This is the html version of the file http://ecommerce.mit.edu/papers/ERF/ERF250.pdf.
G o o g l e automatically generates html versions of documents as we crawl the web.
To link to or bookmark this page, use the following url: http://www.google.com/search?q=cache:Q6Lz6qcpALMJ:ecommerce.mit.edu/papers/ERF/ERF250.pdf+ghose+%22personalized+pricing%22+filetype:pdf&hl=en&ct=clnk&cd=26&gl=us

Google is neither affiliated with the authors of this page nor responsible for its content.
These search terms have been highlighted: ghose personalized pricing 
Internet Enabled Arbitrage
Page 1
Why Pay More? Why Charge Less?
An Empirical Exploration of International Online Price
Differentials
Xue Bai
Heinz School of Public Policy and Management
Carnegie Mellon University, Pittsburgh, PA 15213
xbai@andrew.cmu.edu
Abstract
This paper addresses an emerging phenomenon in the international online book market and
raises questions that are becoming increasingly important to managers in the global e-commerce.
Using data from the international online textbook market, this study finds that average prices for
equivalent items differ significantly across countries. Even after controlling shipping services by
equalizing the shipping time for books, UK online retailers still uniformly offer 22.9 percent
lower prices than US online retailers. Applying a conditional logit model to customer choice, this
study shows that US customers value foreign offers 10.5 percent lower than domestic offers.
There is a 12.4 percent of net price difference between UK and US online retailers after
controlling for all other choice-specific characteristics. This price differential can benefit
consumers, and can provide profitable opportunities for firms involved at each stage of the book
industry.
Key words: cross-country price differential, customer choice, international online market
Page 2
1. Introduction
"...with the invention of the steamship, telegraph, railroad and eventually telephone, it is safe to
say that this first era of globalization before World War I shrank the world from a size 'large' to a
size 'medium'. Now, we have the next round of globalization, thanks to the Internet, starting in
the early '90s. The big difference is the intensity of our current globalization, because vast
numbers of people have access to the information and technology to become a player (unlike
Globalization I)."
Thomas Friedman (2000)
Since its birth, the Internet has had wide range of impacts on every aspect of business. During
the past decade, Internet penetration has been increasing dramatically in countries all over the
world. In combination, the fast developing Internet technologies, such as broadband, mobile
Internet, wireless LANs, and more intuitive interfaces, hold great potential for influencing online
marketing practices. Reduced search costs, real-time access to firms anywhere, abolition of
certain types of intermediaries, etc. strongly reinforce globalization. As it approaches universal
adoption and usage, the Internet has become a catalyst for the globalization of electronic
commerce. Over the past decade, online retailers all over the world have been switching their
focus from exclusively domestic markets to competition in international markets. Almost all
online companies have both local and foreign services. Customers make worldwide searches via
shopbots or search engines. Business managers are seeking ways to restructure and reconstruct
their business to take advantage of the new technology and the new economic systems. The
growing global electronic marketplace is fostering intensive research on how the globalization of
electronic commerce affects market efficiency, the competitive behavior of retailers, and
customer choice behavior, etc.
Despite the fact that the Internet has far-reaching implications for international e-commerce,
there is little systematic research into the principles of Internet economics and pricing in the
international market. Several studies of Internet market efficiency have yielded two main
1
Page 3
findings: First, significant price differentials and price dispersion exist online (Bailey 1998b,
Clemons, Hann and Hitt 1998, Brynjolfsson and Smith 2001). Although the Internet eliminates
space and time barriers and reduces the cost of searching, it does not make online markets
conform to the Law of One Price. No evidence has been found that the dispersion falls over time.
The source of online price dispersion can be classified into e-tailer characteristics and market
characteristics (Pau et al., 2001). Second, following the nature of information goods, retailers
make strategic experiments online in pricing, bundling and product differentiation (Clay,
Krishnan and Wolff 2001). Retailers take advantage of price elasticity of online markets to make
smaller and more frequent changes on price and differentiate themselves on dimensions of brand,
channel, service etc (Smith, Bailey, Brynjolfsson 2000). However, all of the previous research
has focused on the domestic online marketplace. No literature has explored the online markets in
an international context except one recent paper by Tay and Clay which studies cross country
price differentials in the online book market.
Although there are quite a few studies about online communication and customer choice, most
of them have focused on brand and trust online (Shanker and Rangaswamy 1998, Urban, Sultan
and Qualls 1998, Degeratu, Rangaswamy and Wu 1999, Kollock 1999, Brynjolfsson and Smith
1999, 2000). Some studies of international consumer behavior show that cultural differences play
important parts of online buyer’s intentions (Javenpaa and Tractinsky 1999, Cheol Park 2001,
Park and Jun 2002, Patrick et al. 2002). Western cultural values prize individualism and low-
context while oriental ascribes collectivism and high-context (Kim et al. 1998). However, other
researchers have found that many international users of the Internet are similar to U.S. users
(Quelch and Klein 1996). Some observers view Internet-based transactions as essentially culture
2
Page 4
free (Perterson et al. 1997). Systematic research to address online consumer behavior in an
international context is important for a variety of constituencies.
Overall, the application of the existing theory and methodologies, and the development of new
theories are needed in a global context.
This paper studies the international online market for textbooks from two main aspects: the
price levels and price dispersion across countries and across stores, and the customer choice
behavior in the international online market. Based on the results, possible explanations and
profitable opportunities in the international online market are discussed. Using panel data
collected from DealTime.com for 461 English textbooks from seven countries over 23 months,
this study demonstrates that significant price differences exist between different countries and
among bookstores within a country. On average, as a percentage of list prices
1
, the item prices
from UK online retailers are 35.7 percent
2
less than the item prices from US online retailers.
Even after equalizing the shipping time, UK price is still 22.9 percent less than US pirce. On the
customer choice side, the model shows that customers respond to cross-country price
differentials. US customers value foreign offers 10.5 percent lower than domestic offers. They
can take advantage of the net price differential, which on average is 12.4 percent of list price, to
increase the consumer surplus. This price differential can also offer a value proposition for
suppliers and retailers.
1
The list price is the publisher’s recommended price for a book.
2
For convenience, the “percentage” refers to the percentage value of the list price when
comparing two offers in this paper.
