[1I4-GS-4a] Web intelligence

4:40 PM - 4:55 PM JST(7:40 AM - 7:55 AM UTC)

[1I4-GS-4a-05] Maximizing Fan Engagement Inferred from Social Media Data through Post-Timing Optimization

〇Rena Enomoto¹, Taiga Nishimura², Takahiro Hoshino^1,3 (1. Keio University, 2. Keio Graduate School, 3. RIKEN Center for Advanced Intelligence Project)

Keywords:

Social Media,Dynamic Factor Model,Ad-stock effect,User engagement,Posting patterns

While the proliferation of social media has led many companies to disseminate information through official accounts, the lack of a unified definition or metric for user 'enthusiasm' has made its quantitative measurement difficult.This study focuses on the anime industry, where trends are readily reflected in SNS metrics, and defines enthusiasm as a latent factor common to engagement indicators.Using a factor analysis model, we extract enthusiasm from the engagement metrics of official accounts across multiple intellectual properties and estimate the short-term effects of posting frequency and posting intervals on its fluctuations.Furthermore, recognizing that enthusiasm is shaped by the accumulation and decay of interest generated by past posts, we introduce cumulative enthusiasm by drawing on the Ad-stock framework, and test the hypothesis that this cumulative enthusiasm explains variations in search engine query volume.This study operationalizes enthusiasm as a measurable construct and identifies posting patterns that maximize it.