Supply Chain

Demand – Types

2011

Page
 Table of Contents

Dependent versus independent demand.......................................................................................................................1

Examples:.................................................................................................................................................................... 1
Components of Demand.................................................................................................................................................2
Forecast Horizon:............................................................................................................................................................3
Forecasting by Projection Methods:...............................................................................................................................4
Simple Moving Average:.............................................................................................................................................4
Weighted Moving Average:........................................................................................................................................4
Exponential Smoothing - Single:.................................................................................................................................4
Exercise-1: In Session.................................................................................................................................................5
Exercise-2: Take-Home..............................................................................................................................................5
Forecasting by Analytical Methods.................................................................................................................................6
Economics Model:.......................................................................................................................................................6
Input-Output Model:..................................................................................................................................................6
Exponential Smoothing:..............................................................................................................................................6
Forecasting by Qualitative Methods...............................................................................................................................7
Marketing Methods:...................................................................................................................................................7
Futures:.......................................................................................................................................................................7
Simulation Methods:...................................................................................................................................................7
Tracking Signal:...............................................................................................................................................................9
Case Study:................................................................................................................................................................... 10
Annexure-I: Product Planners’ Responses:...............................................................................................................11
Recap of Discussion Points............................................................................................................................................12

Page
Demand – Types
Dependent versus independent demand
The choice of how to control inventory depends upon the fact of whether or not the demand of the items has a
dependent or an independent character. This difference is important in selecting an adequate inventory
management approach.

Dependent demand exists for those items for which demand is linked to the use of other items (for example,
subassemblies which go into a higher-level component or finished item, such as the motherboard for a computer).
The essential aspect of dependent demand is that inventory items can be calculated.

Independent demand exists where the rate of use for an item does not relate directly to the use of another item
(for example, finished goods such as a computer).

It is obvious that most organizations have items which fall into each category, namely the category of dependent
and independent demand. If this is the case, then the organization will have to use both systems, because it would
be a mistake to suppose that you can use a dependent system for an independent item and, the other way around,
to use an independent system for a dependent item.

Examples:
 Independent: PCs.
 Dependent: Micro Processors, Key Boards, Cables.

 Independent: Cars.
 Dependent: Batteries, Tires, Wind Screens, Indicators, Engines.

 Independent: Food Dish.

 Dependent: Salt, Paste, Water, Vegetable, Oil.
 Demand – Components

Components of Demand
There are four basic components of Demand:
1. Trend.
2. Jump or Random Variation.
3. Periodicity or Seasonality.
4. Randomness.

Trend is a unidirectional gradual change (increasing or decreasing) in the average value of the variable. Changes in
nature caused by human activities are the main reason for the over-several-years trends. Trends are usually
smooth, and we should be able to represent it by a continuous and differentiable function of time. Trend is usually
considered to be deterministic and it can be modeled by linear or polynomial functions:
 Linear function,
 Polynomial function,
 Power functions.

Jump is a sudden positive or negative change in the observed values. Human activities and natural disruptions are
the main reasons for jumps in time series.

Periodicity represents cyclic variations in a time series. These variations are repetitive over fixed intervals of time.

Randomness represents variations due to the uncertain nature of the stochastic process. The random component of
time series can be classified as autoregressive or purely random.

Trend Periodicity

Jump Randomnes

Page
 Forecast Horizon

Forecast Horizon:

Forecast horizon varies by organization, time, and situation. It largely depends on the following factors:

1. Objective of Forecasting.
2. Availability of Resources.
3. Ability to Forecast.
4. Relevance of Information.

Forecast can be made for Short, Medium or Long Term, where these terms are defined by the organizational scope
of operations, its life cycle stage and the industry situation. Generally speaking, following are the domain of forecast
with respect to different time horizons.

Horizon Time Span Items

Long Years Product lines
Factory capacities
Planning for new products
Capital expenditures
Facility location or expansion
R&D
Medium Months Product groups
Department capacities
Sales planning
Production planning and budgeting
Short Weeks / Days Specific product quantities
Machine capacities
Planning
Purchasing
Scheduling
Workforce levels
Production levels
Job assignments
 Forecasting by Projection Methods

Forecasting by Projection Methods:

Methods of this type are concerned with a variable that changes with time and which can be said to depend only
upon the current time and the previous values that it took (i.e. not dependent on any other variables or external
factors). If Yt is the value of the variable at time t then the equation for Yt is
Y T =∫ ( Y t −1 , Y t−2 ,Y t−3 , …. , Y 0 , t )
i.e. the value of the variable at time t is purely some function of its previous values and time, no other
variables/factors are of relevance. The purpose of time series analysis is to discover the nature of the function f and
hence allow us to forecast values for Y t.
Time series methods are especially good for short-term forecasting where, within reason, the past behavior of a
particular variable is a good indicator of its future behavior, at least in the short-term. The typical example here is
short-term demand forecasting. Note the difference between demand and sales - demand is what customers want -
sales is what we sell, and the two may be different.

