Harrah’s Case Study Questions

Our academic experts are ready and waiting to assist with any writing project you may have. From simple essay plans, through to full dissertations, you can guarantee we have a service perfectly matched to your needs.

GET A 40% DISCOUNT ON YOU FIRST ORDER

ORDER NOW DISCOUNT CODE >>>> WELCOME40

Harrah’s Case Study Questions
1. Was Harrah’s current method for allocating room capacity adequate? If not, how
should Harrah’s exploit the enhanced capabilities of a new system?
2. The current system was based on the profits Harrah’s derived from customers’ gaming
activities. How should nongaming activity be accounted for?
3. The current system modeled demand from high value customers as being highly
sensitive to room rates, an assumption that held true for gaming customers. However,
it was not yet clear how “experience seeker” demand levels would be affected by
room rates. How should the system account for the price elasticity of demand?
4. At some locations, such as Las Vegas and Atlantic City, Harrah’s owned multiple
properties. How should rooms be allocated across multiple properties within a single
geographic market?
INFORMS
Transactions on Education
Vol. 9, No. 3, May 2009, pp. 160–168
issn1532-0545  09  0903  0160 informs ®
doi 10.1287/ited.1090.0031cs
© 2009 INFORMS
Case
Revenue Management at Harrah’s
Entertainment, Inc.
Narendra Agrawal
Department of Operations and Management Information Systems, Leavey School of Business,
Santa Clara University, Santa Clara, California 95053, nagrawal@scu.edu
Morris A. Cohen, Noah Gans
Operations and Information Management Department, The Wharton School, University of Pennsylvania,
Philadelphia, Pennsylvania 19104 {cohen@wharton.upenn.edu, gans@wharton.upenn.edu}
Introduction
David Norton, Harrah’s Senior Vice President of
Relationship Marketing, considered the cycle of service
interactions the company’s gaming customers
followed:
At a high level you might view a customer interaction
with Harrah’s as taking place in three stages. First,
there is the customer’s decision to visit a property after
being stimulated through direct marketing or advertising.
Next, there’s his or her request for a hotel room
and our allocation of hotel capacity. Finally, there’s
the on-property experience itself, which covers a wide
range of specific gaming and nongaming interactions,
that the customer enjoys over the course of a visit to a
property.
Harrah’s Entertainment had already set world class
standards across these customer processes. In the first
stage, the company had pioneered the use of IT to
track its customers’ gaming behavior and, in turn,
to effectively match its marketing programs to individual
customers’ preferences. Harrah’s used this
capability to entice customers to visit its more than
40 properties. In the last stage, Harrah’s had also made
innovative use of IT to track its customers’ real-time
behavior at its properties and, in turn, to provide them
with exciting and enjoyable gaming experiences.
In the middle stage, in which Harrah’s allocates
hotel rooms to customers, the company had used
revenue management (RM) tools in innovative ways.
Harrah’s used its sophisticated Total Rewards (TR)
customer loyalty system to segment customers based
on their potential gaming value. This value, called
an average daily theoretical (ADT) value, was based
on the games a customer played, the length of time
and the frequency of play, and the average house
advantage on those games. The company then used
this detailed segment information to set prices and
allocate available hotel rooms to customers seeking
reservations.
Recent growth through acquisition had transformed
Harrah’s from a fairly homogenous set of properties,
most catering to a local clientele that patronized its
casino for the excitement of gaming, to a broad collection
of properties that served a wider range of market
segments and geographies: “day trippers” who visited
frequently and often spent small amounts; “high
rollers” who frequented high-end locations, such as
Caesars, visited periodically, and might spend large
amounts on gaming; and cash-paying “experience
seekers” who spent more on nongaming activities and
purchases than they did on gambling. The fact that
customers cut across multiple demographic segments
and ethnicities further complicated the mix.
Given this challenge, David thought that the time
was right to further develop Harrah’s approach to
offering hotel capacity to customers. “We have the
opportunity to allocate rooms in a manner that provides
more value to each property.” As he considered
what capabilities a new system should have, a set of
four questions were foremost in his mind.
—Was Harrah’s current method for allocating room
capacity adequate? If not, how should Harrah’s
exploit the enhanced capabilities of a new system?
—The current system was based on the profits
Harrah’s derived from customers’ gaming activities.
How should nongaming activity be accounted for?
—The current system modeled demand from highvalue
customers as being highly sensitive to room
160
Additional information, including supplemental material and rights and permission policies, is
a
vailable at http://ite.pubs.informs.org.
Agrawal, Cohen, and Gans: Case: Revenue Management at Harrah’s Entertainment, Inc.
