Loading...
CinePulse analyses pre-release promotional content — trailers, songs, teasers and character introductions — to predict how satisfied different audience segments will be before a single ticket is sold.
Existing systems predict box office revenue or general sentiment. CinePulse addresses a different and more specific question — how satisfied will each audience segment be after watching a film, based only on what they have seen before release?
The foundation is a satisfaction gap formula that measures the distance between what promotional content promises to each audience persona and what that persona actually expects. A small gap means high satisfaction. A large gap means disappointment.
We collect YouTube comments from all pre-release promotional content — trailers, teasers, video songs and character introductions. Each content item is assigned a temporal weight based on proximity to release date. Comments are analysed to extract four dimension scores across mass-class position, emotional hook, genre clarity and star power.
Each comment is classified into one of eight cinema persona types. Classification uses engagement weighting — comments with higher likes influence persona distribution more than zero-liked comments. Three prediction modes are computed: raw YouTube derived weights, researcher corrected weights reflecting actual cinema audience distribution, and a blended approach.
For each persona the system computes the gap between the content signal score and that persona's expected score. Satisfaction is derived from this gap using a calibrated formula. Confidence intervals are generated using Monte Carlo simulation with 500 iterations of the demographic weight matrix.
YouTube comment data over-represents highly engaged fans relative to the general cinema audience. Our system accounts for this through a researcher correction layer where domain expertise in South Asian cinema audience psychology is used to calibrate persona population weights. This correction reduced prediction error by 63% in our Border 2 validation study.
The specific methodology including the satisfaction gap formula, persona classification framework, film type engine switching, temporal weighting model and fan community inflation detection is protected under a provisional patent filed March 28 2026. A full research paper detailing the methodology is currently in preparation for submission to an academic journal.
We analyse pre-release promotional content — trailers, songs, teasers and character introductions — using a proprietary combination of audience persona classification, satisfaction gap analysis and demographic weighting to predict how satisfied different audience segments will be after watching a film.
Predictions are published before release with timestamps. After films release we collect verified audience scores from BookMyShow, IMDb and Letterboxd and publish accuracy analyses showing what we got right, what we missed and what we learned.
South Asian cinema represents the world's largest film market by volume with over 1,800 films released annually. Despite this scale no existing tool provides pre-release audience satisfaction prediction specifically calibrated for South Asian cinema audience psychology and the unique mass-class divide that characterises this market.
This is an active research project. The underlying methodology is patent pending as of March 2026. A research paper is currently in preparation. All predictions published on this site are timestamped research outputs — not commercial recommendations.
Research by S.G. — independent AI researcher specialising in entertainment analytics and audience prediction systems.
This research analyses publicly available audience comment data from YouTube for academic purposes. No proprietary studio data was used. All film titles referenced are used solely as research case studies. CinePulse is not affiliated with any film studio, production house or streaming platform.