factors influencing the types of products and services purchased over the internet
TRANSCRIPT
Factors influencing thetypes of products andservices purchased overthe Internet
Ian Phau and
Sui Meng Poon
Introduction
Following the commercialisation of the
Internet in the USA, many Asian countries
have since begun promoting the use of the
Internet by setting up state-run Internet
service providers (ISP) and implementing
policies to encourage educational institutions
and business to go online. The market
forecast is that Singapore will experience a
sharp increase in Internet and World Wide
Web (WWW) usage towards the end of the
century. It is expected to increase from
198,773 users in 1998 to 1,341,050 users in
2001 whereby 17.3 per cent and 34.3 per cent
respectively will buy via the Net. The overall
electronic commerce would also increase
from US$20.97 million to US$855.43 million
for the same period of time. As consumer
confidence grows with the familiarity of
electronic commerce and secure online
transactions, we will see an increase in overall
e-commerce expenditure for the Asia Pacific
region. The study also predicts that the
overall electronic commerce trade will reach
US$16,538.18 million in 2001 for the Asia
Pacific region.
However, the attitudinal effects of in-home
shopping via the Internet have received little
direct research attention so far. Related
research has dealt primarily with the more
traditional direct modes of shopping, such as
mail/phone shopping (Settle et al., 1994), or
more restrictive forms of electronic shopping
such as videotex and television shopping
(Eastlick and Liu, 1997). As electronic
commerce over the Internet increases, it will
become more important that Internet
marketers or cyber mall operators have some
basis to better market their products or
services over the Internet. The more the
Internet shopping malls' marketers
understand the underlying reasons for
differences in the consumer choices, the more
effectively and profitably they can serve their
markets. As such there is an inherent need to
investigate the nature and perceptions of
shoppers, and the suitability of different types
of products and services with respect to
Internet buying in Singapore and to a larger
extent the Asia Pacific region.
The authors
Ian Phau ([email protected]) is at Henley Management
College, Henley-on-Thames, UK.
Sui Meng Poon is at Surrey European Business School, UK.
Keywords
Electronic commerce, Internet, Shopping,
Consumer behaviour, Singapore
Abstract
Presents the findings of an empirical investigation of
Internet shopping in Singapore. Specifically, Internet buying
behaviour is compared between potential Internet buyers
and non-Internet buyers. It was found that the classification
of different types of products and services will significantly
influence the consumer choice between a retail store and
Internet shopping mall. The types of products and services
that are suitable for selling through the Internet are also
identified. Generally, products and services that have a low
outlay, are frequently purchased, have intangible value
proposition, and relatively high on differentiation are more
likely to be purchased via the Internet. Based on the analysis
and results obtained from the study, some recommendations
are presented to retailers who are interested in selling their
products via the Internet.
Electronic access
The current issue and full text archive of this journal is
available at
http://www.emerald-library.com
102
Internet Research: Electronic Networking Applications and Policy
Volume 10 . Number 2 . 2000 . pp. 102±113
# MCB University Press . ISSN 1066-2243
Literature review
The Internet as a marketing channel
It has been reported in prior research that many
factors influence the choice between a retail
store and the in-home shopping method such as
mail/phone order and the Internet. Some of
which include socio-economic and
demographic factors (Cunningham and
Cunningham, 1973; Kono and Buatsi, 1984);
perceived purchase risk (Spence et al., 1970 and
Prasad, 1975); product type and distribution
methods (Akaah and Korgaonkar, 1989 and
Reynolds et al., 1977); personal traits or
characteristics (Cunningham and Cunningham,
1973) and shopping or delivery time (Arora,
1991 and Sharma et al., 1983). Other factors
include intrusiveness; confrontation and
contact control (Brezen
et al., 1987); manufacturer or brand reputation
(Akaah and Korgaonkar, 1989); type and
source of the offer (Korgaonkar, 1984;
Reynolds et al., 1977); price and refund or
exchange privileges (Arora, 1991). When
in-house shopping is extended to the Internet
shopping malls, these factors will be more
apparent. As a marketing channel, the Internet
has both unique characteristics and
characteristics that are shared with other
marketing channels. For instance, it has the
ability to store large amounts of information at
different virtual locations and provide
information to the consumers on demand.
There is also the advantage of a physical
distribution medium for certain goods (for
example, software) with relatively low entry and
establishment costs for sellers. As such, some
frameworks have been developed to help us
understand the process of online marketing.
Welz (1995) illustrates many of these issues
and proposes that online marketing should be
perceived as five components namely:
(1) promotions;
(2) one-to-one contact;
(3) closing;
(4) transaction; and
(5) fulfilment.
The first component, `̀ promotions'' refers to
advertising and creating awareness for users to
learn about the products. For `̀ one-to-one
contact'' the Internet allows for asynchronous
contact via other methods such as e-mail and
discussion groups and these can be used to
develop a relationship with a customer. The
third component `̀ closing'' involves setting the
price and closing the deal. While this can be
achieved through a collection of e-mail
messages, Internet buying is more like ordering
from a catalogue. Conversely, there is an
electronic record of the sale and the deal struck.
The fourth component `̀ transaction'' refers to
the payment for goods. Until recently, such
transactions were difficult over the Internet.
Increasing numbers of companies are
introducing systems to enable secure credit card
transactions on the Internet. Finally,
`̀ fulfilment'' involves the delivery of goods.