INFORMS Transactions on Education 9(3), pp. 160–168, © 2009 INFORMS 161
rates, an assumption that held true for gaming customers.
However, it was not yet clear how “experience
seeker” demand levels would be affected by
room rates. How should the system account for the
price elasticity of demand?
—At some locations, such as Las Vegas and
Atlantic City, Harrah’s owned multiple properties.
How should rooms be allocated across multiple properties
within a single geographic market?
A useful means of addressing many of these issues
would be to “run the numbers” on example properties.
As a start, David thought it would be useful to
look at historical room allocation data for one of its
properties in Atlantic City over the 2007 Labor Day
weekend.
Company Background
In 1937 William Fisk Harrah opened a bingo parlor
in Reno, Nevada. By 1946, Harrah had expanded
to include a lush casino, with blackjack, dice tables,
slot machines, and roulette. In 1955, Harrah bought
a casino in South Lake Tahoe and converted it into
the largest gambling structure in the world. A second
casino followed in 1973.
Phil Satre joined Harrah’s in 1980, and in 1984
he succeeded Bill Harrah as Chairman and CEO.
Satre led Harrah’s expansion to Atlantic City, and as
other states began legalizing gambling operations, the
Harrah’s network expanded widely.
In the late 1990s, competitors such as Mandalay
Resort Group and MGM Mirage invested billions of
dollars in building dazzling casinos that offered a
wide range of amenities, such as spas, premium shopping
malls, and marquee shows. Satre believed that
the novelty of any particular attraction would wane
over time, however, and he feared that Harrah’s could
be lured into a risky war based on competing capitalintensive
attractions.
Instead, Satre directed company resources toward
strengthening Harrah’s relationships with its most
loyal gaming customers. For Harrah’s, these core customers
were not necessarily flashy “high rollers.”
Rather, they were bankers, teachers, doctors, and
other middle-aged adults with discretionary time and
income who loved playing slot machines.
The company began the development of its Winner’s
Information Network (WINet), an IT network
that allowed Harrah’s to better understand and cater
to this core group. WINet enabled Harrah’s to collect
customer data from diverse locations and sources,
to integrate the data to create customer profiles, and
to act on the profile information to provide special
offers that were tailored to the preferences of the regular
gamblers visiting Harrah’s Casinos. For example,
they found that offers of $60 in casino chips
were more effective than an offer of a free room, two
steak meals and $30 in chips (Loveman 2003). In 1997,
Harrah’s launched Total Gold, a customer loyalty program
that made use of the WINet system to allow
regular gamers to earn rewards across all of Harrah’s
properties.
With the launch of Total Gold, the four elements
of Harrah’s corporate strategy were in place: a focus
on delivering great service to Harrah’s loyal base of
gaming customers; a wide, multimarket footprint for
its properties; the infrastructure required to track and
learn about customers’ preferences across its properties;
and the ability to provide individually tailored
offers and incentives to maintain customer loyalty
and increase Harrah’s share of its customers’ gaming
activity.
At the same time, the company’s decentralized
structure—with each property setting its own operating
and marketing policies—made it difficult for
Harrah’s to fully execute on and profit from this
vision. In 1998, Satre brought in Gary Loveman,
a Harvard Business School professor who had been
consulting to the company, as Chief Operating Officer.
Loveman’s mandate was to integrate Harrah’s many
properties’ operations and marketing activities so that
the company could better capitalize on its strategic
assets.
Loveman did just that. Where property managers’
broad discretion and independence once allowed
them to make conflicting offers, even competing for
the same gamers, Loveman instituted corporate coordination
and influence over properties’ marketing
activities and promotions. For example, he implemented
bonus plans that were tied to achieving
improved customer satisfaction scores at a property.
Thus, the bonuses may be denied despite record
breaking financial performance at a location if customer
satisfaction target scores were missed. Property
managers learned to act with the aim of maximizing
Harrah’s total share of a customer’s gaming activity
across all properties. Loveman heightened Harrah’s
overarching brand image, so that customers would
know that, whichever Harrah’s property they patronized,
they would find an exciting gaming experience.
Loveman also continued to aggressively expand
Harrah’s, through the opening of properties in new
markets. Within a six-year period, Harrah’s acquired
Rio (1999), Horseshoe (2004), and Caesars (2005).
Figure 1 shows Harrah’s brand portfolio as of 2007.
By the autumn of 2007, Harrah’s Entertainment
operated a highly diversified network of gambling
facilities within the United States and the United
Kingdom, and it planned major international facilities
in Spain and the Bahamas. It had the broadest
footprint in the industry (Table 1) and was the largest
provider of branded casino entertainment in the
Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.
