ArticlePDF Available

What Is Personalization? Perspectives on the Design and Implementation of Personalization in Information Systems


Abstract and Figures

In e-commerce and mobile commerce, personalization has been recognized as an important approach element in customer relationships and Web strategies. However, there are wide differences in how this concept is defined, characterized, and implemented in the literature. In this article we present a high-level framework for classifying approaches to personalization that delineates fundamental assumptions about personalization in the literature and relates them to strategies for developing personalization systems. The framework consists of 2 parts: (a) a set of perspectives on personalization that guide the design of personalization systems at a general level and (b) a scheme for classifying how personalization can be implemented. The personalization perspectives represent 4 distinct schools of thought on the nature of personalization distilled from the literature of several fields. These perspectives are ideal types and we discuss them in terms of the motivation they supply for personalization, the goals and means of personalization, and the ways in which they conceptualize and model users. The implementation classification scheme is constructed on 3 dimensions of implementation choices. These 3 dimensions pertain to what to personalize (content, interface, functionality, channel), to whom to personalize (individuals or categories of individuals) as well as who does the personalization (implicit or explicit personalization). The personalization perspectives represent particular concepts of personalization that guide general design choices; these choices are implemented via the options described in the implementation classification scheme. The framework contributes to the development of a common theoretical basis for the study of personalization. We discuss implications of the framework for design of personalization systems and future research directions.
Content may be subject to copyright.
This article was downloaded by:
[University of Queensland]
12 April 2011
Access details:
Access Details: [subscription number 907688104]
Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-
41 Mortimer Street, London W1T 3JH, UK
Journal of Organizational Computing and Electronic Commerce
Publication details, including instructions for authors and subscription information:
What Is Personalization? Perspectives on the Design and Implementation
of Personalization in Information Systems
Haiyan Fan; Marshall Scott Poole
Online publication date: 18 November 2009
To cite this Article Fan, Haiyan and Poole, Marshall Scott(2006) 'What Is Personalization? Perspectives on the Design and
Implementation of Personalization in Information Systems', Journal of Organizational Computing and Electronic
Commerce, 16: 3, 179 — 202
To link to this Article: DOI: 10.1207/s15327744joce1603&4_2
Full terms and conditions of use:
This article may be used for research, teaching and private study purposes. Any substantial or
systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or
distribution in any form to anyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation that the contents
will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses
should be independently verified with primary sources. The publisher shall not be liable for any loss,
actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly
or indirectly in connection with or arising out of the use of this material.
What Is Personalization? Perspectives
on the Design and Implementation
of Personalization in Information Systems
Haiyan Fan
Marshall Scott Poole
Department of Information and Operations Management
Mays School of Business
Texas A&M University, College Station
In e-commerce and mobile commerce, personalization has been recognized as an im
portant approach element in customer relationships and Web strategies. However,
there are wide differences in how this concept is defined, characterized, and imple-
mented in the literature. In this article we present a high-level framework for classify-
ing approaches to personalization that delineates fundamental assumptions about
personalization inthe literature and relates them to strategies for developing personal-
ization systems. The framework consists of 2 parts: (a) a set of perspectives on person-
alization that guide the design of personalization systems at a general level and (b) a
scheme for classifying how personalization can be implemented. The personalization
perspectives represent 4 distinct schools of thought on the nature of personalization
distilled from the literature of several fields. These perspectives are ideal types and we
discuss them in terms of the motivation they supply for personalization, the goals and
means of personalization, and the ways in which they conceptualize and model users.
The implementation classification scheme is constructed on 3 dimensions of imple
mentation choices. These 3 dimensions pertain to what to personalize (content, inter
face, functionality, channel), to whom to personalize (individuals or categories of indi
viduals) as well as who does the personalization (implicit or explicit personalization).
The personalization perspectives represent particular concepts of personalization that
guide general design choices; these choices are implemented via the options described
in the implementation classification scheme. The framework contributes to the devel
opment of a common theoretical basis for the study of personalization. We discuss im
plications of the framework for design of personalization systems and future research
personalization, multiparadigm review, classification scheme, ideal types,
design philosophy, electronic commerce, mobile commerce
AND ELECTRONIC COMMERCE 16(3&4), 179–202 (2006)
Correspondence should be sent to Haiyan Fan, Department of Information and Operations Manage
ment, Mays School of Business, Texas A&M University, College Station, TX 77843–4217. Email: hfan@
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
The impulse to personalize environments, tools, and products to fit the unique con
cerns of the individual is as old as human society. In this era of technological inno
vations, the Internet, and new media, personalization is possible on a broader scale
and can be done more quickly and effectively than ever before. As an important so
cial phenomenon that carries great economic value [1, 2], personalization has
drawn increasing research attention from both academia and industry. Personal
ization has been studied in such academic fields as economics, management, mar
keting, information systems (IS), and computer science. In industry, corporate
spending on content personalization is estimated at $6 billion by 2004 [3], and per
sonalization technology providers have mushroomed (e.g., Net Perceptions,
BroadVision, Documentum, Vignette).
However, there is little consensus on how best to characterize the personaliza
tion construct. There is considerable diversity in thinking about the concept across
the various disciplines and researchers who have studied personalization. Such di
versity is advantageous because it offers multiple creative viewpoints on an impor
tant phenomenon. However, the wide range of viewpoints has tended to hinder
accumulation of a foundational body of research on personalization. Most current
research on this topic has centered on the technical level in which the conceptual-
ization of personalization systems depends on the developer or researcher’s partic-
ular view of personalization. This has resulted in studies and systems that are
difficult to relate to one another. Furthermore, empirical studies that have com-
pared and contrasted the effectiveness of different personalization technologies are
rare. The current practice of focusing on “how to do personalization” rather than
“how can personalization be done well” suggests that the field is still in its infancy.
This situation motivated us to develop a high-level framework for classifying
approaches to personalization. The framework delineates fundamental assump-
tions about personalization in the literature and relates them to design strategies
for developing personalization systems. It consists of two parts: (a) a set of perspec
tives on personalization that guide the design of personalization systems at a gen
eral level and (b) a scheme for classifying how personalization can be
implemented. The personalization perspectives represent four distinct schools of
thought on the nature of personalization distilled from the literature of several
fields. These perspectives are ideal types and are discussed in terms of the motiva
tion they supply for personalization, the goals and means of personalization, and
the ways in which they conceptualize and model users. The implementation classi
fication scheme is constructed on three dimensions of implementation choices.
These three dimensions pertain to what to personalize (content, interface, function
ality, channel), to whom to personalize (individual or categories of individuals) as
well as who does the personalization (implicit or explicit personalization). The per
sonalization perspectives represent particular concepts of personalization that
guide general design choices; these choices are implemented via the options de
scribed in the implementation classification scheme. The classification scheme of
fers a descriptive analysis of personalization practices, whereas the
personalization perspectives offer a normative analysis of design possibilities.
Whereas the implementation classifications shed light on the question of how per
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
sonalization can be done, the personalization perspectives provide insight into
possibilities—diverse ways of thinking about personalization and the design of
personalization systems.
We organized this article as follows. We first consider definitions of personaliza
tion ventured by scholars and designers from a range of fields to illustrate some of
the complexities behind the construct. Based on the literature review, we advance a
working definition for personalization that presents a broad view of the phenome
non. Next, we present the classification scheme for methods of implementing per
sonalization. We introduce this prior to the perspectives on personalization
because it flows directly from our analysis of the definitions and also delimits the
means by which personalization is conducted. This provides important context for
the perspectives, which are at a more general level and shape the form and content
of the implementation. Finally, we discuss the implications for design of personal
ization systems as well as for IS research and practice.
We conducted a literature review in electronic databases using the keywords per-
sonalization, variants of the same word stem, and related terms such as customiz-
ation, adaptation, individuation, consumer-centric, and one-to-one relationship. The ini-
tial filtering of over 300 abstracts and book summaries yielded a total of 142
references thatdiscussed orstudiedpersonalization including86 journal articles,35
books or book sections, 13 conference papers, and eight Web references.
These sources represent six general areas in which personalization has been
studied: marketing/e-commerce; computer science/cognitive science; architec-
ture/environmental psychology; information science; and social sciences includ-
ing sociology, anthropology, and communication. Readers may get a sense of the
diversity of concepts in current research from Table 1, which exhibits sample defi
nitions of personalization. This array of definitions reflects the multidisciplinary
nature of personalization research.
At the conceptual level, personalization means different things to different peo
ple in different fields. For architects, personalization means creating functional,
pleasant personal spaces; for social scientists it is a way of enhancing social rela
tionships and building social networks [4, 5]; for some computer scientists, person
alization is a toolbox of technologies to enhance the Web experience through
graphic user interface design. Different conceptualizations in turn dictate different
research methodologies and implementations. Cognitive scientists resort to ex
plicit mental modeling to differentiate users, whereas e-commerce marketers rely
on user profiles and purchase records to segment customers.
The two largest bodies of work in our review were comprised of research by
scholars in the computer science and marketing/e-commerce areas. Each of these
groups has different research agendas and assumptions about personalization. Pri
marily comprised of researchers in marketing, management, economics, and IS,
the marketing/e-commerce group focuses on how to manage customer relation
ships by delivering unique value and benefits to each individual customer. This
group is particularly interested in the use of personalization in Web-enabled com
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
Table 1
Representative Definitions of Personalization
Sample Definitions
Marketing/e-commerce a. “Personalization is the combined use of technology and customer
information to tailor electronic commerce interactions between a
business and each individual customer” [78].
b. “Personalization is about building customer loyalty by building a
meaningful one-to-one relationship; by understanding the needs
of each individual and helping satisfy a goal that efficiently and
knowledgeably addresses each individual’s need in a given
context” ([63], p. 26).
c. “Personalization is the capability to provide users, customers,
partners, and employees, with the most relevant web experience
possible” ([79], p. 15).
d. “Personalization is any behaviors occurring in the interactions
intended to contribute to the individuation of the customer”([80],
p. 87).
e. An enterprise, process, or ideology in which personalized
products and services are integrated and implemented throughout
the organization including all points of sale; other points of
customer contact; and back-end activities and departments such as
inventory, shipping, production, and finance [66].
