Description

For this assignment, plan to identify the process early in the week in order for you to compile data which can then be presented in the final assignment. Submit the required content to include the data and flowchart in one final document.

Please be sure to consider a process that you would perform daily, and only present the data for the week of class.

Please see attached grading guide for criteria and assigned points.

Select a process you perform daily, but would like to spend less time doing, such as driving to work.

Design a flowchart and provide written analysis by using any appropriate tool.

Comment on the factors that affect the process design.

Identify at least one metric to measure the process.

Describe which forecasting methods would be applicable.

Discuss how one could manage this process by using PERT/CPM techniques.

Submit your flowchart for the process and the data collected at the end of the week by collecting data for the identified metric every day of the workweek.

Use APA format to include a reference page with the course text at a minimum.

Individual Assignment: Flowchart for a Process

Purpose of Assignment

The purpose of this assignment is to apply quality-management tools to help examine a process.

Students will design a flow chart, comment on the factors affecting the process design, and identify a

metric to evaluate the success of the process.

Grading Guide

Content

Met

Partially

Met

Not Met

Total

Available

Total

Earned

Comments:

The student selects a process they perform

daily but would like to spend less time doing,

such as driving to work. (5 pts)

The student designs a flowchart and provides

written analysis using an appropriate tool. (10

pts)

The student comments on the factors that

affect the process design. (5 pts)

The student identifies at least one metric to

measure the process. (5 pts)

The student describes which forecasting

methods would be applicable. (10 pts)

The student discusses how one could

manage this process by using PERT/CPM

techniques. (10 pts)

The student submits their flowchart for the

process and the data collected at the end of

the week by collecting data for the identified

metric every day of the workweek. (10 pts)

55

Writing Guidelines

The paperÃ¢â‚¬â€including tables and graphs,

headings, title page, and reference pageÃ¢â‚¬â€is

consistent with APA formatting guidelines and

Met

Partially

Met

Not Met

Comments:

Writing Guidelines

Met

Partially

Met

Not Met

Total

Available

Total

Earned

meets course-level requirements. (5 pts)

Intellectual property is recognized with in-text

citations and a reference page. (5 pts)

Paragraph and sentence transitions are

present, logical, and maintain the flow

throughout the paper. (5 pts)

Sentences are complete, clear, and concise.

(5 pts)

Rules of grammar and usage are followed

including spelling and punctuation. (5 pts)

25

Assignment Total

Additional comments:

80

Comments:

Part 2 Managing Customer

Demand

1270131 – Pearson Education Limited Ã‚Â©

8 Forecasting

1270131 – Pearson Education Limited Ã‚Â©

Accurate forecasting is crucial to maintaining the proper amount of product

in the supply chain. Kimberly-Clark recently incorporated demand-signal data

Ã¢â‚¬â€information on actual consumer salesÃ¢â‚¬â€into its forecasting system and

greatly increased the accuracy of the forecasts. Here a worker moves

pallets of paper products at a Kimberly-Clark warehouse in Beech Island,

South Carolina.

Annette M. Drowlette/Agusta Chronicle/ZUMAPRESS.com/Newscom

1270131 – Pearson Education Limited Ã‚Â©

Kimberly-Clark

What do Kleenex tissues, Huggies diapers, and Scott paper

towels all have in common? They are all produced by

Kimberly-Clark, a $21 billion multi-national company based in

Irving, Texas. With 106 production and warehouse facilities

worldwide one can only imagine the complexity of ensuring

that retail customers located in 175 countries receive their

orders on time. Any time the retailerÃ¢â‚¬â„¢s inventories are out of

sync with production forecasts, it can have a dramatic effect

on Kimberly-ClarkÃ¢â‚¬â„¢s bottom line: empty shelves at the retail

level force consumers to seek out competitorsÃ¢â‚¬â„¢ products

while too much inventory at Kimberly-Clark is very costly. For

example, during the high-volume Ã¯Â¬â€šu season a one-day

reduction in safety stock inventories translates into a $10

million savings across the supply chain network. It is no

wonder that forecast accuracy is a top priority at KimberlyClark. Forecast errors drive the need for safety stocks

(greater forecast errors equate to greater uncertainty in

demands) and result in ineÃ¯Â¬Æ’cient operations and higher

costs. Consequently, Kimberly-Clark undertook a major

project to improve its forecasting performance.

Prior to the onset of the project, forecasts were based on

historical shipment data. The shipments were geared to

satisfy actual customer orders. Intuitively, those data should

be good for making forecasts of future orders. However,

actual shipments may be subject to all sorts of anomalies

such as supply disruptions, factory or transportation capacity

1270131 – Pearson Education Limited Ã‚Â©

limits, or severe weather, all of which could delay the

shipments and miss the dates the customer actually wanted

the product. Perhaps the biggest problem with using past

shipment data for making forecasts is that it is backwardlooking; forecasts assume that what happened in the past

will happen in the future. Such an approach will miss spikes

in consumer demand. For example, Hurricane Sandy

pummeled the east coast in 2012 and caused a drop in the

sales of paper products in the northeast region. Estimating

weekly sales on the basis of historical shipment data when

there was no storm will be fraught with errors; even

managerial judgment to temper the forecasts will not provide

much relief. When would consumers in the northeast turn

their attention from buying generators and portable lighting

products to paper towels and diapers again?

The project to improve forecasting performance was a major

part of a larger project to create a demand-driven supply

chain. Kimberly-Clark reduced the number of production

facilities and warehouses, opened new larger facilities, and

repurposed others to handle a smaller set of customers or to

ship only promotional items. All told, this design not only

improved logistical eÃ¯Â¬Æ’ciency but also simpliÃ¯Â¬Âed the

forecasting effort. The key, however, to creating a demanddriven supply chain was to incorporate demand-signal dataÃ¢â‚¬â€

information about actual consumer purchasesÃ¢â‚¬â€into its plans

to resupply retailers with product. In close collaboration with

Terra Technology, whose Multi-Enterprise Demand Sensing

(MDS) system was chosen for the forecasting tool, KimberlyClark incorporated the point-of-sale data (POS) from three of

its largest retail customers in North America. The software

uses that data along with inventory in the distribution

1270131 – Pearson Education Limited Ã‚Â©

channel, shipments from warehouses, and the retailerÃ¢â‚¬â„¢s own

forecasts to create a daily operational forecast. These inputs

are re-evaluated weekly for their inÃ¯Â¬â€šuence on the forecast.

For example, POS might be the best predictor of a shipment

forecast on a three-week horizon, but actual orders could be

the best predictor for the current week. A new metric for

evaluating forecast performance was created; it was deÃ¯Â¬Âned

as the absolute difference between shipments and forecast

and reported as a percentage of shipments. Using that

metric and the new forecasting system, Kimberly-Clark

observed forecast error reductions as high as 35 percent in

the Ã¯Â¬Ârst week of the horizon and 20 percent on a two-week

horizon. These forecast error reductions can translate into

one to three weeks reduction in safety stocks.

Sources: James A. Cooke, Ã¢â‚¬Å“Kimberly-Clark Connects Its supply Chain to the Store Shelf,Ã¢â‚¬Â DC Velocity, April 10, 2013; Paul Taylor,

Ã¢â‚¬Å“Demand Forecasting Pays Off for Kimberly-Clark,Ã¢â‚¬Â Financial Times, September 10, 2011; Heather Clancy, Ã¢â‚¬Å“Kimberly-Clark Makes

Sense of Demand,Ã¢â‚¬Â CGT, http://consumergoods.edgl.com; Kimberly-Clark Annual Report, 2013, http://www.Kimberly-Clark.com.

1270131 – Pearson Education Limited Ã‚Â©

Learning Goals

After reading this chapter, you should be able to:

1. Explain how managers can change demand patterns.

2. Describe the two key decisions on making forecasts.

3. Calculate the Ã¯Â¬Âve basic measures of forecast errors.

4. Compare and contrast the four approaches to

judgmental forecasting.

5. Use regression to make forecasts with one or more

independent variables.

. Make forecasts using the Ã¯Â¬Âve most common

statistical approaches for time-series analysis.

7. Describe the six steps in a typical forecasting

process.

Balancing supply and demand begins with making accurate

forecasts, and then reconciling them across the supply chain

as shown by Kimberly-Clark. A forecast is a prediction of

future events used for planning purposes. Planning, on the

other hand, is the process of making management decisions

on how to deploy resources to best respond to the demand

forecasts. Forecasting methods may be based on

mathematical models that use available historical data, or on

qualitative methods that draw on managerial experience and

judgments, or on a combination of both.

1270131 – Pearson Education Limited Ã‚Â©

forecast

A prediction of future events used for

planning purposes.

In this chapter our focus is on demand forecasts. We begin

with different types of demand patterns. We examine

forecasting methods in three basic categories: (1) judgment,

(2) causal, and (3) time-series methods. Forecast errors are

deÃ¯Â¬Âned, providing important clues for making better

forecasts. We next consider the forecasting techniques

themselves, and then how they can be combined to bring

together insights from several sources. We conclude with

overall processes for making forecasts and designing the

forecasting system.

Forecasts are useful for both managing processes and

managing supply chains. At the supply chain level, a Ã¯Â¬Ârm

needs forecasts to coordinate with its customers and

suppliers. At the process level, output forecasts are needed to

design the various processes throughout the organization,

including identifying and dealing with in-house bottlenecks.

As you might imagine, the organization-wide forecasting

process cuts across functional areas. Forecasting overall

demand typically originates with marketing, but internal

customers throughout the organization depend on forecasts

to formulate and execute their plans as well. Forecasts are

1270131 – Pearson Education Limited Ã‚Â©

critical inputs to business plans, annual plans, and budgets.

Finance needs forecasts to project cash Ã¯Â¬â€šows and capital

requirements. Human resources uses forecasts to anticipate

hiring and training needs. Marketing is an important source

for sales forecast information because it is closest to

external customers. Operations and supply chain managers

need forecasts to plan output levels, purchases of services

and materials, workforce and output schedules, inventories,

and long-term capacities. Managers at all levels need

estimates of future demands, so that they can plan activities

that are consistent with the Ã¯Â¬ÂrmÃ¢â‚¬â„¢s competitive priorities.

1270131 – Pearson Education Limited Ã‚Â©

Managing Demand

Before we get into the tools and techniques for forecasting demands, it is

important to understand that the timing and sizing of customer demand

can often be manipulated. Accurately forecasting customer demand is a

diÃ¯Â¬Æ’cult task because the demand for services and goods can vary greatly.

For example, demand for lawn fertilizer predictably increases in the spring

and summer months; however, the particular weekends when demand is

heaviest may depend on uncontrollable factors such as the weather. These

demand swings are costly to satisfy for any process, even if they are

predictable. However, managers can often do two things to alleviate the

pains of demand swings. First, understand the demand pattern they are

facing; second, employ one or more options to alleviate any avoidable

swings.