3
Page 5
2. Prior Research
Bakos (1997) conducted one of the first theoretical studies on electronic market efficiency. He
pointed out that the electronic marketplace reduces the search cost, resulting in increased market
efficiency, possibly lower prices and increased competition among sellers. Empirical research
later proved that electronic marketplace does increase competition, but does not necessarily
lower the price comparing to conventional outlets (Bailey 1998a, Brynjolfsson and Smith 1999,
Lee 1997). Ward empirically compared online grocery prices to their offline counterparts and
found the price discrimination mechanism is used less often online than offline. Together with
information on product availability and price levels, these findings are consistent with online
groceries having limited selection, higher costs and lower price-cost margins. These suggest that
the market may only serve a niche of consumers who place a premium on convenience (Ward
2002). Meanwhile, price dispersion online is significant in CDs, books, software markets, and
online travel agents (Bailey 1998b, Clemons, Hann and Hitt 1998, Brynjolfsson and Smith 2001,
Clay and Tay 2001). In order to avoid direct price competition, retailers are experimenting online
with product differentiation and price discrimination on different books according to their
popularity (Clay, Krishnan and Wolff 2001). Pan et al. developed a comprehensive framework of
the drivers of online price dispersion that includes market characteristics such as number of
competitors, consumer involvement, and product popularity, in addition to e-tailer characteristics
and product category differences (Pan, Ratchford, Shankar 2001
)
. Their follow-up paper (Pan,
Ratchford, Shankar 2002) showed empirically that the proportion of the price dispersion
explained by e-tailer characteristics is small. Choudhary et al. developed an analytical framework
to investigate the competitive implications of personalized pricing technologies (PP)
(Choudhary, Ghose, Mukhopadhyay 2002). Maurer et al. pointed out the emergence of winner
4
Page 6
take all is one of the distinctive characteristics in the competitive electronic market (Maurer and
Huberman 2000). The effect of cost transparency is becoming a threat to the companies on the
Internet (Sinha 2000). Violation of the Law of One Price is not confined to the Internet. There is
a great deal of literature about the failure of Law of One Price in both domestic and international
markets. Parsley and Wei (1996) found that convergence rates of Purchasing Power Parity within
US are substantially higher than cross-country rates. Asplund and Rogers (1996) examined the
prices from products sold in Scandinavian duty-free stores and found that the Law of One Price
does not even hold for identical goods sold at the same locations, as long as these goods are
denominated in different currencies. Haskel and Wolf (2001) examined the retail transaction
prices for a multinational retailer and found up to fifty percent deviations from the Law of One
Price.
Studies on consumer choice behavior are from several perspectives. Häubl and Trifts
developed a hypothesis by using a two stage choice model of consumer decision making in
online shopping environment (Häubl and Trifts 1999). Ward et al. found that consumer search
and purchase behavior varies in the context of a multiple channel retail environment (Ward,
Morganosky 2000). Online customers may or may not be more sensitive to price than offline.
While customers can search online for lower prices, the Internet channel cannot provide them
live information which they can exploit in physical stores. Online customers may rely more on
other factors such as brand, appearing to be less price sensitive (Brynjolfsson and Smith 2000).
Chircu et al. conducted field experiments on making travel reservations, and found that trust and
expertise are important in encouraging adoption of electronic commerce (Chircu, Davis,
Kauffman 2000). Study from shopbot customer behavior shows that even though price sensitive,
shopbot customers value brand effects in their choice decision (Brynjolfsson and Smith 2000).
5
Page 7
The brand awareness of competing retailers, brand sensitivity, and the willingness of consumers
to pay a premium for a leading retailer affect the range and level of price dispersion online (Chen
and Hitt 2002). Lee proposed a behavioral model of online purchasers in E-commerce
environment (Lee 2002). Park did a cross-cultural comparison of online buying intentions in
Korea and America, and identified a model of online buying intention explained by Internet
usage, perceived risk, innovativeness, and online buying experience (Park 2002). Lynch et al.
conducted a sample study of advanced Internet users from 200 countries, and found there were
differences in beliefs, attitude, perceptions, and Internet buying behavior depending on user
experience, and home country or region. There were important differences after controlling for
social, cultural and macro-economic factors (Lynch and Beck 2001).
3. Methodology and Data
3.1 Methodology
3.1.1 Cross-country and cross-store price dispersion
We use store and book fixed effects regression to analyze the cross-country price difference. Our
analysis focuses on US, UK and German online bookstores because 44 out of 48 retailers are
from these countries. It is worth mentioning that the tax policy is different in each country.
Online retailers in the United States are exempt from sales tax if the purchase is not within their
local state. So most of the online purchases made through online US retailers are free of tax. The
United Kingdom exempts books from the Value Added Tax. Germany’s seven percent Value
Added Tax is automatically included in the price for any product. So we do not need to care
about tax issues in this analysis.
6
Page 8
Almost all the retailers offered various shipping options. To facilitate the analysis, four main
shipping services are considered: domestic standard shipping, for which the shipping time is
about 3 to 7 workdays; domestic express shipping, for which the shipping time is about 2
workdays; international standard shipping, for which the shipping time is about 3 weeks; and
international express shipping, for which the shipping time is about 2 to 3 workdays.
Unless a costumer is ordering a huge quantity of books, custom duties are not an issue. United
Stated Customs permits printed books to enter the country duty free. Orders below $2,000 are
subject to informal entry. US customers can take advantage of the price difference on the
international book market and save money by purchasing from UK retailers or other bookstores.
3.1.2 Online customer choice
The multinomial logit model (McFadden, 1974) has been extensively applied in discrete choice
analysis of panel data. Conditional (or fixed effect) logit model is a variation of the multinomial
logit model. Conditional logit model deals with choice-specific characteristics (McFadden,
1974). The equation is
∑
=
∑
∑
=
=
=
=
J
j
z
z
K
k
jk
k
K
k
k
jk
e
e
j
y
P
1
1
1
)
(
α
α
(1)
where
distinguishes choice response, and k
J
j
,
,2,
1 L
=
K,
,2,
1 L
=
distinguishes
characteristics.
is the kth characteristic for an offer j, and
jk
z
k
α
is the corresponding
coefficient. In this study, all the explanatory variables are choice dependant. A customer’s last
click is used as the indicator of the maximum utility choice. The probability that a customer
chooses an offer j in a search session can be expressed as:
7
Page 9
∑
=
∑
⋅
+
⋅
+
⋅
+
⋅
+
⋅
+
⋅
+
⋅
∑
⋅
+
⋅
+
⋅
+
⋅
+
⋅
+
⋅
+
⋅
=
=
=
=
J
j
n
Firstscree
Firstoffer
Foreign
Bigthree
DeliveryNA
iverytime
Averagedel
Totalprice
n
Firstscree
Firstoffer
Foreign
Bigthree
DeliveryNA
iverytime
Averagedel
Totalprice
j
K
k
K
k
e
e
lastclick
P
1
1
7
6
5
4
3
2
1
1
7
6
5
4
3
2
1
)1
(
α
α
α
α
α
α
α
α
α
α
α
α
α
α
(2)
In this model framework, it is assumed that the customers value each offer and independently.
