Capacity Management

Overview

 Introduction and context

 Three steps of capacity management

 Strategies for managing capacity

 Level, chase and demand management

 Capacity in service operations

 Yield management

 Conclusions
 Capacity Management

To provide an ‘appropriate’ amount of capacity at any point in time

The ‘appropriateness’ of capacity planning in any part of the

operation can be judged by its effect on …

Costs
Revenue
Working capital
Service level
Source: British Airways London Eye
 Capacity
 Capacity = “the maximum level of value-added activity that an operation, or
process, or facility is capable of over a period of time” (Slack et al, 2007)

 Capacity constraint

 Long term capacity strategy

 Medium-term capacity planning and control

 Short-term capacity planning and control

 Aggregated planning and control

 If the OM system produces a variety of products, then capacity often is

measured in terms of a standard unit of measure.
Objectives of capacity planning and control
Capacity management decisions can affect several different aspects of
performance:

 costs

 revenues

 working capital

 quality

 speed

 dependability

 flexibility
Capacity Planning

Supply Demand
Availability of capacity
to deliver products and
services
The operation The market

Required availability
of products and
Operations resources services Customer requirements
 Supply Does Not Naturally Match Demand
 Inventory results from a mismatch between supply and demand

 Mismatch can take one of the following two forms

 Supply waits for Demand
 Inventory = Finished goods and resources

 Demand waits for Supply

 Inventory = Backorder in manufacturing

 Inventory = Waiting customers in services

 Mismatch happens because

 the demand varies
 the capacity is rigid and finite.
 If the capacity is infinite, products (or services) can be provided at
an infinite rate and instantaneously as the demand happens.
Then there is no mismatch.
 Importance of Capacity in Services

 What happens when capacity planning goes wrong?

 Airline industry
 Retailing
 Service recovery

 How can we manage and adjust service capacity utilization?

 University space
 Airlines/Trains
 Restaurants
 Prosumerism / Servunction

 Queuing Systems – Virtual or Real

 Capacity Issues for Services

 Capacity planning is more difficult for pure service operations

 Services must (generally) wait for customers to arrive
 Services cannot be stored (inventoried)
 Demand is more variable over short periods

 Services focus on matching availability of staff to demand

 Managing customers (e.g., appointments)
 Different methods for scheduling staff
 Pooling resources (e.g., supermarkets)
 Labour flexibility
 Capacity in Manufacturing
 Basic Problem - Costly to produce products that don’t sell, costly not to
produce products that will sell

 Cisco Systems writes off $2.2 billion of excess inventory (April 2001)

 British Airways had seat utilization of 70.3% in the early 2000s. If it could
increase utilization by 0.33% (by flying one more person on a 300 seat
aircraft), it would create additional revenues equal to quarter 2 profits of
2001, which was $65 M

 In 2000, Playstation 2 of Sony were backordered by several weeks due to

high demand. But X-Box of Microsoft did not sell well and was discounted
by $100 per unit.
 Discounting is a symptom of a problem in operations rather than being a
usual practice.
 Capacity Management
 The capacity management process is defined by the business
environment.

 Five shared criteria define it:

 Demand: How many must we deliver?

 Demand Timing: When must we deliver?

 Location: Where should we place our operations?

 Capacity Timing: When must we make capacity adjustment

decisions?

 What are the value trade-offs among the above?

 Measuring capacity I
 Input capacity measure

 Output capacity measure

Operation Input capacity measure Output capacity measure

Air-conditioner plant Machine hours available Number of units per week

Hospital Beds available Number of patients treated per week

Theatre Number of seats Number of customers entertained per week

University Number of students Students graduated per year

Retail store Sales floor area Number of items sold per pay

Airline Number of seats available Number of passengers per week

Electricity company Generator size Megawatts of electricity generated

Brewery Volume of tanks Litres per week

Measuring capacity II

 Design capacity = the capacity of a process or facility as it is

designed to be, often greater than effective capacity

 Effective capacity = the useful capacity of a process or operation

after maintenance, changeover, and other stoppages and loading
has been accounted for.

