How a Counselling Authority Actually Runs a Round

The mechanics of setting up, running, and publishing a counselling round, and exactly what stays entirely in the authority’s own hands throughout.

Setting Up a Round

1. Configure the seat matrix

   How many seats exist, split by programme, category, and any special quotas. This comes directly from the authority, nothing here is inferred or estimated.

2. Set eligibility rules

   Minimum rank or percentile cutoffs, domicile requirements, any exam-specific conditions. These are the authority’s existing rules, entered once and then applied consistently every round.

3. Set the round structure

   How many rounds there will be, what each round’s deadlines are, and how choices carry forward or drop between rounds.

Running a Round

The matching itself runs fast enough that scale isn’t the bottleneck, the same underlying engine can process well over a million applications in under an hour. What comes out the other end doesn’t go to students automatically. It goes to the authority first, for review and sign-off.

What Stays in the Authority’s Hands, Always

Eligibility rules

Set by the authority, applied exactly as configured, never adjusted automatically.

Reservation mandates

Category percentages and quotas are the authority’s own policy, not a default the system supplies.

The seat matrix

How many seats exist and where, entered and updated only by the authority.

Round timing and publication

When a round opens, closes, and when a result actually goes live, all require the authority’s own action.

The Verification Queue an Authority Actually Sees

Documents are scored and triaged before they ever reach a human, most get cleared automatically and only genuinely uncertain cases are surfaced. Exactly how that scoring works is covered on Document Workflows. What an authority sees here is the short list that already needs an actual person’s judgment, not a pile of routine checks.

If Something Needs to Change Mid-Round

Pausing a round, extending a deadline, or rolling back a result that was published in error are all real possibilities, and they’re handled as a distinct set of controls rather than something buried in the normal flow above.