# An Empirical Investigation of Personalization Factors on TikTok

**Authors:** Maximilian Boeker, Aleksandra Urman  
**arXiv:** [2201.12271](https://arxiv.org/abs/2201.12271)  
**Date:** 2022-01-28  
**Summary page:** https://www.summarizepaper.com/en/paper/2201.12271v1/

## Key points

- Study by Maximilian Boeker and Aleksandra Urman on personalization factors on TikTok
- Importance of TikTok's recommendation system for its success
- Use of sock-puppet audit methodology and custom algorithm in the study
- Factors tested: language and location settings, user interactions (follows, likes), viewing duration
- Findings: all tested factors impact recommended content, follow-feature has strongest influence
- Discussion on filter bubbles formation and problematic content proliferation
- Implications for understanding personalization factors in social media user experiences

## Summary

In their study titled "An Empirical Investigation of Personalization Factors on TikTok," Maximilian Boeker and Aleksandra Urman delve into the impact of various personalization factors on the content recommended to users on the popular social media platform. With over 1 billion active monthly users, predominantly from Generation Z, TikTok's recommendation system plays a crucial role in driving its success. However, there has been a lack of empirical analysis concerning TikTok's algorithm. Boeker and Urman aim to bridge this research gap by employing a sock-puppet audit methodology along with a custom algorithm they developed. Through their testing and analysis, they explore how factors such as language and location settings, user interactions like follows and likes, as well as viewing duration of specific posts influence the content suggested to TikTok users. Their findings reveal that all tested factors have an impact on the recommended content, with the follow-feature exerting the strongest influence followed by likes and video view rates. The authors also discuss the implications of their research in relation to filter bubbles formation on TikTok and potential issues surrounding problematic content proliferation. This study not only sheds light on the inner workings of TikTok's recommendation system but also underscores the importance of understanding personalization factors in shaping user experiences on social media platforms. Boeker and Urman's work sets a foundation for further exploration into algorithmic influences on content distribution and user engagement within digital spaces like TikTok.