DEMAND AND

REVENUE
MANAGEMENT

UNIVERSITAS
MERCUBUANA
LEARNING OBJECTIVES
Explain the concept of revenue management
and integrate it with operation strategy

Identify the conditions for and obstacles to

effective revenue management

Design product to achieve customer

segmentation and price discrimination

Analyze time segmentation with dynamic

pricing and capacity constrains

Apply two tools for capacity segmentation and

reservation: overbooking and dynamic capacity
allocation
Demand and Revenue Management: Concepts

• The concept of demand management includes all the

activities that influence the demand of product and services
(marketing activities and sales and services activities)
• In this section how operations can contribute to demand
management
Much operations deals with matching supply and demand,
often by tailoring the production process to the demand
characteristics. But when the existing operational system is
inflexible,operations strategy must match demand to
available supply in a value-maximizing manner
Revenue management refers to the collection of demand
management practices that aims to maximize the revenue
of available supply.
The Growing Importance of Revenue Management

• Revenue monagement takes the age-old business

practices of price discrimination (tailoring the price
to an individual’s willingness topay so as to extract
more revenue)
• Movie theaters,buses, trains, amusement parks,and
restaurants have long practiced revenue management
techniques by offering discount to children,student,or
and senior citizens
• American airlines estimated the quantifiable benefits of
revenue management at $1,4billion over the previous
three years. The expected an annual revenue contribution
of over $500 million to continue into the future.
The Growing Importance of Revenue Management

• 1990revenue management spread to the car

rental and hotel industry (case National Car
Rental , Mariott Hotel)
• By 2000 revenue management practices
had spread to many other industries,
including banking, entertainment,internet
groceries, and real estate.
 1.Evaluate your customer

Revenue
management 2.Design products to segment
follows a four customers and time
steps process
and is
continously 3.Price products
updated

4.Control capacity reservation

 • The opportunities arising from
revenue management techniques
depend on your strategy as well as the
actions of your competitors
Evaluate your • Understanding customer purchasing
strategy,competitio behaviour and valuation is key: what
n,and customer do customers need and value?besides
price, how importantis flexibility in the
purchasing decision?Do valuations
change significantly over customers or
over time?

• If so, there is an opportunity to serve

them better and improve profits by
Design Product designing products toilored to each
group and time period
 • Price the products based on
customer purchasing
Price behaviour, valuations,and
products forecasting demand curves

• Control the capacity that is

Control allocated to the previous
capacity products
Typical Revenue Management Techniques

Revenue
Management

Price Capacity
Control controls

Price Service
Dynamic Yield
customizatio customizatio
pricing management
n n
•Segmentation •Markdown •Overbooking •Priority service
•Products & fences •Promotions •Discount •Mass
•Price differentiation •Peak-load allocation customization
Typical Revenue Management Techniques
Price Controls

• Revenue= pricex quantity

• Price controls are techniques that vary prices in
order to influence demandpricing is arguably the
most intuitive and widely used form of revenue
management (it is useful to make distinction
between strategic and tactical pricing)
• Strategic pricing refers to the setting of long-term,
average prices that take into account competitive
behaviour as well as customer valuation and cost
Typical Revenue Management Techniques
Price Controls

Price discrimination and customization

• It aims to extract a substantial part (if not all) of

each customer’s willingness-to-pay
• It is common to discrimination among groups of
customer’s by offering different products
Typical Revenue Management Techniques
Price Controls
Dynamic pricing : the tactic of varying price
over time in response in demand or supply
conditions
• Markdown pricing: decreasee price over time in order to
clear leftover inventory. In retail, it reflects decreasing
customer valuation near the endof the season
• Promotional pricing includes discounting, normally within
a limited period. The power of promotions derives from
the fact that they can be targeted to particular customer
segments in ways that are not possible throughlist prices
• Peak-load pricing increases prices during periods of high
demand while supply remains constant. It is widely
practiced in industries with fixed supply but changing
demand over the day, week, or season
Typical Revenue Management Techniques
Price Controls

• In addition to these tactics of controlling

posted pricest that customers can see, there
are also price discovery mechanisms, in
which customers determine price through
their own action during the transactions
Typical Revenue Management Techniques
Capacity Controls

Techniques designed to influence demand by restricting the

availability of supply (while keeping prices fixed)
• Uses reservation systems and booking controls to accept and
Yield allocate demand to available capacity
management • This strategy originated in airlines industry

