You are on page 1of 5

Journal of Business Research 65 (2012) 1290–1294

Contents lists available at SciVerse ScienceDirect

Journal of Business Research

Technological expectation and consumer preferences for product form


Seh-Woong Chung a, Jin K. Han a, Yong Seok Sohn b,⁎
a
LKC School of Business, Singapore Management University, 50 Stamford Road, #05-01, 178899, Singapore
b
School of Management, Kyung Hee University, Dongdaemun-Gu, Hoegi-Dong 1, Seoul, Republic of Korea

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

Article history: Increasingly, manufacturers address consumers' fear of product obsolescence by offering products in a mod-
Received 1 September 2010 ular form, hence allowing the consumer to replace or upgrade components as needed. The extant literature,
Received in revised form 1 January 2011 however, has not studied consumers' responses to product modularity. This study explores how consumers
Accepted 1 March 2011
weigh potential benefits (e.g., alleviate obsolescence concerns) vis-a-vis downsides (e.g., magnify complexity
Available online 16 November 2011
perceptions) to arrive at product-form preferences. The study further extends the literature by showing
Keywords:
product lifecycle stages and order-of-entry to moderate the impact of product modularity on consumer
Modularity preferences.
Obsolescence © 2011 Elsevier Inc. All rights reserved.
Technological expectation
Product complexity

1. Introduction modular PDA was somewhat transitory, thereby, leaving some key
questions behind. Specifically, do consumer preferences for product
platform evolve over time, and if so, what are the factors beyond ob-
Creating appropriate modular architectures to support new kinds
solescence issues that may drive the preferences for monolithic vs.
of product strategies is now central to business strategies. Busi-
modular platform?
nesses need to create product and process architectures that are
To this end, this research explores consumer preference at differ-
capable of providing the flexibility…for individuals and to up-
ent points along the technological trajectory and offers a contingency
grade them when better components come along. (Sanchez, 2000)
perspective to product-architecture strategy. The studies in this re-
search demonstrate that preference for product architecture evolves
As a rite of passage, all high-tech product consumers will, at some with technological trajectories and that choice decisions follow a
point, have deliberated over the new/obsolete paradox, pondering high-tech buying hierarchy. This research also explores order-of-
the inevitable fate that today's latest technologies will become out- entry effects to better understand the strategic potential for leverag-
moded in the not-so-distant future (Mick & Fournier, 1998). As the ing product architecture, and finally, discusses managerial implica-
modern-day consumer increasingly looks to high-tech products as tions and directions for future research.
lifestyle solutions—at work as well as for play (e.g., smart phones,
notebooks, personal media devices) (Gerson, 2007), the new/obso- 2. Conceptual development
lete paradox will remain a persistent dilemma from the buyer's
standpoint, which presents a challenge and also perhaps an opportu- 2.1. Scope of modular product architecture: Modularity-in-use
nity for firms in the consumer-technology sector.
In this context, a potential strategic tool for addressing aversion to Both practitioners and scholars alike acknowledge that a paradigm
obsolescence may underlie in the manufacturer's strategy on product shift is taking place in product development: from product creation
architecture (Brusoni & Fontana, 2005; Sanchez, 1999, 2000). A good towards modular platform development (Baldwin & Clark, 1997).
illustration is the success by the upstart PDA-maker Handspring, who According to Sanchez (1999), strategic benefits of modular architec-
adopted a modular architecture for its PDA (i.e., where users can ture may arise from its potential to “introduce technologically im-
readily interchange or upgrade components). This strategy is a stark proved products more rapidly, bring new products to market more
contrast to the incumbent Palm's monolithic platform, where all quickly, and lower costs of creation and realization.” Specifically,
parts are hardwired. Nonetheless, the appeal of Handspring's modularity in product development can occur at three levels:
modularity-in-design, modularity-in-production, and modularity-in-
⁎ Corresponding author. Tel.: + 82 2 961 2152; fax: + 82 2 961 0515.
use (Baldwin & Clark, 1997). To date, the literature focuses dispropor-
E-mail addresses: swchung@smu.edu.sg (S.-W. Chung), jkhan@smu.edu.sg tionate amount of attention on modularity-in-design and -produc-
(J.K. Han), ysohn@khu.ac.kr (Y.S. Sohn). tion, while giving scant attention to modularity-in-use — mainly

