# What are the weighted criteria?


**Use the following default weights for a recruitment-intelligence purchase. They total 100.** The first three criteria account for 55 points: useful evidence, access to the required evidence and correctness.

The anchors are proposed standards. Before the trial, translate “required”, “material” and “acceptable” into the club’s own task-level acceptance conditions.

| Criterion | Weight | Poor: 0–1 | Acceptable: 3 | Strong: 5 |
|---|---:|---|---|---|
| **1. Decision contribution beyond the agreed baseline** | **25** | Repeats existing answers or produces unsupported novelty | Resolves the named gap with supported findings unavailable from the baseline within the agreed task conditions | Repeatedly resolves important gaps across the test cases, including useful exclusions or challenges to the initial view |
| **2. Competition, season and evidence coverage** | **18** | Catalogue claims do not translate into usable priority evidence | Meets the agreed match, history, field and delivery requirements; exceptions are visible | Meets requirements across routine and difficult cases, with transparent missingness and reliable delivery |
| **3. Accuracy, identity and provenance** | **12** | Wrong identities, unexplained discrepancies or untraceable claims | Records and claims meet agreed sample tolerances; sources and definitions are inspectable | Reproducible evidence, strong exception handling and corrections demonstrated in the club’s sample |
| **4. Methodological and contextual validity** | **8** | Rankings or judgements conceal assumptions and unsuitable comparisons | Method, comparison population and uncertainty are explained and appropriate to the task | Validation relevant to the intended population and robustness checks support the proposed use |
| **5. Workflow usefulness and adoption** | **8** | Demonstration works only with supplier assistance or creates more verification work | Intended users complete the full task within agreed time and quality limits | Users repeatedly complete tasks with less handling, low rework and usable outputs for decision meetings |
| **6. Integration and portability** | **8** | Required joins, imports or exports fail; manual reconciliation is opaque | Data or reports enter the club’s workflow through a tested, documented process | Refresh, correction, reconciliation and exit processes work without dependency on one person |
| **7. Licence fit and exit rights** | **7** | Intended use is prohibited or materially ambiguous | Required workflows, users, entities and end-of-term treatment are addressed in the governing documents | Necessary flexibility and exit arrangements are clear, practical and tested where applicable |
| **8. Confidentiality, privacy and security** | **4** | Excessive access or uncontrolled handling of club information | Approved access, retention, processing and incident arrangements are documented | Controls are demonstrated through access, revocation and deletion tests, with appropriate assurance evidence |
| **9. Service reliability and continuity** | **3** | No accountable response or correction process | Support, escalation, delivery responsibilities and update arrangements meet the brief | Trial incidents are resolved effectively; continuity and change processes are demonstrated |
| **10. Fully loaded cost and commercial fit** | **7** | Headline fee omits essential scope or implementation effort | Comparable quote and cost model fit the approved resources and contract horizon | Costs, dependencies and renewal/exit exposure are transparent, with a stronger evidenced trade-off than alternatives |
| **Total** | **100** | | | |

### Can the weights change for different buyers?

**Yes—before supplier results are known, and identically for every option in the same decision.**

- **NWSL recruitment:** consider increasing coverage and contextual-validity weights where women’s competitions are the central uncertainty.
- **Lean MLS department:** consider increasing workflow and delivery weight where staff capacity is the constraint.
- **Analytics-led purchase:** consider increasing integration and accuracy weight where the output must enter a production database.
- **Multi-club consolidation:** consider increasing licence and security weight where entity boundaries and shared access are decisive.

Keep the total at 100 and document the reason for each change. Gates remain gates regardless of weight.

Do not demand an API from a human scouting service when an authorised structured report import meets the requirement. Likewise, do not mark integration “not applicable” when the proposed plan depends on it. If a criterion genuinely does not apply to the decision, remove and redistribute its weight **before** inviting proposals—not after a weak result.