Thus the choice behavior is governed by independence of irrelevant alternatives (IIA). Using the
total price as the baseline, one can compare the relative marginal effect of the other attributes to
the total price. For example, suppose we would like to know the following: Given two offers
with all the other attributes the same, how much should we decrease the price of a foreign retailer
in order to keep the latent utility constant? The answer is:
totalprice
forgeign
p
β
β
−
=
∆
(3)
Similar equations apply to the other coefficients.
3.2 Data source and data characteristics
The dataset is from DealTime.com, a prominent Internet shopbot for books. DealTime is an
international company with local shopbot sites in the U.S., U. K., and Germany. By using
DealTime, customers can make worldwide price comparisons for books and other products from
around 70 different retailers operating in 10 different countries. Customers visiting the site first
identify the book they are interested in by searching on the title, the author, the publisher, or the
ISBN. DealTime then queries distinct book retailers for information on this book. The prices and
the delivery times are queried in real-time and thus represent the most up-to-date data from the
8
Page 10
retailers. Customers then evaluate different options in the price comparison table and click on a
particular offer to complete the purchase at a retailer’s site.
The textbook market is a good choice because several features of this industry conform well to
our research requirements. First of all, it is a relatively mature electronic market. Secondly, the
online book industry is homogenous. Thirdly, textbooks are high value items in the book
industry; customers in the textbook market tend to be price sensitive. Some customers choose to
buy used textbooks from used book markets. The popular way for customers to get lower prices
is to search for and purchase books online. Through searching and price comparison, customers
can get the textbooks for a lower price. Fourthly, the main customers for textbooks are students,
who have less money and are more price sensitive than the public in general. Fifthly, students
have the most frequent and regular access to the Internet, so they are most likely to search for
and purchase books online.
Shopbot data has good characteristics for analyzing price dispersion and consumer choice
behavior, because it allows researchers to observe the exact information offered to the customer
by multiple competitive retailers. It also provides information about the customer’s search and
choice behavior in response to this information. The data used here includes three categories:
offer data, session data and choice data (Table 1). The offer data contains an individual price
quote
3
from a retailer, which in our analysis refers to item price, total price, retailers who offer
the price, shipping cost, delivery time, delivery availability and the position in an offer table.
Session data contains information of an individual search occasion for a book, which refers to the
session identification number and the ISBN number for a book. Choice data includes last click-
through. A customer clicks the link to a particular retailer if he or she is interested in its offer.
3
All the quotes in our dataset are recorded in US dollar value.
9
Page 11
“Last click-through” records information if this click is the last one during this search. By using
click-throughs as the proxy for customer choices, we are able to track the traffic driven to a web
site through the shopbot. Although the fact that the customer visited a store doesn’t guarantee the
actual purchase, Brynjolfsson and Smith (2001) shows that the last click-throughs and the actual
purchase have positive and steady relationship; the conversion rates
4
range from 0.49 to 0.51.
These statistics do not vary significantly across stories.
4. Results and Analysis
4.1 Descriptive statistics
The data set obtained from DealTime has 77,267 observations from 48 retailers in 7 countries
from August 27, 1999 to July 25, 2001. It includes 2,561 distinct customer searches for 461
distinct ISBN’s, an average of 37 offers per search. Table 2 lists the summary statistics of the
dataset. If counting offers from the same retailer as one distinctive observation in each session,
we have 29,764 distinctive observations, among which, 9,718 observations (32.65 percent of the
total observations) are from non US retailers. Among these offers, 6,505 observations are from
UK retailers, 2,352 observations are from German retailers, and 1,277 observations are from
Canadian retailers. It shows that the price dispersion within a search session is large. On average,
4 Conversion rate is the percentage of visitors who complete a desired action. Different sites
have different objectives, so conversion comes in many forms, from registering for a newsletter
to completing a purchase. For example, let's say you get 1,000 visits to your site, and 20 orders.
Your sales conversion rate = 20 /1,000 = 2%.
10
Page 12
the standard deviation of the normalized item price
5
is 0.22, while the difference between the
highest price and lowest price is 57 percent of the list price. The difference between the lowest
price and the second lowest price is about 9 percent of the list price, while the difference between
the lowest price and the fourth lowest price is about 23 percent of the list price. If a customer
chooses the lowest price offer in a session instead of choosing the offer from Amazon (US), he
or she could save approximately 47 percent of the list price. The price difference across countries
is large. In terms of the normalized item price, the average of US retailers is 1.02, while the
average of UK retailers is 0.67. UK retailers charge 35 percent less than US retailers for a book.
4.2 Cross-country and cross-store price dispersion
Table 3 presents the regression results of price variation patterns across countries from US
customers’ point of view. The price differences are analyzed from three perspectives: the
normalized item price, the normalized total price with standard shipping service and the
normalized total price with express shipping service. Column 2, 3 and 4 of Table 3 list the
respective results for the purchase of a single book by US customers. It shows that the price
difference between countries is significant. US online retailers consistently charge the highest
prices for all the three price types while UK bookstores always offer the lowest prices. For
example, as the percentage of list price, the average difference of the item price between the US
and UK is 35.7 percent. German retailers charge 3.1 percent less than US retailers. Retailers
from other countries offer on average 14.1 percent lower prices than US bookstores. Given that
the average list price in our dataset is $73.39, the prices from UK retailers on average are $26.20
lower for a single book than US retailers. A US customer choosing standard shipping can save
5
Normalized item price = Item price / List price
11
Page 13
$24.22 by ordering from a UK retailer. If choosing express shipping, by which the shipping time
is about the same both from US and UK retailers, one can save $16.81 by ordering from a UK
retailer. Column 5 and 6 of Table 3 list the results for normalized total prices of a bundle of five
6
books with standard shipping and express shipping to US costumers. It shows that the savings
per book in a bundle are about the same as for a single book due to the high marginal shipping
cost for overseas shipping. But one can still save a lot in bundle purchase from UK retailers
instead of US retailers.