0148-2963/$ – see front matter © 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jbusres.2011.10.032
S.-W. Chung et al. / Journal of Business Research 65 (2012) 1290–1294 1291

attributable to the topic's engineering-/operations-based origin. performance drive the preferences and choices for higher-order
Nonetheless, as the demand-side (vis-à-vis supply-side) factors ulti- criterion.
mately determine the success or failure of any product or its platform, In the context of product architecture, the above studies show that
the motivation for this research is to bridge the gap in the knowledge the distinction in flexibility between the two platforms may also sig-
on modularity-in-use. This study, therefore, examines how the end- nal performance differences to consumers. Past works show that con-
users perceive the merits of modular architecture. sumers are cognizant of the technological trend of improving
performances with time, and accordingly, today's buying decisions
2.2. Modular vs. Monolithic architecture actually reflect future performance expectations (Bridges, Yim, &
Briesch, 1995; Kim, Han, & Srivastava, 2002). The implications are
Its flexibility is one of the most important advantages of modular two-fold. First, if the future technological expectation is high (i.e.,
architecture (Schilling & Steensma, 2001). Specifically, this open- steep technological trajectory), then consumers will prefer modular-
system architecture bestows the end-user with the flexibility to re- ity's flexibility allowing for performance upgrades vs. static monolith-
place and upgrade to an improved component in a “plug-and-play” ic structure. This scenario is likely to occur in the earlier part of the
style. In other words, consumers are buying into the idea of future- product's life-cycle when the industry's performance standard falls
proofing when they select a modular platform (McLaughlin, 2005). below the consumer's threshold level (e.g., imaging quality of early
Ulrich (1995) similarly underscores this point with upgrade opportu- digital cameras or speed of yesteryear's processor chips). In addition,
nities offered by modular computer systems that enable more power- with the performance criterion activated, the relative advantage of
ful chip enhancements. In contrast, the monolithic product monolithic platform on complexity perception is immaterial due to
architecture is a closed-system, where the entire product has to be its status as lower-order criteria (e.g., reliability and usability).
replaced in any upgrading decisions. The concept of future-proofing In sum, modular product architecture dominates the monolithic
is entirely inapplicable in the monolithic archetype. The overriding counterpart in choice for below threshold cases. H1a: In a choice con-
implication is that, in the introduction stage of the product life text between modular versus monolithic product platform, more con-
cycle, when a high-tech category is undergoing a rapid pace of im- sumers will prefer the modular platform at low levels of technological
provement in its core performance, the appeal of modular architec- performance for a common given feature.
ture should dominate that of monolithic archetype on the basis of Secondly, if the future technological expectation is low (i.e., flat
flexibility criterion. Brusoni and Fontana (2005) echo this point by technological trajectory), then consumers will prefer the less
pointing out that “companies wanting to emphasize…flexibility and seemingly-complex monolithic platform for its reliability and usabil-
upgradability may well choose a modular architecture.” ity dimensions vs. the modular option. This scenario is likely to
The flipside to modularity's advantage may, however, serve as its occur as the technology approaches mature phase of the product's
Achilles heel. Intrinsic to modular architecture is its platform's inter- life-cycle when the industry's performance standard has satisfied
