Vous êtes sur la page 1sur 10

Mass Customization in the Newspaper Industry:

Consumers Attitudes Toward Individualized Media Innovations


D. Schoder,
Mass
Customization
S. Sick, J. in
Putzke,
the Newspaper
and A. M. Industry
Kaplan

Detlef Schoder, Stefan Sick, Johannes Putzke, and Andreas M. Kaplan


University of Cologne, Germany

This article applies the concept of mass customization to the newspaper industry. Although the theory of
mass customization has received considerable attention in recent years, its application to the printed mass
media market has been almost totally neglected. Researchers have not provided any empirical evidence of a
substantial market for mass customized printed newspapers, and we do not know much about customer attitudes toward such innovations. This article contributes to research on both issues. Based on an empirical
survey (n = 2,114), we examine consumer acceptance of an individualized newspaper. We address the 2 most
pressing issues associated with mass-customized products. These are consumer willingness to pay a premium and consumer willingness to devote additional effort to designing such a newspaper. The results,
based on conjoint analysis, suggest that consumers are generally willing to devote effort to customizing
their news. However, only those who are well-educated and belong to the upper socioeconomic strata are
willing to pay extra for individualized newspapers. When introducing mass-customized media, management should focus on these customer groups and their respective preferences.

Mass customization is a concept that has increasingly attracted the attention of both researchers and practitioners, especially over the last few years. According to Du,
Jiao, and Tseng (2003), more than 2,300 academic articles
have been published since Pines (1993) influential work
Mass-Customization: The New Frontier in Business Competition.
About 60% of these articles have appeared within the last
2 years.
Although the theory of mass customization has received considerable attention in recent years, its application to the printed mass media has been almost totally neglected. Furthermore, most of the work on mass
customization in the domains mentioned previously is
conceptual. The few empirical studies mainly approach
mass customization from a company perspective (e.g.
hlstrm & Westbrook, 1999; Feitzinger & Lee, 1997). Only
a minimal number of empirical studies take the consumer
perspective, and these studies deal mainly with perceptions of the codesign process (Franke & Piller, 2004;

Huffman & Kahn, 1998; Liechty, Ramaswamy, & Cohen,


2001).1
We conducted a conjoint analysis on the consumer acceptance of individualized daily newspapers as a prime
example of mass-customized media goods. In the study, a
panel of 2,114 respondents throughout Germany was
questioned by several interviewers in one-to-one interviews.
The remainder of this article is structured as follows.
The next section highlights mass customization from the
consumer perceptive. Individualized printed newspapers
are then introduced as archetypal examples of mass-customized media. Two research hypotheses are developed.
The next two sections describe the methodology and data
analysis. The final section discusses the results, their theoretical and managerial implications, and limitations and
possible directions for future research.

Consumer Acceptance
of Mass-Customized Media Goods
Address correspondence to Detlef Schoder, University of Cologne,
Department of Information Systems and Information Management,
Pohligstrasse 1, 50969 Cologne, Germany. E-mail: schoder@
wim.uni-koeln.de

The International Journal on Media Management, 8(1), 918

Mass customization can be described as a strategy which


creates value by some form of companycustomer interaction at the fabrication/assembly stage of the operations

level to create customized products with production cost


and monetary price similar to those of mass produced
products (Kaplan & Haenlein, 2006). This has become
both feasible and necessary, because technological evolution has lowered the minimum efficient scale of production to an individual level (Peppers & Rogers, 1997).
As early as 1970, before the term itself had been created, in Future Shock, Toffler (1970) described the principle
of mass customization. The term was not used until
1987, when Davis (1987) introduced it in his book Future
Perfect. However, only since Pine, Bart, and Boyntons
(1993) Harvard Business Review article on Pines famous
monograph (Pine, 1993) has greater attention been paid to
this highly relevant topic.
Interestingly, this strategy has so far been observed
mainly from the manufacturers perspective. Although
some scholars have commented that mass customization is
not the best strategy for all firms in all cases (Squire,
Readman, Brown, & Bessant, 2004; Zipkin, 2001), there
remains a lack of research on customer attitudes toward
mass-customization strategies. Therefore, Tu, Vonderembse, Ragu-Nathan, and Ragu-Nathan (2004) have
called for additional studies to deal with customer requirements before adopting a mass-customization strategy. Because there is almost no large-scale empirical research on
mass customization in general (Tseng & Piller, 2003), this
article makes a contribution not only in the field of media
management studies but also in related research areas
such as the management of information systems, marketing, or operations management.

