Content uploaded by Jeannette Ho
Author content
All content in this area was uploaded by Jeannette Ho on Apr 19, 2020
Content may be subject to copyright.
Cornell University School of Hotel Administration
e Scholarly Commons
&.3&1'/1/20*3",*37&2&"1$)4#,*$"3*/.2 ;&&.3&1'/1/20*3",*37&2&"1$)
Implementing Revenue Management in Your
Restaurants: A Case Study with Fairmont Ra"es
Hotels International
Sheryl E. Kimes
Cornell University School of Hotel Administration2&+$/1.&,,&%4
Jeanne#e Ho
Fairmont Raes Hotels International
/,,/63)*2".%"%%*3*/.",6/1+2"3 )>022$)/,"12)*02)"$/1.&,,&%4$)104#2
"13/'3)& /20*3",*37%-*.*231"3*/.".%"."(&-&.3/--/.2
;*213*$,&*2#1/4()33/7/4'/1'1&&".%/0&."$$&22#73)&;&&.3&1'/1/20*3",*37&2&"1$)"3;&$)/,"1,7/--/.23)"2#&&.
"$$&03&%'/1*.$,42*/.*.&.3&1'/1/20*3",*37&2&"1$)4#,*$"3*/.2#7"."43)/1*8&%"%-*.*231"3/1/';&$)/,"1,7/--/.2/1-/1&
*.'/1-"3*/.0,&"2&$/.3"$3 )/3&,,*#1"17$/1.&,,&%4
&$/--&.%&%*3"3*/.
*-&2/-0,&-&.3*.(1&5&.4&-"."(&-&.3*.7/411&23"41".32$"2&234%76*3)"*1-/.3":&2/3&,2
.3&1."3*/.", Cornell Hospitality Report, 19
Implementing Revenue Management in Your Restaurants: A Case Study
with Fairmont Ra"es Hotels International
Abstract
."*1-/.3":&2/3&,2.3&1."3*/.",*-0,&-&.3&%"2723&-6*%&1&5&.4&-"."(&-&.301/(1"-*.
*321&23"41".323"13*.(6*3)".".",72*2/'#"2&,*.&%"3""*1-/.3"00,*&%"<5&23&01&5&.4&-"."(&-&.3
01/$&223/)*(),*()30/3&.3*",1&5&.4&&.)".$&-&.3/00/134.*3*&2&23"41".3-"."(&12".%&-0,/7&&26&1&
*.5*3&%3/24((&233"$3*$2".%231"3&(*&2%1"6.'1/-3)1&&$"3&(/1*&2",,0410/2&231"3&(*&2231"3&(*&2
3/42&6)&."1&23"41".3*2#427)/3".%231"3&(*&23/42&6)&."1&23"41".3*2./3#427$/,%
001/01*"3&,7$)/2&.231"3&(*&26&1&2*-0,&3/*-0,&-&.3*.-/23$"2&2".%(4&232/=&.6&1&0,&"2&%6*3)
3)&/0&1"3*/.",".%-&.4$)".(&2!*3)*."7&"1/'*-0,&-&.3"3*/."*1-/.31&23"41".323)"3*-0,&-&.3&%
1&5&.4&-"."(&-&.3)"%(&.&1"3&%<5&3*-&2-/1&1&5&.4&(1/63)3)".3)/2&./3"00,7*.(3)&01/(1"-
Keywords
1&23"41".321&5&.4&-"."(&-&.3
Disciplines
/20*3",*37%-*.*231"3*/.".%"."(&-&.3
Comments
9/1.&,, .*5&12*37&01*.3&%6*3)0&1-*22*/.,,1*()321&2&15&%
;*2"13*$,&*2"5"*,"#,&"3;&$)/,"1,7/--/.2 )>022$)/,"12)*02)"$/1.&,,&%4$)104#2
CORNELL CENTER FOR HOSPITALITY RESEARCH
Implementing Revenue
Management in Your
Restaurants:
A Case Study with Fairmont Rafes Hotels International
by Sheryl E. Kimes and Jeannee Ho
EXECUTIVE SUMMARY
In 2015, Fairmont Raes Hotels International implemented a system-wide revenue
management program in its restaurants. Starting with an analysis of baseline data,
Fairmont applied a ve-step revenue management process to highlight potential
revenue-enhancement opportunities. Restaurant managers and employees were
invited to suggest tactics and strategies drawn from three categories: (1) all-purpose strategies,
(2) strategies to use when a restaurant is busy (hot), and (3) strategies to use when a restaurant
is not busy (cold). Appropriately chosen strategies were simple to implement in most cases,
and guests often were pleased with the operational and menu changes. Within a year of
implementation, Fairmont restaurants that implemented revenue management had generated
ve times more revenue growth than those not applying the program.
Keywords: restaurants, revenue management
Disciplines: business, food and beverage management, revenue management
Cornell Hospitality Report • August 2019 • www.chr.cornell.edu • Vol. 19 No. 5 1
ABOUT THE AUTHORS
Sheryl E. Kimes, PhD, is professor emerita of service operations management at the School of
Hotel Administration in the Cornell S.C. Johnson College of Business. She has served as interim
dean of the school, and also as the school’s Richard and Monene P. Bradley Director of Graduate
Studies. Kimes teaches revenue management, restaurant revenue management, and service
operations management. She has been named the school’s graduate teacher of the year three
times and was awarded a Menschel Distinguished Teaching Fellowship by Cornell University in
2014.