1270131 – Pearson Education Limited Ã‚Â©

Using Operations to Create Value

Managing Processes

Process Strategy and Analysis

Quality and Performance

Capacity Planning

Constraint Management

Lean Systems

Project Management

Managing Customer Demand

Forecasting

Inventory Management

Operations Planning and Scheduling

Resource Planning

Managing Supply Chains

Supply Chain Design

Supply Chain Logistic Networks

Supply Chain Integration

Supply Chain Sustainability

1270131 – Pearson Education Limited Ã‚Â©

Demand Patterns

Forecasting demand requires uncovering the underlying patterns from

available information. The repeated observations of demand for a service

or product in their order of occurrence form a pattern known as a time

series . There are Ã¯Â¬Âve basic patterns of most demand time series:

time series

The repeated observations of demand for a service

or product in their order of occurrence.

1. Horizontal. The Ã¯Â¬â€šuctuation of data around a constant mean.

2. Trend. The systematic increase or decrease in the mean of the

series over time.

3. Seasonal. A repeatable pattern of increases or decreases in

demand, depending on the time of day, week, month, or season.

4. Cyclical. The less predictable gradual increases or decreases in

demand over longer periods of time (years or decades).

5. Random. The unforecastable variation in demand.

Cyclical patterns arise from two inÃ¯Â¬â€šuences. The Ã¯Â¬Ârst is the business cycle,

which includes factors that cause the economy to go from recession to

expansion over a number of years. The other inÃ¯Â¬â€šuence is the service or

product life cycle, which reÃ¯Â¬â€šects the stages of demand from development

through decline. Business cycle demand is diÃ¯Â¬Æ’cult to predict because it is

affected by national or international events.

1270131 – Pearson Education Limited Ã‚Â©

The four patterns of demandÃ¢â‚¬â€horizontal, trend, seasonal, and cyclicalÃ¢â‚¬â€

combine in varying degrees to deÃ¯Â¬Âne the underlying time pattern of

demand for a service or product. The Ã¯Â¬Âfth pattern,

Figure 8.1 Patterns of Demand

random variation, results from chance causes and thus, cannot be

predicted. Random variation is an aspect of demand that makes every

forecast ultimately inaccurate. Figure 8.1 shows the Ã¯Â¬Ârst four patterns of

a demand time series, all of which contain random variations.

1270131 – Pearson Education Limited Ã‚Â©

MyOMLab

Animation

1270131 – Pearson Education Limited Ã‚Â©

Demand Management Options

Matching supply with demand becomes a challenge when forecasts call

for uneven demand patternsÃ¢â‚¬â€and uneven demand is more the rule than the

exception. Demand swings can be from one month to the next, one week to

the next, or even one hour to the next. Peaks and valleys in demand are

costly or can cause poor customer service. Air New Zealand can lose sales

because capacity is exceeded for one of its Ã¯Â¬â€šights, while another of its

Ã¯Â¬â€šights to the same destination at about the same time has many empty

seats. If nothing is done to even out demand, sales are lost or greater

capacity cushions might be needed. All come at an extra cost. Here we

deal with demand management , the process of changing demand

patterns using one or more demand options.

demand management

The process of changing demand patterns using

one or more demand options.

Various options are available in managing demand, including

complementary products, promotional pricing, prescheduled appointments,

reservations, revenue management, backlogs, backorders, and stockouts.

The manager may select one or more of them, as we illustrate below.

1270131 – Pearson Education Limited Ã‚Â©

Complementary Products

One demand option for a company to even out the load on resources is to

produce complementary products , or services that have similar resource

requirements but different demand cycles. For example, manufacturers of

matzoh balls for the Jewish Passover holiday are in a seasonal business.

The B. Manischewitz Company, a kosher foods manufacturer in Jersey City,

New Jersey, previously experienced 40 percent of its annual sales for the 8day Passover holiday alone. It expanded toward markets with year-round

appeal such as low-carb, low-fat foods, including canned soups and

crackers, borscht, cake mixes, dressing and spreads, juices, and

condiments.

complementary products

Services or products that have similar resource

requirements but different demand cycles.

For service providers, a city parks and recreation department can

counterbalance seasonal staÃ¯Â¬Æ’ng requirements for summer activities by

offering ice skating, tobogganing, or indoor activities during the winter

months. The key is to Ã¯Â¬Ând services and products that can be produced with

the existing resources and can level off the need for resources over the

year.

1270131 – Pearson Education Limited Ã‚Â©

Promotional Pricing

Promotional campaigns are designed to increase sales with creative

pricing. Examples include automobile rebate programs, price reductions

for winter clothing in the late summer months, reduced prices for hotel

rooms during off-peak periods, and Ã¢â‚¬Å“two-for-the-price-of-oneÃ¢â‚¬Â automobile

tire sales. Lower prices can increase demand for the product or service

from new and existing customers during traditional slack periods or

encourage customers to move up future buying.

Prescheduled Appointments

Service providers often can schedule customers for deÃ¯Â¬Ânite periods of

order fulÃ¯Â¬Âllment. With this approach, demand is leveled to not exceed

supply capacity. An appointment system assigns speciÃ¯Â¬Âc times for service

to customers. The advantages of this method are timely customer service

and the high utilization of service personnel.

Doctors, dentists, lawyers, and automobile repair shops are examples of

service providers that use appointment systems. Doctors can use the

system to schedule parts of their day to visit hospital patients, and lawyers

can set aside time to prepare cases. Care must be taken to tailor the length

of appointments to individual customer needs rather than merely

scheduling customers at equal time intervals.

1270131 – Pearson Education Limited Ã‚Â©

Reservations

Reservation systems, although quite similar to appointment systems, are

used when the customer actually occupies or uses facilities associated

with the service. For example, customers reserve hotel rooms,

automobiles, airline seats, and concert seats. The major advantage of

reservation systems is the lead time they give service managers and the

ability to level demand. Managers can deal with no-shows with a blend of

overbooking, deposits, and cancellation penalties. Sometimes overbooking

means that a customer with reservations cannot be served as promised. In

such cases, bonuses can be offered for compensation. For example, an

airline passenger might not only get on the next available Ã¯Â¬â€šight, but also

may be given a free ticket for a second Ã¯Â¬â€šight sometime in the future.

1270131 – Pearson Education Limited Ã‚Â©

Revenue Management

A specialized combination of the pricing and reservation options for

service providers is revenue management. Revenue management

(sometimes called yield management) is the process of varying price at the

right time for different customer segments to maximize revenues

generated from existing supply capacity. It works best if customers can be

segmented, prices can be varied by segment, Ã¯Â¬Âxed costs are high, variable

costs are low, service duration is predictable, and capacity is lost if not

used (sometimes called perishable capacity). Airlines, hotels, cruise lines,

restaurants (early-bird specials), and rental cars are good examples.

Computerized reservation systems can make hour-by-hour updates, using

decision rules for opening or closing price classes depending on the

difference between supply and continually updated demand forecasts. In

the airlines industry, prices are lowered if a particular airline Ã¯Â¬â€šight is not

selling as fast as expected, until more seats are booked. Alternately, if

larger than expected demand is developing, prices for the remaining seats

may be increased. Last-minute business travelers pay the higher prices,

whereas leisure travelers making reservations well in advance and staying

over the weekend get the bargain prices. Southwest Airlines now segments

its customers by creating a Ã¢â‚¬Å“Business SelectÃ¢â‚¬Â ticket class that rewards

more perks to frequent Ã¯Â¬â€šiers willing to pay higher prices.

revenue management

Varying price at the right time for different

customer segments to maximize revenues yielded

by existing supply capacity.

1270131 – Pearson Education Limited Ã‚Â©

Backlogs

Much like the appointments or reservations of service providers, a

backlog is an accumulation of customer orders that a manufacturer has

promised for delivery at some future date. Manufacturers in the supply

chain that maintain a backlog of orders as a normal business practice can

allow the backlog to grow during periods of high demand and then reduce

it during periods of low demand. Airplane manufacturers do not promise

instantaneous delivery, as do wholesalers or retailers farther forward in the

supply chain. Instead, they impose a lead time between when the order is

placed and when it is delivered. For example, an automotive parts

manufacturer may agree to deliver to the repair department of a car

dealership a batch of 100 door latches for a particular car model next

Tuesday. The parts manufacturer uses that due date to plan its production

of door latches within its capacity limits. Firms that are most likely to use

backlogsÃ¢â‚¬â€and increase the size of them during periods of heavy demandÃ¢â‚¬â€

make customized products and tend to have a make-to-order strategy.

Backlogs reduce the uncertainty of future production requirements and also

can be used to level demand. However, they become a competitive

disadvantage if they get too big.

backlog

An accumulation of customer orders that a

manufacturer has promised for delivery at some

future date.

1270131 – Pearson Education Limited Ã‚Â©

Backorders and Stockouts

A last resort in demand management is to set lower standards for

customer service, either in the form of backorders or stockouts. Not to be

confused with a backlog, a backorder is a customer order that cannot be

Ã¯Â¬Âlled when promised or demanded but is Ã¯Â¬Âlled later. Demand may be too

unpredictable or the item may be too costly to hold it in inventory. Although

the customer is not pleased with the delay, the customer order is not lost

and it is Ã¯Â¬Âlled at a later date. In contrast, a stockout is an order that

cannot be satisÃ¯Â¬Âed, resulting in a loss of the sale. A backorder adds to the

next periodÃ¢â‚¬â„¢s demand requirement, whereas a stockout does not.

Backorders and stockouts can lead dissatisÃ¯Â¬Âed customers to do their

future business with another Ã¯Â¬Ârm. Generally, backorders and stockouts are

to be avoided.

backorder

A customer order that cannot be Ã¯Â¬Âlled when

promised or demanded but is Ã¯Â¬Âlled later.

stockout

An order that cannot be satisÃ¯Â¬Âed, resulting in a loss

of the sale.

1270131 – Pearson Education Limited Ã‚Â©

Combinations of demand options can also be used. For example, a

manufacturer of lighting equipment had several products characterized as

Ã¢â‚¬Å“slow movers with spikes,Ã¢â‚¬Â where only 2 or 3 units were sold for several

weeks, and then suddenly there was a huge order for 10,000 units the next

week. The reason is that their product was purchased by commercial

property managers who might be upgrading the lighting in a large oÃ¯Â¬Æ’ce

building. The result was a forecasting nightmare and having to resort to

high cost supply options to meet the demand spikes. The breakthrough in

solving this problem was to combine the pricing and backlog options.

Contractors are now offered a 3 percent discount (the pricing option) on

any order in excess of 10,000 units that are placed Ã¯Â¬Âve or more weeks

before they are needed (the backlog option). The advanced warning allows

the manufacturer to smooth out its production processes, saving millions

of dollars annually.

1270131 – Pearson Education Limited Ã‚Â©

Key Decisions on Making

Forecasts

Before using forecasting techniques, a manager must make two decisions:

(1) what to forecast, and (2) what type of forecasting technique to select

for different items.

1270131 – Pearson Education Limited Ã‚Â©

Deciding What to Forecast

Although some sort of demand estimate is needed for the individual

services or goods produced by a company, forecasting total demand for

groups or clusters and then deriving individual service or product forecasts

may be easiest. Also, selecting the correct unit of measurement (e.g., units,

customers, or machine-hours) for forecasting may be as important as

choosing the best method.

Level of Aggregation

Few companies err by more than 5 percent when forecasting the annual

total demand for all their services or products. However, errors in forecasts

for individual items and shorter time periods may be much higher.

