Simulating the Effective Policies for Improving Demand Response Rate in an Internet Home-made Food Distribution System: a System Dynamics Approach

One of the salient features of today's competitive world is the widespread use of Information Technology. E-business, which has grown significantly in recent years, has many benefits for organizations, customers and the community. The purpose of this study is to find the best policies to improve the demand response rate in online sales in e-business. In this study, the main variables of the study is identified, using system dynamics and the demand response rate is forecasted and described as the core variable. By creating a simulation model, different policies implemented, and the results and consequences of each are studied. Out of them, only those policies that guarantee the growth and success of the business, in reality, are selected.

Using System Dynamics, the main variables of an online home food distribution system has been identified and their interrelationships created in the form of cause and effect loops and state-flow model. The data needed to simulate the system, obtained from interviews with business executives as well as internet search, and the model simulated using the Vensim software for 72 months. After validating the model using appropriate tests, to validate the model, the proposed policies implemented, and their results compared with the current performance of the system. For this purpose, SPSS software and paired comparison test used to analyze the data obtained from the simulation and to find the policy that made a significant difference at the 5% error level under conditions of the current system.  

The simulation results of the proposed policies indicated that it is possible to improve the demand response rate by holding training courses, increasing production to the extent of capacity, as well as combining these two policies. Findings also indicated that increasing the food prices and implementing advertising programs did not affect improving demand response rates. The statistical analysis resulted in an insignificant difference for the second policy at 95% confidence level. The third policy emphasized increasing the number of food produced to the production capacity. The results of this policy indicated a significant difference between this policy and the current system at the 5% error level. The fourth policy suggested an increase in the final price of food. There was a significant difference at the 5% error level. However, due to the value close to the significance level of 0.05, it was not suggested to implement the policy. The last policy implemented in the system was a combination of the first and third policies. At the 95% confidence level, the fifth policy was significantly different from the current state of the system.  

Due to the large size of the model, variables such as revenue and profit entered into the system. The variables of food quality, customer's expected quality and customer's complaints removed from the model because they did not directly affect the behaviour of the main variables. Moreover, the variables of raw material prices and final food prices considered as average since it was not possible to enter different daily prices for more than 80 types of food.

The proposed model helps managers to evaluate the results of their suggested policies efficiently before their implementation and to make effective policy. Making the first policy, i.e. holding a training course, has affected all of the three variables of demand response rate, production capacity and profitability, significantly; hence managers are advised to put significant emphasis on such policy. Also, increasing the production to the maximum capacity was associated with a slight increase in the studied variables; hence, making this policy was not recommended. Furthermore, the simultaneous implementation of the training course and equalization of the number of food produced with the production capacity, each of which alone significantly changes the behaviour of the main variables, and can significantly increase the demand response rate, production capacity and profitability. Therefore, it is a suitable choice for managers and decision-makers to combine and implement these two effective policies simultaneously.  

By the use of System Dynamics approach, the causal relationships between different variables of the internet business system transformed into a dynamic model and the interactions of variables over time simulated. In similar studies on simulated e-business by a system dynamics approach, the simultaneous impacts of the production, demand, sales, and investment subsystems have not investigated.