Agrawal, Cohen, and Gans: Case: Revenue Management at Harrah’s Entertainment, Inc.
162 INFORMS Transactions on Education 9(3), pp. 160–168, © 2009 INFORMS
Figure 1 Harrah’s Portfolio of Gaming Brands
world, with about four million square feet of casino
space, 40,000 hotel rooms, and 85,000 employees.
After the terrorist attacks of September 11, 2001, this
broad distribution network had proven to be a significant
asset. At a time when travel within the United
States was severely curtailed and most gaming companies
took up to a year to bring revenues back to
previous levels, Harrah’s reached those levels within
a week.
Loveman also continued the development of
Harrah’s Total Gold loyalty program. Much like
the frequent flyer programs of airlines, Total Gold
rewarded gamers for the money they wagered. They
would earn credits which could be redeemed for
amenities such as free hotel rooms, meals, and show
tickets.
Loveman realized that, as successful as Total Gold
had been, there remained deficiencies with its original
Table 1 Harrah’s Distribution Network∗
Arizona Casinos
Harrah’s Phoenix Ak-Chin Casino
California Casinos
Harrah’s Rincon Casino
Illinois Casinos
Harrah’s Joliet Casino
Harrah’s Metropolis Casino
Indiana Casinos
Caesars Indiana
Horseshoe Casino Hammond
Iowa Casinos
Harrah’s Council Bluffs Casino
Horseshoe Council Bluffs
Kansas Casinos
Greater Topeka Area Casinos
Louisiana Casinos
Harrah’s Louisiana Downs
Harrah’s New Orleans Casino
Horseshoe Casino Bossier City
Mississippi Casinos
Grand Biloxi
Grand Casino Resort Tunica
Horseshoe Casino Tunica
Sheraton Casino and Hotel (Tunica)
Missouri Casinos
Harrah’s North Kansas City Casino
Harrah’s St. Louis Casino
Nevada Casinos
Bally’s Las Vegas
Bill’s Casino Lake Tahoe
Caesars Palace (Las Vegas)
Flamingo Las Vegas
Harrah’s Lake Tahoe Casino
Harrah’s Las Vegas Casino
Harrah’s Laughlin Casino
Harrah’s Reno Casino
Harveys Lake Tahoe Casino
Imperial Palace
O’Sheas Casino Las Vegas
Paris Las Vegas
Rio All-Suite Hotel and Casino
New Jersey Casinos
Bally’s Atlantic City
Caesars Atlantic City
Harrah’s Resort Atlantic City
Showboat Casino
North Carolina Casinos
Harrah’s Cherokee Casino
Ontario, Canada Casinos
Casino Windsor
Pennsylvania Casinos
Harrah’s Chester
Uruguay Casinos
Conrad Punta del Este
Source. http://www.harrahs.com. ∗List does not include Harrah’s properties in Egypt, the United Kingdom, or South Africa.
implementation. The program was not differentiated
from competitors’ offerings, it lacked uniformity
across Harrah’s properties, and it did not provide
incentives for customers to consolidate their entertainment
spending with Harrah’s.
Nevertheless, Total Gold did prove to be a significant
source of detailed transactional data that was
vital to Harrah’s understanding of its customers and
their spending behavior. Analysis of the data showed
that customers were spending only 36% of their gaming
budget with Harrah’s, and the rest with competitors.
Thus, a customer who spent only a few dollars
at one of Harrah’s properties was not necessarily a
bad customer; it might just have been that the company
did not capture a large enough share of his or
her gaming expenditures.
Similarly, Total Gold allowed Harrah’s to carefully
track the results of marketing activities and learn
Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.
Agrawal, Cohen, and Gans: Case: Revenue Management at Harrah’s Entertainment, Inc.
INFORMS Transactions on Education 9(3), pp. 160–168, © 2009 INFORMS 163
which types of promotions worked for different customers.
For example, offers of free rooms and meals
often proved to be less effective than smaller offers of
free casino chips for core customers.
These types of insights, which were the result of
sophisticated quantitative analysis of customer data,
as well as controlled experiments, became the basis
for Harrah’s targeted marketing programs. They also
paved the way for the next generation of Harrah’s
loyalty program. In 1999 the company expanded the
program to have multiple tiers, with benefits that
became richer with increased activity, and renamed it
Total Rewards (TR).
