Demand Management and

Forecasting
Module 2
 Order Inventory
Processing Control

Warehousing
Elements of Transportation
• Material Handling
Logistics • Packaging

Information
Packaging
and Analytics
Supply Chain Strategies
Make-to-stock (MTS)

Assemble-to-order (ATO)

Made-to-order (MTO)

Engineer-to-order (ETO)
Impact of Strategy on Lead Time
Delivery Lead Time

Design +
Manufacture Assemble Deliver Engineer-to-Order
Purchase
Delivery Lead Time

Purchase Make-to-Order
Manufacture Assemble Deliver
Raw Matl.
Delivery Lead Time
Manufacture
Purchase Component Assemble Deliver Assemble-to-Order
s
Delivery Lead Time

Purchase Manufacture Assemble Deliver Make-to-Stock

Make to Stock
Assemble to
Order
 • Most items may be procured after the order is
Make to obtained
• Forecasting of end-items is not very critical
Order • Sometimes, raw material requirements or imported
items are forecast
Engineer to
Order
Thought – what is the
difference between

Forecasting

Demand
Management
 Demand Management

Demand Management is the

logistics function that Demand Management is not
Demand is usually managed
attempts to synchronise just about measuring
by the Sales and Operations
supply (what we can demand; it is about
Planning process
make/ship) with demand influencing it
(what customers want/order)
Importance of Demand Management

Operational Customer
Financial Health
Stability Satisfaction
• Reduces lost • Reduces the • Ensures that the
sales (stockouts) need for costly logistics system
and prevents emergency can deliver the
obsolescence logistics (e.g., time and
(excess premium freight) dependability
inventory) and excessive elements of
overtime customer service
Factors Influencing Demand

Nature of product

Business, social and economic conditions prevailing

Competitive factors

Government regulation

Organization business goals and strategies

Sources of Demand

Customer orders (and quotations) for end products

Sales promotions and marketing campaigns

Spare part demands for components and assemblies

Other demands such as product testing, trial and demonstration pieces

Characteristics
of Demand
• Demand typically varies for
most products Apparel

• The charts show the demand

for different products

Air-conditioners
Demand
Patterns
• Stable demand tends to
retain the same shape
over a period of time
• Dynamic demand tends
to be erratic
• Demand can also be
lumpy and intermittent
 Forecasting vs Demand Management

Comparison Point Forecasting Demand Management

To influence and synchronise actual

To predict future customer demand based
Primary Objective customer demand with supply chain
on historical data and statistical modelling
capacity

An active strategic process. It answers the

A passive analytical tool - it answers the
Action / Stance question: "How can we make this happen,
question: "What is likely to happen?"
or how can we smooth the requirement?"

Statistical analysis, selection of models Strategic pricing, sales promotions, setting

Key Activities (MA, ES), calculation of error metrics lead-time promises, and capacity
(MAD, MAPE) allocation

Requires cross-functional collaboration,

Primarily driven by the Supply Chain and
Organisational Scope especially with Marketing, Sales, and
Planning functions
Finance
 How does forecasting help?

01 02 03 04
Forecasting can help Forecasting can help Forecasting can help Forecasting can help
reduce investment in increase equipment decide the right estimate
inventory uptime amount of skilled • Cash and fund requirements
people to be hired • Warehouse space
requirements
• Transportation requirements
Forecasts required

What When Where

 Independent vs Dependent Demand

Independent demand - demand for • For example, in an automobile company, the projected sales of
end-products should be forecast various car models should be forecast

Dependent demand (demand that can • The automobile company should not forecast the demand for tyres
be derived from the demand an end- that go on the cars as they can be derived (calculated) from the
product) should not be forecast forecast for vehicles

The supplier of automotive

components should get the projected
demand from the vehicle manufacturer
rather than forecasting demand
 Independent Demand

Example
MONTH MONTH MONTH
1 2 3
Forecast 1000 1500 1600
of
Scooters
Forecast 2000 3000 3200
of Tyres

Dependent Demand
Who needs forecasts?

Marketing needs forecasts to plan promotions, product introductions and other product life-
cycle decisions

Manufacturing needs forecasts to plan capacity, factory space, tooling

Maintenance needs forecasts to plan maintenance capacity

Finance needs forecasts for budgeting and financial planning

Purchasing needs forecasts for planning long lead time materials

Types of Forecasts

Strategic supply chain

Long-term forecasts are
decisions include adding
required for top
new facilities (warehouse
management to take
/ factory), relocating
strategic planning
capacity, developing new
decisions
vendors, make vs buy
Types of Forecasts

Operational supply chain

decisions include changing
Short-term forecasts are
production schedules,
required for tactical and
using alternate materials,
operational management
subcontracting and
workforce adjustments
Demand Patterns

1 2 3
Forecasters need to Cross-sectional data are Time-series data is data
understand demand observations collected at that is recorded over
patterns in order to a single point in time (e.g. successive increments
interpret data age of all equipment in a of time (e.g. monthly,
plant) quarterly, annual sales)
Levels of Level
What is
Time Horizon
Forecasting forecast?

Business Plan / Market

* - Depending upon Years (2-5)*
Strategic Plan Direction
product life-cycle

Product
Production Plan Years (1-3)*
Families

Master
Production End-items Months (2-6)*
Schedule
Principles of Forecasting
Forecasts are usually inaccurate and should include both the expected
value of forecast and a measure of error

Long-term forecasts are usually less accurate than short-term forecasts

Aggregate forecasts are usually more accurate than individual forecasts

In general, the farther the company is from its end-customer, the greater
is the distortion of information it receives (bullwhip effect)
Types of Forecasts
• Quantitative techniques - use
data and apply statistical
models to derive a forecast
• Qualitative techniques - use
judgement of forecasters /
experts to derive a forecast
• Most organizations use a
combination of both
quantitative and qualitative
techniques
Extrinsic and Intrinsic Forecasts

Extrinsic forecasting methods are based on external

patterns from information outside the company

Intrinsic forecasting methods are based on historical

patterns of the data from the company’s own data

Qualitative and quantitative forecasts may be generated

based on either extrinsic or intrinsic forecasts
 • New customers or new business
segments
• Growth plans of major customers
Extrinsic • Competition
• Government and regulatory

Factors policies
• Economic conditions
• Environmental and green issues
• Global trends
 Product life-cycle management

Planned price changes

Changes in the sales force or sales channel

Intrinsic Supply and resource constraints
Factors Sales promotions

Advertising campaigns

Production / quality issues

 Qualitative Techniques
Based on expert opinions and are judgmental

Subjective and are used for business planning and forecasting for new
products

Usually used for medium-term and long-term forecasting

Techniques such as market research, expert opinion, salesforce composite,

focus groups, Delphi method are qualitative techniques
 Qualitative Technique obtaining expert
knowledge