Kimes’s research interests revolve around revenue management in the restaurant, hotel, and golf
industries. She has over 100 articles in leading journals, such as Interfaces, Journal of Operations
Management, Journal of Service Research, Decision Sciences, and the Cornell Hospitality Quarterly.
She was awarded the CHR Award for Industry Relevance in 2010, 2012, and 2014 and was given a lifetime achievement award by
the Production and Operations Management Society in 2010.
Kimes has served as a consultant to many hospitality enterprises around the world, including Chevy’s Fresh Mex Restaurants,
Walt Disney World Resorts, Fairmont Hotels and Resorts, Starwood Asia-Pacic, and Troon Golf. She was given the Vanguard
Award for Lifetime Achievement in Revenue Management by the Hotel Sales and Marketing Association International in 2017.
Jeannette Ho was vice president of of revenue and customer analytics, Fairmont Rafes Hotels
International at the time of the study discussed in the accompanying article. She has since
become vice president Rafes global brand strategy and strategic relationships, AccorHotels.
Jeannette joined Rafes in 2005 and has held leading roles in brand marketing, customer
analytics, distribution, revenue management, and guest experience with Rafes and its parent
companies.
As VP marketing for Rafes, Jeannette played an instrumental role in the transformational
openings of Rafes Dubai, Rafes Beijing during the 2008 Olympics and Le Royal Monceau–
Rafes Paris. Previously, she held various senior-level positions with international companies
such as Singapore Airlines, Banyan Tree, and Starwood Hotels & Resorts.
Jeannette holds a rst class honours degree from the London School of Economics, University of London, and was awarded both
the British High Commission’s Award and the Singapore Airlines Scholarship. She is an active speaker at industry conferences,
guest lecturer for Masters in Innovation at Singapore Management University, and has co-authored numerous papers with leading
services marketing academics from Cornell University and National University of Singapore.
The Center for Hospitality Research • Cornell University2
CORNELL CENTER FOR HOSPITALITY RESEARCH
CORNELL HOSPITALITY REPORT
Implementing Revenue
Management in Your Restaurants:
A Case Study with Fairmont Rafes Hotels International
by Sheryl E. Kimes and Jeannee Ho
Originated by the airline industry, revenue management (RM) has been
applied to restaurants for over 20 years.1
Although the principles are similar,
restaurant RM requires a somewhat dierent approach than that applied by
airlines. The restaurant approach involves implementing the following ve-
step process.2
Restaurants rst need to establish their baseline performance (Step 1) and then
seek to understand the causes for that performance (Step 2). With that knowledge, they can
formulate strategies on how best to drive revenue in their restaurants (Step 3). Subsequently,
they face the challenging task of implementation (Step 4). This implementation involves
strategies that fall into three categories: (1) all-purpose strategies, (2) strategies to use when
your restaurant is busy (hot), and (3) strategies to use when your restaurant is not busy
(cold). Finally, they need to measure whether their strategies were successful (Step 5).
1 Kimes, Sheryl E., Richard B. Chase, Sunmee Choi, Philip Lee, and Elizabeth Ngonzi. 1998. “Restaurant Revenue Management: Ap-
plying Yield Management to the Restaurant Industry,” Cornell Hotel and Restaurant Administration Quarterly. 39 (3): 32-39; Kimes, Sheryl E.
2004. “Restaurant Revenue Management,” Center for Hospitality Research Report. Cornell University.
2 Kimes, Sheryl E., Deborah I. Barrash, and John E. Alexander. 1999. “Developing a Restaurant Revenue Management Strategy,”
Cornell Hotel and Restaurant Administration Quarterly. 40 (5): 18 - 30.
Cornell Hospitality Report • August 2019 • www.chr.cornell.edu • Vol. 19 No. 5 3
This report illustrates the application of the ve-
step RM process and its aendant strategies using as
an example the revenue management program imple-
mented by Fairmont Raes Hotels International.
The Fairmont RRM Journey
Fairmont embarked on its restaurant revenue manage-
ment (RRM) program in 2011, starting with several
pilot restaurants in their Singapore hotels and then
expanding the pilots into restaurants in China, the
U.S., and Canada. By 2018, RRM had been deployed to
over 70 percent of their F&B revenue. The results were
striking. Restaurants using RRM generated ve times
more new revenue growth than restaurants not using
RRM in their rst 12 months of application.
Noisy data. The rst issue to overcome involved
acquiring baseline data and then reducing the noise in
those data. During the pilots, we quickly realized that
data were challenging to obtain, and even when ob-
tained were often “dirty.” We saw numerous instances
of zero cover counts, incorrect starting and ending
times, and obviously incorrect check amounts (e.g.,
zero or impossibly high). Indeed, when we rst started
RRM, zero cover counts accounted for 40 to 50 percent
of all transactions.
Upon inspection, we determined that the zero
cover counts came about for two reasons: inaccurate
server data entry and confusion over the denition of
a cover. We addressed those two issues as follows.