Delivery online is restricted to digital products
such as electronic books and software.
Fulfilment includes not only shipping, but also
effective electronic order processing.
In the same vein Peterson et al. (1997)
propose another model of Internet buying.
They perceive that the marketing activities can
occur through three types of channels:
distribution, transaction and communication
channels. The function of the distribution
channels is to facilitate the physical exchange of
products and services. Transaction channels
generate sales activities between buyers and
sellers. Finally, communication channels enable
the exchange of information between buyers
and sellers.
Putting these frameworks together, the
Internet comes across as a flexible, interactive,
and efficient medium through which economic
parties can communicate. The potential that it
offers for efficiency improvements in channel
functions will vary across different types of
intermediaries. However it is not without
caveats. Some appeals may not be able to be
transcended across the Web page.
Products sellable through the Internet
One weakness of the Internet is that it can
realistically reproduce only two of our five
senses, namely, sight and sound. One cannot
feel, smell or taste products that are advertised
on the WWW. This limitation will restrict the
kind of products that are sellable on the
Internet. Stern (1995) suggested that products
that make use of the hypermedia advantages of
WWW would be suitable. For example,
103
Factors influencing products and services purchased over the Internet
Ian Phau and Sui Meng Poon
Internet Research: Electronic Networking Applications and Policy
Volume 10 . Number 2 . 2000 . 102±113
although the WWW appeals only to the sight
and sound of customers, this capability is
sufficient for customers to sample a certain
category of products. CDNOW! (http://
www.cdnow.com), for instance, offers buyers a
short audio clip on the music or songs available
for sale. One can even purchase art pieces
online by viewing the uploaded pictures (http://
www.kspace.com/).
Despite multimedia presentations, shopping
on the Internet is no substitute for the leisure
experience associated with conventional
shopping. Resnick (1995) claims that `̀ sellers of
mainstream goods who think they will succeed
on the Internet simply because they're making
purchasing more important conveniently miss
the point ± shopping at their sites must be fun to
attract customers''. In some context however,
Internet shopping can be inherently attractive
when compared to normal shopping. For
example, in retailing CDs, it is normal for the
buyer to go through lists of catalogues and
browse through racks of discs during purchase.
In the Internet CD store, a good database
management can provide easy access to this
information. A further benefit to the customer
is that it can be done at the comfort of the home
and at a time conducive to the buyer.
The suitability of the Internet for marketing
to consumers depends to a large extent on the
characteristics of the products and services
being marketed. It is therefore necessary to
explicitly consider product and service
characteristics when evaluating the impact of
the Internet. This can be done by formally
incorporating a product and service
classification into any analysis.
Peterson et al. (1997) suggested a
classification system in which products and
services are categorised along three dimensions:
(1) cost and frequency of purchase;
(2) value proposition; and
(3) degree of differentiation.
Goods vary along the first dimension from low-
cost, frequently purchased goods (e.g.
consumable products such as milk) to high-
cost, infrequently purchased goods (e.g.
durable products such as stereo systems). In
general, when purchase fulfilment requires
physical delivery, the more frequent the
purchase and the smaller the cost (e.g. milk),
the less likely there is to be a good `̀ fit'' between
a product or service and the Internet-based
marketing.
Goods vary along the second dimension
according to their value proposition, that is, if
they are tangible and physical or intangible and
service related. Internet-related marketing is
particularly well suited to certain types of
intangible or service-related goods (i.e. those
based on digital assets). To the extent that the
value proposition is intangible, the greater the
frequency of purchase or use of a good, the
greater the advantage of the Internet as a
transaction and distribution medium.
The third dimension reflects the degree to
which a product or service is differentiable. In
particular, it reflects the extent to which a seller
is able to create a sustainable competitive
advantage through product and service
differentiation. Internet-related marketing can
result in extreme price competition when
products or services are incapable of significant
differentiation. However, when products or
services are capable of significant
differentiation, the Internet can serve as an
effective segmentation mechanism for guiding
buyers to their ideal product or service. Table I
reflects the three dimensions together with
examples of products and services.
Empirical research in Singapore
The Internet shopping experience is also
captured in a study conducted by Temasek
Polytechnic in January 1998. Out of 1,800
respondents, only about 16 per cent of the
respondents had ever bought from the Internet.
Most of the respondents are young, between the
ages of 15 to 29, and over 74 per cent are single.
About 86 per cent viewed the Internet shopping
as either positive or neutral. The survey also
revealed that 15 per cent of the cyber-buyers
cited that, gifts, novelties, flowers and hampers
are the most popular category. Only 2 per cent
of the respondents chose the footwear category
as one of their most frequently purchased
categories. In addition, the study also shows
that convenience (32 per cent) was the most
important element in cyber-shopping followed
by the merchandise (22 per cent) itself. Another
19 per cent said that the promotion offered
would affect their purchase decision and only
12 per cent cited impulsive buying as their
104
Factors influencing products and services purchased over the Internet
Ian Phau and Sui Meng Poon
Internet Research: Electronic Networking Applications and Policy
Volume 10 . Number 2 . 2000 . 102±113
purchase. Of those who did not buy, it was
noted that 26 per cent of the non-cyber-buyers
stated that they preferred the real-life shopping
experience, that is, shopping in the store.