Recognizing this reality, many companies use a two-tier forecasting

system. They Ã¯Â¬Ârst cluster (or Ã¢â‚¬Å“roll upÃ¢â‚¬Â) several similar services or products

in a process called aggregation , making forecasts for families of

services or goods that have similar demand requirements and common

processing, labor, and materials requirements. Next, they derive forecasts

for individual items, which are sometimes called stock-keeping units. A

stock-keeping unit (SKU) is an individual item or product that has an

identifying code and is held in inventory somewhere along the supply chain,

such as in a distribution center.

aggregation

The act of clustering several similar services or

products so that forecasts and plans can be made

for whole families.

1270131 – Pearson Education Limited Ã‚Â©

Units of Measurement

Rather than using dollars as the initial unit of measurement, forecasts often

begin with service or product units, such as individual products, express

packages to deliver, or customers needing maintenance service or repairs

for their cars. Forecasted units can then be translated to dollars by

multiplying them by the unit price. If accurately forecasting demand for a

service or product is not possible in terms of number of units, forecast the

standard labor or machine-hours required of each of the critical resources.

1270131 – Pearson Education Limited Ã‚Â©

Choosing the Type of Forecasting

Technique

Forecasting systems offer a variety of techniques, and no one of them is

best for all items and situations. The forecasterÃ¢â‚¬â„¢s objective is to develop a

useful forecast from the information at hand with the technique that is

appropriate for the different patterns of demand. Two general types of

forecasting techniques are used: judgment methods and quantitative

methods. Judgment methods translate the opinions of managers, expert

opinions, consumer surveys, and salesforce estimates into quantitative

estimates. Quantitative methods include causal methods, time-series

analysis, and trend projection with regression. Causal methods use

historical data on independent variables, such as promotional campaigns,

economic conditions, and competitorsÃ¢â‚¬â„¢ actions, to predict demand. Timeseries analysis is a statistical approach that relies heavily on historical

demand data to project the future size of demand and recognizes trends

and seasonal patterns. Trend projection with regression is a hybrid

between a time-series technique and the causal method.

1270131 – Pearson Education Limited Ã‚Â©

1270131 – Pearson Education Limited Ã‚Â©

A Moto X phone manufactured by Motorola Mobility. Motorola considerably

improved its demand forecasting process by closely collaborating with its

major retailers, obtaining point-of-sale data from them.

Mark Lennihan/Associated Press

judgment methods

A forecasting method that translates the opinions

of managers, expert opinions, consumer surveys,

and salesforce estimates into quantitative

estimates.

causal methods

A quantitative forecasting method that uses

historical data on independent variables, such as

promotional campaigns, economic conditions, and

competitorsÃ¢â‚¬â„¢ actions, to predict demand.

time-series analysis

A statistical approach that relies heavily on

historical demand data to project the future size of

demand and recognizes trends and seasonal

patterns.

1270131 – Pearson Education Limited Ã‚Â©

trend projection with regression

A forecasting model that is a hybrid between a

time-series technique and the causal method.

1270131 – Pearson Education Limited Ã‚Â©

Forecast Error

For any forecasting technique, it is important to measure the accuracy of

its forecasts. Forecasts almost always contain errors. Random error results

from unpredictable factors that cause the forecast to deviate from the

actual demand. Forecasting analysts try to minimize forecast errors by

selecting appropriate forecasting models, but eliminating all forms of

errors is impossible.

Forecast error for a given period t is simply the difference found by

subtracting the forecast from actual demand, or

forecast error

The difference found by subtracting the forecast

from actual demand for a given period.

where

This equation (notice the alphabetical order with

coming before

) is

the starting point for creating several measures of forecast error that cover

longer periods of time.

There are Ã¯Â¬Âve basic measures of forecast error: CFE, MSE,

, MAD, and

MAPE. Figure 8.2 shows the output from the Error Analysis routine in

ForecastingÃ¢â‚¬â„¢s dropdown menu of POM for Windows. Part (a) gives a big

picture view of how well the forecast has been tracking the actual demand.

Part (b) shows the detailed calculations needed to obtain the summary

1270131 – Pearson Education Limited Ã‚Â©

error terms. Finally, Part (c) gives the summary error measures summarized

across all 10 time periods, as derived from Part (b).

1270131 – Pearson Education Limited Ã‚Â©

Cumulative Sum of Forecast Errors

The cumulative sum of forecast errors (CFE)

forecast error:

measures the total

cumulative sum of forecast errors (CFE)

A measurement of the total forecast error that

assesses the bias in a forecast.

CFE is a cumulative sum. Figure 8.2(b) shows that it is the sum of the

errors for all 10 periods. For any given period, it would be the sum of errors

up through that period. For example, it would be

for

period 2. CFE is also called the bias error and results from consistent

mistakesÃ¢â‚¬â€the forecast is always too high or too low. This type of error

typically causes the greatest disruption to planning efforts. For example, if

a forecast is consistently lower than actual demand, the value of CFE will

gradually get larger and larger. This increasingly large error indicates some

systematic deÃ¯Â¬Âciency in the forecasting approach. The average forecast

error, sometimes called the mean bias, is simply

1270131 – Pearson Education Limited Ã‚Â©

Figure 8.2(a) Graph of Actual and Forecast Demand Using Error Analysis

of Forecasting in POM for Windows

Figure 8.2(b) Detailed Calculations of Forecast Errors

1270131 – Pearson Education Limited Ã‚Â©

Dispersion of Forecast Errors

The mean squared error (MSE) , standard deviation of the errors

, and mean absolute deviation (MAD) measure the dispersion of

forecast errors attributed to trend, seasonal, cyclical, or random effects:

mean squared error (MSE)

A measurement of the dispersion of forecast

errors.

standard deviation of the errors ( )

A measurement of the dispersion of forecast

errors.

mean absolute deviation (MAD)

A measurement of the dispersion of forecast

errors.

Figure 8.2(b) shows the squared error in period 1 is 4, and MSE is 87.9

for the whole sample. The standard deviation of the errors, shown as

29.648 in Figure 8.2(b) , is calculated using a separate function available

in Excel. The absolute value of the error in period 2 is 6, and MAD is 8.1

across the whole sample.

1270131 – Pearson Education Limited Ã‚Â©

The mathematical symbol is used to indicate the absolute valueÃ¢â‚¬â€that is,

it tells you to disregard positive or negative signs. If MSE, , or MAD is

small, the forecast is typically close to actual demand; by contrast, a large

value indicates the possibility of large forecast errors. The measures do

differ in the way they emphasize errors. Large errors get far more weight in

MSE and because the errors are squared. MAD is a widely used measure

of forecast error and is easily understood; it is merely the mean of the

absolute forecast errors over a series of time periods, without regard to

whether the error was an overestimate or an underestimate.

1270131 – Pearson Education Limited Ã‚Â©

Mean Absolute Percent Error

The mean absolute percent error (MAPE) relates the forecast error to

the level of demand and is useful for putting forecast performance in the

proper perspective:

mean absolute percent error (MAPE)

A measurement that relates the forecast error to the

level of demand and is useful for putting forecast

performance in the proper perspective.

For example, an absolute forecast error of 100 results in a larger

percentage error when the demand is 200 units than when the demand is

10,000 units. MAPE is the best error measure to use when making

comparisons between time series for different SKUs. Looking again at

Figure 8.2(b) , the percent error in period 2 is 16.22 percent, and MAPE,

the average over all 10 periods, is 17.062 percent.

Finally, Figure 8.2(c) summarizes the key error terms across all 10 time

periods. They are actually found in selected portions of Figure 8.2(b) .

For example, CFE is

which is in the error column of Figure 8.2(b)

in the TOTALS row. MAD is 8.1, found in the

column and AVERAGE

row. Finally, MAPE is 17.062%, which is in the

column and

AVERAGE row.

Figure 8.2(c) Error Measures

1270131 – Pearson Education Limited Ã‚Â©

Example 8.1 Calculating Forecast Error Measures

The following table shows the actual sales of upholstered chairs for a

furniture manufacturer and the forecasts made for each of the last 8

months. Calculate CFE, MSE, , MAD, and MAPE for this product.

Month,

Demand,

Forecast,

Error,

t

Error,

Absolute

Absolut

Squared,

Error,

Percent

Error,

(100)

1

200

225

625

25

12.5%

2

240

220

20

400

20

8.3

3

300

285

15

225

15

5.0

4

270

290

400

20

7.4

5

230

250

400

20

8.7

6

260

240

400

20

7.7

7

210

250

1,600

40

19.0

8

275

240

1,225

35

12.7

5,275

195

Total

20

35

81.3%

1270131 – Pearson Education Limited Ã‚Â©

Solution

Using the formulas for the measures, we get

Cumulative forecast error (bias):

Average forecast error (mean bias):

Mean squared error:

Standard deviation of the errors:

Mean absolute deviation:

Mean absolute percent error:

A CFE of

indicates that the forecast has a slight bias to

overestimate demand. The MSE, , and MAD statistics provide

measures of forecast error variability. A MAD of 24.4 means that the

average forecast error was 24.4 units in absolute value. The value of ,

27.4, indicates that the sample distribution of forecast errors has a

standard deviation of 27.4 units. A MAPE of 10.2 percent implies that,

on average, the forecast error was within about 10 percent of actual

1270131 – Pearson Education Limited Ã‚Â©

demand. These measures become more reliable as the number of

periods of data increases.

Decision Point

Although reasonably satisÃ¯Â¬Âed with these forecast performance results,

the analyst decided to test out a few more forecasting methods before

reaching a Ã¯Â¬Ânal forecasting method to use for the future.

Computer Support

Computer support, such as from OM Explorer or POM for Windows, makes

error calculations easy when evaluating how well forecasting models Ã¯Â¬Ât

with past data. Errors are measured across past data, often called the

history Ã¯Â¬Âle in practice. They show the various error measures across the

entire history Ã¯Â¬Âle for each forecasting method evaluated. They also make

forecasts into the future, based on the method selected.

1270131 – Pearson Education Limited Ã‚Â©

Judgment Methods

Forecasts from quantitative methods are possible only when there is

adequate historical data, (i.e., the history Ã¯Â¬Âle). However, the history Ã¯Â¬Âle may

be nonexistent when a new product is introduced or when technology is

expected to change. The history Ã¯Â¬Âle might exist but be less useful when

certain events (such as rollouts or special packages) are reÃ¯Â¬â€šected in the

past data, or when certain events are expected to occur in the future. In

some cases, judgment methods are the only practical way to make a

forecast. In other cases, judgment methods can also be used to modify

forecasts that are generated by quantitative methods. They may recognize

that one or two quantitative models have been performing particularly well

in recent periods. Adjustments certainly would be called for if the

forecaster has important contextual knowledge. Contextual knowledge is

knowledge that practitioners gain through experience, such as cause-andeffect relationships, environmental cues, and organizational information

that may have an effect on the variable being forecast. Adjustments also

could account for unusual circumstances, such as a new sales promotion

or unexpected international events. They could also have been used to

remove the effect of special one-time events in the history Ã¯Â¬Âle before

quantitative methods are applied. Four of the more successful judgment

methods are as follows: (1) salesforce estimates, (2) executive opinion,

(3) market research, and (4) the Delphi method.