First, we needed to resolve the denition of a
cover. Like many hotels and restaurant chains, Fair-
mont dened a cover as the sale of an entrée. This
meant that almost all cover counts in the lounges
were zero, no maer how busy the lounges actually
were. Operators recorded incorrect party-size informa-
tion for parties that shared an entrée or just ordered
starters. The issue of how to dene a cover has since
been rectied by the Uniform Systems of Accounts for
the Lodging Industry, 10th edition, which now clearly
states that a cover is a customer, regardless of whether
that customer orders an entrée.
With the denition of a cover claried, the next
goal was to reduce the incidence of zero cover counts.
To address this issue, Fairmont emphasized the impor-
tance of entering an accurate cover count in its server
training. By 2016, the zero-cover-count percentage had
dropped to 10 to 11 percent (most probably due to
split checks). Only 1 to 2 percent of zero-cover-count
transactions remain unexplained.
We found another source of dirty data at buets.
The manager of one of the buet restaurants com-
plained about how long it took guests to complete
their meals. However, when we looked at the data, we
noticed that check opening and closing data indicated
that a substantial portion of guests took less than ten
minutes to dine (which we considered impossible).
Investigating this apparent contradiction, we discov-
ered that the servers were not opening the check when
guests were seated, but were instead waiting until
guests requested the check.
These data issues are certainly not conned to just
Fairmont. Over the past fteen years, we’ve noticed
the same problems in restaurants around the world.
The source of the noise is employee-entered data,
which is more likely to be error-prone than computer-
or machine-generated data regardless of industry.
Fairmont, like other companies facing the same dirty
data issues, began to stress the importance of entering
data correctly and provided training to ensure that
correct entry occurred.
Geing Started
Before embarking on their RRM journey, Fairmont’s
restaurants needed to determine their baseline perfor-
mance. To do this, we developed an RRM dashboard
for each restaurant and used it as a basis for identify-
ing the appropriate RRM strategies to deploy.
We measured the following ve key metrics: table
occupancy, seat occupancy, average check per person,
meal duration, and RevPASH (revenue per available
seat-hour, which we dene below). All metrics were
calculated by day of week and time of day.
Table occupancy gives a clear indication as to
how busy a restaurant is. Logically, if the table oc-
cupancy nears 100 percent, there will almost certainly
be customers waiting for tables. Table occupancy is
calculated by dividing the number of table-hours used
(# of covers multiplied by the average meal duration)
by the number of table-hours available (# of tables
multiplied by the number of hours in question).
Seat occupancy gives an indication of how com-
pletely the restaurant’s tables are being used. The idea
here is that occupied tables should not have numerous
empty seats. Seat occupancy is calculated by dividing
the number of seat-hours used (number of customers
served multiplied by the average meal duration) by
the number of seat-hours available (number of seats
multiplied by the number of hours in question). Note
that a restaurant can have a high table occupancy, yet
have a fairly low seat occupancy, for example, when
numerous singletons or deuces are occupying four-
tops. This is an indication that the restaurant has a
poor table mix.
The Center for Hospitality Research • Cornell University4
Average check per person is a commonly used
metric that is simple to calculate, assuming that the
data are available and accurate. It is simply the total
check amount divided by the associated party size.
As discussed below, if the check amount is incorrect
or if the party size is missing, it may be impossible to
develop an accurate estimate of the average check per
person.
Meal duration is typically calculated from POS
data and is calculated by subtracting the opening time
of the check from the closing time of the check. As we
noted above, meal duration calculations will be inac-
curate if checks are not opened and closed in a timely
fashion.
RevPASH (revenue per available seat-hour) is a
measure akin to RevPAR (revenue per available room)
commonly used in the hotel industry. This measure
indicates how well a restaurant is using its inventory
of seats. RevPASH can be calculated in two ways. The
simplest way to calculate it is to divide the revenue
earned by the number of seat-hours available (number
of seats multiplied by the number of hours in ques-
tion). The other approach is to multiply the average
check per person by the seat occupancy and divide by
the meal duration.
One of the key challenges that we faced was
helping restaurant operators understand the
dierence between being operationally busy and
revenue-management busy. We found that many of
the operators stated that they were quite busy, and
that was true for some restaurant sections. But we also
noticed that some of the other sections in a particular
restaurant were closed, and sometimes reservations
were being turned away even when there was
available capacity.
The Strategies
Fairmont applied one or more of the three categories
of strategies to improve its restaurant operations.
Again, the strategies are (1) all purpose, (2) hot strate-
gies, and (3) cold strategies. For each strategy, we
will describe the available tools and then provide an
example of successful implementation.
All-Purpose Strategies
Three all-purpose strategies can help restaurants gen-
erate incremental revenue regardless of how busy they
are. The strategies are (1) menu engineering, (2) menu
design, and (3) server mentoring and upselling. We
will briey describe each strategy and then provide an
example of successful implementation at Fairmont.
(1) Menu engineering. Menu engineering has
been examined in considerable detail since it was
introduced three decades ago.3 While there are a
number of variations, the essential approach involves
determining the contribution margin (selling price less
food cost) and the sales volume of each menu item
by menu category (e.g., starter, entrée, or dessert).
For each menu category, the classic approach assigns
items to one of four quadrants. Menu items are classi-
ed as Stars (above average contribution margin and
sales volume), Cash Cows (below average contribu-
tion margin and above average sales volume), Puzzles
(above average contribution margin and below aver-
age sales volume), and Dogs (below average contribu-
tion margin and sales volume).