Cognitive science f. Personalization is “a system that makes explicit assumptions
about users’ goals, interests, preferences and knowledge based on
an observation of his or her behavior or a set of rules relating
behavior to cognitive elements” [81].
g. Personalization is the process of providing relevant content based
on individual user preferences or behavior [12, 66].
h. Personalization is the“explicit user model that represents user
knowledge, goals, interests, and other features that enable the
system to distinguish among different users” ([82], p. 31).
i. Personalization is the understanding of “the user, the user’s tasks,
and the context in which the user accomplishes tasks and goals”
([83], p. 50).
Social science j. Technology that reflects and enhances social relationships and
social networks [4, 5].
k. “Technology that provide experiences that bridge cultures,
languages, currencies, and ideologies” ([77], p. 14).
Computer science l. “Personalization is a toolbox of technologies and application
features used in the design of an end-user experience” ([84], p. 44).
m. “Personalization system is any piece of software that applies
business rules to profiles of users and content to provide a
variable set of user interfaces”[13].
n. Machine-learning algorithms that are integrated into systems to
accommodate individual user’s unique patterns of interactions
with the system [21].
o. “Computer networks that provides personalized features,
services and user interface portability across network boundaries
and between terminals” ([25], p. 128).
p. Unifying platform embedded in any type of computing devices
that support individualized information inflow and outflow [85].
q. Presenting customers with services that are relevant to their
current locations, activities, and surrounding environments [22].
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
merce and mobile commerce. On the other hand, with an interest in making com-
puter technologies more usable by people, computer scientists in the computer-
human interaction (CHI) group view personalization as a way to close the gap be-
tween user and computer. The assumption in this case is that systems that are de-
signed and adapted to user requirements will facilitate user goal attainment. These
two groups hold considerably different understandings about what personaliza-
tion is, and each group has developed its own terminology, methodology, and tool
kits. Moreover, as Table 1 shows, there are differences within each group as well.
Differences in views of personalization among fields and among researchers in the
same field make it difficult to relate personalization studies to one another and to
cumulate knowledge about personalization.
The concept of personalization is intuitive but also slippery. As Table 1 indi
cates, the term has been used in so many ways that it is difficult to discern core fea
tures by just scanning the definitions. A thematic analysis of the definitions in
Table 1 suggests that most definitions of personalization include (a) a purpose or
goal of personalization, (b) what is personalized (interface, content, etc.), and (c)
the target of personalization (user, consumer, etc.). Although most definitions also
include a statement of the means by which personalization is implemented, we be
lieve that to be useful to the many fields involved in personalization research, a def
inition must be neutral as to means of personalization. It is clear that there are
many ways to do personalization whether via knowledge representation, a specific
product, or a Web page, and a general definition should not favor any particular
approach. To this end, we adapt Blom’s [6] general concept and define personaliza
tion as a process that changes the functionality, interface, information access and
content, or distinctiveness of a system to increase its personal relevance to an indi
vidual or a category of individuals.
This definition is framed around the goal of increasing personal relevance to
avoid dependence on particular motivations for personalization that limit the
scope of many of the definitions in Table 1. Definitions built around specific
Table 1 Continued
Sample Definitions
r. Consumer-centric infomediary that act on behalf of users to
perform online shopping, searching and information-gathering
services [23, 24].
environmental psychology
s. “The relationship between persons and the spatial dimensions of
the environment that effects the cognitive, affective and socio-
cultural components of the individual” ([39], p. 142).
t. Deliberate decoration or modification of an environment to reflect
the occupants’ identities by increasing the usability and aesthetic
value of the space [38].
Information science u. Fine-tuning and prioritizing information based on criteria that
include timeliness, importance, and relevance to the audience [86].
v. “Delivering to a group of individuals relevant information that is
retrieved, transformed, and/or deduced from information
sources” ([87], p. 30).
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
goals—such as tailoring electronic commerce transactions, delivery of business
processes, and enhancing social relationships—or for specific targets such as cus
tomers, partners, and employees are too context bound to be useful in understand
ing the entire field. Definitions that emphasize certain techniques or tools of
personalization—such as user models, machine learning algorithms, computer
networks, or infomediaries—predecide how personalization is to be done and also
limit one’s view of the phenomenon. Our definition attempts to provide a more
general view of personalization. In our view, an adequate definition of personal
ization should open up the field of personalization research to many perspectives
and methods and avoid specific assumptions about motives, context, or method.
At this point, it is useful to explore a few terms that are closely related to person
alization and sometimes used synonymously for it. The term customization is fre
quently used interchangeably with personalization. Most work in which
customization is used [7, 8] has referred to it as customer-initiated personalization
actions. Somewhat analogous to ordering from a menu, customization is often
comprised of a suite of template-driven, finite set of options from which users
choose. Examples are personal portal sites such as “My Yahoo!” [9] or sites that of
fer custom-made apparel such as “Lands End” [10]. Because users are in direct con-
trol, customization is advantageous with respect to high predictability and low
intrusiveness. In relation to the definition of personalization just advanced, cus-
tomization would be one approach to implementing personalization.
Another closely related term is adaptation. In the CHI literature, this term has
been used to refer to the properties of a system that can automatically adjust its be-
havior and interaction to suit the user’s needs [11]. Specifically, an adaptive system
employs explicit mental or cognitive modeling of the user to enable the system to
distinguish among different users. Adaptable systems, on the other hand, require the
user to explicitly specify how he or she wants the system to be different. The dis-
tinction between adaptable and adaptive systems is similar to the distinctions be-
tween explicit and implicit personalization [7, 12–14] and static and dynamic
personalization [15] that can be found in descriptions of commercial systems for
personalization (see also Karat et al. [16]). This terminological duplication is proba
bly due to different perspectives. The CHI research is system oriented and focuses
on what the system can do as compared to the application-oriented business re
search that focuses on applying personalization technology in e-commerce or mo
bile commerce. In terms of our definition of personalization, the adaptive or
adaptable system is the means to accomplish personalization. In this article, we
adopt the widely accepted classification of implicit versus explicit personalization
from the e-commerce literature to reflect the adaptive–adaptable distinction.
Personification refers to endowing inanimate computer objects with human qual
ities or human form [6, 17, 18]. In CHI terminology, personification is often imple
mented as anthropomorphic software agents [19, 20]. This term does not relate to
our discussion of personalization.
Beyond terminological issues, diversity in the technologies used in building per
sonalization systems also constitutes a barrier to mutual understanding among
personalization researchers. There exist many approaches to personalization rang
ing from computational algorithms to less rigorous applications of various types.
For example, machine-learning algorithms have been integrated into systems to ac
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
commodate the individual user’s unique patterns of interactions with the system
[21]. Ubiquitous computing and context-aware computing provide customers with
services relevant to their current locations, activities, and surrounding environ
ments [22]. Agent technology has been used in building consumer-centric
infomediaries that act on behalf of users to perform various tasks [23, 24]. Com
puter networks can provide personalized features, services, and user interface por
tability across network boundaries and between terminals [25].
To get a perspective on the variety of approaches to personalization, it is useful
to define general dimensions underlying the implementation of personalization
systems. These dimensions can be used to construct a classification scheme for per
sonalization systems that relates them to theoretical and practical concepts.
The schemeis constructedalongthree dimensionsof implementation implicitin the
previous section: (a) the aspect of the information system that is manipulated to
provide personalization (what is personalized), (b) the target of personalization (to
whom to personalize), and (c) whodoesthepersonalization (i.e., the user or the sys-
tem). This classification scheme draws on several previous classification systems.
Blom [6] distinguished three motivations to personalize: to access information, to
accomplish workgoals,and to accommodate individualdifferences. Rossi et al. [26]
made a distinction betweenbaseinformation and behavior, what the user perceives
and how the user perceives. This framework is largely concerned with system-level
elements such as personalization for links, navigation structure, and navigation
context. Instone [13] and Wu et al. [27] classified personalization on e-commerce
Web sites into a two-by-two grid with implicit versus explicit personalization on
one dimension and Web content versus Web interface on the other dimension.
In terms of the first dimension, what is personalized, we can distinguish four as
pects of IS that can be personalized: the information itself (content), how the infor
mation is presented (user interface), the media through which information is
delivered (channel/information access), and what users can do with the system (func
tionality). These represent the basic elements of IS that can be manipulated in a per
sonalization system to make the system more personally relevant to the user. This
dimension focuses on the particular parts of the system that deliver personaliza
tion to the user.
The second dimension, the target of personalization, can be either a category of
individuals or a specific individual. One option is to implement personalization for
a particular category of user such as women, single-child families, or members of a
club. Insofar as an individual user identifies with this category, he or she is likely to
perceive that the system is personalized for them. Another option is to design sys
tems to adapt and cater to the needs of a single user. Individuated personalization is
targeted to a specific individual, and its goal is to deliver goods, services, or infor
mation unique to each individual as an individual.
Research on social identity [28, 29] has shown that people may think of them
selves either as members of a social group (a category) or as individuals, depend
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
ing on the social cues available in a particular context. Furthermore, research has
indicated that people react differently when they are focused on their unique iden
tity as an individual (individuated) as opposed to how they act if their focus on
their identity as members of a social group (categorized). When people focus on
category membership, their motivation revolves around values and concerns of the
social group; they are more influenced by group norms than by individual consid
erations; they tend to make judgments based on perceived group standards; and
they may stereotype members of outgroups, groups they view as opposed or dif
ferent from their own. When people are individuated, their motivation is largely
driven by their particular individual needs; they are not as strongly influenced by
norms but make decisions on individual bases, and they are more likely to see oth
ers as individuals as well and not as members of other social groups. Personaliza
tion systems based on categories are likely to give categorical cues (e.g., “This site is
specially designed for members of the Blackwell Club”) and are likely to elicit quite
different user reactions than are individuated systems.
