Netflix Engagement Patterns

Global Distribution and Viewer Engagement

Proud Mink
Reinesse Wong, Hannah Wang, Cindy Wang

2026-03-02

Introduction

Netflix releases biannual What We Watched reports covering ~99% of global viewing hours. This project analyzes four reporting periods from late 2023 through mid 2025.

The combined dataset includes:

  • Title
  • Content type (Movie / Show)
  • Release date
  • Reporting period
  • Availability strategy (available_globally)
  • Adjusted viewing hours

Dataset characteristics:

  • ~99% of global Netflix viewing hours
  • Movies and television shows
  • Engagement measured via adjusted_hours_viewed
  • Titles classified by global availability

Goal: Understand how distribution strategy and release patterns relate to engagement.

1.1: Engagement by Title Age

Question: Is Netflix engagement driven more by new releases or by older catalog titles across reporting periods?

Key Insights

  • Catalog dominates total hours

  • TV shows lean more catalog-heavy than movies

  • New releases contribute, but not the majority

1.2: Engagement Intensity

Why normalize? Older titles have more time to accumulate hours.

views_per_month = hours_viewed / months_since_release

Key Insights

  • Normalization matters
  • 0–3 months has the highest median hours/month
  • Intensity drops after the first year

2.1: Global and Regional Engagement

Question: How does engagement differ between globally distributed and regionally released titles across reporting periods and content type?

Key Insights

  • Global titles have higher median engagement
  • More high-end outliers for global movies and shows.
  • Strong overlap between global and regional distributions.

2.2: Global and Regional Engagement by Period

Question: How does engagement differ between globally distributed and regionally released titles across reporting periods and content type?

Key Insights

  • Global titles show higher median engagement than regional

  • The engagement gap widens for TV shows

  • Movie performance remains stable across releases.

Thank you!