Journal of Business Research 63 (2010) 519–527

Contents lists available at ScienceDirect

Journal of Business Research

Technology revenue management system for customer groups in hotels☆

José Guadix ⁎, Pablo Cortés, Luis Onieva, Jesús Muñuzuri
University of Seville, Spain

a r t i c l e i n f o a b s t r a c t

Article history: This paper discusses revenue management; a technique that focuses on decision making that will maximize
Received 1 April 2008 profit from the sale of perishable inventory units. New technologies management plays an important role in
Received in revised form 1 March 2009 the development of revenue management techniques. Each new advancement in technology management
Accepted 1 April 2009
leads to more sophisticated revenue business capabilities. Today decision support revenue management
systems and technologies management are crucial factors for the success of businesses in service industries.
Keywords:
Revenue management
This paper addresses the specific case of customer groups in hotels. This paper introduces a new decision
Customer groups support system that sets the revenue maximization criteria for a hotel. The aforementioned system includes a
Hotels set of demand forecasting methods for customers and addresses a general case considering individual guests
and customer groups. The system also incorporates deterministic and stochastic mathematical programming
models that help to make the best decisions. The actual revenue depends upon which reservation system the
hotel uses. A simulation engine makes a comparison between different heuristics of room inventory control:
the results include performance indexes such as occupancy rate, efficiency rate, and yield; it compares results
and chooses one of them. The system proves its suitability for actual cases by testing against actual data and
thus becoming an innovative and efficient tool in the management of hotels' reservation systems.
© 2009 Elsevier Inc. All rights reserved.

1. Introduction also consider overbooking and allowing for passenger adjustments.

Belobaba and Weatherford (1996) perform a comparison of various
Many firms seek to use revenue management techniques to decision making rules incorporating passenger adjustments.
maximize profitability in capacity-constrained situations. As firms In this way, they define revenue management as the sale of the
seek out revenue management techniques to squeeze profits from right inventory unit to the right customer at the right time. Their
increasingly more efficient business processes, researchers respond research focuses on hotels' revenue management as hotels use this
to these needs. In the past, different industries used most of the type of system to determine the number of available rooms at dif-
characteristics underlying this technique. Perishable firms, such as ferent rates. Rothsein (1971) performs the early work on overbooking
bakers, grocers, fresh fruit vendors, or theater managers regulated of hotel reservations. Liberman and Yechiali (1978) consider customer
demand by varying prices during specific periods of time. cancellations in a 24-hour period. Orkin (1988) outlines some of the
Following the US Airline Deregulation Act established in 1978, any ideas behind revenue management for hotels and provides examples
airline can now operate any route at any time with whichever fares of the different types of calculations. Bitran and Mondschein (1995)
they choose, point out Smith, Leimkuhler and Darrow (1992). These model hotel reservations including multiple day stays, and Bitran
facts have led the scientific community to develop a new management and Gilbert (1996) extend previous models to incorporate uncertain
approach called revenue management. Initially, revenue management arrivals.
techniques assumed that passengers chose from one particular fare Revenue management applies to the service industry when it
class, without moving to a lower fare if it became available. Companies meets the following five conditions (Kimes, 2000), each specifically
adopted differentiated pricing in order to compete for price sensitive adapted for hotels.
travelers, without giving up the revenue from their existing, full fare
1. Limited capacity. The design of revenue management target
customers. The later extension of these techniques allows for pas-
capacity-constrained services firms. The units of inventory sell in
senger flexibility amongst fare classes. Bodily and Weatherford (1995)
a short period of time with a fixed capacity, measured by the
number of rooms.
2. Market segmentation. Service industries make use of segmentation
☆ The author acknowledges the financial support given by the Ministerio de
because they can choose between different types of customers.
Educación y Ciencia (project ref. DPI2005-09132-C04-01), Spain.
⁎ Corresponding author. University of Seville, The School of Industrial Engineering,
They do not allow arbitrary pricing, so the service should have
Camino de los Descubrimientos, s/n, Seville 41092, Spain. some distinguishing characteristic so that it uses the same unit
E-mail address: guadix@esi.us.es (J. Guadix). of capacity to deliver many different services. Hotels usually use