Interestingly, the actual implementation of individuated personalization may
be based on categorical analysis. If it is desirable to capture the unique individual
ity of a person, this can be defined as the unique intersection of a variety of catego-
ries representing the individual’s important characteristics (e.g., female, Hispanic,
professional, living in Idaho, 25 years old, one child, etc.) and utilizing enough cat-
egories to define the individual uniquely. Although categories are used, this sys-
tem functions for all intents and purposes as an individuated personalization
system. In general, as this example illustrates, individuated personalization takes
more system resources than categorical personalization.
The third dimension pertains to degree to which personalization is automated.
Personalizationin whichtheuser participatesbymakingchoicesorprovidinginfor-
mation to give the system guidance as to how to adapt is termed explicit personaliza-
tion. Personalization that is done automatically by the system is termed implicit
personalization. As wenotedin the previous section, this distinctionparallels the dif
ferentiation of system-initiated versus user-initiated personalization, adaptive ver
out personalization but also because users are likely to react differently to a system
they knowthey control (explicit personalization)andone that seemstohave a lifeof
has suggested that people react to systems that display agency on the same basis as
they respond to other human beings,whereasasystem that is dependent on human
input—and thus clearly responsive rather than proactive—is more likely to be
viewed as nonhuman. Hence, implicit personalization would be expected to affect
users differently than would explicit personalization.
Together, the three dimensions capture key implementation choices involved in
personalization, yielding the scheme depicted in Table 2. The dimension along the
top of the table, aspects of IS used to personalize, breaks out options pertaining to
technical implementation. The target and automation dimensions, which run along
the side, highlight implementation choices with different implications for user re
developedbyAmoroso andReinig [30].AmorosoandReinig classifiedtechnologies
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
Table 2
Classification Scheme for Personalization Systems With Examples
Automation Content Functionality User Interface
Information Access
Implicit Individuated Recommendation
of new
destination sent
by travel
company in a
greeting card to
customers each
year around the
time they took
their last trip
based on the
analysis on
groups of other
travelers [72]
Mobile wireless
equipped with
systems acting
as a personal
tourist guides
that can
adjust to users’
interests and
changes in
systems that
grouping, or
look and feel
according to
user’s real-time
action [21]
computing in
support such as
flight schedule
change alert
sent to
digital assistant
(PDA) (www.
Categorical Recommender
suggests related
titles to all
customers who
browsed or
bought Science
of the Artificial
company Web
site designed
with different
for users of
different roles:
partners, truck
drivers, and
case study on
3D interactive
changes its
presentation for
groups of
visitors such as
big kids,
romance &
gatherings, etc.
service: call
center, Web,
ATM, wireless
support, etc.
Explicit Individuated Personalized
digital library
(my book bag,
my bookmark);
web portal (My
apparel (Lands’
Customer built
personal virtual
model by
specifications of
(Lands’ End);
cosmetics tool
with which
customer can
with myriads of
configure the
look and feel of
My page saved
on the server
virtual home
(3rd generation
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
used in personalization management systems into four broad categories. User-
behavior tracking technologies include cookies, clickstream tracking, and hover tech-
nologies.Asthename implies,thistypeoftechnology providesmechanismstoiden-
tify users and to monitor user online behavior in the background. Personalization
database technologies are built on large database systems and require intensive com-
puting power. This type of technology includes statistical analysis, data mining,
webhousing, intelligent agent, recommender systems, collaborative filtering, and
user profiling. Personalized user interface technologies include user interface design,
human-oriented digital design, and adaptive hypermedia. Finally, customer support
ing on a user’s current location, activity, and surrounding environment. This
scheme, whichcaptures complex bundlesoftechnology used inpersonalization, of
make the scheme too complex.
Thus far, we have undertaken the descriptive task of structuring research on
personalization by defining the construct and offering a classification of key
choices in the implementation of personalization systems. In the following section,
we present a normative analysis of general perspectives on personalization. Nor
mative perspectives guide design by developing a vision of what personalization
could be that articulates the purpose of personalization and criteria for realizing
that purpose. Current approaches to personalization in both industry and acade
mia tend to adopt relatively narrow, specialized views of the subject. However, we
believe that creative and effective design will best be facilitated by considering
widely different approaches. The commercial and functional approaches that dom
inate the marketing and computer science schools on personalization are only two
of several approaches that can be taken to the design and execution of personaliza
tion. Alternative norms to guide design can be identified through exploring alter
native ways in which personalization is characterized, conceptualized, and
practiced in diversified disciplines.
Table 2 Continued
Automation Content Functionality User Interface
Information Access
Categorical Online interest
groups and
support groups
on various
health topics
such as
pregnancy, and
moderated by
interactive tools
such as quizzes,
games for kids,
parents, and
different look
and feel of the
web site for
Local restaurant
or vending
available on
user’s PDA
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
Different schools of thought can be discerned withinthediversepersonalization lit
erature. To capture the characteristic features of these logically consistent ap
proaches to thinking about personalization, we distilled four perspectives from the
literature onpersonalization. In conducting thismultiparadigm review, we utilized
two metatriangulation techniques discussed by Lewis and Grimes [31] to uncover
paradigmatic disparity and complementarity. First we used paradigm bracketing to
differentiate and articulate various sets of assumptions. In some cases, these as
sumptions underlie prevailing thought about personalization, and we attempt to
make themmoreexplicit. In other cases,these assumptions are moreexplicitbut are
embedded in less well known paradigms, and we argue that they should be consid
ered by personalization researchers [32]. Second, paradigm bridging suggests “tran
sition zones” whereparadigmaticboundaries become fuzzy and new viewsperme
ating across paradigms are synthesized. In this section, we present the results of
paradigm bracketing that enabled us to identify four perspectives on personaliza
tion. At the end of this section and in the following section, we use paradigm bridg
ing to identify commonalities and differences among the perspectives and to ex-
plore design implications.
We relied on Weber’s [33] ideal type theory in defining the personalization
perspectives. Weber argued that social, economic, and historical research can
never be fully inductive or descriptive, as one always approaches it with a con-
ceptual apparatus. This conceptual apparatus Weber defined as the ideal type,
which is an abstraction of essential features of a particular social or economic
phenomenon. The ideal type is useful for studying personalization for two rea-
sons. First, it focuses on the development of internally coherent perspectives on
the subject. It is important to study each of the distinct schools of thought on per-
sonalization in its “pure form” so as to capture their respective central character-
istics. Ideal type theory provides a methodology for analyzing the typical or
logically consistent features of social institutions or behaviors [33]. Second, there
is a strong need to establish a common frame of reference against which the cur
rent practice of personalization can be evaluated. Although the ideal type does
not describe any particular concrete course of action, it does describe what
Weber referred to as “objectively possible” courses of action. The ideal type is an
analytical tool for comparing the extent to which a concrete example of practice
is similar to or different from the defined ideal. In this sense, ideal types can
serve as guidelines for conducting and evaluating personalization systems in
light of alternative approaches.
From our literature review, we distilled four ideal types of personalization as
shown in Table 3: the architectural, relational, instrumental, and commercial per
spectives. Each perspective represents a different philosophy concerning the moti
vation behind personalization and what personalization tries to accomplish (its
goal). Each perspective also implies a different strategy for personalization, differ
ent means for carrying out this strategy, and different user modeling techniques.
Finally, each perspective implies different criteria for evaluating personalization
systems. In discussing the four perspectives, we use Web enabled e-commerce or
mobile commerce sites as focal cases.
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
4.1 Architectural Personalization
Architectural personalization is most generally associated with the fields of archi
tecture, environmental psychology, and urban planning. This approach is reflected
in the definitions under the architecture category in Table 1. Architecture has long
been used asareference discipline for CHI research, primarilyinthe graphic design
and visualization areas [34]. As a recognized reference discipline for management
IS research, the social processes and research methodologies of architecture have
been studied [35, 36].Forexample,Kim et al. [37] applied architectural constructs to
measure the architectural quality of Internet business. In this article, we explore
ways in which architects personalize physical environments that are applicable to
the Web-enabled environment.
We define architectural personalization as the construction of the digital environ
ment to create a pleasant user space and a unique experience for the user through
arrangement and design of digital artifacts in a way that meet the user’s needs and
reflect his or her style and taste. Because architectural personalization is concerned
with building digital environments, it relates particularly to the interface aspect of
the system.
The motive of architectural personalization is to fulfill the user’s needs and to
enable him or her to express himself or herself through design of the online envi
ronment. The goals for personalization in this view are twofold: (a) to create a func
Table 3
Personalization Ideal Types
Architectural Instrumental
Motive: To fulfill a human being’s needs for
expressing himself/herself through the
design of the built environment
Motive: To fulfill a human being’s needs for
efficiency and productivity
Goals: To create a functional and delightful
Web environment that is compatible with a
sense of personal style
Goals: To increase efficiency and productivity
of using the system
Strategy: Individualization Strategy: Utilization
Means: Building a delightful Web
environment and immersive Web
Means: Designing, enabling, and utilizing
useful, usable, user-friendly tools
User model: Cognitive, affective, and social-
cultural aspects of the user
User model: Situated needs of the user
Relational Commercial
Motive: To fulfill a human being’s needs for
socialization and a sense of belonging
Motive: To fulfill a human’s beings needs for
material and psychic welfare
Goals: To create a common, convenient
platform for social interaction that is
compatible with the individual’s desired
level of privacy
Goals: To increase sales and to enhance
customer loyalty
Strategy: Mediation Strategy: Segmentation
Means: Building social interactions and
interpersonal relationships
Means: Differentiating product, service, and
User model: Social context and relational
aspects of the user
User models: User preference or demographic
profiling; user online behavior and user
purchasing history
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
tional and delightful Web environment that provides aesthetic value and reflects
the user’s personal style and (b) to help the user cultivate a sense of personal and
social identity within the space [38].