Individualized Printed Newspapers


as Mass-Customized Media Goods
Pine, Peppers, and Rogers (1995) argue that anything
digitized (p. 108) can also be customized. In particular,
products that are purchased frequently (e.g., newspapers)
and reveal a discernible pattern of personal interest
would be ideally suited to customercompany interaction.
Consequently, several researchers cite personalized news
services as a perfect example of mass-customized products
(Franke & Schreier, 2002; Liechty et al., 2001; Pine et al.,
1995; Shapiro & Varian, 1999).
Currently, traditional newspapers must analyze, select, and synthesize raw information (e.g., news from
press agencies) for their readers (Iksal & Garlatti, 2002). In
a traditional newspaper, the relative importance of news
is determined by an editor. This is certainly not the best
way to serve clients with diverse interests (Lai, Liang, &
Ku, 2003, p. 225) and regular printed newspapers accordingly lose many readers (Iksal & Garlatti, 2002).
In contrast, individualized newspapers allow customers to tailor news according to their needs and preferences
by selecting and defining certain areas of interest. Thus,

10

the reader does not need to browse through unwanted information but obtains a newspaper that conforms
perfectly to his requirements (Franke & Schreier, 2002)
and that might even contain more information on topics
of interest than available in a traditional newspaper. As a
result, the time and effort invested in the initial design
process are compensated for by the time savings during
the news selection process. Thus, personalized information services simultaneously add value both for consumers and producers and could be traded as part of
value-added publishing strategies (Berghel, 1999). Even if
most researchers implicitly assume a Web-based environment (Ihlstrm & Palmer, 2002), customized printing is
now technically feasible at low cost (Pitta, 1998).
As an example, Pine et al. (1995) highlight the case of
Individual, Inc. and their First! service in which an editorial manager helps the client to determine his or her particular interests. Subsequently, an information retrieval
system takes over. Finally, the individually composed
news is delivered to the customer. Due to a constant rating
of articles by the customer, their perceived relevance rises
from 40% to 60% in the 1st week to 80% to 90% in the 4th
week.
Nevertheless, Sunstein (2001) and Collins and Butler
(2003) made critical statements about individualized
newspapers, their dangers, and negative implications for
society. The problem is that readers can limit themselves
to certain very specific points of view. These critical sentiments thus argue that the freedom to choose and customize news messages has undesired welfare consequences.
Furthermore, it is almost impossible to discover new areas
of interest by browsing through the newspaper: If all my
information is personalized, how will I ever learn anything new? (Silverstone, 1999, p. 20). This point of view is
supported by Tewksbury and Althaus (2000) who, in a
study of university students, found those who read the online edition of the New York Times read fewer international
and national political reports than readers of the traditional printed version.
Several researchers deal with news personalization
techniques. Mizzaro and Tasso (2002) provide a good overview of these techniques. In addition, applying the mass
customization paradigm, Ritz (2000) developed a formal
modeling method describing the production of personalized information services from both abstract and modular
perspectives.

Research Hypotheses
The development of research hypotheses takes into account the two most important drawbacks of mass-customized products that were identified through a literature review: consumer willingness to pay more and consumer
willingness to devote effort to designing the product.

D. Schoder, S. Sick, J. Putzke, and A. M. Kaplan

Tu, Vonderembse, and Ragu-Nathan (2001) define


mass-customized products as products that meet specific
customer needs at close to mass-production prices (p.
202). In such a case, close means that even if higher prices
are charged, customization does not change the nature of
market segments as in traditional customization approaches that entail additional costs (Piller, Moeslein, &
Stotko, 2004).
Nevertheless, several researchers (Kotha, 1995; Peppers & Rogers, 1997) argue that most mass-customized
products are somewhat more expensive than standardized products. The cost drivers of mass-customized goods
are also described in detail in the literature (Zipkin,
2001). Such costs include higher capital investment due
to higher setup costs and investment in information systems and flexible production units, higher labor costs
due to the need for better qualified staff, quality control,
and increased complexity in production planning and
control (Piller et al., 2004). Although a mass-customized
production system may increase manufacturing costs,
these should, nonetheless, be far below the value added
for the customer (Murakoshi, 1994).
Examining the cases of Mathushita Electric Company,
Washington Shoes, and Pola Cosmetics, Murakoshi (1994)
illustrated this fact by investigating the shift of Japanese
companies from conventional manufacturing strategies
to individual made-to-order systems. Therefore, some researchers suggest that customers are willing to pay a price
premium for customized products, because they perceive
higher value due to a more appropriate product specification in terms of their needs than the best standard product available (Peppers & Rogers, 1997; Pine, 1993; Pine et
al., 1995).
On the other hand, some researchers claim that consumer willingness to pay may even decrease, or they at
least doubt the assumption of consumer willingness to
pay more (e.g., McCutcheon, Raturi, & Meredith, 1994;
Zipkin, 2001). So far, there does not seem to be any convincing empirical evidence to support either the latter
point of view or the converse.
An experiment (Franke & Piller, 2004) and a
small-scale empirical study (Bardakci & Whitelock,
2004) addressing consumer willingness to pay for
mass-customized products provide additional insight.
Although both studies make a valuable contribution to
this area, their explanatory power is limited, because
the applicability of the results relating to the products
(watches and automobiles) to media industries is doubtful. Also the applicability of the results of a small-scale
study (N = 75) by Ihlstrm and Palmer (2002), dealing
with the consumers willingness to pay for personalized
online news services, to the printed newspaper market
and mass-customized goods in general is questionable.
Hence, additional research is needed to test the following hypothesis:

Mass Customization in the Newspaper Industry

H1. Customers value customized products. Thus, in


the purchasing decision, the type of daily newspaper (regular or individualized) is more important
than the price.
If consumers are willing to pay a premium price for a
product, they must also devote time and energy to specifying their product requirements. Urban and von Hippel
(1988) state that the more benefit a user expects to gain
from a new product, the greater the effort he or she will
be willing to devote to finding an appropriate solution.
For that reason, the next relevant question of interest is
whether the customer is really willing to commit to that
effort.
From the customer perspective, the effort involved in
obtaining a mass-customized product consists of both
time effort and cognitive effort. In the literature, the latter is often discussed under the buzzword mass confusion. This concept refers to consumer information overload when confronted with too many choices during the
decision-making process. Cognitive effort has indeed been
examined by many researchers (e.g., Huffman & Kahn,
1998). Although several studies suggest that consumers
are willing to spend time specifying their product requirements for high-involvement products such as clothes or
automobiles (e.g., Bardakci & Whitelock, 2004), there is no
evidence to suggest that customers would also be prepared make such an effort in designing other low-involvement products such as newspapers. In contrast, Manber,
Patel, and Robison (2000) state that the majority of active
My Yahoo! users do not customize their online news pages.
Consequently, the authors question the consumers willingness to personalize their news services and note that
the consumers might prefer to obtain the same news everyone else is getting. This suggests the following:
H2. Customers are willing to make an effort designing
their products (at least for the first transaction)
and have the capabilities to do so. Thus, the kind of
newspaper is more important than the effort necessary to specify the product requirements.

Method
Participants
Being part of a larger research project, the authors conducted a survey of individualized printed daily newspapers between January 30, 2004 and October 2, 2004.2 In
the study, several interviewers questioned a panel of 2,114
respondents in Germany in one-to-one interviews. In addition to 19 other research questions in the survey, the authors used a conjoint analysis to test the two hypotheses.

11

The decision to employ conjoint analysis was taken to


strengthen the validity and reliability of the empirical
study. By using conjoint analysis, there is less danger of
misinterpreting the attributes during the rating process
and a lower potential bias caused by social desirability
(Montgomery, 1986). Furthermore, conjoint analysis has a
greater range sensitivity than compositional methods
(Nitzsch & Weber, 1993).
By utilizing one-to-one interviews, good data quality
was obtained. Moreover, the participants taking part in
the study were selected in such a manner that the study
can validly be regarded as representative for Germany.
Before conducting the final conjoint analysis, the respondents were questioned (on a scale from 1 to 10) on the
likelihood that they would test read a customized printed
newspaper. If the respondents answered 3 or lower (n =
614), they were excluded from the conjoint analysis, because they did not belong to the potential readers of individualized newspapers.
From the 2,114 respondents, 51 (2.4%) made no statements at all about their trial probability; therefore, 665
respondents (31.5%) were excluded from further analysis
because they did not belong to the potential readers of an
individualized newspaper. The remaining 1,449 (68.5%)
respondents were included as a subsample in the conjoint
analysis.

With respect to the effort required, we wanted to allow


some personal interpretation, although effort had been
specified in a previous question of the general survey as (a)
selecting topics from a list, (b) communicating selected
topics by telephone, and (c) changing topics at any time by
telephone.
However, to assess the relative importance of a factor in
conjoint analysis, it is unimportant which of the two
available attribute levels is favored by a given respondent.
It is only necessary that there is a significant perceived difference between the two attribute levels (Schoder &
Haenlein, 2004).

Attributes and Attribute Levels

Data Collection

In conjoint analysis, respondent accuracy decreases as the


number of questions increases. Therefore, the authors created three super-attributes (Green & Srinivasan, 1978):
(a) type of newspaper, (b) price, and (c) effort. To avoid a
number-of-levels effect, which means that attributes defined at more levels than other attributes will be biased
upward in importance (Green & Srinivasan, 1990), each attribute comprised the same number of attribute levels. Because reliability declines with a greater number of attribute levels (Green & Srinivasan, 1978), their number was
kept low (two for each attribute). The final attributes and
their levels are depicted in Table 1.
Because soft categories were used for two of the attributes, we would like to comment on this decision: This research was part of a larger study about individualized
daily newspapers. To test the proposed hypotheses, a dichotomous variable for the type of newspaper was suffi-

This study followed the recommendation of Green and


Srinivasan (1978, 1990) to select the data-collection
method depending on the number of attributes. The
study used a full factorial design consisting of 23 stimuli.
Advantages of the full profile method include a high level
of task realism and thus higher reliability and validity
(Malhotra, 1982; Segal, 1982; Wind & Spitz, 1976). Furthermore, it is highly suitable for market segmentation methods and superior in assessing the relative importance of
attributes.
Nevertheless, to avoid any cognitive burden on the respondents (which was evident in pretesting), the following stimulus representation mode and ranking procedure
was used. The interviewer presented a pack of eight stimulus cards to respondents and asked them to divide it into
two packs of four cards. After allowing an appropriate
amount of time for consideration, respondents were

Table 1.
Level
#1
#2

12

cient, because respondents were familiar with the product characteristics of individualized (as well as regular)
newspapers. Additionally, the simple operationalization
reduces the cognitive burden on the respondent. The following text, which specifies individualized newspapers,
was presented to each interviewee (in German):
Assume it would be possible to get an individualized printed newspaper fully adapted to your personal preferences. For example, if
you are particularly interested in politics, sports, or science, the
newspaper would contain mainly articles from these areas. The
newspaper would therefore be adapted to your personal preferences
in terms of content and size.

