From Uber to Alibaba to Airbnb, the spectacular growth of innovative platform-based companies abruptly changed their industries. Social Networks shape many aspects of how people and organizations interact, take decisions, and ultimately perform. Our research taps on social network data to draw insights and provide actionable recommendations for business. We have experience in working with established firms as well as startups, and our expertise lies in the areas of online platforms, media and telecommunications. We use large-scale social network data to tackle problems such as pricing, network growth, product adoption, and customer churn. We use a variety of methods and frameworks, ranging from machine learning, randomized experiments, simulation, and econometrics to analyze the problems, design, deploy, and evaluate our solutions.
Platforms are generally faced with the challenge of how best to attract and retain participants to achieve and maintain a critical mass and ultimately turn a profit. This cold-start problem is the main challenge in achieving and maintaining platform participation and growth. In a recent study, we work with an exclusive online dating platform to study if a personalised referral programme would help to stimulate network growth. Companies can amplify social contagion and accelerate product purchases by directly requesting members to invite their friends and acquaintances. In this study, we examined the underlying processes, dynamics, and implications of personalized referral policies for individual behaviors and platform growth. We collected data from a large-scale randomized field experiment over three years, in which different referral policies were used to request existing users to invite referrals to join the platform. Our findings show that referral programs can work as a double-edged sword. On the one hand, asking users to invite more new users provides benefits in terms of increased number of successful referrals and total payment. On the other hand, these benefits appear to come at the cost of reduced level of user engagement. Furthermore, in collaboration with the platform, we developed a method for designing personalized referral policies that account for consumer heterogeneity to maximize consumer life-time value. We demonstrated in real-life how such personalized policies lead to considerable improvements in customer life-time value and company profit, when compared to randomly assigned policies.