The general strategy of architectural personalization is individualization. Re
search in architecture has shown that personalized design that incorporates the
needs and requirements of users has significantly improved the quality and func
tion of the built environment [39–43]. Personalization of domestic and work spaces
strives to make them true reflections of the occupant’s personal and social identity,
particularly with respect to social-cultural positions such as ethnic, socio
demographic, socioeconomic, or socioprofessional background [39]. Architects
seek to honor individual experience sui generis, and therefore, architectural per
sonalization is in direct contrast with commercial personalization, which starts and
ends with the premise that personalization must enhance the marketability of the
products and profitability of the business transaction.
Intransferring thearchitectural perspectiveto theWeb-enabled environment,re
searchers are confronted with a question: What constitutes the space? We contend
thatthis spaceisthe digitalcounterpartofanalogspace. Novaketal. [44]arguedthat
“the Internet is best thought of not as a simulation of the ‘real world’, but as an al-
ternative real, yet computer-mediated environment in which the online customer
of physical artifactssuchas buildings, furniture, and otherobjects[45], digital space
iscomprisedof human-madedigital artifactssuchasthe structureof aWeb site,nav-
igation components, hyperlinks, layout, and site flow [46]. Architectural personal-
form, andfunction, and itscentral reference pointis a balancethat is capturedby the
phrase “aesthetic functionality.”
Theories of behavior–environment congruence advance the premise that ma-
nipulating physical space provides an effective means for influencing the cogni-
tive, affective, and social-cultural aspects of residents [47–49]. Personalization
systems designed following the architectural design philosophy employ user mod
els that map the cognitive, affective, and social-cultural aspects of users. Most cur
rent research has explored principles for constructing digital spaces that afford
easy navigation, intelligent presentation, and aesthetic delight [34, 50, 51]. Less re
search has been devoted to how to utilize a user’s individual style and taste as
shaped by his or her individual and social identity. A good example of architec
tural personalization would be the L
ORÉAL® Web site. The site is designed with
a different look and feel for different countries. The Japanese site is presented with
the fresh pure look of oriental lotus, the Brazilian site is imbued with passionate
dashes of red, and the French site is enlivened by an avant-garde-looking model.
The variety brings in intrigue, mood, and added value to a site.
4.2 Instrumental Personalization
In Marxist philosophy, a human being is defined as the creature capable of creating
and using tools. Human history is a history of creation and use of increasingly pow
erful toolsandmachines. From the industrialrevolutionto the information age,me
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
chanical tools followed by electronic and digital machines have emerged at an ex
ponential rate, becoming an indispensable part of daily life. Instrumental
personalization attempts to facilitate human use of computer systems as tools.
Instrumental personalization correlates with the goals of the traditional systems
designer and is exemplified by definitions under the computer science, cognitive
science, and IS categories in Table 1. It refers to the utilization of IS to enhance effi
ciency and personal productivity by providing, enabling, and delivering useful,
usable, user-friendly tools in a way that meet the user’s situated needs. Instrumen
tal personalization focuses on the functionality of the system. The assumption in
this case is that users will find systems that are designed and tailored to their par
ticular requirements more relevant. Regardless of the type or sophistication of the
machines, the purpose for instrumental personalization nevertheless is singular—
to support users in accomplishing their goals. Unlike architectural personalization
in which function and form balance each other, instrumental personalization em
phasizes functionality and usability and treats aesthetics as a secondary consider
ation to be addressed once instrumental standards are met.
There are three aspects of instrumental personalization: providing tools, design
ing tools, and utilizing tools. Each aspect takes a different perspective on the per-
sonalization issue and entails different research interests. Providing tools is
concerned with creating devices for personalized use that can be delivered through
the appropriate channels. Channels for provision of services include the wired and
wireless Webs, personal digital assistants, interactive TV, and voice portals among
others. Devices deployed in wired or wireless applications offer personalized func-
tions ranging from Hallmark’s® interactive calendar that sends reminders of im-
portant dates to personal agents capable of conducting business transactions [24,
52]. Designing tools is concerned with making tools and machines usable, useful,
and user friendly, the traditional domain of software engineers.
Utilizing tools is concerned with choosing the appropriate channels and devices
to deliver relevant content effectively. The challenge lies in identifying the proper
vehicle to carry out the service through multiple channels. For example, ubiquity,
localization, and convenience have been often cited as key mobile value proposi
tions [53, 54]. Mobile wireless agents equipped with Global Positioning Systems
are suitable for personal tourist guides that can dynamically adjust to users’ inter
ests and changes in environment (e.g., indicating when museums are open during
times convenient for the user) [55]. Web-based shopping agents are capable of per
forming complicated price, utility, and functionality comparison among brands
[56]. The challenge lies in identifying the proper vehicle to carry out the service
through the “multi-channel zigzag” [57]. An important task for instrumental per
sonalization is the integration of different computing devices across platforms.
Truly personal control over the flow of information across the boundaries of net
works, platforms, and devices can be realized through the creation of personalized
communication networks such as 3rd Generation Partnership Project’s “Personal
Service Environment” and “Virtual Home Environment” [25].
Instrumental personalization highlights the importance of the user’s situated
needs. Deviating from traditional artificial intelligence research that treats the per
sonasa rational,linear informationprocessing system,studies ofsituated needsand
actionargue that“everycourse ofactiondependsin essentialwaysuponits material
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
cumstances and represent it as a rational plan, the approach is to study how people
usetheir circumstancestoachieve intelligentaction”([58], p.7).Personalization sys
tems designed under the instrumental perspective utilize information about the
user’s contextsuch as time,location,and surrounding environmentalparameters to
make inferences or predictions and to act accordingly [59].
4.3 Relational Personalization
Another way to personalize one’s world is to create a unique web of social relation
ships. Thisapproach ismostclosely associatedwithsociology, communication,and
anthropology and is reflected in the definitions in the social science category in Ta
ble 1. Positive social relationships give individuals a sense of well-being by creating
support and a sense that they are not alone and are valued. In a very real sense, they
lend an aura of the personal to one’s world.
Relational personalization can be defined as the mediation of interpersonal rela
tionships and utilization of relational resources to facilitate social interactions by
providing a convenient platform for people to interact with others in a way that is
compatible with the individual’s desired level of communality and privacy. The
motivation behind relational personalization is to fulfill the user’s needs for social-
ization and a sense of belonging. The goal of relational personalization is twofold:
(a) to enhance the effectiveness of interpersonal interactions and (b) to help gener-
ate “social capital” [4] by providing new opportunities for strengthening social re-
lationships and maintaining social networks. Relational personalization takes a
myriad of forms, ranging from personalized gifts to computer-mediated interper-
sonal communication (MIT Media Lab).
Personalization systems designed according to the relational perspective focus
on a strategy of mediation. They seek to provide a common, convenient platform
for interpersonal communication and community building that emphasizes design
on the basis of what Preece [60] termed sociability. Once a social network has
emerged, the designer can use this critical mass to further enlist users and increase
the relational potential of the network. Applications amenable to relational person
alization vary greatly in size and complexity. They can be as simple as providing an
“e-mail to a friend” button to notify others of one’s flight schedule after booking
tickets online or as complicated as a conglomeration of online information portal
and activity center in a “Digital City” that engages residents or visitors [61].
The relational perspective models the user’s relational needs and the social con
text that satisfies them. Preece [60] discussed several aspects of social context that
are important in meeting users’ needs for sociability. These include a clearly stated
purpose that attracts people with similar goals and interests to the community,
provision of people who play key roles in the community (moderators, mediators),
and community governance policies that make participation safe and preserve pri
vacy. In this model, the individual’s attempt to achieve a desired level of privacy on
one hand is balanced by his or her attempt to maintain a sense of community on the
other [62]. A relationship that maintains a desired level of privacy ensures good
community building, whereas good community building enhances trust.
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
4.4 Commercial Personalization
One of the most important human activities is the consumption of goods and ser
vices. Personalization driven by the commercial perspective is reflected in the defi
nitions in the marketing/e-commerce categories in Table 1. Adopting Riechen’s
[63] definition, wedefine commercial personalization as thedifferentiationof prod
uct, service, and information to increase sales and to enhance customer loyalty by
segmenting customers in a way that efficiently and knowledgeably address each
user or group of users’ needs and goals in a given context. Commercial personaliza
tion is strongly technology driven. Information technology makes mass personal
ization possible through the personalized channel of “addressable media.” The en
tire business paradigm shifts from mass-produced goods and standardized
services to an emphasis on one-to-one contact as discussed by Peppers and Rogers
[64] in The One To One Future: Building Relationships One Customer at a Time, a corner
stone book for much recent activity in the personalization industry.
The motivation of commercial personalization is to fulfill users’ material needs
and thus contribute to their psychic welfare [65]. Commercial personalization pri
marily focuses on the content of the system. The assumption is that product, ser-
vice, and information of high relevance to the consumer yields a satisfying
shopping experience and loyal adherence to the Web site as well as the organiza-
tion behind it. The goal of commercial personalization is to increase sales directly
and through cross sales [14] and to increase customer loyalty and build brands [63].
Customers benefit from customized products, individualized services, and an en-
hanced experience [10]. Cultivating a one-to-one relationship makes future trans-
actions smoother and more efficient, benefiting both parties in the long run.
