Ch 18: Forecasting

Forecasting is essential to planning for ___ 

all areas of busines 

tracking signal should be 

< -3 or >3

The lower the MAD, the:

better

Who developed CPFR? 

Walmart 

Tracking signal finds:

bias

a vital function and effects every significant management decision

forecasting

tactical forecasts are used to guide:

day to day decisions

point at which some form of inventory is stored that allows supply chain partners/entities to operate independently

decoupling point

otherwise known as an inventory buffer (line for services) and is represented by an upside-down triangle

decoupling point

the choice of the decoupling point in a supply chain is: 

strategic 

Forecasting helps determine the___ needed at the decoupling point

level of inventory

What are the four basic types of forecasting

Qualitative
Time Series Analysis
Causal Relationships
Simulation

relies on the future behaving, at least to some extent, similarly to the past

time series analysis

What are the components of demand? (6) 

Average demand for period of time 
Trend
Seasonal element 
Cyclical elements 
Random Variation 
Autocorrelation

Means errors correlate together (this is hard to measure)

autocorrelation

What are two examples of seasonal element

swimsuits and snowblowers

What are the common trend types?

Linear
S curve
Asymptotic
Exponential

What is the most common type of trend? 

S curve 

using the past to predict the future

time series analysis

what are the types of time series analysis

short term
medium term
long term

time series analysis used for forecasting for less than 3 months

short term 

time series analysis used for forecasting for 3 months to 2 years 

medium term 

time series analysis used for forecasting for more than 2 years

long term

time series analysis used mainly for tactical decisions 

short term 

time series analysis used to develop a strategy that will be implemented over the next 6 to 18 months (meeting demand)

medium term

time series analysis used for detecting general trends and identifying major turning points

long term

Choosing the appropriate forcasting model depends on: (5)

Time horizon to be forecasted
Data available
Accuracy required
Size of forecasting budget
Availability of qualified personnel

What are the forecasting methods? (5)

Simple moving average 
Weighted moving average (and simple exponential smoothing) 
Exponential smoothing with trend 
Linear regression 
Trend and seasonal models 

What is the data pattern for simple moving average

stationary (no trend or seasonality)

What is the data pattern for weighted moving average and simple exponential smoothing

stationary (cyclical)

forecast is the average of a fixed number of past periods 

simple moving average 

useful when demand is not growing or declining rapidly and no seasonality is present; removes some of the random fluctuation from the data; selecting the period length is important

simple moving average

longer periods provide more ____

smoothing

shorter periods react to ___ more quickly

trends

What is the denotation for simple moving average? 

MA(n)

the ___ formula implies equal weighting for all periods

simple moving average

with weighted moving average there is ___ weighting or prior time periods (cyclical)- usually higher weights in periods farther in past

unequal

How do you select weights?

Experience
Trial and error
Recent past is best indicator

for weighted moving average, if the data is seasonal or cyclical, weights should: 

reflect this appropriately 

The weighted average method that includes all past data in the forecasting calculation; most used of all forecasting techniques; an integral part of computerized forecasting

exponential smoothing

exponential smoothing is well accepted for 6 reasons

Surprisingly accurate
Formulating model is relatively easy
Easy to understand
Little computation required
Computer storage requirements are small
Tests for accuracy are easy to compute

a trend in data causes the exponential forecast to:

always lag the actual data

The lag in the actual data can be corrected somewhat by adding in a trend adjustment 

exponential smoothing 

To correct the trend, we need what two smoothing constants?

Smoothing constant alpha
Trend smoothing constant delta

___ depends on how much random variation is present

alpha

___ depends upon how steady the trend is

delta

used to identify the functional relationship between two correlated variables, usually from observed data 

linear regression 

determines the parameters a and b such that the sum of the squared errors is minimized

least squared method

chronological ordered data are referred to as a time series, which may contain one or many elements (trend, seasonal, cyclical, autocorrelation, and random), separating these elements and separating the time series data into these components =

decomposition

may either be additive or multiplicative and varies over time

seasonal variation

ratio or index of the amount sold during each season divided by the average for all seasons 

seasonal factor

we use ___ to create the final forecast of decomposition

regression

the difference between the forecast value and what actually occurred

forecast errors

All forecasts contain some level of___

error

What are the sources of forecasting errors? 

Bias
Random 

When a consistent mistake is made

bias

errors that are not explained by the model being used

random

Measures of error (3)

Mean absolute deviation (MAD)
Mean absolute percent error (MAPE)
Tracking signal

larger values of MAD indicate : 

a less accurate model

MAPE scales the forecast error to the:

magnitude of demand

indicates whether forecast errors are accumulating over time (either positive or negative errors)

Tracking signal

forecasting that uses independent variables other than time to predict future demand

causal relationship

The independent variable in a causal relationship forecast must be a: 

leading indicator 

in most cases, more than one independent variable will be valid predictor of future demand… but in the case of causal relationship forcasting , the forecast analyst may utilize___

multiple regression

analogous to simple linear regression analysis but with multiple independent (explanatory variables)

multiple regression

generally used to take advantage of expert knowledge; useful when judgement is required, when products are new, or if the firm has little experience in a new market

qualitative forecasting

What are examples of qualitative forecasting techniques

market research
panel consensus 
historical analogy 
delphi method (expert knowledge)

web based process used to coordinate the efforts of a supply chain and depends on the exchange of internal info to provide a more reliable view of demand

CPFR

What does CPFR stand for?

Collaborative Planning, Forecasting and Replenishment

CPFR integrates ___ members of a supply chain

all

What are the steps of CPFR? (5) 

Creation of front end partnership agreement (NDA)
Joint business planning 
Development of demand forecast
Sharing forecasts 
Inventory replenishment 

What are the principles of forecasting (5)

Forecasting is fundamental step in any planning process
All functional areas rely on the forecast
Forecast effort should be proportional to the magnitude of the decisions being made
Web based (CPFR) growing in importance and effectiveness
All forecasts have errors