Understanding consumer online shopping behaviour from the perspective of transaction costs

2.4.2 Major Research on TCs of Online Shopping 
Compared to the field of economics, relatively few IS studies have investigated the roles of 
TCs on individual consumer decision-making in the e-commerce context. Only limited 
studies applied TCT in online shopping, for exmaple Liang and Huang (1998), Teo et al. 
(2004), Teo and Yu (2005), Kim and Li (2009b), Kim et al. (2011), Kim et al. (2013b), Yen 
et al (2013) and Wu et al. (2014). Their studies provide valuable insights into the 
understanding of the roles of TCs in determining online consumer behaviour. The following 
section will review their work.

---

66
Liang and Huang (1998) applied TCT in the study of e-commerce adoption and examined 
what products are more suitable for marketing electronically and why. Their study posited 
that whether a customer would buy a product electronically is determined by the TCs of the 
channel and the TCs of a product on the web are determined by the uncertainty and asset 
specificity. In total, 86 Internet users were involved and five products with different 
characteristics (book, shoes, toothpaste, microwave oven, and flower) were used in the study. 
The TCs were measured by the costs associated with each of the stage of the online 
transaction process, including search cost, comparison cost, examination cost, negotiation 
cost, payment cost, delivery cost and post-service cost. The findings showed that the 
customer acceptance is determined by the TCs, which are in turn, determined by the 
uncertainty and asset specificity, and while experienced shoppers were concerned more about 
the uncertainty in e-shopping, inexperienced shoppers were concerned with both.
Although Liang and Huang’s (1998) findings had been encouraging and useful, the research 
was not without limitations. Firstly, the study only chose five products with different 
characteristics to represent all types of products. It might be problematic to draw a conclusion 
that different products have different customer acceptance simply based on five products. It is 
unclear whether the findings can be generalizable to all products and services on the web. A 
more valid classification of products may be needed. Secondly, as online environments have 
some notable differences compared to traditional shopping environments, only focusing on 
uncertainty and asset specificity may limit our understanding towards online consumer 
behaviour. Online shopping decision-making is a complicated process that is risky and 
uncertain. Importantly, financial risks (Hong and Cha 2013), product quality risks (Kim and 
Lennon 2010) and privacy risks (Ruiz-Mafé
 et al.
2009) have all been suggested as major 
inhibitors to online purchase. Although Liang and Huang’s (1998) study highlighted the 

---

67
importance of uncertainty, they failed to examine the specific uncertainty factors that 
consumers may be concerned about, in this regard, this study offers limited implications for 
e-commerce practitioners to alleviate consumers’ perceived uncertainty. Their study would 
provide more valuable insights into the production and causes of TCs if it could decompose 
uncertainty and asset specificity. Thirdly, buying frequency is a key factor in TCT. However, 
the study did not present the relationship between buying frequency and TCs.
Teo and colleagues’ study (2004) extended Liang and Huang’s work (1998) by examining 
more antecedents (six instead of two) of TCs and testing the model among Internet users 
across two countries, i.e., the USA and China. The study hypothesized that consumers’ TCs 
of online shopping were affected by six antecedents: product uncertainty, behavioural 
uncertainty, convenience, economic utility, dependability, and asset specificity. In turn, TCs 
had a negative relationship with consumers’ willingness to buy online. The results showed 
that behavioural uncertainty and asset specificity were positively related to TCs whilst 
convenience and economic utility were negatively related to TCs among US consumers and 
those in China. Product uncertainty and dependability were negatively related to TCs among 
US consumers but not consumers in China. TCs were found to have a positive influence on 
willingness to buy online among US consumers and those in China. The study extends the 
extant literature by explicating the relative importance of the antecedents to TCs and 
validating the relationship between TCs and online purchase intention in two countries.
However, Teo et al.’s (2004) findings revealed that the model accounts for only one third of 
the total variance of online buying behaviour, which implies that there might exist other 
factors that can affect TCs of online shopping. In addition, contrary to the TCT and related 
studies, product uncertainty was shown to be negatively related to TCs. Nevertheless, the 

