Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Analyzing User Engagement with TikTok's Short Format Video Recommendations using Data Donations

Savvas Zannettou, Olivia-Nemes Nemeth, Oshrat Ayalon, Angelica Goetzen, Krishna P. Gummadi, Elissa M. Redmiles, Franziska Roesner

TL;DR

It is found that the average daily usage time increases over the users’ lifetime while the user attention remains stable at around 45% and users like more videos uploaded by people they follow than those recommended by people they do not follow.

Abstract

Short-format videos have exploded on platforms like TikTok, Instagram, and YouTube. Despite this, the research community lacks large-scale empirical studies into how people engage with short-format videos and the role of recommendation systems that offer endless streams of such content. In this work, we analyze user engagement on TikTok using data we collect via a data donation system that allows TikTok users to donate their data. We recruited 347 TikTok users and collected 9.2M TikTok video recommendations they received. By analyzing user engagement, we find that the average daily usage time increases over the users' lifetime while the user attention remains stable at around 45%. We also find that users like more videos uploaded by people they follow than those recommended by people they do not follow. Our study offers valuable insights into how users engage with short-format videos on TikTok and lessons learned from designing a data donation system.

Analyzing User Engagement with TikTok's Short Format Video Recommendations using Data Donations

TL;DR

It is found that the average daily usage time increases over the users’ lifetime while the user attention remains stable at around 45% and users like more videos uploaded by people they follow than those recommended by people they do not follow.

Abstract

Short-format videos have exploded on platforms like TikTok, Instagram, and YouTube. Despite this, the research community lacks large-scale empirical studies into how people engage with short-format videos and the role of recommendation systems that offer endless streams of such content. In this work, we analyze user engagement on TikTok using data we collect via a data donation system that allows TikTok users to donate their data. We recruited 347 TikTok users and collected 9.2M TikTok video recommendations they received. By analyzing user engagement, we find that the average daily usage time increases over the users' lifetime while the user attention remains stable at around 45%. We also find that users like more videos uploaded by people they follow than those recommended by people they do not follow. Our study offers valuable insights into how users engage with short-format videos on TikTok and lessons learned from designing a data donation system.
Paper Structure (25 sections, 13 figures, 1 table)

This paper contains 25 sections, 13 figures, 1 table.

Table of Contents

  1. Introduction
  2. Background & Related Work
  3. TikTok
  4. User Engagement
  5. Data Donation
  6. Remarks
  7. Data Donation System
  8. Requesting Data from TikTok
  9. Data Anonymization.
  10. Data Customization.
  11. Data Donation & Survey.
  12. Data Collection
  13. User Recruitment
  14. Video Metadata Collection
  15. Assessing the "Quality" of Donations
  16. ...and 10 more sections

Figures (13)

  • Figure 1: Percentage of donations that opt-in to donate to each field included in the TikTok data.
  • Figure 2: CDF of video durations/inferred viewing duration.
  • Figure 3: Changes in the volume of watched videos and time spent on TikTok over time. We observe that over time, there is an increase in both the volume of videos watched per day and the daily time spent on TikTok.
  • Figure 4: Number of participants that are considered for each day in our temporal analysis. The first day considers all participants while the 120th day considers 26% of them.
  • Figure 5: Number of following accounts and number of videos that are viewed from following accounts over time. The aggregate number of following accounts increases over the users' tenure, while the percentage of videos that are watched from following accounts remains stable over the users' tenure.
  • ...and 8 more figures