Managers then use these classications to deter-
mine possible actions to take with each menu item.
For example, with Star menu items, recommendations
might involve highlighting them on the menu, featur-
ing them as a signature dish, or perhaps raising the
price. On the other hand, possible courses of action for
Dogs might be to bundle them with other menu items,
drop them from the menu, or even raise the price of
the menu item (to gain more contribution margin from
the relatively scant sales).
The menu engineering process need not be oner-
ous. In its simplest form, it just involves meeting every
month or two (or whenever the menu is about to be
changed), reviewing the classications, and using
those classications to guide some of the menu chang-
es. The resulting changes in sales volume and contri-
bution will help determine whether the changes were
eective or whether further revisions are required (or
should be undone).
Menu engineering at Fairmont. Fairmont de-
veloped an Excel-based tool for their restaurants to
analyze their menu items (see Exhibit 1, overleaf). The
tool also gives guidance on which actions to take for a
particular menu item.
For example, the rm’s potential tactics for Cash
Cow menu items included reducing the portion size or
3 The seminal work on menu engineering was done by
Michael Kasavana and Donald Smith (see: Kasavana, Michael L.,
and Donald A. Smith. 1990. Menu Engineering: A Practical Guide to
Menu Analysis. Revised Edition. Okemos, Mich.: Hospitality Publica-
tions, Inc.). For other research on menu engineering, see: Atkinson,
Helen, and Peter Jones. 1993. “Menu Engineering: Managing the
Foodservice Micro-Marketing Mix,”Journal of Restaurant & Food-
service Marketing. (1): 37–55 or Leo Yuk Lun. 2005. “The Applica-
tion of Menu Engineering and Design in Asian Restaurants,”Inter-
national Journal of Hospitality Management,24(1): 91-106.
Cornell Hospitality Report • August 2019 • www.chr.cornell.edu • Vol. 19 No. 5 5
E
XHIBIT
1
Menu engineering tool
Dinner Entrées Matrix
E
XHIBIT
2
Lunch menu at Jaan
宏季季官季季宑季季宆季季宋季季季室宷季季季宍季季宄季季宄季季宑季
Artisanal Cuisine
3 course menu 68
Including coffee
季
宄季季宓季季宓季季守季季宗季季完季季宖季季守季季宕季季宖季
RICE CURED ATLANTIC MACKEREL
Caviar, rattes, horseradish
QUAIL & FOIE GRAS BALLOTINE
Pickled onion, pumpkin, samosa
(Supplement $15)
JAAN’S GARDEN
Season’s best composition, wild herbs
季
宐季季宄季季完季季宑季季季季季季宆季季宒季季官季季宕季季宖季季守季季宖季
CONFIT LINE CAUGHT RED SNAPPER
Carrots “Saveurs d’Orient” crayfish, saffron
SALT MARSH WELSH LAMB
« Provence » Asparagus, « Nicoise » olives, barley
(Supplement $15)
bundling them with menu items with a higher contri-
bution margin (CM). For menu items classied as Stars,
the rm tested highlighting them on the menu or
raising their prices. Dogs were divided into Strategic
Dogs and True Dogs. Strategic Dogs provide balance
to the menu (perhaps by oering vegetarian options)
or support the restaurant concept, while True Dogs
might eventually be dropped from the menu if sales or
margins could not be improved. To adjust items clas-
sied as Puzzles, Fairmont proposed such strategies
as changing the menu item description, dropping the
price, or highlighting the item on the menu.
(2) Menu design. Menu engineering can be used
to determine which menu items to highlight (or hide),
but other menu design tools can be used to help
restaurants generate more revenue from their menus.
Designing a menu for revenue generation involves
four key issues: (1) how to name the menu item, (2)
how to describe it, (3) where to place it, and (4) how
and where to display the price. Numerous studies
have examined the eects of how an item is presented
on a menu.4
Menu design at Fairmont. Jaan, a well-regard-
ed modern French restaurant at the Swissôtel in
Singapore,5 introduced a supplement of $15 to $25 for
high demand items on their lunch menu. In addition,
they highlighted those items on the menu to aract
aention (Exhibit 2). The results were positive, as
average food check per person increased by $16 and
revenue exceeded budget by 36 percent.
(3) Server mentoring and upselling. Another all-
purpose tool is to improve servers’ selling skills. For
example, consider a restaurant with 10 servers that has
an average check per person of $16. Say that the top-
performing server has an average check per person of
about $20, while the boom-performing server has an
4 For a good review of menu design fundamentals, see: Yang,
Sybil S., Sheryl E. Kimes, and Mauro M. Sessarego. 2009. “Menu
price presentation inuences on consumer purchase behavior in
restaurants,”International Journal of Hospitality Management. 28(1):
157-160.
5 Winner of the S. Pellegrino Young Chef for Southeast Asia
in 2015.
The Center for Hospitality Research • Cornell University6
E
XHIBIT
3
Average check per server
average check of around $13 (Exhibit 3). If the boom-
performing server can increase his or her average
check per person up to the overall average, the restau-
rant can generate an additional $3 per check. (Plus, if
the restaurant permits tipping, the server receives an
additional 60 cents in tips per check.)
