Description
It’s a fact that a successful company not only put customers first, but put customers at the center of the organization because the changes in customer behavior determines unpredictable profitability and may be the cause for inefficient marketing planning.
3261
IMPROVING CUSTOMER RELATIONSHIP MANAGEMENT
IN HOTEL INDUSTRY BY DATA MINING TECHNIQUES
MIRELA DANUBIANU, VALENTIN CRISTIAN HAPENCIUC
Mirela DANUBIANU, Lecturer Ph. D. Eng, Ec.
“Stefan cel Mare” University of Suceava
Valentin Cristian HAPENCIUC, Associate Professor, Ph.D. Ec.
“Stefan cel Mare” University of Suceava
Keywords CRM, data mining hotel industry
1. Introduction
It’s a fact that a successful company not only put customers first, but put customers
at the center of the organization because the changes in customer behavior determines
unpredictable profitability and may be the cause for inefficient marketing planning.
The main goal of CRM is the capability to handle customer interaction across dif-
ferent channels and functions, for building loyal and profitable customer relationships.
Although cost cutting and competitive pricing strategies may attract customers
from competitors, in many services industries price advantages are not a sufficient
reason for customers moving between suppliers. In these situations successful
competitive strategies include developing strong relationships with customers and
cross-selling them other services.
Data mining - techniques for exploration and analysis of large quantities of data in
order to discover meaningful patterns and rules - helps businesses sift through layers of
seemingly unrelated data for meaningful relationships, where they can anticipate, rather
than simply react to, customer needs.
2. An overview of CRM
Customer Relationship Management (CRM) is an enterprise customer-centric
approach that uses different techniques to understand and influence consumer behavior.
It is a process which has two objectives:
• to impact all aspects to the consumer relationship (improve customer
satisfaction, enhance customer loyalty or increase profitability)
• to ensure that employees within an organization are using CRM tools. The need
for greater profitability requires an organization to proactively pursue its relationships
with customers [5]
In the real world, acquiring, building, and retaining customers are becoming top
priorities. For many companies, the quality of their customer relationships provides
their competitive edge over other businesses. In addition, the definition of customer has
been expanded to include immediate consumers, partners and resellers - in other words,
everyone who participates, provides information, or requires services from the
company.
Companies are beginning to realize that surviving an intensively competitive and
global marketplace requires closer relationships with customers.
In turn, enhanced customer relationships can boost profitability in three ways:
• by attracting more suitable customers,
• by generating profits through cross-selling and up-selling activities, and
3262
• by extending profits through customer retention.
Generally CRM may be defined by a framework composed by four elements:
know, target, sell and service[6].
It requires the company to know and understand its markets and customers. This
involves detailed customer informations in order to select the most profitable customers
and identify those no longer worth targeting. CRM also entails development of the
offer: which products to sell to which customers and through which channel. In selling,
firms use campaign management to increase the effectiveness of the marketing
departments. Finally, it seeks to retain its customers through services such as call
centers and help desks.
CRM is an association of several components. Before the beginning of the process,
the company must have customer information. These informations may proceed from
internal sources (summary tables that describe customers, customer surveys or
behavioral data contained in transactions systems) or the data can be purchased from
outside sources.
A critical component of a successful CRM strategy is an enterprise data warehouse.
Then, it must analyze the data using statistical tools, OLAP, and data mining. The last
component is campaign execution and tracking.
3. Data mining
Data mining is the exploration and analysis, by automatic or semiautomatic means,
of large quantities of data in order to discover meaningful patterns and rules[3]. So, data
mining is defined as the process of extracting interesting and previously unknown
information from data, and it is widely accepted to be a single phase in a complex
process known as Knowledge Discovery in Databases (KDD).
This process consist of a sequence of the following steps [12]:
• after analysing the goals of the end user and receiving all necessary prior
knowledge, one selects a target data set. This means focusing on a subset of variables
or on data samples.
• the target data are preprocessed and cleaned in order to remove noise or
outliers. One also has to decide how to handle missing data fields.
• useful features have to be found to represent the data, depending on the goal of
the discovery task. The dimensionality is reduced, i.e. one has to find the effective
number of variables under consideration, or invariant representations for the data.
• the primary goals of the knowledge discovery process are predicting the future
values of interesting variables or finding human-interpretable patterns in data.
According to this goal an appropriate data mining algorithm is chosen and applied.
There are algorithms for association, classification, clustering, sequence-based analysis,
and other tasks.
• the patterns are interpreted and evaluated, for example with the aid of
visualisation tools.
After each step, one can return to any other step prior to the current step. Thus, the
knowledge discovery process may contain many loops between any two steps.
In order to ensure that the extracted information generated by the data mining
algorithms is useful, additional activities are required, like incorporating appropriate
prior knowledge and proper interpretation of the data mining results.
In general, CRM promises higher returns on investments for businesses by
enhancing customer-oriented processes such as sales, marketing, and customer service.
3263
Data mining helps companies build personal and profitable customer relationships
by identifying and anticipating the needs of customers throughout the customer
lifecycle.