However, 44 per cent of them were worried
about the merchandise quality (18 per cent),
security in Internet shopping (15 per cent) and
trustworthiness of the company (11 per cent).
The survey suggests that the public is
generally positive towards cyber-shopping but
they do not purchase often in the Internet. They
usually buy through the Internet due to
convenience, wider merchandise selection and
promotional offers. Generally, they prefer to
purchase items with no sizing requirements.
Summary and research objectives
Previous research on in-home shopping has
certainly provided some understanding of the
in-home shoppers. However, most are done
using Western samples (For example, Aldridge
et al., 1997; Eastlick and Lotz, 1999; White and
Manning, 1998). Although there is an
increasing growth of electronic commerce in
Singapore and the Asia Pacific region, research
in this region is still conspicuously deficient. To
date, only one cross-national study has been
uncovered. This compares Internet strategies
between Japanese and US banks. Simeon
(1999) found that the contrasting approaches to
Web design information and service delivery are
due to cultural and institutional differences.
With this evidence, it can be assumed that
shoppers from different cultural and national
backgrounds may have influenced Internet
purchase intention. Further, Johnston and Johal
(1999) highlight that the Internet has developed
into a heterogeneous, dynamic and antagonistic
population and thus is an interesting subject for
cultural analysis. While the attempt to build on
Hofstede's cultural model offers some
directions, empirical studies involving Asian
culture are still deficient. The study conducted
by Temasek Polytechnic has shed some light on
generalising some of the demographic and
attitudinal attributes of Singapore Internet
shoppers. However, some questions remain
unanswered. For instance, how different are
cyber shoppers across different nations and
cultures? Do we have a `̀ global'' tribe of cyber
shoppers? Does Internet shopping differ
significantly by product types?
What is lacking is the specificity of a
framework in which strategists, marketers and
academics alike, can draw useful information to
generalise the buying behaviour of cyber-
shoppers. Building on Peterson et al.'s (1997)
classification of product and services, the
consumer choice between a retail store and
Internet shopping mall may be influenced by
the type of products and services offered.
Furthermore, price, frequency of purchase,
product tangibility and differentiation may be
factors affecting the mode of purchase.
Building on the preceding discussion, the
following hypotheses are proposed:
H1: Product and service type classification
will significantly influence the
consumer choice between a retail store
and Internet shopping mall.
H2a: Products and services that are relatively
expensive and infrequently purchased
are more amenable to be purchased via
the Internet.
H2b: Products and services that have
intangible value proposition are more
amenable to be purchased via the
Internet.
Table I Product classification
Dimension 1 Dimension 2 Dimension 3 Products and services
Low outlay, frequently Value proposition tangible or physical Differentiation potential high Wines, soft drinks, cigarettes
purchased goods Differentiation potential low Milk, eggs
Value proposition intangible or Differentiation potential high Online newspapers and magazines
informational Differentiation potential low Stock market quotes
High outlay, infrequently Value proposition tangible or physical Differentiation potential high Stereo systems, automobiles
purchased goods Differentiation potential low Precious metal ingot of known weight and purity
Value proposition intangible or informational Differentiation potential high Software packages
Differentiation potential low Automobile financing, insurance
105
Factors influencing products and services purchased over the Internet
Ian Phau and Sui Meng Poon
Internet Research: Electronic Networking Applications and Policy
Volume 10 . Number 2 . 2000 . 102±113
H2c: Products and services that are relatively
high on differentiation are more
amenable to be purchased via the
Internet.
Research methodology
Research instrument
The questionnaire was divided into two sections:
Section A of the questionnaire consisted of a list
of questions that gathered the demographic
variables of the respondents. In addition, one
question was designed to find out the likelihood
of the respondents in making purchases over the
Internet. Respondents were asked to indicate
their intention to buy over the Internet on a six-
point scale (1 = very unlikely and 6 very likely).
Based on a median split approach, we could then
classify the respondents into two groups, namely
potential Internet buyers (i.e. willing to consider)
and non-Internet buyers (i.e. not willing to
consider).
Section B of the questionnaire asked the
respondents to indicate how likely they were to
purchase each of the 20 listed products through
the Internet on a five-point scale (1 = very
unlikely, to 5 = very likely). The list was
adopted from Peterson et al.'s (1997)
classification of product and services. The
sample products were classified into various
categories according to costs, purchase
frequency, product tangibility and
differentiation. The Appendix reflects the entire
list. In addition, the respondents were also
asked directly on their preference on each of the
classification parameters, i.e. cheap-expensive,
frequently-infrequently purchase, intangible-
tangible, low-high differentiation.
Data collection
A pilot test to check on the readability and
reliability of the survey instrument was
conducted before the official data collection. A
total of 280 electronic mails were sent out with
a link to the Web site where the questionnaire
was displayed. The population coverage of the
questionnaire included the general public (20
per cent), students (23 per cent), and
employees of public and private organisations
(37 per cent) and patrons of some commercial
electronic commerce Web sites (20 per cent).
The overall response rate was around 60.6 per
cent, which yielded a total of 183 usable
questionnaires. Hence, the effective response
rate was 55.4 per cent. This is considered
adequate for analysis and reporting (Babbie,
1990).
Of those who gave usable information, 85
were females. The average age of the
respondents was 23.4, and their average weekly
Web use was 6.06 hours. A total of 47
respondents had purchased products over the
Internet while 136 had never tried before.