Salesforce estimates are forecasts compiled from estimates made

periodically by members of a companyÃ¢â‚¬â„¢s salesforce. The salesforce is the

1270131 – Pearson Education Limited Ã‚Â©

group most likely to know which services or products customers will be

buying in the near future and in what quantities. Forecasts of individual

salesforce members can be combined easily to get regional or national

sales estimates. However, individual biases of the salespeople may taint

the forecast. For example, some people are naturally optimistic, whereas

others are more cautious. Adjustments in forecasts may need to be made

to account for these individual biases.

salesforce estimates

The forecasts that are compiled from estimates of

future demands made periodically by members of

a companyÃ¢â‚¬â„¢s salesforce.

Executive opinion is a forecasting method in which the opinions,

experience, and technical knowledge of one or more managers or

customers are summarized to arrive at a single forecast. All of the factors

going into judgmental forecasts would fall into the category of executive

opinion. Executive opinion can also be used for technological

forecasting . The quick pace of technological change makes keeping

abreast of the latest advances diÃ¯Â¬Æ’cult.

1270131 – Pearson Education Limited Ã‚Â©

executive opinion

A forecasting method in which the opinions,

experience, and technical knowledge of one or

more managers are summarized to arrive at a

single forecast.

technological forecasting

An application of executive opinion to keep abreast

of the latest advances in technology.

Market research is a systematic approach to determine external

consumer interest in a service or product by creating and testing

hypotheses through data-gathering surveys. Conducting a market research

study includes designing a questionnaire, deciding how to administer it,

selecting a representative sample, and analyzing the information using

judgment and statistical tools to interpret the responses. Although market

research yields important information, it typically includes numerous

qualiÃ¯Â¬Âcations and hedges in the Ã¯Â¬Ândings.

1270131 – Pearson Education Limited Ã‚Â©

market research

A systematic approach to determine external

consumer interest in a service or product by

creating and testing hypotheses through datagathering surveys.

The Delphi method is a process of gaining consensus from a group of

experts while maintaining their anonymity. This form of forecasting is

useful when no historical data are available from which to develop

statistical models and when managers inside the Ã¯Â¬Ârm have no experience

on which to base informed projections. A coordinator sends questions to

each member of the group of outside experts, who may not even know who

else is participating. The coordinator prepares a statistical summary of the

responses along with a summary of arguments for particular responses.

The report is sent to the same group for another round, and the participants

may choose to modify their previous responses. These rounds continue

until consensus is obtained.

Delphi method

A process of gaining consensus from a group of

experts while maintaining their anonymity.

In the remainder of this chapter, we turn to the commonly used quantitative

forecasting approaches.

1270131 – Pearson Education Limited Ã‚Â©

1270131 – Pearson Education Limited Ã‚Â©

Causal Methods: Linear

Regression

Causal methods are used when historical data are available and the

relationship between the factor to be forecasted and other external or

internal factors (e.g., government actions or advertising promotions) can

be identiÃ¯Â¬Âed. These relationships are expressed in mathematical terms and

can be complex. Causal methods are good for predicting turning points in

demand and for preparing long-range forecasts. We focus on linear

regression, one of the best known and most commonly used causal

methods.

In linear regression , one variable, called a dependent variable, is related

to one or more independent variables by a linear equation. The dependent

variable (such as demand for door hinges) is the one the manager wants

to forecast. The independent variables (such as advertising

expenditures and new housing starts) are assumed to affect the dependent

variable and thereby Ã¢â‚¬Å“causeÃ¢â‚¬Â the results observed in the past. Figure 8.3

shows how a linear regression line relates to the data. In technical terms,

the regression line minimizes the squared deviations from the actual data.

1270131 – Pearson Education Limited Ã‚Â©

linear regression

A causal method in which one variable (the

dependent variable) is related to one or more

independent variables by a linear equation.

dependent variable

The variable that one wants to forecast.

independent variables

Variables that are assumed to affect the dependent

variable and thereby Ã¢â‚¬Å“causeÃ¢â‚¬Â the results observed in

the past.

1270131 – Pearson Education Limited Ã‚Â©

Figure 8.3 Linear Regression Line Relative to Actual Demand

In the simplest linear regression models, the dependent variable is a

function of only one independent variable and, therefore, the theoretical

relationship is a straight line:

where

1270131 – Pearson Education Limited Ã‚Â©

The objective of linear regression analysis is to Ã¯Â¬Ând values of a and b that

minimize the sum of the squared deviations of the actual data points from

the graphed line. Computer programs are used for this purpose. For any set

of matched observations for Y and X, the program computes the values of

a and b and provides measures of forecast accuracy. Three measures

commonly reported are (1) the sample correlation coeÃ¯Â¬Æ’cient, (2) the

sample coeÃ¯Â¬Æ’cient of determination, and (3) the standard error of the

estimate.

The sample correlation coeÃ¯Â¬Æ’cient, r, measures the direction and strength of

the relationship between the independent variable and the dependent

variable. The value of r can range from

A correlation

coeÃ¯Â¬Æ’cient of

implies that period-by-period changes in direction

(increases or decreases) of the independent variable are always

accompanied by changes in the same direction by the dependent variable.

An r of

means that decreases in the independent variable are

always accompanied by increases in the dependent variable, and vice

versa. A zero value of r means no linear relationship exists between the

variables. The closer the value of r is to

, the better the regression

line Ã¯Â¬Âts the points.

The sample coeÃ¯Â¬Æ’cient of determination measures the amount of variation

in the dependent variable about its mean that is explained by the regression

line. The coeÃ¯Â¬Æ’cient of determination is the square of the correlation

coeÃ¯Â¬Æ’cient, or

. The value of

equations with a value of

ranges from 0.00 to 1.00. Regression

close to 1.00 mean a close Ã¯Â¬Ât.

The standard error of the estimate,

, measures how closely the data on

the dependent variable cluster around the regression line. Although it is

similar to the sample standard deviation, it measures the error from the

dependent variable, Y, to the regression line, rather than to the mean. Thus,

1270131 – Pearson Education Limited Ã‚Â©

it is the standard deviation of the difference between the actual demand

and the estimate provided by the regression equation.

1270131 – Pearson Education Limited Ã‚Â©

Example 8.2 Using Linear Regression to Forecast

Product Demand

The supply chain manager seeks a better way to forecast the demand

for door hinges and believes that the demand is related to advertising

expenditures. The following are sales and advertising data for the past

5 months:

MyOMLab

Active Model 8.1 in MyOMLab provides insight on

varying the intercept and slope of the model.

Month

Sales (Thousands of Units)

Advertising (Thousands of $)

1

264

2.5

2

116

1.3

3

165

1.4

4

101

1.0

5

209

2.0

The company will spend $1,750 next month on advertising for the

product. Use linear regression to develop an equation and a forecast for

this product.

1270131 – Pearson Education Limited Ã‚Â©

Solution

We used POM for Windows to determine the best values of a, b, the

correlation coeÃ¯Â¬Æ’cient, the coeÃ¯Â¬Æ’cient of determination, and the standard

error of the estimate.

The regression equation is

and the regression line is shown in Figure 8.4 . The sample

correlation coeÃ¯Â¬Æ’cient, r, is 0.98, which is unusually close to 1.00 and

suggests an unusually strong positive relationship exists between sales

and advertising expenditures. The sample coeÃ¯Â¬Æ’cient of determination,

, implies that 96 percent of the variation in sales is explained by

advertising expenditures.

1270131 – Pearson Education Limited Ã‚Â©

Figure 8.4 Linear Regression Line for the Sales and Advertising Data

Using POM for Windows

Decision Point

The supply chain manager decided to use the regression model as input

to planning production levels for month 6. As the advertising

expenditure will be $1,750, the forecast for month 6 is

, or 183,016 units.

Often several independent variables may affect the dependent variable.

For example, advertising expenditures, new corporation start-ups, and

residential building contracts all may be important for estimating the

demand for door hinges. In such cases, multiple regression analysis is

helpful in determining a forecasting equation for the dependent variable

as a function of several independent variables. Such models can be

analyzed with POM for Windows or OM Explorer and can be quite useful

for predicting turning points and solving many planning problems.

1270131 – Pearson Education Limited Ã‚Â©

1270131 – Pearson Education Limited Ã‚Â©

Time-Series Methods

Rather than using independent variables for the forecast as regression

models do, time-series methods use historical information regarding only

the dependent variable. These methods are based on the assumption that

the dependent variableÃ¢â‚¬â„¢s past pattern will continue in the future. Time-series

analysis identiÃ¯Â¬Âes the underlying patterns of demand that combine to

produce an observed historical pattern of the dependent variable and then

develops a model to replicate it. In this section, we focus on Ã¯Â¬Âve statistical

time-series methods that address the horizontal, trend, and seasonal

patterns of demand: simple moving averages, weighted moving averages,

exponential smoothing, trend projection with regression, and multiplicative

seasonal method. Before we discuss statistical methods, let us take a look

at the simplest time-series method for addressing all patterns of demandÃ¢â‚¬â€

the naÃƒÂ¯ve forecast.

1270131 – Pearson Education Limited Ã‚Â©

NaÃƒÂ¯ve Forecast

A method often used in practice is the naÃƒÂ¯ve forecast , whereby the

forecast for the next period

equals the demand for the current

period

. So if the actual demand for Wednesday is 35 customers, the

forecasted demand for Thursday is 35 customers. Despite its name, the

naÃƒÂ¯ve forecast can perform well.

naÃƒÂ¯ve forecast

A time-series method whereby the forecast for the

next period equals the demand for the current

period, or

.

The naÃƒÂ¯ve forecast method may be adapted to take into account a demand

trend. The increase (or decrease) in demand observed between the last two

periods is used to adjust the current demand to arrive at a forecast.

Suppose that last week the demand was 120 units and the week before it

was 108 units. Demand increased 12 units in 1 week, so the forecast for

next week would be

. The naÃƒÂ¯ve forecast method

also may be used to account for seasonal patterns. If the demand last July

was 50,000 units, and assuming no underlying trend from one year to the

next, the forecast for this July would be 50,000 units. The method works

best when the horizontal, trend, or seasonal patterns are stable and random

variation is small.

1270131 – Pearson Education Limited Ã‚Â©

Horizontal Patterns: Estimating the

Average

We begin our discussion of statistical methods of time-series forecasting

with demand that has no apparent trend, seasonal, or cyclical patterns. The

horizontal pattern in a time series is based on the mean of the demands, so

we focus on forecasting methods that estimate the average of a time

series of data. The forecast of demand for any period in the future is the

average of the time series computed in the current period. For example, if

the average of past demand calculated on Tuesday is 65 customers, the

forecasts for Wednesday, Thursday, and Friday are 65 customers each day.

Consider Figure 8.5 , which shows patient arrivals at a medical clinic

over the past 28 weeks. Assuming that the time series has only a horizontal

and random pattern, one approach is simply to calculate the average of the

data. However, this approach has no adaptive quality if there is a trend,

seasonal, or cyclical pattern. The statistical techniques that do have an

adaptive quality in estimating the average in a time series are (1) simple

moving averages, (2) weighted moving averages, and (3) exponential

smoothing. Another option is the simple average, but it has no adaptive

capability.