Data mining can help to reduce information overload and improve decision
making. This is achieved by deriving and refining useful knowledge through a process
of searching for relationships and patterns from the extensive data collected by
organizations. The extracted information is used to predict, classify, model, and
summarize the data.
Data mining technologies, such as rule induction, neural networks, genetic
algorithms, fuzzy logic, and rough sets, are used for classification and pattern
recognition in many industries [4][10][11].
By example, a supermarket organizes its merchandise stock based on purchase
patterns of shoppers, an airline reservation system uses travel patterns of customers and
trends to increase seat utilization., or the web pages alter their organizational structure
or visual appearance based on information about the person who is requesting the
pages.
Data mining builds models of customer behavior by using statistical and machine-
learning techniques. The basic objective is to construct a model for one situation in
which the answer or output is known and then apply that model to another situation in
which the answer or output is desired. The best applications of the above techniques are
integrated with data warehouses and other interactive, flexible business analysis tools.
Hence, data-mining applications can help companies to identify market segments
containing customers with high profit potential, by searching for patterns among the
different variables that serve as effective predictors of purchasing behaviors.
Marketers can then design and implement campaigns that will enhance the buying
decisions of a targeted segment. To facilitate this activity, marketers feed the data-
mining outputs into campaign management software that focuses on the defined market
segments.
Regarding the three ways of boosting profitability discussed in the above section,
the data mining techniques may be used as following:
• for attracting more suitable customers: Data mining can help firms understand
which customers are most likely to purchase specific products and services, thus
enabling businesses to develop targeted marketing programs for higher response rates
and better returns on investment.
• for better cross-selling and up-selling: Businesses can increase their value
proposition by offering additional products and services that are actually desired by
customers, thereby raising satisfaction levels and reinforcing purchasing habits.
• for better retention: Data-mining techniques can identify which customers are
more likely to defect and why. A company can use this information to generate ideas
that allow them to maintain these customers.
Moreover, there are additional ways in which data mining supports CRM
initiatives.
• Database marketing: Data mining helps database marketers develop campaigns
that are closer to the targeted needs, desires, and attitudes of their customers. If the
necessary information resides in a database, data mining can model a wide range of
customer activities. The key objective is to identify patterns that are relevant to current
business problems. For example, data mining can help answer questions such as
“Which customers are most likely to acquire a certain tourist’s service?” Answering
3264
these types of questions can boost customer retention and campaign response rates,
which ultimately increases sales and returns on investment.
Table 1.
Possible association between data mining techniques and CRM operations
Data mining technique CRM operation
Association rules
Information from customer-purchase histories is used to
formulate probabilistic rules for subsequent purchases.
Decision trees
Automatically constructed from data, these yield a sequence
of step-wise rules; good for identifying important predictor
variables, non-linear relationships, and interactions among
variables.
Descriptive statistics
Averages, variation, counts, percentages, crosstabs, simple
correlation; used at the beginning of the data-mining process
to depict structure and identify potential problems in data.
Genetic algorithms
Use procedures modeled on evolutionary biology to solve
prediction and classification problems or develop sets of
decision rules.
Neural networks
Applications that mimic the processes of the human brain;
capable of learning from examples (large training sets of
data) to discover patterns in data.
Query tools
Provide summary measures such as counts, totals, and
averages.
Regression-type models
Ordinary least-squares regression, logistic regression,
discriminant analysis; used mostly for confirmation of models
built by “machine-learning” techniques.
Visualization tools
Histograms, box plots, scatter diagrams; useful for
condensing large amounts of data into a concise,
comprehensible picture
• Customer acquisition: The growth strategy of businesses depends heavily on
acquiring new customers, which may require finding people who have been unaware of
various products and services, who have just entered specific product categories (for
example, new parents and the diaper category), or who have purchased from compe-
titors. Although experienced marketers often can select the right set of demographic
criteria, the process increases in difficulty with the volume, pattern complexity, and
granularity of customer data. Highlighting the challenges of customer segmentation has
resulted in an explosive growth in consumer databases. Data mining offers multiple
segmentation solutions that could increase the response rate for a customer acquisition
campaign. Marketers need to use creativity and experience to tailor new and interesting
offers for customers identified through data-mining operations.
• Campaign optimization: Many marketing organizations have a variety of
methods to interact with current and prospective customers. The process of optimizing a
marketing campaign establishes a mapping between the organization set of offers and a
given set of customers that satisfies the characteristics and constraints of a campaign,
defines the marketing channels to be used, and specifies the relevant time parameters.
Data mining can elevate the effectiveness of campaign optimization processes by
modeling the channel- specific responses of customers to marketing offers.
3265
The different data mining techniques may be associated with CRM tasks. Table 1
present some of these associations.
4. Data Mining Application for the Hotel Industry
Information technology was initially viewed by the hotel industry as a back-office
function that supports the finance and accounting areas. The industry has advanced far
beyond this view during the past decade.
