This is the site's published evidence, with the source, the date, and — for every entry — the limits of what that source actually establishes. It is the first thing here with full provenance, and it is deliberately not a findings paper: no pilot has been run, and nothing here claims otherwise.
What this is not
This is not a research paper, and it does not report original findings. Every figure below was published by someone else and is reproduced here with its limits attached.
The project’s own evidence is currently zero published findings. The gap list further down says exactly what would close each one.
The Problem page states that numbers are not added without a source. That rule only means something if it applies to the site's own claims about itself, so it is applied here, in full, to everything published below.
Each entry states the bounded claim, who published it and when, links to the source itself, and records what that source does not establish. The last line of each is the one that does the work.
India's media and entertainment sector directly employs about 2.8 million people and supports roughly 10 million more in indirect roles.
EY, A studio called India: Content and media services for the world · May 2025 · Read the source (opens in a new tab)
What this establishes — and what it does not: Covers the sector as a whole. The report does not break this figure down by medium, so it does not establish that any single medium is the largest employer.
Regional cinema contributed more than 65% of all films produced in India in 2025. Hindi accounted for 226 of 1,972 theatrical releases — about 11.5%.
FICCI-EY Media & Entertainment Industry Report 2026 · March 2026 · Read the source (opens in a new tab)
What this establishes — and what it does not: Counts theatrical releases. Some secondary coverage of the same report gives the release count as 1,900 rather than 1,972; both figures are in circulation and neither has been reconciled here.
Regional languages were 56% of the content produced for streaming platforms in 2025, up from 52% in 2024, and their share of streaming consumption rose from 27% in 2020 to 56% in 2025.
FICCI-EY Media & Entertainment Industry Report 2026 · March 2026 · Read the source (opens in a new tab)
What this establishes — and what it does not: Measures share of catalogue and share of consumption. It does not measure revenue by language, and it does not distinguish independent production from studio output.
Television reached about 745 million people in a typical week in 2025, while digital became the largest segment of the sector by revenue.
FICCI-EY Media & Entertainment Industry Report 2026 · March 2026 · Read the source (opens in a new tab)
What this establishes — and what it does not: Reach and revenue are different measures. Television leads on weekly reach; digital leads on revenue. Neither is a measure of employment.
In a peer-reviewed survey of 52 mid-level crew in Malayalam cinema, 55.8% were paid per project and a further 21.2% per day — 77% in all on non-salaried terms — and 73.1% held no workplace-accident insurance.
George, Paul & Suresh Babu, South India Journal of Social Sciences 23(1):98–106 · March 2025 · Read the source (opens in a new tab)
What this establishes — and what it does not: A single study, n=52, in one language industry and one crew tier. It is strong evidence about the population it surveyed and a starting point for others — not a national figure, and not transferable without further measurement.
About 77% of mid-level Malayalam crew surveyed had been excluded from the credit list of a film they worked on, and 50% had a credit attributed to someone else.
George, Paul & Suresh Babu, South India Journal of Social Sciences 23(1):98–106 · March 2025 · Read the source (opens in a new tab)
What this establishes — and what it does not: Self-reported credit experience, n=52, one language industry. This is the specific problem the site proposes to measure. No independent replication exists yet, and none is claimed here.
Only 17.3% of that crew said federations have a 'high influence' on their remuneration, and 53.9% said 'minimal influence'.
George, Paul & Suresh Babu, South India Journal of Social Sciences 23(1):98–106 · March 2025 · Read the source (opens in a new tab)
What this establishes — and what it does not: A measure of perceived influence by the crew surveyed, not of formal guild power. Both readings are consistent with it, and the distinction matters.
The sector is consolidating rather than churning: FY2026 saw a series of equity deals including Dharma Productions, Excel Entertainment, Sikhya Entertainment and Bhansali's Sanzar Entertainment, with further consolidation forecast.
FICCI-EY Media & Entertainment Industry Report 2026 · March 2026 · Read the source (opens in a new tab)
What this establishes — and what it does not: Documents transaction activity, not churn rates across the whole sector. It does not establish what happens to crew-level credit and employment as ownership concentrates.
India has 6.8 cinema screens per million people, against 109 in the US, 95 in France, 66 in the UK and 64 in China. 16,350 of roughly 19,000 pin codes have no screen at all.
