The file is a computation,
not a writing project.
Marks, attendance, faculty records, admissions, placements — the self-assessment report is arithmetic over data the college already has. Nobody connected the two, because nobody owned both ends.
What it costs today
₹3–15 lakh a programme.
And the numbers trace back to nothing.
Excel plus a consultant, three to nine months of a department's life, and a 300-page Word file. The weakness of that market is not that consultants are bad — it is that the report is a computation nobody connected to the systems holding the inputs. We own the marks and the faculty file, so we compute it continuously.
Consultant fee for assembling what already exists
Paid again at the next cycle, because nothing that was assembled stayed connected to a system.
Spent collecting evidence that was generated all along
Mostly hunting for dates: the work happened, but nothing recorded when, or who was accountable.
Between knowing and being able to fix it
A continuously computed attainment tells a head of department in March what a September cycle will say. Two more faculty, or a professor confirmed, is a decision with a deadline.
How the award is actually decided
Grades and gates.
Never a total.
Change a grade, or put the head of department on contract, and watch the ceiling move. One failed gate caps the award on its own, whatever the marks say.
Built against the primary PDFs — SAR Format January 2025 (rev. 24-12-2025), User Manual 2026 Edition (21 May 2026), Evaluation Guidelines (7 July 2025).
Nine criteria · 1000 marks click a grade to change it
Y Fully compliant · ≥ 75%C Concern · 60–75%W Weakness · 40–60%D Deficiency · < 40%
The four gates that veto everything
Six years is blocked by 1 criterion at W, SFR 20.4:1 (six years needs 20:1), one Professor (six years needs two, or one plus an Associate).
Y 7 · W 1 · D 0
No W, no D, six criteria at Y
Plus PhD faculty ≥ 30% of required, SFR ≤ 20:1, two Professors (or one plus an Associate) regular with PhD, and a regular HoD with a PhD.
No D, three criteria at Y
W is permitted. PhD faculty ≥ 10%, SFR ≤ 25:1, and Tier-II relaxes to one Professor or Associate Professor. Tier-I does not relax it.
Not accredited
Reapply after one year. Borderline cases go to a Moderation Committee whose observations you answer within 10 days; appeal within 30.
The clock starts early
A three-year award files a compliance report at least six months before expiry; a six-year award files a fresh application at least five months before.
The nine criteria
Eight at programme level.
One at institution level.
Roughly 580 of the 1000 marks are computable rather than authored — they are arithmetic over marks, attendance, faculty records and admissions. The rest is workflow: an owner, a due date and a piece of dated evidence.
Traceability
Every number clicks down
to the person who produced it.
This is the one line the product is built to keep. A standalone accreditation tool has to beg for a CSV every cycle; it can show you a figure but never where the figure came from.
Each step names the product that authored it. Do not license that product and the metric does not disappear — it falls back to manual entry, and says so on the figure. An unlicensed neighbour is absent, never an error.
Sourced, not summarised
Four things the market
repeats that are wrong.
Every fact in this product is tagged with the document it came from and that document's date, because NBA revised the SAR, both User Manuals, both Evaluation Guidelines and the Pre-Qualifiers between January 2025 and July 2026 without changing the visible version label.
Ten criteria — seven programme, three institute
Nine criteria, 1000 marks. Eight programme-level, one institution-level.
Verified from Part B Criteria Summary, page 8, of the Tier-II SAR (January 2025). The ten-criteria split describes the superseded format, and a pack built on it would be wrong from the first cell.
750 of 1000 gets you six years
There is no aggregate threshold. The award is per-criterion grades plus hard gates.
From §4.2 of the User Manual for UG Engineering Tier-II, 2026 Edition. A single W blocks six years and a single D blocks everything, whatever the total says.
Twelve programme outcomes
Eleven, under GAPC v4.0.
Old PO6 (Engineer and Society) and PO7 (Environment and Sustainability) merged into PO6 The Engineer and the World. A tenant carrying a 12-PO articulation matrix must be migrated explicitly, never silently reinterpreted.
Tier-I is a different product
Tier-I is the same nine criteria, the same marks and an identical Criterion 5.
Only Criterion 6 differs, plus one gate: Tier-II relaxes the three-year professor requirement and Tier-I does not. It is a pack edit, not a second engine.
One engine, many packs
The framework changes.
The evidence does not.
Attainment, evidence, owners and dates are the same machinery whichever body is asking. A pack is the questions, the weights and the exhibits — not a second product.
NBA Tier-II · UG Engineering
Specification complete against the January 2025 SAR, the 2026 User Manual and the July 2025 Evaluation Guidelines.
NBA Tier-I · UG Engineering
Same nine criteria and marks; Criterion 6 and one professor gate differ.
NAAC · affiliated colleges
Seven criteria, 1000 points. The spine is built; the binary submission format waits on a primary source we can read.
CBSE SQAA · schools
Schools call this quality and compliance, not accreditation. One engine, a different pack.
IB · CIS · international schools
Evidence and review cycles, in the vocabulary those schools actually use.
NIRF
Five parameters confirmed; sub-weights not yet from a primary source, so it is not a pack yet.
Coaching centres see no surface from this product at all. Depth appears where it is needed and never leaks down.
Bring your last SAR.
We'll show you where the numbers live.
A working session against your own programmes and your own faculty file — not a slide deck. You will know within the hour which gate is your problem.