Understanding Your Analysis
Scores & Grades Explained
ClearStaq surfaces several scores in different places. This page is the single reference for what each one measures, its range, and where it appears.
Document-level
| Score | Range | Where it appears | What it measures |
|---|---|---|---|
| Fraud Score | 0–100, banded Low / Medium / High | Documents list, document Fraud Detection tab | Tampering likelihood from StaqGuard's findings — structural (metadata, producer, timestamps) plus ClearStaqAI visual analysis when enabled. 0 = no findings. Higher = more / stronger findings. |
| Health Grade | A+ through F | Documents list | Overall financial health of the statement: balances, cash flow, NSF activity, negative days. A+ = strong, clean account activity. Independent of fraud — a statement can be A+ healthy and still fraud-flagged, or honest but financially weak. |
Client-level (aggregated across all statements)
| Score | Range | Where it appears | What it measures |
|---|---|---|---|
| Overall Risk Score | 0–100 | Client header, Dashboard "Risk Score" column | The worst statement in the batch — deliberately pessimistic. One doctored statement makes the whole client worth reviewing, no matter how clean the rest are. |
| Risk Level | Low / Medium / High (or N/A while analysis is running) | Client Risk Profile tab | The worst of metadata + ClearStaqAI findings across all statements. |
| Max Fraud Score | 0–100 | Client Risk Profile tab | Highest document fraud score on file for this client. |
Specialized metrics
| Metric | Range | Where | What it measures |
|---|---|---|---|
| Balance Volatility Index | 0–1 | Balance Insights | Stability of daily balances: near 0 = steady, near 1 = wild swings. Compare merchants with similar averages. |
| Stacking Risk | LOW / MEDIUM / HIGH | Bank & Debt Summary | Combined read of detected MCA positions, monthly debt service, and debt-to-revenue ratio. |
| Withhold % | 0–100% | MCA History monthly breakdown | Share of revenue a specific funder is pulling in a given month. |
How they fit together
A useful mental model:
- Fraud Score answers: can I trust this document?
- Health Grade answers: is this account financially sound?
- Stacking Risk answers: how leveraged is this merchant already?
- Client Risk Score answers: should a human look closer before anything moves forward?
A fundable deal generally wants: low fraud scores across all statements, B-or-better health grades, and stacking risk consistent with your position appetite. Any single red number is a review trigger, not necessarily a decline.
FAQ
Why does my client show "Risk Score 20" and "N/A" at the same time? The number is the worst statement's computed score; N/A indicates the combined risk level is still being computed (fraud analysis in progress across the batch). When the run completes, the level fills in.
Can two statements from the same bank get different fraud scores? Yes — scores are per document. One original PDF and one re-saved-through-an-editor PDF will score very differently, which is exactly the point.
Do health grades affect fraud scores or vice versa? No — they're independent axes (financial health vs document authenticity). Read both.
Can I set my own thresholds? Custom fraud rules and customizable scorecard weights are available on StaqScale.