Using the median split approach, 70.5 per cent
of the respondents were classified as potential
non-Internet buyers, while 29.5 per cent were
classified as potential Internet buyers. The most
likely purchased categories of products were
CDs, online videos/music, online paid
subscription to financial reports and stock
market quotes and computer software.
Findings and analysis
The results of products that are likely to be
purchased through the Internet are tabulated in
Table II. The results show that the means of
most products were higher for Internet buyers
than for non-Internet buyers. Independent
samples t-tests reveal that some products were
found to be significantly more likely to be
purchased by Internet buyers than non-Internet
buyers. These include flowers, online paid
subscription to newspapers and financial
information, online videos/music, computer
software, consultancy services, car loans and
insurance. Conversely, there were also some
products found to be significantly more unlikely
to be purchased by Internet buyers than non-
Internet buyers. These include milk, eggs and
vegetables.
To better understand the influence of each of
the product attribute, namely low outlay and
frequently purchased versus high outlay and
infrequently purchased, tangible or physical
versus intangible or informational, and low
differentiation versus high differentiation, the
data was recoded according to each product
attribute as shown in Table III.
Paired sample t-tests were then conducted to
evaluate the consumer receptiveness of each
product class. The results are shown in Table IV.
106
Factors influencing products and services purchased over the Internet
Ian Phau and Sui Meng Poon
Internet Research: Electronic Networking Applications and Policy
Volume 10 . Number 2 . 2000 . 102±113
Significant differences in each of the product
classes were found. Low outlay and frequently
purchased goods are better received than high
outlay and infrequently purchased goods.
Similarly, intangible or informational goods are
better received than tangible or physical goods,
while products with high differentiation are more
suited to be sold on the Internet than products
with low differentiation.
In general, the analysis suggested that
computer-related products (with the exception
of flowers) could be marketed through the
Internet. This could be due to the high
technology orientation of the Internet users.
Also, the Internet medium can become the
delivery medium for software. Thus a software
package can be delivered digitally and be used
by the customer literally in minutes. Another
reason is that the Internet enables online trial of
such products. Internet buyers, for example, are
able to try out the demonstration version of
computer software, or be given trial periods of
the online newspapers or video/music, before
making a purchase decision. This reduces
uncertainty in purchase decision and stimulates
purchases.
It also appears that products and services with
low outlay, frequently purchased, tangible or
physical, and with a low differentiation
potential, i.e. category B products, are
unsuitable for selling through Internet in
Singapore. In fact, both Internet buyers and
non-Internet buyers had relatively low purchase
intentions for such goods. Expensive items such
as cars, stereo systems and jewellery are also not
suitable to be sold over the Internet. These
products require more than visual inspection.
For example, a car needs to be test-driven and a
jewellery piece needs to be physically examined
by the potential buyer in order to appreciate its
value. In general, the analysis suggests that
products capitalising on the advantages of the
Internet are more likely to be bought through
the Internet.
Discriminant analysis
To analyse the relationship between the
product categories and shopping behaviour,
discriminant analysis was applied to predict the
relationship between the product categories and
Internet buying. The data on product categories
was summed accordingly so as to obtain the
Table II T-tests of products between non-Internet buyers and Internet buyers
Non-Internet buyer Internet buyer p-value
Products Mean Standard deviation Mean Standard deviation (two-tail)
Wines 1.9922 1.2593 2.3148 1.3150 0.121
CDs 3.2713 1.3214 3.6111 1.3929 0.120
Flowers 2.6822 1.4251 3.4815 1.3138 0.001*
Milk 1.9535 1.3685 1.4074 1.0003 0.009*
Eggs 1.9845 1.3919 1.5556 1.2079 0.050*
Vegetables 1.8140 1.2548 1.4074 1.0003 0.036*
Online newspapersa 2.9845 1.5461 3.6667 1.3318 0.005*
Online videos/music 3.0930 1.4222 3.7037 1.2074 0.006*
Online financial informationa 3.0388 1.4969 3.7963 1.2190 0.001*
Stock market quotes 3.0078 1.5537 3.4444 1.4879 0.081
Weather information 2.8915 1.5270 3.2407 1.7149 0.175
Stereo systems 1.8140 1.1711 1.9630 1.2733 0.445
Automobiles 1.5581 0.9430 1.4074 0.5993 0.279
Jewellery 1.7752 1.2577 1.9074 1.1205 0.504
Gold, silver ingots/bars 1.4729 0.8011 1.5741 0.9635 0.465
Pigs/cows, livestock 1.4109 0.8625 1.3704 0.7842 0.767
Computer software 2.8760 1.3694 3.8889 1.1271 0.000*
Consultancy services 2.3643 1.3047 2.9630 1.3593 0.006*
Car loans 1.9845 1.0968 2.4630 1.5627 0.019*
Insurance 1.8915 1.0770 2.5370 1.4885 0.001*
Notes: 1 = very unlikely, 5 = very likely; apaid subscription; *significant at p 5 0.5
107
Factors influencing products and services purchased over the Internet
Ian Phau and Sui Meng Poon
Internet Research: Electronic Networking Applications and Policy
Volume 10 . Number 2 . 2000 . 102±113
composite scores for each product category.