Simple Moving Average:

One, very simple, method for time series forecasting is to take a moving average (also known as weighted moving
average). The moving average over the last ‘n’ periods ending in period t is calculated by taking the average of the
values of last ‘n’ periods:
At −1 + At −2+ A t−3 +…+ A t−n
F t=
n

Weighted Moving Average:

One disadvantage of using moving averages for forecasting is that in calculating the average all the observations are
given equal weight (namely 1/L), whereas we would expect the more recent observations to be a better indicator of
the future (and accordingly ought to be given greater weight).

n
w 1 At −1+ w2 A t−2 + w3 A t−3 +…+ wn A t−n
F t=
W
∑ W i=1
i=1

Exponential Smoothing - Single:

Exponential Smoothing is a Statistical Technique used for detecting significant changes in data. In exponential
smoothing (as opposed to in moving averages smoothing) older data is given progressively-less relative weight
(importance) whereas newer data is given progressively-greater weight. It is employed in making short-term
forecasts.
F t=F t−1 +∝ ( A t−1−F t−1 )

Multiple Exponential Smoothing can also be used to incorporate more than just-last period of Demand. The multiple
exponential smoothing method requires more multiple smoothing variables.

F t=F t−1 +∝ ( A t−1−F t−1 ) + β ( A t −2−Ft −2 )+ …+∁ ( A n−F n )

Forecasting by Projection - Exercise

Exercise-1: In Session

Exercise-2: Take-Home
 Forecasting by Analytics

Forecasting by Analytical Methods

Economics Model:
The use of this approach bases demand forecasting on certain economic indicators, eg. Construction contracts
sanctioned for the demand of building materials, say, cement; Personal income for the demand of consumer goods;
Agricultural income for the demand of agricultural inputs, implements, fertilizers, etc.; and Automobile registration
for the demand of car accessories, petrol, etc.

Steps for economic indicators:

1. See whether a relationship exists between the demand for the product and certain economic indicators.
2. Establish the relationship through the method of least squares and derive the regression equation. (Y= a + bx)
3. Once regression equation is derived, the value of Y (demand) can be estimated for any given value of x.
4. Past relationships may not recur. Hence, need for value judgement.

Limitations:
1. Finding an appropriate economic indicator may be difficult.
2. For new products – no past data exists.
3. Works best when the relationship of demand with a particular indicator is characterized by a time lag. Eg.,
construction contracts will result in a demand for building materials but with a certain amount of time lag.

Input-Output Model:

The Input-Output methods is essentially based on ‘The theory of Input-Output analysis’ as an area of econometrics.
The input-output table is a tabular compilation of all monetary transactions among industries for one year.

Mathematically it can be expressed as

X11 + X12 + X13 + … + X1n + Y1 = X1
X21 + X22 + X23 + … + X2n + Y2 = X2
...
Xn1 + Xn2 + Xn3 + … + Xnn + Yn = Xn

where there are n sectors in the economy,

Xij is the sale from sector i to sector j,
Yi is the final demand for the product of sector i,
Xi is the total output of sector i,
and i and j can take values between 1 and n.

Exponential Smoothing:
Exponential smoothing method smooth-outs the coming forecast in reaction to previous period’s deviation. The
smoothing constant ∝ is used to control the speed of adjustment. ∝ usually ranges from 0 to 1 but can be any
number.
F t=F t−1 +∝ ( A t−1−F t−1 )
Multiple values of ∝ can be taken for incorporating multiple previous periods.
F t=F t−1 +( ∝¿ ¿ 1 ( At −1−Ft −1) )+(∝¿ ¿ 2 ( At −2−F t−2 ) )¿ ¿
 Forecasting by Qualitative Methods

Forecasting by Qualitative Methods

Qualitative methods are usually based on judgments about causal factors that underlie the demand of particular
products or services. They do not require a demand history for the product or service, therefore are useful for new
products/services. These approaches vary in sophistication from scientifically conducted surveys to intuitive hunches
about future events. Methods of this type are primarily used in situations where there is judged to be no relevant
past data (numbers) on which a forecast can be based and typically concern long-term forecasting. One approach of
this kind is the Delphi technique.

Marketing Methods:

Delphi method: it consists of an effort to arrive at a consensus in an uncertain area by questioning a group of
experts repeatedly until the results appear to converge along a single line of the issues causing disagreement are
clearly defined. Developed by Rand Corporation of the U.S.A in 1940s by Olaf Helmer, Dalkey and Gordon. Useful in
technological forecasting (non-economic variables).

