Demand estimates for products and services are
the starting point for all the other planning in
operations management.
Management teams develop sales forecasts
based in part on demand estimates.
The sales forecasts become inputs to both
business strategy and production resource
forecasts.
Qualitative Approaches
- Usually based on judgments about
causal factors that underlie the demand
of particular products or services
- Do not require a demand history for the
product or service, therefore are useful
for new products/services
- Approaches vary in sophistication from
scientifically conducted surveys to
intuitive hunches about future events
Some Reasons Why Forecasting is Essential
- The approach/method that is
- New Facility Planning–It can take 5 appropriate depends on a product’s life
years to design and build a new factory cycle stage
or design and implement a new
Qualitative Methods
production process.
- Production Planning–Demand for - Educated guess
products vary from month to month - Executive committee consensus
and it can take several months to - Delphi method
change the capacities of production - Survey of sales force
processes.! - Survey of customers
- Workforce Scheduling–Demand for - Historical analogy
services (and the necessary staffing) - Market research scientifically
can vary from hour to hour and
Marketing Research Approaches to Demand
employees weekly work schedules
Estimation
must be developed in advance
- Consumer Surveys (data from survey
Forecasting Methods
questions)
- Qualitative Approaches - Observational Research (data from
- Quantitative Approaches observed behavior)
- Consumer Clinics (data from laboratory
experiments)
- Market Experiments (data from real - Regression Line: Minimizes the sum of
market tests) the squared vertical deviations (et) of
each point from the regression line.
Quantitative Forecasting Approaches
- Ordinary Least Squares (OLS) Method
I. Based on the assumption that the
“forces” that generated the past
demand will generate the future
demand, i.e., history will tend to repeat
itself
II. Analysis of the past demand pattern
provides a good basis for forecasting
future demand
III. Majority of quantitative approaches fall
in the category of time series analysis
IV. Time spans usually greater than one
year
V. Necessary to support strategic decisions
about planning products, processes,
and facilities
Quantitative Forecasting Approaches
1. Regression analysis
2. Time Series analysis
3. Moving Average
4. Exponential Smoothing
5. Barometric Techniques
Regression Analysis
- Regression Line: Line of Best Fit
- Autocorrelation: Consecutive error
terms are correlated.
Linear Regression…….a review
- Linear regression analysis establishes a
relationship between a dependent
variable and one or more independent
variables.
- In simple linear regression analysisthere
is only one independent variable.
- If the data is a time series, the
independent variable is the time period.
- The dependent variable is whatever we
wish to forecast.
Simple Linear Regression
- Regression Equation This model is of the
form:
Y = a + bX
Y = dependent variable
X = independent variable
a = y-axis intercept
b = slope of regression line
Simple Linear Regression
- Constants a and b. The constants a and b
are computed using the following
equations
Problems in Regression Analysis Simple Linear Regression
- Multicollinearity: Two or more - Once the a and b values are computed, a
explanatory variables are highly future value of X can be entered into the
correlated. regression equation and a
- Heteroskedasticity: Variance of error corresponding value of Y (the forecast)
term is not independent of the Y can be calculated.
variable.
Example 1: College Enrollment
Simple Linear Regression
At a small regional college enrollments
have grown steadily over the past six
years, as evidenced below. Use time
series regression to forecast the student
enrollments for the next three years
Equation of Regression line
S = 44.82 –0.641P
S(when P = 30) = 44.82 –0.641*30
= 25.29 (‘000 units)
Time Series Analysis
- A time series is a set of numbers where
the order or sequence of the numbers is
important, e.g., historical demand
Y7 = 2.387 + 0.180(7) = 3.65 or 3,650 students
- Analysis of the time series identifies
Y8 = 2.387 + 0.180(8) = 3.83 or 3,830 students patterns
- Once the patterns are identified, they
Y9 = 2.387 + 0.180(9) = 4.01 or 4,010 students
can be used to develop a forecast
Note: Enrollment is expected to increase by 180
Components of Time Series
students per year.
- Trends are noted by an upward or
Example 2: ABC company Ltd
downward sloping line
Fit a linear regression line to the following data - Seasonality is a data pattern that repeats
and estimatethedemandatprice=Rs.30 itself over the period of one year or less
- Cyclical fluctuations is a data pattern
that repeats itself... may take years
- Irregular variations are jumps in the level
of the series due to extraordinary events
- Random fluctuation from random
variation or unexplained causes
- Representative Historical Data Set
- Compute the Seasonal Indexes
Seasonal Patterns
- De seasonalize the Data
Seasonalized Time Series Regression Analysis
- Select a representative historical data - Perform Regression on Deseasonalized
set. Data
- Develop a seasonal index for each
season.