face with modular components. This very structural element that the consumer's threshold level (e.g., present-day imaging quality of
gives rise to flexibility may also serve to raise the perceived complex- digital cameras or current standard speed of computer processor
ity of the product (Ernst, 2004). Past studies show that consumers chips). Performance has become what Sanchez (1999) refers to as
tend to associate higher levels of product complexity in technology “threshold attributes” in this phase. Threshold attributes are func-
products with lower perceptions of product reliability (e.g., Leek & tions provided by components that are essential to the overall func-
Kun, 2006) as well as degradation in perceived usability (e.g., tioning of a product but are not perceived by consumers as adding
Liljegren, 2006; Mukherjee & Hoyer, 2001). That is, for complex prod- value to the product as long as they function adequately. In essence,
ucts, consumers tend to infer high likelihood and frequency of fail- components that provide threshold attributes have the strategic role
ures. Complexity is also detrimental on the ease-of-use judgment— of functioning at a threshold level of acceptable performance and
with inferred learning costs as a key factor. Altogether, given that reliability.
modularity's flexibility enhances the platforms appeal whereas com- With performance criterion no longer serving a discriminating
plexity the opposite, what would the net impact be, on balance role, consumers will now apply the lower-order criteria (reliability
(refer to Table 1)? To shed light on this issue, the next section dis- and usability) in forming their preferences for one type of product
cusses Christensen (2000)'s high-tech buying hierarchy. platform over another, and thus, prefer the less seemingly-complex
monolithic platform (hence, seemingly-superior in reliability and us-
2.3. High-tech buying hierarchy and technological trajectory ability) to the modular platform.
In sum, monolithic product architecture dominates in the choice
According to Christensen (2000), consumers typically make buy- over the modular counterpart for above the threshold cases. H1b: In
ing decisions on high-tech products by following a buying hierarchy, a choice context between modular vs. monolithic product platform,
with performance criterion at the top, followed by reliability and us- more consumers will prefer the monolithic platform at high levels
ability considerations, respectively. To elaborate, performance differ- of technological performance for a common given feature.
ences among the alternatives would dominate the choice decision. If
this dimension does not play a discriminating role, then reliability 3. Experiment 1
subsequently becomes the next choice criterion, followed by the
lower-order criterion of usability. A recent study by Han, Chung, 3.1. Method
and Sohn (2009) on product-form choice lend support to
Christensen (2000)'s model of high-tech buying hierarchy. They cor- Experiment 1 explores consumers' choices between the monolith-
roborate the hierarchy effects in high-tech choice decision and fur- ic and the modular product forms (H1a and H1b). One hundred twen-
ther show that consumers' perceived vs. objective differences in ty one business students participated in this experiment in exchange
for a course credit. The participants received questionnaires contain-
Table 1 ing information two options of product forms. The first option was a
Product architecture: flexibility vs. complexity. purchase of a PDA integrated with a digital camera as a ‘two-in-one’
product (i.e., monolithic model). The second option was a PDA with
Flexibility Complexity
a digital camera module that the user can plug into the PDA (i.e.,
Modular + −
modular model). The subjects' task was to indicate which of the two
Monolithic − +
models they would choose.
1292 S.-W. Chung et al. / Journal of Business Research 65 (2012) 1290–1294