EY and Multiplex Association of India, The story of film exhibition in India 2025 · January 2026 · Read the source (opens in a new tab)
What this establishes — and what it does not: A measure of physical exhibition infrastructure and reach. It says nothing about distribution discovery, streaming reach, or the economics of any individual title.
India recorded about 868 million cinema admissions in 2025, meaning fewer than 150 million people — under 10% of the population — entered a cinema hall during the year.
EY and Multiplex Association of India, The story of film exhibition in India 2025 · January 2026 · Read the source (opens in a new tab)
What this establishes — and what it does not: Admissions are not unique viewers; repeat visits are counted. The population share is an order-of-magnitude statement, not a precise audience figure.
Casting fraud is being reported at production-company level: Aashirvad Cinemas filed a police complaint over fake casting calls made in its name, and the makers of Thaai Kizhavi issued a public fraud alert on the same pattern in May 2026.
Telangana Today; Times of India · May 2026 · Read the source (opens in a new tab)
What this establishes — and what it does not: Two named incidents. No prevalence data exists for casting fraud in India, and none is claimed here — the pattern is established, its scale is not.
Crew payment disputes are resolved by confrontation rather than by matching: FWICE halted the shoot of Ram Gopal Varma's Police Company on day one over roughly ₹1.32 crore owed to technicians for work on Officer, some of it unpaid for around nine years.
Times of India; NDTV; Moneycontrol · August 2026 · Read the source (opens in a new tab)
What this establishes — and what it does not: Reported figures differ between outlets on the length of the halt (two hours or five) and the amount paid (₹1.32 crore or ₹1.11 crore). The pattern is well attested; the exact figures are reported, not settled.
Professional listing is already cheap in India: LetsFAME lists over 100,000 film professionals and 500+ postings a week, with premium access from ₹999 a year.
LetsFAME (published service terms and public listings) · Accessed October 2026 · Read the source (opens in a new tab)
What this establishes — and what it does not: The site's own published figures for its own service, not an audited count. Taken as evidence that the layer is occupied and priced, not as a market-share measurement.
Indian entertainment professionals now have a free professional-tools incumbent: IMDbPro gave Indian entertainment professionals 12 months of Premium membership free.
Economic Times · 31 August 2026 · Read the source (opens in a new tab)
What this establishes — and what it does not: An introductory offer for professionals with existing IMDb credits. It constrains what a paid professional-profile product can charge in India; it does not establish what it will cost after the period ends.
For each load-bearing claim, a disconfirming search was run. Where nothing was found, that is recorded as 'not found' — never as support. This section is where a reader can check whether the evidence has been read fairly rather than selectively.
A gap list is only worth publishing if it says what would close each gap. Below, every missing piece of evidence is paired with the specific thing that would produce it, and the level of the site's own evidence ladder it would reach.
The 2025 study is the only credit-accuracy measurement located, and it covers one language industry and one crew tier.
What would close it
The same instrument — an agreed baseline definition of a credit, collected the same way — applied in at least two other language industries.
Level it would reach
OBSERVED across industries, rather than in one
Nothing has been tested. The mechanism is proposed; the effect is not measured.
What would close it
A defined intervention, a baseline, and a comparison — with the unfavourable result published alongside the favourable one.
Level it would reach
DEMONSTRATED INTERVENTION
Sector-wide figures are available and are published on this site, but they are not broken down by crew tier, language industry, or employment form.
What would close it
A sampling design that can carry a claim about mid-level crew specifically, rather than about the sector as a whole.
Level it would reach
VALIDATED PATTERN for a defined population
Disputes are documented case by case. There is no measured distribution of how often payment is late, or for whom.
What would close it
A self-reported, non-identifying measure of payment timing across a defined cohort of crew.
Level it would reach
OBSERVED, then VALIDATED with a stated sample
No conversation with a prospective buyer has taken place, and no price has been tested.
What would close it
Ten conversations with producers and financiers, published in full whether they end in a sale or a rejection.
Level it would reach
Not applicable — this is a commercial question, not an evidence claim
Where to go next
Every hypothesis on the Opportunities page names the evidence that would test it. Read against this page, it is clear which of them the published evidence already touches, and which are still entirely open.
Nothing on this site claims to measure what the ecosystem already knows. That is the gap this project exists to close.