In two sessions sponsored by the International Hotel and Restaurant Association
(IH&RA), one in Singapore in 1997 and the second in Nice, France, in 1998, hotel-
industry leaders pondered the role of technology. Among the conclusions reached were:
“Going forward, technology will be the most competitive weapon for any hospitality
company.
If hospitality organizations want to compete successfully, they must do so by using
technology to drive value to both the customer and to the firm.”[7]
In the hotel industry knowing the guests - where they are from, how much they
spend, and when and on what they spend it- can help a company to formulate marketing
strategies and maximize profits. Due to technological development hotel companies
have accumulated large amounts of customer data, which can be organized and
integrated in databases that can be used to guide marketing decision.
Because identifying important variables and relationships located in these
consumer-information systems can be a difficult task, some hotel companies have
attempted to raise the power of information by using data mining technologies that
exploits the data regarding the consumer.
Such data-mining technology allows hotel companies to predict consumer-behavior
trends, which are potentially useful for marketing applications. For example, Best
Western marketing staff can run reports and analysis on customer and occupancy data
stored in a data warehouse that combines customer and transaction information from all
company properties.
Such information indicates where live the customers who visit a specific hotel. If
the data reveal that the Best Western in Gura Humorului experiences a abundant in
visitors from Iasi in April, for instance, hotel marketers can increase promotional efforts
in Iasi during the late winter months.
Related the hotel industry the tasks performed by data mining can be grouped into
the following five categories.
• Classification arranges customers into pre-defined segments that allow the size
and structure of market groups to be monitored. Also, predictive models can be built to
classify activities. Classification uses the information contained in sets of predictor
variables, such as demographic and lifestyle data, to assign customers to segments.
• Clustering groups customers based on domain knowledge and the database, but
does not rely on predetermined group definitions. This function is beneficial because it
aids hoteliers in understanding who are their customers. For example, clustering may
reveal a subgroup within a predetermined segment with homogenous purchasing
behavior (a subgroup of holiday shoppers within the transient segment) that can be
targeted effectively through a specific ad campaign with the scope that the members of
the subgroup will increase their number of stays or become more loyal. On the other
hand, clustering may indicate that previously determined segments are not
parsimonious and should be consolidated to increase advertising efficiency. Information
such as demographic characteristics, lifestyle descriptors, and actual product purchases
3266
are typically used in clustering.
• Deviation detection uncovers data anomalies, such as a sudden increase in
purchases by a customer. Information of this type can prove useful if a hotel corporation
wants to thank a guest for her or his recent increase in spending or offer a promotion in
appreciation. Marketing managers may also attempt to draw correlations between
surges in deviations with uncontrollable business-environment factors that are not
represented in the database .
• Association entails the detection of connections between records, driven by
association and sequence discovery. For example, a possible detected association may
be that a particular segment’s average length of stay increases after a specific
advertising campaign. Another association task could be employed in an effort to
determine why a specific promotion was successful in one market, but ineffective
elsewhere. Specific information regarding customer-purchase histories is necessary to
formulate probabilistic rules pertaining to subsequent purchases.
• Forecasting predicts the future value of continuous variables based on patterns
and trends within the data. For instance, the forecasting function can be used to predict
the future size of market segments. With forecasting one can also use data trends to
project which hotel amenities are of growing importance to consumers and will be key
drivers of the future perception of value of consumers.
An important task of data mining application for CRM is building appropriate
segmentation and predictive models. For good results it is essential an extensive
knowledge of the hotel business. In [8] there are presented some ways that hotel guests
can be segmented. These refers to demographic aspects like: age, life-cycle stage,
gender or income, or to psychographics aspects like: social class, life-style, personality,
behavior, user status (potential, former, first time) or loyalty status.
The different kind of hotels (transient hotels, convention hotels, extended-stay
hotels) segment guests differently. Furthermore, guest segmentation is distinctive for
most hotel properties. For instance, Best Western or Holiday Inn property-management
systems segment and code markets at the property level, since each location has its own
particular segments. A given property may serve a set of clients, a group of government
clients, and social clients (weddings or reunions).
The segment categories discussed above can be linked into a large set of
combinations. Furthermore, a guest could potentially fit into several categories, which
poses a challenge for current data-mining techniques.
Once a data-mining model is built, it must be tested to assess its predictive
accuracy. For instance, a model designed to predict who will respond to a promotion
should be based on a prior offering in which it is known who did or did not respond.
After the model is constructed, a sample group from a previous promotion can be
analyzed to verify reliability. If the sample predictions do not replicate the results of the
past promotion, then the model may not be significantly predictive. To further enhance
accuracy, a score can be assigned to the model based on the level of agreement between
the sample group and the entire group.
Accurate data collection is critical for successful data mining. Data problems lead
to a decrease in the value of any data warehouse, in addition to decreasing the value of
proposed models.
The first possible problem involves missing or inaccurate data. For example, when
occupation information is available for only 10 percent of a data set, it is difficult to
create a profile of customer occupations. Then again, it is a problem if the data file
3267
contains occupation information for 90 percent of the population, but the accuracy of
the information is poor. Hotel companies can reduce inaccuracy of this kind by asking
guests for their current occupation.