Used in uncertain or complex situations

Delphi
Forecasting Harnesses collective wisdom
Technique
Valuable for long-term forecasts,
technological advancements and high
uncertainty
 Panel Selection

Anonymity
Delphi
Process Iterative Rounds

Convergence
Forecasting in a B2C Environment

Demand Typical
Customer Base Data Sources Time Horizon
Patterns Challenges
• Large, diverse • More volatile • Point-of-sale • Shorter • Managing
customer and data, social forecasting demand
base with influenced by media trends, horizons due spikes due to
varying trends, marketing to faster promotions
preferences promotions, campaigns, consumption and new
and and cycles and product
seasonality economic changing launches
factors preferences
Forecasting in a B2B Environment

Customer Demand Typical

Data Sources Time Horizon
Base Patterns Challenges
• Fewer, larger • More stable • Historical • Longer • Impact of
customers and sales data, forecasting large order
with long- predictable, customer horizons due fluctuations,
term often based orders, and to planned longer lead
contracts on industry and periodic times
contractual trends orders.
agreements
Summary
Demand management is the process of synchronizing customer demand
with supply chain capacity

The demand for a product is affected by multiple factors (both internal

and external)

Forecasting is required as, lead times for the fully manufacturing cycle
are less than lead time require by the customer (except for ETO products)

Forecasts are needed by marketing, manufacturing, maintenance,

finance and purchasing
Summary
Forecasts are usually wrong

Aggregate forecasts are more accurate than individual forecasts

Forecasts for nearer time periods tend to be more accurate than forecasts for distant time
periods
Forecasts should include an estimate of error

Forecasts may be Quantitative – based on statistical manipulation of data or Qualitative –

based on intuition and informed opinion hence often subjective
Both qualitative and quantitative forecasts may be generated based on information from
external source (extrinsic) or internal sources (intrinsic)
 What are the different types of supply chain strategies –
explain each of them with an industry example?
Compare demand management and forecasting.

What factors influence demand? Use an industry example to

illustrate your answer.
What is the difference between dependent and independent
demand and how does this impact forecasting?
Questions What are the key principles of forecasting?

Compare extrinsic and intrinsic factors that affect forecasts.

Give an example of a qualitative forecast technique and where

it may be applicable.
Compare forecasting between a B2C and B2B company using
industry examples.
Quantitative
Forecasting
Techniques
Module 2
Topics

Quantitative
Data Error
Forecasting
Collection Measurements
Techniques
 Most forecasting techniques use past data
collected to predict future trends

Data
Data collection is therefore a very important
aspect of the overall forecasting process

Collection Garbage-in, Garbage-out (GIGO)

Data collection should also record the specific

circumstances under which data is being
collected in order to avoid flawed analysis
 Many automation tools are available for data
collection

Companies use internal sources such as ERP

systems and retail outlet POS systems for data

Data
collection
Companies may also use market research
agencies to collect data

Collection E-commerce sites are also used an important

source of data collection

Companies collect large amounts of data in data

lakes and data warehouses for building
forecasting models
Considerations • Sources of Data • Amount of history
required
• Data Quality
for Data • Data Relevance
• Trends, seasonality,
Collection • Consistency of Data
anomalies, outliers
• Forecasting horizon,
• Availability of Data sample size
• Frequency of Collection • Validation
• Cost and time issues • Ethical considerations
• Order date vs. ship date • Level of aggregation
• Product units vs. financial • Customer partnering
units
 Types of Forecasting Models

Judgmental Models (Qualitative forecasts)

Time-Series Models

Causal Models

Simulation Models
 Different Techniques for Intrinsic and Extrinsic
Forecasts

Long-term,
Short-term,
Use past Extrinsic, Use external
data Intrinsic drivers
Causal

Moving Exponential Regression

Averages Smoothing Analysis
Basic Approach to Forecasting

Integrate the demand

Identify the major
Understand the planning process
factors that influence
objective of forecasting throughout the supply
the forecast
chain

Goal of forecasting:
Forecast at the Establish performance Predict the systematic
appropriate level of and error levels for the component of demand
aggregation forecast and estimate the
random component
Review of Fundamental Statistical Concepts
Most quantitative forecasting techniques are based on statistical concepts

Statistical procedures make inferences about the items of interest being examined

The entire set of items of interest is called the population

The sub-set which is used for study is called a sample

Careful selection of a representative sample from the population is key to generating a forecast
with a low degree of error (risk)
In forecasting, we use historical data, select a sample set, manipulate the data and create a
forecast based on the results
 Time-series Models

Statistical models
include - moving
Based on historical data Assumes that the past
averages, exponential
available with the will be repeated in the
smoothing, Holt's model,
company future
Winter's model are used
to develop forecasts
Time-
Series
Trends
• Represents the growth or decline
over time
• Shows the long-term movement
• Trends may be upward or positive,
downward or negative or even
stable
Seasonality
• The time-series component that
repeats and is predictable
• Influenced by factors such as
weather, temperature, festive
seasons, events
• Useful for adjusting forecast data
Cyclicals

• Fluctuations that do not have a fixed

period
• Irregular and typically span multiple
years
• Do not have a consistent pattern
• Influenced by economic, political and
social factors
• Helps the forecaster understand long-
term business cycles
 Residual component or “noise” in
the time-series
Residual Cannot be explained by trend,
Component seasonality and cyclic components

May result from measurement

errors and random occurences
 Needed to identify patterns
and anomalies
Importance of
Analyzing
Components Techniques such as time-
series decomposition may
be applied to further
understanding
Methods to forecast demand
Goal of forecasting: Predict the systematic component of
demand and estimate the random component
Multiplicative: Systematic demand = Level x Trend x Seasonal
factor
Additive: Systematic demand = Level + Trend + Seasonal factor

Mixed: Systematic demand = (Level + Trend) x Seasonal factor

Static vs Adaptive Methods

Static methods assume that the estimates of level,

trend and seasonality within the systematic component
do not change as new demand is observed

In adaptive methods, the estimates of level, trend and

seasonality are updated after each demand observation
Quantitative Models in Practice

Seasonal
Exponential Holt-Winters
Moving Average Decomposition
Smoothing Method
of Time Series

Box-Jenkins Neural
ARIMA Deep Learning
Methodology Networks
Choosing a Forecasting Technique

Why is a forecast needed?

Who will use the forecast?

What are the characteristics of the available data?

What time period is to be forecast?

How much accuracy is required?

What will be the cost of creating the forecast?