---

68
authors did not pay much attention to explaining the unexpected findings. Another limitation 
of their study lies in the conceptualization of TCs. TCs were conceptualized as the time spent 
on searching for information about online stores and monitoring online stores for product 
delivery. According to their definition, the total TCs incurred by consumers in online 
shopping are represented by the search costs and the monitoring costs. Although search costs 
and monitoring costs play significant roles in determining the overall TCs, other costs (e.g., 
comparison cost, examination cost, negotiation cost, payment cost, delivery cost and post-
service cost) identified by Liang and Huang (1998) have largely been neglected. This 
conceptual limitation should be addressed in future research. The last limitation of their study 
is that no particular attention has been paid to the effects of different product categories on 
the hypothesized relationships. Given the importance of product categories in affecting 
consumer attitude and perception (Lian and Lin 2008, Lian
 et al.
2012), an area that warrants 
attention is the comparison of the relationships between TCs and its antecedents across 
various product categories.
In a follow up study, Teo and Yu (2005) further argued that information existing on the 
Internet could reduce information asymmetry and TCs. The study developed and empirically 
tested a consumer choice model based on TCT. Consistent with Teo et al (2004), the results 
confirmed that TCs were negatively related to willingness to buy online. To measure TCs, the 
authors highlighted three kinds of the costs involved in the online buying processes. They 
were searching cost (time and effort used to search for relevant products or services 
information and compare prices or other attributes among different online stores), monitoring 
costs (time and effort used to ensure that the terms of the contract have been met), and 
adapting costs (time and effort related to changes and customer service and support during 
the period of contract). The study extended prior research on online TCs by adding online 

---

69
trust and buying frequency as new antecedent variables that influenced TCs. By examining 
various types of uncertainties, the results demonstrated that different kinds of uncertainties 
had different impacts on TC. They found that three kinds of uncertainties, namely 
performance uncertainty of products, behavioural uncertainty of online stores and 
environmental uncertainty of online stores, positively affected TCs of online buying, whereas 
branding uncertainty of online stores was not found to be positively related to TCs. The study 
proposed that trust composed of two components (dependability of online stores and privacy 
policy). The findings showed that dependability of online stores was negatively related to 
TCs while the relationship between privacy policy and TCs was insignificant. Moreover, the 
study filled the gap in the literature by testing and validating the negative relationship 
between buying frequencies and TCs.
Despite their contributions, the findings of Teo and Yu’s (2005) study have several 
limitations. The major limitation lies in the sample selection. As the results indicated, more 
than 50 per cent of Internet users had not experienced online shopping. In this case, it would 
be difficult for those Internet users to respond to the questions on the survey regarding their 
perceptions of TCs, online stores’ performance, brand and behavioural uncertainty, and 
privacy policy toward a specific online store. These questions were designed to be answered 
by Internet users who have experienced at least one instance of online shopping. To obtain 
reliable responses, respondents should answer all the questions based on their actual online 
purchase experience by following the method outlined and used in the majority of online 
consumer behaviour studies (O'Cass and Fenech 2003, O'Cass and Carlson 2010). However, 
in Teo and Yu’s (2005) study, only general internet users were recruited to fill the survey, 
which could make the findings dubious and would further limit the generalizability of the 
findings as Internet users are not fully representative of the ultimate online shoppers.

---

70
Secondly, the search costs, monitoring costs and adapting costs may not fully tap into the 
concept of TCs occurring in the whole online transaction process. This conceptualization was 
not in line with previous research (Liang and Huang 1998, Teo
 et al.
2004). The 
measurement scales can be futher refined in future research. Thirdly, asset specificity was not 
considered in the study because the researchers claimed that physical and human asset were 
not the specific investments that consumers made for online shopping only. This view was 
contrary to previous research (Liang and Huang 1998, Teo
 et al.
2004) which confirmed the 
significant role of asset specificity in increasing TCs. In addition, as the researchers 
acknowledged, other potential antecedent variables of TCs they did not test in their study 
should be included in future research. Finally, the study examined TCT in an Asian context 
and demonstrated its applicability in a non-western context. However, as the sample was 
collected in Singapore only, the study’s results might be different if the model was retested in 
other Asian cultures (e.g., China, India) or a different cultural environment.
Another study by Kim and Li (2009b) applied the TCT to the online travel market and 
investigated the influences of TCs on customer satisfaction and loyalty. They conceptualized 
TCs as a three-dimension factor, consisting of searching costs, comparing costs and 
monitoring costs. The findings echoed previous research (Liang and Huang 1998, Teo and 
Yu 2005) in the effects of uncertainty and buying frequency on TCs. The study identified 
personal security as a new antecedent of TCs and found a positive relationship. The result 
also revealed the negative effects of TCs on customer satisfaction and loyalty with regard to 
online travel purchases. The study further suggested that TCs served as a key mediator 
between uncertainty, personal security, buying frequency and customer satisfaction, which 
was considered as the major theoretical contribution of the study.