1270131 – Pearson Education Limited Ã‚Â©

Figure 8.5 Weekly Patient Arrivals at a Medical Clinic

1270131 – Pearson Education Limited Ã‚Â©

Simple Moving Averages

The simple moving average method simply involves calculating the

average demand for the n most recent time periods and using it as the

forecast for future time periods. For the next period, after the demand is

known, the oldest demand from the previous average is replaced with the

most recent demand and the average is recalculated. In this way, the n

most recent demands are used, and the average Ã¢â‚¬Å“movesÃ¢â‚¬Â from period to

period.

simple moving average method

A time-series method used to estimate the average

of a demand time series by averaging the demand

for the n most recent time periods.

SpeciÃ¯Â¬Âcally, the forecast for period

can be calculated at the end of

period t (after the actual demand for period t is known) as

where

1270131 – Pearson Education Limited Ã‚Â©

Example 8.3 Using the Moving Average Method to

Estimate Average Demand

a. Compute a three-week moving average forecast for the arrival of

medical clinic patients in week 4. The numbers of arrivals for the

past 3 weeks were as follows:

Week

Patient Arrivals

1

400

2

380

3

411

b. If the actual number of patient arrivals in week 4 is 415, what is

the forecast error for week 4?

c. What is the forecast for week 5?

MyOMLab

Active Model 8.2 in MyOMLab provides insight on

the impact of varying n using the example in Figure

8.5 .

MyOMLab

Tutor 8.1 in MyOMLab provides another example to

practice making forecasts with the moving average

method.

1270131 – Pearson Education Limited Ã‚Â©

Solution

a. The moving average forecast at the end of week 3 is

b. The forecast error for week 4 is

c. The forecast for week 5 requires the actual arrivals from weeks 2

through 4, the 3 most recent weeks of data.

Decision Point

Thus, the forecast at the end of week 3 would have been 397 patients

for week 4, which fell short of actual demand by 18 patients. The

forecast for week 5, made at the end of week 4, would be 402 patients.

If a forecast is needed now for week 6 and beyond, it would also be for

402 patients.

The moving average method may involve the use of as many periods of

past demand as desired. Large values of n should be used for demand

series that are stable, and small values of n should be used for those that

are susceptible to changes in the underlying average. If n is set to its

lowest level (i.e., 1), it becomes the naÃƒÂ¯ve method.

1270131 – Pearson Education Limited Ã‚Â©

Weighted Moving Averages

In the simple moving average method, each demand has the same weight

in the averageÃ¢â‚¬â€namely,

. In the weighted moving average method ,

each historical demand in the average can have its own weight. The sum of

the weights equals 1.0. For example, in a three-period weighted moving

average model, the most recent period might be assigned a weight of 0.50,

the second most recent might be weighted 0.30, and the third most recent

might be weighted 0.20. The average is obtained by multiplying the weight

of each period by the value for that period and adding the products

together:

weighted moving average method

A time-series method in which each historical

demand in the average can have its own weight; the

sum of the weights equals 1.0.

For a numerical example of using the weighted moving average method to

estimate average demand, see Solved Problem 2 and Tutor 8.2 of OM

Explorer in MyOMLab.

The advantage of a weighted moving average method is that it allows you

to emphasize recent demand over earlier demand. (It can even handle

seasonal effects by putting higher weights on prior years in the same

season.) The forecast will be more responsive to changes in the underlying

average of the demand series than the simple moving average forecast.

1270131 – Pearson Education Limited Ã‚Â©

Exponential Smoothing

The exponential smoothing method is a sophisticated weighted moving

average method that calculates the average of a time series by implicitly

giving recent demands more weight than earlier demands, all the way back

to the Ã¯Â¬Ârst period in the history Ã¯Â¬Âle. It is the most frequently used formal

forecasting method because of its simplicity and the small amount of data

needed to support it. Unlike the weighted moving average method, which

requires n periods of past demand and n weights, exponential smoothing

requires only three items of data: (1) the last periodÃ¢â‚¬â„¢s forecast; (2) the

actual demand for this period; and (3) a smoothing parameter, alpha

,

which has a value between 0 and 1.0. The equation for the exponentially

smoothed forecast for period

is calculated

exponential smoothing method

A weighted moving average method that calculates

the average of a time series by implicitly giving

recent demands more weight than earlier demands.

The emphasis given to the most recent demand levels can be adjusted by

changing the smoothing parameter. Larger values emphasize recent

levels of demand and result in forecasts more responsive to changes in the

underlying average. Smaller values treat past demand more uniformly

and result in more stable forecasts. Smaller values are analogous to

increasing the value of n in the moving average method and giving greater

weight to past demand. In practice, various values of are tried and the

one producing the best forecasts is chosen.

1270131 – Pearson Education Limited Ã‚Â©

Exponential smoothing requires an initial forecast to get started. There are

several ways to get this initial forecast. OM Explorer and POM for Windows

use as a default setting the actual demand in the Ã¯Â¬Ârst period, which

becomes the forecast for the second period. Forecasts and forecast errors

then are calculated beginning with period 2. If some historical data are

available, the initial forecast can be found by calculating the average of

several recent periods of demand. The effect of the initial estimate of the

average on successive estimates of the average diminishes over time.

1270131 – Pearson Education Limited Ã‚Â©

Example 8.4 Using Exponential Smoothing to Estimate

Average Demand

a. Reconsider the patient arrival data in Example 8.3 . It is now

the end of week 3, so the actual number of arrivals is known to

be 411 patients. Using

, calculate the exponential

smoothing forecast for week 4.

b. What was the forecast error for week 4 if the actual demand

turned out to be 415?

c. What is the forecast for week 5?

MyOMLab

Active Model 8.2 in MyOMLab provides insight on

the impact of varying using the example in

Figure 8.5 .

MyOMLab

Tutor 8.3 in MyOMLab provides a new practice

example of how to make forecasts with the

exponential smoothing method.

1270131 – Pearson Education Limited Ã‚Â©

Solution

a. The exponential smoothing method requires an initial forecast.

Suppose that we take the demand data for the Ã¯Â¬Ârst 2 weeks and

average them, obtaining

as an initial

forecast. (POM for Windows and OM Explorer simply use the

actual demand for the Ã¯Â¬Ârst week as a default setting for the initial

forecast for period 1, and do not begin tracking forecast errors

until the second period). To obtain the forecast for week 4, using

exponential smoothing with

, and

, we calculate the forecast for week 4 as

Thus, the forecast for week 4 would be 392 patients.

b. The forecast error for week 4 is

c. The new forecast for week 5 would be

or 394 patients. Note that we used

, not the integer-value

forecast for week 4, in the computation for

. In general, we

round off (when it is appropriate) only the Ã¯Â¬Ânal result to maintain

as much accuracy as possible in the calculations.

Decision Point

Using this exponential smoothing model, the analystÃ¢â‚¬â„¢s forecasts would

have been 392 patients for week 4 and then 394 patients for week 5

and beyond. As soon as the actual demand for week 5 is known, then

the forecast for week 6 will be updated.

Because exponential smoothing is simple and requires minimal data, it is

inexpensive and attractive to Ã¯Â¬Ârms that make thousands of forecasts for

1270131 – Pearson Education Limited Ã‚Â©

each time period. However, its simplicity also is a disadvantage when the

underlying average is changing, as in the case of a demand series with a

trend. Like any method geared solely to the assumption of a stable

average, exponential smoothing results will lag behind changes in the

underlying average of demand. Higher values may help reduce forecast

errors when there is a change in the average; however, the lags will still

occur if the average is changing systematically. Typically, if large values

(e.g.,

) are required for an exponential smoothing application,

chances are good that another model is needed because of a signiÃ¯Â¬Âcant

trend or seasonal inÃ¯Â¬â€šuence in the demand series.

1270131 – Pearson Education Limited Ã‚Â©

Trend Patterns: Using Regression

Let us now consider a demand time series that has a trend. A trend in a

time series is a systematic increase or decrease in the average of the

series over time. Where a signiÃ¯Â¬Âcant trend is present, forecasts from naÃƒÂ¯ve,

moving average, and exponential smoothing approaches are adaptive, but

still lag behind actual demand and tend to be below or above the actual

demand.

Trend projection with regression is a forecasting model that accounts for

the trend with simple regression analysis. To develop a regression model

for forecasting the trend, let the dependent variable, Y, be a periodÃ¢â‚¬â„¢s

demand and the independent variable, t, be the time period. For the Ã¯Â¬Ârst

period, let

; for the second period, let

; and so on. The

regression equation is

One advantage of the trend projection with regression model is that it can

forecast demand well into the future. The previous models project demand

just one period ahead, and assume that demand beyond that will remain at

that same level. Of course, all of the models (including the trend projection

with regression model) can be updated each period to stay current. One

apparent disadvantage of the trend with regression model is that it is not

adaptive. The solution to this problem comes when you answer the

following question. If you had the past sales of Ford automobiles since

1920, would you include each year in your regression analysis, giving equal

weight to each yearÃ¢â‚¬â„¢s sales, or include just the sales for more recent years?

You most likely would decide to include just the more recent years, making

your regression model more adaptive. The trend projection with regression

model can thus be made more or less adaptive by the selection of

historical data periods to include in the same way that moving average

(changing n) or exponential smoothing (changing ) models do.

1270131 – Pearson Education Limited Ã‚Â©

The trend projection with regression model can be solved with either the

Trend Projection with Regression Solver or the Time Series Forecasting

Solver in OM Explorer. Both solvers provide the regression coeÃ¯Â¬Æ’cients,

coeÃ¯Â¬Æ’cient of determination

, error measures, and forecasts into the

future. POM for Windows has an alternative model (we do not cover in the

textbook, although a description is provided in MyOMLab) that includes the

trend, called the Trend-Adjusted Smoothing model.

MyOMLab

The Trend Projection with Regression Solver focuses exclusively on trend

analysis. Its graph gives a big-picture view of how well the model Ã¯Â¬Âts the

actual demand. Its sliders allow you to control when the regression begins,

how many periods are included in the regression analysis, and how many

periods you want forecasted into the future. The Time Series Forecasting

Solver, on the other hand, covers all time series models, including the trend

projection with regression. It also computes a combination forecast, which

we cover in a subsequent section on using multiple techniques.

1270131 – Pearson Education Limited Ã‚Â©

Example 8.5 Using Trend Projection with Regression to

Forecast a Demand Series with a Trend

Medanalysis, Inc., provides medical laboratory services to patients of

Health Providers, a group of 10 family-practice doctors associated with

a new health maintenance program. Managers are interested in

forecasting the number of blood analysis requests per week. Recent

publicity about the damaging effects of cholesterol on the heart has

caused a national increase in requests for standard blood tests. The

arrivals over the last 16 weeks are given in Table 8.1 . What is the

forecasted demand for the next three periods?

Table 8.1 Arrivals at Medanalysis for Last 16 Weeks

Week

Arrivals

Week

Arrivals

1

28

9

61

2

27

10

39

3

44

11

55

4

37

12

54

5

35

13

52

6

53

14

60

7

38

15

60

8

57

16

75

1270131 – Pearson Education Limited Ã‚Â©

Solution

Figure 8.6(a) shows the results using the Trend Projection with

Regression Solver when all 16 weeks are included in the regression

analysis, with Figure 8.6(b) showing the worksheet that goes with it.