Server mentoring at Fairmont. To assist the poor-
performing servers, Fairmont drew on an innova-
tive approach called Single Server Mentoring (SSM),
developed by Avero (www.averoinc.com). SSM has
been adopted in over 10,000 restaurants, and restau-
rants using SSM have generated over US$40 million in
incremental revenue.6 Restaurants using this method
extract data from the POS system to analyze menu
item sales by server. They can then pinpoint areas in
which a server is either below or above average and
give managers specic advice on how to mentor indi-
vidual servers on how to improve.
Restaurants can, of course, train servers to upsell
without a formal SSM program. The formal program
provides specic nancial information regarding
upselling. Otherwise, managers must rely on telltale
signs that upselling might help increase revenue, for
example, when guests are just ordering the lowest cost
or simplest menu items with no add-ons or starters.
6 www.averoinc.com/products/view/single-server-mentoring
After Fairmont adopted SSM in their restau-
rants, the rm realized an annualized US$3.5 million
incremental uplift in 2015, as the program brought up
average checks of 419 lower performing servers. The
training brings an equivalent of US$8,273 of revenue
uplift per selected server per year. In addition, cus-
tomer satisfaction increased.
Sales of foie gras, a high-margin signature dish,
provide an example. A manager using the SSM ap-
proach noticed that the stronger servers sold foie gras
to six of every ten guests, but that the lowest perform-
ing server sold foie gras only to two out of ten guests.
When the manager pointed this out to the server, she
explained that since she hated liver, she did not want
to suggest that guests order foie gras. Once the server
realized that most guests liked the foie gras, she shad-
owed some of the more successful servers and learned
how to pair it with wine. As a result, she became one
of their top performing servers.
A key challenge Fairmont faced was how to
motivate the servers to participate in SSM. In coun-
tries where tips are common (e.g., the U.S.), it was not
dicult to provide motivation since a higher aver-
age check results in higher tips. The question, how-
ever, was how to implement an upselling program
in restaurants where tips are not customary. At rst,
the rm oered rewards such as hotel vouchers and
monetary awards, but then they realized that the serv-
Cornell Hospitality Report • August 2019 • www.chr.cornell.edu • Vol. 19 No. 5 7
Sample dashboard
E
XHIBIT
4
ers enjoyed the competition and liked the recognition.
Now the restaurants show sales results on a weekly
basis so servers can see how they are performing ver-
sus their peers.
Is It Hot or Is It Cold?
In addition to the all-purpose tools that we just out-
lined—tools that can be applied no maer how busy
a restaurant is—certain revenue management tools
work beer if a restaurant is extremely busy or not
busy at all. Thus, the choice of RM tools we discuss
next depends upon how busy your restaurant is. As
a starting point, we have found it eective to classify
dierent time periods as either hot (busy) or cold (not
busy). This simple approach works well and, in our
experience, makes it easier for restaurants to deter-
mine which tools to deploy at which times.7
Telltale signs of hot periods are full tables, queues,
and declined reservations. Conversely, cold periods
are easy to spot—too many empty tables. Typically, a
restaurant will have some hot periods and some cold
periods. The trick is to identify when they occur.
7 This characterization was borrowed from erstwhile coau-
thor John Alexander, former CEO of CBORD, a provider of menu
systems, food production management, and other computerized
support for congregate dining facilities. See: Kimes et al. 1999.
To identify hot and cold periods at
Fairmont, we calculated table occu-
pancy by day of week and month. Hot
periods were typically dened as hours
in which the table occupancy was over
80 percent, while cold periods were
usually dened as hours in which the
table occupancy was under 50 percent.
Each restaurant involved with the
RRM program was provided with a
simple dashboard that allowed them to
quickly see the percentage of time the
restaurants had hot or cold periods (Ex-
hibit 4). Once these had been identied,
the restaurant manager, the director
of food and beverage, and the director
of revenue management could begin
to determine which tools to deploy at
which times.
Tools to deploy during hot times
are adopting a beer table mix, beer
managing reservations, restricting pro-
motions, and implementing premium
pricing. Suggestive selling can also
be used, but only if it does not extend
meal duration, since it would it most probably be bet-
ter to seat another party rather than sell espresso and
dessert.
During cold (or not-hot) periods, the operator
should focus on making the best of the situation by
maximizing its use of distribution channels (online
and mobile reservations or ordering) and oering
targeted promotions and discounts. Servers should
also use suggestive selling since it really doesn’t mat-
ter how long guests stay at a table. On the other hand,
trying to nd a beer table mix is not really an issue
since tables are empty anyway. Exhibit 5 summarizes
the appropriate tools to be deployed for hot and cold
periods
Hot Tools
Let’s look at the “hot tools” in more detail.
Adjusting table mix. The optimum table mix
matches the mix of table sizes and availability to the
mix of party sizes. Thus, telltale signs that indicate
that the table mix should be changed are when there’s
a mismatch between table and seat occupancy or a
mismatch between the party-size mix and the table-
size mix, and when there’s a queue because all tables
are occupied even though there are plenty of empty
seats.
The Center for Hospitality Research • Cornell University8
A study on the impact of the optimal table mix at
Chevys FreshMex restaurants found that the optimal
table mix would allow a restaurant to serve up to 35
percent more customers while maintaining the same
waiting time.8 Clearly, an improved table mix has
great promise for busy restaurants.