---

71
Notwithstanding the compelling findings, Kim and Li’s (2009b) study contains a number of 
limitations. Firstly, although personal security as a new antecedent of TCs was added into the 
TCT, other key antecedents (e.g., asset specificity and convenience) identified and confirmed 
in previous studies were overlooked. Secondly, the proposed conceptual model was only 
tested in online travel market by collecting data in South Korea. The results may not be 
generalized in other types of contexts, such as online apparel market in other countries. 
Thirdly, the study did not take into account the effects of different products on consumers’ 
perceptions of online shopping. In their research design, the respondents were asked to 
identify the main travel agency for the product they purchased online. As a result, the variety 
of travel products with differing purchase cycles and product groups with different levels of 
competitive intensity could have impacts on the findings. 
In the online tourism shopping context, Kim et al. (2011) further developed a theoretical 
model to examine which factors influenced trust, satisfaction and loyalty. The results of the 
study indicated that TCs, navigation functionality and perceived security had a significantly 
positive effect on satisfaction. However, they found that TCs had no effect on trust. Only 
navigation functionality and perceived security were found to have a significantly positive 
effect on trust. They further pointed out that satisfaction had a significantly positive effect on 
trust and both factors positively affected loyalty, which in turn influenced consumer 
behavioural intentions with respect to tourism products and services online. The study argued 
that many online customers hesitate at the last moment to click the final order button for a 
purchase when they confirm the TCs associated with online purchasing. To proceed with a 
transaction, consumers search for information and monitor the process to ensure the best deal. 
The costs involved in all such transaction-related activities were called TCs in their study. 
The TCs were mainly to do with saving money online.

---

72
Although Kim et al.’s (2011) study offered greater theoretical and empirical insights into trust 
formation in online tourism shopping context, some potential limitations that had not been 
acknowledged by the authors should not be neglected. Firstly, the study adopted a relatively 
narrow perspective to conceptualize TCs by emphasizing the price reduction and money-
saving. For instance, some measurement items of TCs were “Online purchasing can save 
money as compared to offline purchasing”, “E-commerce can provide more discount than 
offline purchasing” and “Online shopping is the right choice when price and other expenses 
are considered”. Given that previous studies have considered TCs as a broad concept that 
covers all the costs incurred by consumers at each stage of the online transaction process (e.g., 
search cost, evaluation cost, monitoring cost, etc.) (Teo
 et al.
2004, Teo and Yu 2005, Kim 
and Li 2009b), this focus on saving money to depict TCs would constrain our understanding 
regarding the role of TCs in the online shopping for goods and services. Picking up on this 
point, future research may consider adopting a more comprehensive measurement approach 
which includes a wider array of TCs-relevant items in order to improve the statistical power 
and to reduce the potential for error in this type of research. Secondly, while TCs are the 
important factors that determine behavioural outcomes, this study did not explore the 
antecedents of TCs, which offer avenues for future research. Thirdly, the generalization of the 
results is limited by the context of online tourism shopping in South Korea: all of the 
observations were from South Korea and online retailing for tourism products in other 
countries and other product categories may not resemble those in South Korea. 
In a subsequent study, Kim et al. (2013b) partially addressed the aforementioned limitations. 
The study identified factors that affect trust in online tourism shopping based on the TCT, 
including transaction security, navigation functionality, and cost-effectiveness.
 
The results 

---

73
indicated that these factors positively affected trust, which in turn positively influenced 
repurchasing intentions. To explore the factors affecting trust, the authors largely relied on 
the TCT and its relevant studies. As they stated, TCs could be defined as transaction fees, 
time, and search effort, which can be viewed as a more comprehensive and advanced 
conceptualization as compared with that in their previous study (Kim
 et al.
2011). The TCs 
could be classified into price-type costs (e.g., parking fees, installation fees, taxes, transaction 
fees), time-type costs (e.g., travel time, waiting time, searching time, general shopping time, 
and delivery time), and psychological-type costs (e.g., ease of use, inconvenience, frustration, 
annoyance, anxiety, depression, dissatisfaction, disappointment) (Devaraj
 et al.
2002, Chircu 
and Mahajan 2006). After identifying the specific types of TCs, Kim et al. (2013b) then 
selected three key factors, namely cost-effectiveness, navigation functionality and transaction 
security, to represent the price-type costs, time-type costs and psychological-type costs, 
respectively. The three factors were thought to eliminate uncertainties in transactions. Since 
alleviating uncertainty can improve trust in online shopping according to the TCT, the 
authors thus posited that cost-effectiveness, navigation functionality and transaction security 
were positively related to trust. The results supported this notion by showing significant 
positive effects.
Although the factors influencing trust were identified based on a strong theoretical foundation, 
Kim et al.’s (2013b) study failed to examine the relationship between TCs and trust. 
Therefore, future research could test the effects of overall TCs on trust and other behavioural 
consequences, such as satisfaction and loyalty. Moreover, whether the results can be 
generalizable to all products in online environments are unknown. In future studies, 
presenting a broader range of online goods or services and comparing the different product 
categories may provide more insighs for academic and managerial fields. 