Attributes and Their Levels

Type of Newspaper

Price

Regular newspaper
Individualized newspaper

1.00 Euro
1.40 Euro

Effort
Low
Slightly greater

D. Schoder, S. Sick, J. Putzke, and A. M. Kaplan

asked to arrange the single cards within the two packs according to their preferences.
In the study, a ranking procedure was applied rather
than a rating procedure, because the former is more realistic, reliable, and valid (Leigh, MacKay, & Summers, 1981;
Russel & Gray, 1994). A further disadvantage of the rating
method is the danger of obtaining too few discriminating
judgments due to the absence of forced-choice simulation
(Russel & Gray, 1994) and, therefore, a lack of balance between the stimuli. Because the stimuli were presented verbally, particular attention was paid to the order of attributes on the stimulus card, because this could influence
the evaluation (Kumar & Gaeth, 1991).

Estimation of the Partworths


Ordinary least square (OLS) was used to estimate the
partworths.3 If nonmetric data are collected, OLS is theoretically inappropriate, because it assumes equidistant
rankings and homogenous variance of the residuals. Nevertheless, several researchers have pointed out that OLS
leads to the same results as nonnonmetric procedures
(Darmon & Rouzies, 1994; Huber, 1975; Wittink & Cattin,
1981). Furthermore, the preference rating scale quality is
not a mathematical requirement of the OLS procedure
and must only be taken into account to interpret the results. In this context, the analyst must distinguish between mathematical interpretation problems and the assumptions underlying the mathematical method.
Therefore, a program using OLS is appropriate.
In this study, it may be best not to analyze the individual partworths obtained but rather the preferences of customer groups. To analyze the latter, the individual
partworths were aggregated by rescaling

lsaia =

laia - min ( laia )


la

(
)
l
ia
l
ia
max
a
a
min

la

l
a=1
a
A

lsaia = (re)scaled partworth of level l of attribute a for respondent i and averaging them.
The partworths so obtained were then used to assess
the importance of the factors. If all levels of an attribute
have similar partworths, this factor is of little relative (but
possibly great absolute) importance, because the attribute
levels do not influence the respondents preference. Thus,
it was reasonable to compute the relative importance of
each factor as follows. The range of partworth for this factor with respect to value was divided by the sum of all
range values. In this context, the relative importance of

Mass Customization in the Newspaper Industry

the factors can be interpreted as a quantitative measure


for the impact of each factor on the respondents utility
function (in percent). The range of partworth for each factor reflects the variance in the utility function, when the
levels of a factor change. By dividing this range value
through the sum of all range values, the measure is standardized so that the relative importance of the prize, the
effort, and the type of newspaper adds up to one in total.

Results
From the 1,449 respondents included in the conjoint analysis, 382 made either no or incomplete preference
revelations, so that the final conjoint estimation was
based on 1,067 (73.6%) data sets. To evaluate the cognitive
burden on respondents (and therefore the data quality),
the interviewers were requested to provide information
about the problems that the respondents encountered in
completing the task.
Because of the ranking procedure used, only 124
(11.6%) of the 1,067 respondents had substantial problems
completing the task, 397 (37.2%) some problems, 271
(25.4%) hardly any problems, and 268 (25.1%) no problems. The interviewers did not make any statements at all
with only 7 (0.1%) respondents. Consequently, these results indicate good data quality, particularly because the
interviewers were able to assist the respondents if any
problems occurred.
SPSS 12.0.0 was used for the final estimation. After
which, the estimation quality was assessed by means of
Pearsons R and Kendalls tau-b. Because the estimation of
partworths was conducted at the individual respondent
level, citing all the results for the single respondents was
impossible, even if all measures showed very high values
and thus high quality. Instead, we used Kendalls tau-b
and Pearsons R for the estimation at the level of the entire
sample, which led to the same results as the aggregation
of individual partworths. In this case, both Pearsons R
and Kendalls tau-b yielded maximum values (1.00/1.00),
indicating outstanding estimation quality.
Data analysis across the entire sample resulted in a relative importance for the purchasing decision of 34.6% for
type of newspaper, 36.4% for price, and 29.1% for effort.
Paired-samples t tests were used to compare the relative
importance of the three factors. The results of the first
paired-samples t test (t = 1.470, df = 1066, p = .14) revealed
no statistically significant difference between the relative
importance of the type of newspaper and the price.4 Thus,
H1 must be rejected, and it cannot be concluded that the
relative importance of the type of newspaper exceeds the
relative importance of the price.
In contrast, the paired-samples t test employed to test
H2 (t = 4.947, df = 1066, p = .000) yielded strong evidence to