Table 4 presents the store and book fixed effects regression results by online retailers
7
for
normalized item price and normalized total price with express shipping to US customers. One
can see more specifically that the price dispersion is significant both across stores and across
countries. According to column 2 of Table 4, the item price of a book from StudentBookWorld
(UK) can be 51.6 percent lower than from Amazon.com (US). Considering that the average list
price in our dataset is $73.39, a US customer would save $37.87 by ordering a book from
StudentBookWorld (UK) instead of from Amazon.com (US). Column 3 of Table 4 displays the
regression results for normalized total price with express shipping cost where the shipping time
is about the same for orders either from US retailers or from foreign retailers. It shows that after
eliminating the shipping time difference, most UK retailers still offer lower prices to US
customers than the average US retailers. Specifically, the lowest price is from bol (UK), which is
6 Two scenarios apply when customers buy books online: single book purchase and multiple
books purchase. Here an amount of 5 is used as a simplified proxy for multiple books purchase
to study the price differences in UK and US markets.
7
Half.com is excluded from the analysis because Half.com is known as a used book sale
website.
12
Page 14
34.3 percent lower than Amazon (US); and the second lowest price is from Blackwells (UK),
which is 25.2 percent lower than Amazon (US). If choosing express shipping for one single
book, a US customer can save $25.25 by ordering from bol (UK) instead of ordering from
Amazon (US). We also found that a German retailer, bol.de, offers even lower prices (52.7
percent on average) to US customers. According to this, a US customer can save about $38.68
for a book by ordering it from bol.de rather than from Amazon (US). However, in general,
German retailers charge total prices 18 percent higher than US retailers.
4.3 Online customer choice
Table 5 lists the estimates of the parameters in conditional logit models for US customer choice.
The model selection criteria used are the pseudo-R
2
, Akaike information criteria (AIC), Schwartz
Bayesian information criteria (SBIC) and Bozdogan’s index of informational complexity
(ICOMP). The pseudo-R
2
selects the model that minimizes the log-likelihood value in maximum
likelihood estimation. The best model maximizes the criterion. AIC, SBIC and ICOMP select the
“best” model compromising an adequate goodness of fit and a small number of parameters by
adding a penalty for over parameterization to the lack of fit measure. The best model minimizes
the criterion.
The coefficients listed in Table 5 are interpreted as preference weights in a latent utility
function. The values of the criterion for each model specification are listed at the bottom of the
table. It shows that either model 5 or model 6 is the best fit. As one can see from the results, the
total price, average delivery time, delivery time uncertainty and foreign factor have consistently
negative effects, which means higher price, longer delivery time, less shipping guarantee, and
purchase from foreign retailers will reduce the latent utility when customers evaluate an offer.
13
Page 15
An interesting result is that the absolute value of the total price and average delivery time
become much less significant factors compared to the other factors such as the delivery
uncertainty, foreign factor, and positions in the offer table, which is not consistent to the results
from previous studies on domestic markets that online customers are very sensitive to price. The
results show that in the international online market, US customers value delivery certainty
guarantees 2.22 times as much as the price and 40.4 times as much as the average delivery time
when making choices. The “foreign” parameter has a significant negative effect on a US
customer’s latent utility. In model 6, the coefficient of “Foreign” is -0.700, which is 7.69 times
of the coefficient of the total price. On the other hand, the position where an offer is located in
the price comparison table strongly affects a customer’s evaluation of an offer. Customers tend
to value offers that appear in the first screen of the price comparison table more than those
appear later. This is consistent with the customer’s high sensitivity to the price rank of retailers
found by Ellison and Ellison (2001) when they examined the market for commodity memory
modules sold via a price search engine.
Comparing the relative marginal effect to the total price, being a foreign retailer has $7.69
price disadvantage, which means—given that all other factors the same—a foreign retailer must
charge a US customer at least $7.69 (10.5 percent of the list price) less for a book in order to
have this customer click through its offer for the book. Delivery uncertainty has a $2.22
disadvantage (3.1 percent of the list price), which means if a retailer can not provide the specific
delivery time, the price it offers should be at least $2.22 less to keep a customer’s latent utility
constant.
14
Page 16
5. Discussion
The empirical results show that cross-country price dispersion online is remarkable. What are the
reasons for the price differences? One possible reason could be that it is simply due to the
weakness of the British pound relative to the US dollar during the time the data were collected.
To investigate this issue, we collected the monthly currency exchange rates of GBP to USD for
the same period our data were collected. The record shows moves in the exchange rate in favor
of USD during the period (Figure 1), which should lead to stronger purchasing power for US
dollar and larger price differences for US customers. The synchronized price data from UK
online book stores does show a very weak down slope of the average normalized price over the
whole period. But the price fluctuates so much that the effect of the currency exchange ratio is
almost over shadowed. The regression result also does not show a significant relationship
between the currency exchange rate and the average normalized price of UK retailers.
One important reason for this price differential is that there has been a substantial discounting
by UK retailers of all kinds since the abolition of the Net Book Agreement in the UK. According
to Book Marketing Ltd figures, 52% of all titles are discounted by one retailer or another
whereas in USA, discounting both in terrestrial and online bookstores tends to be applied only to
current bestsellers.
Moreover, it appears that the European market competes aggressively with US market not only
within the retailing markets but also from the publishing industry. Both the paperback lists and
the hardback lists reflect the advantage UK publishers achieve by their decisions about
appropriate price points. A B format paperback of Nick Hornby’s How to be Good at £6.99
($10.34) competes against a US trade paperback priced at $13. Janet Evanovich’s Hard Eight,
15
Page 17
priced at £ 12.99 ($19.92), discounted to £ 9.99 ($14.79) in Waterstone’s would cost a US
customer £22.47 ($32.26) in Barnes & Noble.
Competition between stores within a country could also escalate price differences between
countries. For example, WH Smith competes aggressively against Amozon.co.uk by offering
particularly steep discounts online; Amazon (US) and Barnes and Noble headed for the price war
against each other on price range and price levels.
On the other hand, the timing difference between US and UK publications could reduce the
opportunity for direct competition.
Manifold factors make online price dispersion a complex issue. There may be other interesting
reasons which can be explored by future researchers. More empirical findings and theoretical
support are expected to provide us intuitions for developing analytical models of cross-country
price competition.