Table mix at Fairmont. While an optimal table
mix can help increase revenue, Fairmont noticed that
many of their restaurants were not busy enough to
justify the investment unless they were undergoing a
renovation. The rm viewed an optimal table mix as
an ideal, but from a practical perspective they instead
chose to focus on providing a exible table mix that
could be recongured by meal period and day of week,
on other restaurant design features, and on selecting
the right mix of reservations.
By having a exible table mix, Fairmont could
change their table mix according to expected party
size mix for each meal period (either from reservations
data or from historical data). This is not the same as
changing their table mix as parties arrive. As Thomp-
son has shown, changing the table mix “on the y” is a
suboptimal solution for larger restaurants (dened as
200 seats) since it results in idle tables.9
Fairmont has also been designing their restaurants
so that the various spaces ow into each other. For ex-
ample, the bar might ow into the restaurant. During
breakfast, they install soft separators such as ower-
pots to separate the restaurant from the bar, but for
other meal periods, they remove the soft separators so
that they are able to fully use both spaces.
Beer manage reservations. As with hotel RM
during high demand periods, it is important to make it
easy for people to make reservations, select the “right”
reservation requests, and manage both the arrival
uncertainty (late-shows and no-shows) and duration
uncertainty (length of meal).
Reservations Management at Fairmont
Select the right reservation requests. Fairmont has
adopted the approach of selecting the party-size mix
that best ts their table mix. For example, the Imperial
8 California-based Chevys FreshMex operates three dozen
restaurants, most of them in the southwestern United States. See:
Kimes, Sheryl E., and Gary M. Thompson. 2004. “Restaurant Reve-
nue Management at Chevys: Determining the Best Table Mix,”De-
cision Sciences,35(3): 371-392; and Kimes, Sheryl E. and Gary M.
Thompson. 2005. “An Evaluation of Heuristic Methods for Deter-
mining the Best Table Mix in Full-service Restaurants,”Journal of
Operations Management 23.6 (2005): 599-617.
9 Thompson, Gary M. 2002. “Optimizing a Restaurant’s Seat-
ing Capacity: Use Dedicated or Combinable Tables?,”Cornell Hotel
and Restaurant Administration Quarterly.43(4): 48-57.
E
XHIBIT
5
Appropriate RRM tools
Tool ‘Hot’ ‘Cold’
Menu Engineering Yes Yes
Menu Design Yes Yes
Server mentoring/
upselling Yes Yes
Table mix Yes No
Reservation
management Better manage Maximize
distribution
channels
Promotions Restrict Yes
Suggestive selling Only if doesn’t
extend duration Yes
Pricing Premium Selected
discounts
Bar at the Royal York in Toronto is quite busy dur-
ing happy hour (between 6:00 and 8:00 p.m.) for most
days of the week. Most of their tables are designed
for four or more guests. Since the GM did not want
to change the table mix until the Imperial Bar had to
undergo a renovation, they decided to not accept party
sizes of less than three during happy hour. Similarly,
Singapore’s Jaan restaurant started to select more
parties of four so that they could beer match their
party-size mix to their table mix.
Make it easy to buy. At the restaurants at the
Fairmont and Swissôtel in Singapore, we noticed that
about half of guests made their reservations the same
day as they dined and that there were more reserva-
tions on weekends. In addition, there were signicant
same-day reservation aempts made between 7:00 and
8:00 p.m. However, we observed that the reservation
oce closed at 7:00 p.m., at which point calls were
then directed to the restaurants (with a high likelihood
of not being answered!). By extending the reservations
oce closing time by one hour, the rm was able to
generate S$50,000 in incremental revenue per month.
Overbooking. Equinox, a ne-dining restaurant
also at the Swissôtel Singapore, had a 40-percent no-
show and cancellation rate. Given that there was no
penalty levied for no-shows, we analyzed the no-show
and cancellation rates in detail and developed appro-
priate overbooking levels. To reduce no-shows, we
had the reservations sta call to conrm reservations.
Cornell Hospitality Report • August 2019 • www.chr.cornell.edu • Vol. 19 No. 5 9
E
XHIBIT
6
Pricing. Pricing can be used
to help build demand during slow
Bow Valley Grill: Seat occupancy by day of week and hour of day
periods but also to capitalize on high
Hour of day
As a result, the cancellation rate dropped to about 21
percent.
Reduce arrival and duration uncertainty. We
noticed that Fairmont’s Chinese restaurants were
extremely busy for the traditional Chinese New Year
dinner. Oering 6:00 and 8:00 p.m. seatings controlled
meal duration to some extent, but the restaurants
still experienced no-shows and late-shows. As a rst
step, the restaurants implemented a non-refundable
pre-payment for the prix xe meal. As a result, no-
shows dropped signicantly. Management also did
not allow substitutions in the menu items and made
sure that food delivery began promptly at 6:00 or 8:00
regardless of whether the guests were on time for their
seating.
Don’t turn away potential business. The Jazz
Bar at the Fairmont Peace Hotel in Shanghai was
extremely busy during peak periods, and manage-
ment often turned customers away at the door. Rather
than continue to lose this business, they opened up the
adjoining Cin-Cin room for food and beverage service.