Looking at the Results sheet of Figure 8.6(a) , we see that the Y

intercept of the trend line (a) is 28.50 and the slope of the line (b) is

2.35. Thus, the trend equation is

, where t is the time

period for which you are forecasting. The forecast for period 19 is

The error terms are

(which is to be

expected when the regression begins at the same time that error

analysis begins),

, and

. The coeÃ¯Â¬Æ’cient of determination

is

decent at 0.69. The trend line is rising gently and reaches 73 for period

19. Each period the forecast predicts an increase of 2.35 arrivals per

week.

1270131 – Pearson Education Limited Ã‚Â©

Figure 8.6(a) Trend Projection with Regression Results

1270131 – Pearson Education Limited Ã‚Â©

Figure 8.6(b) Detailed Calculations of Forecast Errors

1270131 – Pearson Education Limited Ã‚Â©

Seasonal Patterns: Using Seasonal

Factors

Seasonal patterns are regularly repeating upward or downward movements

in demand measured in periods of less than one year (hours, days, weeks,

months, or quarters). In this context, the time periods are called seasons.

For example, customer arrivals at a fast-food shop on any day may peak

between 11 . . and 1 . . and again from 5 . . to 7 . .

An easy way to account for seasonal effects is to use one of the

techniques already described, but to limit the data in the time series to

those time periods in the same season. For example, for a day-of-the-week

seasonal effect, one time series would be for Mondays, one for Tuesdays,

and so on. Such an approach accounts for seasonal effects, but has the

disadvantage of discarding considerable information on past demand.

Other methods are available that analyze all past data, using one model to

forecast demand for all of the seasons. We describe only the multiplicative

seasonal method , whereby an estimate of average demand is multiplied

by seasonal factors to arrive at a seasonal forecast. The four-step

procedure presented here involves the use of simple averages of past

demand, although more sophisticated methods for calculating averages,

such as a moving average or exponential smoothing approach, could be

used. The following description is based on a seasonal pattern lasting one

year and seasons of one month, although the procedure can be used for

any seasonal pattern and season of any length.

1270131 – Pearson Education Limited Ã‚Â©

multiplicative seasonal method

A method whereby seasonal factors are multiplied

by an estimate of average demand to arrive at a

seasonal forecast.

1. For each year, calculate the average demand per season by dividing

annual demand by the number of seasons per year.

2. For each year, divide the actual demand for a season by the average

demand per season. The result is a seasonal factor for each season

in the year, which indicates the level of demand relative to the

average demand. For example, a seasonal factor of 1.14 calculated

for April implies that AprilÃ¢â‚¬â„¢s demand is 14 percent greater than the

average demand per month.

3. Calculate the average seasonal factor for each season, using the

results from step 2. Add the seasonal factors for a season and

divide by the number of years of data.

4. Calculate each seasonÃ¢â‚¬â„¢s forecast for next year. Begin by forecasting

next yearÃ¢â‚¬â„¢s annual demand using the naÃƒÂ¯ve method, moving

averages, exponential smoothing, or trend projection with

regression. Then, divide annual demand by the number of seasons

per year to get the average demand per season. Finally, make the

seasonal forecast by multiplying the average demand per season by

the appropriate seasonal factor found in step 3.

1270131 – Pearson Education Limited Ã‚Â©

Example 8.6 Using the Multiplicative Seasonal Method to

Forecast the Number of Customers

The manager of the Stanley Steemer carpet cleaning company needs a

quarterly forecast of the number of customers expected next year. The

carpet cleaning business is seasonal, with a peak in the third quarter

and a trough in the Ã¯Â¬Ârst quarter. The manager wants to forecast

customer demand for each quarter of year 5, based on an estimate of

total year 5 demand of 2,600 customers.

1270131 – Pearson Education Limited Ã‚Â©

Solution

The following table calculates the seasonal factor for each week.

It shows the quarterly demand data from the past 4 years, as well as the

calculations performed to get the average seasonal factor for each

quarter.

YEAR 1

Quarter

Demand

YEAR 2

Seasonal Factor

Demand

YEAR 3

Seasonal Factor (2)

Demand

(1)

1

45

70

100

2

335

370

585

3

520

590

830

4

100

170

285

Total

1,000

1,200

1,800

Average

For example, the seasonal factor for quarter 1 in year 1 is calculated by

dividing the actual demand (45) by the average demand for the whole

year

. When this is done for all 4 years, we then can

average the seasonal factors for quarter 1 over all 4 years. The result is

a seasonal factor of 0.2043 for quarter 1.

Once seasonal factors are calculated for all four seasons (see last

column in the table on the previous page), we then turn to making the

forecasts for year 5. The manager suggests a forecast of 2,600

customers for the whole year, which seems reasonable given that the

annual demand has been increasing by an average of 400 customers

each year (from 1,000 in year 1 to 2,200 in year 4, or

1270131 – Pearson Education Limited Ã‚Â©

. The computed forecast demand is found by extending that trend, and

projecting an annual demand in year 5 of

customers. (This same result is conÃ¯Â¬Ârmed using the Trend Projection

with Regression Solver of OM Explorer.) The quarterly forecasts are

straight-forward. First, Ã¯Â¬Ând the average demand forecast for year 5,

which is

Then multiple this average demand by the

average seasonal index, giving us

Quarter

Forecast

1

2

3

4

Figure 8.7 shows the computer solution using the Seasonal

Forecasting Solver in OM Explorer. Figure 8.7(b) , the results,

conÃ¯Â¬Ârms all of the calculations made above. Notice in Figure 8.7(a) ,

the inputs sheet that a computer demand forecast is provided as a

default for year 5. However, there is an option for user-supplied demand

forecast that overrides the computer-supplied forecast if the manager

wishes to make a judgmental forecast based on additional information.

1270131 – Pearson Education Limited Ã‚Â©

Figure 8.7 Demand Forecasts Using the Seasonal Forecasting Solver

of OM Explorer

Decision Point

Using this seasonal method, the analyst makes a demand forecast as

low as 133 customers in the Ã¯Â¬Ârst quarter and as high as 1,300

customers in the third quarter. The season of the year clearly makes a

difference.

An alternative to the multiplicative seasonal method is the additive

seasonal method , whereby seasonal forecasts are generated by adding

or subtracting a seasonal constant (say, 50 units) to the estimate of

average demand per season. This approach is based on the assumption

that the seasonal pattern is constant, regardless of average demand. The

1270131 – Pearson Education Limited Ã‚Â©

amplitude of the seasonal adjustment remains the same regardless of the

level of demand.

additive seasonal method

A method in which seasonal forecasts are

generated by adding a constant to the estimate of

average demand per season.

1270131 – Pearson Education Limited Ã‚Â©

Criteria for Selecting Time-Series

Methods

Of all the time series forecasting methods available, which should be

chosen? Forecast error measures provide important information for

choosing the best forecasting method for a service or product. They also

guide managers in selecting the best values for the parameters needed for

the method: n for the moving average method, the weights for the weighted

moving average method, for the exponential smoothing method, and

when regression data begins for the trend projection with regression

method. The criteria to use in making forecast method and parameter

choices include (1) minimizing bias (CFE); (2) minimizing MAPE, MAD, or

MSE; (3) maximizing

for trend projections using regression; (4) using a

holdout sample analysis; (5) using a tracking signal; (6) meeting

managerial expectations of changes in the components of demand; and (7)

minimizing the forecast errors in recent periods. The Ã¯Â¬Ârst three criteria

relate to statistical measures based on historical performance, the fourth is

a test under realistic conditions, the Ã¯Â¬Âfth evaluates forecast performance

and the potential need to change the method, the sixth reÃ¯Â¬â€šects

expectations of the future that may not be rooted in the past, and the

seventh is a way to use whatever method seems to be working best at the

time a forecast must be made.

1270131 – Pearson Education Limited Ã‚Â©

Using Statistical Criteria

Statistical performance measures can be used in the selection of which

forecasting method to use. The following guidelines will help when

searching for the best time-series models:

1. For projections of more stable demand patterns, use lower values

or larger n values to emphasize historical experience.

2. For projections of more dynamic demand patterns using the models

covered in this chapter, try higher values or smaller n values. When

historical demand patterns are changing, recent history should be

emphasized.

1270131 – Pearson Education Limited Ã‚Â©

Using a Holdout Sample

Often, the forecaster must make trade-offs between bias (CFE) and the

measures of forecast error dispersion (MAPE, MAD, and MSE). Managers

also must recognize that the best technique in explaining the past data is

not necessarily the best technique to predict the future, and that

Ã¢â‚¬Å“overÃ¯Â¬ÂttingÃ¢â‚¬Â past data can be deceptive. A forecasting method may have

small errors relative to the history Ã¯Â¬Âle, but may generate high errors for

future time periods. For this reason, some analysts prefer to use a holdout

sample as a Ã¯Â¬Ânal test (see Experiential Learning Exercise 8.1 at the

end of this chapter). To do so, they set aside some of the more recent

periods from the time series and use only the earlier time periods to

develop and test different models. Once the Ã¯Â¬Ânal models have been

selected in the Ã¯Â¬Ârst phase, they are tested again with the holdout sample.

Performance measures, such as MAD and CFE, would still be used but they

would be applied to the holdout sample. Whether this idea is used or not,

managers should monitor future forecast errors, and modify their

forecasting approaches as needed. Maintaining data on forecast

performance is the ultimate test of forecasting powerÃ¢â‚¬â€rather than how well

a model Ã¯Â¬Âts past data or holdout samples.

holdout sample

Actual demands from the more recent time periods

in the time series that are set aside to test different

models developed from the earlier time periods.

1270131 – Pearson Education Limited Ã‚Â©

Using a Tracking Signal

A tracking signal is a measure that indicates whether a method of

forecasting is accurately predicting actual changes in demand. The

tracking signal measures the number of MADs represented by the

cumulative sum of forecast errors, the CFE. The CFE tends to be close to 0

when a correct forecasting system is being used. At any time, however,

random errors can cause the CFE to be a nonzero number. The tracking

signal formula is

tracking signal

A measure that indicates whether a method of

forecasting is accurately predicting actual changes

in demand.

Each period, the CFE and MAD are updated to reÃ¯Â¬â€šect current error, and the

tracking signal is compared to some predetermined limits. The MAD can

be calculated in one of two ways: (1) as the simple average of all absolute

errors (as demonstrated in Example 8.1 ) or (2) as a weighted average

determined by the exponential smoothing method:

If forecast errors are normally distributed with a mean of 0, the relationship

between and MAD is simple:

1270131 – Pearson Education Limited Ã‚Â©

where

This relationship allows use of the normal probability tables to specify

limits for the tracking signal. If the tracking signal falls outside those

limits, the forecasting model no longer is tracking demand adequately. A

tracking system is useful when forecasting systems are computerized

because it alerts analysts when forecasts are getting far from desirable

limits. Figure 8.8 shows tracking signal results for 23 periods plotted on

a control chart. The control chart is useful for determining whether any

action needs to be taken to improve the forecasting model. In the example,

the Ã¯Â¬Ârst 20 points cluster around 0, as we would expect if the forecasts are

not biased. The CFE will tend toward 0. When the underlying characteristics

of demand change but the forecasting model does not, the tracking signal

eventually goes out of control. The steady increase after the 20th point in

Figure 8.8 indicates that the process is going out of control. The 21st

and 22nd points are acceptable, but the 23rd point is not.