Results from customer choice analysis show that—given all the other characteristics the
same—in order to keep a customer’s latent utility constant, the acceptable price difference
between a foreign retailer and a domestic retailer should be $7.69, which is 10.5 percent of the
average list price of our dataset. What are the reasons for the price disadvantage of foreign
retailers?
One important reason could be the awareness of foreign retailers. Customer awareness is
known to be as important as physical location in conventional Bricks-and-Mortar markets.
Searching for a retailer online can be very difficult because of the sheer volume of information
online. Given that there are millions of Internet sites available online and customers are less
aware of foreign retailers, the search cost to locate foreign retailers can be very high.
16
Page 18
Trust could be another important reason because of the spatial and temporal separation
between buyers and sellers imposed by the medium (Brynjolfsson and Smith 1999). Consumers
may be willing to pay a premium to purchase a product from a retailer who they trust in favor of
an unknown retailer. Trust can come from the previous experience about the robustness of the
online communities, unbiased product information, good service, and conventional brand name
and reputation, or simple bias against foreign retailers. Customers have less experiences dealing
with foreign retailers, so US customers give less credit to foreign retailers than to US retailers.
Another reason could be the immaturity of international electronic commerce. The
globalization of e-commerce is a long term, complex and systematic process. It is not simply
about standardization and homogenization (Guillen, Mauro F. 2001). Given the cross-national
differences in infrastructure, regulation, language, consumer demographics and behavior,
payment system and currencies, internationalizing companies need to standardize their
managerial strategies on a worldwide basis and take into consideration both global integration
and local responsiveness. Online globalization is still in the middle of transition. Business
managers are looking for ways to reconstruct their business. Academic researchers attempt to
find the new rules and principles for the emerging systems. The disadvantage for foreign
companies may be expected to shrink with the development of electronic commerce and the
increasing popularity of cross-country commercial activities.
As one can see, in the online market, the average price difference between US retailers and UK
retailers is 22.9 percent. Given the average list price $73.39, the difference in dollar format is
$16.81. For US customers, the acceptable price difference between a foreign retailer and a US
retailer is $7.69, which is 10.5 percent of the average list price. There is a net difference of $9.12
between the value in the marketplace and the value from the customer’s choice, which is 12.4
17
Page 19
percent of the list price. According to the UK/USA Price Comparison Index reports from the
Publishers Association (PA)
8
, UK publishers deliberately set their price points to compete
against US publishers. So this cross-country price difference can provide profitable opportunities
for expansion among wholesaler-distributors, the supply chain partners and retailers, and
welfare-enhancing opportunities for consumers. For example, wholesaler-distributors and the
supply chain partners can use cross-channel price discrimination strategies to gain extra profits;
US retailers can purchase goods from foreign wholesalers at a lower cost. US customers can buy
a book from UK retailers online and get the same book in about the same time as from US
retailers but at a cheaper price.
6. Conclusions and Open Questions
The Internet and E-commerce have driven the competitive transformation and globalization of
the book industry. New economic phenomena occur in the international competitive structure of
online markets. This study empirically demonstrates that significant price differences exist online
across countries and among bookstores within a country. Book prices from the UK online
retailers are uniformly lower than prices from US online retailers. The usual three dominant
bookstores in the US book market are no longer dominant in the global online market. Customers
value foreign online retailers less when they make choices in the global online market. There is
12.4 percent of net price difference between UK and US online retailers after controlling for all
8 The Publishers Association is the UK’s representative body for the book and serials industries.
It regularly surveys the comparative prices of books in the UK and the USA both in bookshops
and through online sources
.
18
Page 20
other choice-specific characteristics. This price differential can benefit consumers, and can
provide profitable opportunities for firms at each stage of the book industry. Why does this price
differential exist? How long will it last? How will it develop? What does this mean to
international book publishers, wholesales, distributors, supply chain partners, online retailers and
online customers? Is this the common phenomenon in other online markets? Are there
generalized rules and principles for this new economic phenomenon? Given the opportunities
and challenges in the international e-commerce, what are the short- and long-term strategic
business models for firms operating internationally? What are the implications on national and
international business policy? This paper is only a starting point of exploration of the
international e-commerce. There are a wide range of topics and areas where future research can
add to our understanding.
19
Page 21
Table 1: Data Category
Offer Data
Total Price
Total price offered by retailers( total price=item price + shipping cost)
Item Price
The unit price a retailer offers on its web
Retailer_ID
A number to identify a retailer
Shipping Price
The price for shipping
Maximal Delivery Time
Maximum of delivery time range
Minimal Delivery Time
Minimum of delivery time range
Delivery NA
Dummy variable which takes value of 1 if a retailer can’t quote an
acquisition time (time to get book from distributor)
First Offer
Dummy variable which takes value of 1 if an offer is the first one in the
price comparison table
First Screen
Dummy variable which takes value of 1 if an offer is listed in the first
screen in the comparison table
Session Data
ISBN
A number, which can uniquely identify a book. One can get book
variables(category, binding, pub date)
Session_ID
Uniquely identifies each search session
Choice Data
Last Click-Through
==1 if a customer at last click through certain retailer
20
Page 22
Table 2 Summary Statistics
Variable
Observations
Mean
Std. Dev
Min
Max
Prices and Offers (461 unique ISBNs)
Publisher’s
Recommended
Price
77,267
73.39
22.80
16.91
109.84
Item Price
77,267
61.42
24.50
8.75
128.81
Normalized Price
77,267
0.94
0.22
0.38
1.77
Offers per Session
77,267
36.76
12.48
1
61
Normalized
1
Price Dispersion Within a Session
Standard
Deviation
77,267
0.18
0.09
0
0.50
Max-Min Price
77,267
0.57
0.20
0
1.11
2
th
Lowest Price -
Lowest Price
77,267
0.09
0.11
0
0.58
3th
Lowest Price -
2
th
Lowest Price
77,237
0.06
0.09
0
0.72
4
th
Lowest Price -
3
th
Lowest Price
77,024
0.08
0.12
0
0.77
Amazon Price -
Lowest Price
64,704
0.47
0.22
0
1.19
Normalized Cross Country Prices
Total Price
2
29,764
6
72.73
24.95
12.54
150.06
US(28 retailers)
20,046
1.02
0.14
0.39
1.77
UK(10 retailers)
6,011
0.67
0.23
0.39
1.47
GER(6 retailers)
2,077
1.00
0.22
0.43
1.47
Others(4 retailers)
1,775
0.88
0.14
0.39
1.48
Customer Choice Parameters
Total Price
3
77,267
73.73
25.07
11.7
156.04
AveDeliveryTime
77,267
12.56
11.91
1
38.5
Delivery NA
77,267
0.38
0.49
0
1
Foreign
77,267
0.21
0.41
0
1
Big Three Stores
77,267
0.19
0.39
0
1
First Offer
77,267
0.03
0.17
0
1
FirstScreen Offers
77,267
0.31
0.46
0
1
Price and Competitive Market Structure Parameters
Time
4
26,057
6
233.56
123.28
1
425
Number of US
Stores
26,057
9.36
2.97
0
15
Number of UK
Stores
26,057
2.97
1.80
0
8
Average US
Price
5
29,742
1.02
0.09
0.66
1.67
Average UK Price
26,774
0.72
0.23
0.38
1.46
Notes:
1
Normalized price = Item price / List price;
2
Total price data here is the subset for price and price
dispersion regression;
3
Total price data here is for customer choice model regression;
4
Time is a discrete
variable. The first day in our data collection period is coded as 1; the second day is coded as 2, and so on so
forth.
