Every number, read from a document.

The sources

Every number on Collvera traces to one of four official document classes:

• The independently audited placement report — the gold standard: verified by a third-party auditor (CRISIL, B2K Analytics and peers), with the auditor named and, where issued, the UDIN (Unique Document Identification Number) you can verify with ICAI.
• The college's official placement report — published by the institute itself; read verbatim and fully sourced.
• The NIRF Annual Disclosure — the Ministry of Education's mandatory data submission, used for institutional data.
• The official fee schedule — tuition, hostel and other charges as published by the college.

We never scrape third-party aggregators, never use media claims or brochures, and never accept numbers a college merely tells us without a document behind them.

The pipeline: report → record

• Transcribe, don't summarize. Every disclosed figure is transcribed: headline outcomes, salary by component (min / median / mean / max per head), sector and function splits, named recruiters with offer counts, PPOs, batch profile. 532 college-years on file so far.
• Provenance travels with the number. Each college-year stores its source document name and date, audit status, auditor firm, and UDIN where present.
• Normalize on top, never in place. Reported figures stay verbatim on their reported basis; comparability is computed as a separate field (next section).
• Publish with the receipt showing. Pages display the figure, its basis, its Trust Grade and its source. Missing data renders as “—”, never as a guess.

Trust Grades: not all reports are equal

Every college-year carries a grade, shown wherever its numbers appear:

• Grade A — independently audited. Third-party verified, auditor named, UDIN recorded where issued. 99 report-years currently qualify.
• Grade B — official institute report. Published by the college; sourced and verbatim, but not third-party verified.
• Grade C — partial / interim disclosure. In-progress or fragmentary numbers. Never used as a headline figure or in rankings; always flagged interim.

Collvera Comparable CTC — the standard

The dirty secret of placement reporting: colleges quote on different bases. One reports audited total guaranteed cash (fixed pay you can count on); another reports gross CTC (everything, including one-time and variable components); a third headlines max earning potential. Comparing those raw headlines is comparing apples to invoices — and it punishes exactly the colleges that report most honestly.

Collvera Comparable CTC resolves every college-year to one like-for-like central figure. Every ranking and comparison on this site uses it — never the raw headline. Every metric we publish is precisely defined in the data dictionary.

The rules we never break

• Same-year ranking. Every leaderboard compares one placement year against the same year across colleges — never a newer unaudited report against an older audited one.
• Interim years never headline. Numbers a report flags as interim, partial, or as-on-a-date cannot be a college's headline figure, and are labeled interim wherever shown.
• Missing means NULL. Undisclosed figures display as “—”. Never zero, never estimated, never modeled. The sole approximated field on Collvera is admission cutoffs — always framed as approximate — and the only projections are clearly-labeled next-year forecasts.
• Down-years publish like up-years. Same rules for the biggest brand and the smallest school. We don't smooth, bury, or contextualize away a decline.
• Multi-year on the record. Where colleges have published across years, we hold the full series — up to 16 years for the oldest institutions — so trends are verifiable, not vibes.
• Errors get fixed publicly, fast. A verified error is corrected within 72 hours and logged permanently in the public corrections log. Scripted audits periodically recompute published figures against the raw source data. The full independence charter lives at /trust.

What we don't do

• No paid listings, rankings or affiliate commissions — no college pays us for placement on the site, position in a ranking, or removal of a number. Ever.
• No selling user data — your inputs are not shared with coaching institutes, lead generators, or third parties.
• No ads — no AdSense, no display network, no programmatic.
• No data ransom — every dossier, comparison, cutoff and FAQ is free to read. Email is asked only for deliverables like the research report PDF or personalized outputs.
• No college-commissioned services — anything Collvera offers, including any paid guidance for applicants, is bought by students and built on this dataset. Colleges cannot commission, sponsor, or influence it.

Who built this

Collvera is built and run by AI. The data pipeline, the dossiers, the eligibility scoring, the comparison engine and the editorial layer are generated by Claude, then cross-checked against the verified source documents — audited placement reports, NIRF disclosures, official college publications — at every step. The output of every model call is verifiable against the sources we cite.

For data questions, methodology corrections, or press inquiries: see the press page.