A second problem is poorly coded data. Databases must have standards regarding
data formats, text case, and redundant codes. Problems can occur when data-input
sources are added over an extended time and no one has ensured that the data entering
the warehouse is properly formatted. This would occur, if, when original data-mining
technology was installed, predictions were made based on the reservations system and
the property-management system, but then a subsequent decision was made to input
data from guest-satisfaction surveys. Problems occurs also when additional data inputs
are not standard or are coded improperly. For example, some models require continuous
and ordinal data, while others demand categorical data fields or binary constructs.
Due to the fact that numerous analytical tools can be employed to transform data
into useful information it is very important to select the appropriate tools for analysis
and prediction. Each methodology has strengths and weaknesses, and each is
appropriate for a specific scenario.
It is obvious that technology must serve the purposes of managers, rather than dic-
tate processes. Along that line, data mining cannot capture all the information relating
to what drives consumer behavior. Data mining is simply one of a number of research
methods that help predict the demand trends of travelers. Therefore, data-mining tech-
nology should be used in conjunction with other forecasting and research techniques.
Data mining is a useful tool, but managers should be aware of the following
limitations of data-mining technology:
• Data mining analyzes only data collected from existing customers. Data-mining
software generates information by analyzing data patterns derived from the company’s
reservation, property-management, and guest-loyalty program systems. Patterns thus
detected can help predict the actions of current guests in the system and of those with
similar needs and wants. Data-mining technology does not, however, provide
information about market segments not found in the company’s databases. Moreover, a
market segment that is currently small but is on the limit of experiencing substantial
growth may not be detected by data mining.
• Databases used in the mining process are often hotel-brand specific. Just as
data mining cannot analyze the markets of competitors, it also creates prediction models
that are brand specific. Thus, companies that operate multiple brands often must create
a data warehouse and conduct data mining for each brand. This is also true for the
franchisees that may have a portfolio comprising, say, six Holiday Inns and four Best
Western.
• Data mining may not segment travelers by psychographics traits. Segmenting
consumers based on psychographics traits, such as personality and lifestyle, can be
useful in the hotel industry. This is because psychology and emotion play significant
roles in the decision process of hotel guests . That is, a traveler may select a destination
for a variety of psychological reasons (education, escape, relaxation, social interaction).
One limitation of data mining is that common system inputs do not account for
psychological factors that influence the purchase decision of a traveler.
5. Conclusion
Data-mining technology can be a useful tool for hotel corporations that want to
understand and predict guest behavior. Based on information derived from data mining,
3268
hotels can make well informed marketing decisions—including who should be con-
tacted, to whom to offer incentives (or not), and what type of relationship to establish.
REFERENCES
1. Berry, M., Linoff, G., (1997) Data Mining Techniques for Marketing, Sales and
Customer Support, John Whiley &Sons;
2. Danubianu M. (2003) Determinarea metodei optime de explorare a datelor
pentru un sistem inteligent de evaluare a activit??ii în turism, volumul „Tehnologii
Informa?ionale”, Suceava, pag.156-163, ISBN 793-666-059-1;
3. Danubianu M.(2005) Determinarea claselor pe baza tehnicii SBA, volumul
seminarului “Procesare distribuit?”, pag. 86-90, ISBN 973-666-177-6;
4. Danubianu M. (2006) Using data mining techniques for decision support
systems –Proceedings of the International Conference on Signal/Image Processing and
Pattern Recognition -"UkrObraz-2006",pag. 19-22, August 28-31, 2006, Kiev, Ucraina;
5. Fleisher, C. S., & Blenkhom, D. (2003) Controversies in competitive
intelligence: The enduring issues. Westport, CT: Praeger;
6. IDC & Cap Gemini. Four elements of customer relationship management. Cap
Gemini White Paper;
7. Kotler, P., Bowen, J., and Makens, J., (1999)Marketing for Hospitality and
Tourism, second edition ,Upper Saddle River, NJ: Prentice-Hall;
8. Olsen, M., Connolly, D., (1999) Antecedents of Technological Change in the
Hospitality Industry, Tourism Analysis, Vol. 4, p. 29;
9. Pentiuc., Ghe-St, MORARIU, N. Morariu, M, Pentiuc. L (2001) Intelligent
System For Impact Prognosis Of The Economic Decisions At District Level, Advances
in Electrical and Computer Engineering ISSN 1582-7445 - Volume 1(18), Number
1(15), 2001;
10. Zhao, L. J., & Zhu, D. (2003). Workflow resource selection from UDDI
repositories with mobile agents. Proceedings of Web2003, USA;
11. Zhu, D., Premkumar, G., Zhang, X., & Chu, C. (2001). Data mining for network
intrusion detection: A comparison of alternative methods. Decision Sciences, 32(4),
635-660;
12. Wirth, R. and Hipp, ( 2000) J. CRISP-DM: Towards a standard process model
for data mining. In Proceedings of the 4th International Conference on the Practical
Applications of Knowledge Discovery and Data Mining, pages 29-39, Manchester, UK.
doc_193659959.pdf
It’s a fact that a successful company not only put customers first, but put customers at the center of the organization because the changes in customer behavior determines unpredictable profitability and may be the cause for inefficient marketing planning.
