Use gates before scores
Start with requirements that must be met: product specification, necessary qualification, acceptable origin boundary and required delivery window. A candidate that fails a mandatory condition should not pass because it scores well on unrelated dimensions. Agree the gates with engineering, procurement and the decision owner before reviewing supplier names.
Public datasets can support discovery but differ in purpose. The USGS occurrence database, for example, records geological information; its presence in a research list is not a supplier approval signal.
USGS: Global rare earth element occurrence database (2022) ↗
Score the evidence as well as the finding
For each criterion, record the conclusion, source, date and confidence basis. Use an explicit unresolved state rather than assigning a neutral numerical score to a missing answer. Keep desirability separate from evidence quality: an appealing proposed project can have weak evidence of near-term delivery.
Only weight criteria after the mandatory gates are addressed. Show how the ranking changes under a reasonable alternative weighting. If a small adjustment reverses the result, the decision is sensitive and should be discussed as such rather than presented as a precise winner.
Illustrative scoring failure
A supplier earns high marks for geography and price but has no demonstrated qualification for the required grade. A simple average can place it above a qualified candidate. The correct output is 'qualification unresolved' with a named follow-up, not an inflated composite score. The gate protects the decision from attractive but irrelevant strengths.
Minimum scorecard fields
| Field | Purpose |
|---|---|
| Gate status | Pass, fail or unresolved mandatory requirement |
| Finding and evidence | Conclusion linked to a dated source |
| Next action | Question, owner and decision deadline |
Your review checklist
- Agree mandatory gates before weighting preferences.
- Keep unknown separate from satisfactory.
- Retain the basis of each score.
- Test whether the ranking is sensitive to weights.
Investigate with Oreline
Use Oreline to assemble and compare the researched evidence behind the scorecard. The product supports the investigation; your team sets its own purchasing criteria and retains responsibility for the final qualification and weighting decisions.
Follow the evidence through the supply chain.
Explore Oreline’s guided investigations. No account required.
Explore the sandbox →Frequently asked questions
Does Oreline assign a universal supplier score?
No. Its records and analytical tools support your question and evidence review rather than prescribing one ranking for every buyer.
Should missing data receive a low score?
Show it as unresolved first. A low score implies an adverse finding, while missing evidence may instead require a targeted verification step.