force composite- Estimates from
rate of technological progress, impacts individual salespersons are reviewed for
development of new products reasonableness, then aggregates
 Consumer Market Survey- Ask the
 DEMAND FORECASTS- predicts sales of customer
existing product and services
 QUANTITATIVE METHODS
STRATEGIC IMPORTANCE OF FORECASTING
- Used when situation is stable and
 HUMAN RESOURCES- Hiring, training, historical data exist: Existing products,
laying off workers Current technology
 CAPACITY- Capacity shortages can
result in undependable delivery, loss of - Involves mathematical techniques:
customers, loss of market share forecasting sales of color television
 SUPPLY CHAIN MANAGEMENT- Good  OVERVIEW OF QUANTITATIVE
supplier relations and price advantages METHODS

SEVEN STEPS IN FORECASTING  TIME SERIES MODELS:

1. Determine the use of the forecast o Naïve Approach
2. Select the items to be forecasted
o Moving Averages
3. Determine the time horizon of the
o Exponential Smoothing
forecast
 ASSOCIATIVE MODEL
4. Select the forecasting model(s)
o Trend Projection
5. Gather the data
o Linear Regression
6. Make the forecast
7. Validate and implement results TIME SERIES FORECASTING

NOTE: REALITIES  Set of evenly spaced numerical data

 Forecast based only on past values, no
 Forecasts are seldom perfect
other variables important
 Most techniques assume an underlying
stability in the system TIME SERIES COMPONENTS
 Product family aggregated forecasts
are more accurate than individual  TREND COMPONENT- persistent,
product forecasts changes due to population,tech etc.,
typically several years duration
FORECASTING APPROACHES  SEASONAL COMPONENT- Regular
pattern of up and down fluctuations,
 QUALITATIVE METHODS
due to weather, occurs within a single
- Used when situation is vague and little year
data exist: New products, New technology  CYCLICAL COMPONENT- Repeating up
and down movements, affected by
- Involves intuition, experience: Forecasting business cycle, political etc., multiple
sales on internet years duration
 OVERVIEW OF QUALITATIVE METHODS
  RANDOM COMPONENT- Erratic, EXAMPLE
unsystematic, residual fluctuations, due
to random or unforeseen events, short
duration and non-repeating

 NAÏVE APPROACH
- assumes demand in next period is
the same as demand in most recent
period
- sometimes cost effective and
efficient, can be a good starting
point

POTENTIAL PROBLEMS WITH MOVING

 MOVING AVERAGE METHOD AVERAGE:

 MA is a series of arithmetic means  Increasing n smooth the forecast but

 Used if little or no trend makes it less sensitive to changes
 Used often for smoothing  Do not forecast trends well
 Provides overall impression of data over  Require extensive historical data
time
 EXPONENTIAL SMOOTHING
Moving Average=
∑ demand ∈ previous n periods  Form of weighted moving average
EXAMPLE: n
o Weights decline exponentially
o Most recent data weighted
most
 Requires smoothing constant ()
o Ranges from 0 to 1
o Subjectively chosen
 Involves little record keeping of past
data

 WEIGHTED MOVING AVERAGE

 Used when some trend might be
present
 Older data usually less important
 Weights based on experience and
intuition

IMPACT OF DIFFERENT ALPHA

  Chose high values of  when
underlying average is likely to
change
 Choose low values of  when
underlying average is stable

CHOOSING ALPHA

- We generally do this by selecting the

model that gives us the lowest forecast

error.

Forecast error = Actual demand -

Forecast value

= At - F t

 COMMON MEASURES OF ERROR

 

Example:

Example:
Example:

 EXPONENTIAL SMOOTHING WITH

TREND ADJUSTMENT