The primary strategy of commercial personalization is segmentation. Commer-
cial personalization is ultimately effective only to the extent that the offerings pro-
vide value to the target market segments by differentiating the product, service,
and information provided. Business goals are sometimes in direct conflict with the
interests of consumers, who are money conscious, time conscious, and sensitive to
privacy infringement. Personalization strategies merely for the benefit of the busi
ness are not sustainable even if they result in an initial sales boost.
Rich knowledge about the personalizee is a prerequisite for success in commer
cial personalization. This requires continuous learning about each individual, un
derstood as a systemic entity in terms of personal preferences and interests [66],
cognitive ability, motives, demographic or psycho-cultural profiles [14], user be
haviors [12], and specific contexts. Two types of contextual information are impor
tant for adaptive personalization. One type pertains to users’ intent, preferences,
and purchasing history, whereas the other relates to environmental factors such as
time and location of the user [67]. Effective personalization takes into account these
contextual elements to better anticipate customer needs and predict the goods and
services that will satisfy them.
4.5 Implications for Design
In this section, we discuss paradigmatic similarities, differences, and
complementarities among the four perspectives. The purpose of such exercise is
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
to anchor the perspectives to design strategies. A close examination of the four
personalization ideal types reveals that the four types can be further classified
intoa2×2grid. First, the perspectives can be differentiated in terms of utilitar
ian or affective orientation. The instrumental and commercial perspectives em
phasize task achievement and commercial transactions and therefore are
oriented to utilitarian issues, whereas architectural and relational perspectives
place more emphasis on users’ feelings, both aesthetic and socioemotional. The
perspectives can also be differentiated in terms of the basic premise of use,
whether the user primarily engages the system as an individual or through an in
teraction. Both architectural and instrumental personalization are concerned with
individual use of an artifact, be it a building, an information system, or a Web
site. Design emphasis is on an individual’s interaction with the artifact. On the
contrary, relations among multiple entities and the management of the relations
are of paramount importance in relational and commercial personalization.
Commercial and instrumental personalization, predominantly used for infor
mation retrieval, transaction processing, and content management, belong to the
class of productivity applications [68]. The purpose for personalization is utilitar
ian oriented, the goal of which is to get something done. Hence, content, function-
ality, and usability are given priority in design. In contrast, architectural and
relational personalization—primarily used for creating an attractive Web environ-
ment, an interactive social network, and a sense of psychological and social well-
being—belong to the class of entertainment applications. The purpose for person-
alization is affect oriented, the goal of which lies in the experience itself. Hence, a
balance between form and function as well as meaning of the using the system is
emphasized. Affective design is process oriented, whereas utilitarian design is re-
sults oriented [68]. Note that the distinction between productivity and entertain-
ment applications is based on the intended use of the software, not the intention of
the user. For example, a recommender system that suggests potentially interesting
DVD titles to the user would be a productivity application because the built-in
function of the recommender is to reduce information overload by focusing on in
formation relevant to the specific user. The difference between productivity appli
cations and entertainment applications is important because a series of design
decisions are contingent on the nature of the application.
The majority of existing personalization systems are designed to enhance pro
ductivity, whether in the form of one-click ordering (e.g., or
wireless, just-in-time, personalized information services such as stock, weather,
and local traffic information (e.g., DoCoMo). People use these systems to get things
desired (such as relevant information, quality product, or service) but not the expe
rience of using such systems (see Figure 1). Hence, the utility function is to maxi
mize convenience and efficiency [53]. Productivity applications are results
oriented and task oriented [68], with a focus on results such as fast checkout, high-
quality information retrieval, and immediate response. Design guidelines for these
types of task-oriented applications are similar to those for designing tools such as
the principles advanced by Norman [69] for designing everyday things in which
content, functionality, and usability are emphasized. Key usability issues for pro
ductivity applications are ease of use, clarity, consistency, freedom from ambigu
ity, and error. The aspect of ease of use includes both the use of the application itself
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
and the setup and configuration to make personalized features functional. Consis
tency helps users better orient themselves to the site and alleviates cognitive effort.
For example, in Amazon, the shopping cart is always on the upper right-hand cor
ner, browsing history is always displayed on the left column, and recommendation
list always appears after the user places an item in the shopping cart.
Nevertheless, in addition to “getting things done,” people also have a need to
“simply enjoy things.” This implies that personalization systems may not only ful
fill the functional aspects of human needs but also their entertainment aspects. En-
tertainment-oriented personalization applications capitalize on the process and
experience of using the systems. They are designed to stimulate thinking and to in-
voke feelings. The results are not tangible, but the process itself is critical in creat-
ing an engaging, fulfilling user experience. The principle of consistency may not be
sufficient to invoke feelings or engage users on the site for an extended amount of
Architectural personalization and relational personalization have provided in-
sight into designing for affect. A key design principle of architectural personaliza-
tion is the balance of function and form. The idea of function and form as one was
advocated by the influential Bauhaus movement in architecture. Later articulated
by Frank Lloyd Wright, “form was to display the functionality of a building in an
organic way” ([70], p. 162). Rather than “looking functional,” “forms in organic ar
chitecture are uniquely suited to their purposes” ([70], p. 162). The organic, holistic
view of forms and function implies that decisions related to form should be moved
forward in the design life cycle to be considered together along with function. Rela
tional personalization, on the other hand, lays an emphasis on meaning, which is
derived from social interactions with different circles of life from close friends and
family, to immediate community, and to the society at large [71].
Although the ideal types represent distinctive paradigms of design strategy,
there exists great potential to combine multiple paradigms in a way that best meet
different needs of users. A design that combines function and form; embeds mean
ing in use; and integrates productivity, education, and entertainment is more likely
to fulfill human needs. A good example of combining form and function is found in
Disney World Web site. The “select an experience” bar presents to the users differ
ent segments of audience, that is, “preschoolers,” “teens,” “big kids,” “romance
and relaxation,” and so forth. Once an experience is selected with a sparkling
magic wand (the cursor), the site takes the user to a three-dimensional map where
users conveniently locate points of attractions and activity spots that are suitable
for that specific audience group. Combining productivity, education, and enter
Figure 1. Personalization design paradigms.
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
tainment is another effective way to promote use of personalization. The Web site
for the Public Broadcasting System teenage reality show American High is an excel
lent example of utilizing relational personalization for teen education. The Web
site is an innovative way of bringing students, teachers, parents, educators, and
artists together, making sure every voice gets heard and every role benefits from
this technological and artistic collaboration. Students can meet and chat with char
acters in the show, talk with the filmmaker, and experiment with a personalized
online yearbook. Teachers of media arts or social studies can find lesson plans that
provide a framework for creating student-produced video diaries and explore so
cial issues related to reality TV ( These two
examples, a business organization and a nonprofit organization, both show that
there exist tremendous possibilities for creating a personalized experience by com
bining perspectives.
In this article, we attempted to achieve two purposes. First, we have advanced a
working definition for personalization and proposed a classification scheme to
frame personalization research and practice. Second, we have developed a norma-
tive framework of personalization ideal types that distinguishes four distinct per-
sonalization design philosophies.
The goal of the classification scheme was to give a clearer structure to the diverse
and rapidly developing field of personalization. The scheme focuses on a general
description of personalization in terms of (a) the elements of a system that can be
personalized (content, functionality, interface, channel), (b) who initiates the per-
sonalization activity (user or system), and (c) the target of personalization activity
(individual or group). We believe these properties span the space of current ap-
proaches to personalization. These represent a set of core design choices for system
developers. They are also useful as a structure for organizing previous research
and to situate future projects. The goal of the definition and classification scheme is
to help to establish a common language for talking about personalization that can
help to integrate this multidisciplinary field.
The significance of the personalization ideal types is twofold. First, they define
different lenses for personalization research and practice. Although it is useful to
describe possibilities for personalization systems, in the end, personalization is a
practice that is shaped by the designer’s motives for personalization and viewpoint
on “what personalization really is.” The personalization ideal types, although ad
mittedly abstract, are useful because they identify relatively consistent tendencies
in personalization theory, research, and practice. Each ideal type describes a differ
ent philosophy that is built around a view of the particular goals for personaliza
tion and what it means to satisfy those goals. We suspect that different
organizations and different activities would best fit particular ideal types and so
the match between ideal type and the context in which the personalization system
operates would influence its effectiveness.
The ideal type scheme implies that no single standard or approach to personal
ization is “the best.” Each ideal type employs different criteria for evaluating how
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
well the system succeeds in delivering the desired effect. Each type also has a dif
ferent model of the user, which suggests that different user modeling techniques
and implementation tools should be employed. Researchers and developers who
operate within a single perspective on personalization have tended to generalize
the motives, viewpoint, and approaches of that perspective to the entire field. This
has the potential to limit creativity and confine the development of personalization
systems to a few well-defined trajectories. In defining multiple ideal types for per
sonalization, we are attempting to open a wider space for researchers and design
ers that will enable them to imagine other ways of doing personalization than their
accepted approach. Currently, the instrumental and commercial approaches to
personalization are the most commonly employed. Most of the prototypes and
methodologies that have been described in the literature on personalization have
been embedded in these two perspectives. This is, to some extent, appropriate be
cause IS research is situated at the intersection of management, organizational, and
computer sciences. However, being embedded in particular ways of thinking can
blind one to other possibilities. The architectural perspective, with its emphasis on
balancing aesthetics and functionality, and the relational perspective, which ar
gues that personalization is best handled by creating a personal social world on-
line, offer quite different approaches to personalization. They throw prevailing
thinking into perspective and suggest novel approaches to personalization.
