Fake views removal and popularity on YouTube

Area of Science:

• Social Sciences
• Computer Science
• Media Studies

Background:

• YouTube engagement metrics, specifically view counts, are crucial for content visibility and platform algorithms.
• The platform actively removes artificial or illegitimate views, termed "fake views," to maintain metric integrity.

Purpose of the Study:

• To analyze the patterns and extent of YouTube's view count corrections.
• To investigate the timing and correlation between view corrections and video popularity.
• To explore the implications of fake views and corrections on content diffusion and online debate.

Main Methods:

• Analysis of 1.5 years of data from 1000 French YouTube channels.
• Tracking and quantifying view count corrections over time.
• Developing experimental techniques to circumvent API limitations for data extraction.

Main Results:

• Massive view count corrections affect the majority of channels and over 78% of videos in the corpus.
• Corrections occur in daily batches, typically around 5 p.m., not continuously.
• Later corrections are significantly correlated with higher video popularity, suggesting a delayed impact on metrics.

Conclusions:

• Fake views can artificially inflate perceived popularity, potentially skewing content recommendation and public discourse.
• The timing of corrections, particularly late in a video's life, warrants further investigation into its impact on content diffusion.
• Greater transparency from YouTube regarding view correction mechanisms is essential for understanding online attention cycles and combating misinformation.