3261
IMPROVING CUSTOMER RELATIONSHIP MANAGEMENT
IN HOTEL INDUSTRY BY DATA MINING TECHNIQUES
MIRELA DANUBIANU, VALENTIN CRISTIAN HAPENCIUC
Mirela DANUBIANU, Lecturer Ph. D. Eng, Ec.
“Stefan cel Mare” University of Suceava
Valentin Cristian HAPENCIUC, Associate Professor, Ph.D. Ec.
“Stefan cel Mare” University of Suceava
Keywords CRM, data mining hotel industry
1. Introduction
It’s a fact that a successful company not only put customers first, but put customers
at the center of the organization because the changes in customer behavior determines
unpredictable profitability and may be the cause for inefficient marketing planning.
The main goal of CRM is the capability to handle customer interaction across dif-
ferent channels and functions, for building loyal and profitable customer relationships.
Although cost cutting and competitive pricing strategies may attract customers
from competitors, in many services industries price advantages are not a sufficient
reason for customers moving between suppliers. In these situations successful
competitive strategies include developing strong relationships with customers and
cross-selling them other services.
Data mining - techniques for exploration and analysis of large quantities of data in
order to discover meaningful patterns and rules - helps businesses sift through layers of
seemingly unrelated data for meaningful relationships, where they can anticipate, rather
than simply react to, customer needs.
2. An overview of CRM
Customer Relationship Management (CRM) is an enterprise customer-centric
approach that uses different techniques to understand and influence consumer behavior.
It is a process which has two objectives:
• to impact all aspects to the consumer relationship (improve customer
satisfaction, enhance customer loyalty or increase profitability)
• to ensure that employees within an organization are using CRM tools. The need
for greater profitability requires an organization to proactively pursue its relationships
with customers [5]
In the real world, acquiring, building, and retaining customers are becoming top
priorities. For many companies, the quality of their customer relationships provides
their competitive edge over other businesses. In addition, the definition of customer has
been expanded to include immediate consumers, partners and resellers - in other words,
everyone who participates, provides information, or requires services from the
company.
Companies are beginning to realize that surviving an intensively competitive and
global marketplace requires closer relationships with customers.
In turn, enhanced customer relationships can boost profitability in three ways:
• by attracting more suitable customers,
• by generating profits through cross-selling and up-selling activities, and
3262
• by extending profits through customer retention.
Generally CRM may be defined by a framework composed by four elements:
know, target, sell and service[6].
It requires the company to know and understand its markets and customers. This
involves detailed customer informations in order to select the most profitable customers
and identify those no longer worth targeting. CRM also entails development of the
offer: which products to sell to which customers and through which channel. In selling,
firms use campaign management to increase the effectiveness of the marketing
departments. Finally, it seeks to retain its customers through services such as call
centers and help desks.
CRM is an association of several components. Before the beginning of the process,
the company must have customer information. These informations may proceed from
internal sources (summary tables that describe customers, customer surveys or
behavioral data contained in transactions systems) or the data can be purchased from
outside sources.
A critical component of a successful CRM strategy is an enterprise data warehouse.
Then, it must analyze the data using statistical tools, OLAP, and data mining. The last
component is campaign execution and tracking.
3. Data mining
Data mining is the exploration and analysis, by automatic or semiautomatic means,
of large quantities of data in order to discover meaningful patterns and rules[3]. So, data
mining is defined as the process of extracting interesting and previously unknown
information from data, and it is widely accepted to be a single phase in a complex
process known as Knowledge Discovery in Databases (KDD).
This process consist of a sequence of the following steps [12]:
• after analysing the goals of the end user and receiving all necessary prior
knowledge, one selects a target data set. This means focusing on a subset of variables
or on data samples.
• the target data are preprocessed and cleaned in order to remove noise or
outliers. One also has to decide how to handle missing data fields.
• useful features have to be found to represent the data, depending on the goal of
the discovery task. The dimensionality is reduced, i.e. one has to find the effective
number of variables under consideration, or invariant representations for the data.
• the primary goals of the knowledge discovery process are predicting the future
values of interesting variables or finding human-interpretable patterns in data.
According to this goal an appropriate data mining algorithm is chosen and applied.
There are algorithms for association, classification, clustering, sequence-based analysis,
and other tasks.
• the patterns are interpreted and evaluated, for example with the aid of
visualisation tools.
After each step, one can return to any other step prior to the current step. Thus, the
knowledge discovery process may contain many loops between any two steps.
In order to ensure that the extracted information generated by the data mining
algorithms is useful, additional activities are required, like incorporating appropriate
prior knowledge and proper interpretation of the data mining results.
In general, CRM promises higher returns on investments for businesses by
enhancing customer-oriented processes such as sales, marketing, and customer service.
3263
Data mining helps companies build personal and profitable customer relationships
by identifying and anticipating the needs of customers throughout the customer
lifecycle.