A second goal for the ideal types is to provide a first step toward formulating a
systematic methodology for the design and development of personalization sys-
tems. Jupiter Research, Forrester Research, and Mainspring Research, among oth-
ers, identified several major obstacles to the effective implementation and
evaluation of personalization systems [72] including low return on investment,
lack of measurement methodologies, low levels of technology adoption, and
mounting technical difficulties. The ideal type system can help address at least the
first two of these. In this analysis, we specified four distinct kinds of user motives
for using personalization systems: aesthetic value for architectural personalization,
social welfare/psychological well-being for relational personalization, productiv
ity/efficiency for instrumental personalization, and material and psychic well-
being for commercial personalization. These motive types suggest different stan
dards for assessing the effectiveness of personalization, which should help re
searchers and IS professionals better focus instruments for measuring the impacts
of personalization. Although return on investment and click-to-buy rates are the
most widely used measures of personalization effectiveness on Web sites [72], our
analysis suggests that it is not reasonable to measure everything using a single
yardstick. Other measurement constructs should be developed to suit different
The ideal types are theoretical constructs that can guide research and practice
and do not represent the realities of practice itself. It is possible, and even likely,
that two perspectives might be combined in designing particular personalization
applications. For example, the popular online role-playing games such as
Everquest® seem to combine the architectural and relational perspectives to enable
users to create shared worlds that to many users seem more real and desirable than
“real life.” This seems to represent a relatively harmonious mix of types, but it is
also possible that perspectives could be combined in a dissonant fashion.
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
The frameworks wepresentedin this article suggest several directionsforfuture re
search. First, from this multiparadigm review, one can see that personalization is a
multidimensional construct. Measurement of such a multidimensional construct is
always a challenge. Existing literatures in e-commerce and marketing tend to adopt
a monolithic approach [73, 74]; therefore, personalization dimensions that have
been evaluated in those studies might not have captured personalization values
that users/customers have [44, 75]. We are currently developing and validating
such a measurement instrument that incorporated all paradigms we reviewed here
in this article. Second,there is the question of how existingpracticesin the personal
ization industrymap ontothetypologies. Whichideal types aremost common, how
are they combined, and what is their effectiveness? Which aspects of personaliza
tion systems contribute to their effectiveness? What factors, for example, contribute
to the effectiveness of adaptive versus adaptable systems? How might one validate
the typologies? Third, there are different levels of understanding of the perspec
tives. As we have noted, the commercial and instrumental types have enjoyed a
great deal of attention. The architectural and relational types require further explo-
ration and development. We suspect that a number of practitioners of personaliza-
tion have pursued the architectural and relational approaches despite the fact that
they have not been discussed much in the academic IS literature. Study of these
practitioners seems likely to yield insights into personalization designs and meth-
ods that are different from those currently described in the literature.
Personalization is one of those subjects that will always be with humans. In dif-
ferent forms and guises, it continues to maintain currency. Personalization will
continue to be an important dimension if IS because it has a central place in a soci-
ety embracing heterogeneity and diversity, an economy increasingly individual
oriented, and a capitalism remarkably personalized [76]. Because of its inherently
human element, personalization cannot be reduced to a technical undertaking.
When one realizes that personalization does not necessarily have to be solely profit
driven and that personalization is not merely a technical issue, one can free their
imagination to make personalization richer, more meaningful, and more relevant.
As Brooks [77] pointed out, “we can use the softer elements of our humanity to de
sign the harder mechanisms of our technology” (p. 15). On Brooks’s view, true bal
ance is achieved through “pas de deux” of social science and engineering, the same
interaction that underlies the frameworks advanced here.
Both authors contributed to this article equally. An earlier version of thisarticlewas
presented at the American Conference on Information Systems in Tampa, Florida,
in August 2003.
[1] T. Davenport and J. Beck, The Attention Economy. Boston: Harvard Business School Press, 2001.
[2] B.J.Pineand J. H. Gilmore, The Experience Economy. Boston:HarvardBusinessSchool Press, 1999.
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
[3] J. Ledford and R. Lawler. Content Personalization Market Trends. Pennsauken, NJ: Faulkner Infor
mation Services, 2002.
[4] B. Wellman, “Designing the internet for a networked society,”Comm. of the ACM, vol. 45, no. 5, pp.
91–96, 2002.
[5] J. Cummings, B. Butler, and R. Kraut, “The quality of online social relationships,” Comm. of the
ACM, vol. 45, no. 7, pp. 103–107, 2002.
[6] J. Blom, “Personalization—A taxonomy,” presented at CHI2000 Conf. on Human Factors in Com
puting Systems, The Hague, The Netherlands, 2000.
[7] C. Wachob. (2002). What are personalization and customization? [Online]. Available: http://workz.
[8] J. Nielson. (1998, October 4). “Personalization is over-rated” [Online]. Jakob Nielsen’s Alertbox
[9] U. Manber, A. Patel, and J. Robison, “Experience with personalization on Yahoo!,” Comm. of the
ACM, vol. 43, no. 8, pp. 35–39, 2000.
[10] B. Ives and G. Piccoli, “Custom made apparel and individualized service at Lands’ End,” Comm. of
the AIS, vol. 11, no. 1, pp. 11–26, 2003.
[11] M. Schneider-Hufschmidt, T. Kuhme, and U. Malinowski, Adaptive User Interface: Principles and
Practice. Amsterdam: North Holland, 1993.
[12] Vignette Corp., Personalization Strategies-Fit Technology to Business White Paper. Austin, TX:Author,
[13] K.Instone,InformationArchitecture and Personalization. Ann Arbor, MI: ArgusAssoc.,Inc.,2000.
[14] J. Nejezchieb, Personalization: A View of the Current Landscape. San Francisco: Vertebrae, 2001.
[15] D. Rubini, “Overcoming the paradox of personalization: Building adoption, loyalty and trust in
digital markets,” Design Management Journal, vol. 12, no. 2, pp. 49–54, 2001.
[16] C. M. Karat, C. Brodie, J. Karat, J. Vergo, and S. R. Alpert, “Personalizing the user experience on,” IBM Systems Journal, vol. 42, no. 4, pp. 686–701, 2003.
[17] J. Blom and A. Monk, “One-to-one e-Commerce: Who’s the one?,” presented at CHI Conf. on Hu-
man Factors in Computing Systems, Seattle, WA, 2001.
[18] B. Reeves and C. Nass, The Media Equation. Cambridge, England: Cambridge University Press,
[19] E. Andre and T. Rist, “Presenting through performing: On the use of multiple lifelike characters in
knowledge-basedpresentationsystems,”Intelligent User Interface, vol. 9, no. 2, pp. 147–165, 2000.
[20] D.C.Dryer,C.Eisbach,andW. S. Ark, “At what cost pervasive? A social computing view of mobile
computing systems,” IBM Systems Journal, vol. 38, no. 4, pp. 652–676, 1999.
[21] H. Hirsh, C. Basu, and B. Davison, “Learning to personalize,” Comm. of the ACM, vol. 43, no. 8, pp.
102–106, 2000.
[22] A. K. Dey, “Understanding and using context,” Personal and Ubiquitous Computing Journal, vol. 5,
no. 1, pp. 4–7, 2001.
[23] G. Adomavicius and A. Tuzhilin, “An architecture of e-butler—A consumer-centric online person
alization system,” International Journal of Computational Intelligenceand Applications, vol. 2,no. 3, pp.
313–327, 2002.
[24] P. Maes, R. Guttman, and A. Moukas, “Agents that buy and sell,” Comm. of the ACM, vol. 42, no. 3,
pp. 81–91, 1999.
[25] O. Tomarchio, A. Calvagna, and G. DiModica, “Virtual Home Environment for Multimedia Ser
vices in 3rd Generation Networks,” 3rd Generation Partnership Project, Tech. Report, 2002.
[26] G. Rossi, D. Schwabe, andR. Guimaraes, “Designing personalizedweb applications,” presentedat
the 10th Int. Conf. on World Wide Web, Hong Kong, 2001.
[27] D. Wu, I. Im, M. Tremaine, K. Instone, and M. Turoff, “A framework for classifying personalization
scheme used on e-commerce web sites,” presented at 36th Hawaii Int. Conf. on System Sciences,
Big Island, Hawaii, 2003.
[28] R. Spears and M. Lea, “Panacea or panopticon?: The hidden power in computer-mediated
communication,” Communication Research, vol. 21, no. 4, pp. 427–459, 1994.
[29] M. A. Hogg and D. Abrams, Social Identifications. London: Routledge, 1992.
[30] D. Amoroso and B. Reinig, “Personalization management systems,” presented at 36th Hawaii Int.
Conf. on System Sciences, Big Island, Hawaii, 2003.
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
[31] M.LewisandA.Grimes,“Metatriangulation: Building theory from multiple paradigms,” Academy
of Management Review, vol. 24, no. 4, pp. 672–690, 1999.
[32] G. Burrell and G. Morgan, Sociological Paradigms and Organizational Analysis. Portsmouth, NH:
Heinemann, 1979.
[33] R. Rogers, Max Weber’s Ideal Type Theory. New York: Philosophical Library Co., 1969.
[34] B. Shneiderman, Designing the User Interface: Strategies for Effective Human-Computer-
Interaction. Reading, MA: Addison-Wesley, 1993.
[35] R. Kling, “Assimilating social values in computer-based technologies,” Telecommunications Policy,
vol. 8, no. 2, pp. 127–147, 1984.
[36] A. S. Lee, “Architecture as a reference discipline for MIS,” in Information Systems Research: Contem
porary Approaches and Emergent Traditions, H. E. Nissen, H. K. Klein, and R. Hirschheim,
Eds. North-Holland, The Netherlands: Elsevier Science, 1991.
[37] J. Kim, J. Lee,K. Han, andM. Lee,“Business asbuildings: Metrics forthe architecturalquality ofin
ternet business,” Information Systems Research, vol. 13, no. 3, pp. 239–254, 2002.