Choosing a Forecasting Technique

1 2 3 4
Define the nature of the Explain the different Understand the Develop specific
forecasting problem types of data under capabilities and criteria to for selecting
investigation limitations of the technique
forecasting techniques
being considered
 • The process of, evaluating the accuracy and
Measuring reliability of a forecasting model or method
• Done by comparing, the forecasted values with

Forecast the actual observed values

• Goal of forecast performance measurement is,

Performance to determine the effectiveness of the

forecasting method, and identify any areas for
improvement
 Mean Absolute Deviation (M A D)

Mean Squared Error (M S E)

Root Mean Squared Error (RMSE)

Measuring
Forecast Error Mean Absolute Percentage Error (MAPE)

Tracking Signal

Forecast Bias

Forecast Accuracy
 • Where:

Formula for • At = Actual Demand in period t

• Ft = Forecasted Demand for period t

RMSE
• n = Number of periods being evaluated
• The lower the RMSE, the higher is the accuracy of forecast

Calculation
 Managing the Forecast Process

Identify the various Select the

Adjust the data if
Clean the data components of Forecasting
required
demand Technique

Review the
Project the series
Determine Trend Measure the feedback and
and generate the make necessary
and Seasonality random error
forecast
changes
 Forecasting for New Products
Technique Key Feature Best For
Market Research Direct customer feedback Consumer products

Historical Analogy Using similar products’ data Products similar to existing ones

Bass Diffusion Model Predicting adoption rates Technology and innovative products

Judgmental Forecasting Expert opinions and experience High uncertainty products

PLC Analysis Based on product life cycle stages Long-term forecasting

Customer Intent Data Pre-orders or crowdfunding metrics Online-focused products

Conjoint Analysis Customer ranking of product features Multi-feature products

Simulation Models Market scenario simulations High uncertainty environments

Test Markets Limited release to gauge sales Phased product launches

Sales Force Composite Sales team provides demand estimates B2B or direct sales

Analogous Industry Data Using data from a similar industry Highly innovative or novel products

Social Media/Web Analytics Monitoring online activity to gauge interest Digital and younger demographic products
 “Looks-like” Forecasting
Look for comparable product launches

• Create month by month (week by week) sales record

• Use the percent of total sales as guide to trajectory

Structured Analogy

• Create database of past launches (sales over time)

• Characterize each launch by
• Product type
• Season of introduction
• Price
• Target market demographics
• Physical characteristics
• Use characteristics of new product to query old launches
• Avoid only including successful launches!
 Spare Parts Planning

Technique Key Feature Best For

Demand Classification Segmentation based on demand patterns Prioritizing critical parts and inventory control
Statistical Forecasting Time series, seasonality, and trend-based Parts with consistent demand patterns
forecasting
Machine Learning Advanced data analytics and predictive modeling Large datasets with complex demand
variables
Inventory Optimization EOQ, safety stock, multi-echelon inventory Minimizing costs while maintaining service
management levels
Failure Data Analysis Forecasting based on failure rates and maintenance High-cost, critical parts with preventive
schedules maintenance
Warranty Data Analysis Using warranty claims data for forecasting Parts replaced under warranty
Scenario-Based Forecasting Forecasting based on future scenarios and Uncertain markets and economic factors
simulations
 Data Collection is an important part of the
forecasting process and is essential to
improve the quality of forecasts

There are several important considerations

for data collection including reliability,

Summary
quality, consistency and availability

Statistical models include: Moving averages,

Exponential smoothing, Regression analysis
and Econometric modelling

Some methods used to estimate error are

MAE, MSE, RMSE, and MAPE
 • Explain the role of data collection in forecasting in detail.
• With the help of a chart, briefly discuss the components of
a time-series forecast.
• What is the difference between Static and Adaptive
methods of forecasting?
• List key quantitative forecasting techniques used by
Questions companies.
• What are the considerations for selecting a forecasting
technique?
• Why is forecast measurement important? Discuss the
following forecast measurement techniques – RMSE and
MAPE in detail.
• With the help of a chart explain how the forecasting
process is managed.
Module 2

Forecast Models –
Moving Averages
Topics

Moving Building a Measuring

Average Moving Model
Concepts Average Model Performance
Quantitative Techniques Recap
• When do we rely on numbers?
• When historical data is reliable, stable, and the product is mature
(high-volume, established life cycle)
• Data Preparation: The importance of cleaning data—removing
outliers (e.g., sales during a one-off promotion or a stockout
event) before modelling
 Moving Averages

A simple technique that uses Based on calculating the

Helps smooth out
a fixed number of data points average, of a fixed number of
fluctuations in time-series
and computes a mean based consecutive data points over
data and identify underlying
on the most recent time, and using this average
trends
observations to forecast future values

Trade-Off: The value of 'N'

The forecast for a specific (the number of periods)
day, is the moving average • A small N (e.g., N=3) is responsive
calculated for the previous but volatile
day • A large N (e.g., N=12) is stable but
slow to react to real shifts in demand
 Moving average works well when historical data is
reliable, stable, and the product is mature

Moving This is found in products which are high-volume

and have an established life cycle
Average
Usage Data cleaning is required to remove outliers before
modelling (e.g., sales during a special promotion
or a stockout event)

Types of moving average – simple moving average,

weighted moving average
 • The simplest form of time series forecasting
• The MA is best used when demand is
relatively stable (little trend or seasonality)
and the manager wants to smooth out
Simple random noise (random spikes or dips) without
overreacting
Moving • Forecast for the next period is simply the
Average (MA) average of the last 'N' actual demand periods

Method • The logistics manager must decide on 'N' (the

number of periods)
• A small N (e.g., N=3) is responsive but volatile
• A large N (e.g., N=12) is stable but slow to
react to real shifts in demand
 • For N periods, where the demand for each
period is Ai:
Equations • Ϝ𝑡 + 1 = ∑#!"#$%&' 𝐴𝑖 ∕ 𝑁
• Example : If N = 4
• F5 = (A4 + A3 +2 + A1) / 4
Moving Average Example
• Consider a company which Month Sales
has Monthly sales for 6 1 120
months as shown. 2 110
• Calculate the forecast using 3 130
the moving average for N = 3 4 125
(i.e; the number of periods = 3 5 140
months) 6 135
7
Moving Average Example
• Ϝ𝑡 + 1 = ∑#!"#$%&' 𝐴𝑖 ∕ 𝑁 Month Sales Forecast
1 120
• For month 7, consider 2 110
the data for the previous 3 130
3 months: 4 125
• Month 6: 135 5 140
• Month 5: 140 6 135
• Month 4: 125 7

• Calculate the average:

• 400 / 3 = 133.33
• 133 units (Rounded)
 • Smoothing the Noise: Notice that in Month 5, there
was a spike to 140 units. If we used a 1-period
forecast (naïve method), we would have predicted 140
for Month 6. By using a 3-period average, the forecast
for Month 7 (133) is "smoother" and doesn't overreact
to that single high-demand month

Interpretation • The "Lag" Effect: Simple Moving Averages always "lag"

behind a trend. If demand continues to rise every
month, the SMA will always be slightly lower than the
actual demand
• Logistics Impact: This 133-unit forecast tells the
Warehouse Manager how much labour to schedule for
picking and tells the Transport Planner how much
vehicle capacity is required for the upcoming month
Error Measurement using RMSE