0148-2963/$ – see front matter © 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jbusres.2009.04.013
520 J. Guadix et al. / Journal of Business Research 63 (2010) 519–527

purchase restrictions and refund requirements to help segment the Sections comprise the remainder of the paper. Section 2 presents
market between leisure and business customers. a new methodology used for tackling the problem in service in-
3. Future demand is uncertain. Revenue management must have the dustries. Section 3 addresses demand forecasting models that airlines
ability to forecast the demand variability so that managers can traditionally use and their adaptation for the hotel sector. Section 4
increase prices during periods of high demand and decrease prices presents the problem of optimizing room distribution. A new sto-
during periods of low demand. Hotels must set aside rooms for chastic model is the basis of the problem, with or without groups'
business customers, to protect them from the lower prices acquired option. Section 5 describes a simulation model where it defines
by leisure customers before they know how many business rooms arrivals under three different policies for room inventory control.
will sell. Section 6 discusses computational results and their comparisons.
4. Perishable units of inventory. Inventory distinguishes service firms This section includes the comparison of performance indexes for
from manufacturing firms. The units of inventory unsold after a heuristics, including occupancy rate, efficiency rate, and yield. Finally,
specific date go to waste in service industries, because services cannot Section 7 draws conclusions.
be stored. This special characteristic leads to the sale of services in
advance. Hotels cannot store rooms for use by tomorrow's customer. 2. Methodology
5. Appropriate cost and pricing structure. Many service firms have
a fixed cost capacity expense and a demand that cannot rapidly The TRM system comprises of three management levels (Jones and
adjust. In the same way, the additional cost of adding a new cus- Lockwood, 1998):
tomer to the available capacity is very low.
• Strategic level addresses the long-term and generally focuses at
This paper studies revenue management models including group the head office. TRM system data establishes market segmentation
acceptance in hotels. Customer groups for hotels have their own set of criteria and overall pricing policy in long-term, structural decisions.
characteristics that require a slightly different set of strategic levers • Tactical level deals with the intermediate-term running of individual
from the typical approaches in use for the individual customer. operating units. TRM system data establishes target occupancies for
Therefore, this study models the customer typology as an individual or different market segments in the intermediate-term.
as a group. The study tests a variety of different rooms' optimization • Operational level concerns itself with the short-term conduct of
algorithms, based on deterministic and stochastic programming tech- the operating system, such as the sales office or the front desk.
niques. The research intends to test a Technology Revenue Manage- Human capital constitutes a key determinant of the operational
ment (TRM) system in a hotel chain and to identify factors associated office in service industries, Arribas and Vila (2007). TRM system
with the management of different customer typologies. data decides what price to offer and what reservations to accept in
The hospitality industry needs to use technology management the short-term.
for its survival, and several studies show evidence of this necessity.
Donaghy, McMahon-Beattie and McDowel (1997) raise a 10-step Following this structure, we propose an original methodology as
model which stresses the use of technology management in the described in the figure below that features a brief description of the
segmentation of clients and the use of their characteristics in each architecture of the TRM system. Fig. 1 introduces the key components
market segment. Emeksiz, Gursoy and Icoz (2006) present a model in and gives an overview of information flows, decision and design, and
5 steps comparing those hotels using the technology management and the test stage. Shoemaker (2003) also includes “tactical level” within
those not using the system. It is also necessary to devise an asset to the “strategic management’, distinguishing the use of price changes in
clients as long term. Therefore, it is necessary to manage revenue the hotel. Later sections describe in detail each TRM system module.
management with CRM systems, Noone, Kimes and Renaghan (2003), TRM system follows four steps:
to ensure the provision of quality service, and customer loyalty for
the future. 1. Demand forecasting must come from historical data. Based on
However, businesses using different prices for the same service occupation rates from historical data, the company can forecast
offer to customers should do it very carefully. An example of such an future demand in a short-term period of time. The accuracy of
occurrence took place in 2000, Enos (2000), when Amazon.com sold forecasted demand is of special importance because it condi-
DVDs at different prices, and was offering discounts between 20% and tions the effectiveness of TRM system. Frequent updates to
40%, in accordance with the geographic area in which the customer historical data improve the accuracy of the model. Results from
was purchasing the product. The customers using ICTs and Internet this module.
could check the different prices for the same film. The experiment had 2. Optimal room distribution. The system uses forecasted data as
a negative impact on the company. In other sectors, such as the airline input to the application of the capacity models, so the forecasted
or hotel industry, price variations are higher and have not created quantity distributes among the different categories subject to the
any negative perceptions of the companies so far. This is because daily capacity of the hotel. A room distribution optimization model
the service that airlines and hotels offer at different prices is well sets booking limits at diverse fare levels.
differentiated by its characteristics, so that the customer receives 3. Room inventory control. Two differentiated phases make up this
tangible differences in the products or services offered. step: the arrival generation and the reservation system. First a
Six hotels in Andalusia (Spain) become the test sites of the pro- simulation engine generates arrival processes of customers, whose
posed decision support system, implementing the TRM system. These data helps set up the arrival generation submodule within the
hotels are part of a 4-star hotel chain with an average of 160 bedrooms room inventory control process. Conversely, the previously stated
per hotel and with locations on the southern coast of Spain, a optimal room distribution process, along with the arrival genera-
destination where the tourism industry is important on an interna- tion submodule, are inputs for the reservation system submodule.
tional level, Guzman, Moreno and Tejada (2008). TRM system focuses The room inventory control process states the rooms' sell mode and
on Marbella Hotels. These hotels stay open all year round, and the the reservation system. The sales manager must receive the defined
organization owns another hotel in Marbella. If necessary, guests can criterion to determine whether to accept or reject a request when a
move from one hotel to another. This hotel chain obtains high cus- customer arrives.
tomer satisfaction results, a necessary factor in service industries, 4. Real assignment. As a final step, the sales office offers room prices
Fullard (2007). Lindenmeier and Tscheulin (2008) address the same to individual customers and negotiates rates for group customers
aspects in another paper, but dealing with the airline industry. with tour operators and travel agents.
 J. Guadix et al. / Journal of Business Research 63 (2010) 519–527 521