5
The average US price refers to the average price of US retailers in a particular session.
6
A certain
retailer may show more than one offer in the same session table if it offers several kinds of shipping
services for the book. To avoid this redundancy, we count only one observation per retailer per session both
in descriptive statistics and price dispersion analysis.
Page 23
Table 3 Normalized Price Regressions by Countries for US customers
(ISBN Fixed Effects)
Independent
Variables
Normalized Item
Price
Normalized
Total Price
Standard
Shipping
4
to US
Normalized
Total Price
Expressed
Shipping
5
to US
Normalized
Bundle Total
Price (5 books)
Standard
Shipping to US
Normalized
Bundle Total
Price (5 books)
Express
Shipping to US
Constant
1
1.022***
2
(0.001
3
)
1.104***
(0.001)
1.282***
(0.002)
1.062***
(0.001)
1.100***
(0.001)
UK Retailers
-0.357***
(0.002)
-0.330***
(0.003)
-0.229***
(0.005)
-0.318***
(0.002)
-0.223***
(0.003)
GER Retailers
-0.031***
(0.004)
0.005
(0.004)
0.180***
(0.006)
-
-
Other Retailers
-0.141***
(0.004)
-0.114***
(0.006)
-0.010
(0.012)
-0.143***
(0.004)
-0.070***
(0.006)
Book Fixed
effects
Yes
Yes
Yes
Yes
Yes
Numbers
Observations
29,764
29,764
27,180
25,131
20,869
Adj. R-Squared
0.4098
0.3212
0.1285
0.3977
0.1651
Notes:
1
The omitted indicator variable is for US retailers.
2
***: Results are significant at p< 0.01; **:
Results are significant at p< 0.05; *: Results are significant at p< 0.1.
3
Standard errors are listed in the
parenthesis.
4
Standard shipping to US customer usually takes 30 work days for delivery.
5
Expressed
shipping to US customer usually takes 2-3 work days for delivery.
22
Page 24
Table 4 Normalized Price Regression by Retailers (ISBN Fixed Effects)
Independent Variables
Normalized Item Price
Normalized Total Price
(Time-Equal
2
Shipping to US)
Constant
1
1.069***
(0.002)
1.324***
(0.004)
US Retrailers
1BookStreet
0.054***
(0.004)
0.020***
(0.006)
A1Books
-0.128***
(0.008)
-0.116***
(0.007)
AllBooks4Lesscom
-0.059
(0.010)
-0.100***
(0.011)
AlphaCraze
-0.101***
(.007)
-0.024***
(0.005)
Buycom
-0.063
(0.008)
-.111***
(0.008)
BCYbookloft
-0.173***
(0.045)
-
BNcom
0.023***
(0.007)
0.005
(0.009)
BookbuyersOutlet
0.062***
(0.009)
0.179***
(0.010)
Borders
-0.044***
(0.007)
-0.097***
(0.006)
Brians
-0.039***
(0.137)
-0.153
(0.251)
Classbook
-0.025***
(0.007)
0.063***
(0.005)
Cherryvalley
-0.166**
(.007)
-0.105*
(0.060)
DoubleDiscount
-0.123***
(0.010)
-0.153***
(0.012)
Ecampus
-0.110***
(00.007)
-0.064***
(0.006)
Elgrande
0.040***
(0.012)
-0.227***
(0.014)
FatBrain
-0.046***
(0.006)
-0.145***
(0.005)
Magusbooks
-0.052
(0.082)
-
Page1Book
-0.061***
(0.006)
-
Powells
-0.059
(0.015)
0.133***
(0.017)
Seekbooks
0.076***
(0.009)
0.162***
(0.009)
Shopping
-0.162***
(0.009)
-0.227**
(0.009)
Textbookcom
-0.079**
(0.010)
-0.057***
(0.012)
Textbookx
-0.181***
(0.009)
-
23
Page 25
TheBigStore
-0.108***
(0.011)
-
Wordsworth
0.010***
(0.007)
-0.049***
(0.060)
UK Retailers
Alphabetstreet
-0.465***
(0.007)
-0.212***
(0.006)
Amazon.co.uk
-0.306***
(0.007)
0.069***
(0.006)
Blackwells
-0.307***
(0.015)
-0.252***
(0.018)
Countrybookstore
-0.435***
(0.007)
-
Davista
-0.511***
(0.013)
-
InternetBookShop
-0.275***
(0.007)
-
StudentBookWorld
-0.516***
(0.008)
-
UKbol
-0.501***
(0.007)
-0.343***
(0.007)
WaterstonesOnline
-0.497***
(0.021)
-0.164***
(0.017)
WHSmithOnline
-0.378***
(0.009)
-0.080***
(0.010)
GERMAN Retailers
Amazon.de
-0.052*
(0.006)
0.154***
(0.005)
Buchkatalog
0.018***
(0.018)
-
Bol.de
-0.567***
(0.019)
-0.527***
(0.023)
Bbooxtra
-0.217
(0.145)
-
Lesezone
-0.192***
(0.012)
-
Retailer_id41
-0.273***
(0.078)
-
Retailers from Other Countries
ChaptersGlobe
-0.200***
(0.007)
-
Dymocks(AUS)
-0.380***
(0.015)
-0.093***
(0.019)
Lioncc (AUT)
-0.165***
(0.008)
-
Retailer__48
-0.166***
(0.008)
-0.036***
(0.008)
Book Fixed effects
Yes
Yes
Numbers Observations
29,764
27,180
R-Squared
0.6416
0.6859
Notes:
1
The omitted store is Amazon.com.