The information collected through the TR program
was used to determine targeted and customized marketing
promotions. For instance, for a given property,
a hotel-voucher might be awarded to an out-of-state
guest, while a free show ticket might be best for someone
who had made a day trip. Similarly, specific promotions
could be designed to increase frequency of
visiting, profitability per visit, or cross-market play.1
Thus, Harrah’s used the TR program to provide
regular gamers with a strong incentive to consolidate
their play with the company. Its presence in multiple
markets, along with its attractive rewards program,
had driven strong growth in its best customers’
cross-market play and had allowed Harrah’s to capture
a larger share of their gaming revenues. Nearly
one-third of Harrah’s revenues were due to crossmarket
play, and its share of customers’ gaming budgets
reached 45%.
As of October 2007, Harrah’s was the largest and
the most profitable casino entertainment company
(Table 2), with a market capitalization of $16.45 billion.
Its stock price had risen steadily from $39.60 in
December 2002 to $88.75 in December 2007 (Center for
Research in Security Prices 2009).
Mechanics of the Total
Rewards Program
In the TR program, loyalty cards were the main vehicle
customers used to record their rewards, as well as
for Harrah’s to collect detailed gaming data. A customer
playing slot machines who wished to accrue
rewards would place his or her TR card into a slot
in the machine at the start of a gaming session. The
machine would then record the value of each wager
made. Similarly, a customer playing a table game,
such as roulette or poker, would allow the dealer to
swipe his or her card and to track his or her wagers
and enter them into a Harrah’s IT system.
1 A gamer’s cross-market play is the extent to which he or she gambles
at Harrah’s Entertainment properties in multiple geographic
markets.
Table 2 Selected Financial Data
All amounts in millions of dollars, except per share amounts
Dec. 2006 Dec. 2005 Dec. 2004
Annual income statement
Revenue 96739 71110 45483
Cost of goods sold 48568 36941 24067
Gross profit 48171 34169 21416
Gross profit margin (%) 498 481 471
SG&A expense 25527 19034 9881
Depreciation and amortization 7114 5338 3616
Operating income 15530 9797 7920
Operating margin (%) 161 138 174
Nonoperating income 513 47 95
Nonoperating expenses 6705 4812 2718
Income before taxes 8348 5032 5288
Income taxes 2956 2278 1906
Net income after taxes 5392 2754 3381
Continuing operations 5239 2635 3295
Discontinued operations 119 271 382
Total operations 5358 2364 3677
Total net income 5358 2364 3677
Net profit margin (%) 55 33 81
Diluted EPS from total net income ($) 285 157 326
Dividends per share 152 139 126
Annual balance sheet
Assets
Current assets
Cash 7996 7244 4890
Net receivables 6017 7245 2066
Inventories 630 595 256
Other current assets 1665 1207 660
Total current assets 16308 16291 7872
Net fixed assets 144083 129561 52476
Other noncurrent assets 62458 59324 25509
Total assets 222849 205176 85856
Liabilities and shareholders’ equity
Current liabilities
Accounts payable 17898 15914 7522
Short-term debt 4512 70 18
Other current liabilities — — —
Total current liabilities 22410 15984 7540
Long-term debt 116387 110388 51511
Other noncurrent liabilities 23341 22153 6453
Total liabilities 162138 148525 65504
Shareholders’ equity
Preferred stock equity — — —
Common stock equity 60711 56651 20352
Total equity 60711 56651 20352
Shares outstanding (mil.) 1861 1838 1127
Source. Harrah’s Entertainment, Inc. 2006 Form 10-K.
Each wager recorded in the TR system earned the
customer credits. For those playing reel slot machines
or video poker, the benefit was explicitly defined:
every $5 “coin-in” on a reel slot machine or $10 coinin
on video poker would earn one credit. For those
playing table games, the formula for granting credits
was based on Harrah’s house advantage and was not
explicitly communicated to customers. Nevertheless,
as customers played, the credits would accrue in a
Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.
Agrawal, Cohen, and Gans: Case: Revenue Management at Harrah’s Entertainment, Inc.
164 INFORMS Transactions on Education 9(3), pp. 160–168, © 2009 INFORMS
Figure 2 Summaryof Total Rewards Benefits
BENEFITS
Gold PLATINUM DIAMOND
Ability to earn comps, cash, and offers based on play
10% discount at participating gift shops
Ability to earn bonus comps with Total Rewards Visa
partner program (available for U.S. residents only).
Special birthday gift
Monthly Reward Credit Multiplier Days
(Platinum-2x or more and Diamond-2x or more)
500 Bonus Reward Credits each month
Tickets to shows in Las Vegas, Lake Tahoe, Reno, Laughlin,
and New Orleans
(Platinum-2 for 1 and Diamond-2 free)
Monthly exclusive values on select Rewards Catalog Merchandise
Exclusive gift during yearly tier status renewal period
(March/April)
Free tournament entry and hotel stay for the Summer Fest
and Winter Fest slot tournaments
(Platinum and Diamond member events in Las Vegas)
(Platinum-2 Nights and Diamond-3 Nights)
Members-only access to Diamond Lounges (where available)
Guaranteed priority service at participating restaurants, clubs,
hotel front desk, Total Rewards center, cashier cage, and
slots services
Invitations to exclusive events and tournaments at casinos
participating in the Total Rewards program
Note. http://www.harrahs.com
predictable fashion. The TR credits customers earned
counted in two ways.