• Where:
• At = Actual Demand in period t
• Ft = Forecasted Demand for period t
• n = Number of periods being evaluated
Moving Average Example with RMSE
• Consider the same example Month Sales
with Monthly sales for 10 1 120
months 2 110
• Calculate the moving average 3 130
for N = 3 (i.e; the number of 4 125
periods = 3 months) and 5 140
determine the RMSE 6 135
7 145
8 150
9 148
10 155
 Calculation of RMSE

Month Actual (At) Forecast (Ft) Error (At−Ft) Error Squared

(At−Ft)2
1 120 - - -
2 110 - - -
3 130 - - - MSE = 683.05/7=95.78
4 125 120.0 +5.0 25.0
RMSE = Sqrt (97.78) = 9.88
5 140 121.7 +18.3 334.89
6 135 131.7 +3.3 10.89
7 145 133.3 +11.7 136.89
8 150 140.0 +10.0 100.0
9 148 143.3 +4.7 22.09
10 155 147.7 +7.3 53.29
Sum 683.05
 • Accuracy: On average, our 3-period SMA forecast
is off by approximately 9.88 units
• The "Outlier" Impact: Notice Month 5 had a high
error (+18.3). In RMSE, this single error contributed
334.89 to the sum—nearly half of the total error!
This shows how RMSE highlights months where the
Interpretation forecast was significantly wrong
• Comparison Tool: If you were to calculate the
RMSE for a 5-period SMA and it resulted in 12.5,
you would conclude that the 3-period SMA is a
better model for this specific data because its
RMSE (9.88) is lower
Yummy Pizza
• Yummy Pizza is a new pizza chain which started about 10
years ago, first in Bengaluru and then expanded to Mangaluru
and Mysuru. The company has grown the number of outlets
across the three cities to 12.
• The company’s service objective was not to make its
customers wait for more than 12 minutes. So, the outlets
always had sufficient quantity of popular pizza combinations
in a form that could be quickly baked and served to the
customers. These combinations were made in a central
facility and sent to each outlet based on their requirement
• While growth has been good, the company’s profit margin has
not been up to expectations. An internal study revealed that
while the pricing and costs are comparable to its competitors,
the company has been having a much higher level of waste,
especially because of its service objective. So, the
management wanted to find out how the waste could be
reduced WITHOUT affecting the customer service.
 • A new MBA intern, Kavya was brought to work on this
initiative. After discussing with both the marketing
manager and the supply chain manager, Kavya found
that each of the outlets decided upon the number of
pre-made popular combinations for a week, based on
the outlet managers experience. Due to that each outlet
had its own formula for forecasting
• Kavya decided to concentrate on the Jayanagar outlet,
which was the largest selling unit. She collected 100
Yummy Pizza weeks worth of sales data and planned to do a simple
moving average analysis.
• The marketing manager advised her to consider only the
last 4 weeks worth of data for the moving average, but
the supply chain manager suggested that she use 6
weeks data.
• Your task is to help Kavya model the data and present
the results.
Outline Steps

Use the given sheet YummyPizza01

Using the sales data:

• Model the moving average using 4 weeks

• Model the moving average using 6 weeks
• Create a line chart for the last 10 weeks (91-100) showing both the forecasts as
well as the actual sales.
• Measure the error for both using RMSE (Note: Consider only the last 10 weeks
for the error calculation)
• What is your recommendation?
Solution
Moving Average Forecast
180

160

140

120

100
RMSE-MA4: 13.5
RMSE-MA6: 14.9
80

60

40 Based on this, how would you

20 Interpret the results in terms
0
91 92 93 94 95 96 97 98 99 100 of its impact on wastage?
Will the conclusions change
Week

Sales Fcst MA4 Fcst MA6

if we use the last 25 weeks
data?
 Industry/Product Instance Scenario Context Why MA is Used

Demand is very stable throughout

the year. The manager wants to
Example Product: Standard issue
ensure inventory is based on the
1. Fast-Moving Consumer tinned vegetables or basic
average of the last 4 weeks to
Goods (FMCG) cleaning supplies (e.g., floor
avoid overstocking due to a
cleaner).
temporary panic-buy spike one
week ago. MA smooths the noise.

Industry Example Product: Standard oil

Long product life cycle with stable,
predictable replacement demand.
Logistics needs a forecast for

Example 2. Automotive Spare Parts

filters or windscreen wiper blades
for established car models (not
new launches).
warehouse picking and stock
ordering. They use a 6-month MA
to smooth out monthly volatility
caused by garages placing batch
orders.

Demand is consistent and driven

by the customer's production rate,
Example Product: Bulk lubricants which is usually steady. A 12-
or standard nuts and bolts used period (month) MA is used for
3. Industrial Consumables
continually in manufacturing annual capacity planning to
processes. establish a smooth baseline and
ignore short-term, month-to-
month fluctuations.
 The WMA is used when there is a detectable, slow trend (up or
down) and the manager believes that the most recent data is
more predictive than older data, making the forecast more
responsive

Weighted A direct improvement over the MA - the idea is that recent sales
data is a better predictor of the future than old data
Moving Average
(WMA) Method The more recent figures are multiplied by a factor (weight) to
adjust the value (higher or lower)

The Key Decision: The manager now decides on both N and the
weightage factors
 • For N periods, where the demand for each
period is Ai and the weights are Wi:
• Ϝ𝑡 + 1 = ∑#!"#$%&' 𝑊𝑖𝐴𝑖 ∕ 𝑁
Equations • Example : If N = 4
• F5 = (W4 .A4 + W3 A3 + W2 A2 + W1 A1) / 4
• Note: The sum of weights must always be
equal to 1
Weighted Moving Average Example
Month Sales
• Consider a company which
1 120
has Monthly sales for 10
2 110
months as shown 3 130
• The forecaster wants to use 4 125
the Weighted Moving Average 5 140
for the past 4 periods as – 6 135
15%, 20%, 30% and 35% 7 145
8 150
• Calculate the forecast using
9 148
the weighted moving average
10 155
for N = 3 for the 7th month.
Weighted Moving Average Example
• Ϝ𝑡 + 1 = ∑#!"#$%&' 𝑊𝑖𝐴𝑖 ∕ 𝑁
• For month 7, consider the data for the previous 4
months:
• Month 6: 135, Weight 35%
• Month 5: 140, Weight 30%
• Month 4: 125, Weight 20%
• Month 3: 130, Weight 15%
• Calculate the weighted moving average:
• (135 x .35) + (140 x .30) + (125 x .20) + (130 x .15)
• 134 units (Rounded)
Month Actual (At) Forecast Calculation (Ft) Forecast Value