Fig. 1. Technology revenue management system process flow.

Vinod (2004) raises a revenue management system applicable casting method depending on the time of year due to the strong
to the hospitality industry, stressing the technology needs of each of seasonal component. In general, regression model, linear, or loglinear
the modules that comprise the aforementioned industry. Along the regression should provide dependable data. Unpublished studies use
same line, Chiang, Chen and Xu (2007) address the importance of combination forecasting or specific methods as a pick-up model.
technology management in revenue management techniques. Group forecasts calculate the number of rooms available to in-
Historical data module updates automatically by incorporating dividual guests. There are two types of group demands; ad hoc and
data from sales and reservations. Also, data updates thanks to Internet series. Ad hoc groups consist of guests that are not regular in terms
and technology management, play an important role in revenue and of repetition of travel patterns (dates and/or services). They use a
pricing management. Nowadays, it is easier for customers to compare specified number of rooms and services for specific nights. A typical ad
prices amongst competitors, whilst service providers can get detailed hoc request might be a single or a few one-time rooms. Series groups
information about customer behavior much quicker. typically stay longer and come from tour operators or travel agencies.
These customers might request rooms in specific blocks of time or
3. Demand forecasting nights and reallocate them through tour packages.
If the group forecast is inaccurate, the total number of rooms
Revenue management depends highly upon an accurate forecast- available will be inaccurate, and the TRM system proposals may lead
ing needed for efficient reservation systems, and as input data for real- to poor decisions. Inaccurate group forecasts have a greater impact
life oriented optimization models. For a comprehensive literature during high occupancy periods of time. If group forecasts are too high,
review on forecasting models see McGill and Van Ryzin (1999), Talluri any mistake in the detection of such groups could lead to unused
and Van Ryzin (2004), Pai and Hong (2005) or Fernández-Morales rooms. Unfortunately individual guests, had there been prior knowl-
and Mayorga-Toledano (2008). edge, could have booked these rooms, instead of leading to un-
The TRM system uses the customers' demand forecasting as input necessary waste. The experimental results section presents the results
to obtain an optimal allocation of rooms. Usually the system calculates of the forecasting module for the different analyzed cases.
demand forecasting from historical arrival information taking into
consideration the length of the stay and room category. Different 4. Optimal room distribution
methods can work, from traditional approaches to advanced and/or
combined booking models, Lee (1990). Using the forecasting the guests' arrival, the system relies upon
Traditional forecasting techniques include moving average book- filling the available capacity by charging the highest price. This
ings, exponential smoothing, or ARIMA time series models amongst ensures that those customers most willing to pay for a room can do so.
other well known statistical approaches, Makridakis, Wheelwright Most of the optimization models follow the Williamson (1992)
and Hyndman (1998). Advanced booking models predict customer models, maximizing revenue using a deterministic mathematical
pickup. They consider the incremental booking received during a programming model, originally created for the airline industry. In the
certain time interval. Hybrid models include regression methods in hotel industry, the objective is to allocate rooms to maximize revenue,
which the independent variable is the number of reservations on hand while satisfying capacity constraints.
for a particular day and the dependent variable is the economics The optimal room distribution uses four models. The first is a
parameters from customer countries taking the final number of rooms deterministic model (DP), which accounts for the number of rooms in
sold. each category, taking into consideration individual guests only. The
There is not an agreement on the best method. In fact, every hotel deterministic group problem (DGP) considers the DP scenario but also
has its own particular characteristics, and a hotel may use a fore- customer group arrivals. The system determines the opportunity cost
522 J. Guadix et al. / Journal of Business Research 63 (2010) 519–527