First, the number of credits earned over the course
of a calendar year would determine a customer’s
rewards tier for the following year: customers with
0–3,999 credits would become Gold; those with
4,000–9,999 credits would be Platinum; customers
with 10,000–99,999 credits would become Diamond;
and those with 100,000 or more credits would become
Seven-Star customers.
Each TR tier level was associated with special
tier-specific privileges. For example, members with
Diamond and Platinum status would enjoy access to
premium lounges, preferred reservations at restaurants,
and shorter lines during hotel check-ins. These
benefits were clearly communicated to all customers,
and they were delivered in a way that created a visible
differentiation in the level of customer service across
tiers. These distinctions provided an aspirational element
to the program (Figure 2).
Because customers’ tier status was earned over
the calendar year, New Year’s Eve was typically
an extremely busy night. Gamers, who were often
acutely aware of the number of remaining credits they
needed to reach the next tier-level, would play to
cross the hurdle, so that they could enjoy the associated
“perks” all through the following calendar year.
Second, in addition to these tier credits, customers
also received rewards and bonus credits based on
their level of play. These credits had a monetary value
attached to them and could be used to purchase goods
and services, such as free hotel rooms, meals, tickets
to shows, etc. Most customers redeemed their rewards
credits for food and beverages, and a smaller percentage
used them for hotel rooms, merchandise, and
other categories. Like money, these credits could be
Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.
Agrawal, Cohen, and Gans: Case: Revenue Management at Harrah’s Entertainment, Inc.
INFORMS Transactions on Education 9(3), pp. 160–168, © 2009 INFORMS 165
banked and did not expire as long as the customer
remained an active gambler.
The TR card could be used at any Harrah’s property,
regardless of where it was originally issued.
Similarly, the credits and comps2 could be earned,
tracked, and redeemed at any location. More importantly,
a TR member’s tier status was valid—and
valued—across all of Harrah’s many properties and
brands. The company considered these features to be
crucial in providing customers with an incentive to
increase their gaming expenditure, as well as to consolidate
their play with Harrah’s, thereby increasing
Harrah’s share of their total gaming activity.
In 2007, Total Rewards was the leading loyalty program
in the gaming industry and was accepted at
more than 40 casinos. It boasted nearly 40 million
members, of which about 8 million had been active
in the previous twelve months. Across the company,
nearly 80% of gaming revenue could be tracked to a
TR card.
Hotel Revenue Management
at Harrah’s
Harrah’s used an extensive set of RM tools to allocate
hotel rooms at its properties around the country.
These tools were based on a “clearing price” system
based on customer’s projected gaming activities, and
they differed from the allocation methods typically
used in hospitality and airline RM systems.
Traditionally, hospitality companies viewed hotel
rooms as their most important asset and sought
to maximize hotel revenues. In contrast, Harrah’s
viewed a hotel room as a facilitating good and aimed
to offer hotel rooms as a means of supporting gaming
activity and profits. Rather than allocating rooms
to customer segments based on room rates, Harrah’s
used the expected profit generated by customers’ projected
gaming activities as the basis of its rates.3
The system, developed in 1999, tracked the availability
of rooms at a given property, and it constructed
detailed demand forecasts for that property
using the average daily theoretical (ADT) gaming
profit data distilled from its customers’ TR activity.
2 “Comps” are complimentary benefits—such as free drinks, meal
vouchers, tickets to shows, and hotel rooms—that casinos offer
gamers.
3 It was traditional in the gaming industry not to consider the revenue
from room charges when allocating rooms to high-value customers.
Free rooms were often used as a means to lure high-value
customers onto a property to use the casino’s gaming facilities.
The giving up of a room to a lower-valued customer would be
traded off against the chance that a more valuable customer would
then seek the room at a later date. High rollers, in particular, were
known to make last-minute decisions, and it was not uncommon
for a property to keep a large number of rooms open until the day
before arrival, even as it denied access to lower-valued customers.
A customer’s ADT value represented Harrah’s estimate
of the profit it would clear on one day of his
or her gaming activity. Harrah’s segmented customers
into more than 20 ADT profit tiers, and the RM system
produced demand forecasts by ADT tier, arrival
date, and length of stay.