1 120 - -

2 110 - -

3 130 - -

4 125 - -

5 140 (125x .35) + (130 x .30) + (110 x.20) + (120x.15) 122.75

6 135 (140 x .35) + (125 x .30) + (130 x .20) + (110 x .15) 129.00

7 145 (135 x .35) + (140 x .30) + (125 x .20) + (130 x .15) 133.75

8 150 (145 x .35) + (135 x .30) + (140 x .20) + (125 x .15) 140.00

9 148 (150 x .35) + (145 x .30) + (135 x .20) + (140 x .15) 144.00

10 155 (148 x .35) + (150 x .30) + (145 x .20) + (135 x .15) 146.05
 Interpretation

Accuracy (WMA vs SMA) Responsiveness Logistics Impact

• The 3-period SMA had an • Consider Month 10. The • The WMA is better at
RMSE of 9.88Our 4-period actual demand was 155. catching the upward trend.
WMA has an RMSE of 10.46 The WMA forecast (146.05) If a logistics manager
• Conclusion: Even though is climbing faster toward expects growth, the WMA
the WMA is more the actual value than a helps prevent stockouts
sophisticated, in this simple average would, better than a standard SMA,
specific dataset, the SMA because we gave a 35% even if the overall error
actually performed slightly weight to the most recent (RMSE) is slightly higher
better. This shows that month
"more complex" isn't always
"more accurate"
 • After reviewing the results of the simple moving average
method, Kavya has decided to take the marketing
managers choice of 4 periods. However, she wants to
see if she can improve upon this. One approach would
be to weigh the nearer periods with a higher factor than
the periods far away
• For this, the following weights were selected – 12%,
Yummy Pizza 18%, 30%, 40% (40 being the value for the period just
past)
• Your task is to help Kavya build the weighted moving
average model using these assumptions and compare
them to the earlier model used (simple moving average).
• Which model do you think is more accurate?
 Industry/Product Instance Scenario Context Why WMA is Used

Demand is increasing week-over-

week as the weather warms. The
manager wants to predict next
Example Product: Basic t-shirts
week's demand based mostly on
1. Seasonal Apparel/Fashion being introduced ahead of the
last week's sales (e.g., 60% weight
summer season.
on the last week), to catch the
rising trend quickly. WMA allows
for responsiveness to a trend.

Industry Example Product: Fresh dairy

Inventory holding costs are
extremely high due to spoilage
risk. Logistics needs a very up-to-

Examples 2. Short Shelf-Life/Perishable

Goods
products or certain prepared
foods stocked in Distribution
Centres.
date forecast. A WMA places high
weight on the last 1-2 days of
sales to minimize overstocking
risk, assuming demand volatility is
high.

The manager needs to forecast

the decline accurately to phase
Example Product: A generic out inventory and avoid
component that is nearing the obsolescence. They use a WMA
3. Electronics Components
end of its product life cycle with higher weights on recent
(demand is slowly declining). periods to ensure the forecast is
trending downwards quickly
enough.
 Useful for identifying trends, and making short to medium-
term forecasts
Might not capture, sudden changes or seasonality in the
data effectively
Different window sizes can be used depending on the level
of smoothing required and the data characteristics
Too small an interval, makes the forecast overly responsive
to noise
Summary Too large an interval, averages out the noise, but misses
step jumps in demand
Moving averages always lag the changes that are happening

MA: Provides Stability (ignoring recent volatility to keep

operations smooth)
WMA: Offers Responsiveness (prioritising recent
information to catch emerging trends)
 Demand
Planning in
Practice
Topics

Components
Building the Performance
of Demand
Demand Plan Measurement
Planning
 Demand planning is the process of forecasting
customer demand for products and services in order to
ensure that the supply chain has the right amount of
inventory at the right time to meet that demand

Demand Demand planning helps balance supply and demand,

ensuring that production, procurement, and
distribution strategies are optimized to avoid both
Planning shortages and excess inventory

It is a collaborative process that involves different

departments—sales, marketing, finance, and
operations—working together to generate a consensus
forecast. This helps align company-wide goals and
ensures everyone is working with the same data and
assumptions
Demand Planning

Effective demand planning aids in inventory management by minimizing carrying costs, improving
service levels, and reducing the need for urgent procurement

Many companies use advanced software tools and techniques such as machine learning, artificial
intelligence, and data analytics to improve the accuracy and responsiveness of their demand
planning processes

Demand planning also helps in managing supply chain risks. By anticipating demand fluctuations,
companies can prepare for disruptions, mitigate risks such as stockouts or excess inventory, and
better handle sudden changes in market conditions.

Demand planning is an iterative process. As forecasts are made, they need to be continually
reviewed and adjusted based on actual sales data, changes in market conditions, and other external
factors.
Improved Customer Reduced Inventory by
Service by planning for stocking it at the right
availability of product place

Reduced forecasting
Optimized Utilization of
cycle times which allows
Manufacturing
organizations to react to
Resources
market changes faster

Benefits of Demand Planning

Components of Demand Planning

DATA FORECASTING PLANNING PROCESS

TECHNIQUES AND COLLABORATION
 Demand Planning Process
Demand Planning is a part of Tactical Supply Chain management (medium to short term)

Unlike Strategic Planning which is more long term in nature, Demand Planning is more tactical and
the typical planning horizon is weekly, monthly or quarterly

The process owner for Demand Planning is usually the Supply Chain department

The key elements of Demand planning include:

• Statistical forecasting
• Consensus forecasting
• Demand analysis
• Performance measurement
 Demand Planning Process
Project the series Collaborate –
Build the model and generate the Internally &
forecast Externally

Finalise & Sales & Evaluate

Communicate Operations Performance &
Forecast Planning Finetune process
 Sales - customer orders

Marketing - promotions, new product

Business introductions, competitive activities
Functions
involved in Supply Chain - distribution constraints,
Demand Planning service levels
Manufacturing / Operations -
production constraints
Finance - financial constraints, cash
flow
 Most forecasting tools or techniques require
historical sales data for generating future forecasts

Historical A min. of 8-10 data points is necessary to generate

a reliable and accurate statistical forecast
Sales data
collection and Raw historical sales data has “noise” which needs
cleansing to be removed

The Demand Planning team has the responsibility

of data cleansing

It is standard practice to use historical sales data

closest to the end customer or consumer
 Forecasting techniques or machine learning
algorithms are used to generate a statistical sales
forecast using the historical sales data

Creating In B2B industries like Auto Ancilliary, Industrial

Sales Goods where OEM/Customer demand drives
downstream planning, customer orders of
Forecasts forecasts are also incorporated in the sales
forecast
Before the statistical forecasts are reviewed by
other departments, the Demand Planning team
reviews the forecasts and makes necessary
adjustments
 The Demand Planning team releases or publishes the Statistical
Forecast for key departments like Sales, Marketing & Finance to review
and provide feedback/update on any changes

A formal process called “Collaborative Planning” is usually managed

by the Demand Planning team where these departments provide
Building specific input (called “Market Intelligence”)