Data mining can help to reduce information overload and improve decision
making. This is achieved by deriving and refining useful knowledge through a process
of searching for relationships and patterns from the extensive data collected by
organizations. The extracted information is used to predict, classify, model, and
summarize the data.
Data mining technologies, such as rule induction, neural networks, genetic
algorithms, fuzzy logic, and rough sets, are used for classification and pattern
recognition in many industries [4][10][11].
By example, a supermarket organizes its merchandise stock based on purchase
patterns of shoppers, an airline reservation system uses travel patterns of customers and
trends to increase seat utilization., or the web pages alter their organizational structure
or visual appearance based on information about the person who is requesting the
pages.
Data mining builds models of customer behavior by using statistical and machine-
learning techniques. The basic objective is to construct a model for one situation in
which the answer or output is known and then apply that model to another situation in
which the answer or output is desired. The best applications of the above techniques are
integrated with data warehouses and other interactive, flexible business analysis tools.
Hence, data-mining applications can help companies to identify market segments
containing customers with high profit potential, by searching for patterns among the
different variables that serve as effective predictors of purchasing behaviors.
Marketers can then design and implement campaigns that will enhance the buying
decisions of a targeted segment. To facilitate this activity, marketers feed the data-
mining outputs into campaign management software that focuses on the defined market
segments.
Regarding the three ways of boosting profitability discussed in the above section,
the data mining techniques may be used as following:
• for attracting more suitable customers: Data mining can help firms understand
which customers are most likely to purchase specific products and services, thus
enabling businesses to develop targeted marketing programs for higher response rates
and better returns on investment.
• for better cross-selling and up-selling: Businesses can increase their value
proposition by offering additional products and services that are actually desired by
customers, thereby raising satisfaction levels and reinforcing purchasing habits.
• for better retention: Data-mining techniques can identify which customers are
more likely to defect and why. A company can use this information to generate ideas
that allow them to maintain these customers.
Moreover, there are additional ways in which data mining supports CRM
initiatives.
• Database marketing: Data mining helps database marketers develop campaigns
that are closer to the targeted needs, desires, and attitudes of their customers. If the
necessary information resides in a database, data mining can model a wide range of
customer activities. The key objective is to identify patterns that are relevant to current
business problems. For example, data mining can help answer questions such as
“Which customers are most likely to acquire a certain tourist’s service?” Answering
3264
these types of questions can boost customer retention and campaign response rates,
which ultimately increases sales and returns on investment.
Table 1.
Possible association between data mining techniques and CRM operations
Data mining technique CRM operation
Association rules
Information from customer-purchase histories is used to
formulate probabilistic rules for subsequent purchases.
Decision trees
Automatically constructed from data, these yield a sequence
of step-wise rules; good for identifying important predictor
variables, non-linear relationships, and interactions among
variables.
Descriptive statistics
Averages, variation, counts, percentages, crosstabs, simple
correlation; used at the beginning of the data-mining process
to depict structure and identify potential problems in data.
Genetic algorithms
Use procedures modeled on evolutionary biology to solve
prediction and classification problems or develop sets of
decision rules.
Neural networks
Applications that mimic the processes of the human brain;
capable of learning from examples (large training sets of
data) to discover patterns in data.
Query tools
Provide summary measures such as counts, totals, and
averages.
Regression-type models
Ordinary least-squares regression, logistic regression,
discriminant analysis; used mostly for confirmation of models
built by “machine-learning” techniques.
Visualization tools
Histograms, box plots, scatter diagrams; useful for
condensing large amounts of data into a concise,
comprehensible picture
• Customer acquisition: The growth strategy of businesses depends heavily on
acquiring new customers, which may require finding people who have been unaware of
various products and services, who have just entered specific product categories (for
example, new parents and the diaper category), or who have purchased from compe-
titors. Although experienced marketers often can select the right set of demographic
criteria, the process increases in difficulty with the volume, pattern complexity, and
granularity of customer data. Highlighting the challenges of customer segmentation has
resulted in an explosive growth in consumer databases. Data mining offers multiple
segmentation solutions that could increase the response rate for a customer acquisition
campaign. Marketers need to use creativity and experience to tailor new and interesting
offers for customers identified through data-mining operations.
• Campaign optimization: Many marketing organizations have a variety of
methods to interact with current and prospective customers. The process of optimizing a
marketing campaign establishes a mapping between the organization set of offers and a
given set of customers that satisfies the characteristics and constraints of a campaign,
defines the marketing channels to be used, and specifies the relevant time parameters.
Data mining can elevate the effectiveness of campaign optimization processes by
modeling the channel- specific responses of customers to marketing offers.
3265
The different data mining techniques may be associated with CRM tasks. Table 1
present some of these associations.
4. Data Mining Application for the Hotel Industry
Information technology was initially viewed by the hotel industry as a back-office
function that supports the finance and accounting areas. The industry has advanced far
beyond this view during the past decade.
