Repeat Purchase Behaviour of Online Customers for Grocery Products
Prasad, Vindhyalakshmi A
MetadataShow full item record
Anecdotal evidence, from industry reports, suggest that consumption patterns in India are increasingly moving away from functional needs to lifestyle considerations. However, there is not much empirical evidence regarding this. Studies have also established the profitability of retaining customers and the importance of understanding and linking the demographic profiles of customers to behavioural pattern. In this study, we make an attempt to understand the antecedents of behavioural loyalty of customers shopping online for groceries in India. We work towards developing a novel framework for quantifying the habitual component of behavioural loyalty. We obtained the anonymised transaction history of about 8000 customers from an online retailer. At the outset we examined the differences in buying behaviour in the demographic variables, in terms of the purchase frequency, purchase volume, mode of payment, voucher usage, inter-purchase time and the number of complaints raised. We found a significant difference between the order values before and after a complaint was raised by the customer. Paradoxically, the mean in the group before complaint was found to be higher than after the complaint. We examined this further with the backdrop of literature examining the existence of habitual component in behavioural loyalty. We developed a habit score to quantify habit. We plotted the estimated log order value on the Habit Score and the Log day difference between orders. From the plot, we saw that higher the habit score, higher the improvement of order value by a larger gap between purchases. In the subsequent Latent Growth Curve model we clearly saw that with the inclusion of the habit score, the model fit indices improved. The logit regression model, fitted based on the variables identified by the study was accurately able to predict the behaviourally loyal customers. We then drew out the managerial and policy implications based on the results of the analysis.