• The tactic of adjusting a service(or even a good) to an

individual customer.
• While service cutomization shares some characteristics with
Service yield management (such as the contol of availlable capacity),
customizatio it doses not requieres the use of reservation systems
n • Services customization make to order or service to order
enviroments
• FedEx’s overnight versus regular delivery
OBSTACLES TO REVENUE MANAGEMENT

Data mining, forecasting,and

optimization

Competition and automated

revenue management

Treat your best customers well

Social concerns and fairness (non-

markets factors)
OBSTACLES TO REVENUE MANAGEMENT
Data mining, forecasting,and optimization

• Revenue management techniques stand or fall with

the availability of representative and detailed
forecasts
• Substantial valid data sets are required to forecast
demand curves by market segment, product, and
time of purchase
OBSTACLES TO REVENUE MANAGEMENT
Competition and automated revenue management

• Prices depend not only on demand characeristics,

inventories,and capacities,but also on competitor’s
actions. Competitive revenue management can
quickly become complex and is not yet full
understood in a network setting
• For example,consider in the airlines industry, where
two airlines may compete for customers on a single
flight –leg (horizontal competition), but may also
serve on connecting flights (vertical integration).
Airlines may form alliances that allow each airlines
to sell tickets for flights operate by any alliance
partner
OBSTACLES TO REVENUE MANAGEMENT
Treat your best customers well

• It is crucial that the best (most valuable) customers

do not resent your revenue management techniques.
• Good revenue management increases firm profits
while improving service along some dimension that
is important to customer that pay the high price
the important message is not heeded and revenue
management is perceived as unfair, strong customer
reactions follow.
OBSTACLES TO REVENUE MANAGEMENT
Social concerns and fairness (non-markets factors)

• Few customers enjoy discrimination of other revenue

techniques such as overbooking and dynamic pricing.
• Krugman (2000) summarizes the main criticism: “dynamic
pricing is undeniably unfair: some people pay more just
because of who they are”. In addition to unfairness, the
practice of purposely degrading a product just to separate
the market is not easy defensible from a social perspective
• Example: In February 2005, the Ueropean Union
mandated specific, minimal compensations from airlines
to customers who are denied boarding due to overbooking
or flight cancelation. Similar regulations are to follow
for international train and boat traffic.
Customer Segmentation and Product Design
Example: Club Bora-bora

• Club Bora-bora is an exclusive vacation resort in French Polynesia.

• Collecting information on how much customers value a day at a club
( for example, we sampled 5% of the people who expressed interest in
staying at Club Bora on a given day)
Custome
r C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 SUM
$ $
WTP 323,00 151 534 378 358 284 50 113 225 4562.872,00

$600.00

$500.00
f(x) = − 50.4848484848485 x + 564.866666666667
Willingness to pay

$400.00

$300.00

$200.00

$100.00
Total Surplus
$-
C3 $ 2872
C10 C4 C5 C1 C6 C9 C2 C8 C7
Customer (labe)
Customer Segmentation and Product Design
Example: Club Bora-bora