 Utilization = actual output

design capacity

 Efficiency = actual output

effective capacity

 Overall Equipment Effectiveness (OEE) = a method of judging the

effectiveness of how operations equipment is used.
How Capacity and Demand are Measured

Efficiency = Actual output

Planned Effective capacity
loss of 59
hours

Design Avoidable
capacity loss - 58
Effective hours per
168 hours capacity week
per week
109 hours Actual
per week output -
51 hours
per week

Utilization = Actual output

Design capacity
 The Capacity Planning Hierarchy

 Long-term
 Decisions about maximum rate of inputs / outputs
 Issues of capacity expansion

 Medium-term
 Utilisation of existing capacity
 Size of workforce
 Seasonal variations

 Short-term
 Control of resources and flows
 Balancing resource utilisation

 Scheduling people (shifts) and the use of

facilities/equipment
 Timing of LT Capacity Expansion

 Lead - adding capacity in advance of demand growth

 Lag - adding capacity after demand growth

 Average - trying to maintain average capacity

Walmart
Lead Lag Average
Capacity Demand Demand

Demand
Capacity Capacity
 Sizing Capacity

 “Size” is not the same as “Scale”

 Economies of scale
 Learning by doing

 Diseconomies of scale
 Diseconomies of distribution
 Diseconomies of bureaucratisation
 Diseconomies of vulnerability
 Steps in Capacity Planning

Step 1 - Measure aggregate

capacity and demand.

Step 2 - Identify the alternative

capacity plans.
Aggregated output

Forecast demand
Step 3 - Choose the most
appropriate capacity
plan.

Estimate of current capacity

Time
 1. Forecasting Demand

 Forecasting provides an estimate of future demand

 The goal is to minimize forecast error.

 What factors may influence demand?





 Improved forecasts benefit all trading partners in the supply chain

 Better forecasts result in lower inventories, reduced stock-outs,

smoother production plans, reduced costs, and improved customer
service.
 Good forecasts are essential for effective capacity planning …
… but so is an understanding of demand uncertainty, because it allows
you to judge the risks to service level
Only 5% chance of demand
Distribution of demand being higher than this
DEMAND

DEMAND
Only 5% chance of demand
being lower than this
TIME TIME

When demand uncertainty is high, the risks to service level of under-

provision of capacity are high
 Forecasting Techniques
 Qualitative forecasting is based on opinion and intuition.
 judgment forecasting
 herd behaviour

 black swan events

 prediction markets

 Quantitative forecasting uses mathematical models and historical data to make

forecasts
 Time series models are the most frequently used forecasting models.
 Moving average forecasting

 Exponential smoothed forecasting

 Causal modelling

 Forecast accuracy (Et) = actual value (At) – forecast value (Ft) [at time period t ]
 Mean Absolute Error (MAE)
 Mean Absolute Percentage Error (MAPE)

 Root Mean squared error (RMSE)

Causes of seasonality

Climatic Festive Behavioural Political Financial Social

Construction materials Travel services

Beverages (beer, cola) Holidays
Foods (ice-cream, Christmas cake) Tax processing
Clothing (swimwear, shoes) Doctors (influenza epidemic)
Gardening items (seeds, fertilizer) Sports services
Fireworks Education services
 DELL Computers
 Historically, Dell’s supplier-management processes relied on manual
mechanisms, limiting their ability:
 to scale its global business
 to maintain an optimal balance between supply and demand
 to react quickly to marketplace changes

 Dell introduced factory-scheduling & demand-planning capabilities:

 Dell (servers, storage systems and supply chain expertise)
 Accenture (program planning and management, business process and
systems design, technical architecture, and integration services)
 i2 (supply chain planning and management applications).

 Key points:
 Forecasting demand management
 Flexibility through modular production
 Advanced supply chain management techniques for Build-to-Order supply
chains
 Zara’s Counter-Intuitive Model
 Zara business model:
 Short lead times = More fashionable clothes
 Lower quantities = Scare Supply
 More styles = More choice, more chance of hitting it right!

 Operations Issues:
 React rather than predict
 Flexibility, not forecasting
 Modularity of design
 Close management of the supply chain
Zara’s Fast Response

 Nonlinear relationship exists

between capacity utilisation,
demand variability and
responsiveness.

 Queuing Theory - As capacity

utilization begins to increase from
low levels, waiting times increase
gradually. But at some point, as
the system uses more of the
available capacity, waiting times
accelerate rapidly.

 As demand becomes ever more

variables, this acceleration starts
at lower and lower levels of
capacity utilization.
 2. Reconciling Capacity and Demand

• How to manage fluctuations in demand and supply?

Demand Demand Demand

Capacity Capacity Capacity

Level capacity Chase demand Demand

management
Ways of reconciling capacity and demand
How do you cope with
fluctuations in demand?