---

74
Prior research has showed that the TCT has not only tested in Business-to-Customer (B2C) 
commerce, but also extended to Customer-to-Customer (C2C) environment. A prime 
example of C2C is the online auction. As the construct of repurchase intention has been 
gaining prominence both in the literature and in business practice, Yen et al. (2013) 
investigated the factors influencing it. The study integrated the TCT and ECT (expectancy 
confirmation theory) to understand the determinants of online bidders’ repurchase intention 
in online auctions. The findings indicated that TCs were negatively associated with online 
bidders’ repurchase intention, while satisfaction had a significantly positive influence on 
repurchase intention. An online auction website’s asset specificity and product uncertainty 
were positively associated with the bidder’s perceived TCs. The interaction frequency 
between bidder and seller negatively affected the bidder’s perceived TCs. Bidders’ 
satisfaction was determined by confirmation and by the e-service quality of both online 
auction sites and sellers. The study contributed to a better understanding of repurchase 
intention by demonstrating that when online bidders make repurchasing decisions, they 
undertake a cost-benefit analysis to determine which option offers the greatest net benefits 
and the least costs. The research results provided a novel approach to understanding bidders’ 
benefits and costs dimensions in online auction marketplaces.
Despite that, the study needs to be considered in the light of specific limitations. The study 
did not provide a clear definition of TCs. It further measured TCs as a uni-dimensional 
construct without considering the individual cost involved at each stage of the online auction 
process. Given that the narrow perspective was adapted to measure TCs of online auction, 
future research may consider specifying the components of TCs by looking into the cost 
incurred at each transaction stage. While the study identified product uncertainty as a 

---

75
dimension of TCs, other uncertainties that bidders could encounter were largely overlooked 
by the authors. Bidders have expressed concerns about exposure of credit card numbers and 
personal information (Yeh
 et al.
2012a). This level of concerns highly suggests that it is 
necessary to take into account the security and privacy uncertainties in the online auction 
research. Apart from that, other uncertainties in relation to e-service quality, website 
performance, seller reputation and external environment may also need to be explored as they 
are likely to exert influences on TCs. Accordingly, research addressing the effects of a 
comprehensive set of uncertainty factors on TCs is urgently needed. The findings may not be
suitable to online B2C studies due to the significant differences in transaction processes and 
environments between C2C online auction and general B2C online shopping.
A more recent online shopping study by Wu et al. (2014) using the TCT proposed an 
integrative framework to examine the impacts of both TCs-related and value-related factors 
on repurchase intention from online shoppers' perspective. The study defined the construct of 
TCs as a three-component conceptualization that consists of information searching costs, 
moral hazard costs, and specific asset investments. Based upon evidence from survey of 887 
online shoppers, the results showed that each cost component and perceived value were 
positively related to repurchase intention. Importantly, information searching costs exerted 
the most significant influences on repurchase intentions. It also found that perceived costs 
positively affected perceived values. The study addressed the research gap by investigating 
the relative importance of the three distinct costs on perceived value and consumer 
repurchase intention. Each cost exhibited different impacts on the perceived value and 
repurchase intentions, which explained why employing a three-component conceptualization 
of costs to understand consumers' retention and value creation was important.

---

76
Beyond theoretical significance, Wu et al.’s (2014) study had several limitations. For 
example, it did not include the discussion of product types, product complexity and product 
attributes, all of which would likely affect the TCs. Given that different product categories 
have different influential determinants of behavioural intention, future research should 
investigate the effects of product categories on the relationships between TCs, their 
antecedents and behavioural consequences. The study did not include the possibility of other 
crucial factors, for example product quality concerns, environment uncertainties, and buying 
frequency, affecting the TCs. In addition, consumers' personal factors may affect the TCs. 
For example, consumers’ shopping motives (e.g., convenience- oriented) and internet skills 
(e.g., internet expertise) may affect their perceptions of TCs and, hence, their online shopping 
behaviour (Keaveney and Parthasarathy 2001, Bart
 et al.
2005). To explore these factors, 
future research can focus on the impacts of individual consumers’ characteristics on their 
perception of online TCs.