• All good revenue management starts with a forecast. If we

cannot have perfect customer information, we can at least
model and predict their reservation prices using the best
fitting function to our data.
• For our limited data set, a simple linear regression using a
spreadsheet shows that the best linear inverse demand
function is : p =565-50qjadi :q = (565-p)/50
assuming 5% customer sample is representative of the entire
customer population, our total demand function is
5/100 * D(p) = (565-p)/50 D(p) = 20 * ((565-p)/50)
= 226-0,4 p
Customer Segmentation and Product Design
Example: Club Bora-bora (Charging a single price)
• Charging a single price: if we treat customer equally, we must charge each the same price.
The optimal price maximizes total revenue and is thus derived from a simple optimization
problem.
For a given price p, the total sales quantitiy is thus D(p) with associated revenue p
(Dp)= 226p-0,4 p2
Simple price optimization yields the:
a = q max and b = qmax/pmax  pmax = qmax/b
optimal single price = a/2b = ½ p max = 226/(2*0,4)
= $282,5
associated sales quantity = 226-(0,4 x 285,5)= 113
associated revenue = 113 x $ 282,5 = $31,923
Customer Segmentation and Product Design
Example: Club Bora-bora (Segmenting customer into two
groups)
• Segmenting customer into two groups: low valuation customers are
more price-sensitive and maybe willing to purchase in advance in lower
price, while high value customers may buy at the last moment at full
price.
• In this case, it is optimal to sort and group customers and charge them
different prices by creatng two “products” advance (p2) and regular
purchase (p1) the quantity of regular segment is D(p1) and the
quantity demanded by advance segment is given by D(p2)-D(p1)
Total revenue = regular purchase revenue +advance purchase revenue
= p1 D (p1) + p2(D(p2)-D(p1))
= p1 (226-0,4p1)+ p2(p1-p2)*0,4
Customer Segmentation and Product Design
Example: Club Bora-bora (Segmenting customer into two
groups)
Total revenue = regular purchase revenue +advance purchase revenue
= p1 D (p1) + p2(D(p2)-D(p1))
= p1 (226-0,4p1) + p2(p1-p2)*0,4
a = q max and b = qmax/pmax  pmax = qmax/b
P1 = 2/3 p max ; p2 = 1/3 p max; q1= q2 = 1/3 q max
Price optimization yields:
Optimal price: p1= 2/3 (226/0,4)= 2/3 (565)= $376,67 ; p2 =1/3 (565)=$188,33
Sales quantity: 226- 0,4 p1= 75,33 and (p2-p1)*0,4 =75,33
(total 150,67 vs113, increase 33%)
Associated reveneue: (75,33 x$376,67) + (75,33 x 188,33)= $42.563 (increase revenue by
$10.640 atau 33% relative to uniform pricing)
It has two effects: raising the price for customers with a high willingness to pay
and allows the company to give a discount to the other customers
Customer Segmentation and Product Design
Example: Club Bora-bora (Customer segmentation with capacity
constraints)
• Customer segmentation with capacity constraints: such an increase in
sales, however, cannot be accomodated by the limited room capacity of
120 at Club Bora Bora.
• To mitigate capacity shortages, revenue management typically raise
price.
• Again,optimal prices maximize revenue but now a capacity constraint it
added to the price optimization below:
Maximize Revenue= regular revenue +advance revenue
Subject to Sales = regular sales +advance sales
≤capacity
Customer Segmentation and Product Design
Example: Club Bora-bora (Customer segmentation with capacity
constraints)
Maximize Revenue= regular revenue +advance revenue
Subject to Sales = regular sales +advance sales ≤capacity

Revenue= p1 D(p1) + p2 (D(p2)-D(p1))

Sales = D(p1) +(D(p2)-D(p1))
Sales = (226-0,4p1)+ ((226-0,4p2)-(226-0,4p1))
Sales = 226-0,4p1+(226-0,4p2-226+0,4p1)
Sales = 226-0,4p1+(226-0,4p2-226+0,4p1)
Sales = 226-0,4p2
capacity constraint 226 – 0,4 p2 ≤ 120

Constraint price optimization yields: (see appendix)

a = q max and b = qmax/pmax  pmax = qmax/b

optimal prices: p1 = p max – (K/2b) = (qmax/b)-(K/2b)= (226/0,4)-(120/(2*0,4))=415

p2 = (a-K)/b= (226-120)/0,4= $265
sales quatity: 226-(0,4*415)=60 and (p1-p2)*0,4= 60
associated rev: (60 x $415) + (60 x $265) = $ 40.800 (bandingkan dengan
perhitugan sebelumnya: $42.563 )
Customer Segmentation and Product Design
Example: Club Bora-bora (The value of adding more price optios)

• The example showed that

Optimal
revenue segmenting the customer base
% of total into two groups and offering
surplus two price points increased
revenues by one-third
Marginal gain
of adding last • The incremental value of
price point additional price points quickly
decrease, however, because
there is less and less customer
surplus to be captured
• Revenue gains quickly taper
1 2 3 4 5 6
off as more price points are
added
Customer Segmentation and Product Design

• Product design: • Demand • When should we

fences and modelling and consider
leakage forecasting customer
segmetation and
different pricing?
Customer Segmentation and Product Design
Product design: fences and leakage