Consensus Typical Market Intelligence consists of actual or likely Customer

orders, Planned Sales Promotions, Customer Offers, New Product

Forecasts Introduction plans & competitive activity in the market that is likely to
impact demand
The outcome of the Collaborative Planning process is to finalise a
Consensus Forecast which is commonly agreed to by all departments

This will be used for downstream supply planning and against which
actual sales will be reviewed for any corrections etc.
 With increase in size and complexity of SKUs, it
is not possible to review and modify forecasts
for all SKUs

The Collaborative Planning process is therefore

Building used mainly for the critical/high velocity Class
Consensus “A” SKUs

Forecasts During this process, Finance ensures that the

finalised Consensus Forecasts are in line with
the overall Financial goals of the organisation

In a monthly demand planning cycle, the

Consensus Forecasts are finalised by the
2nd/3rd working day of the month
 The forecasts are then matched with Supply Plans
and incase there are constraints like Manufacturing
capacity or shortage of raw materials etc. due to
which Supply may not meet Demand, changes are
made in the forecasts
Sales and The Final Forecasts are then shared & published by
Operations Demand Planning for all key users for planning and
scheduling their respective activities/tasks
Planning
This process of Supply & Demand Collaboration is
typically known as Sales & Operations Planning
which is usually conducted by the first week of the
month by Supply Chain department
 Demand Planning and S&OP
Demand Statistical Strategic
Safety Stock Inventory
Planning Calculation Management
Days Coverage

Demand History Inventory Turns

Market Intelligence Customer Service
Customer Collaboration Level

Freight cost
Warehousing cost

Revenue Manufacturing Cost

Profits Capacity Utilisation

Sales & Manufacturing

Logistics &
Operations Planning &
Distribution
Planning Procurement
 Inventory Distribution Production Procurement
Planning Planning Planning Planning

Ensuring Ensuring Ensuring that Ensuring that

adequate inventory is production is RM/PM are

Downstream
inventory is distributed in planned and procured and
available to the most timely scheduled at made available
meet demand and efficient Plants in at the
despite manner to alignment with respective

Planning
fluctuations in downstream downstream Plants to meet
actual demand nodes like demand their
and lead time Depots, Production
of supply Distributors & Plan
Retailers to
meet end
consumer
demand
 Problems with historical data

• Stock-outs
• Data entry error

Causes of Problems with manual intervention

Errors • Too much optimism / pessimism

• Lack of trust in supply chain

Way of doing business

• One off orders

• Last minute incentives

Wrong choice of forecast method

 The ultimate goal of the Demand Planning process is to generate
accurate and reliable forecasts that are agreed upon collaboratively

Actual Sales/Demand vs Forecasted Sales/Demand of the previous

planning cycle is carried out by the Demand Planning team

Typical measures used to measure Forecast Accuracy are :

Review
Performance Mean Absolute Percentage Error (MAPE) = (Forecast – Actuals) x 100 /
Forecast

% Good Forecasts – This is a measure of the # of SKUs which achieved a

Forecast Accuracy of >80% of the total # of SKUs that were forecasted

A good practice is to also measure & compare the forecast accuracy of

the Statistical Forecast and the Consensus Forecast. This helps
establish if the Market Intelligence provided by various departments is
improving the final forecast or not
 The key components of Demand Planning
include Data, Forecasting Techniques
and the Collaborative Process

Summary
The collaborative Demand Planning
process results in a better consensus
forecast which is easier to implement

Performance Measurement of Forecasts

is key to process improvement
Order Management and
Fulfillment

Module 2
Topics

Order
Order
Sales Orders Management
Fulfillment
Process
Types of Sales Methods

Direct sales to customers via catalogues

Direct sales to customers via company owned stores

Direct sales to customers via e-commerce

Channel sales: Manufacturer - Distributor - Retail

Business to Business Sales

 Catalogue ordering

Ordering through salesperson

E-commerce order
Order E-mail orders
Methods
EDI / website / system-to-system

RFP based selection

Tenders
Channel Members

Manufacturers
Wholesalers / Distributors
C&F Agents
Retailers
Other Members
Distribution Options

Manufacturer storage with direct shipping

Manufacturer storage with direct shipping and in-transit merge

Distributor storage with carrier delivery

Distributor storage with last-mile delivery

Manufacturer / distributor storage with customer pickup

Retail storage with customer pick-up

 Generation of Leads

Qualifying Leads

Making offers or quotations

Typical B2B Negotiations

Sales Process Processing the Sales order

Fulfillment

Invoicing

Collections
 Order Processing
• Customer places order
• Order received by manufacturer
• Order processed
Steps in Order
Order Fulfillment
Management
• Items for delivery picked from stores and
packed
• Documentation for shipping prepared
• Shipment
• Delivery to customer
 Depends upon how orders are
received (e-commerce, direct, B2B)

Order validation and customer

confirmation

Order Receipt Order acknowledgment

Order confirmation
 Check for
inventory and
Order entry
production
schedule

Order Check for credit Application of

Processing limit discounts

Check terms and Forecast

conditions consumption
 Direct order entry

Conversion from quotations

Order Entry RFP / Tender

Rate Contract from

Government
Inventory Check
• The ATP quantity (ATP = Available-to-Promise) is
calculated based on warehouse stock, planned
receipts (production orders, purchase orders,
planned orders, and so on), and planned issues (sales
orders, deliveries, reservations, and so on)
• Preferred warehouse locations for customers
• Alternate / substitute products
• Alternate locations
• Master schedule
 Role of inventory control
Inventory in fulfilling orders
Management
in Order
Fulfillment
Real-time stock visibility

Impact on order accuracy

and delivery timelines
 Credit limit check

Credit limit calculation methods

Credit Limit
Check Credit hold

Credit release
 Quality and Inspection terms

Price basis, payment terms

Terms and Delivery terms

Conditions
Documentation, taxes, import/export
regulations
Safety, environmental considerations
Inventory Check and Issue Steps

Packing – Documentation
Picking from depending upon (invoice / delivery
warehouse customer and documents / bill
shipment mode of lading)

Delivery to
Dispatch Managing returns
customer
 Warehouse management during
picking
Automated picking, steps to minimize
errors
Picking and Labelling and packaging
Packing
Accessories, manuals

Packing processes to minimize

damage and maximize space efficiency
Importance of Order Management

Impact on customer satisfaction

Reduction of errors and delays

Efficient resource utilization

 Select transportation model
and carrier

Shipping and Coordination with third-party

Delivery logistics (3PL) providers

Delivery methods and

tracking
 • Custom orders
• Rush deliveries
• Special packaging
• Export requirements (third-party
Exceptions inspection, customs
documentation)
• Managing complex customer
requirements in logistics
 Inventory shortages,
incorrect shipments, delays
Challenges in
Order Handling returns and
Management exchanges

Communication gaps
between departments
 Impact of software systems
Role of like ERP, TMS, and OMS on
Technology in order fulfillment
Order
Management
Use of AI, machine learning,
and data analytics for better
decision-making
 Order accuracy
Key Metrics to
Monitor in On-time delivery rates
Order
Management
Return rates

Customer satisfaction
 Integrating systems across
the supply chain

Best
Practices for
Efficient Prioritizing customer
experience
Order
Management
Continuous monitoring and
process optimization
 There are multiple methods used by companies to
take orders from customers including direct orders,
catalogs, channel sales, e-commerce orders

Based on the company’s distribution strategy,

different methods of distribution channels are used

B2B sales process, used typically for high-value

Summary orders is usually multi-step and may involve many
participants

The two key steps in order management are order

processing and order fulfillment

Order management is a critical step in determining

response to customer and customer satisfaction
Questions

What are the different sales methods and order methods used by organizations?

