Forecasting

(2nd chapter)

Introduction to forecasting
Purpose: ...to match supply to demand...Shortage VS Excess
How much capacity?

Two aspects of forecasts are important:

Expected level of demand
Degree of accuracy

Expected level of demand and degree of accuracy depends on:

Structural variation (trend and seasonal)

Demand
Random variation
Unforseen event (Irregular variation)

Time horizon
Short-term forecast VS long-term forecast

Introduction to forecasting
Yield management:
Percentage of capacity being used?

Forecasting of customer demand (Unit sales,

Stockout)

Experience, judgment, and technical expertise

Science and intuition

Features common to all

forecasts...
...the same underlying causal system that existed in
the past will continue in the future...
Forecasts are not perfect...
Forecasts for groups of items tend to be more accurate
than forecasts for individual items...
Forecasts accuracy decreases as the time horizon
increases...

Elements of a good forecast...

Timely
Accurate (degree of accuracy)
Reliable
Meaningful units (financial planner VS production
planner)
In writing
Simple to understand and use
Cost-effective

Steps in the forecasting process...

1.
2.
3.
4.
5.
6.

Purpose
Time horizon
Data (data may need to be cleaned)
Forecasting technique
Make the forecast
Monitor (reexamine)

Forecast accuracy...

Forecast accuracy...

Forecast accuracy...two objectives...

Approaches to forecasting...
Judgemental forecast
Time-series forecast
Associative model

Approaches to forecasting...qualitative
forecasts
Executive opinion
Knowledge
The view of one person will prevail
Less pressure to produce a good foreast

Salesforce opinion
Any plans the customers may be considering for the
future
What they actually will do?
Influenced by recent experiences (pessimistic vs
optimistic)
Sales quota

Approaches to forecasting...qualitative
forecasts
Consumer survey

Information that might not be available elsewhere

Knowledge
Expensive
Time-consuming
Irrational behavior pattern
Low reponse rate

Delphi method
Very little hard information, single-time forecast,
analytical technique
Technological foreast

Time series data

Time series...observations...regular interval
No attempt is made to identify variables that influence
the series
Plotting the data
Types:

Trend
Seasonality
Cycle
Irregular variations
Random variations

Naive method
Stable series
Seasonal variation
Trend
Weakness?
Benefits?

Techniques for averaging

Random variation (white noise)
Two kinds of variation...

Techniques for averaging

Moving average
Most recent actual data values

3 period moving average

5 period moving average

Techniques for averaging

Moving average
Most recent actual data values

Few data point VS more data point

Sensitive (responsive)
More data point will smooth more

Techniques for averaging

Moving average
Advantages:
Easy to compute
Easy to understand

Disadvantages:
All values in the average are weighted equally (the old value has the
same weight as the most recent value)

Techniques for averaging

Moving average

Techniques for averaging

Moving average

Techniques for averaging

Weighted moving average

Techniques for averaging

Weighted moving average
It assigns more weight to the most recent values
Note that the weights must sum to 1
Advantage:
More reflective of the most recent occurrences
Trial and error

Techniques for averaging

Exponential smoothing

Responsiveness VS smoothing

Starting forecast
Several period back
One period back

Techniques for averaging

Exponential smoothing

Techniques for averaging

Exponential smoothing

Other forecasting methods

Focus forecasting
Several forecasting method
Historical data
Highest accuracy

Diffusion model

New product
New Service
Historical data
Established product
Mathematical diffiusion model

Techniques for trends

Linear trend
Non-linear trend:

Parabolic trend
Life-cycle trend
Exponential trend
Growth curve

Linear trend:
Trend equation
Expansion of exponential smoothing

Techniques for trends

Linear trend equation
y=a+bx

Techniques for trends

Linear trend equation

Predict sales for week 11 and 12

Associative model
Simple linear regression
Equation of a straight line
Least square line

Y=a+bx
Outside the range of observed values

Associative model
Simple linear regression
Standard error of estimate

Logical explanation

Associative model
Simple linear regression
Correlation

r
Strength and direction
-1 to + 1

Square of the correlation coefficient

r2
Percentage of variability explained by the independent variable
0 to 1