Private beta30 seats for the $9 test film.Check a niche

Disasters & engineering failures

YouTube niche report, collected 3 November 2026: who outperforms their own median, which titles they use, and what viewers ask.

Saturation (our judgement)

medium

Look-alikes / distinct channels

15 / 27

Strongest repeatable outlier

11.4×

Collected

2026-11-03

Summary

We searched YouTube for “engineering disaster documentary what went wrong” (50 videos published after 2025-11-03). They came from 27 distinct channels; 15 of them look like faceless long-form channels in this niche (our count). We analysed 6 channels in depth.

Strongest repeatable outlier: Andy R Animations (74,000 subscribers), whose best video reached 1,859,143 views, 11.4× its long-form median of 163,121 (“Apollo 13: What Went Wrong”).

Our notes, judgement

Two sub-formats perform: DisastersUncovered's 15-25 min human-narrated stories ('No AI' in every title, median 87k) and Andy R's multi-hour 3D 'What Went Wrong' films (Apollo 13 at 1.86M views, outlier 11.4x). The four other independents analyzed are 2025-2026 entrants with medians from 79 to 2,275 views; two of them (The Yield Point, COLLAPSE POINT) stopped uploading within 1-2 months. The 'No AI' label is a notable positioning signal against AI-made competitors (judgement). Six channels were analyzed instead of four because the first picks were too small to be informative. Faceless judgements are from names, titles and descriptions only; thumbnails were not viewed.

Outlier channels

ChannelSubscribersLong-form medianBest video ÷ medianBest videoUploads / week
DisastersUncoveredlikely faceless (judgement)38,60087,4193.69×Too Low & Too Slow. The Shoreham Airshow Disaster. No AI. A Disaster Documentary.322,770 views, 25.2 min, 2026-08-142.25
Andy R Animationslikely faceless (judgement)74,000163,12111.4×Apollo 13: What Went Wrong1,859,143 views, 120.5 min, 2025-04-040.11
Forgotten Disasterslikely faceless (judgement)7,2402,2745.02×How Did 739 People Die in the 1995 Chicago Heat Wave?11,418 views, 25 min, 2026-06-072.74
The Yield Pointlikely faceless (judgement)162344.45×The Big Blue Crane Collapse: He Was Fired For Slowing Down The Project1,042 views, 15.4 min, 2026-04-171.31
Collapse Logiclikely faceless (judgement)4822.6×The Hyatt Regency Collapse: How Did Engineers Miss This?213 views, 9.2 min, 2026-10-072.42
COLLAPSE POINTlikely faceless (judgement)3791.86×The Engineer Who Tried to Stop Challenger. And the Eight Words That Overruled Him.147 views, 12.2 min, 2026-06-300.47

Long-form median: median views of the uploads over 3 minutes among the last 30. Best video ÷ median is the outlier factor; on a median under 1,000 it is noise. Uploads per week count the last 10 uploads. Channel links open on YouTube.

Title patterns

Our judgement
  • The [Hidden Flaw / Simple Mistake] That [Destroyed X]

    • “The Hidden Flaw That Made This 300-Meter Bridge Collapse | Massive Engineering Mistakes”
    • “This Simple Mistake Dropped an Entire Highway | Massive Engineering Mistakes Season 2 Episode 10”
    • “China’s Top Secret: How the Deadliest Dam Collapse was Caused by a Minor Flaw”
    • “The Real Reason This Bridge Tore Itself Apart”
  • They saw the warning - nobody acted

    • “They Saw the Cracks Coming — Nobody Stopped Them | Engineering Disasters Documentary”
    • “They Saw the Cracks. Nobody Closed the Road.”
    • “The Turbine That Kept Running Despite the Warnings”
    • “He Knew the Brakes Were Failing But Kept on Going. Vernon, NJ Bus Crash. A No AI Documentary”
  • <Event>: What Went Wrong

    • “Apollo 13: What Went Wrong”
    • “The Challenger Disaster: What Went Wrong”
    • “The Columbia Disaster: What Went Wrong”
    • “Top 5 Engineering Disasters — What Went Wrong?”
  • Death toll as the hook (N Dead / How Did N People Die in X?)

    • “233 Dead in Just 40 Minutes. The MV Christena Disaster. No AI. A Disaster Documentary”
    • “How Did 739 People Die in the 1995 Chicago Heat Wave?”
    • “114 People Died at the Kansas City Hyatt — Nobody Was Watching”
    • “13 Engineering Disasters That Killed Hundreds (And Were 100% Preventable)”
  • Superlative listicle (The N Worst / Deadliest ...)

    • “The 5 Worst Engineering Mistakes Ever Made (Explained in 9 Minutes)”
    • “10 Engineering Disasters That Should Never Have Happened”
    • “The Worst Engineering Disasters in History”
    • “America's Biggest Megaproject Disasters”

Read these as patterns, not templates. Across all 15 niches, channels that copied a title word for word sat at long-form medians of 4 to 31 views.

Topic ideas, as the viewer asks them

Our judgement
  1. 01Why did the Teton Dam fail the first time it was filled?
  2. 02How did one design change double the load on the Hyatt Regency walkways?
  3. 03Why did NASA launch Challenger after engineers warned about the O-rings?
  4. 04How could fallen leaves bring down a steel bridge?
  5. 05Why do construction cranes collapse over busy city streets?

Formats

Our judgement
  • case-study documentary (10-25 min)
  • long 3D-animated investigation (2-3 h)
  • listicle / ranking
  • sleep compilation
  • shorts as teasers

Saturation

Our judgement

medium, 15 look-alike channels among 27 in the results

15 of 27 distinct channels look like faceless long-form disaster/engineering channels, at the top of the medium band. But the result set is dominated by TV catalogue: Banijay Science alone has 21 of 50 results, plus National Geographic UK, PBS and Free Documentary. Many faceless look-alikes are 2026 micro entrants (The Yield Point, Collapse Logic, COLLAPSE POINT with under 20 subs), so supply is growing fast while the proven independents (DisastersUncovered, Andy R) are few. Judgement.

Method and date

One search on the YouTube Data API v3 (type video, order relevance, 50 results, region DE), then the last 30 uploads of 6 channels for medians and outliers and their last 10 uploads for upload rate. Every number is copied from an API response; saturation, title patterns, topic ideas, formats and faceless calls are our judgement. It is a sample, not a census. No RPM or revenue figures: there is no public source for them. Full method.

Source: YouTube Data API v3, collected 3 November 2026.