The samples were again divided into two parts
the analysis sample (70 per cent) and the
holdout sample (30 per cent). The analysis
sample was used to estimate the discriminant
function while the holdout sample was used to
validate the discriminant function. A two-group
discriminant analysis was used where a
dichotomous variable was created to separate
the respondents into potential Internet buyers
and non-Internet buyers. A median-split was
applied in which 100 respondents were
identified to be non-Internet buyers while 37
were grouped as Internet buyers.
The data were analysed using the SPSS
program DISCRIMINANT function, adopting
a stepwise procedure. At each step, the variable
that minimised the overall Wilks' Lambda was
entered. The eigenvalue was 0.227 and Wilks'
Lambda = 0.815. The procedure indicated
product Category A (0.533), Category B
(±0.438), Category E (±0.765), Category F
(0.418), Category G (0.435), and Category H
(0.330) as the six discriminating product
Table III Recoding of product data
Product attribute Categories data to be drawn from
Low outlay and frequently purchased goods Category A ± wines, CDs, flowers
Category B ± milk, eggs, vegetables
Category C ± online newspapers, videos/music, financial reports
Category D ± stock market quotes, weather information
High outlay and infrequently purchased Category E ± stereo systems, automobiles, jewellery
Category F ± gold, silver ingots/bars, pigs/cows, livestock
Category G ± computer software, consultancy services
Category H ± car loans, insurance
Tangible or physical Category A ± wines, CDs, flowers
Category B ± milk, eggs, vegetables
Category E ± stereo systems, automobiles, jewellery
Category F ± gold, silver ingots/bars, pigs/cows, livestock
Intangible or information Category C ± online newspapers, videos/music, financial reports
Category D ± stock market quotes, weather information
Category G ± computer software, consultancy services
Category H ± car loans, insurance
High differentiation Category A ± wines, CDs, flowers
Category C ± online newspapers, videos/music, financial reports
Category E ± stereo systems, automobiles, jewellery
Category G ± computer software, consultancy services
Low differentiation Category B ± milk, eggs, vegetables
Category D ± stock market quotes, weather information
Category F ± gold, silver ingots/bars, pigs/cows, livestock
Category H ± car loans, insurance
Table IV Paired sample t-tests on product class
Product class Mean Standard deviation P-value
Low outlay and frequently purchased 2.6885 0.7934
High outlay and infrequently purchased 2.0012 0.76880.000*
Tangible or physical 1.9826 0.7046
Intangible or informational 2.8640 0.99500.000*
Low differentiation 2.0650 0.6938
High differentiation 2.6364 0.80670.000*
Notes: 1 = very unlikely; 5 = very likely; *significant at p 5 0.5
108
Factors influencing products and services purchased over the Internet
Ian Phau and Sui Meng Poon
Internet Research: Electronic Networking Applications and Policy
Volume 10 . Number 2 . 2000 . 102±113
categories between the Internet and non-
Internet buyers. The highest p-value was only
3.962 6 10-4 (the unstandardised coefficients
are shown in parenthesis). The group means
(centroids) of the discriminant function is -
0.274 for non-Internet buyers and 0.814 for
Internet buyers. The results give us the
following estimated discriminant function:
D = ±1.942 + 0.533 Category A ±
0.438 Category B± 0.768 Category E + 0.418
Category F+. 0.435 Category G + 0.330
Category H.
To determine the predictive accuracy, a
classification matrix was constructed. The
discriminant function had correctly classified
68.1 per cent of respondents in the analysis
sample and 68.8 per cent in the holdout
sample. The proportional chance criterion was
employed to assess the usefulness of the
function. Cpro was found to be 60.57 per cent,
which is lower than both the prediction
accuracy of 68.1 per cent for the analysis
sample and the 68.8 per cent for the holdout
sample. Therefore, the prediction accuracy was
considered acceptable (Hair et al., 1995) and
validated the usefulness of the discriminant
function. Table V presents the statistical details.
The relative importance of the predictors can
also be obtained by examining the structure
correlation of the discriminant function
obtained from the analysis. As a result, products
in Category G were most successful in
differentiating Internet buyers from
non-Internet buyers while products in Category
E had the least discriminating power. Category
G are essentially products and services that are
relatively expensive, infrequently purchased,
have intangible value proposition, and relatively
high on differentiation are more amenable to be
purchased via the Internet. Table VI presents
the statistical details.
Discussion
The findings of this research help to profile a list
of products that are suitable for sale in an
interactive electronic shopping medium. In so
doing, it assists marketers in introducing and
promoting certain products that have more
success potential in this mode of retail. Generally,
the preceding section on the findings cohere very
well with Peterson et al. (1997) hypotheses that
Internet-relating marketing is more suitable for
intangible or service-related, as well as high
differentiation goods. Conversely, the hypothesis
that the greater the frequency of purchase and the
smaller the cost, the less likely there is to be a
good `̀ fit'' between a product or service and the
Internet-based marketing is not supported. As a
summary, Table VII provides the results of the
hypotheses testing.
