Essays on Network Effects, Service Pricing Plans and Online Message Virality
The first part of my dissertation investigates information technology and examines the matching between buyers and sellers in IT outsourcing markets, and how the social network between firms explains the observed matches. Specifically, I examine how interfirm connections and the information and resources accruing to firms from such connections-i.e., firm social capital - affect firms’ decision to sign a contract with a specific firm. The idea is that information spillovers from these connections constrain economic activities and that firms’ competitiveness changes by changing their network position. Using a unique panel data of 49,072 contracts signed between buyers and sellers in 1989-2013, I construct the network such that the nodes are the firms which are connected because they collaborate on shared projects. I use the network measures and clientspecific characteristics in a two-sided matching model to quantify the change in the created joint value in the match when the firms’ network position changes. Results suggest that a network of vendors and clients is more valuable as the size of their network grows (i.e., higher joined social capital is more valued) and that clients with higher social capital derive disproportionally more value from a vendor’s network. Moreover, although synergies created from the similarity in network positions are valuable in most cases, there are exceptions to this which are explained by transaction costs specific to such instances. The findings help firms design their competitive strategy by predicting the effect of repositioning in the network. In my second chapter, I use a randomized field experiment to investigate customers’ reaction to pricing and tariff design in the internet service industry. Specifically, I examine the effect of promotions and tariff structure (3-part tariff) on internet subscriber revenue and churn using a field experiment. Researchers have reported cases in which the impact of pricing decisions on profits is beyond what economic interpretations justify. However, behavioral effects of pricing and tariff structure on post-purchase outcomes are ignored in empirical studies even though they provide useful insight into consumer behavior and can have significant policy implications. I propose a parsimonious model that allows pricing, tariff structure, and new service introduction to impact two relevant behaviors in contractual settings: the level of transactions after plan choice and the decision to churn. My model also accounts for customer heterogeneity by including “level of pricesensitivity” and comparing the behavior of high and low price-sensitive segments. I use data from an Internet Service Provider (ISP) in a natural field experiment setting and find that promotions attract new customers, but they do so at the cost of increasing customers’ price sensitivity and lowering their inertia. Moreover, I find that customers who are not exposed to promotions spend more on plan and credit purchases even after I control for possible self-selection of more pricesensitive customers into the group that is exposed to promotions. I also find that high pricesensitive customers respond differently to constant exposure to promotions compared with low price-sensitive segments. I conclude that doing less relevant and targeted promotions wins pricesensitive customers, but it also encourages showrooming and comparison behavior. In other words, a marketing campaign that attracts new customers may also hurt customer lifetime value. I recommend that firms need to consider this trade-off when designing pricing policies. In the last chapter, I develop an empirical model to study how users’ social capital, mediated by image motives (i.e., driven by the perception of others) and intrinsic motives (i.e., driven by personal satisfaction rather than posting consequences), influences the propensity to post/retweet positive or negative contents online. My findings show that the identity of users can explain their motivations to post on online platforms and the receivers' engagement with the posted contents. Results show that there is an inverted U-shaped relationship between the number of followers and motives. The breakdown of motives reveals that both image and intrinsic motives are highest for users with a medium number of followers. Moreover, I find that as the number of followers increases, users are more likely to post due to intrinsic motives than image-related motives. Even though users with a higher number of followers believe negative content can lead to more virality, they do not post negative content mainly because they do not care as much about the reciprocity from followers as they care about the intrinsic satisfaction that they derive from posting contents. The results of my model show that intrinsic motives (in order of importance: validate thoughts, have fun, be a listener, amplify news, entertain/inform, save tweets) are more conducive to posting of positive content. On the other hand, image motives (in order of importance: look clever/expert, identify with a group, gain followers) are associated with the negative content but have no significant association with the positive content. The findings have implications for designing social media content strategy and fostering reader engagement.