Abstract

Social networks allow individuals and organizations to connect, grow their audience and share content, shaping public opinion and societal outcomes. Unfortunately, the way networks are structured and used often leads to polarization and competition for attention. In this talk, we present novel methods to address these two problems. In the first part of the talk, we focus on the effect of the competition for attention on acquisition (i.e., trying to connect with new users) and on engagement (i.e., trying to obtain endorsements or reposts from existing connections). We propose a dynamic-programming model that helps new users allocate their effort (e.g., time) by balancing these two actions conditional on attention. We illustrate our model using a dataset of all artists who joined a leading social media music platform, tracking their actions and network evolution over three years. Our findings indicate that our method outperforms traditional influencer-marketing strategies, including those that focus on low indegree, high indegree, and indegree asymmetry. In the second part of the talk, we focus on a novel method to disseminate content in polarized settings. The method addresses the trade-off between targeting community centers and bridges between communities. We illustrate our method on empirically-grounded polarized-community data, and on Facebook data. Our findings suggest that by accounting for the dynamics of polarization, our method achieves higher content-sharing performance than traditional approaches based on indegree (including community-based methods, such as bridges-only and centers-only), community-agnostic methods (such as eigenvector centrality,) and similar metrics.

Please contact dip.mkt@unibocconi.it if you wish to attend.