Two empirical studies on purchase behavior and referral engagement

Abstract

Both existing customer retention and new customer acquisition are of critical importance to the growth of the business. Loyal programs and referral programs are usually used to stimulate repeated purchase and referral engagement. However, the relationship between customers' repeated purchase and referral behavior is underexplored, and the empirical studies are even rare. By using a unique empirical dataset that includes the customer's demographic information, historical purchase, and referral behavior in 2019-2020 from a direct selling company, along with the COVID-19 data, the first study of this paper examines the impact of the COVID-19 pandemic on direct sales and whether there are heterogeneous effects of the pandemic on experienced and inexperienced direct- sellers. By utilizing the difference-in-differences regression research method, the results show an insignificant impact of COVID-19 on the overall consumption and referral activities of users in this direct-sales network. However, when decomposing consumption into different product categories, I find that the consumption for health products increases after COVID-19, where luxury products such as skincare products will experience an adverse demand shock; for obvious reasons that people have less need skincare products in times of lockdown. I also find heterogeneous effects of the pandemic on experienced and inexperienced direct- sellers. Experienced direct-sellers (i.e., the top two quartiles of users concerning historical consumption amount and/or network size) are purchasing (significantly) less after COVID-19. The second study of this paper investigated the factors that influence people’s repeated purchase and referral behavior and the relationship between consumers’ repeated purchase and referral behavior by using the data in 2016-2018. This paper also considers the customer’s heterogeneous characteristics and can distinguish the effect of customer characteristics on both customers’ purchase intention and their actual purchase behavior. Moreover, this paper not only focuses on pure consumers' behavior but also pays a lot of attention to the direct sellers (i.e., referrers) who play a significant role in recommending more new customers and motivating more repeat purchases. In the view of direct sellers (i.e., referrers), this paper also investigated the mutual effect of their repeated purchase and referral behaviors. For each research question mentioned above, the corresponding research hypotheses are proposed based on previous literature and the research context of this paper. The econometric regression models are built and estimated to test the proposed research hypotheses. Before going into the econometric regression models part (quantitative research method), this paper also uses a qualitative research method (including literature review and case study) to lay a foundation for understanding the research background. There are several essential conclusions found in this paper. (1) Female customers and older customers are more likely to make repeated purchases. But from the product category view, the two groups of users are found to own heterogenous product preferences. Female customers are inclined to make purchases in skincare and personal care products. In contrast, older customers are more likely to purchase health-related products, such as health care and health supplies. (2) In terms of the effect of gender and age on customer’s referral probability and actual referral performance, the results show that female customers and middle-aged customers (around 39-yearold) are more likely to make referrals. (3) For the effect of self-reported job on the repeated purchase probability and purchase ability, the unemployed users are 69% less likely to make a purchase, and the retired users tend to purchase 59 RMB less. In addition, this paper also finds that retired customers are 27.3% less likely to make referrals. (4) Age still has a positive and significant effect on the total reward amount, which means the older customers are more likely to get higher reward amounts by making successful referrals. In addition, the users who self-reported that they are freelance get more rewards than the baseline group of users. More interestingly and importantly, the absolute number of recommended users and the total purchase amounts of invited users are positively significant, revealing that the referral-reward schema in this research context is very different from the traditional referral-reward programs. The result shows an additional factor that influences customers' referral behavior – the referral program reward design. In this research context, both quantity (absolute number of customers recommended by the referrer) and quality (the consumption-ability of customers recommended by the referrer) matter for the realization of the reward. (5) By only focusing on the behaviors of referrers/salespeople in our user sample, this paper finds a positive effect of repeated purchase on referral and also a positive impact of referral on repeated purchase. Understanding the influential factors that play roles in customers’ purchase and referral decisions is very important for firms. And the identification of the heterogeneous need of their customers helps make precise customer segmentation and targeting. The positive correlations between customers’ repeated purchase and referral behaviors can provide insights for firms in leveraging on one of the two past customers' behaviors well and achieving mutual transformation. In summary, this paper offers rich managerial implications about motivating customers’ repeated purchase and referral engagement and precise customer segmentation and targeting.