Harrah’s system then converted its projections into
forecasts by ADT tier and occupancy date and compared
the tiers’ demands to the number of available
rooms for each night. Demand was sorted by
ADT value, highest to lowest, and allocated to that
night’s available rooms. The “clearing price” for a
night was the maximum ADT value at which the
forecast demand would occupy all free rooms, and
the room rate quoted to a gamer for each night of a
requested stay reflected the difference between his or
her ADT tier value and the clearing price.
To better understand how Harrah’s RM system
functioned, we sketch some of its details. We begin
with the supply of rooms and then move to the
demand side (see Table 3).
On a given night, each of a property’s rooms (Phys)
may have been in one of several states. For example,
it may have been out-of-order due to ongoing
renovations (Ooo), “manager held” (Mgr Held) for
special guests or a special purpose, or reserved for a
group block (Grp Blk). Rooms that were unavailable
for these reasons were excluded from the total number
of rooms available to book (Avail T) on that night.
Available rooms may have been booked or free, and
to account for cancellations and no-shows, the system
allowed Harrah’s to designate a percentage of booked
rooms to be considered free (Free) as well. The system
could also indicate additional rooms as being free
due to overbooking, as determined by the overbooking
(Ovbk) level.
Table 3 shows the forecast of room supply for the
Labor Day Weekend 2007, as of August 1, at an
Atlantic City property. Examination of the “Booked”
and “Cancel” rows shows that the cancellation rate
used was 10%.
On the demand side, the RM system began with a
segmentation scheme that was based on customers’
Table 3 Room Availabilityfor Labor DayWeekend as of
August 1, 2007
8/31 9/1 9/2 9/3 9/4
Phys 1570 1570 1570 1570 1570
Ooo — — 2 2 4
Mgr Held —————
Grp Blk 222 269 260 227 163
Avail T 1348 1301 1308 1341 1403
Ovbk 00110
Booked 187 266 251 990 1011
Cancel 19 27 25 99 101
Free 1180 1062 1082 450 493
Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.
Agrawal, Cohen, and Gans: Case: Revenue Management at Harrah’s Entertainment, Inc.
166 INFORMS Transactions on Education 9(3), pp. 160–168, © 2009 INFORMS
Table 4 ADT Tier Values
Tier ADT
0 $881
1 $285
2 $69
3 Non-TR
gaming activity. Specifically, the system placed each
customer into one of 20 different segments that were
based on the ADT profit categories into which customers
fell.
Table 4 provides a small example with only four
tiers. Here ADT values range from $881 per day for
Tier 0 customers, down to $69 per day for Tier 2
gamers. Tier 3 captures customers who are not a part
of the TR program (Non-TR), for whom gaming profits
cannot be estimated.4 Note that the ADT tiers listed
in Table 4 are somewhat distinct from TR tiers, which
measured gaming activity on annual basis. For example,
while a given customer might have earned the
10,000 TR credits required for the Diamond tier by
coming often and by playing $50 a day, that customer
would only land in ADT Tier 2.
For a given night the RM system forecasted unconstrained
demand for each tier, by arrival date and by
length of stay (LOS).5 Given 20 ADT tiers, one year’s
worth of future arrival dates, and lengths of stay that
varied from 1 to 13 nights, at any given time the system
maintained nearly 95,000 separate forecasts for a
single property.
Although length of stay was explicitly used as a
forecast criterion, the system did not use LOS information
when allocating rooms. Instead, forecasts for
LOS of n > 1 nights were split up into n distinct
single-night demands. Tables 5 and 6 provide a small
example of how the system translated LOS-based
forecasts into total numbers of rooms demanded each
night.
Each cell in Table 5 represents a forecast, as of
August 1, 2007, for the number of rooms demanded
by a single ADT tier with LOS equal to 1, 2, or 3 over
a five-night period that covers the Labor Day weekend.
For instance, the extreme right cell in the row
corresponding to 8/31 of Table 5 shows that Harrah’s
4 If a non-TR customer were to attempt to reserve a room via the
Harrah’s website, they would have been informed of availability
and quoted a price. Since priority was given to TR customers, it
was often the case that, during busy periods, the non-TR customer
would have been directed to nearby hotels at various price points.
If a price were quoted it would have been at least as high as the
current clearing price for the night(s) in question. For simplicity’s
sake, our description ignores the impact of non-TR customers.
5 A tier’s unconstrained demand was a measure of potential
demand. It assumed that Harrah’s accepted every booking request
coming from a customer in that tier.