In two sessions sponsored by the International Hotel and Restaurant Association
(IH&RA), one in Singapore in 1997 and the second in Nice, France, in 1998, hotel-
industry leaders pondered the role of technology. Among the conclusions reached were:
“Going forward, technology will be the most competitive weapon for any hospitality
company.
If hospitality organizations want to compete successfully, they must do so by using
technology to drive value to both the customer and to the firm.”[7]
In the hotel industry knowing the guests - where they are from, how much they
spend, and when and on what they spend it- can help a company to formulate marketing
strategies and maximize profits. Due to technological development hotel companies
have accumulated large amounts of customer data, which can be organized and
integrated in databases that can be used to guide marketing decision.
Because identifying important variables and relationships located in these
consumer-information systems can be a difficult task, some hotel companies have
attempted to raise the power of information by using data mining technologies that
exploits the data regarding the consumer.
Such data-mining technology allows hotel companies to predict consumer-behavior
trends, which are potentially useful for marketing applications. For example, Best
Western marketing staff can run reports and analysis on customer and occupancy data
stored in a data warehouse that combines customer and transaction information from all
company properties.
Such information indicates where live the customers who visit a specific hotel. If
the data reveal that the Best Western in Gura Humorului experiences a abundant in
visitors from Iasi in April, for instance, hotel marketers can increase promotional efforts
in Iasi during the late winter months.
Related the hotel industry the tasks performed by data mining can be grouped into
the following five categories.
• Classification arranges customers into pre-defined segments that allow the size
and structure of market groups to be monitored. Also, predictive models can be built to
classify activities. Classification uses the information contained in sets of predictor
variables, such as demographic and lifestyle data, to assign customers to segments.
• Clustering groups customers based on domain knowledge and the database, but
does not rely on predetermined group definitions. This function is beneficial because it
aids hoteliers in understanding who are their customers. For example, clustering may
reveal a subgroup within a predetermined segment with homogenous purchasing
behavior (a subgroup of holiday shoppers within the transient segment) that can be
targeted effectively through a specific ad campaign with the scope that the members of
the subgroup will increase their number of stays or become more loyal. On the other
hand, clustering may indicate that previously determined segments are not
parsimonious and should be consolidated to increase advertising efficiency. Information
such as demographic characteristics, lifestyle descriptors, and actual product purchases
3266
are typically used in clustering.
• Deviation detection uncovers data anomalies, such as a sudden increase in
purchases by a customer. Information of this type can prove useful if a hotel corporation
wants to thank a guest for her or his recent increase in spending or offer a promotion in
appreciation. Marketing managers may also attempt to draw correlations between
surges in deviations with uncontrollable business-environment factors that are not
represented in the database .
• Association entails the detection of connections between records, driven by
association and sequence discovery. For example, a possible detected association may
be that a particular segment’s average length of stay increases after a specific
advertising campaign. Another association task could be employed in an effort to
determine why a specific promotion was successful in one market, but ineffective
elsewhere. Specific information regarding customer-purchase histories is necessary to
formulate probabilistic rules pertaining to subsequent purchases.
• Forecasting predicts the future value of continuous variables based on patterns
and trends within the data. For instance, the forecasting function can be used to predict
the future size of market segments. With forecasting one can also use data trends to
project which hotel amenities are of growing importance to consumers and will be key
drivers of the future perception of value of consumers.
An important task of data mining application for CRM is building appropriate
segmentation and predictive models. For good results it is essential an extensive
knowledge of the hotel business. In [8] there are presented some ways that hotel guests
can be segmented. These refers to demographic aspects like: age, life-cycle stage,
gender or income, or to psychographics aspects like: social class, life-style, personality,
behavior, user status (potential, former, first time) or loyalty status.
The different kind of hotels (transient hotels, convention hotels, extended-stay
hotels) segment guests differently. Furthermore, guest segmentation is distinctive for
most hotel properties. For instance, Best Western or Holiday Inn property-management
systems segment and code markets at the property level, since each location has its own
particular segments. A given property may serve a set of clients, a group of government
clients, and social clients (weddings or reunions).
The segment categories discussed above can be linked into a large set of
combinations. Furthermore, a guest could potentially fit into several categories, which
poses a challenge for current data-mining techniques.
Once a data-mining model is built, it must be tested to assess its predictive
accuracy. For instance, a model designed to predict who will respond to a promotion
should be based on a prior offering in which it is known who did or did not respond.
After the model is constructed, a sample group from a previous promotion can be
analyzed to verify reliability. If the sample predictions do not replicate the results of the
past promotion, then the model may not be significantly predictive. To further enhance
accuracy, a score can be assigned to the model based on the level of agreement between
the sample group and the entire group.
Accurate data collection is critical for successful data mining. Data problems lead
to a decrease in the value of any data warehouse, in addition to decreasing the value of
proposed models.
The first possible problem involves missing or inaccurate data. For example, when
occupation information is available for only 10 percent of a data set, it is difficult to
create a profile of customer occupations. Then again, it is a problem if the data file
3267
contains occupation information for 90 percent of the population, but the accuracy of
the information is poor. Hotel companies can reduce inaccuracy of this kind by asking
guests for their current occupation.