Table 1 To represent the mathematical formulation of the problem,

List of parameters and variables of the models considered for optimizing the room Table 1 presents the data, parameters, and variables to deal with the
distribution in the TRM system.
different models to consider in the TRM system, and those previously
DP DGPa presented.
Data k Length of stay (in days) λg Length of group stay (in days) Once introducing Table 1, one can formulate the different models
pj Fare price (category j) μg Group size (customers) previously described (Fig. 2).
bi Hotel capacity on day i cg Fare group
To formulate the models, one follows the next hypothesis. Data
dijk Forecasted demand on day i,
staying k days at fare category j. updates automatically in order to solve the model with the latest
Variables xijk Integer variable. Number of rooms xg Binary variable. It represents information. This leads to a situation where cancellations have a very
reserved for the guest of ijk the possible acceptance of low impact because the system incorporates eventual cancellations
characteristics (arrival day i, fare j group, g. into the demand forecasting module varying the input data of the
and length of stay k)
optimal room distribution module that has the ability to re-run.
SP/SGPa, b Additionally, the system does not account for overbooking. Over-
Data r Number of alternative scenarios being considered depending on booking occurs when a hotel accepts more reservations than available
different customers' arrival process. It varies from 1 to S rooms. Depending on the country it could cause different legal issues
dijk,r Forecasted demand on day i, staying k days at fare category j according
when hotel managers use airline overbooking as a justification for the
to scenario r.
Dijk Demand taken from a discrete set of values {dijk,1 b dijk,2 b…b dijk,r} practice. However, the legal framework of the airline and the hotel
Variables xijk,r Integer variable. It represents the part of the demand Dijk falling into the industries differs. In actual practice, hotels overbook less frequently
interval (dijk,r − 1, dijk,r] than airlines.
a
In DGP and SGP problems, subscript i⁎ means the arrival of a group on day i⁎ at Fig. 2 presents the four models dealt with in this paper. First, the
difference from the subscript, i, applicable to arrivals of individual customers. DP model attempts to maximize average profit per available unit by
b
SP considers the same set of data and variables in problem DP plus those specific for anticipating the price sensitivity of different customers and by
stochastic problems at the box below. The same happens with respect to SGP and DGP
anticipating the possibility of reserving a room for the customers
problems.
willing to pay the highest price. The model selects a number of rooms
the guest can reserve of ijk characteristics, which stands for the arrival
due to the assignment of a set number of rooms to a group instead of on day i, at a fare j, and for a stay of k days. The constraints of the
individual customers. Individual customers usually pay more expen- model include the daily capacity (in rooms) of the hotel and the
sive rates than customer groups, but individual customers have a expected demand.
higher probability of no-shows, so there is greater uncertainty of them Svrcek (1991) introduces an extension of the DP model including
arriving. group reservation. Groups are special clients because they make
The stochastic problem (SP) considers the possibility of an arrival bookings in advance, include blocks of rooms, and sometimes need
differing from the mean, taking into account the natural variability of conference rooms. Groups are also sensitive about price. In actual
demand. The main problem corresponds with situations where more situations, tour operators or travel agents negotiate the group rate.
requests exist than what appears as the mean value. In such cases, During negotiation, tour operators contact the reservation supervisor
there are more customers willing to stay at the hotel than the requesting a specific number of rooms for a specific period of time. In
expected amount. On such occurrences, the probability of customers addition, the group usually needs extra services such as food and
accepting higher rates is greater than usual in deterministic models beverages, conference rooms, etc. With these requests the hotel
and therefore revenues would increase. Afterwards, it presents a requires the minimum amount for a room to remain profitable in
stochastic demand model including groups of customers; this is the order to accept or reject decisions. Group requests can replace in-
stochastic group problem (SGP) which considers the SP problem, plus dividual customers that could pay higher fares. Some group customers
group consideration. may occupy rooms with higher expected marginal revenue than other

Fig. 2. Optimal room distribution models.