The between-subject experimental design consists of two condi-


tions (low-performance vs. high-performance). The dependent vari-
able is subjects' product-form choice between the two models
(monolithic vs. modular). The low-performance condition employs
1-megapixel resolution for both the integrated and the modular prod-
uct forms. The high-performance condition uses a 5-megapixel reso-
lution. The 1-megapixel level represented the lower-end of digital
camera technology at the time of the data collection in 2004. The
5-megapixel reflected the higher-end of technological standards.
According to an industry analyst, “the sweet spot for mass sales of
digital cameras last year was 3 to 4 megapixels, and this year it will
likely rise again” (Hughlett, 2005). The industry standard for standa-
lone digital camera is generally higher than the functional equivalent
feature in the camera phone. This procedure is also consistent with
Han et al. (2009)'s study in setting low vs. high performance level
for digital camera category. This study demonstrates that the perfor-
mance reservation value is indeed higher for the digital camera
Fig. 1. Platform choice percentage by performance level.
(about 3 megapixel) than for the digital camera function in camera
phones (about 1.5 megapixel). The price level is equivalent across
the two product forms To summarize, the results of Experiment 1 indicate that the tech-
nological performance level of the given feature moderates con-
sumers' choice of product form between the monolithic vs. modular
3.2. Measurement
platform. When the performance level is below the consumers' per-
formance reservation value, consumers seek flexibility, and thereby,
On the second page of the questionnaire, subjects provide, in writ-
choose the modular form. However, when the performance level is
ing, the reasons for their choices, and provide the performance reser-
above the threshold value, flexibility for future performance is no lon-
vation value for the digital camera function. Specifically, they indicate
ger important. Hence, with the modular platform appearing relatively
the minimum acceptable level of picture resolution on a seven-point
more complex, the monolithic option becomes the more preferred
scale ranging from 1-megapixel to 7-megapixel. Finally, subjects an-
option due to higher perceived convenience, ease-of-use, and/or reli-
swer questions measuring product perceptions on the dimensions
ability—all arising from relatively less-complex product perceptions.
of flexibility and complexity.
As aforementioned in the PDA market, Palm (the incumbent) had
adopted the monolithic platform whereas Handspring (the new en-
3.3. Results and discussion trant) had chosen the modular platform. Although Handspring initial-
ly experienced substantial success, Palm recaptured the dominance
The performance reservation value serves as a manipulation later in the product lifecycle. These findings, in the context of
check. Lending support to the manipulation, the average performance Handspring's strategy, suggest that new entrants may potentially
reservation value is 3.3. This validates the use of 1-megapixel level overcome pioneering advantage if they adopt a contingent product-
and 5-megapixel level as low-performance condition and high- architecture strategy. The next study addresses this very issue.
performance condition, respectively. H1a and H1b posit that more
consumers would choose the modular (monolithic) product form in
the low-performance (high-performance) condition. In support of 4. Experiment 2
these hypotheses, subjects are significantly more likely to choose
the modular model in the low-performance condition than in the By sheer order-of-entry, pioneers in new markets often enjoy in-
high-performance condition. Eighty-four percent of the subjects (51 herent advantages over subsequent entrants in a variety of ways
of 61 subjects) prefer the modular model in the low-performance (e.g., established preferences, supra-normal returns, market share,
condition compared with forty-eight percent (29 of 60 subjects) in customer loyalty, quality perceptions, preemption of resources). The
the high-performance condition (χ 2 = 14.60, p = .0001)(Fig. 1). literature offers numerous examples of the pioneer advantage across
Analyses of consumers' product perceptions on flexibility and multiple disciplines including economics, management, and market-
complexity dimensions provide insight on the underlying rationale ing (for reviews, see Kalyanaram, Robinson, & Urban, 1995; Kerin,
for the revealed preferences. On the flexibility dimension, subjects Varadarajan, & Peterson, 1992; Lieberman & Montgomery, 1988,
rate modular option to be significantly higher than the monolithic 1991). The incumbent advantage is persistent and difficult to over-
counterpart (mean = 4. 7 vs. 2.8 for modular and monolithic options, come, which leaves a challenging proposition for coming into the
respectively; t = 11.08, p = .000); on the complexity dimension, sub- market henceforth. Although the findings from Experiment 1 support
jects rate modular option relatively higher as well (mean = 4. 6 vs. 2.9 the notion that a firm should consider contingent product-
for modular and monolithic options, respectively; t = 8.10, p = .000). architecture strategy (modular and monolithic) as technological per-
The written open-ended responses for the choice further corroborate formance trajectory evolves higher over time, whether such a
the premise that the tradeoff between the complexity-related issues product-platform strategy can overcome pioneering advantage re-
and the future flexibility is the driving factor for the observed choice mains to be seen.
pattern. Nearly all subjects who choose the monolithic option men- To this end, this study explores the contingencies of product-
tion convenience, ease-of-use, or reliability as the reason for the platform strategy (monolithic vs. modular) by a new entrant to gain
choice. Typically, these responses include statements such as: ‘easier access to the market that is dominated by the incumbent. For this
to carry,’ ‘there is less hassle,’ ‘easy to use,’ etc. In contrast, subjects purpose, the next experiment examines two situations; when the in-
who choose the modular product option list the flexibility as the cumbent is currently using (1) a monolithic strategy and (2) a modu-
most important reason: For example, ‘expand in the future,’ ‘techno- lar strategy. The new entrant seeking to differentiate itself would
logical advancement is rapid these days,’ ‘transfer to another module then adopt a modular strategy and a monolithic strategy in these
later,’ ‘can detach the module when not needed,’ etc. two conditions, respectively.
S.-W. Chung et al. / Journal of Business Research 65 (2012) 1290–1294 1293