[38] F. Becker, “User participation, personalization, and environmental meaning: Three field studies,”
in Program in Urban and Regional Studies. Ithaca, NY: Cornell University, 1977.
[39] M. Bonnes and G. Secchiaroli, Environmental Psychology: A Psycho-Social Introduction. London:
Sage, 1995.
[40] I. Altman, The Environment and Social Behavior: Privacy, Personal Space, Terminality and
Crowding. Monterey, CA: Brooks/Cole, 1975.
[41] D. Canter, Psychology for Architects. London: Applied Science Publication, 1974.
[42] F. Hendricks and M. MacNair, Concepts of environmental quality standards based on life styles. Pitts
burgh, PA: University of Pittsburgh, 1969.
[43] C. J. Holahan, Environment and Behavior: A Dynamic Perspective. New York: Plenum, 1978.
[44] Novak, D. Hoffman, and Y.-F. Yung, “Measuring the customer experience in online environment:
A structural modeling approach,” Marketing Science, vol. 19, pp. 22–42, 2000.
[45] R. Maxwell, Contexts of Behavior: Anthropological Dimensions. Chicago: Nelson-Hall, 1983.
[46] L. Barfield, The User Interface: Concepts and Design. Workingham, England: Addison-Wesley,
[47] S. Van der Ryn and M. Silverstein, “Dorms in Berkeley: An environmental analysis,” Center for
Planning and Developmental Research, University of California, Berkeley, 1967.
[48] W. R. Rosengren and S. Devault, “The sociologyof time and space inan obstetrical hospital,” in En-
vironmental Psychology, W. H. I. H. M. Proshansky and L. G. Rivilin, Eds. New York: Holt,
Rinehart & Winston, 1970.
[49] H. Osmond, “The Relationship between architect and psychiatrist,” in Psychiatric Architecture, C.
Goshen, Ed. New York: Doubleday-Anchor Press, 1959.
[50] I. Graham, A Pattern Language for Web Usability. London: Addison-Wesley, 2003.
[51] N. Sgouros, G. Papakonstantinou, and P. Tsanakas, “Dynamic dramatization of multimedia story
presentation,” presented at 2nd Int. Conf. on Intelligent User Interfaces, Orlando, FL, 1997.
[52] E. Andre and T. Rist, “From adaptive hypertext to personalized web companions,” Comm. of the
ACM, vol. 45, no. 5, pp. 43–46, 2002.
[53] N. Sadeh, M-Commerce: Technologies, Services, and Business Models. New York: Wiley Computer
Publishing, 2002.
[54] I. Clarke, “Emerging valuepropositions form-commerce,” Journalof BusinessStrategies, vol. 18, no.
2, pp. 133–148, 2001.
[55] K. Cheverst, K. Mitchell, andN. Davies, “Therole ofadaptive hypermedia ina context-aware tour
ist guide,” Comm. of the ACM, vol. 45, no. 5, pp. 47–51, 2002.
[56] L. Ardissonon, A. Goy, G. Petrone, and M. Segnan, “Personalization inbusiness-to-customer inter
action,” Comm. of the ACM, vol. 45, no. 5, pp. 52–53, 2002.
[57] D. Rose, “Experience architecture: A Framework for designing personalized customer interac
tions,” Design Management Journal, vol. 12, no. 2, pp. 68–77, Spring 2001.
[58] L. Suchman, Plans and Situated Action. New York: Cambridge University Press, 1987.
[59] P. Dourish, “A foundational framework for situated computing: Position paper,” presented at the
CHI 2000 Workshop on Situated Computing: A Research Agenda, The Hague, The Netherlands,
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
[60] J. Preece, Online Communities: Designing Usability, Supporting Sociability. Chichester, England:
Wiley, 2000.
[61] I. Toru, “Digital city Kyoto,” Comm. of the ACM, vol. 45, no. 7, pp. 76–81, 2002.
[62] C. Alexander, A Pattern Language. New York: Oxford University Press, 1977.
[63] D.Riecken,“Personalizedviews of personalization,” Comm. of ACM, vol. 43, no. 8, pp. 26–28, 2000.
[64] D. Peppers and M. Rogers, The One To One Future: Building Relationships One Customer at a
Time. New York: Currency and Doubleday, 1993.
[65] M. Douglas and B. Isherwood, The World of Goods. New York: Basic Books, 1979.
[66] S. Larsen and S. Tutterow,” Developing the personalization-centric enterprise through collabora
tive filtering and rules-based technologies,” CRM Project, vol. 1, 1999.
[67] M. Bayler and D. Stoughton, “Design challenges in multi-channel digital markets,” Design Manage
ment Journal, vol. 12, no. 2, pp. 29–36, 2001.
[68] R. J. Pagulayan, K. Keeker, D. Wixon, R. L. Romero, and T. Fuller, “User-centered design in
games,” in The Human–Computer Interaction Handbook: Fundamentals, Evolving Technologies and
Emerging Applications, A. Sears, Ed. Mahwah, NJ: Lawrence Erlbaum Associates, Inc., 2003, pp.
[69] D. Norman, The Psychology of Everyday Things. New York: Doubleday, 1988.
[70] D. Norman and S. Draper,User-Centered System Design:New Perspectives onHuman–Computer Inter
action. Hillsdale, NJ: Lawrence Erlbaum Associates, Inc., 1986.
[71] B.Shneiderman,Leonardo’sLaptop:Human Needs and the New Computing Technologies. Cambridge,
MA: MIT Press, 2003.
[72] P. R. Hagen, Smart Personalization. Cambridge, MA: Forrester Research, Inc., July 1999.
[73] S. S. Srinivasan, R. Anderson, and K. Pnnocolu, “Customer loyalty in e-commerce: An exploration
of its antecedents and consequences,” Journal of Retailing, vol. 78, no. 1, pp. 41–50, 2002.
[74] J. Palmer, “Web site usability,design, and performance metrics,” Information Systems Research, vol.
13, no. 2, pp. 151–167, 2002.
[75] P. Chen and L. Hitt, “Measuring switching costs and the determinants of customer retention in in-
ternet-enabled business: A study of the online brokerage industry,” Information Systems Research,
vol. 13, no. 3, pp. 255–274, 2002.
[76] S. Zuboff and J. Maxmin, The Support Economy: Why CorporationsAre Failing Individuals and The Next
Episode of Capitalism. New York: Viking, 2002.
[77] K. Brooks, “Pas de deux—The dance of digital design,” Design Management Journal, vol. 12, no. 2,
pp. 10–16, 2001.
[78] Personalization Consortium (2003), “What is personalization,” Personalization Consortium, [On
line]. Available:
[79] B. Kasanoff, Making It Personal. Cambridge, MA: Perseus Publishing, 2001.
[80] C. Surprenant and M. Solomon, “Predictability and personalization in the service encounter,”
Journal of Marketing, vol. 51, no. 2, pp. 86–96, 1987.
[81] A. Kobsa, “User modeling as a key factor in system personalization,” presented at CHI 2000, The
Hague, The Netherlands, 2000.
[82] P. Brusilovsky and M. T.Maybury, “From adaptivehypermedia tothe adaptive web,”Comm. of the
ACM, vol. 45, no. 5, pp. 30–33, 2002.
[83] J. Karat, C. Marat, and J. Ukelson, “Affordances, motivations, and the design of user interfaces,”
Comm. of the ACM, vol. 43, no. 8, pp. 49–51, 2000.
[84] J. Kramer, S. Noronha, and J. Vergo, “A user-centered design approach to personalization,” ACM
Computing Surveys, vol. 43, no. 8, pp. 44–48, 2000.
[85] D. Riechen, “Personalized communication networks,” Comm. of the ACM, vol. 43, no. 8, pp. 41–42,
[86] W. Bender, “Twenty years of personalization: All about the ‘daily me’,” EDUCAUSE Review,
37, no. 1, pp. 20–29, 2002.
[87] W. Kim, “Personalization: Definition, status, and challenges ahead,” Journal of Object Technology,
vol. 1, no. 1, pp. 29–40, 2002.
Downloaded By: [University of Queensland] At: 04:32 12 April 2011
... In the first cycle, we first conducted interviews with youth suffering from depression to gain an in-depth understanding of the problem, their needs, and preferences. Based on the interviews, CBT and IPT, and theories of personalization [18,21], we derived two initial design principles (DPs) for personalized CAs to treat depression. Next, we instantiated these two initial design principles in four prototypes, which were evaluated in interviews with five experts and five potential users. ...
... Users appreciate personalization features because they can improve ease of use, efficiency, and provide users with a feeling of being in control [23]. Our work draws on the frameworks of personalization approaches of Fan and Poole [21] and Kocaballi et al. [18]. Depending on the specific field of research and discipline, personalization is often used synonymously with adaptation, customization, and tailoring [21]. ...
... Our work draws on the frameworks of personalization approaches of Fan and Poole [21] and Kocaballi et al. [18]. Depending on the specific field of research and discipline, personalization is often used synonymously with adaptation, customization, and tailoring [21]. We decided to use the term personalization because it is commonly used in the medical and health literature [17]. ...
Conference Paper
Full-text available
Depression is a large-scale and consequential problem in youth and young adults. Conversational agents (CAs) can contribute to addressing current barriers to seeking treatment, such as long waiting lists, and reduce the high dropout rates reported for other digital health interventions. However, existing CAs have not considered differences between youth and adults and are primarily designed based on a ‘one-size-fits-all’ approach that neglects individual symptoms and preferences. Therefore, we propose a theory-driven design for personalized CAs to treat depression in youth and young adults. Based on interviews with patients (i.e., people diagnosed with depression), we derive two design principles to personalize the character of the CA and its therapeutic content. These principles are instantiated in prototypes and evaluated in interviews with experts experienced in delivering psychotherapy and potential nondiagnosed users. Personalization was perceived as crucial for treatment success, and autonomy and transparency emerged as important themes for personalization. We contribute by providing design principles for personalized CAs for mental health that extend previous CA research in the context of mental health.