1270131 – Pearson Education Limited Ã‚Â©

Figure 8.8 Tracking Signal

1270131 – Pearson Education Limited Ã‚Â©

Forecasting as a Process

Often companies must prepare forecasts for hundreds or even thousands

of services or products repeatedly. For example, a large network of health

care facilities must calculate demand forecasts for each of its services for

every department. This undertaking involves voluminous data that must be

manipulated frequently. However, software can ease the burden of making

these forecasts and coordinating the forecasts between customers and

suppliers. Many forecasting software packages are available, including

Manugistics, Forecast Pro, and SAS. The forecasting routines in OM

Explorer and POM for Windows give some hint of their capabilities.

Forecasting is not just a set of techniques, but instead a process that must

be designed and managed. While there is no one process that works for

everyone, here we describe two comprehensive processes that can be quite

effective in managing operations and the supply chain.

1270131 – Pearson Education Limited Ã‚Â©

A Typical Forecasting Process

Many inputs to the forecasting process are informational, beginning with

the history Ã¯Â¬Âle on past demand. The history Ã¯Â¬Âle is kept up-to-date with the

actual demands. Clarifying notes and adjustments are made to the

database to explain unusual demand behavior, such as the impact of

special promotions and closeouts. Often the database is separated into

two parts: base data and nonbase data. The second category reÃ¯Â¬â€šects

irregular demands. Final forecasts just made at the end of the prior cycle

are entered in the history Ã¯Â¬Âle so as to track forecast errors. Other

information sources are from salesforce estimates, outstanding bids on

new orders, booked orders, market research studies, competitor behavior,

economic outlook, new product introductions, pricing, and promotions. If

point-of-sale data are used, as is done by Kimberly-Clark in the opening

vignette, then considerable information sharing will take place with

customers. For new products, a history database is fabricated based on the

Ã¯Â¬ÂrmÃ¢â‚¬â„¢s experience with prior products and the judgment of personnel.

Outputs of the process are forecasts for multiple time periods into the

future. Typically, they are on a monthly basis and are projected out from six

months to two years. Most software packages have the ability to Ã¢â‚¬Å“roll upÃ¢â‚¬Â

or Ã¢â‚¬Å“aggregateÃ¢â‚¬Â forecasts for individual stock-keeping units (SKUs) into

forecasts for whole product families. Forecasts can also be Ã¢â‚¬Å“blown downÃ¢â‚¬Â

or Ã¢â‚¬Å“disaggregatedÃ¢â‚¬Â into smaller pieces. In a make-to-stock environment,

forecasts tend to be more detailed and can get down to speciÃ¯Â¬Âc individual

products. In a make-to-order environment, the forecasts tend to be for

groups of products. Similarly, if the lead times to buy raw materials and

manufacture a product or provide a service are long, the forecasts go

farther out into the future.

1270131 – Pearson Education Limited Ã‚Â©

The forecast process itself, typically done on a monthly basis, consists of

structured steps. These steps often are facilitated by someone who might

be called a demand manager, forecast analyst, or demand/supply planner.

However, many other people are typically involved before the plan for the

month is authorized.

Step 1. The cycle begins mid-month just after the forecasts have

been Ã¯Â¬Ânalized and communicated to the stakeholders. Now is the

time to update the history Ã¯Â¬Âle and review forecast accuracy. At the

end of the month, enter actual demand and review forecast accuracy.

Step 2. Prepare initial forecasts using some forecasting software

package and judgment. Adjust the parameters of the software to

Ã¯Â¬Ând models that Ã¯Â¬Ât the past demand well and yet reÃ¯Â¬â€šect the demand

managerÃ¢â‚¬â„¢s judgment on irregular events and information about future

sales pulled from various sources and business units.

Step 3. Hold consensus meetings with the stakeholders, such as

marketing, sales, supply chain planners, and Ã¯Â¬Ânance. Make it easy

for business unit and Ã¯Â¬Âeld sales personnel to make inputs. Use the

Internet to get collaborative information from key customers and

suppliers. The goal is to arrive at consensus forecasts from all of

the important players.

Step 4. Revise the forecasts using judgment, considering the inputs

from the consensus meetings and collaborative sources.

Step 5. Present the forecasts to the operating committee for review

and to reach a Ã¯Â¬Ânal set of forecasts. It is important to have a set of

forecasts that everybody agrees upon and will work to support.

Step 6. Finalize the forecasts based on the decisions of the

operating committee and communicate them to the important

stakeholders. Supply chain planners are usually the biggest users.

1270131 – Pearson Education Limited Ã‚Â©

As with all work activity, forecasting is a process and should be continually

reviewed for improvements. A better process will foster better

relationships between departments such as marketing, sales, and

operations. It will also produce better forecasts. This principle is the Ã¯Â¬Ârst

one in Table 8.2 to guide process improvements.

1270131 – Pearson Education Limited Ã‚Â©

Table 8.2 Some Principles for the Forecasting Process

Better processes yield better forecasts.

Demand forecasting is being done in virtually every company, either formally

or informally. The challenge is to do it wellÃ¢â‚¬â€better than the competition.

Better forecasts result in better customer service and lower costs, as well as

better relationships with suppliers and customers.

The forecast can and must make sense based on the big picture, economic

outlook, market share, and so on.

The best way to improve forecast accuracy is to focus on reducing forecast

error.

Bias is the worst kind of forecast error; strive for zero bias.

Whenever possible, forecast at more aggregate levels. Forecast in detail only

where necessary.

Far more can be gained by people collaborating and communicating well

than by using the most advanced forecasting technique or model.

Source: From Thomas F. Wallace and Robert A. Stahl, Sales Forecasting: A New Approach (Cincinnati, OH: T. E. Wallace &

Company, 2002), p. 112. Copyright Ã‚Â© 2002 T.E. Wallace & Company. Used with permission.

1270131 – Pearson Education Limited Ã‚Â©

Using Multiple Forecasting Methods

Step 2 of the forecasting process relates to preparing an initial forecast.

However, we need not rely on a single forecasting method. Several different

forecasts can be used to arrive at a forecast. Initial statistical forecasts

using several time-series methods and regression are distributed to

knowledgeable individuals, such as marketing directors and sales teams,

(and sometimes even suppliers and customers) for their adjustments. They

can account for current market and customer conditions that are not

necessarily reÃ¯Â¬â€šected in past data. Multiple forecasts may come from

different sales teams, and some teams may have a better record on

forecast errors than others.

Research during the last two decades suggests that combining forecasts

from multiple sources often produces more accurate forecasts.

Combination forecasts are forecasts that are produced by averaging

independent forecasts based on different methods, different sources, or

different data. It is intriguing that combination forecasts often perform

better over time than even the best single forecasting procedure. For

example, suppose that the forecast for the next period is 100 units from

technique 1 and 120 units from technique 2 and that technique 1 has

provided more accurate forecasts to date. The combination forecast for

next period, giving equal weight to each technique, is 110 units (or

). When this averaging technique is used

consistently into the future, its combination forecasts often will be much

more accurate than those of any single best forecasting technique (in this

example, technique 1). Combining is most effective when the individual

forecasts bring different kinds of information into the forecasting process.

Forecasters have achieved excellent results by weighting forecasts equally,

and this is a good starting point. However, unequal weights may provide

1270131 – Pearson Education Limited Ã‚Â©

better results under some conditions. Managerial Practice 8.1 shows how

Fiskars Corporation successfully used combination forecasts.

combination forecasts

Forecasts that are produced by averaging

independent forecasts based on different methods,

different sources, or different data.

1270131 – Pearson Education Limited Ã‚Â©

Managerial Practice 8.1 Combination

Forecasts and the Forecasting Process

Fiskars Corporation, which generated more than $1.1 billion in sales

in 2013, is the second oldest incorporated entity in the world and

produces a variety of high-quality products such as garden shears,

pruners, hand tools, scissors, ratchet tools, screwdrivers, and the

like. Business is highly seasonal and prices quite variable. About 10

percent to 15 percent of the annual revenue comes from one-time

promotions, and 25 percent to 35 percent of its products are new

every year. Quality is very important at Fiskars; its scissors were

selected as the OÃ¯Â¬Æ’cial Net-Cutting Scissors of the NCAA National

Championship in 2014.

1270131 – Pearson Education Limited Ã‚Â©

Fiskars Brands, Inc. totally overhauled its forecasting process for

products such as those shown here. It introduced time-series

methods, with much emphasis placed on combination forecasts.

Sales staff added their judgmental modiÃ¯Â¬Âcations, which were

combined with forecasts from several time-series techniques to

produce more accurate forecasts down to the level of the product.

HANDOUT/MCT/Newscom

Given the highly volatile demand environment, Fiskars Brands, Inc., a

subsidiary of Fiskars Corporation located in Madison, Wisconsin,

needed to improve its forecasting process. It serves 2,000

customers ranging from large discounters to local craft stores

providing about 2,300 products. Fiskars Brands introduced a

statistical-based analysis in its forecasting process along with a

Web-based business intelligence tool for reporting. It put much

1270131 – Pearson Education Limited Ã‚Â©

more emphasis on combination forecasts. Instead of asking

members of the sales staff to provide their own forecasts, forecasts

were sent to them, and they were asked for their validation and

reÃ¯Â¬Ânement. Their inputs are most useful relative to additions,

deletions, and promotions. Converting multiple forecasts into one

number (forecasts from time-series techniques, sales input, and

customer input) creates more accurate forecasts by product.

FiskarsÃ¢â‚¬â„¢s software has the ability to weigh each input. It gives more

weight to a statistical forecast for in-line items, and inputs from the

sales staff get much more weight for promoted products and new

items.

It also segments products by value and forecastability so as to

concentrate forecasting efforts on products that have the biggest

impact on the business. High-value items that are easier to forecast

(stable demand with low forecast errors to date) tend to do well

with the time-series techniques, and judgmental adjustments are

made with caution. High-value items that are diÃ¯Â¬Æ’cult to forecast get

top priority in the forecasting effort, and spark the need for

collaboration with customers and suppliers. Much less attention is

given to improving forecasts for low-value items for which there is

some history and fairly steady demand.

judgmental adjustment

An adjustment made to forecasts from one

or more quantitative models that accounts

for recognizing which models are

performing particularly well in recent past, or

take into account contextual information.

1270131 – Pearson Education Limited Ã‚Â©

Finally, Fiskars instituted a Web-based program that gives the entire

company visibility to forecast information in whatever form it needs.

For example, Finance wants monthly, quarterly, and yearly

projections in dollars, whereas Operations wants projections in units

as well as accuracy measures. Everybody can track updated

forecast information by customer, brand, and product.

Source: David Montgomery, Ã¢â‚¬Å“Flashpoints for Changing Your Forecasting Process,Ã¢â‚¬Â The Journal of Business

Forecasting, (Winter 2006Ã¢â‚¬â€œ2007), pp. 35Ã¢â‚¬â€œ37; http://www.Ã¯Â¬Âskars.com, April 15, 2014.

OM Explorer and POM for Windows allow you to evaluate several

forecasting models, and then you can create combination forecasts from

them. In fact, the Time-Series Forecasting Solver of OM Explorer

automatically computes a combination forecast as a weighted average,

using the weights that you supply for the various models that it evaluates.