Survey of buyers’ intentions: also known as Opinion surveys. Useful when customers are industrial producers. Not a
very useful tool for household consumers. Passive and “does not expose and measure the variables under
management’s control”

Expert Opinion Method: To ask “experts in the field” to provide estimates, eg., dealers, industry analysts, specialist
marketing consultants, etc. It is very simple and quick method and have no danger of a “group-think” mentality.

Collective Opinion Method: Also called “sales force polling”, salesmen are required to estimate expected sales in
their respective territories and sections. It is simple, based on first-hand knowledge, quite useful in forecasting sales
of new products, yet has some disadvantages. It is most completely subjective and its usefulness is restricted to
short-term forecasting where salesmen may be unaware of broader economic changes.

Naïve forecasting models: are based exclusively on historical observation of sales (or other variables such as
earnings, cash flows, etc). They do not explain the underlying casual relationships which produces the variable being
forecast. This is inexpensive to develop, store data and operate, but does not consider any possible causal
relationships that underlie the forecasted variable.

Futures:
Futurology is the study of postulating possible, probable, and preferable futures and the worldviews and myths that
underlie them. It is considered as a topic in philosophy. Futures studies seeks to understand what is likely to
continue, what is likely to change, and what is novel. Part of the discipline thus seeks a systematic and pattern-based
understanding of past and present, and to determine the likelihood of future events and trends. Unlike Science
where a narrower, more specified system is studied, futurology studies a much bigger and complex world system.

Simulation Methods:
Under this method, an effort is made to vary separately certain determinants of demand which can be manipulated,
(price, advertising, etc.) and conduct the experiments assuming that the other factors remain constant. These
experiments are expensive as well as time consuming and carry the risk of leading to unfavorable reaction on
dealers, consumers, competitors.
 Forecasting Accuracy

Short-range forecasting models are evaluated on the basis of three characteristics:

1. Impulse response.
2. Noise-dampening ability.
3. Accuracy:
 Error & Bias.
 Mean Percent Error – MPE.
 Mean Absolute Percent Error – MAPE.
 Tracking Signal.

Forecasts that respond quickly to changes in data are said to have a high impulse response. A forecast system that
responds quickly to data changes necessarily picks up a great deal of random fluctuation (noise). If forecasts have
little period-to-period fluctuation, they are said to be noise dampening. Hence, there is a trade-off between high
impulse response and high noise dampening.

Accuracy is the typical criterion for judging the performance of a forecasting approach. Accuracy is how well the
forecasted values match the actual values. Accuracy of a forecasting approach needs to be monitored to assess the
confidence you can have in its forecasts and changes in the market may require reevaluation.

Accuracy can be measured in several ways

 Error & Bias of the forecast.
 Mean Percent Error - MPE
 Mean Absolute Percent Error – MAPE
 Tracking Signal – TS.

Some common formulae are as under:

Percent Error
A t−F t
E=
At

Mean Percent Error - MPE

n
1 At −F t
M= ∑
n t =1 A t

Mean Absolute Percent Error - MAPE

n
1 A t−F t
M= ∑ | |
n t =1 At

Tracking Signal - TS
 Forecasting Accuracy - Exercise

∑ ( A i−F i ) Tracking Signal:

TS= i=1
MAD
Tracking Signal is a measure that indicates whether the forecast average is keeping pace with any genuine upward or
downward changes in demand. The TS measures the cumulative forecast error over n periods in terms of MAD.

Example:

Exercise:

Perfect-Pharma has implemented a linear regression method to forecast sales. Actual sales for the months of
January 2009 through January 2010 are given in Table below along with their corresponding forecasts. Perfect-
Pharma would like to employ a tracking signal to measure the performance of its forecasting method.
Case Study on Demand Forecasting

Case Study:
Mr. Ali has recently joined a fast paced growing pharmaceutical company as a “Demand Planner” for a broad range
of products. He is responsible for making consolidated demand forecast on monthly basis for an organization that
used to have separate demand-planners for each product-line using his own formula / tools & techniques.

In his previous employment, he was considered a person with strong analytical skills in demand forecasting, he
looked after several beverage product lines, planning about their future demands. His formula for demand
forecasting proved to be a success as it produced close-to-accurate results. He also had a very supportive team
working under his supervision that used to maintain the data in a state-of-the-art demand forecasting software. The
software was capable of handling a large variety of SKUs on the basis of customized forecasting formula set by the
Demand Planner.

Mr. Ali, through the software, was able to forecast demand and supply needs for beverages with a single customized
formula, as the company was dealing in cold beverages only.