Table 5 August 1, 2007 Forecast for 3 ADT Tiers byArrival Date and
LOS
Tier 0 Tier 1 Tier 2
Arrival
date LOS 1 LOS 2 LOS 3 LOS 1 LOS 2 LOS 3 LOS 1 LOS 2 LOS 3
8/31 432 149 107 280 213 157 369 148 49
9/1 486 147 84 251 168 118 333 47 16
9/2 655 189 98 321 210 146 310 170 57
9/3 20 7 6 60 48 36 372 215 72
9/4 20 8 7 44 35 26 352 164 55
expected 49 customers from ADT Tier 2 would request
rooms for August 31, wishing to stay for three nights.
Each cell in the Tier 2 section of Table 5 then corresponds
to a full row in Table 6. For example, the
Tier 2 forecast of 49 for 8/31 in Table 5 maps to third
row in Table 6, where the demand of the 49 customers
described above is recorded for each occupancy night.
Note that by totaling the demand numbers in a
column of Table 6, the RM system translated ADT
Tier 2’s LOS-based forecasts into total demands for
each night.
Given each ADT tier’s forecast by occupancy date,
the system then compiled an overall forecast by ADT
tier and occupancy. For example, the Total row in
Table 6 would become the ADT Tier 2 row in Table 7.
The RM system then used the data on room availability
(Table 3), the ADT tier data (Table 4), and
the forecast data by occupancy date (Table 7) to
determine a clearing price for each occupancy date.
Table 8 shows how the clearing price calculation was
performed.
The first two columns of Table 8 come directly from
the ADT tiers and values listed in Table 4, and the
first three rows of the remaining columns are cumulative
versions of the rows in Table 7. For example,
Table 6 August 1, 2007 Forecast for ADT Tier 2 byOccupancyDate
ADT Tier 2 demand for each occupancy date Arrival
date LOS 8/31 9/1 9/2 9/3 9/4
8/31 1 369
8/31 2 148 148
8/31 3 49 49 49
9/01 1 333
9/01 2 47 47
9/01 3 16 16 16
9/02 1 310
9/02 2 170 170
9/02 3 57 57 57
9/03 1 372
9/03 2 215 215
9/03 3 72 72
9/04 1 352
9/04 2 164
9/04 3 55
Total 566 593 649 902 915
Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.
Agrawal, Cohen, and Gans: Case: Revenue Management at Harrah’s Entertainment, Inc.
INFORMS Transactions on Education 9(3), pp. 160–168, © 2009 INFORMS 167
Table 7 August 1, 2007 Forecast for All ADT Tiers byOccupancyDate
Demand by tier per occupancy date
ADT tier 8/31 9/1 9/2 9/3 9/4
0 688 973 1280 404 146
1 650 907 1120 618 335
2 566 593 649 902 915
Total 1904 2473 3049 1924 1396
the demand numbers for ADT Tier 1 of Table 8 are
obtained by adding the demands for Tiers 0 and 1
from Table 7. In the row below that of Tier 2, marked
Free Rooms, Table 8 lists the numbers of rooms free
that night, as calculated in Table 3.
The clearing price on a given night was determined
by finding the ADT at which the free rooms would
sell out. For example, on September 1, the 1,062 free
rooms could have been filled using only Tiers 0 and 1
customers, so the ADT value of Tier 2, or $285, would
become the clearing price for that night.
In most cases, the room rate quoted to gamers making
hotel reservations became the maximum of zero
and the difference between the clearing price and their
ADT tier value. For example, if a Tier 2 customer
were to make a reservation for September 1, then
he or she would be quoted a room rate of max[$0,
Clearing Price−ADT Value] = max$0 $285−$69
= $216 for that night. If a Tier 0 customer were
to make a reservation for that same night he or
she would be offered a complimentary room, since
max$0 $285−$881 = $0. For customers making
reservations for multiple-night stays the rate quoted
would be the sum of the quotations for the individual
nights. For instance, a Tier 2 customer wanting
a room for the nights of September 3 and 4 would
be quoted a rate of max$0$285−$69+max$0
$69−$69=$216+$0=$216.
The RM system included a provision to override
the standard clearing price calculation, shown above,
and in some cases Harrah’s used this additional logic.
In particular, the system imposed minimum prices
and maximum room rates that were quoted. For
example, a non-TR customer making a reservation for
the night of September 4 might have been quoted a
Table 8 August 1, 2007 Calculation of the Clearing Price byOccupancy
Date
Cumulative demand by tier per occupancy date ADT Tier
tier value ($) 8/31 9/1 9/2 9/3 9/4
0 881 688 973 1 280 404 146
1 285 1338 1880 2400 1022 481
2 69 1904 2473 3049 1924 1396
Free rooms 1180 1062 1082 450 493
Clearing price ($) 285 285 881 285 69
room rate of $150, even though the clearing price for
that night was $69. Conversely, if the clearing price
became too high then, rather than quoting a room rate
that would be beyond a gamer’s expectations, the system
would simply respond that the property was sold
out that night.