13

Table 2.
#

Paired Samples t Test-to-Test Assumption of the Superiority of Personalized Newspapers


Type of Newspaper

6
7

Individualized newspaper
Regular newspaper

support the hypothesis.5 Thus, it can be concluded that


the relative importance of the type of newspaper exceeds
the relative importance of the effort.
The second hypothesis assumed implicitly that people
who attach greater relative importance to the type of
newspaper than to the effort involved prefer the individualized version to the regular one. However, this fact has
not yet been proven. Therefore, a paired samples test was
conducted comparing the average ranks of the following
two newspapers. The results of the t test (t = 6.701, df =
1066, p = .000) provide very strong evidence to support the
implicit assumption (see Table 2). Thus, H2 is supported.
Because H1 could not be supported for the entire sample, the authors tested the hypotheses for demographic
subsamples of the population. For this purpose, a test
was conducted as to whether there was a significant statistical difference between the subsamples regarding the
relative importance of one of the three factors (type of
newspaper, price, and effort). If a significant difference
was found, (multiple) paired comparisons were conducted between the relative importance in support of H1
and H2 for each of the subsamples. By so doing, the authors took into consideration the accumulated alpha error by reducing the significance criterion to * = 1 (1
)1/c. For a large number of paired comparisons, * can
be computed alternatively by Bonferroni adjustment
which means *= /c.
The procedure of estimating partworths with conjoint
analysis and grouping the respondents according to external factors has been criticized by several researchers, because the external factors are often not associated with
the utility function (e.g., Moore, 1980). Thus, some

Table 3.

Price

Effort

1.00 Euro
1.00 Euro

Low
Low

researchers recommend a tandem approach (Green &


Krieger, 1991), which means estimating the partworths
and subsequent cluster analysis. However, as a result of
the dissimilarity of the optimization criteria of OLS
partworths estimation and cluster analysis, the tandem
approach is also problematic (DeSarbo, Wedel, Vriens, &
Ramaswamy, 1992). Consequently, a priori segmentation
provides a higher external validity. Accordingly, we decided to employ this procedure.
The employed tests did not indicate any difference with
respect to the relative importance of the three attributes
between the subsamples classified by sex, age, labor market status, urbanization type, number of household members, number of working (earning) household members,
net monthly household income, the net monthly income
of main wage-earners in the household, and federal state.
Differences were found between the subsamples classified by population of hometown, job level, and private
Internet connection. The subsequent tests to support H2
for these subsamples provided the same results as for the
entire sample. However, the subsequent paired-samples t
test did not support H1 for any of the subsamples.
H1 was supported for two subsamples. Data analysis
across educational achievement and strata revealed statistically significant differences between the three factors.
The tests of H2 for these subsamples provided the same results as for the entire sample. In the subsample classified
according to educational achievement, the results of the t
tests (see Table 3) support H1 for the subsample of postgraduates but not for any other subsample.
In the subsamples classified according to strata, the t
tests (see Table 4) support H1 for the upper middle class at

Data Analysis by Educational Achievement:


t Tests to Support H1

Subsample
8 years of school without any school
qualifications
Basic secondary school education
GCSE qualifications
General National Vocational Qualification
A-Levels
Postgraduate

t Value

Degrees of
Freedom

Significance

3.0968

22

0.0053

2.2254
1.7691
0.9046
0.1761
3.1159

267
442
60
121
149

0.0269
0.0776
0.3693
0.8605
0.0022

Note. GCSE = General Certificate of Secondary Education.

14

D. Schoder, S. Sick, J. Putzke, and A. M. Kaplan

Table 4.
Subsample
Upper class
Upper middle class
Lower middle class
Lower class

Data Analysis by Stratum: t Tests to Support H1

t Value

Degrees of Freedom

Significance (2-tailed)

0.4989
2.9996
2.7047
2.7934

16
269
657
121

0.6246
0.0030
0.0070
0.0061

a very high level of significance but not for any of the


other three strata.
Aggregating the relative importance of the factors by
averaging them leads to a high variance of the means. This
indicates a majority fallacy, because of the respondents
heterogeneous preference. However, this procedure is appropriate, because aggregating the participants into particular subsamples was based on sound theoretical considerations (Blaho, 2001). The use of a regression-based
estimation method for estimating the partworths also increases the error variance and thus the instability of the
estimated parameters (Green & Srinivasan, 1990).