Certainly, at least within the Singapore context,
expensive goods like automobiles, jewellery, and
stereo systems are not ready for Web selling. The
monetary risks involved in buying these products
are too great. These products also require more
than visual inspection. This reflects some of the
Table V Classification resultsa, b
Predicted group membership
Non-Internet buyer Internet buyer Total
Original cases selected
Count Buyer Non-Internet buyer 70.0 31.0 101
Internet buyer 12.0 22.0 34.0
Percentage Buyer Non-Internet buyer 69.3 30.7 100.0
Internet buyer 35.3 64.7 100.0
Cases not selected
Count Buyer Non-Internet buyer 20.0 8.0 28.0
Internet buyer 7.0 13.0 20.0
Percentage Buyer Non-Internet buyer 71.4 28.6 100.0
Internet buyer 35.0 65.0 100.0
Notes: a68.1 per cent of selected original grouped cases correctly classified; b68.8 per cent of unselected original grouped cases correctlyclassified; Cpro = p2 + 1 ±p)2 where p = proportion of individuals in group 1; (1±p) = proportion of individuals in group 2
109
Factors influencing products and services purchased over the Internet
Ian Phau and Sui Meng Poon
Internet Research: Electronic Networking Applications and Policy
Volume 10 . Number 2 . 2000 . 102±113
findings in prior studies (such as Eastlick, 1991;
Shim and Mahoney, 1991) whereby perceived
risk tends to impede the adoption of non-store
retailers such as online shopping systems.
Furthermore, Internet shopping is generally still
unfamiliar to most Internet users, and there exists
a certain amount of uncertainty regarding the
retailers' trustworthiness and credibility. Relating
this to prior studies, Eastlick and Lotz (1999)
found that their US sample of non-adopters of
Internet shopping generally have higher negativity
towards this mode of shopping as compared to
adopters. While it is difficult to generalise this to
the current sample, it is not surprising that the
findings of this current sample reflect similar
concerns.
Category B products, which are those with
low outlay, are frequently purchased, tangible
or physical, and with a low differentiation
potential, are unsuitable for selling through the
Internet in Singapore. This is because these
products are easily available in the
neighbourhood shops at low prices. Moreover,
these products are usually bought when the
need arises and very little search for pre-
purchase information is needed. This seems to
reflect the findings of White and Manning
(1998) whereby more consumers are more
likely to buy food and drinks from retail outlets
(75.2 per cent) than through online sources
(29.7 per cent).
For a consumer wishing to experience the
merchandise prior to purchase, Internet-based
marketing would seem to be a poor substitute for
traditional retail channels, where the good is
available for inspection. However, a consumer
might use a traditional retail channel to
experience the merchandise and then revert to an
Internet-based channel when acquiring it.
Additionally, as a communication channel, the
Internet will be increasingly able to offer
perceptual experiences that far transcend verbal
descriptions of goods. Pictures of flower bouquets
or fruits can be presented in great detail, and
music from a compact disc can be sampled
online. When products are capable of significant
differentiation, the Internet can serve as an
effective segmentation mechanism for guiding
buyers to their ideal product or service. For
example, consider a prospective buyer in the
market for computer virus protection software.
Searching on the Internet enables the prospective
buyer to obtain information on various competing
products, possibly sample the products for free,
and select the one that best meets his or her
requirements. Compared to conventional
retailing, such detailed search and personal
sampling would be costly. Finally, for frequently
purchased goods, in which a consumer has
considerable personal experience, the Internet
may serve as an effective communication and
transaction medium since the perceived risks
would be very much reduced.
In general, the analysis suggests that products
capitalising on the advantages of the Internet
are more likely to be bought through the
Internet. Electronic retailers should therefore
focus on computer-related products, or
Table VII Summary of results of hypotheses
Hypothesis Description p < 0.05
H1 Product and service type classification will significantly influence the
consumer choice between a retail store and Internet shopping mall
Supported
H2a Products and services that are relatively expensive and infrequently purchased
are more amenable to be purchased via the Internet
Not supported
H2b Products and services that have intangible value proposition are more amenable
to be purchased via the Internet
Supported
H2c Products and services that are relatively high on differentiation are more
amenable to be purchased via the Internet
Supported
Table VI Structure coefficients of the discriminant function
Variables Structure coefficients
Category G 0.662
Category H 0.524
Category A 0.471
Category C 0.409
Category D 0.317
Category B ±0.314
Category F 0.170
Category E ±0.014
110
Factors influencing products and services purchased over the Internet
Ian Phau and Sui Meng Poon
Internet Research: Electronic Networking Applications and Policy
Volume 10 . Number 2 . 2000 . 102±113
products which allow trial sampling or those
that are high on information content. Examples
of these include computer software, CDs,
online newspapers, online videos/music, online
financial and stock information.
Since the search costs for product
information through the Internet are low,
buyers can easily compare similar products
across different Web site companies with a click
of the mouse. As Internet buyers value
convenience, electronic retailers should keep
their Web sites open 24 hours a day and
equipped with an online ordering mechanism
for easy impulse buying. Thus buyers can easily
obtain pre-purchase information and compare
products at any time of the day to make
immediate purchases.
The study also advocates that retailers carry
familiar brands and speciality products, a point
highlighted in Eastlick and Lotz's (1999) study.
It is logical to infer that Internet buyers will
prefer well-known brands, as they are confident
of the quality assurance. Otherwise, buyers may
easily switch to another Internet shopping mall
since the search costs are low.
Another way of providing a wider range of
merchandise is by pooling the resources of
different electronic retailers to establish a virtual
shopping mall. A virtual shopping directory that
classifies the retailers according to the products
offered will make it easier for the buyer to shop.
Since the shopping mall is virtual, new retailers
can also be easily added to the mall.