For any given night, the number of free rooms and
the clearing price became the information used in
Harrah’s hotel reservation system (called the Lodging
Management System). When a customer requested a
number of rooms over a certain number of nights,
these data were used to check availability and to
quote a price.
Thus, the RM system allows Harrah’s to price
rooms by ADT tier segment using a convenient clearing
price mechanism. Rather than constructing the
price using forecasts of demand at various room rates,
the system uses demand forecasts for customers with
various gaming values.
Finally, it is important to note that the above
description of the system presents only a snapshot of
the rates quoted for a single property as of a single
point in time—August 1. As inquiries would come
in for a property, Harrah’s reservation system would
update room availability in real time. These demand
data would then be fed back to the system, which
would compare actual requests to the forecast, adjust
the forecast accordingly, and update its clearing price
calculations. The updating process happened once
daily, and Harrah’s was planning to increase the frequency
to several times per day.6
Continued Development of Harrah’s
Hotel Revenue Management
Capabilities
The existing RM system had served Harrah’s well.
However, as sophisticated as it was, it lacked a number
of important features that David believed to be
essential to Harrah’s going forward. Thus, in 2006,
Harrah’s began working to develop a next-generation
RM system.
At the most basic level, the move would allow for
the development of a more comprehensive RM system.
The new system would have expanded capabilities
to accommodate many more control segments,
differentiate among room types, and provide multiple
forecast updates per day. David was also concerned
that, because clearing prices did not correctly use LOS
information, they did not accurately reflect rooms’
(opportunity) value to Harrah’s on any given night.
The new system would use sophisticated optimization
software that would allow LOS-based forecasts
to directly drive the clearing price calculation.
6 In theory, forecasts and prices should be updated after every room
assignment or a cancellation.
Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.
Agrawal, Cohen, and Gans: Case: Revenue Management at Harrah’s Entertainment, Inc.
168 INFORMS Transactions on Education 9(3), pp. 160–168, © 2009 INFORMS
With the acquisition of rival gaming companies,
such as Bally’s, Caesars, Flamingo, and Horseshoe,
Harrah’s had also become the owner of multiple
properties in both Las Vegas and Atlantic City, and
David viewed Harrah’s ability to own and manage
multiple properties within a given location as having
a potentially strategic advantage. That is, he believed
that, for Harrah’s to reap the benefits of owning multiple
properties, the RM process must be managed at
the location, rather than property, level. The development
project had provided an ideal opportunity to
extend the RM system’s scope.
With the acquisition of Caesars, in particular,
nongaming activities had also become a signifi-
cant source of revenue for Harrah’s. David was
keenly interested in understanding how these revenues
should be accounted for when making hotel
RM decisions. Should Harrah’s redefine the ADT tier
values that were used to determine room rates? How
should Harrah’s integrate nongaming and TR information
to define segment boundaries and set clearing
prices?
Other members of Harrah’s RM team also saw
the potential to expand the scope of Harrah’s RM
activities to valuable resources beyond hotel room
allocation. How could Harrah’s use an RM approach
to allocate capacity for restaurant seating, show tickets,
and other venues?
When asked about the company’s long-term vision
for its RM system, a member of Harrah’s RM team
became thoughtful for a moment. He responded that,
while all of the near and longer-term developments
described above were critical, he thought there was
one important aspect of all RM systems that needed
further development. “What I’d really like to know—
and I pose this as a question for researchers in
revenue management—is how to integrate information
about price elasticity into these systems. Clearly,
changes in price affect the level of demand we experience.
However, none of the systems we are familiar
with capture this effect.”
Acknowledgments
We gratefully acknowledge the assistance of Harrah’s executives,
David Norton, Andy Chance, and Steven Pinchuk,
as well as the valuable comments of the area editor and
anonymous referees.
Reference
Loveman, G. 2003. Diamonds in the data mine. Harvard Bus. Rev.
81(5) 109–113.
Additional information, including supplemental material and rights and permission policies, is available at http://ite.pubs.informs.org.

Our academic experts are ready and waiting to assist with any writing project you may have. From simple essay plans, through to full dissertations, you can guarantee we have a service perfectly matched to your needs.

GET A 40% DISCOUNT ON YOU FIRST ORDER

ORDER NOW DISCOUNT CODE >>>> WELCOME40

 

 

Posted in Uncategorized