Summary and Discussion


Data analysis across the entire sample led to a relative
importance of 34.6% for type of newspaper, 36.4% for
price, and 29.1% for effort. The subsequent t test supported H2 at a very high level of significance for the entire
sample. By contrast, H1 could only be supported for two
subsamples.
The results of H2 were as expected. Therefore, no further interpretation is required. The interpretation of H1 is
much more relevant and important. The general population is not willing to pay a price premium for individualized newspapers. These results correspond to those obtained by Ihlstrm and Palmer (2002) who found that only
20% of a population was willing to pay a price premium
but contradict those results obtained by Bardakci and
Whitelock (2004) who found 58.8% of the population was
willing to pay a premium. Franke and Piller (2004) also
found customers are willing to pay price premiums of almost 100%.
However, this is only superficially contradictory. A
closer look at Franke and Piller (2004) reveals that the
participants were students. The results consequently correspond to our research, which found that only welleducated people are willing to pay a price premium for
individualized daily newspapers. This can be explained
by the fact that most well-educated people are better off,
have a distinctive interest profile, and know what material corresponds to their needs. Consequently, the
cognitive burden of the codesign process mutates into a

Mass Customization in the Newspaper Industry

pride of authorship (Franke & Piller, 2004, p. 412)


experience.
The differences with respect to the study of Bardakci
and Whitelock (2004) must be attributed to the particular
products. The authors expect the pride of authorship to
be much higher in the case of automobiles than for newspapers. For many people, automobiles are a status symbol
and they are incomparably more expensive than newspapers. Future research should comprehensively examine
the relation between pride of authorship and willingness
to pay more.
This point of view is supported partly by the other results of our survey. People belonging to the upper middle class were found to be willing to pay a price premium for the customization of their newspapers.
Furthermore, the results also suggest that with an increasing sample size, H1 could be supported for members of the upper class. This can be explained by the fact
that a prestige orientation leads to an increasing individualization of demand. Future research is needed to verify
these results as well.
Recently, mass customization has emerged as a new
paradigm that is attracting substantial attention from
both academic researchers and managers from all
kinds of industries. Up to the present, there has been
little analysis of this new concept, because most people
only look at companies already using this new concept
successfully. Consequently, managers attempting to implement a mass-customizing strategy are concerned
only with technical or organizational aspects. Yet, before implementing a mass-customizing strategy, it is essential for management to know whether customers actually appreciate this new concept and are willing to
pay for it.
As already mentioned, H2 (type of newspaper is more
important than effort) could be supported at a very high
level of statistical significance across the entire sample.
Thus, manufacturers should not focus disproportionately
on technical or organizational issues as in the past (facilitating the design process) but deal with the possibly increasing complexity in order to lower cost. This becomes
all the more important, because customers seem unwilling to pay a price premium for individualized daily newspapers.

15

The results of the study did not support H1 across the


entire sample. However, H1 could be supported for some
subsamples: Postgraduates and the upper middle class are
willing to pay price premiums for individualized daily
newspapers. Thus, managers of media companies should
focus on these target groups when launching their personalized news services.
Although not all people seem willing to pay a price premium for individualized newspapers, mass-customized
media can, in principle, succeed. The data collected during the customization process allows for targeted marketing and thus increased advertising revenues. Hence, even
for customers who are not willing to pay a price premium,
mass-customized media services may become feasible. A
feasibility study of individualized daily newspapers remains an issue worthy of further research.
As with any empirical study using conjoint analysis,
this work is subject to limitations. However, most of the
limitations can be constrained.
First, conjoint analysis rests, inter alia, on two basic
assumptions: utility maximizing rational decision makers and completeness of relevant stimulus design attributes. In this study, we did not include any attribute addressing privacy issues as those mentioned by Pitta,
Franzack, and Laric (2003). Further research might also
address the issues of privacy and trust and test the proposed hypotheses for other product categories (e.g., personalized periodicals). When testing the hypotheses for
other product categories, particular attention should be
paid to the trade-off between price and customization. A
further study with different price levels (more than two)
would be particularly desirable. For other product categories, the willingness to wait for the customized product might also play an important role (e.g., Bardakci &
Whitelock, 2003, 2004).
Second, in interpreting the results, the reader has to
bear in mind that approximately 30% of the respondents
were excluded from the final conjoint analysis, because
they did not belong to the potential readers of individualized newspapers. Also Ihlstrm and Palmer (2002) found
approximately 20% of the population to be not at all interested in the personalization of their online newspaper.
Thus, the obtained results can be, strictly speaking, only
applied to 70% of the population.
Third, the analysis was limited to Germany and theories dealing with the selection of media suggest that the
particular country or culture is important in the context
of these decisions (McQuail, 2000). Although the survey
can be seen as representative for Germany, these results
cannot necessarily be generalized to other countries.
Nevertheless, the results can at least provide a useful indicator of consumer acceptance in other European countries, because the factors driving the news selection process might be similar for most European countries.

16

Moreover, with 323 newspapers per 1,000 inhabitants


over 14 years and an audience penetration of 76%, German newspaper coverage is close to the median for all
European countries. However, future research should
specifically examine the consumer acceptance of
mass-customized media in other countries and cultural
areas.
We hope that this research project will assist others in
conducting this kind of study and form the basis for substantial future research dealing with the mass customization or individualization of media services from the customer perspective.

Notes
1. An extensive literature review of the available (empirical)
work in the area of mass customization can be found in
Franke and Piller (2003).
2. The survey was conducted in cooperation with Institut fr
Demoskopie Allensbach (IFD).
3. For the advantages of regression based methods, see Hauser
and Rao (2003).
4. All the following statistical inferences are based on a level of
significance of = 5% unless otherwise specified.
5. A significance level of .000 means that the significance level
is less than .0005.