With the help of industry examples, explain the typical distribution options used by organizations.

Draw a chart to illustrate a typical B2B sales order process.

Using a B2B example, discuss the Order Processing and Order Fulfillment stages in order management.

Explain the role of inventory in the Order Management cycle.

What are the main challenges in Order Management and what best practices may be used to manage them?

List KPIs in Order Management – what are the formulas used for each of the KPIs.
Customer Service and
Logistics

Module 2
 • Customer Service Perception and Gaps
• Basic Logistics Service Capabilities
Topics • Metrics for Customer Service
• Customer Profitability Analysis (CPA)
• Value-Added Services
 The Customer Gap

Expected Service

Customer Gap

Perceived Service
Customer Gap
Difference between customer expectations and perceptions

Expectations – what the customer believes will happen

Perception – what the customer believes has happened

Closing the gap is critical to delivering quality service

Provider Gap

Gap 2: The Service

Gap 1: The Listening
Design and Standards
Gap
Gap

Gap 3: The Service Gap 4: The

Performance Gap Communication Gap
Service Design Standards

• Provider needs to have service designs and standards that meets expectations
• Challenge in translating customer expectations into service delivery
• Organization may believe that expectations are unrealistic or that delivery
performance is variable
• Causes:
• Poor service design
• Absence of customer-driven standards
• Inappropriate physical evidence and servicescape
 Logistics Support during the PLC stages

Sl PLC Stage Customer Requirements Logistics Response Requirements

1 Introduction High availability, small volumes Flexibility, high inventory, rapid replenishment,
quick delivery, LTL shipments, very high costs, ship
from factory
2 Growth High availability, high volumes, Delivery / cost balance, support larger delivery
more markets areas, high logistics cost, develop warehouses
3 Maturity Reasonable availability, Price- Standard deliveries (FTL), lower logistics costs,
focus regional and local distribution centres, longer lead
times
4 Decline Limited availability, Margin-focus Slower response balanced by risk, lower customer
service levels, longer lead-times, lower logistics
costs
Customer Expectations

• Timely delivery
• Accurate order fulfillment
• Communication and transparency
• Flexibility and responsiveness
• Returns Management
• Documentation
• Follow-up on subsequent processes
Examples of Logistics Service Failures

Late Missing Wrong Damaged

deliveries material material material

Lack of Personnel
Defective
information Higher costs failures
packaging
to customers (driver issues)
Issues that Impact Customer Perception of
Service Quality

Competence Reliability Responsiveness

Transaction
Trustworthiness Access
security
Customer Service Phases

Pre-Transaction Transaction Post-Transaction

• Customer service policy • Order fulfilment reliability • Order status information

• Service organization • Delivery consistency • Customer complaint, claim
• Structuring service • Order convenience and return
• Customer education • Order postponement • Installation, commissioning
• System design • Product substitute and technical support
• Education and training
Basic Service Capability

Product availability

Logistics performance

Logistics reliability
 Safety Stock Trade-off: Increasing
service levels (e.g., 95% to 99%)
follows an exponential cost curve

Stock-out Frequency: No. of times

Product product is unavailable to fulfill
Availability demand

Fill Rate: The percentage of

demand met by immediate stock
availability
 • Stocks are held to meet the customer orders
• Service Level - desired probability of not running
out of stock between order placement and
receipt of material
SERVICE • 90% service level indicates that over a specified
LEVEL period, the organization is able to meet the
customer demand on 90% of the time
POLICIES • Every incremental gain in availability requires a
disproportionate investment in inventory
• The concept of service level seeks to achieve the
objective of customer service, by deciding the
amount of safety stock to be maintained
Logistics Performance

Speed Flexibility Consistency Service Failures

• Delivery time • Changes in • Predictable • Recovery after

impacts customer destination/source, deliveries failures
inventory levels new products,
(and costs) specific packaging
requirements,
disruption in supply
 Customer Service Metrics
Order Fill Rate: Total Number of Customer Orders Shipped / Number of Customer Orders Received x 100

Perfect Order Rate: (Percent of orders delivered on time) x (Percent of orders complete) x (Percent of orders damage
free) x (Percent of orders with accurate documentation)

On-time Shipping: (Number of On-Time Items / Total Items) x 100

Back-order Rate: Number of undeliverable orders/by the total number of orders x 100

Order Cycle-time: Actual Delivery Date – Customer Order Date

Rate of Return: Total Items Returned / Total Items Shipped

 CLV is a metric that estimates the total revenue a business
can reasonably expect from a single customer account
throughout the entire business relationship

Relevance to Supply Chain

Customer • When inventory is scarce (a "stock-out"), managers must decide who gets

Lifetime
the remaining product. CLV helps prioritize partners who provide long-term
stability over "transactional" buyers who only shop for the lowest price

Value (CLV) CLV = (Average Order Value x Purchase Frequency)

x Customer Relationship Span

• Example: If a retailer orders 1,00,000 of goods, 4 times a year, and usually

stays with the company for 5 years,
• CLV = (1,00,000 x 4) x 5 = 20,00,000
Other Customer Service Attributes
Logistics

Distribution aspects Commercial Aspects

• Order processing time • Offering credit facility

• Delivery consistency • Financing the stocks
• Delivery frequency • Service support
• Stock availability • Handling complaints
Reliability
Predictability

Availability of accurate information

Detailed tracking of customer orders

Physical integrity of the goods (safety and packaging)

Continuous improvement
Customer Profitability Analysis (CPA)

• Not all customers deserve the same service

level. A strategic approach requires
segmenting customers based on their Cost-to-
Serve (CTS)
• The Whale Curve
• The top 20% of customers usually
generate the majority of a company's
profit
• The middle 60% break even
• The bottom 20% lose money because
their service requirements (small orders,
frequent returns, custom requests)
exceed the margin they provide
 Customer Segmentation Matrix

Customer Category Action Strategy

Protect: Provide the highest "Perfect
High Margin / Low Cost-to-Serve
Order" levels.
Engineer: Work with them to reduce
High Margin / High Cost-to-Serve logistics costs (e.g., consolidate
orders).
Maintain: Standardize service; keep
Low Margin / Low Cost-to-Serve
costs lean.
Re-examine: Increase prices,
Low Margin / High Cost-to-Serve
change service terms, or terminate
Value-Added • Customized transportation
• Payment collection on
• Cross docking
• Reverse logistics
Services delivery
• Consignment tracing and
• Price marking and labelling tracking
• Product mixing and • Drop-shipment
packaging
• Logistical audit
• Point-of-sale presentations
• Government liaison
• Vendor-managed inventory
 • Explain the GAP model and how it is relevant to logistics as a
service.
• Why should logistics providers have service design standards?
• How does logistics customer service vary with the Product Life
Cycle? Use a mobile phone company to support your answer.
• What are typical customer expectations of logistics support?
• Discuss logistics service failures using a B2B automotive industry
example.
• Which issues affect the customer perception of quality?

