Original research
Fifteen separate reports for questions creators already argue about: hooks, captions, opening clarity, comments, niche benchmarks, safe zones, zero views, recuts, and what people change after a flop.
Why publish now
Each report names what was measured or still needs to be measured, what remains unknown, which media can be used, how disagreement is preserved, and which claims stop publication.
That means a future null result still counts. It also prevents a convenient metric from becoming a platform claim after the charts are drawn.
How quickly the subject, language, and visible information become usable to a cold viewer.
Caption-event timing audit
When does meaningful on-screen text appear relative to speech, meaning, and the visual it explains?
Blinded 100-opening teardown
How quickly can a first-time reviewer reconstruct the subject, topic, promise, or tension from an opening?
Silent-opening cue audit
How often could one coder identify a Short's premise from three opening stills without hearing the audio?
Opening-language transcript pilot
What kind of information appeared in the first five recovered words of the sampled YouTube Shorts?
Short-form text-placement safe-zone proxy audit
How often did primary opening text enter a conservative edge-band proxy, and what is required before calling that a platform safe-zone violation?
How an edit progresses, where it stalls, and how a documented recut could test a structural change.
Longitudinal annotation benchmark
Where does the first reviewable structural problem appear, and can independent reviewers agree on it?
Promise-proof-payoff annotation study
How far does a video travel between its opening promise, first proof, and primary payoff, and what happens in between?
Consented short-form recut case series and blinded viewer experiment
What changes after a preregistered structural recut, and how can viewer perception be separated from live platform distribution?
What creators infer after a weak result, what the evidence supports, and which next action remains testable.
Verified paired-outcome survey
What did creators change before reposting, and how did the second upload compare after baseline and natural variation were considered?
Creator decision and evidence-alignment study
After an underperforming video, how closely does a creator's first action align with the evidence available at the time?
YouTube Shorts advice evidence audit
Which familiar YouTube Shorts advice claims have first-party grounding, and which turn context into unsupported universal rules?
Public zero-view resolution case pilot
What evidence is required before a reported recovery from zero views can be assigned a cause?
Creator, expert, and analyzer diagnosis agreement protocol
Where do creators, independent experts, and a structural analyzer disagree about the weakest moment in a short-form video?
How comments and niche labels become reproducible evidence instead of convenient categories.
Viewer-confusion comment classifier feasibility test
Can a multilingual question-word lexicon identify comments that show viewers did not understand a video's premise?
Stratified short-form structural benchmark protocol
How do observable short-form video structures vary by niche without turning descriptive medians into universal targets?