Absorb Adjust output Change

demand to match demand
demand
Level capacity Demand
management
Chase demand
 Level Capacity Strategies
 An approach to medium-term capacity management that attempts to
keep output from an operation or its capacity, constant, irrespective of
demand.
 Produce and store outputs in advance of demand
 Relies on building inventory.

Absorb
Absorbdemand
demand

Keep
Keepoutput
outputlevel
level

Have
Have Make
Make
Make
Maketo to customer
excess
excess customer
stock
stock wait
capacity
capacity wait
Part finished, Queues
Finished Goods, or Backlogs
Customer Inventory
 Chase Demand Strategies

An approach to medium-term capacity management

that attempts to adjust output and/or capacity to reflect
fluctuations in demand. Source: Corbis/Photocuisine

Adjust output to
match demand
Hire Fire

Temporary labour Lay-off

Overtime Short time

Subcontract Third-party work

Demand Management Strategies
• An approach to medium-term capacity management
that attempts to change or influence demand to fit
available capacity.
• Try to change demand to smooth high and low periods

Change
Change
demand
demand

Change
Changepattern
patternof
of Develop
Developalternative
alternative
demand
demand products
products//services
services
Pricing i.e. establish new demand
Promotions
 Strategies for Matching Demand and
Capacity in Service Operations
C o m p le m e n ta r y
s e r v ic e s

P a r tit io n in g
d e m a n d

M a n a g in g R e s e r v a t io n s
d e m a n d

P r ic e in c e n t iv e s

O ff-p e a k
M a n a g in g d e m a n d
Y ie ld
s e r v ic e m a n a g e m e n t
c a p a c ity S h a rin g
c a p a c it y

C u s to m e r
p a r tic ip a tio n
M a n a g in g C r o s s -t r a in in g
s u p p ly e m p lo y e e s

P a r t- tim e
e m p lo y e e s

A d ju s ta b le
c a p a c it y

Based on Fitzsimmons and Fitzsimmons (1997), p. 387, Figure 13.1

 Yield Management
 A collection of methods that can be used to ensure that an operation
(usually with a fixed capacity) maximizes its potential to generate profit.

“allocating the right type of capacity to the right type of customer at the
right price and time to maximise revenue or yield”

 Effective when:
 Capacity is relatively fixed
 The market can be fairly clearly segmented
 The service cannot be stored in any way
 The services are sold in advance
 The marginal cost of making a sale is relatively low
Key Issues in Yield Management
 Pricing structures must appear logical to the customer and
justify the different prices

 Must be able to handle variability between customers through

accurate forecasting of demand

 Service processes must be managed

 Workers and managers must be trained to deal with customers

Price
Fixed Variable
Quadrant 1 Quadrant 2
Predictable
Duration

Movies Hotels

Quadrant 3 Quadrant 2
Unpredictable
Restaurants Long-term hospitals
Capacity planning as a queuing problem
Queuing Theory
 In service operations, the output cannot be stored. Thus, waiting
time created by fluctuations in demand and variability in service
times

 Queuing theory helps explore the trade-off between the amount of

capacity and the level of demand

 Too much capacity is costly, too little capacity results in long waiting
times and irate customers

 Examples
 Bank customers, airplanes circling airport, patients seeing a
doctor and parts waiting to be processed at a machining centre
Simple queuing system

Low variability -
narrow
distribution of
process times

Time
High variability -
wide distribution
of process times

Time
 Simple queuing system

Server 1

Distribution of
Distribution of
processing times
arrival times

Reneging
Rejecting Balking
Server 2

Source of
customers

Queue or Served
“waiting line” customers
Server m
Boundary
of system
The psychology of queues
 Unoccupied time feels longer than occupied time
 Pre-process waits feel longer than in-process waits
 Anxiety makes waits seem longer
 Uncertain waits are longer than known, finite waits
 Unexplained waits are longer than explained waits
 Unfair waits are longer than equitable waits
 The more valuable the service, the longer the customer
will wait

 Solo waits feel longer than group waits

Scheduling
Introduction to Scheduling
 Scheduling establishes the timing of the use of specific
resources of the organisation (ie. start and finish times)

 A schedule shows the use of equipment, facilities, and

human activities over a period of time.
 Manufacturers schedule production
 Hospitals schedule admissions, surgery and nursing
assignments
 Universities schedule classrooms, instructions, and students

 Scheduling is the final step in the transformation process

before actual output occurs. It is constrained within the
context of capacity decisions, equipment selection,
training of workers, design of products and services.
Hierarchy Of Company Plans In Operations