Questions • What are the 3 factors that contribute to the Basic Service
Capability of logistics? Explain each of the factors in detail.
• List 5 metrics used to monitor customer service performance
along with their method of calculation.
• Which are the factors that are critical to the Operational
Performance of logistics?
• Explain customer profitability analysis in detail.
• Using a 3PL service provider as an example, compare the activities
that are important in each of the phases of customer service.
Customer Service
Failures and
Recovery
Module 2
 • The "Paperwork" Bottleneck: Even with
GST and E-Way bills, inter-state check-
posts or incorrect documentation can
lead to shipments being impounded for
days
Typical • Infrastructure Fragility: Monsoon-related
Logistics flooding or highway landslides that create
"Black Hole" delays where tracking stops
Failures updating
• The Address Problem: Lack of
standardized geocoding leads to "Address
Not Found" failures, particularly in high-
density areas
 • Direct Costs: Extra freight (expedited
shipping), return handling (reverse
logistics), and administrative labor to fix

Costs of the error

• Penalty Costs: Contractual "liquidated
damages" or SLA fines (common in

Failure retail and automotive sectors)

• Opportunity Costs (The "Silent Killer"):
Loss of future sales. Research shows it
costs 5x to 25x more to acquire a new
customer than to keep an existing one
Customer Responses to Stockouts

Substitute: They
buy a different Delay: They wait
brand (Loss for for the stock
the (Best case
manufacturer, scenario)
not the retailer)

Permanent
Defection: They
Switch: They go
never return
to a competitor
(Maximized loss
(Loss for both)
of Lifetime
Value)
Effective Service Recovery – Triple-A
Framework
Acknowledge & Apologize: Speed is critical. The "Listening Gap" grows the
longer the firm stays silent

Amend : Transparency is key. Don't just say "it's coming"; provide a real-
time tracking link or a specific new ETA

Appreciate : Distinguish between Outcome Fairness (did the customer get

what they paid for?) and Interactional Fairness (was the customer treated
with respect?
 • The Scenario: A high-end hospital in
Jayanagar is waiting for a critical MRI
cooling component from Germany
• The Failure: The part arrived at KIA
(Bangalore Airport) at 8:00 AM but is stuck
in customs due to a technical glitch
Group
• The Clock: It is now 4:00 PM. The hospital
Activity has 12 patients scheduled for tomorrow
morning
• The Gridlock: If the part is released at 5:00
PM, it must face the peak-hour traffic
surge from Devanahalli to South
Bangalore
Activity – Corrective Action:

Recovery How do you get the

part there faster than
the standard time?

Plan

Compensatory
Action: What do you Communication Plan:
offer to make up for Who do you call, and
the 24-hour what is the "Script"?
downtime?
 • The "Service Recovery Paradox" proves that a
logistics failure isn't the end of a customer
relationship; it is a critical touchpoint
• A customer who experiences a transparent,
empathetic, and swift resolution is statistically more
likely to remain loyal than a customer who never
experienced a problem at all
• A successful recovery must satisfy the customer on
Conclusion three distinct levels:
• Outcome (Distributive): Providing a tangible fix or
compensation.
• Process (Procedural): Making the "complaint-to-
resolution" path frictionless.
• Empathy (Interactional): Ensuring the human
interaction is respectful and proactive
• Communication and systematic response to failure
recovery are critical
 Give examples of typical logistics
failures.

What are the costs of failures?

Questions
How do customers respond to failures?

Explain the recommended Triple-A

framework for effective service recovery.
Case Study Assignment: The Vayu Express Seasonal
Surge Challenge
Time Allotted: 60 Minutes

Focus Areas: Demand Forecasting, Order Management, Customer Service & Recovery.

1. Scenario Overview
Vayu Express is struggling with the Bullwhip Effect. While consumer demand for the "Z-
Phone" is stable, order signals are becoming distorted as they move up the supply chain.
Furthermore, the volatility has led to "Stock-Outs," requiring the Order Management team
to handle disgruntled retailers and implement service recovery protocols.

2. Learning Objectives
● Analyze demand distortion (The Bullwhip Effect)
● Develop a 3-month moving average forecast
● Apply Service Recovery frameworks to supply chain failures

3. Demand Forecasting & Analysis

Data Set (Last 6 Months - Units in 000s)
Month Consumer Retailer Factory
Sales Orders Orders
Jan 100 105 115
Feb 102 95 80
Mar 104 120 145
Apr 106 90 60
May 108 130 170
Jun 110 85 50

Tasks:
1. Volatility Assessment: Identify which stage of the chain has the highest variance
2. July Forecast: Calculate a 3-Month Simple Moving Average (SMA) for Factory Orders to
determine the July production requirement
3. Behavioural Analysis: Explain why the Factory Orders (50) in June are so much lower
than actual Consumer Sales (110)
4. Order Management & Prioritization
Scenario: A sudden monsoon flood has delayed a primary shipment, leaving you with only
500 units available for the next 4 days. You have three pending orders:

1. Order A: 300 units for a "Gold-Tier" Corporate Client (Contractual penalty of

₹50,000 for late delivery).
2. Order B: 250 units for the Flagship Retail Store (Stock-out will result in lost walk-in
customers and high dissatisfaction).
3. Order C: 100 units for an E-commerce VIP Pre-order (Guaranteed "Next-Day"
delivery).

Tasks

• How would you allocate the 500 units? Provide a rationale based on Total Cost of
Service and Customer Lifetime Value (CLV)
• (Note:There is no single "right" answer, but your logic must account for both
financial penalties and brand equity)

5. Service Recovery Strategy

Scenario: Due to a system glitch in the Order Management System (OMS), 50 customers in
South Mumbai were sent the wrong colour variant of the Titan-X phone. They have already
unboxed the devices and are venting their frustration on social media.

Task

• Design a Service Recovery Plan using the "Triple-A Framework"

• How will you address the situation if there is a lot of negative comments on Social
Media?