• Selling each product as a different price, customers can choose

or self select, thereby partially revealing their valulation
• The key idea behind designing these differentiated products is
that they attract different customer groups that are correlated
by willingness to pay, while their marginal cost is roughly the
Product same
design:
fences • As in the publishng industry where pricier hardcover books
and attract reader who value novelty while some soft cover books
leakage are sold at discount (their marginal production cost differs by a
few dollars at best, much less than their price differences)
• Often, however, product design for revenue management
purposes leaves the physical asset unchanged, instead adding
qualifying restrictions or “fence” to its purchase or usage
Customer Segmentation and Product Design
Product design: fences and leakage
Time of purchase or usage restrictios
• Advance purchases (discount price)vs spot purchase; martinee tickets can only be used for
morning or afternood shows
Location to purchase
• Car rental companies change their prices depending on the country that is selected when
making reservation
Purchase restrictions
• Cancelation or changes fee increase product separation and aim to discourage people from
switching among products
Purchase volume restriction
• Discount products are often only available in individual quantities to prevent resale
• Special discount only apply to groups as a form of quantity discounts
Duration of usage restrictions
• Single day rentals are usually more expensive than the price per day associated with a
weekly rate
Customer affiliation and loyalty program, which can be used to separate and retain customers
Customer Segmentation and Product Design
Demand modelling and forecasting

• Nearly all revenue management techniques rely on

forecasts of demand curves to help design products and
optimize price and capacity controls.
• In addition to describing a relationship between demand
Demand and price, demand curve also include the dependence of
modelling that relationship on various exogenous parameters such
and as income, advertising spending, gender, age, wether,
forecasting time, and competitors prices. When any of this
parameters change, the demand function may “move”
• Forecast should include an expected value (point
forecast), as well as a measure of accuracy (forecost
error or standard deviation)
Customer Segmentation and Product Design
When should we consider segmentation and differential pricing?

• The valuation of an identical asset unit varies

considerably over the customer base
Implementing
• Customer can effectively be segmented into
customer groups with similar willingness to pay in
segmentation advance, location, etc effective segmentation
and price means that customers self-select the product
discrimination
can be valuable designed for them without much leakage (or buy
when down)
following • Customer valuation of each segment (product)
conditions are
met can be forecasted with reasonable accuracy
• Differential pricing is particularly attractive for
high fixed product, low marginal cost operation
Typical Revenue Management Techniques

Revenue
Management

Price Capacity
Control controls

Price Service
Dynamic Yield
customizatio customizatio
pricing management
n n
•Segmentation •Markdown •Overbooking •Priority service
•Products & fences •Promotions •Discount •Mass
•Price differentiation •Peak-load allocation customization
 Time Segmentation and Dynamic Pricing

Dyamic pricing is the tactic of varying

price over time in response to changing
supply and demand conditional
Other reasons for
implementing dynamic
Markdown and
pricing are stochastic
promotional pricing
demand variability and
Seasonal or peak-load apply price changes
lack of perpect
pricing adjus prices for when the supply is
information about
temporal change in changing (clearing
underlying market
demand while supply leftover inventory) or
conditions (example,
remains contanst when it differs from the
can be “learned” in the
forecast, while demand
first weeks and then
may also be chaning
used for pricing in
remaining weeks)
Time Segmentation and Dynamic Pricing
Seasonal discount and peak-load pricing
• Club Bora Bora experieces lower demand during the fall season.
• Assume that the lower valuation and demand quantities for a single night
are reflected the lower demand curve: D low season(p)= 180-0,4p
• If the same prices (p1=415 and p2=265) as before would be used for the
off season, sales and revenue would be:
sales quantity : 180-(0,4 x 415)=14 and 180-(0,4x265)-14=60
associated revenue: (14x $415)+(60x $265)= $21.710
Utilities fall from 74 /120=61,67% revenue fall $21.710 or 53%
Both can improved by adjusting the price to off season demand

Time Segmentation and Dynamic Pricing
Seasonal discount and peak-load pricing
Both can improved by adjusting the price to off season demand
a = q max and b = qmax/pmax  pmax = qmax/b
optimal prices : p1 = p max – (K/2b) = (qmax/b)-(K/2b)
= (180/0,4)-(120/(2*0,4))=$300
p2 = (a-K)/b= (180-120)/0,4= $150
sales quantity: 180- 0,4 p1= 60 and (p2-p1)*0,4 =60
associated revenue (60x 300)+(60x 150)= $27.000

By offering a low-season discount of $115, room utilization increase to 100%, and the
revenue improves significantly (lower pricing power : $27.000 vs $40.800 in peak season)