In his new position at Perfect Pharma, he feels overwhelmed by the complexity and variation found in the industry.
There are primarily four types of product lines to forecast for:
1. Products having stable demand and consistent in trends – generally on-the-counter medicines.
2. Products showing seasonal shifting and variation – like cough syrup, etc.
3. Products that are only for specific need (Cancer, etc.) & emergency cases (Drip, Blood-Bag, etc).

Mr. Ali has only one week to develop a forecast for all product ranges of the company. being the Demand-Planner,
he has called an urgent meeting to determine the single basis of forecasting for all the product lines.

As the meeting commenced, Mr. Ali has a very clear agenda in his mind – ‘Decision on the basis of forecasting’. He
was sure all product-planners would agree on simplicity and this would also bring efficiency in the system while
removing extra workload from the product-planners.

He put forward a simple question, to start with, before his product-planners on how they forecast for their product
lines. It was an open forum, so each product-planner shared his views and argued in it favor. Meeting went on and
the team had a heated discussion and lot of idea sharing. Each product-planner presented before the Demand-
Planner sound rationale for his own formula. He started noting down all the suggested formulae on a piece of paper
(see Annexure-I for details).

Till the half time, Mr. Ali was even more confused than before. He quickly adjourned the meeting and moved back to
his office. As he approached his desk he saw a card in front of him that he didn’t notice before.

This was ‘your’ visiting card; you closely worked with him in his last organization as a supply-chain consultant. You
happened to be an expert in Demand Planning. In desperation to meet his deadline, he picked up your visiting card
and called you to help him come out of the dilemma he was facing.

You agreed to help him find out the best mechanism for the demand planning in an organization with such a wide
variety of product line.

You promised him to share your views very soon.

 Case Study on Demand Forecasting

Annexure-I: Product Planners’ Responses:

1. Mr. Arshad’s response was a simple one, and apparently valid too – “It should be simply the average of last 6 months”.

At −1+ A t−2 + A t −3 + …+ A t −n A 7 + A 6 + A 5 + A 4 + A 3+ A 2
F t= F 8=
n 6

2. Mr. Ghazanfar partially agreed with Mr. Arshad’s view, but he insisted to keep the period up to just three months only,
instead of taking six months, and also to assign weights to the past month. He argued that industry has seen many changes
in the last quarter and going back further is of no use. “It should be the weighted average of last 3 months. Highest weight
(0.5) should be given to the immediate past month, followed by 0.3 and 0.2 for the other two months”.

W 1 A t−1 +W 2 A t −2+W 3 At −3+ …+W n At −n W 1 A7 +W 2 A6 +W 3 A5

F t= F8 =
W W

3. Mr. Sajjad used to take the last year values to form his forecast. He suggested “the next month’s forecast should equal the
actual demand of last year comparable month”.

F t=La st Year Actualt F08 /2010 = Actual08/ 2009

4. Mr. Atiq based his argument on growth pattern. He suggested that “the growth trend of last-year-next-month (Y LMN) over
last-year-current-month (YLMC) will help us find the forecast value for the current-year-next-month (Y CMC) using current-
year-current-month (YCMC)”.

Last Year Actualt Last Year Actual8

F t= ( Last Year Actualt−1 )
× ( Current Year Actualt−1 ) F8 =(
Last Year Actual7
) ×(Current Year Actual7 )

5. Mr. Moin argued that taking growth is the right approach, as mentioned by Mr. Atiq, however, he suggested that only the
“Cumulative Year-To-Month growth over last year will be the right basis for forecasting. Apply the Cumulative growth
percentage over last-year-next-month (YLMN) to find the current-year-next-month forecast value (Y CMN)”.

Cumulati ve Sales Current Year Upto Montht −1

F t=( )×( Last Year Montht )
Cumulative Sales Last Year Upto Montht −1

Cumulative Sales Current Year Upto Month7

F 8=( )×(Last Year Month8 )
Cum ulative Sales Last Year Upto Month7

6. Mr. Saleem presented the most complex method of forecasting which appeared hard to understand but had some merit.
He argued that one must tale last-3-month-average and compare it with last-6-mont-average. If the ratio between smaller-
period to longer-period is greater than 2, start decreasing your forecast by a certain degree, as listed in a table with him. If
the ratio between shorter-period to the longer-period is lesser than 0.5, take reverse action – start increasing the forecast
using the same very table.

If Actual ∑ 3 ❑ ¿ ¿ ≥2.0 , F t=F t−1 ×(1−0.2)

Actual
t=1 ¿
∑
❑ 6
t=1¿

If Actual ∑ 3 ❑ ¿ ¿ ≤ 0.5 , F t=F t −1 ×(1+ 0.2)

Actual
t=1 ¿
∑6
❑
t=1¿
Recap

Recap of Discussion Points