The product-form differentiation strategy of the new entrant like- Prior literature across many disciplines presents ample evidence
ly will be effective only to the extent that the strategy matches con- that the pioneering advantage is a strong, persistent phenomenon.
sumers' preference pattern observed in Experiment 1. In specific, a The results of this experiment show that the new entrant may over-
modular strategy will be efficient only when the incumbent is offer- come this seemingly insurmountable obstacle by adopting the
ing an integrated product at the low level of technological perfor- product-form differentiation strategy. More importantly, any firm
mance. In contrast, when the incumbent is using the modular adopting the product-form differentiation will have to tailor its ef-
strategy, the new entrant may successfully differentiate itself using forts by considering both the incumbent strategy and the technolog-
the monolithic strategy provided the technological performance is ical performance of the feature.
at a high level. H2: When the incumbent is using the monolithic
(modular) strategy, the modular (monolithic) strategy is a successful 5. Summary and discussion
product platform strategy when the technological performance of a
given feature is low (high). The objective of this research is to provide the initial groundwork
and motivation for studying the product-platform preferences in the
context of technological evolution. Results of experiments reveal im-
4.1. Method
portant regularities in consumer preferences. Most importantly, con-
sumers prefer modular product platform over monolithic ones when
The sample consists of one hundred one business students. Similar
the level of technological performance is relative low; however, the
to Experiment 1, the questionnaire contains information about two
pattern of preference reverses at high performance levels. These re-
companies (one company the incumbent, and the other a new en-
sults are due to the fact that consumers view modular vs. monolithic
trant) respectively offering a PDA integrated with a digital camera
platform as a trade-off between flexibility and complexity.
and a PDA that comes with a digital camera module. The experiment
One theoretical account for the observed patterns of results is the
consists of 2 (differentiation: modular vs. integrated) by 2 (perfor-
performance reservation value explanation. When the technological
mance: low vs. high) between-subjects design. In the modular differ-
performance level is below the performance reservation value, con-
entiation condition, the cover story describes that the new entrant
sumers see added value in the flexibility provided by the modular
offers a modular PDA in response to the incumbent company offering
product. As the performance level begins to exceed consumers' reser-
an integrated PDA. Conversely, the description in the monolithic dif-
vation value, the complexity (inclusive of convenience, usability, reli-
ferentiation condition depicts a situation where the new entrant of-
ability) factor outweighs the concern for the flexibility, and therefore,
fers a monolithic PDA in response to the incumbent selling a
leads to the preference for the monolithic form.
modular PDA. The manipulation of the performance remain the
These results expand the extant understanding of the consumer
same as that used in Experiment 1; namely, 1-megapixel vs. 5-
decision making processes in high-tech product purchases. The
megapixel.
high-tech buying hierarchy model (Christensen, 2000) stipulates
that consumers first examine the performance criteria when making
4.2. Results and discussion choice decisions. The findings of the current study extend this
model by showing that consumers not only consider the current per-
The results are consistent with H2. In the ‘modular differentiation’ formance differences among the alternatives, but also take into ac-
condition, sixty-two percent of the subjects (31 of 50 subjects) prefer count future performance expectations.
the new-entrant's product when the technological performance is Furthermore, results also indicate that the ‘differentiation-by-
low. However, when the technological performance is high, only product-form’ may provide an effective product strategy for new en-
twenty-eight percent of the subjects (14 of 50 subjects) prefer the trants to the technology market; however, any firm using the product
new-entrant's product (χ2 = 11.09, p = .0009). The choice patterns form as a market-entry strategy must consider the performance res-
are reverse in the ‘monolithic differentiation’ condition. Seventy-one ervation value held by consumers and the performance level of the
percent of the subjects (36 of 50 subject) prefer the new entrant's given technology. In short, provided that the incumbent is currently
product in the ‘high performance’ condition, while only sixteen per- offering a monolithic product, the modular product strategy is appro-
cent (8 of 51 subjects) do so in the ‘low performance’ condition priate for the new entrant only in low technological performance
(χ 2 = 28.12, p b .0001) (Fig. 2). conditions. In contrast, if the incumbent presently has a modular

Fig. 2. Platform choice percentage: New entrant strategy and performance level.
1294 S.-W. Chung et al. / Journal of Business Research 65 (2012) 1290–1294