2
If a customer chooses express shipping when making an
order, the shipping time is about the same either from a US store or from a foreign store.
24
Page 26
Table 5 Customer Choice Models
1
2
3
4
5
6
5
Total Price
-0.116***
(0.006)
3
-0.119***
(0.006)
-0.134***
(0.007)
-0.133***
(0.005)
-0.090***
(0.005)
-0.091***
(0.007)
Average
Delivery
Time
-0.004
(0.003)
-0.0004
(0.004)
-0.0002
(0.004)
-0.005
(0.004)
-0.005
(0.005)
Delivery
NA
1
-0.060
(0.082)
-0.085
(0.081)
-0.090
(0.081)
-0.189**
(0.084)
-0.202***
(0.109)
Foreign
2
-1.064***
(0.135)
-1.073***
(0.135)
-0.822***
(0.131)
-0.700***
(0.783)
Big Three
-0.079
(0.129)
-0.010
(0.133)
First Offer
0.973***
(0.082)
1.019***
(0.086)
First Screen
1.599***
(0.129)
1.382***
(0.133)
Number of
observations
30,778
30,778
30,778
30,778
30,778
30,778
Store Fixed
Effects
No
No
No
No
No
Yes
4
Log
Likelihood
-2182.729
-2181.751
-2146.830
-2146.637
-1981.805
-1921.715
Pseudo-R
2
0.3324
0.3235
0.3340
0.3340
0.3765
0.3950
AIC
4367.4589
4369.5022
4301.6607
4303.2758
3977.6111
3895.4313
SBIC
4375.7934
4394.5058
4334.9989
4344.9485
4035.9530
4112.1298
ICOMP
4365.4589
4372.6262
4304.8555
4304.9494
3975.7490
3881.1757
Notes:
1
The dummy variable which takes value 1 if a retailer can not decide the specific delivery time.
2
The dummy variable is 1 for stores outside US.
3
Standard error is listed in parenthesis.
4
I create
dummies for the top 20 stores with the highest number of click_through in the dataset.
5
In model 6, instead
of using the dummy variable “BigThree” as what we did in model 5, we use store fixed effects regression.
25
Page 27
Figure 1
The Trend of the Currency Exchange Rate and the
Average Price of UK Book Stores
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
0
1
2
3
4
5
6
7
8
9 10 11 12 13 14
month
The av erage
normalized price
of UK book
stores
USDs to 1 GBP
26
Page 28
Reference:
1. Adamic, Jada A., Bernardo A. Hubernam. 1999. The nature of markets in the World Wide Web.
Quarterly Journal of Electronic Commerce 1(1) 5-12
2. Bailey, Loseph P. 1988a. Intermediation and electronic markets: Aggregation and pricing in
Internet commerce. Ph.D. Thesis, Technology, Management and Policy, Massachusetts Institute
of Technology, Cambridge, MA.
3. Bailey, Loseph P. 1988b, Electronic commerce: Prices and consumer issues for three products:
Books, Computer Discs, and Software, Organization for Economic Co-Operation and
Development OCDE/GD 98(4)
4. Bakos, J.Yannis. 1997. Reducing buyer search costs: Implications for electronic marketplaces.
Management Science 43(12) 1613-1630
5. Ben-Akiva, Moshe, Steven R. Lerman. 1985. Discrete Choice Analysis: Theory and Application
to Travel Demand. The MIT press, Cambridge, MA.
6. Brown, Jeffery R., Austan Goolsbee. 2000. Does the Internet make markets more competitive”
Evidence from the life insurance industry. National Bureau of Economic Research Working
Paper #7996, October.
7. Brynjolfsson, Erik, Michael D. Smith. 2000. The Great Equalizer? The Role of Shopbots in
Electronic Markets. Working Paper, MIT Sloan School of Management, Cambridge, MA
8. Burdett, Kenneth, Kenneth Judd. 1983. Equilibrium price dispersion. Econometrica 51(4) 955-
969
9. Carlson, John A., Donn R. Pescatrice. 1980. Persistent price distributions. Journal of Economics
and Business 33(1) 21-27.
27
Page 29
10. Charles Engel, John H. Rogers. 1996. How Wide is the Border? The American Economic
Review. 86(5). Pp 1112-1125
11. Chen, Hitt. 2001. Brand Awareness and Price Dispersion in Electronic Markets, Proceedings of
the 21st Annual International Conference on Information Systems (ICIS), New Orleans,
Louisiana, 2001
12. Cheol Park, “A Cross-Cultural Study on Consumer Values of Web Navigation and the Internet
Shopping Behaviors,” European Advances in Consumer Research, Vol. 5, 281-285, A.
Groeppel-Klein and F. Esch (eds.), Valdosta, GA: Association for Consumer Research,
December, 2001.
13. Cheol Park and Jongkun Jun, “A Cross-Cultural Comparison of Online Buying Intention: Effects
of Internet Usage, Perceived Risks, and Innovativeness," The Proceedings of the 8th Australian
WWW Conference, 472-484, A. Treloar and A. Ellis eds., Sunshine Coast, Queensland:
Australia, July 2002
14. Choudhary, Ghose, Mukhopadhyay 2002, Personalized Pricing and Quality Differentiation on
the Internet, eCommerce Research Forum, MIT
15. Chircu, Davis, Kauffman 2000, Trust, Expertise and E-Commerce Intermediary Adoption,
eCommerce Research Forum, MIT
16. Christ, Mario. 2001. On web site loyalty and web site stickiness in the World Wide Web.
Working Paper, Carnegie Mellon University, Pittsburgh, PA
17. Clay K., R. Krishnan, E. Wolff. 2001. Prices and Price Dispersion on the Web: Evidence from
Online Book Industry. National Bureau of Economic Research Working Paper# 8271, May.
28
Page 30
18. Clemons, Eric K., II-Horn Hann, Lorin M. Hitt. 1998. The nature of competition in electronic
markets: An empirical investigation of online travel agent offers. Working Paper, The Wharton
School of the University of Pennsylvania, Philadelphia, PA.
19. Devine, D. G. and Marine, B. W., 1979, The Influence of Consumer Price Information on Retail
Pricing and Consumer Behavior, American Journal of Agriculture Economics, 61, pp, 228-37.