References
hlstrm, P., & Westbrook, R. (1999). Implications of mass customization for operations management. International Journal of
Operations and Production Management, 19, 262274.
Bardakci, A., & Whitelock, J. (2003). Mass-customization in marketing: The consumer perspective. Journal of Consumer Marketing, 20, 463479.
Bardakci, A., & Whitelock, J. (2004). How ready are customers
for mass customization? An exploratory investigation. European Journal of Marketing, 38, 13961416.
Berghel, H. (1999). Value-added publishing. Communications of the
ACM, 42(1), 1923.
Blaho, R. (2001). MassenindividualisierungErstellung integrativer
leistungen auf massenmrkten [Mass individualizationProducing integrative commodities for mass markets]. Bratislava,
Slovakia: Sevt Verlag.
Collins, N., & Butler, P. (2003). When marketing models clash
with democracy. Journal of Public Affairs, 3, 5262.
Darmon, R. Y., & Rouzies, D. (1994). Reliability and internal validity of conjoint estimated utility functions under error-free versus error-full conditions. International Journal of Research in Marketing, 11, 465477.
Davis, S. (1987). Future perfect. Reading, MA: Addison-Wesley.
DeSarbo, W. S., Wedel, M., Vriens, M., & Ramaswamy, V. (1992). Latent class metric conjoint analysis. Marketing Letters, 3, 273288.

D. Schoder, S. Sick, J. Putzke, and A. M. Kaplan

Du, X., Jiao, J., & Tseng, M. M. (2003). Identifying customer need
patterns for customization and personalization. Integrated
Manufacturing System, 14, 387396.
Feitzinger, E., & Lee, H. L. (1997). Mass customization at
Hewlett-Packard: The power of postponement. Harvard Business
Review, 75(1), 116121.
Franke, N., & Piller, F. T. (2003). Key research issues in user interaction with user toolkits in a mass customization system. International Journal of Technology Management, 26, 578599.
Franke, N., & Piller, F. T. (2004). Value creation by toolkits for user
innovation and design: The case of the watch market. Journal of
Product Innovation Management, 21, 401415.
Franke, N., & Schreier, M. (2002). Entrepreneurial opportunities
with toolkits for user innovation and design. International Journal on Media Management, 4, 225235.
Green, P. E., & Krieger, A. M. (1991). Segmenting markets with
conjoint analysis. Journal of Marketing, 55(1), 2031.
Green, P. E., & Srinivasan, V. (1978). Conjoint analysis in consumer research: Issues and outlook. Journal of Consumer Research, 5, 103123.
Green, P. E., & Srinivasan, V. (1990). Conjoint analysis in marketing: New developments with implications for research and
practice. Journal of Marketing, 54(10), 319.
Hauser, J. R., & Rao, V. (2003). Conjoint analysis, related modeling,
and applications. In J. Wind & P. E. Green (Eds.), Marketing research
and modeling: Progress and prospects (pp. 141168). Boston: Kluwer.
Huber, J. (1975). Predicting preferences on experimental bundles
of attributes: A comparison of models. Journal of Marketing Research, 12, 290297.
Huffman, C., & Kahn, B. (1998). Variety for sale: Mass customization or mass confusion? Journal of Retailing, 74, 491513.
Ihlstrm, C., & Palmer, J. (2002). Revenues for online newspapers:
Owner and user perceptions. Electronic Markets, 12, 228236.
Iksal, S., & Garlatti, S. (2002). Adaptive special reports for online
newspapers. In S. Mizzaro, & C. Tasso, (Eds.), Personalization techniques in electronic publishing on the web: Trends and perspectives proceedings of the AH2002 workshop on personalization techniques in electronic publishing (pp. 3144). Mlaga, Spain: Universidad de
Mlaga.
Kaplan, A. M., & Haenlein, M. (2006). Toward a parsimonious definition of traditional and electronic mass customization. Journal of Product Innovation Management, 23(2), 168182.
Kotha, S. (1995). Mass customization: Implementing the emerging paradigm for competitive advantage. Strategic Management
Journal, 16, 2142.
Kumar, V., & Gaeth, G. J. (1991). Attribute order and product familiarity effects in decision tasks using conjoint analysis. International Journal of Research in Marketing, 8, 113124.
Lai, H., Liang, T., & Ku, Y. C. (2003). Customized Internet news services based on customer profiles. ACM international conference
proceeding series. Proceedings of the 5th international conference on
electronic commerce (pp. 225229). Pittsburgh, PA: ACM.
Leigh, T. W., MacKay, D. B., & Summers, J. O. (1981). On alternative
experimental methods for conjoint analysis. Advances in Consumer Research, 8, 317322.
Liechty, J., Ramaswamy, V., & Cohen, S. (2001). Choice-menus for
mass customization: An experimental approach for analyzing
customer demand with an application to a web-based information service. Journal of Marketing Research, 38, 183196.