Company Strategy

Sales Forecast

Production Master Schedule

Daily Production Planning

Loading

Time
 The "Dilemmas" Of Scheduling Criteria
"Minimize" Versus "Maximize"

 Minimize Production Costs

 Minimize Storage Costs
 Minimize Stock Investment
 Minimize Cash Outflow

 Maximize Labour Utilization

 Maximize Plant Utilization
 Maximize Customer Satisfaction
 Maximize Operator Morale
 Factors In Scheduling Operations
 Due Dates For Finished Items
 Similarities Between Jobs
 Capacities Of Departments
 Efficiencies Of Departments
 Maintenance Plan
 Holidays
 Forecast Absenteeism
 Existing Commitments
 Availability Of Materials
 Priorities Set On Jobs
 Allowances For Scrap
 Scheduling Rules "GOOD IDEAS ...."
 Schedule First Those Jobs With The Shortest First
Operation

 Schedule First Those Jobs With The Shortest Last

Operation

 Schedule Jobs According To Their Total Work Content

 Schedule Jobs According To Their Date Of Receipt

 Prioritise Jobs By Customer

 Scheduling Rules - More "GOOD IDEAS ...."
•First come, first served (FCFS)- the first job or customer to arrive at a
workstation will be the next one processed.
•This is a fair rule when people or jobs arrive randomly and have similar
requirements.
•All customers are treated equally in the sense that priorities are assigned in order
of arrival, but no allowance is made for the fact that some jobs or customers are
more important or need to be finished sooner than others.

•Earliest due date (EDD) - the job or customer with the earliest due date will be
processed next.
•This minimises the total lateness of all jobs or customers being processed.
•This rule highlights the importance of due dates, and therefore may be more in line
with customer needs.
•On the other hand, where service is unreliable or consistently late customers often
learn to "play the system" and submit jobs with artificially early due dates.
 Scheduling Rules - More "GOOD IDEAS ...."

 Shortest processing time (SPT)- the job or customer that will take
the least time to process is the next processed.
 This minimises the total waiting time, but long jobs or more urgent
jobs may not be processed quickly.
 This rule maximises the throughput measured as number of jobs
processed, and is therefore commonly used by operations whose
goal is to maximise cash flow, since these flows come earlier in the
process and therefore are discounted less.

 Longest processing time (LPT)- this is the opposite of shortest

processing time.
 This rule may be applied when the operation is not concerned with
early cash flows, but can be associated with interim (stage)
payments for partly completed work.
Scheduling Rules – Even More "GOOD IDEAS ...."
•Last arrived, first processed (LAFP) - the opposite of first come, first
served.
- This is seldom an efficient rule, because it means that jobs currently in the
system will have to wait even longer to be processed.
- However, overworked administrators often apply this rule since the job
that has arrived latest is usually associated with a living, breathing
customer!

•Least slack time - the job with the least time between the time it will take
to process the job and the due date (slack or float time) will be processed
next.

•Critical ratio (CR) - is a more sophisticated version of least slack time,

since it computes the ratio of time remaining to the work remaining, so that
jobs or customers with varying processing times can be compared more
easily.
Scheduling Rules – Even More "GOOD IDEAS"
 Start the job with the shortest first processing time – This is managed
by breaking up the total work content into the operations that are required.
 The operations manager then chooses the job with the shortest first
operation.
 The rationale for this is to ‘get up and running’ with jobs and this ploy may
be used where firms have invested recently in new technology.
 The problem with this approach is that, like others we have discussed, this
rule pays no attention to customer requirements.
 The other problem is that although firms may wish to be ‘busy’ and to utilise
technology this may simply encourage work in process and not result in
finished goods.
 This in turn may mean that the firm cannot invoice and so cash may be
drained.

 Start the job with the longest last operation - This is quite difficult to
execute but some companies do this.
 The reason for this is that the last thing that a firm wants is to progress jobs
and then have them held up at the last process.
 What this rule tries to do is to avoid bottlenecks occurring at the last stage
(which is the most expensive stage of the overall job because all other costs
have been accrued by this stage).
 Conclusions
 Capacity planning as a strategic issue

 Objectives of capacity planning

 Match demand and supply

 Three steps in managing capacity

 Forecast, Plan, Choose capacity strategy

 Strategies for matching supply & demand

 Level capacity, chase demand, demand management

 Yield management in service organisations

 Queuing problem and scheduling