Seasonal or peak-load pricing is attractive when demand exhibits

strong seasonality
This explains the popularity of seasonal pricing in the entertainment,
transportation, rental, restaurant, recreation, and utility industries under
the forms of matinees, early bird specials, weekend rate, etc
Time Segmentation and Dynamic Pricing
Selling off a fixed supply: markdowns and promotions
• Assume retailer stocked 144 units of a summer safari-look women’ s shirt to sell
over an 8 week season.
• Market research forecasted a demand function: D(p)=60-1,2p
• With sample stock, the optimal retail price and sales quantity would be $25 (price
optimal = a/2b= 60/ (2x1,2)) and 30 unit ( sales = ½ a=1/2x60)
• But, this would be exceed the expected weekly availability of 18 unit (144/8
weeks), the optimal stock constrained price can be found through the equatiion
18=60-1,2p atau p = (60-18)/1,2 = $35
• Halfway into the season, the retailer reviews inventory and finds that sales have
been slower than expecteed
– Expected inventory after 4 weeks = 144- (4x18)= 72 units but the actual is 88 units...
– Assuming the demand curve remains valid for the remaining weeks, it si then optimal
to markdown the shirts and sell 22 units per week (88/4=22 units) to clear the
remaining inventory by the end of season
– The associated optimal markdown price is then
pmark-down = (60-actual remaining weekly inventory)/1,2= (60-22)/1,2 = $31,67
Time Segmentation and Dynamic Pricing
When should we consider time segmentation and dyamic
pricing?

It can be • Demand quanitites are seasonal, while capacity is fixed.

Off-season promotions can move peak demand to fill
valuable demand valley, thus improving average utilization and
to reducing capacity shortage
implemen • Customer valuation for the asset are seasonal
t dynamic • Demand quantities, valuations, or underlying conditions
are uncertain
pricing • Price can easily be changed. Information technology
when one increasingly reduces the cost of changing prices
or more • Asset supply is perishable and demand is volatile.
following Markdown improve the likelihood that assets are sold
condition before perishing
• Assets are sold both in advance and on the spot market
s are met
Capacity Reservation and Overbooking
Capacity pre-sales, yield management , and operation strategy

• In practice, demand is rarely known with certainty; posted price control can still
result in demand-supply mismatches.
• To reduce mismatches, a firm can use a combination of excess capacity (let assets
wait for customers), excess inventory (let products wait for customers), or excess
time (let customer wait on a reservation or backlog list)
• Yield management is the practice of pre-selling capacity and making customers
wait on a reservation list  after capacity is size, but before all demand is known,
we can use yield management to reduce capacity surplus and shortage costs
• Yield management uses capacity controls, which are techniques that influence
demand and maximize revenue by restricting the availability of supply ( while
keeping total capacity and price fixed)
• We will approach the solution in two steps:
– Decide how many reservation to accpet in hotel. If we anticipate order cancelations,
then selling more reservations than actual available capacity can improve actual
capacity utilization and revenue. This called is also called overbooking
– Decide how to segment capacity and how many restrictions to accept for each
capacity segment
Capacity Reservation and Overbooking
Example: overbooking at Club Bora Bora
• For the main session, club Bora Bora has listed an advance purchase price
$265 and a regular price $ 415. It expects to sell about 60 rooms to each
segment (lihat sebelumnya...)
• While the total room capacity is 120, past experience has shown that, on
average, 10 room reservation are either cancelled late or their cutomers do not
show up
•  as solution, the resort can overbook and accept 130 reservations, 10 beyond
capacity, in order to compesate for late cancelations and no-shows  the risk is
that there will be fewer cancellations that predicted; then, a customer with a
confirmend reservation would be denied a room and would have to be lodged
elsewhere
• Club Bora Bora estimates that associated oversale cost, after refunding any
reservation fees and accounting for the compesation for incovenience, the cost
of other accomodation, and lost of goodwill, is $ 640
• ???????? what is the value of this overbooking policy and haow can the
resort do better???
Capacity Reservation and Overbooking
Example: overbooking at Club Bora Bora

Without overbooking (120-10=110 rooms)

• Average price of ($265+$415)/2=$340 per room revenue

110 x $340 = $37.400
With overbooking, 10 additional reserve are sold, increase
revenue by $340 x 10 = $3.400 or 9% to $40.800
(37.400+3.400)
• The value of overbooking = $40.800-$37.400= $3.400 or
9%
In reality, however, the number of cancelations is uncertain and the firm faces two
risk: capacity surplus and shortage risks. When more cancelations occur than
compesate for, capacity underutilized and revenue is lost. In constrast, when fewer
cancellation occur, the firms incurs a net cost for ecxess demand
Capacity Reservation and Overbooking
Example: overbooking at Club Bora Bora
Optimal overbooking
• The optimal overbooking level balances the risks of capacity
surplus and capacity shortages
• If we make one additional reservation, we increase the risk
being overbooked but decrease the risk of having surplus
capacity or spoilage
• Let p be the probability that we will be overbooked
In the Club Bora Bora, the marginal impact on the expected
mismatch cost is the expected shortage cost of $640 plus the
expected surplus cost of $340. The optimal condition sets the
marginalprofit to zero and yields
Opt. shortage probability = reg. margin /(reg. Margin+
shortage penalty)
Capacity Reservation and Overbooking
Example: overbooking at Club Bora Bora
Optimal overbooking