 J. Guadix et al. / Journal of Business Research 63 (2010) 519–527 523

customers. However, the total group revenue may be higher than 5. Room inventory control
selling these rooms to individual customers.
DGP model maximizes the profitability of individual guests and In the previous section, the mathematical models allocated the finite
customer groups. The model modifies the capacity constraint for the rooms' inventory to the demand. The next step defines the operational
days expecting groups of customers. The hotel must have a large work, when a customer requests a room. In such a situation, the
enough capacity to lodge such groups along with individual guests. The reservation supervisor must decide whether or not to accept this guest.
model uses a variable binary to accept or reject the group requests. He/she must analyze the profit of reserving the room at that moment in
However, in practice the demand is stochastic. Stochastic demand time or waiting for another potential customer to arrive in the near
means that the number of allocated rooms could be different from future and pay a higher fare.
the forecasted amount of requested rooms. The study considers a Below is a developed set of heuristics taking into account the
stochastic programming model, SP, with a simple resource problem. acceptance or denial of such requests depending on a few parameters
These particular stochastic problems do not cause severe computa- in the TRM system developed.
tional difficulties, Kall and Wallace (1994). De Boer, Freling and
1. First-Come First-Serve (FC FS). This simple rule evaluates reservation
Piersma (2002) introduce a stochastic model for the airline industry,
request based on the well known first-come first-serve criterion. This
assuming that discrete values are possible scenarios depending on
rule disregards any room distribution. Whoever requests the room
customer demand.
first gets the room.
Therefore, the model divides the number of rooms reserved xijk
2. Distinct. This heuristic considers the protection of rooms according
into possible scenarios, that they rename as decision variables xijk,r.
to the optimal room distribution proposed by the four models. The
Such variables differ from zero when xijk,r − 1 is equal to dijk,r − 1, that is
arrival simulation engine allows for the selection of the better
Pr(xijk = dijk,r − 1) = Pr(xijk = dijk,r − 1). However, the sum of xijk,r rooms
solution from the four models in the simulation.
sold to customers in S scenarios must agree with the daily capacity
3. Nested. This method clusters the number of fare prices into smaller
constraint.
buckets. Williamson (1992) proposes this method, suggesting a
Following De Boer et al. (2002), the assumption is that three
procedure to book rooms that considers higher fares and in turn
demand scenarios are enough to capture most of the extra revenue
utilizes the rooms reserved for the cheaper fare but charging the
generated by excess customers. The forecasted mean calculates these
higher price. The highest fare price class has an inventory limit
demands by adding up and taking away the standard deviation, thus
equal to the daily capacity.
generating a three-value band for every price.
Although the study presents a stochastic model for individual Using a rolling horizon simulation of the reservation and a non-
customers, we develop an original model for stochastic demand con- homogeneous Poisson arrival process they run tests using the three
sidering groups, SGP. This consideration does not appear in scientific heuristic rules, suggested by Lewis and Shedler (1979) three decades
literature thus far consulted. As an objective, the model searches ago, and still considered today a common basis for arrival generation.
for the better method for the assignment of rooms, taking into There is a comparison between the results of the heuristics
consideration the arrival of individual guests and customer groups, simulations and a basic scenario case where they choose the arrival
and accounting for the stochasticity of the demand. rate of individual customers function, λ(t), from the historical daily
The individual customer demands must agree with the three bands pattern and positive correlates for fares (for example, the arrival rate
previously discussed, and corresponding constraints state such con- of customers is higher during the afternoon than in the evening).
sideration. Additionally, the daily capacity of the hotel must be sufficient In the customer groups case, guests arrive in batches, instead of
enough to lodge the stochastic arrival of individual customers and arriving one at a time. Using a discrete distribution that arranges
groups. successive batches into their sizes, they construct the arrival process of
Integer programming models comprise all of the problems. How- such groups. Also, they create the number of each customer batch
ever, the model can set the individual guests' variability to continuous with a random variable (Fig. 3).
due to the unimodularity property of the constraint coefficient matrix.
Consequently, they can all reformulate as linear problems (cases of 6. Results and discussion
DP and SP) or mixed integer linear problems (cases of DGP and SGP),
considering deterministic or stochastic demand depending on the To test the suitability of the TRM system, the experiment uses
model. historical data from an actual Spanish hotel chain with six hotels on

Fig. 3. Heuristics of room inventory control.