Guests could enjoy their reasonably priced drinks and
snacks while still listening to the music in the bar.
Promotions. Given that promotions are designed
to build demand during slow periods, they should not
be oered during hot periods since you don’t need the
extra demand. If a restaurant oers promotions dur-
ing busy periods, it might end up giving unnecessary
discounts.
demand periods by charging pre-
mium prices. During busy periods, a
restaurant might be able to charge a
premium or possibly increase prices
on popular menu items. Research has
shown that customers consider time-
of-day and day-of-week pricing to be
relatively fair, especially if framed as
a discount (that is, full price during
busy times and a reduced price dur-
ing slow periods).10
Pricing at Fairmont
Bow Valley Grill. Seat occupancy for
the Saturday and Sunday brunches at
the Bow Valley Grill in Ban, Alberta,
Canada, topped 90 percent. Other
times of day and days of week were
not all that busy, except for breakfast
between 8:00 and 10:00 a.m. (Exhibit
6). The director of F&B maximized his turns during
the brunches and increased the Sunday brunch price
by $3 (about a 10-percent increase). As a result, Sun-
day brunch prot increased by 6 to 7 percent.
City Space and Equinox. Similarly, City Space
and Equinox at the Swissôtel Singapore, which oered
an excellent view, employed premium pricing for their
window tables by instituting a $20 charge for non-
hotel guests using those tables. As a result, Equinox
generated nearly $100,000 per year from the window
table charges. In addition, the average check per per-
son for guests paying to sit at the window seats was
over $5 higher ($142.57) than those at non-window
seats ($136.92).
Suggestive selling. Suggestive selling should be
applied judiciously during hot periods. For example,
when only one or two guests at a table order a par-
ticular course, the server should suggest that the other
guests do so as well, or if guests only order one or
two drinks during their meal, eective servers should
ask all guests if they would like another drink. But
during busy times, it is probably unwise for servers to
push dessert (or any additional course) if no one has
ordered one since all this would do is increase meal
10 Kimes, Sheryl E. and Jochen Wir. 2003. “When Does Rev-
enue Management Become Acceptable?,” Journal of Service Research.
7 (2): 125-135; and Wir, Jochen and Sheryl E. Kimes. 2007. “The
Moderating Role of Familiarity in Fairness Perceptions of Revenue
Management,” Journal of Service Research. 9 (3): 229-240.
The Center for Hospitality Research • Cornell University10
E
XHIBIT
7
Jazz Bar: Number of checks
duration and preclude other guests from being seated
at the table.
Suggestive selling at Fairmont. With those
caveats in mind, Fairmont restaurants developed and
deployed a number of innovative methods of sugges-
tive selling during hot periods. For example, Prego in
Singapore oered a set menu. If one person at a table
chose that prix xe option, the server would suggest it
for everyone. Similar approaches were used at other
Fairmont restaurants.
Cold Tools
Tools for cold periods are intended to encourage
greater sales in slow times. As we mentioned above,
adjusting the table mix isn’t one of those tools, but
improving reservations, pricing, and upselling can be
valuable.
Reservations and distribution. When you’re cold,
be sure to make it easy for your customers to make a
reservation, and if you don’t take reservations make
it easy for them to order food from you. Well-chosen
outside distribution channels can also generate addi-
tional revenue (even if they add cost) since customers
can make reservations or order food whenever they
want, regardless of whether you’re open and answer-
ing the phone. In addition, food-service distribution
sites increase awareness of your restaurant and may
result in new customers who want to give you a try.
Promotions. Cold-period promotions could
include oering live music, developing special menus,
and building aliate programs. The important thing
to remember is to carefully target the promotions so
that they are only available during cold periods, and
they aract customers who might not normally have
come to your restaurant at all or at least would not
have come at that time.
Promotions at Fairmont. The Peace Hotel’s Jazz
Bar oered two music sets: a popular Old World
Shanghainese band that played from 8:00 to 11:00 p.m.
and an international modern jazz band that played
from 11:00 p.m. to 2:00 a.m. (Exhibit 7).
Restaurant managers noticed that while de-
mand was high (often with long queues) from 8:00
to 11:00, trac was relatively slow from 5:00 to 8:00
p.m. and after 11:00. They worked with travel agents
to develop group packages during the early evenings
and replaced the jazz band with a second Old World
Shanghainese Band. As a result, the Jazz Bar achieved
revenue 14 percent above budget.
Pricing. Pricing tactics during cold periods could
involve oering lower prices at certain times of day or
days of week. The important point here, of course, is
not to lower prices for guests who were going to buy
at full price. Thus, discounts must be fenced, meaning
that customers must meet certain conditions in order
to obtain the special price. Rate fences come in all
Cornell Hospitality Report • August 2019 • www.chr.cornell.edu • Vol. 19 No. 5 11
E
XHIBIT
8
Low lunchtime seat occupancy at Mikuni Roboyaki counter
forms including physical (e.g., table location), transac-
tion-based (e.g., time of day or day of week), custom-
er-based (e.g., age or group aliation), or controlled
availability (e.g., promotional code).