Opt. shortage probability = reg.

margin /(reg. Margin+shortage penalty)
• opt. Shortage probability = $340/($340+
$640)=35% the hotel will be overbooked 35% of
the days
The optimal overbooking level= norminv
(0,35,10, 5)= 8 rooms
• From forecast mean= 10 and standard deviation = 5
Capacity Reservation and Overbooking
When and how to overbook?

Overbooking should be considered when:

We have reliable forecast of
The net shortage penalty is no
cancelations, no shows, and
too large compared to the
shortage penalty costs. It is
regular margin (when there is
important to track past
no capacity constraint)
cancelation and no-show data

The optimal level of overbooking depends on the ratio of

addtional revenue to oversale cost, and the uncertainty in
the cancelations forecast. As the shortage cost rises
relative to the regular margin, however, one should be
more conservative and reduce the overbooking level
Capacity Reservation and Overbooking
Reducing negative impact of overbooking

Cancelation rate
• Cancelation rate can be decreased by limiting a customer’ s cancelation
flexiblitiy. This is the main idea behind a non-refundable deposit when
making a reservation or adding restrictions
Surplus costs can be decreased by last minute discount
and stand by ticket to fill surplus capacity at the last
minute

Shortage penalty can be decreased by:

• Having process to deal with overbooked customer
• Having back ups such as exernal capacity providers or other internal
capacity
• Using more flexible products
Capacity Segmentation and Allocation
Optimal capacity segmentation

• Let p be the probability that the room will be sold at the

regular price. Revenue then is maximized by accepting the
advance purchase as long as $265p ≥$415p
• It is optimal to keep accepting advance reservation as long
as their profit exceed the expected profit of a regular
reservation (Littlewood’s rules)
• Optimal regular shortage probability p= advance margin/regular margin=
265/415= 64%
• Optimal protection level = norminv (36%, 60, 30)= 49 rooms (36% is
optimal probability that a regular request can be accomodated; 60=
mean of regular demand; 30 = standard deviation of regular demand)
• This means the optimal advance booking limit is 128-49=79 rooms
Capacity Segmentation and Allocation
Capacity segmentation for capacity buyer and sellers

Suppliers
• Suppliers can implement capacity segmentation when they have
both long term, contractual customers as well as spot customer
• For exampe: a specialty steel mill may have long term contracts that
guarantee a maximal response time to customer orders, but may also
accept orders on the spot market
• What the optimal portofolio of contracts ? see exampmle 8.6
Buyers
• Buyers deciding on an optimal sourcing strategy face a similar
problem
• When actual demand is uncertain, how much sourcing should be
commited to in advance, and how much sourcing should be decided
at the last minute at higher price (be it from the same supplier,
another supplier, or the spot market)
• Example 8.7
When and How Much Should We Allocate Capacity at
Discount?

Allocating a chunk of capacity to discount

sales and protecting the rest to regular sales
is valuable when
• We have an updated forecast regular future demand
(and of the sell up probability)
• Regular demand is insufficient to consume all available
capacity
• Discount sales increase the overall capacity utilization
and the expected profit
• The value of allocation optimization increases when
regular demand is highly volatile and the discount
differs from 50%
When and How Much Should We Allocate Capacity at Discount?

The efficiency of capacilty alloction depends on the price

and the effectiveness of the the fences used to separate the
discount differs from the regular price
• If the discount segment is not priced properly, it could lead to a large
fractionof regular demand switching to the discount segment (buy
down)
The amount of capacity allocated to the discount segment
depends on the discount (as a fraction of regular price)
and the regular demand forecast volatility
• When the discount is about 50%, it is a good practice to protect the
mean regular demand
• As the discounts become smaller reduce the protection level
• Optimal regular protection level also depends on the volatility of the
regular demand forecast decrease as volatility rises if the discount is
less than 50% and increases otherwise