524 J. Guadix et al. / Journal of Business Research 63 (2010) 519–527

Table 2 Table 3 shows the average results for a 30 day period and the
Individual price classes. twelve alternatives and they compare with the real optimum
Class Price (€) distribution. The aforementioned table contains the obtained average
Premiere/luxury fare 250 daily incomes, sorted by capacity distribution model, and the room
Business/superior fare 175 assignment method for a non-homogeneous Poisson process.
Standard/normal fare 125 As Table 3 indicates, the best room distribution is a combination
Economy/discount fare 90
of group models (DGP/SGP) with assignment rule based on nested
Supereconomy/superdiscount fare 75
heuristic. The results of such a combination show an average error of
less than 5% with respect to the actual optimal distribution. On the
contrary, models not based on customer groups consideration report
the southern coast of Spain. In the company, an analyst is responsible errors higher than an average of 8%, nearly 3,000 Euro daily. Also, the
for making the daily decisions that the TRM system supports, and a efficiency, occupancy, and yield factors reveal the convenience of such
sales team is responsible for checking the outputs of the system, using an approach because it provides more adjusted rates. In fact, group
such information to deal with groups and negotiate prices. consideration is of higher importance when considering the groups of
The company provides historical data that is the input information customers.
needed for the demand forecasting module. The company carries However, a detailed analysis is necessary. To do so, one must
out the forecasting for a 30 day-rolling-horizon because company consider Figs. 4 and 5, which include the daily analysis. They consider
managers consider a month as the longer horizon including reliable the results for the four performance indexes: incomes, occupancy,
data to be forecasted and planned. The forecasting shows how great efficiency, and yield. The figures analyze such results with respect to
volatility makes it extremely difficult to achieve accurate forecasts. the optimal room distribution models (Fig. 4) and with respect to
They use the forecasted demand for each day to obtain the optimal room inventory control heuristic rules (Fig. 5).
room distribution, considering the four mathematical models: DP, Fig. 4 depicts the daily evolution of the four indexes with respect to
DGP, SP, and SGP. Each model produces a different proposal that the the four different optimal room distribution mathematical models.
TRM system considers. Models work using CPLEX 8.0. DGP and SGP (group models) lines are always on top of the DP and SP
They consider a target hotel of 200 available rooms because it lines that consider only individual customers. DGP performs best most
represents the standard hotel of the company. The interval [0, 21] of the time. This result is mainly due to the consideration of all of
randomly generates the length of the stay, k in mathematical models. the typologies of customers, and this allows for a better adaptation to
Individual guests have the ability to book at five different fares, which the demand and the behavior of customers. However, some days show
Table 2 describes. poorer results due to no-shows. For example, refer to day 6 in the
For the stochastic models, they take into account three different figure in question.
scenarios: low track line, average, and high track line. It corresponds to Also, the deterministic approaches show better performance
the r = 1,…,S in the models. Following De Boer et al. (2002) we set a related to the occupancy, efficiency, and yield rates. The difference
probability for each scenario equal to p1: 0.8/0.6/0.4; p2: 0.6/0.4/0.2 between deterministic and stochastic models is the expected value of
and p3: 0.7/0.5/0.3. perfect information, EVPI. It shows how much someone could expect
The arrival of customers provided by the demand forecasting module to earn if they were told what would happen before making their
corresponds to the daily arrival. Therefore, they must distribute this decision. It measures the value of randomness, but it does not show
value through the day by hours. They carry out this distribution by using that the deterministic models are dysfunctional. A small EVPI means
a simulation engine based on ARENA simulation software. that randomness will play a minor role in the model, whereas with a
These arrivals are a non-homogeneous Poisson process with an large EVPI randomness plays a major role.
arrival rate λ(t) depending on the time. They construct an actual daily Despite this, the stochastic model considering groups (SGP) ob-
pattern by taking into account the expert opinion of the people in- tains very good results regarding incomes, although not as good as the
charge in the hotel chain. deterministic model, DGP. After analyzing the global behavior, one can
The reservation system uses the arrival generation together with see that the deterministic group room distribution model presents the
the four proposals from the optimal room distribution to propose best alternative of the analyzed options.
the room assignments. To do so the systems use FC FS, distinct Fig. 5 presents the daily evolution of the four performance indexes
and nested heuristics for the four proposals from the optimal room related to the three assignment heuristics of the room inventory
distribution. The TRM system must analyze and compare twelve dif- control. Generally, the nested line shows the better performance.
ferent proposals. However, the distinct method sometimes provides better assignments
The expected incomes from the twelve alternatives are compared which is the case between approximately days 4 to 10 due to the fact
among them and with a value referred to as “real optimum dis-
tribution”. Such real optimum distribution corresponds to better dis-
Table 3
tribution after analyzing the “a posteriori” actual overall number of
Comparison of average results.
customer arrivals knowing all the information.
The following expressions calculate the percentages of occupancy, DP DGP SP SGP
efficiency, and yield: FC FS Incomes 20,670.70 22,837.50 20,670.70 22,837.50
Occupancy 65.03% 69.75% 65.03% 69.75%
Efficiency 78.21% 86.41% 78.21% 86.41%
number of rooms occupied Yield 59.06% 65.25% 59.06% 65.25%
Occupancy = × 100
maximum daily capacity DISTINCT Incomes 21,881.04 24,150.00 22,023.43 24,386.25
Occupancy 67.67% 62.61% 67.98% 63.13%
number of customers accepted
Efficiency = × 100 Efficiency 82.79% 91.38% 83.33% 92.27%
total number of rooms Yield 62.52% 69.00% 62.92% 69.68%
actual rooms income NESTED Incomes 22,782.86 25,278.75 22,972.71 25,291.88
Yield = × 100 Occupancy 69.64% 65.07% 60.05% 65.10%
potential rooms income
Efficiency 86.21% 95.65% 86.92% 95.70%
Yield 65.09% 72.23% 65.64% 72.26%
Yield rate indicates the real incomes with respect to the maximum RODa Incomes 23,732.14 26,250.00 23.732,14 26,250.00
a
possible income assuming all of the rooms sell at the full rack rate. Supposed real optimum distribution after real requesting by customers.
 J. Guadix et al. / Journal of Business Research 63 (2010) 519–527 525

Fig. 4. Daily optimal room distribution models.