Pricing at Fairmount. At Mikuni, the lunch time
seat occupancy at the Robotayaki counter was quite
low (Exhibit 8), while the tables in the rest of the
restaurant were relatively busy. This restaurant also
turned to a prix xe approach, instituting a special
S$58 set lunch that was only available at the Ro-
botayaki counter. As a result, counter seat occupancy
increased by 25 percent, counter sales increased by 105
percent, and Mikuni achieved a 15-percent year-over-
year performance over budget.
Suggestive selling. Upselling and suggestive
selling are excellent tactics during cold periods. Since
the restaurant is slow anyway, it doesn’t particularly
maer how long customers occupy a table, meaning
that servers can suggest additional courses. If guests
don’t order an appetizer, for instance, servers should
recommend a starter. The same approach can be used
for espresso and desserts, as well as for after-dinner
drinks.
Suggestive selling at Fairmont. The Tonga Room,
a restaurant and bar in the San Francisco Fairmont, ap-
plied menu engineering principles to determine their
Cash Cows. When the analysis identied the Mai Tai
cocktail as a Cash Cow, the restaurant featured it as a
signature item. Servers would make a point of asking
guests whether they would like a second drink when
they were nishing o their rst drink. They also in-
creased the price of a Mai Tai from US$12 to US$13. As
a result, monthly sales increased by 47 percent.
Summary and Conclusion
Eective data collection and analysis are the key factors
in all the tactics and strategies that we have outlined in
this report. With the proper revenue management data
in hand, managers can have ready their all-purpose
strategies, their strategies to use when the restaurant is
busy, and strategies to use when things are slow.
We must also emphasize that implementing RRM
often involves overcoming signicant organizational
challenges. Given that RRM is a dierent way of think-
ing, it is typical to encounter some resistance. In imple-
menting the revenue management approach, it is also
important to be sensitive to the operational pressures
that the F&B team faces.
Fairmont sought to minimize any possible resis-
tance by involving F&B teams in the RRM process. The
company rst developed the dashboard that illustrated
the baseline performance. They then conducted short
property-based RRM seminars for the F&B operators,
presented the dashboard for specic outlets, and then
asked the F&B team from that outlet to look at the dash-
board and come up with two or three initiatives to test.
The revenue management team positioned itself as
the provider of analytical support and left the ideas and
implementation to the F&B teams. The RM group also
made sure to celebrate the teams’ successes both in per-
son, by reporting directly to the hotel general manager,
and through social media.
As a result, by 2016, RRM had been applied to over
70 percent of Fairmont’s restaurants’ F&B revenue
sources. Within 12 months of implementation, restau-
rants using RRM generated ve times more revenue
growth than restaurants not using RRM. n
The Center for Hospitality Research • Cornell University12
CHR Advisory Board
Cornell Hospitality Report
Vol. 19, No. 5 (August 2019)
Sco Berman ’84
Principal, Real Estate Business Advisory
Services,Industry Leader, Hospitality & Leisure
PwC
Nathalie Corredor
Senior Vice President, Strategy
Hilton Worldwide
Susan Devine ’85
Senior Vice President, Strategic Development
Preferred Hotels & Resorts
Chuck Floyd, P ’15 and ’18
Global President of Operations
Hya
RJ Friedlander
Founder and CEO
ReviewPro
Steve Hood
Senior Vice President of Research
STR
Taimur Khan MENG ’93
Vice President, GM Travel, Transportation,
Hospitality Solutions Team
Salesforce
Sanjeev Khanna
Vice President
Tata Consultancy Services
Balaji Krishnamurthy
Vice President of Global Strategy, Corporate
Development, and Business Intelligence
Sabre Hospitality Solutions
Faith Marshall
Director, Business Development
NTT DATA
Craig A. Mason ’85
Senior Vice President, Asset Management
Host Hotels and Resorts
Dan O’Sullivan
Vice President of Sales, EMEA
Translations.com
Abhijit Pal ’02
Head of Research & Strategic Initiatives, Global
Partner Group
Expedia Lodging Partner Services
©2019 Cornell University. This report may not be
reproduced or distributed without the express
permission of the publisher.
Cornell Hospitality Report are produced for the ben-
et of the hospitality and service industries by
The Center for Hospitality Research
at Cornell University
Glenn Withiam, Editor
Linda Canina, Academic Director
Carol Leitch-Zhe, Program Manager
Kate Walsh, Dean, E.M. Statler Professor,
School of Hotel Administration
The Center for Hospitality Research
School of Hotel Adminstration
Cornell SC Johnson College of Business
Cornell University
Statler Hall
Ithaca, NY 14853
607-254-4505
chr.cornell.edu
Dave Roberts, ENG ’87, MS ’88 (ENG)
Senior Vice President,Consumer Insight and
Revenue Strategy
Marrio International
Dan Skodol, MMH ’04
Vice President of Revenue Analytics
Rainmaker
John W. Spencer
Senior Managing Director, Global Hospitality
Leader
Accenture
Berry van Weelden, MMH ’08
Director, Reporting and Analysis
priceline.com
Adam Weissenberg ’85
Global Travel, Hospitality and Leisure Leader
and National Managing Partner, Clients &
Industries
Deloie & Touche USA
Jon Wright
President and Chief Executive Ocer
Access Point Financial
Cornell Hospitality Report • August 2019 • www.chr.cornell.edu • Vol. 19 No. 5 13