... The level of personalization can be expressed in terms of differentiation in online communication with the customer and categorized into three forms (Fan & Poole, 2006;Vesanen, 2007). First, mass marketing represents the zero level of personalization, as this strategy involves no differentiation in customer communication. ...
... Third, an even higher level of personalization involves individualized communication. The content can be adapted to the customer's needs because the company has a deeper data-driven understanding of the customer (e.g., Fan & Poole, 2006). With any form of personalization, customers' perceived benefits from the incentive are based on subjective evaluations gleaned by comparing the benefits with their motives and goals Wilson & Valacich, 2012). ...
... Individualization goes a step further, with communication and offers tailored to an individual's interests and needs. In general, personalization constructs a superficial understanding of customers (e.g., knowing their names), while individualization is more in-depth in incorporating individual needs and behavior based on the data collected (Fan & Poole, 2006;Vesanen, 2007). ...
E-commerce companies are beginning to use data analytics to individually address their customers, as such firms are strongly dependent on customers’ willingness to disclose (WTD) personal information. Conceptually known as privacy calculus, a cognitive comparison of the predicted benefits and risks of such a disclosure results in subjective expected utility (SEU). As this construct has not been elucidated in the context of situational factors, namely personalization and information sensitivity, this research utilizes a quantitative experimental study design to analyze these factors and evaluate the impact of SEU on customers’ WTD. Based on an online survey, the study reveals a positive but decreasing effect of personalization on SEU, while identifying information sensitivity as a negative influential factor. In general, the results emphasize the importance of a fair and transparent exchange relationship to foster a data disclosure setting without acute data privacy concerns.
... Personalization supports the co-creation of experiences by providing products and services that fit the customer's context, preferences, and tastes (Fan & Poole, 2006). When making a purchase, customers can require considerable search time for information to make the right decision (Nieto-García, Muñoz-Gallego, & Gonzalez-Benito, 2020). ...
... To advance the understanding of the relationship between CeoP and WTP, four conceptual dimensions frame the proposed typology: Personalization, WTP, customer philosophy, and novelty-familiarity continuum. This study adopts the types of personalization proposed by Fan & Poole (2006), focusing on information systems and Vesanen (2007), which adopted a socio-marketing perspective. Each personalization type represents a set of core design choices grounded in the strategy, motive, time of use, and customer involvement and describes them as architectural, relation, instrumental, commercial (Fan & Poole, 2006), cosmetic, adaptive, transparent, and collaborative (Vesanen, 2007). ...
... This study adopts the types of personalization proposed by Fan & Poole (2006), focusing on information systems and Vesanen (2007), which adopted a socio-marketing perspective. Each personalization type represents a set of core design choices grounded in the strategy, motive, time of use, and customer involvement and describes them as architectural, relation, instrumental, commercial (Fan & Poole, 2006), cosmetic, adaptive, transparent, and collaborative (Vesanen, 2007). Table 1 depicts the ideal type and motives of personalization. ...
Personalization drives value co-creation and willingness to pay for customers. Consumers are keen to receive personalized services but have various willingness to pay for the personalization process. The willingness to pay is influenced by motives for customer purchase behavior and personalization expectations in a specific context. It also depends on disposable income and the availability of resources, as well as the severity of requirements. The results indicate that customers comprise a heterogeneous market concerning their personalization expectations and willingness to pay. The paper proposes a customer typology based on a conceptual framework that includes personalization, willingness to pay, customer philosophy, and novelty-familiarity continuum. By analyzing data from thirty-eight semi-structured interviews, six customer types are proposed, namely: Budget Adventurer, Family Explorer, Relation Seeker, Relaxation Seeker, Delight Seeker, and Must-Have Customer. The findings suggest that revenue managers should understand customer personalization preferences for each type in order to develop effective pricing strategies.
... And emotional intelligence reflects the empathy and emotional expression ability of chatbot [60]. The final personalization represents whether the chatbot can self-adjust for different types of users and has the ability to serve different users and meet their preferences [22]. ...
... equal supervision 1. Personalization is a term that encompasses a breadth of techniques that use personal data. Here, we use it to describe approaches that target groups rather than individuals -i.e., "categorization" rather than "individualization" as per the taxonomy of Fan and Poole [36]. ...
Full-text available
The standard approach to personalization in machine learning consists of training a model with group attributes like sex, age group, and blood type. In this work, we show that this approach to personalization fails to improve performance for all groups who provide personal data. We discuss how this effect inflicts harm in applications where models assign predictions on the basis of group membership. We propose collective preference guarantees to ensure the fair use of group attributes in prediction. We characterize how common approaches to personalization violate fair use due to failures in model development and deployment. We conduct a comprehensive empirical study of personalization in clinical prediction models. Our results highlight the prevalence of fair use violations, demonstrate actionable interventions to mitigate harm and underscore the need to measure the gains of personalization for all groups who provide personal data.
... First, the main task is not prediction, but rather in RL, the main goal is to select intervention actions so that average rewards across time are maximized for each user. We call this the goal of Personalization [15]. We generalize the PCS framework to include an evaluation of the ability of an online RL algorithm to personalize. ...
Full-text available
Online reinforcement learning (RL) algorithms are increasingly used to personalize digital interventions in the fields of mobile health and online education. Common challenges in designing and testing an RL algorithm in these settings include ensuring the RL algorithm can learn and run stably under real-time constraints, and accounting for the complexity of the environment, e.g., a lack of accurate mechanistic models for the user dynamics. To guide how one can tackle these challenges, we extend the PCS (Predictability, Computability, Stability) framework, a data science framework that incorporates best practices from machine learning and statistics in supervised learning (Yu and Kumbier, 2020), to the design of RL algorithms for the digital interventions setting. Further, we provide guidelines on how to design simulation environments, a crucial tool for evaluating RL candidate algorithms using the PCS framework. We illustrate the use of the PCS framework for designing an RL algorithm for Oralytics, a mobile health study aiming to improve users' tooth-brushing behaviors through the personalized delivery of intervention messages. Oralytics will go into the field in late 2022.
... These advantages allow GP to model both (a) intradriver stochasticity (variation of driving behaviors within the same driver, through variance modeling), as well as (b) interdriver stochasticity (through individual training). Instead of the explicit personalization (i.e., offering drivers to choose from a number of predefined system settings), we focus on the implicit personalization (estimating the drivers' preferences based on their past behaviors) [36]. GP regression can be utilized to identify the relationship between input (driver's perceived information) and output (desired acceleration), and hence provides personalized guidance towards driving. ...
Full-text available
Advanced driver-assistance systems (ADAS) havematured over the past few decades with the dedication toenhance user experience and gain a wider market penetration.However, personalization features, as an approach to make thecurrent technologies more acceptable and trustworthy for users,have been gaining momentum only very recently. In this work,we aim to learn personalized longitudinal driving behaviors viaa Gaussian Process (GP) model. The proposed method learnsfrom individual driver’s naturalistic car-following behavior, andoutputs a desired acceleration profile that suits the driver’s pref-erence. The learned model, together with a predictive safety filterthat prevents rear-end collision, is used as a personalized adaptivecruise control (PACC) system. Numerical experiments show thatGP-based PACC (GP-PACC) can almost exactly reproduce thedriving styles of an intelligent driver model. Additionally, GP-PACC is further validated by human-in-the-loop experiments onthe Unity game engine-based driving simulator. Trips driven byGP-PACC and two other baseline ACC algorithms with driveroverride rates are recorded and compared. Results show thaton average, GP-PACC reduces the human override duration by60% and 85% as compared to two widely-used ACC models,respectively, which shows the great potential of GP-PACC inimproving driving comfort and overall user experience.
... Although there are several examples of cybersecurity training [28], personalization enabling the customization of training materials seems to be a perspective approach. Personalization in information systems research was initially focused on customer relationships in order to improve the user experience [10]. Personalization approaches for enriching the learning process with the help of information systems later emerged which enabled the development of skills process through elearning [20]. ...
Full-text available
Information Systems (IS) represent an integral part of our lives, both in the organizational and personal sphere. To use them securely, users must be properly trained. The main problem is that most training processes still use the one-size-fits-all approach where users receive the same kind of learning material. In addition, personalized training may be a more suitable approach however a comprehensive process for IS user profiling and personalized IS user training improvement has not been introduced yet. This paper proposes a novel approach for personalized user training for secure use of IS to fill in this gap. The proposed approach focuses on three key dimensions (i.e., the personalization process, selection of training tools and materials, and participants) and is composed of five phases covering the identification of key IS security elements, IS user profiling and personalization of IS security training. It is scalable to all company sizes and aims to lower both the IS training costs and optimization of outcomes. As a side-effect, it also helps to lower user resistance to participation in IS security training.
User interface design is one of the most important and one of the most difficult aspects of designing a computer system. It is the contact point between the user and the system and determines to a large extent the usefulness and effectiveness of the system. In this paper, we examine the tools and techniques used for designing user interfaces. As user interface design is to a large extent an art, our goal is to highlight important issues in user interface design and not to prescribe a recipe for designing user interfaces.
Service marketers are confronted with two conflicting goals when designing service delivery systems, efficiency and personalization. The relative importance of each factor is determined by the nature of the specific service to be rendered, and by participants' expectations about degree of personalization. A study was conducted to test two assertions: (1) service personalization is a multidimensional construct and (2) all forms of personalization do not necessarily result in greater consumer satisfaction with the service offering. Three types of personalization strategies were proposed and operationalized in a simulated banking setting. Evaluations of service encounters that differed in the degree and type of personalization employed indicate that personalization is not a unitary phenomenon and must be approached carefully in the context of service design.
Recognizing patterns of behavior helps systems predict your next move.