Netflix Content Longevity & Global Reach

Analyzing Engagement Patterns from Netflix Reports

Steven Xie, Ziqing Yang, Simon Yan, Xinyi Wang

Motivation & Research Questions

  • Netflix invests billions in content — but which titles retain value beyond launch week?
  • Key question: what content transitions into “evergreen” assets for long-term catalog value?
  • Q1: How do post-release “hype curves” differ across genres and languages?
  • Q2: Is global availability associated with higher engagement?

Dataset Overview

  • Netflix Engagement Reports spanning late 2023 – mid 2025
  • 36,000+ movie records and 27,000+ show records across 4 reporting periods
  • Key variables: hours_viewed, release_date, available_globally, runtime
  • Data source: TidyTuesday Netflix dataset

Q1 Approach

  • Standardized temporal axis: periods_since_release (6-month intervals) to align movies across release years
  • OMDB API for official Genre & Language metadata (top 500 movies) – overcomes keyword-based guessing
  • LOESS-smoothed line plots with log-scaled y-axis to handle variance between blockbusters and standard titles

Hype Curves by Genre

  • All genres peak at P1, sharp decline by P3
  • Action & Comedy: rapid drop, flat baseline – upfront novelty
  • Drama: moderate decay, stable mid-range floor
  • Animation: resurgence P6–P8 – a measurable “second wind” driven by family re-watchability

Language Factor

  • English: higher peak, then remarkably stable baseline from P4 – structural backbone of retention
  • Non-English: steeper decay, extreme later volatility – driven by rare outliers
  • Netflix can launch international hits, but sustaining them is harder

Q2 Approach

  • Global distribution is a strategic decision – does it pay off in engagement?
  • Time-adjusted metric: normalized_hours = hours_viewed / months_since_release to control for exposure bias
  • Movies and shows analyzed separately – films are single-session; series have longer engagement cycles
  • Boxplots for group comparison + scatter plots for continuous patterns

Engagement by Availability

  • Globally available titles show higher median normalized engagement
  • Effect consistent across both movies and shows
  • Some regional titles perform well, but global titles dominate the upper distribution

Engagement vs Runtime

  • Global titles cluster in the upper range regardless of runtime
  • Pattern holds for both short and long content
  • Caveat: correlation, not causation – global titles likely have larger budgets and broader appeal

Key Takeaways

  • Animation is uniquely resilient – invest in animated content for long-term catalog value
  • English content provides a stable, predictable catalog backbone for retention
  • Global availability correlates with stronger engagement, but confounders exist
  • Content strategy should balance novelty hits with long-term evergreen assets

Limitations & Future Work

  • Only top 500 movies enriched with OMDB metadata – may not generalize to long-tail content
  • available_globally is binary – does not capture partial regional rollouts
  • No causal inference – observational data cannot isolate the effect of release strategy
  • Future: incorporate budget/marketing data, apply survival analysis for content lifespan modeling

Thank You!

Questions?