20. Degeratu, Rangaswamy, Wu, 1999, Consumer Choice Behavior in Online and Traditional
Supermarkets: The Effects of Brand Name, Price, and other Search Attributes, C
21. Ellison, Glen and S. F. Ellison. 2001. Search, Obfuscation, and Price Elasticities on the Internet.
Working Paper, MIT, Cambridge, MA, January.
22. Friberg, Richard and Asplund, Marcus. September 2001, The law of one price in Scandinavian
duty-free stores. American Economic Review, 91 (4), pp. 1072-1083.
23. Ghosh, Atish and Wolf, Holger. Pricing in international markets: Lessons form the Economist.
Working Paper EC-94-21, Stern School of Business, N.Y.U.
24. Goldberg, Peter M., John D.C. Little. 1983. A Logit Model of Brand Choice Calibrated on
Scanner Data. Marketing Science 2(3) 203-238
25. Guillen, Mauro F. 2001, Is Globalization Civilizing, Destructive or Feeble? A Critique of Five
Key Debates in the Social-Science Literature. Annual Review of Sociology 27
26. Häubl and Trifts 1999, Consumer Decision Making in Online Shopping Environments: The
Effects of Interactive Decision Aids, eCommerce Research Forum, MIT
27. Indrajit Sinha, Timothy Devinney, 2000, Cost Transparency: The Net's Real Threat to Prices and
Brands, eCommerce Research Forum, MIT
28. Jarvenpaa, Srikka L. and Noam Tractinsky (1999), Consumer Trust in an Internet Store: A
Cross-Cultural Validation, Journal of Computer-mediated Communication, 5 (2)
29
Page 31
29. Jonathan Haskel, Holger Wolf, 2001, The Law of One Price – A Case Study. National Bureau of
Economic Research Working Paper#8112, February.
30. Kim, Donghoon, Yigang Pan, and H. S. Park. 1998, High- versus Low Context Culture: A
Comparison of Chinese, Korean, and American Cultures, Psychology & Marketing, 15 (6), 507-
521
31. Kollock, Peter. 1999. "The Production of Trust in Online Markets." Advances in Group
Processes (Vol. 16), edited by E. J. Lawler, M. Macy, S. Thyne, and H. A. Walker. Greenwich,
CT: JAI Press
32. Lee, Ho Geun. 1997. Do Electronic Marketplace Lower the Price of the Goods. Communications
of the ACM. 41(12)
33. Lee Pui-Mun. 2002, Behavioral Model of Online Purchasers in E-Commerce Environment,
Electronic Commerce Research, Volume 2 Issue 1-2
34. Lohse, Gerald L., Steven Bellman, Eric Johnson. 2000. Consumer buying behavior on the
Internet: Findings from panel data. Journal of Interactive Marketing 14(1)15-29
35. Patrick D Lynch, John C Beck. 2001, Profiles of Internet buyers in 20 countries: Evidence for
region-specific strategies, Journal of International Business Studies. Washington: Fourth Quarter
2001. Vol. 32, Iss. 4; p. 725 (24 pages)
36. Patrick Y. K. Chau, Melissa Cole, Anne P. Massey, Mitzi M. Montoya-Weiss, Robert M.
O'Keefe: Cultural differences in the online behavior of consumers. CACM 45(10): 138-143
(2002)
37. Pan, Ratchford, Shankar 2001, Why Aren t the Prices of the Same Item the Same at Me.com and
You.com?: Drivers of Price Dispersion Among E-Tailers, MIT Sloan School of Management,
Cambridge, MA
30
Page 32
38. Pan, Ratchford, Shankar 2002, Can price Dispersion in Online Markets be Explained by
Differences in e-Tailer Service Quality? MIT Sloan School of Management, Cambridge, MA
39. Peterson, Robert A., S. Balasubramanian, and B. J. Bronnenberg (1997), "Exploring the
Implications of the Internet for Consumer Marketing, Journal of the Academy of Marketing
Science, 25, 329-346
40. P. McCullagh and J.A. Nelder, Generalized Linear Model, 2
nd
edition, Chapman & Hall/CRC,
chapter 4
41. Pinelopi K. Goldberg, Frank Verboven. 2001. Market Integration and Convergence to the Law of
One Price: Evidence from European Car Market. NBER Working Paper#8402
42. Quelch, John A. and Lisa R. Klein (1996), The Internet and International Marketing, Sloan
Management Review, Spring, 60-75.
43. Salop, S., J. E. Stiglitz. 1982. The theory of sales: A simple model of equilibrium price
dispersion with identical agents. The American Economic Review 72(5) 1121-1130
44. Sarah Spiekermann, Martin Strobel, Dirk Temme, 2001, Drivers and Impediments of Consumer
Online Information Search: Self-controlled versus Agent-based Search in a High Involvement
Context, eCommerce Research Forum, MIT
45. Sebastian Maurer, Bernardo Huberman, 2000, Competitive Dynamics of Web Sites, eCommerce
Research Forum, MIT
46. Smith Michael. 2001. The Law of One Price? Price dispersion and Parallel Pricing in Internet
Markets. Working Paper, Carnegie Mellon University, Pittsburgh, PA.
47. Smith, Michael D., Joseph Bailey, Erik Brynjolfsson. 2000. Understanding Digital Markets:
Review and Assessment, Brynjolfsson and Kahin eds. Understanding the Digital Economy, MIT
Press, Cambridge, MA, 99-136.
31
Page 33
48. Stiglitz, Joseph E. 1987. Competition and the number of retailer in a market: Are duopolies more
competitive than atomistic market? Journal of Political Economy 96(5)1041-1061
49. Tim Futing Liao, Interpreting Probability Models Logit, Probit, and Other Generalized Linear
Models, a SAGE university paper 07-101
50. Urban, Glen; Sultan, Fareena; Qualls, William. 1998, Trust Based Marketing On The Internet,
MIT Sloan School of Management, Cambridge, MA
51. Varian, Hal R. 1980. A model of sales. The American Economic Review 70(4) 651-659
52. Varian, Hal R. 2000. Market Structure in the Network Age. E. Brynjolfsson, B. Kahin, eds.
Understanding the Digital Economy, MIT Press, Cambridge, MA
53. Ward, Morganosky 2000, Does Online Retailing Cannibalize Other Marketing Channels?
eCommerce Research Forum, MIT
32