The models include the naÃƒÂ¯ve, moving average, exponential smoothing, and

regression projector methods. Alternately, you can create a simple Excel

spreadsheet that combines forecasts generated by POM for Windows to

create combination forecasts. The Time Series Forecasting Solver also

allows you evaluate your forecasting process with a holdout sample. The

forecaster makes a forecast just one period ahead, and learns of given

actual demand. Next the solver computes forecasts and forecast errors for

the period. The process continues to the next period in the holdout sample

with the forecaster committing to a forecast for the next period. To be

informed, the forecaster should also be aware of how well the other

forecasting methods have been performing, particularly in the recent past.

Another way to take advantage of multiple techniques is focus

forecasting , which selects the best forecast (based on past error

measures) from a group of forecasts generated by individual techniques.

Every period, all techniques are used to make forecasts for each item. The

forecasts are made with a computer because there can be 100,000 SKUs at

a company, each needing to be forecast. Using the history Ã¯Â¬Âle as the

1270131 – Pearson Education Limited Ã‚Â©

starting point for each method, the computer generates forecasts for the

current period. The forecasts are compared to actual demand, and the

method that produces the forecast with the least error is used to make the

forecast for the next period. The method used for each item may change

from period to period.

focus forecasting

A method of forecasting that selects the best

forecast from a group of forecasts generated by

individual techniques.

1270131 – Pearson Education Limited Ã‚Â©

Adding Collaboration to the Process

In step three of the forecasting process we try to achieve consensus of the

forecast. One way to achieve that consensus in a formal way is to employ

collaborative planning, forecasting, and replenishment (CPFR) , a

process for supply chain integration that allows a supplier and its

customers to collaborate on making the forecast by using the Internet.

Traditionally, suppliers and buyers in most supply chains prepare

independent demand forecasts. With CPFR, Ã¯Â¬Ârms initiate customerfocused operations teams that share with retailers their real-time data and

plans, including forecasts, inventories, sales to retailersÃ¢â‚¬â„¢ shelves,

promotions, product plans, and exceptions. CPFR involves four interactive

activities:

collaborative planning, forecasting, and replenishment

(CPFR)

A process for supply chain integration that allows a

supplier and its customers to collaborate on

making the forecast by using the Internet.

Strategy and Planning: establish the ground rules for the collaborative

relationship such as business goals, scope of collaboration, and

assignment of roles and responsibilities.

Demand and Supply Management: develop sales forecasts, procedures

for order planning, and inventory positions.

Execution: manage the generation of orders between supplier and

customers and the production, shipment, and delivery of products for

1270131 – Pearson Education Limited Ã‚Â©

customer purchase.

Analysis: monitor the planning process and operations for out of bound

conditions and evaluate to achievement of business goals.

Many Ã¯Â¬Ârms have used CPFR to coordinate forecasts and plans up and

down the supply chain. CPFR enables Ã¯Â¬Ârms to collaborate with their

retailersÃ¢â‚¬â„¢ distribution centersÃ¢â‚¬â„¢ customers and increase their ability to

forecast effectively. The real key to a successful implementation of CPFR

is the forging of a cultural alliance that involves peer-to peer relations and

cross-functional teams.

1270131 – Pearson Education Limited Ã‚Â©

Ever wonder how CPFR came into existence? Walmart has long been known

for its careful analysis of cash register receipts and for working with

suppliers to reduce inventories. To combat the ill effects of forecast errors

on inventories, Benchmarking Partners, Inc. was funded in the mid-1990s by

Walmart, IBM, SAP, and Manugistics to develop a software package.

Walmart initiated the new system with Listerine, a primary product of

Warner-Lambert (now produced by Johnson & Johnson). How did it work?

Walmart and Warner-Lambert independently calculated the demand they

expected for Listerine six months into the future, taking into consideration

factors such as past sales trends and sales promotions. If the forecasts

differed by more than a predetermined percentage, they exchanged written

comments and supporting data. They went through as many cycles as

needed to converge to an acceptable forecast. The program was

successful; Walmart saw a reduction in stockouts from 15 percent to 2

percent, increased sales, and reduced inventory costs, while WarnerLambert beneÃ¯Â¬Âtted from a smoother production plan and lower average

1270131 – Pearson Education Limited Ã‚Â©

costs. The system was later generalized and called collaborative planning,

forecasting, and replenishment, or CPFR.

Mark Peterson/CORBIS

Forecasting as a Nested Process

Forecasting is not a stand-alone activity, but instead part of a larger

process that encompasses the remaining chapters. After all, demand is

only half of the equationÃ¢â‚¬â€the other half is supply. Future plans must be

developed to supply the resources needed to meet the forecasted demand.

Resources include the workforce, materials, inventories, dollars, and

equipment capacity. Making sure that demand and supply plans are in

balance begins in the next chapter, Chapter 10 , Ã¢â‚¬Å“Inventory Management,Ã¢â‚¬Â

and continues with Chapter 11 , Ã¢â‚¬Å“Operations Planning and Scheduling,Ã¢â‚¬Â

and Chapter 12 , Ã¢â‚¬Å“Resource Planning.Ã¢â‚¬Â

1270131 – Pearson Education Limited Ã‚Â©

6 Lean Systems

1270131 – Pearson Education Limited Ã‚Â©

An Aldi discount grocery store.

Kristoffer Tripplaar/Alamy

1270131 – Pearson Education Limited Ã‚Â©

Aldi

Aldi is a discount supermarket chain with headquarters in

Germany and over 8000 stores worldwide including Australia,

Europe, Great Britain, Ireland, and the United States. With

roots and distribution in several countries throughout Europe,

it is a different kind of retailer that prides itself in displaying

key dietary and nutritional information on the front of their

packaging to enable customers to make informed choices

about their food. Aldi also makes its packaging from recycled

materials to keep the planet green. Its emphasis on core

values of simplicity, consistency, and corporate responsibility

are closely tied to the principles of lean production, which Aldi

uses to keep costs down in all areas, provide customers

more value for their money, and remain more competitive in a

business with razor thin margins.

AldiÃ¢â‚¬â„¢s waste reduction efforts start with training its

employees to do many different tasks, which improves

Ã¯Â¬â€šexibility and lowers staff costs. In addition, consistent with

total quality management (TQM) principles, all workers have

the responsibility to get it right the Ã¯Â¬Ârst time, whether it is

accurate pricing or ordering the appropriate replenishment

stocks. In return, they are paid some of the better wages in

the United States in the grocery industry. In the stores, all

items have bar codes in a number of places to save time in

Ã¯Â¬Ânding them, which makes the checkout process more

eÃ¯Â¬Æ’cient. Aldi is also known for having smaller stores, which

are made possible by the fact that it sells fewer variations of

1270131 – Pearson Education Limited Ã‚Â©

each product and so less space is used for display. It also

means that Aldi can get quantity discounts and economies of

scale in sourcing products. Holding only the stock that is

needed for each product is further facilitated through a justin-time ordering and delivery system. Products are delivered

as needed in display-ready cases; some of them are even

sold directly from a pallet or a platform to minimize handling

and increase the eÃ¯Â¬Æ’ciency of getting a large volume into the

store quickly. In contrast to several of its 24-hour

competitors, Aldi stores are only open from 8 . . to 8 . .

on most days, which reduces the use of energy and staff

salary costs. By limiting the use of credit cards, except

Discover in some stores in the United States and Visa and

MasterCard in Ireland, and using only cash or debit cards

saves Aldi the surcharge fees levied by most credit card

companies. Finally, AldiÃ¢â‚¬â„¢s shopping carts utilize a 25Ã‚Â¢ (in the

United States) or a Ã¢â€šÂ¬1 (in Europe) coin system to make sure

that customers return them to the parking stations near the

store, which saves labor costs in collecting the carts that

would otherwise be left scattered across the parking lots or

potentially be lost or stolen.

1270131 – Pearson Education Limited Ã‚Â©

Using Operations to Create Value

Managing Processes

Process Strategy and Analysis

Quality and Performance

Capacity Planning

Constraint Management

Lean Systems

Project Management

Managing Customer Demand

Forecasting

Inventory Management

Operations Planning and Scheduling

Resource Planning

Managing Supply Chains

Supply Chain Design

Supply Chain Logistic Networks

Supply Chain Integration

Supply Chain Sustainability

AldiÃ¢â‚¬â„¢s lean philosophy extends into the supply chain as well.

Up to 60 percent of its fruits and vegetables are sourced

locally to save on transportation costs and time. As part of

its inventory reduction policies, suppliers are not allowed to

hold more than one month of normal orders and

1270131 – Pearson Education Limited Ã‚Â©

requirements of AldiÃ¢â‚¬â„¢s private label products in inventory at

any given point of time, unless Aldi submits a written

authorization for a temporary or permanent change in

suppliersÃ¢â‚¬â„¢ inventory levels. Due to its relentless focus on lean

principles, it is no wonder that Aldi is a clear leader in prices

among leading grocery brands according to a study of 6,200

consumers conducted in May 2014 by Market Force

Information, a customer intelligence solutions Ã¯Â¬Ârm. AldiÃ¢â‚¬â„¢s

products can be as much as 30 percent cheaper than its

competitors in some cases. In addition, Publix and Aldi were

ranked second and third in customer satisfaction in North

America after Trader JoeÃ¢â‚¬â„¢s due to their courteous service, fast

checkouts, and the quality of their private label brand

products. All these initiatives have contributed to AldiÃ¢â‚¬â„¢s

explosive growth globallyÃ¢â‚¬â€a new store opens roughly every

week in the United Kingdom alone.

Sources: Ã¢â‚¬Å“Competitive Advantage through EÃ¯Â¬Æ’ciency: An Aldi Case Study,Ã¢â‚¬Â http://businesscasestudies.co.uk/aldi/competitiveadvantage-through-eÃ¯Â¬Æ’ciency/introduction.html#axzz39GsjgbiQ; http://en.wikipedia.org/wiki/Aldi; https://corporate.aldi.us; http:/

/www.producenews.com/news-dep-menu/test-featured/13168-consumer-study-reveals-top-grocery-stores (August 2, 2014).

1270131 – Pearson Education Limited Ã‚Â©

Learning Goals

After reading this chapter, you should be able to:

1. Describe how lean systems can facilitate the

continuous improvement of processes.

2. Identify the strategic supply chain and process

characteristics of lean systems.

3. Explain the differences between one-worker, multiplemachine (OWMM) and group technology (GT)

approaches to lean system layouts.

4. Understand value stream mapping and its role in

waste reduction.

5. Understand kanban systems for creating a production

schedule in a lean system.

. Explain the implementation issues associated with

the application of lean systems.

1270131 – Pearson Education Limited Ã‚Â©

Aldi is a learning organization and an excellent example of an approach for

designing supply chains known as lean systems, which allow Ã¯Â¬Ârms like Aldi

to continuously improve its operations and spread the lessons learned

across the entire corporation. Lean systems are operations systems that

maximize the value added by each of a companyÃ¢â‚¬â„¢s activities by removing

waste and delays from them. They encompass the companyÃ¢â‚¬â„¢s operations

strategy, process design, quality management, constraint management,

layout design, supply chain design, and technology and inventory

management and can be used by both service …

Purchase answer to see full

attachment