Fig. 5. Daily results for heuristic assignment in the room inventory control.
526 J. Guadix et al. / Journal of Business Research 63 (2010) 519–527

Table 4 selling rooms (volume of sales), on some occasions not making a sale
Comparison of computational times (in seconds). could be more suitable, because it could increase revenues. In fact, this
Average time Maximum time Minimum time Standard deviation revenue objective can lead to lower room sales. The TRM system takes
DP 0.91 2.45 0.51 0.65 into account such aspects, and although the sales team could be
DGP 1.41 2.25 0.93 0.42 recommending increases of room sales at their own discretion, TRM
SP 2.45 3.06 1.85 0.38 system would be preventing the former from offering such discounts
SGP 3.65 4.81 2.70 0.66
to wait for customers willing to pay more in the near future.
In terms of future work we are focusing this approach on many
other service industries, to which this system can adapt considering
that ultimately they do not reach expected demand. Consequently, their particular characteristics. Another issue we are analyzing is
many rooms were not sold to first-come first-serve customers, mainly conceiving group auction setting. This work will involve other func-
economy fare customers. Ultimately the rooms remain empty. The FC tional areas of the company, as pricing analysts and product-design
FS method is a basis method when one does not take action for groups. In this way, we are exploring different alternatives of price
distributing rooms. This method shows as having a worse trend than negotiations among travel agencies, tour operators, and hotels
the others. The global behavior leads to the recommendation of the owners. In addition, we are exploring customer behavior and demand
nested method as the best alternative. models based on individual customer choice, random-utility models,
Figs. 4 and 5 allow the observation of the daily evolution as and aggregate market-demand, product interactions with demand for
function of the optimal room distribution models and the assignment other products and dependence on historical products attributes
heuristics. This analysis goes beyond the average results shown in incorporated in its specification, Konecnik and Gartner (2007).
Table 3 depicting daily limit situations that allows analysis based on Another limitation of this system concerns knowledge manage-
maximum and minimum deviations and not only on average results. ment. Improving information processing that allows for an extensive
The final interesting parameter of the models considered, is the use of knowledge transfer, knowledge reuse, storage and production
computational time. The models run on a PC Pentium IV 3 GHz with of knowledge is necessary. Hallin and Marnburg (2008) suggest new
2 Gb RAM memory, and use CPLEX 8.0 as optimization software. All lines to explore such aspects.
tested approaches obtain feasible times, all executed in less than 5 s.
Table 4 summarizes the computational times related to the average References
time, maximum and minimum times, and standard deviations.
Arribas I, Vila J. Human capital determinants of the survival of entrepreneurial service
firms in Spain. Int Entrep Manage J 2007;3:309–22.
7. Conclusions and further research
Belobaba P, Weatherford L. Comparing decision rules that incorporate customer diversion
in perishable asset revenue management situations. Decis Sci 1996;27:343–63.
This paper presents a decision support revenue management system Bitran G, Gilbert S. Managing hotel reservations with uncertain arrivals. Oper Res
as a sophisticated technology helping managers to make decisions in the 1996;44:35–49.
Bitran G, Mondschein S. An application of yield management to the hotel industry
framework of the hotel industry. The situation presented corresponds to considering multiple day stays. Oper Res 1995;43(3):427–43.
an inventory perishable problem under limited capacity, which price Bodily S, Weatherford L. Perishable-asset revenue management: generic and multiple-
policies differentiate. price yield management with diversion. Omega 1995;23(2):173–85.
Chiang W-C, Chen J, Xu X. An overview of research on revenue management: current
The TRM system includes a demand forecasting module to es- issues and future research. Int J Revenue Manage 2007;1(1):97–128.
timate the arrival of customers from historical data, an optimal room De Boer S, Freling R, Piersma N. Mathematical programming for network revenue
distribution based on mathematical models to distribute the fore- management revisited. Eur J Oper Res 2002;137:72–92.
Donaghy K, McMahon-Beattie U, McDowel D. Implementing yield management:
casted demand into different categories subject to the daily capacity of lessons from the hotel sector. Int J Contemp Hosp Manag 1997;9(2):50–4.
the hotel, a room inventory control module consisting of an arrival Emeksiz M, Gursoy D, Icoz O. A yield management model for five-star hotels: computerized
