6. Conclusions
Recent studies have investigated Internet shopping carriers and provider issues and their effects on consumer services and operating strategies. Most of these empirical studies dealt these issues by collecting empirical data and testing hypotheses. This study further develops a mathematical programming model that can determine the optimal number and duration of service cycles for Internet shopping by exploring demand–supply interaction and time-dependent consumer demand. This study shows how demand–supply interaction can be carefully considered in advance of solving delivery service problems. This study also shows how variations in consumer socioeconomic, temporal and spatial distributions influence consumer demand for Internet store goods and, thereby, profit.
The results show discriminating service strategy yields better objective values than uniform service strategy, from which indicates that the Internet store operator and consumers may benefit from spacing service cycles according to time-dependent consumer demand. This finding also suggests that in practice an Internet store operator should employ frequent and short service cycles for periods with increased demand and long service cycles when demand is very low. The results further show that when transportation cost increases, the optimal frequent service cycles remains the same or increases. This finding indicates that the impact of reduced consumer demand for Internet store goods on profit is more significant than the increased logistics cost and, therefore, Internet store operators should employ more frequent service cycles to attract consumers and offset the influence of increasing costs.
The results show that variations in consumer socioeconomic, temporal and spatial characteristics play important roles in determining the optimal number and duration of service cycles and that not considering these variables yields reduced profit. This finding implies that the Internet store operator should carefully investigate the temporal and spatial distribution of consumer demand, income and needs and provide a delivery service strategy tailored to these criteria. For example, service cycles could be intensely spaced for a consumer area or region with numerous retail stores or during periods of large consumer demand. The finding also implies that consumers with high income are more sensitive to delivery delay than to the price of goods and, thus, serving these consumers with frequent service cycles for high price of goods could yield increased profit.
Conversely, this study shows that without considering demand–supply interaction, the Internet store operator typically minimizes average logistics cost per item by assuming inelastic demand and then applying least-frequent service cycles. However, this strategy yields lower profit than strategies that consider demand–supply interaction. In this study, demand–supply interaction is examined in a way that reduces logistics cost due to a large accumulation of goods based on long and less frequent service cycles; however, this strategy also results in an increased delay in receiving ordered goods, thus reducing consumer intention to shop via the Internet. Consequently, this finding in this study implies that the delivery service strategy may not only affect consumer demand for Internet store goods, but also operator logistics costs. In practice, Internet store operators may investigate the effects of service cycles on consumer demand for Internet store goods and its relationship with logistics costs.
This study can be extended in several ways. On the demand side, this study focused only on choice probabilities for two shopping modes rather than that among shopping stores within each mode. Future studies may use the joint or nested logit models to determine consumer choice probabilities for a specific Internet store. Second, the case study is based on an Internet store selling flowers in Taiwan with a study period of one operating day. Future studies may apply the model to different goods, such as computers and extend the study period beyond one day. Such studies would need to examine the impact of different characteristics of goods on consumer intention toward Internet shopping and calibrate a consumer demand function. Finally, as Chen (2001) suggested, profit may be improved by segmenting the market and then serving different market segments with different combinations of prices and service cycle frequencies. Future studies may expand this study’s model and address this issue by determining an optimal segmenting strategy and investigating the relative influences of the price of goods and delay in receiving ordered goods on consumer intention to shop via the Internet in the contexts of these different segments.