Measuring Gamification Elementss Effects on User Content Generation: An Empirical Study of Stackoverflowws Two Sided Platformms Big Data

Abstract

The cornerstone of the new marketing era consists of user generated content. This information is useful for reducing consumer uncertainty, generating new ideas for new products, and managing the customer relationship. To motivate users to generate content, practitioners use video game elements such as badges, leaderboard, and reputation points for user achievements, in an approach called Gamification. To allow Gamification platforms to target their users, I profile user segments by an ensemble method over LDA, mixed-normal and k-mean clustering, and then I develop a model of state-dependent choices of content generating users. This model captures long tail distribution of user heterogeneity by Dirichlet Process, and investigates the effects of fun and social elements of Gamification, reputation points, rank in the leaderboard, and badges (i.e. gold, silver, bronze) on the users’ probabilities to contribute content. I used a big data set of approximately 11,000,000 choices made by 36,000 users across 250 days on Stackoverflow to estimate the mixed binary logit model of users’ content contribution choices. I show that estimating the model on smaller random samples generate biased results. The estimation results demonstrate that users show heterogeneous significant positive and negative inertia, reciprocity, intrinsic motivation, and responses to badges, reputation points, and leaderboard ranks. I found interesting sensitivity patterns to Gamification elements for users with different nationality, which allows the Gamification platform to create targeted messages. The counterfactual analysis suggests that the Gamification platform can increase the number of contributions by making earning badges more difficult.