Fact-Checking Policy
Verification here is a set of steps rather than an attitude. This page describes the steps, including the ones that kill stories, which is most of what they are for. Last updated: August 2026.
What happens to a piece before it publishes
The writer checks their own copy first, against the sources and not against memory. Names, designations, figures, dates, scores, the spelling of a film’s title and which studio actually released it. Anything the writer did not establish personally is marked in the file with where it came from.
A section editor then reads for accuracy before anything else, asking whose word each sentence rests on and whether the sentence says so. A second editor reads it when the stakes justify that: anything alleging wrongdoing, anything naming a person in a way that could damage them, a first file from a new contributor, anything standing on a single source. Both sign-offs are recorded before publication. On a fast-breaking story the process compresses rather than disappears. We publish what two people have checked and mark what is unconfirmed.
Primary sources, and the number nobody audits
The judgment in full, not the operative paragraph a lawyer sent over. The official scoresheet rather than a rival site’s write-up of it. The transcript of what was said rather than the twelve seconds of it going around. A press note is a claim by an interested party, and so is the quote handed identically to every outlet in a junket room. If six sites carry the same sentence from the same actor, none of us reported anything.
Box office is our hardest case. India has no independent audited body publishing ticket sales, so every opening-day figure originates with a party that benefits from it being large. We call exhibitors and ask what they actually saw, and where the only number available comes from the film’s own side, we publish it as their claim with their name in the sentence, or we do not publish it. Readers dislike that answer. We have not found a better one that is also true.
Why our stories are full of links
Where a claim rests on a document, a dataset or a public statement, we link to that thing in the sentence using it, not to the homepage of the body that published it. Where the source is a PDF or a court record, we say what it is and quote the line we are relying on.
Where we cannot link, we say why. The document is not public. The source is a person. The record was shown to us on condition it stayed unpublished. Nobody should have to email us for a source.
Screenshots, clips, and whatever arrived on a forward
Most viral claims here arrive as a cropped screenshot or a clip with its caption stripped off. Before we describe what a clip shows, we look for the earliest upload we can find, run frames through reverse image search, read whatever metadata survived the re-encoding, and test the visible details against the claim: signage, the language on the boards, uniforms, whether the crowd is dressed for the right season.
An old clip relabelled with a new date is the commonest thing we catch, and footage from another country presented as Indian is close behind. If we cannot establish where and when something was shot, we either leave it alone or write about the claim itself and say in the caption that the footage is unverified. Stills from a film’s promotional material are never used as news photography.
Figures, studies and surveys
A statistic is read at source. We go to the dataset or the published paper rather than the release summarising it, and check that the finding in the summary is the finding in the work. Sample size, how the sample was drawn, whether responses were self-reported and who paid for the research all feed into whether we use the number.
Comparisons get checked for their base. A large percentage jump on a tiny base is a small thing in a costume. Growth measured against a disrupted year is not a trend. When a brand-commissioned survey lands on the desk, the commissioner is named in the sentence if we use it, and mostly we do not.
Posts, accounts, and things people said online
A post is a source like any other. Before we quote one we establish that the account belongs to the person it appears to, that the post exists rather than existing only as a screenshot of itself, and that it has not been edited since. We archive it.
A statement from an official account is evidence that the statement was made, which is not evidence that it is true. When a studio, a board or a player’s team posts something, we report it as their position and check the substance separately. Trending volume is not proof that anything happened.
The standard does not move with the speaker
A claim is checked the same way whether it comes from the governing party or the opposition, a superstar’s team or a first-time director’s, the board or a players’ body. We have no party affiliation and take no political or state money, which is set out on our ownership and funding page.
We do not balance a checked fact against an unchecked assertion so a piece looks even-handed. Where the evidence points one way we say so, and where two accounts genuinely conflict we set out both and name what we could not resolve. We work to the IFCN principles on non-partisanship, sourcing, transparency of method and open corrections.
When the check fails anyway
It does. A source turns out to have been wrong, a document is not what it appeared to be, a figure gets misread on the desk at eleven at night. The piece is corrected on the page with a dated note, and where the failure was ours rather than a source’s, the note says that too. Timings and sign-off are in our corrections policy.
The desk then works out which step let it through, because one wrong number is an error and the same wrong number twice is a fault in the process. The rules these checks enforce are in our editorial standards. If a claim in one of our stories does not hold up, write to support@cover365.in with the headline and the line, and we will show you what we checked.