A second problem is poorly coded data. Databases must have standards regarding
data formats, text case, and redundant codes. Problems can occur when data-input
sources are added over an extended time and no one has ensured that the data entering
the warehouse is properly formatted. This would occur, if, when original data-mining
technology was installed, predictions were made based on the reservations system and
the property-management system, but then a subsequent decision was made to input
data from guest-satisfaction surveys. Problems occurs also when additional data inputs
are not standard or are coded improperly. For example, some models require continuous
and ordinal data, while others demand categorical data fields or binary constructs.
Due to the fact that numerous analytical tools can be employed to transform data
into useful information it is very important to select the appropriate tools for analysis
and prediction. Each methodology has strengths and weaknesses, and each is
appropriate for a specific scenario.
It is obvious that technology must serve the purposes of managers, rather than dic-
tate processes. Along that line, data mining cannot capture all the information relating
to what drives consumer behavior. Data mining is simply one of a number of research
methods that help predict the demand trends of travelers. Therefore, data-mining tech-
nology should be used in conjunction with other forecasting and research techniques.
Data mining is a useful tool, but managers should be aware of the following
limitations of data-mining technology:
• Data mining analyzes only data collected from existing customers. Data-mining
software generates information by analyzing data patterns derived from the company’s
reservation, property-management, and guest-loyalty program systems. Patterns thus
detected can help predict the actions of current guests in the system and of those with
similar needs and wants. Data-mining technology does not, however, provide
information about market segments not found in the company’s databases. Moreover, a
market segment that is currently small but is on the limit of experiencing substantial
growth may not be detected by data mining.
• Databases used in the mining process are often hotel-brand specific. Just as
data mining cannot analyze the markets of competitors, it also creates prediction models
that are brand specific. Thus, companies that operate multiple brands often must create
a data warehouse and conduct data mining for each brand. This is also true for the
franchisees that may have a portfolio comprising, say, six Holiday Inns and four Best
Western.
• Data mining may not segment travelers by psychographics traits. Segmenting
consumers based on psychographics traits, such as personality and lifestyle, can be
useful in the hotel industry. This is because psychology and emotion play significant
roles in the decision process of hotel guests . That is, a traveler may select a destination
for a variety of psychological reasons (education, escape, relaxation, social interaction).
One limitation of data mining is that common system inputs do not account for
psychological factors that influence the purchase decision of a traveler.
5. Conclusion
Data-mining technology can be a useful tool for hotel corporations that want to
understand and predict guest behavior. Based on information derived from data mining,
3268
hotels can make well informed marketing decisions—including who should be con-
tacted, to whom to offer incentives (or not), and what type of relationship to establish.
REFERENCES
1. Berry, M., Linoff, G., (1997) Data Mining Techniques for Marketing, Sales and
Customer Support, John Whiley &Sons;
2. Danubianu M. (2003) Determinarea metodei optime de explorare a datelor
pentru un sistem inteligent de evaluare a activit??ii în turism, volumul „Tehnologii
Informa?ionale”, Suceava, pag.156-163, ISBN 793-666-059-1;
3. Danubianu M.(2005) Determinarea claselor pe baza tehnicii SBA, volumul
seminarului “Procesare distribuit?”, pag. 86-90, ISBN 973-666-177-6;
4. Danubianu M. (2006) Using data mining techniques for decision support
systems –Proceedings of the International Conference on Signal/Image Processing and
Pattern Recognition -"UkrObraz-2006",pag. 19-22, August 28-31, 2006, Kiev, Ucraina;
5. Fleisher, C. S., & Blenkhom, D. (2003) Controversies in competitive
intelligence: The enduring issues. Westport, CT: Praeger;
6. IDC & Cap Gemini. Four elements of customer relationship management. Cap
Gemini White Paper;
7. Kotler, P., Bowen, J., and Makens, J., (1999)Marketing for Hospitality and
Tourism, second edition ,Upper Saddle River, NJ: Prentice-Hall;
8. Olsen, M., Connolly, D., (1999) Antecedents of Technological Change in the
Hospitality Industry, Tourism Analysis, Vol. 4, p. 29;
9. Pentiuc., Ghe-St, MORARIU, N. Morariu, M, Pentiuc. L (2001) Intelligent
System For Impact Prognosis Of The Economic Decisions At District Level, Advances
in Electrical and Computer Engineering ISSN 1582-7445 - Volume 1(18), Number
1(15), 2001;
10. Zhao, L. J., & Zhu, D. (2003). Workflow resource selection from UDDI
repositories with mobile agents. Proceedings of Web2003, USA;
11. Zhu, D., Premkumar, G., Zhang, X., & Chu, C. (2001). Data mining for network
intrusion detection: A comparison of alternative methods. Decision Sciences, 32(4),
635-660;
12. Wirth, R. and Hipp, ( 2000) J. CRISP-DM: Towards a standard process model
for data mining. In Proceedings of the 4th International Conference on the Practical
Applications of Knowledge Discovery and Data Mining, pages 29-39, Manchester, UK.
doc_193659959.pdf