product, the new entrant can use the integrated strategy only in high Christensen Clayton M. The innovator's dilemma. Boston, MA: Harvard Business School
Press; 2000.
technological situations. Ernst Dieter. Searching for a new role in East Asian regionlization: Japanese production
A number of issues still remain for future research. One issue is networks in the electronics industry. Economics Study Area Working Papers, 68.
that this study used a single digital component—PDA with a digital East–West Center: Economics Study Area; 2004.
Gerson Bob. CEA forecasts $160 billion in CE sales for 2007–08. TWICE: This Week in
camera function. Future work may include multiple components Consumer Electronics 2007;22(24):56.
(e.g., digital camera, MP3 player, GPS, etc.) to examine whether/ Han Jin K, Chung Seh Woong, Sohn Yong Seok. Technology convergence: when do con-
how the additional components shift the preference patterns. sumers prefer converged products to dedicated products? Journal of Marketing
2009;73(4):97-108.
By manipulating the level of technological performance of the dig- Hughlett Mike. Rise of camera phone builds synergy between industries. Chicago Tri-
ital camera function, the current study shows that the preference for bune, Business News 2005:1. March 18.
product platforms would depend on the stage of the product life cycle Kalyanaram Gurumurthy, Robinson William T, Urban Glen L. Order of market entry:
established empirical generalizations, emerging empirical generalizations, and fu-
(i.e., emerging vs. mature) as the technological performance of the
ture research. Marketing Science 1995;14(3):G212–21. Part 2.
function would improve along the product life cycle. These findings Kerin Roger A, Varadarajan P Rajan, Peterson Robert A. First-mover ddvantage: a syn-
are in line with prior literature (e.g., Han et al., 2009) which shows thesis, conceptual framework, and research propositions. Journal of Marketing
that consumers form technological expectations on the digital- 1992;56:33–52. October.
Kim Namwoon, Han Jin K, Srivastava Rajendra K. A dynamic IT adoption model for the
camera technology for camera phones at all levels of technology — SOHO market: PC generational decisions with technological expectations. Manage-
even though the standalone digital camera technology possessed a ment Science 2002;48(2):222–40.
higher performance upperbound. Future studies may extend the find- Leek S, Kun D. Consumer confusion in the Chinese personal computer market. The Jour-
nal of Product and Brand Management 2006;15(3):184–93.
ings by replicating the results with separate technological products in Lieberman Marvin B, Montgomery David B. First-mover advantages. Strategic Manage-
different stages of the product life cycle. One would expect that the ment Journal 1988;9:41–58. Summer.
preference for modularity would be the greatest in products that Lieberman Marvin B, Montgomery David B. Strategy of market entry: to pioneer or fol-
low? In: Glass Harold E, editor. The Handbook of Business Strategy. Boston, MA:
are in emerging stages of the product life cycle, and vary across prod- Warren, Gorham, and Lamont; 1991.
uct categories in later stages of the product life cycle. Liljegren E. Usability in a medical technology context assessment of methods for us-
ability evaluation of medical equipment. International Journal of Industrial Ergo-
nomics 2006;36(4):345–52.
Acknowledgments McLaughlin Laurianne. Future-proof your network. http://www.networkworld.com/
research/2005/071805-future-proofing.html2005.
YongSeok Sohn acknowledges the generous support of the Nation- Mick David Glen, Fournier Susan. Paradoxes of technology: consumer cognizance,
emotions, and coping strategies. Journal of Consumer Research 1998;25:123–43.
al Research Foundation in Korea (NRF-B00237), and Jin K. Han would
September.
like to thank Kyung Hee University's International Scholar research Mukherjee Ashesh, Hoyer Wayne D. The effect of novel attributes on product evalua-
support. The authors appreciate helpful comments from Dae Ryun tion. The Journal of Consumer Research 2001;28(3):462–72.
Chang and Shijin Yoo on the earlier version of the manuscript. Sanchez Ron. Modular architectures in the marketing process. Journal of Marketing
1999;63(Special Issue):92-111.
Sanchez Ron. Modularity: upgrading to the next generation design architecture. http://
References www.connected.org/media/modular.html2000.
Schilling Melissa A, Steensma H Kevin. The use of modular organizational forms: an
Baldwin Carliss Y, Clark Kim B. Managing in an age of modularity. Harvard Business Re- industry-level analysis. Academy of Management Journal 2001;44(6):1149–68.
view 1997:84–93. September. Ulrich Karl. The role of product architecture in the manufacturing firm. Research Policy
Bridges Eileen, Yim Chi Kin, Briesch RA. A high-tech product market share model with 1995;24:419–40. May.
customer expectations. Marketing Science 1995;14:61–81. Winter.
Brusoni S, Fontana R. Modularity as an entry strategy: the invasion of new niches in the
LAN equipment industry. CESPRI Working Papers, 171. Milano, Italy: Universita'
Bocconi; 2005.. July.

You might also like