Review now
High fit means the published signals align strongly enough to prioritise human review.
Methodology
A deterministic 100-point tender-fit rubric
The score is a transparent prioritisation aid. Every point comes from a published field or an explicit profile rule; no generative model decides the result.
Five visible scoring dimensions
| Dimension | Weight | How it behaves |
|---|---|---|
| Sector match | 35 | Full points when a configured sector appears in published category or sector labels; otherwise zero. |
| Keyword match | 25 | Proportional to configured included terms found in record text. A matched excluded term suppresses keyword points and forces skip. |
| Value fit | 20 | Full points in range, 10 within 20% of a boundary, zero outside. An unpublished value receives 10 because fit is unknown. |
| Location fit | 10 | Full points for a preferred state. If no state preference is configured, the factor is neutral at 10. |
| Eligibility signal | 10 | Checks only published MSME or startup exemption flags. Missing flags receive 5; this is not an eligibility decision. |
Recommendation bands
Review
Medium fit has partial alignment or important unknowns that need manual inspection.
Skip
Low fit, closed records, and records matching an exclusion rule move down the queue.
Deadline behavior
A record is marked closing soon when its published deadline is between zero and 72 hours after the evaluation time. A deadline in the past marks the record closed and forces the recommendation to skip. Missing deadlines remain unknown.
What the score cannot decide
- Legal, technical, financial, or portal-specific eligibility.
- Whether a notice or amendment is complete, current, or authentic.
- Whether a bid is commercially sensible or likely to win.
- Whether the issuing authority will accept an exemption or credential.
Always inspect the official notice and its latest amendments. TenderFit does not submit bids.
Reproducibility
The same valid profile, tender data, and evaluation timestamp produce the same score, reasons, urgency flags, and recommendation. Review the input and response reference for field limits and formats.