SUPPLY CHAIN FORECASTING - QUANTS

LEARNING OBJECTIVES
1. Understand how forecasting is essential to supply chain
planning.
2. Evaluate demand using quantitative forecasting models.
3. Apply qualitative techniques to forecast demand.
4. Apply collaborative techniques to forecast demand.
ROLE OF FORECASTING IN STARBUCKS
• Starbucks, the largest coffee chain in the world with over 30000 stores in 80 countries.
• Variety of products beyond coffee, coffee beans, salads, sandwiches, mugs, etc.
• Product offerings varies by season and some are store location-specific.
• Many are perishable, some runs the risk of becoming obsolete.
• Starbucks branded coffee and ice cream are sold in grocery stores.
• Need for forecasting regional and global demand as well as store-specific or demand.

• Far-flung Supply chain at Starbucks.

• The company brings raw beans from Latin America, Africa and Asia in giant shipping containers to USA and
European 6 storage locations near roasting facilities .
• Roasted coffee is packaged and shipped to distribution centers (DCs) 4 in USA, 2 in Europe and 2 in Asia).
• DCs also store many more items that are sold through Starbucks besides coffee.

• Various types of forecasting (global, regional, store-specific) will help make decisions as to the quantity to be
bought and shipped and processed through various steps in the supply chain.
• Analytical exercise on Forecasting SC demand –Starbucks corp. gives a numerical illustration of this interesting
forecasting challenge.
FORECASTING IN OPERATIONS AND SUPPLY CHAIN MANAGEMENT

• Forecasting is vital to every business organization and impacts every significant

management decision.
• Forecasting is the basis of corporate planning and control .
• Finance and accounting use forecasts as the basis for budgeting and cost control.
• Marketing relies on forecasts to make key decisions such as new product planning
and personnel compensation.
• Production uses forecasts to select suppliers, determine capacity requirements, and
to drive decisions about purchasing, staffing, and inventory.

• Different activities require different forecasting approaches.

• Decisions about overall directions require strategic forecasts.
• Tactical forecasts are used to guide day-to-day decisions and the goal is to estimate
demand in short term.
 TYPES OF FORECASTING
• Basic types of forecasts:
• Qualitative.
• Quantitative.

• Time series analysis (primary focus of this chapter).

• Causal relationships.
• Simulation.

• Time series analysis is based on the idea that data relating

to past demand can be used to predict future demand.
 COMPONENTS OF DEMAND
• Average demand for the period

• Trend

• Seasonal element

• Cyclical elements

• Random variation

• Autocorrelation
TIME SERIES FORECASTING OVERVIEW
• Trend line is the usual starting point of the time series
forecasting.
• Trend line is adjusted for
• Seasonal effects.
• Cyclical elements.
• Any other expected elements that may influence the final forecast.

• Estimation of autocorrelation allows us the improve forecasting

accuracy.
• Estimation of random effects allows us to give a range for the
forecast.
COMPONENTS OF DEMAND – GROWTH &
SEASONAL
• Exhibit 3.1

• Access the text alternative for slide images.

 TRENDS
• Identification of trend lines is a common starting point when
developing a forecast.
• Common trend types include linear, S-curve, asymptotic, and
exponential.

• Exhibit 3.2

• Access the text alternative for slide images.

 TIME SERIES ANALYSIS
• Using the past to predict the future.

Short term – forecasting less than three months

• Used mainly for tactical decisions (for example, replenishing inventory).
Medium term – forecasting three months to two years
• Used to develop a strategy which will be implemented over the next six to
eighteen months (for example, meeting demand).
Long term – forecasting greater than two years
• Useful for detecting general trends and identifying major turning points.
 MODEL SELECTION
• Choosing an appropriate forecasting model depends upon:
1. Time horizon to forecast.
2. Data availability.
3. Accuracy required.
4. Size of forecasting budget.
5. Availability of qualified personnel.

• Other factors may also be considered.

1. Degree of flexibility (can the firm react quickly if the forecast is
inaccurate?).
2. Consequence of a bad forecast (important or costly decisions
require a good forecast).
FORECASTING METHOD SELECTION GUIDE
Amount of Historical
Forecasting Method Data Data Pattern Forecast Horizon
Simple moving average 6 to 12 months; weekly Stationary (i.e. no Short
data are often used trend or seasonality)
Weighted moving 5 to 10 observations Stationary Short
average and simple needed to start
exponential smoothing
Exponential smoothing 5 to 10 observations Stationary and trend Short
with trend needed to start
Linear regression 10 to 20 observations Stationary, trend, and Short to Medium
seasonality
Trend and seasonal 2 to 3 observations per Stationary, trend and Short to Medium
models season seasonality
 SIMPLE MOVING AVERAGE
• Forecast is based on average demand over the most recent periods
• 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.
• Longer periods provide more smoothing.
• Shorter periods react to trends more quickly.