generation and a reservation system, and finally a real assignment and non-computerized implementation. Int J Hosp Manag 2006;25:536–51.
Enos L. Amazon apologizes for pricing blunder. E-Commerce Times; 2000. Sept 28.
module helping the sales office to offer room prices to individuals and Fernández-Morales A, Mayorga-Toledano MC. Seasonal concentration of the hotel demand
group customers. in Costa del Sol: a decomposition by nationalities. Tour Manag 2008;29:940–9.
Literature on group customers is scarce by and less agreeable. First, Fullard F. A model to evaluate the effectiveness of enterprise training programs. Int
Entrep Manage J 2007;3:263–76.
we consider a special case for the problem, which models as deter-
Guzman J, Moreno P, Tejada P. The tourism SMEs in the global value chains: the case of
ministic programming. Then, we use stochastic programming to solve Andalusia. Service Business 2008;2:187–202.
the same case. The consideration of such a customer groups model is an Hallin CA, Marnburg E. Knowledge management in the hospitality industry: a review of
original idea in the scientific literature dealing with the hotel industry. empirical research. Tour Manage 2008;29:366–81.
Jones P, Lockwood A. Operations management research in the hospitality industry. Int J
The study experiments with several models. The analysis of the Hosp Manag 1998;17(2):183–202.
experimental results concludes that the room distribution based on Kall P, Wallace S. Stochastic Programming. John Wiley & Sons; 1994.
group models together with a nested inventory control assignment Kimes S. A strategic approach to yield management. In: Ingold A, McMahon-Beattie U,
Yeoman I, editors. Yield Management: Strategies for the service industries. London:
method provides the best results. Continuum; 2000. p. 3–14.
This TRM system needs a special implementation of IT department, Konecnik M, Gartner WC. Customer-based brand equity for a destination. Ann Tour Res
special in the sense that it is based on particular models that are highly 2007;34(2):400–21.
Lee A. Airline Reservations Forecasting: Probabilistic and Statistical Models of the
data-fragile. This IT system will not perform in good order without Booking Process. Massachusetts Institute of Technology, M.I.T. Libraries Theses
data or with not worth gathering data, and the system would not Collection; 1990.
perform correctly as a consequence. TRM system needs data collected Lewis P, Shedler G. Simulation of nonhomogeneous Poisson processes by thinning. Nav
Res Logist Quart 1979;26:403–13.
at lowest level and stored for a relatively long time in operational Liberman V, Yechiali U. On the hotel overbooking problem. Manage Sci 1978;24:1117–26.
databases. The TRM system follows a wide spectrum of technology Lindenmeier J, Tscheulin DK. The effects of inventory control and denied boarding on
management focusing on planning, organizing, staffing, implement- customer satisfaction: the case of capacity-based airline revenue management. Tour
Manage 2008;29:32–43.
ing, and monitoring and evaluating stages oriented on how to use
Makridakis S, Wheelwright SC, Hyndman RJ. Forecasting: Methods and Applications.
technology to gain profit. John Wiley and Sons; 1998.
The proposed TRM system provides a suitable alternative for McGill J, Van Ryzin G. Revenue management: research overview and prospects. Transp
the management of any inventory perishable problem under limited Sci 1999;33(2):233–56.
Noone BM, Kimes SE, Renaghan LM. Integrating customer relationship management
capacity, concretely for every hotel located in every place of the and revenue management: a hotel perspective. J Revenue Pricing Manag 2003;2
world. Although some hotel chains usually focus their energy on (1):7–22.
 J. Guadix et al. / Journal of Business Research 63 (2010) 519–527 527

Orkin E. Boosting your bottom line with yield management. Cornell Hotel Restaur Adm Talluri K, Van Ryzin G. The Theory and Practice of Revenue Management. Kluwer Academic
Q 1988;28(4):52–6. Publishers; 2004.
Pai PF, Hong WC. An improved neural network model in forecasting arrivals. Ann Tour Vinod B. Unlocking the value of revenue management in the hotel industry. J Revenue
Res 2005;32(4):1138–41. Pricing Manag 2004;3(2):178–90.
Rothsein M. An airline overbooking model. Transp Sci 1971;5:180–92. Williamson E. Airline Network seat inventory control: Methodologies and revenue
Shoemaker S. Future of revenue management: the future of pricing in services. J Revenue impacts. Massachusetts Institute of Technology, M.I.T. Libraries Theses Collection;
Pricing Manag 2003;2(3):271–9. 1992.
Smith B, Leimkuhler J, Darrow R. Yield management at American airlines. Interfaces
1992;22(1):8–31.
Svrcek T. Modeling Airline Group passenger Demand for Revenue Optimization.
Massachusetts Institute of Technology, M.I.T. Libraries Theses Collection 1991.