Mass Customization in the Newspaper Industry

Malhotra, N. K. (1982). Structural reliability and stability of


nonmetric conjoint analysis. Journal of Marketing Research, 19,
199207.
Manber, U., Patel, A., & Robison, J. (2000). Experience with personalization on Yahoo! Communications of the ACM, 43(8), 3539.
McCutcheon, D. M., Raturi, A. S., & Meredith, J. R. (1994). The customization responsiveness squeeze. Sloan Management Review,
35, 8999.
McQuail, D. (2000). Mass communication theory (4th ed.). London:
Sage.
Mizzaro, S., & Tasso, C. (Eds.). (2002). Personalization techniques in
electronic publishing on the web: Trends and perspectives. Proceedings
of the AH 2002 workshop on personalization techniques in electronic
publishing (pp. 3144). Mlaga, Spain: Universidad de Mlaga.
Montgomery, D. B. (1986). Conjoint calibration of the customer/
competitor interface in industrial markets. In K. Backhaus & D.
T. Wilson (Eds.), Industrial marketing: A German-American perspective (pp. 297319). Berlin: Springer.
Moore, W. L. (1980). Levels of aggregation in conjoint analysis:
An empirical comparison. Journal of Marketing Research, 17,
516523.
Murakoshi, T. (1994). Customer-driven manufacturing in Japan.
International Journal of Production Economics, 37, 6372.
Nitzsch, R. V., & Weber, M. (1993). The effect of attribute ranges on
weights in multiattribute utility measurements. Management
Science, 39, 937943.
Peppers, D., & Rogers, M. (1997). Enterprise one to one. New York:
Currency-Doubleday.
Piller, F. T., Moeslein, K., & Stotko, C. M. (2004). Does mass customization pay? An economic approach to evaluate customer integration. Production Planning & Control, 15, 435444.
Pine, B. J. (1993) Mass-customization: The new frontier in business competition. Boston: Harvard Business School Press.
Pine, B. J., Bart, V., & Boynton, A. (1993). Making mass customization work. Harvard Business Review, 71(5), 108118.
Pine, B. J., Peppers, D., & Rogers, M. (1995). Do you want to keep
your customers forever? Harvard Business Review, 73(2), 10314.
Pitta, D. A. (1998). Marketing one-to-one and its dependence on
knowledge discovery in databases. Journal of Consumer Marketing, 15, 468480.
Pitta, D. A., Franzack, F., & Laric, M. (2003). Privacy and one-to-one
marketing: Resolving the conflict. Journal of Consumer Marketing, 20, 616628.
Ritz, T. (2000). Personalized information servicesAn electronic
information commodity. e-Business Strategy Management, 2,
105114.
Russell, P. A., & Gray, C. D. (1994). Ranking or rating? Some data
and their implications for the measurement of evaluative response. British Journal of Psychology, 85, 7992.
Schoder, D., & Haenlein, M. (2004). The relative importance of different trust constructs for sellers in the online world. Electronic
Markets, 14, 4857.
Segal, M. N. (1982). Reliability of conjoint analysis: Contrasting
data collection procedures. Journal of Marketing Research, 19,
139143.
Shapiro, C., & Varian, H. R. (1999). Information rules: A strategic
guide to the network economy. Boston: Harvard Business School
Press.
Silverstone, R. (1999). Why study the media? London: Sage.

17

Squire, B., Readman, J., Brown, S., & Bessant, J. (2004). Mass customization: The key to customer value? Production Planning and
Control, 15, 459471.
Sunstein, C. (2001). Republic.com. Princeton, NJ: Princeton University Press.
Tewksbury, D., & Althaus, S. L. (2000). Differences in knowledge
acquisition among readers of the paper and online versions of
a national newspaper. Journalism and Mass Communication Quarterly, 77, 457479.
Toffler, A. (1970). Future shock. New York: Bantam Books.
Tseng, M. M., & Piller, F. T. (2003). New directions for mass customization. In M. Tseng & F. T. Piller (Eds.), The customer centric enterprise (pp. 519535). New York: Springer.
Tu, Q., Vonderembse, M. A., & Ragu-Nathan, T. S. (2001). The impact of time-based manufacturing practices on mass customization and value to customer. Journal of Operations Management,
19, 201217.

18

Tu, Q., Vonderembse, M. A., Ragu-Nathan, T. S., & Ragu-Nathan, B.


(2004). Measuring modularity-based manufacturing practices
and their impact on mass customization capability: A customer-driven perspective. Decision Sciences, 35, 147168.
Urban, G. L., & von Hippel, E. (1988). Lead user analyses for the development of new industrial products. Management Science, 34,
56982.
Wind, Y., & Spitz, L. (1976). Analytic approach to marketing decisions in health-care organizations. Operations Research, 24,
973990.
Wittink, D. R., & Cattin, P. (1981). Alternative estimation methods
for conjoint analysis: A Monte Carlo study. Journal of Marketing
Research, 18, 101106.
Zipkin, P. (2001). The limits of mass customization. Sloan Management Review, 42, 8187.

D. Schoder, S. Sick, J. Putzke, and A. M. Kaplan

Vous aimerez peut-être aussi