In Search of a TikTok Baseline - An empirical study of shared cultural experiences on a highly personalised digital platform

Abstract

This study investigates whether a shared cultural experience—what we term the "TikTok Baseline"—exists among Australian TikTok users. While existing research suggests social media platforms’ recommendation systems contribute to homogenizing users' cultural experiences, TikTok's highly personalized, responsive algorithmic system remains understudied. To map the "TikTok Baseline", we employed a methodological approach that minimized personal data exposure and interaction with content. Data collection occurred four times daily over a three-month period (May-July 2024) from multiple Australian locations, resulting in metadata from 5,100 unique videos from TikTok’s generic For You Page (FYP). We developed and validated an AI-driven video analysis tool using Google Gemini's 2.0 Flash multimodal model to enhance traditional metadata analysis. This paper reports on the preliminary phase of a comprehensive study of Australian experiences of algorithmic culture on TikTok. At the heart of the project is a comprehensive data donation-based study of how Australian content creators and users experience TikTok’s recommender system. This study makes three key contributions. First, we establish whether the concept of a reasonably stable TikTok Baseline manifests in real-world data, secondly, we examine the fundamental characteristics of such a baseline, and thirdly we suggest a rigorous computational methodology for examining TikTok baseline in the hope that our approach can be replicated in other territories and contexts. By addressing challenges in algorithmic observability and AI-driven content analysis, our findings offer critical implications for platform governance and regulatory efforts. This research advances our understanding of algorithmic culture, demonstrating how TikTok's recommender system both personalizes and standardizes user experiences.