- Use the seasonal indexes to de
seasonalize the data.
- Perform linear regression analysis on
the de seasonalized data.
- Use the regression equation to compute
the forecasts.
- Use the seasonal indexes to reapply the
seasonal patterns to the forecasts
Example: Computer Products Corp.
Seasonalized Times Series Regression Analysis
An analyst at CPC wants to develop next
year’s quarterly forecasts of sales
- Compute the Deseasonalized Forecasts
revenue for CPC’s line of Epsilon
Computers. She believes that the most Y9 = 8.357 + 0.199(9) = 10.148
recent 8 quarters of sales (shown on the Y10 = 8.357 + 0.199(10) = 10.347
next slide) are representative of next Y11 = 8.357 + 0.199(11) = 10.546
year’s sales. Y12 = 8.357 + 0.199(12) = 10.745
Note: Average sales are expected to increase
- Seasonalized Times Series Regression by .199 million (about $200,000) per quarter
Analysis
- Seasonalize the Forecasts
Short-Range Forecasts
- Time spans ranging from a few days to a
few weeks
- Cycles, seasonality, and trend may have
little effect
- Random fluctuation is main data
component
Short-Range Forecasting Methods
- (Simple) Moving Average
- Weighted Moving Average
- Exponential Smoothing Weighted Moving Average
Simple Moving Average - This is a variation on the simple moving
average where the weights used to
- An averaging period (AP)is given or
compute the average are not equal.
selected
- This allows more recent demand data to
- The forecast for the next period is the
have a greater effect on the moving
arithmetic average of the AP most
average, therefore the forecast.
recent actual demands
- The weights must add to 1.0 and
- It is called a “simple” average because
generally decrease in value with the age
each period used to compute the
of the data.
average is equally weighted
- The distribution of the weights
- It is called “moving” because as new
determines the impulse response of the
demand data becomes available, the
forecast
oldest data is not used
- By increasing the AP, the forecast is less
responsive to fluctuations in demand
(low impulse response and high noise
dampening)
- By decreasing the AP, the forecast is
more responsive to fluctuations in
demand (high impulse response and low
noise dampening)
Barometric Methods
- National Bureau of Economic Research
- Department of Commerce
- Leading Indicators
- Lagging Indicators
- Coincident Indicators
Exponential Smoothing - Composite Index
- The weights used to compute the - Diffusion Index
forecast (moving average) are
exponentially distributed.
- The forecast is the sum of the old
forecast and a portion (w) of the forecast
error
Ft+1 = wA1 + (1-w)Ft
- The smoothing constant, w, must be
between 0.0 and 1.0.
- A large w provides a high impulse
response forecast.
- A small w provides a low impulse
response forecast.
- If the need is to forecast sales of a new
product, then a customer survey may
not be practical; instead, historical
analogy or market research may have to
be used
Time Span
- What operations resource is being
forecast and for what purpose?
- Short-term staffing needs might best be
forecast with moving average or
exponential smoothing models.
Criteria for Selecting a Forecasting Method - Long-term factory capacity needs might
best be predicted with regression or
- Cost executive-committee consensus
- Accuracy methods.
- Data available
- Time span Nature of Products and Services
- Nature of products and services - Is the product/service high cost or high
- Impulse response and noise dampening volume?
Cost and Accuracy - Where is the product/service in its life
cycle?
- There is a trade-off between cost and - Does the product/service have seasonal
accuracy; generally, more forecast demand fluctuations?
accuracy can be obtained at a cost.
- High-accuracy approaches have Impulse Response and Noise Dampening
disadvantages: - An appropriate balance must be
o Use more data achieved between:
o Data are ordinarily more - How responsive we want the
difficult to obtain forecasting model to be to changes in
o The models are more costly to the actual demand data
design, implement, and operate - Our desire to suppress undesirable
o Take longer to use chance variation or noise in the demand
- Low/Moderate-Cost Approaches– data
statistical models, historical analogies,
executive-committee consensus Reasons for Ineffective Forecasting
- High-Cost Approaches–complex - Not involving a broad cross section of
econometric models, Delphi, and people
market research - Not recognizing that forecasting is
Data Available integral to business planning
- Not recognizing that forecasts will
- Is the necessary data available or can it always be wrong
be economically obtained? - Not forecasting the right things
- Not selecting an appropriate forecasting
method
- Not tracking the accuracy of the
forecasting models