Alternatively, electronic retailers selling similar
or complementary products could form a
specialist shopping mall. For instance, a one-
stop travel Web site could be established that
links up all the travel agencies, providing
comprehensive information on the various
destinations, travel tips, hotels, tour packages,
as well as offering online booking of tours,
flights, or hotels. It can even take a step further
to facilitate business-to-business transactions
between the various tour agencies, hotels and
airlines. The Internet buyers can then browse
through all the products via a single Web site.
This will help generate more hits at the Web site
because it provides convenience to the buyer, a
factor that is highly valued as shown in the
earlier study.
The main managerial implication of this
study is that only certain profiles of goods are
suitable for an interactive environment in
Singapore. Peterson et al.'s (1997) classification
framework of goods sold in the interactive mode
is still at its infancy in Western studies, thus
direct comparison of this study with others is
not possible. However, some inference has been
drawn with relevant and related studies
wherever feasible. Through the comparisons
discussed in the preceding sections, it appears
that there are a number of similarities in terms
of shopping motivation and the appropriateness
of goods sold over the Internet. While Simeon
(1999) concluded that there are significant
differences between Japanese and US Web-site
strategies, Internet shoppers may well exhibit
several demographic and psychographic
similarities. This trait is captured in the concept
of the Internet as a virtual cultural region as
suggested by Johnston and Johal (1999).
Certainly, more cross-national studies have to
be conducted to achieve a more conclusive
finding.
What can be concluded however, is that the
design of the Internet shopping environment
must be able to deliver potential advantages
over other alternative shopping formats such as
retail stores. `̀ Convenience'', a factor
commonly cited in Eastlick and Lotz (1999)
and the study conducted by Temasek
Polytechnic (1998) has to be dealt with caution.
While goods such as gifts (for instance, flowers
and cards) fall into this category for both
studies, Category B products should not be
assumed to achieve similar responses from
Singapore and the Western samples.
Geographically, Singapore is a very densely
populated country whereby neighbourhood
shops are within easy reach, with many within
walking distance. This may not be true in the
rural areas of the Western countries with more
diverse geography. This probably explains why
Amazon Bookstore is less popular and
successful in Singapore. In the same vein, it
may be possible to generalise the findings of this
study across other Asian countries of similar
social and economic development such as Hong
Kong, Taiwan and Japan. However, less
developed economies such as Indonesia and the
Philippines may not reflect the same responses
due to a lower exposure of the Internet and thus
the adoption of interactive shopping may be
severely restricted. In short, marketers must
111
Factors influencing products and services purchased over the Internet
Ian Phau and Sui Meng Poon
Internet Research: Electronic Networking Applications and Policy
Volume 10 . Number 2 . 2000 . 102±113
demonstrate to non-Internet consumers that
interactive shopping is convenient, safe and
simple to use.
Conclusion
This study has presented empirical findings of
Internet shopping in Singapore comparing the
Internet and non-Internet buyers. Product and
service type classification will significantly
influence the consumer choice between a retail
store and the Internet shopping mall. In
addition, the types of products or services that
are suitable for selling through the Internet were
also identified. In summary, the products and
services that have a low outlay, are frequently
purchased, have intangible value proposition,
and are relatively high on differentiation are
more likely to be purchased via the Internet.
Based on the analysis and results obtained from
the study, some suggestions were made for
retailers who are interested in selling their
products via the Internet.
A number of limitations are apparent. First,
the convenience sample based on a quota,
rather than a truly random sample. Some
selection bias may still have intruded the
results. Second, the selection of image items is
elicited from previous research on both mail/
phone order marketing and on purchase
decision making for retail store shopping goods
in Settle et al. (1994). Limitations on the survey
response task precluded the use of more than
eight items, and those included were judged to
be the most likely to influence the Internet/retail
shopping medium purchase decision. Future
research should try to eliminate these issues to
provide more generalisable and useful results.
Studies should also be extended to find out
whether products with higher brand awareness
would enjoy better sales in Internet purchase
compared to retail store purchase. The other
obvious lead will be to replicate this study in
Western and other Asian cultures to make
explicit comparisons.
References
Akaah, I. and Korgaonkar, P.K. (1989), `̀ The influence ofproduct manufacturer, and distributor characteristics
on consumer interest in direct marketing offerings'',Journal of Direct Marketing, Vol. 3, Summer,
pp. 27-33.Aldridge, A., Forcht, K. and Pierson, J. (1997), `̀ Get linked or
get lost: marketing strategy for the Internet'', InternetResearch, Vol. 7 No. 3, pp. 161-9.
Arora, R. (1991), `̀ Formulating direct marketing offers with
conjoint analysis'', Journal of Direct Marketing, Vol. 5,Winter, pp. 48-56.
Babbie, E. (1990), Survey Research Methods, WadsworthPublishing Company, Belmont, CA.
Brezen, T.S., Block, M. and Schultz, D.E. (1987),`̀ Consumers' perception of direct marketing
techniques'', Journal of Direct Marketing, Vol. 1,Winter, pp. 38-49.
Cunningham, C.M. and Cunningham, W.H. (1973), `̀ The
urban in-home shopper: socio-economic andattitudinal characteristics'', Journal of Retailing,Vol. 49, Fall.
Eastlick, M.A. (1991), `̀ Catalog shoppers as potential
adopter of Videotex'', in King, R.L. (Ed.), `̀ Retailing:reflections, insights and forecasts'', The Academy ofMarketing Science, Richmond, VA, pp. 9-13.