• - forecast for the coming period t.

• – number of periods to be averaged.
• - actual occurrence in the past period.
• , , and - actual occurrences two periods ago, three periods ago, and so on up to n
periods ago.
 SIMPLE MOVING AVERAGE – EXAMPLE (P.
52 TO 54)
• Note that no forecast is possible until “n” periods have passed
• Exhibit 3.4
Week Demand 3 -Week 9 -Week Week Demand 3-Week 9-Week
1 800 16 1,700 2,200 1,811
2 1,400 17 1,800 2,000 1,800
3 1,000 18 2,200 1,833 1,811
4 1,500 1,067 19 1,900 1,900 1,911
5 1,500 1,300 20 2,400 1,967 1,933
6 1,300 1,333 21 2,400 2,167 2,011
7 1,800 1,433 22 2,600 2,233 2,111
8 1,700 1,533 23 2,000 2,467 2,144
9 1,300 1,600 24 2,500 2,333 2,111
10 1,700 1,600 1,367 25 2,600 2,367 2,167
11 1,700 1,567 1,467 26 2,200 2,367 2,267
12 1,500 1,567 1,500 27 2,200 2,433 2,311
13 2,300 1,633 1,556 28 2,500 2,333 2,311
14 2,300 1,833 1,644 29 2,400 2,300 2,378
15 2,000 2,033 1,733 30 2,100 2,367 2,378

• Access the text alternative for slide images.

 WEIGHTED MOVING AVERAGE
• The simple moving average
formula implies all periods
are equally important.
• W1 – weight to be given to the
• A weighted moving average
allows unequal weighting of actual occurrence for the period
prior time periods. t – 1.

• The sum of the weights • W2 – weight to be given to the

must be equal to one. actual occurrence for the period
t − 2.
• More recent data
(periods) are given more • W1 – weight to be given to the
significance (higher actual occurrence for the period
weights) than older data. t − n.
• n – total number of prior
periods in the forecast.
 CHOOSING WEIGHTS
• Experience and trial and error are the simplest approaches.
• The most recent past is the most important indicator of what to
expect in the future, so weights are generally higher for more
recent data.
• If the data are seasonal, weights should be established
appropriately.
• The weighted moving average has an advantage over the simple
moving average.
 EXPONENTIAL SMOOTHING
• The importance of data diminishes as the past becomes more distant.
• The most logical and easiest method to use.
• An integral part of all computerized forecasting programs.
• Well accepted for six major reasons:
1. Exponential models are surprisingly accurate.
2. Formulating an exponential model is relatively easy.
3. The user can understand how the model works.
4. Little computation is required to use the model.
5. Computer storage requirements are small.
6. Tests for accuracy are easy to compute.
 EXPONENTIAL SMOOTHING MODEL

Ft  Ft 1   ( At 1  Ft 1 )

• Ft • -forecast value for time period t.

• F(t 1) • -forecast value for the previous time period t − 1.

• A(t 1) • -actual occurrence for time period t − 1.

•  • -alpha, the smoothing constant.

EXPONENTIAL SMOOTHING EXAMPLE

Week Demand Forecast

F2  F1   ( A1  F1 )  820  0.2(820  820)
1 820 820
2 775 820 F4  F3   ( A3  F3 )  811  0.2(680  811)
3 680 811
4 655 785 F6  F5   ( A5  F5 )  759  0.2(750  759)
5 750 759
6 802 757 F8  F7   ( A7  F7 )  766  0.2(798  766)
7 798 766
8 689 772 F10  F9   ( A9  F9 )  756  0.2(775  756)
9 775 756
10 760
EXPONENTIAL SMOOTHING WITH TREND
• The presence of a trend in the data causes the exponential
smoothing forecast to always lag behind the actual occurrence.
• This can be corrected by adding a trend adjustment.
• The trend smoothing constant is delta

( ).

• Both alpha and delta reduce the impact of the error that occurs
between the actual and the forecast.
 CHOOSING ALPHA AND DELTA

• Relatively small values for 𝛼 and 𝛿 are common.

• Usually in the range 0.1 to 0.3.

• 𝛼 depends upon how much random variation is present.

