Financial Aid Gap Analysis

A gap flag is a quality finding: this enrolled student's cited unmet need exceeds the retention-risk threshold aid and enrollment agreed to use. The finding must cite the aid package and the need calculation. It is not an emergency grant, not a revised package, and not a student-facing promise.

Unmet need means cited cost of attendance minus the cited Student Aid Index (or a legacy Expected Family Contribution still on that ISIR, labeled as EFC), minus cited packaged aid. Every term must point at a record. If the ISIR is missing, or packaging has not been run in the system of record, unmet need is unknown. Unknown does not become a number. The flag stays empty.

Enrollment may treat a large gap as a retention event, the way a quiet student shows up in an early engagement alert. The response is an aid queue, not a disbursement. Aid still owns eligibility, documentation, fund rules, and professional judgment.

The payload must let a counselor reopen the same package and need worksheet: student, term, award year, ISIR transaction number, packaging identifier, COA, SAI or EFC, packaged-aid total, cited gap, threshold, and as-of times for both sources.

Do not compute until both sides of the comparison are cited.

Need side. Use the COA that matches the student's actual enrollment intensity and housing plan for that award year, not a catalog maximum. Use the SAI (or labeled EFC) on the ISIR transaction currently imported to the student information system. Record the transaction number and import datetime. A blank SAI is not zero and is not last year's index.

Package side. Use the current packaging record: Title IV aid, state grants, institutional grants and scholarships the office treats as packaged, loans, and work-study if they count in your policy. Cite the packaging document or batch ID and its datetime. Do not treat an estimated-award letter, a counselor note, or a CRM scorecard as the package.

Population. The student must be enrolled for the term you are measuring. A deposited student who has not enrolled belongs under enrollment melt prediction, not this flag. Align census or freeze rules so a full-time COA is not compared to a part-time schedule.

Where records live. ISIR import and packaging usually live in Ellucian or Workday. Inquiry, deposit, and counselor workflow often live in Slate (Technolutions). Read need and packaging from the SIS or aid module. Do not reconstruct a package from CRM fields. CRM fields go stale when packaging changes.

If any required cite is missing, stop. Empty stays empty.

With cites in hand, compute in this order.

Cited need equals cited COA minus cited SAI or EFC.

Cited gap equals cited need minus cited packaged aid. If packaged aid meets or exceeds need, the gap is zero or negative. Do not flag.

Compare cited gap to the written retention-risk threshold. That number is policy and budget, not something the model should learn from last year's emergency spend. The model may order students who already clear the cutoff. It may not move the cutoff.

Then write the flag with every cite above.

A returning sophomore is enrolled full time. The SIS holds ISIR transaction 03, imported after a correction, with an SAI present. COA is the full-time budget for that award year. Packaging shows accepted Pell, a state grant, an institutional grant, and a subsidized loan. The cited gap sits above the hardship-review cutoff aid published internally. The model writes a flag that points at that transaction and that packaging record, and marks the student for emergency-grant review: enrolled, gap above cutoff, no emergency award already on the package. It does not add a fund, change packaging, or email an award notice. The counselor confirms transaction 03 is the latest import, confirms packaging matches that snapshot, then applies emergency-grant rules: documentation, remaining need, stacking, fund balance, and professional judgment if it applies.

If scholarship runs can close the same gap, keep this flag downstream of actual packaging. Coordinate with scholarship optimization ML so a pending scholarship is not counted as packaged until it is packaged, and so you do not flag a gap a same-night packaging job will close.

No ISIR. The student has not filed, the record never imported, or verification has the file stuck before a usable transaction. Do not backfill SAI from a prior year, a sibling, a ZIP-code average, or a predictive estimate. That is invented EFC or SAI.

Unusable ISIR. Treat a rejected transaction, a blocking verification hold, or a stored transaction older than the last CPS send as missing until the current transaction is in the SIS.

Stale ISIR under a fresh-looking package. Packaging from an old transaction while a newer ISIR sits unimported, or imported but not re-packaged, is a failure mode. SAI, Pell, and verification can all have moved. Compare only when packaging and ISIR vintages match the office rule, typically the latest imported transaction.

No packaging record. ISIR is in, the student is enrolled, and packaging was never completed in Ellucian or Workday. Do not infer last year's awards. Do not promote a Slate estimated-aid letter to a package.

Mixed timestamps. The packaging datetime predates the ISIR you would cite. Wait or re-package rather than blending vintages.

Partial packaging. Follow local rules: wait for a complete package, or flag only with an explicit partial-package cite so nobody treats the gap as final.

Do not map emptiness to high risk or to no gap. Either cite both sides or output nothing.

The flag may place a student in a review pool; it may not award.

Eligibility belongs to aid: fund-specific rules, remaining need after other aid, hardship documentation, stacking and overaward checks, SAP or enrollment-intensity rules if the fund requires them, and any Title IV constraints that apply to that fund. A cited gap is not a determination.

Send the roster to the aid queue with cites, not to disbursement. Aid must approve any student-facing template so it is not read as an award. A multi-channel enrollment nudge can ask the student to check the aid file or talk to a counselor. It is not a grant.

The grant becomes real only when aid posts it in the packaging system (Ellucian or Workday, not the CRM). After review, write back a disposition: pending documents, ineligible, eligible and packaged, or already covered. Clear or retain the flag from that disposition so nightly jobs do not re-queue the same student.

Inventing EFC or SAI. Filling a blank index with a prior year, a model estimate, or a peer average produces a gap you cannot cite to the need calculation. If the index is missing, the flag stays empty.

Treating the flag as an award. Auto-packaging an emergency grant, posting a credit, or telling the student they have been awarded because the gap cleared a threshold. The number on the flag is a cited remainder. Authorization is a human act.

Packaging from a stale ISIR. Reading a package built on an old transaction after a newer ISIR arrived. Re-package, then compare. Until then, empty is safer than a confident wrong gap.

Related slips: CRM estimates counted as packaged aid, deposited students flagged as enrolled, a model-moved threshold, and skipped verification or professional judgment.

If you keep one rule: cite the package and the need calculation, or write nothing. Emergency-grant eligibility stays with aid.