Eastlick, M.A. and Liu, M. (1997), `̀ The influence of store
attitudes and other non-store shopping patterns onpatronage of television shopping programmes'',
Journal of Direct Marketing, Vol. 11 No. 3, pp. 14-24.Eastlick, M.A. and Lotz, S. (1999), `̀ Profiling potential
adopters and non-adopters of an interactiveelectronic shopping medium'', International Journalof Retail and Distribution Management, Vol. 27 No.6, pp. 209-23.
Hair, J.F., Anderson, R.E., Tatham, R.L. and Black, W.C.
(1995), Multivariate Data Analysis with Readings,Prentice-Hall, Englewood Cliffs, NJ.
Johnston, K. and Johal, P. (1999), `̀ The Internet as a `virtualcultural region': are extant cultural classification
schemes appropriate?'', Internet Research, Vol. 9No. 3, pp. 178-86.
Kono, K. and Buatsi, S. (1984), `̀ A comparative study ofthree direct marketing modes: catalogues, direct
mail, and direct ads on radio/television'', in Klein,D.M. and Smith, A.E. (Eds), Marketing Comes ofAge, Southern Marketing Association, Carbondale,IL, pp. 71-4.
Korgaonkar, P.K. (1984), `̀ Consumer shopping orientation,non-store retailers, and consumers' patronage
intentions: a multivariate investigation'', Journal of theAcademy of Marketing Science, Vol. 12, Winter,
pp. 11-12.Peterson, R.A., Balasubramanian, S. and Bronnenberg, B.J.
(1997), `̀ Exploring the implications of the Internet for
consumer marketing'', Journal of the Academy ofMarketing Science, Vol. 25 No. 4, pp. 329-46.
Prasad, V.K. (1975), `̀ Socioeconomic product risk andpatronage preferences of retail shoppers'', Journal ofMarketing, Vol. 39, July, pp. 42-7.
Resnick, R. (1995), `̀ Business is good, not'', Internet World,
June, pp. 71-3.
112
Factors influencing products and services purchased over the Internet
Ian Phau and Sui Meng Poon
Internet Research: Electronic Networking Applications and Policy
Volume 10 . Number 2 . 2000 . 102±113
Reynolds, F.D., Martin, W.S. and Martin, W.K. (1977),`̀ Media habits of in-home buyers'', Journal ofAdvertising, Vol. 6, Spring, pp. 32-5.
Settle, R.B., Airlock, P.L. and McCorkle, D.E. (1994),`̀ Consumer perceptions of mail/phone order shoppingmedia'', Journal of Direct Marketing, Vol. 8 No. 3,pp. 30-45.
Sharma, S., Bearden, W.O. and Teel, J.E. (1983), `̀ Differentialeffects of in-home shopping methods'', Journal ofRetailing, Vol. 59, pp. 24-51.
Shim, S. and Mahoney, M. (1991), `̀ Electronic shoppers andnonshoppers among Videotex users'', Journal of DirectMarketing, Vol. 3, Summer, pp. 29-38.
Simeon, R. (1999), `̀ Evaluating domestic and internationalWeb-site strategies'', Internet Research, Vol. 9 No. 4,pp. 297-308.
Spence, H.E., Engel, J.F. and Blackwell, R.D. (1970),
`̀ Perceived risk in mail order and retail store buying'',
Journal of Marketing Research, Vol. 7, August,
pp. 364-69.Stern, J. (1995), World Wide Web Marketing: Integrating the
Internet into Your Marketing Strategy, John Wiley &
Sons, New York, NY.Temasek Polytechnic (1998), `̀ Internet shopping in
Singapore'', unpublished report.Welz, G. (1995), `̀ Information highway revs up'', Computer
Reseller News, January, pp. 27-8.White, G.K. and Manning, B.J. (1998), `̀ Commercial WWW
site appeal: how does it affect online food and drink
consumers purchasing behaviour?'', Internet Research,
Vol. 8 No. 1, pp. 32-8.
Appendix
Table AI Classification of products
Product
categories Description
Products included
in questionnaire
Category A Goods that have a low outlay frequently purchased
tangible or physical and have a high differentiation
potential
Wine
CDs
Flowers
Category B Goods that have a low outlay frequently purchased
tangible or physical and have a low differentiation
potential
Milk
Eggs
Vegetables
Category C Goods that have a low outlay frequently purchased
intangible or informational and have a high differentiation
potential
Online newspapers
Online videos/music
Financial reports
Category D Good that have a low outlay frequently purchased
intangible or informational and have a low differentiation
potential
Stock market quotes
Weather information
Category E Goods that have a high outlay infrequently purchased
tangible or physical and have a high differentation
potential
Stereo systems
Automobiles
Jewellery
Category F Goods that have a high outlay infrequently purchased
tangible or physical and have a low differentiation
potential
Gold, silver ingots/bars
Pigs/cows, livestock
Category G Goods that have a high outlay infrequently purchased
intangible or informational and have a high differentiation
potential
Computer software
Consultancy services
Category H Goods that have a high outlay infrequently purchased
intangible or informational and have a low differentiation
potential
Car loans
Insurance
113
Factors influencing products and services purchased over the Internet
Ian Phau and Sui Meng Poon
Internet Research: Electronic Networking Applications and Policy
Volume 10 . Number 2 . 2000 . 102±113