• 𝛿 depends upon how steady the trend is.
• Measurements of forecast error can be used to select values
of 𝛼 and 𝛿 to minimize overall forecast error.
 EXERCISE
• Given the following history, use a three-quarter moving average to forecast the demand for the third
quarter of this year. Note that the 1st quarter is Jan, Feb, and Mar; 2nd quarter Apr, May, Jun; 3rd quarter
Jul, Aug, Sep; and 4th quarter Oct, Nov, Dec..
 EXERCISE
• Here are the data for the past 21 months for actual sales of an ev battery for e bikes:

Develop a forecast for the fourth quarter using a three-quarter, weighted moving average. Weight the most recent
quarter .5, the second most recent .25, and the third .25. Solve the problem using quarters, as opposed to
forecasting separate months.
 EXERCISE
• The following table contains the number of complaints received in a department store for the first 6 months of
operation. If a three-month moving average is used to smooth this series, what would have been the forecast for May?
 EXERCISE
• The following tabulations are actual sales of units for six months and a starting forecast in January.
a. Calculate forecasts for the remaining five months using simple exponential smoothing with a = 0.2.

• Ft+1 = Ft + (At – Ft),  = .20

 LINEAR REGRESSION ANALYSIS
• Regression - the functional relationship
between two or more correlated Yt  a  bt
variables, usually from observed data.
• One variable (the dependent variable) is
predicted for given values of the other
variable (the independent variable).
• Linear regression is a special case which
assumes the relationship between the
variables can be explained with a
straight line.
 FORECAST ERRORS
• Forecast error - the difference between the forecast value and what
actually occurred.
• All forecasts contain some level of error.
• Sources of error.
• Bias – when a consistent mistake is made.
• Random – errors that are not explained by the model being used.

• Measures of error.
• Mean absolute deviation (MAD).
• Mean absolute percent error (MAPE).
• Tracking signal.
 MEASUREMENT OF ERROR
• Ideally, MAD will be zero (no
• MAPE scales the forecast error to
forecasting error).
the magnitude of demand.
• Larger values of MAD
indicate a less accurate model. MAD
MAPE 
Average Demand

• Tracking signal indicates whether

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

Running sum of forecast errors

TS 
Average mean absolute deviation
 QUALITATIVE FORECASTING TECHNIQUES
• Generally used to take advantage of expert knowledge.
• Useful when judgment is required, when products are new, or if the firm has little
experience in a new market.
• Examples of techniques:
• Market research.
• Panel consensus.
• Historical analogy.
• Delphi method.
 COMPUTING FORECAST ERROR – EXAMPLE (P. 69)

• Exhibit 3.12

Month Forecast Actual Deviation RSFE Abs. Dev. Sum of Abs. Dev. MAD TS
1 1,000 950 −50 −50 50 50 50 −1
2 1,000 1,070 +70 +20 70 120 60 0.33
3 1,000 1,100 +100 +120 100 220 73.3 1.64
4 1,000 960 −40 +80 40 260 65 1.2
5 1,000 1,090 +90 +170 90 350 70 2.4
6 1,000 1,050 +50 +220 50 400 66.7 3.3
Overall
C O LL A B O R AT IV E PLA N N IN G , FO R EC A STIN G , A N D R E PL EN IS H M E N T (C P F R )

• An Internet tool used to coordinate the efforts of a supply chain.

• Demand forecasting.
• Production and purchasing.
• Inventory replenishment.

• Integrates all members of a supply chain – manufacturers,

distributors, and retailers.
• Depends upon the exchange of internal information to provide a
more reliable view of demand.
N-TIER SUPPLY CHAIN

Access the text alternative for slide images.

 C P F R STEPS
• Step 1
• Creation of a front-end partnership agreement.
• Step 2
• Joint business planning.
• Step 3
• Development of demand forecasts.
• Step 4
• Sharing forecasts.
• Step 5
• Inventory replenishment.
 EXERCISE
• Zeus Computer Chips, Inc., used to have major contracts to produce the Centrino-type chips. The market
has been declining during the past three years because of the quad-core chips, which it cannot produce, so
Zeus has the unpleasant task of forecasting next year. The task is unpleasant because the firm has not been
able to find replacement chips for its product lines. Here is demand over the past 12 quarters:
• Use the regression and seasonal indexes to forecast demand for the next four quarters.
 EXERCISE
• The sales data for two years are as follows. Data are aggregated with two months of sales in each “period.”

a. Plot the data.

b. Fit a simple linear regression model to the sales data.
c. In addition to the regression model, determine multiplicative seasonal indexes. A full cycle is assumed to be a
full year.
d. Using the results from parts (b) and (c), prepare a forecast for the next year.
 EXERCISE
• Tucson Machinery, Inc. manufactures numerically controlled machines, which sell for an average price of
$0.5 million each. Sales for these NCMs for the past two years were as follows:

a. Find a line using regression in Excel.

b. Find the trend and seasonal indexes.
c. Forecast sales for next year.