# How can the club run a bounded supplier trial?


**Use a ten-business-day pilot as a planning template, not a guaranteed implementation timetable.** Complete permissions and scope agreement before the clock begins.

A proposed starting sample is eight assessment players, three priority competition–season combinations and 24 independently selected required matches. Add a separate discovery brief, an identity challenge set for integration purchases and a claim-level sample for AI reports. Increase or redesign the sample where risk demands it; these quantities have no claimed statistical power.

| Stage | Activity | Required output |
|---|---|---|
| Before day 1 | Permissions, scope, gates, weights and cost horizon agreed | Signed trial brief and authorisation record |
| Days 1–2 | Capture baseline answers; select routine and difficult cases | Baseline ledger and frozen sample |
| Days 3–5 | Equivalent discovery and assessment tasks | Results, evidence references and task timings |
| Days 6–7 | Coverage, identity, import/export and provenance checks | Coverage map and defect register |
| Day 8 | Repeat selected cases; test support and corrections | Reproducibility and resolution records |
| Day 9 | Reconcile quotes, rights and implementation dependencies | Cost model and approval conditions |
| Day 10 | Independent scoring, adjudication and decision meeting | Retain/add/replace/consolidate/build/no-buy memo |

Keep a **routine-use sample** separate from a deliberately difficult **challenge set**. Report both; do not present challenge-set failure rates as representative of all supplier records.

Give suppliers an opportunity to explain or correct discrepancies, but preserve the first observation. Record whether the fix is general, a one-off manual intervention or a future commitment. Retest affected cases before changing a score.

### What should the scoring workbook contain?

Use eight tabs:

1. Decision brief and baseline.
2. Entitlements, rights and gates.
3. Competition–season coverage map.
4. Tasks and incremental-evidence ledger.
5. Accuracy, identity and AI claim checks.
6. Workflow, integration and support observations.
7. Comparable cost model and loss register.
8. Scores, unresolved issues and recommendation.

Copy this header into the scoring tab:

```text
Option | Criterion | Weight | Rating_0_to_5_or_U | Weighted_points | Evidence_status | Evidence_reference | Observation_date | Scope_and_version | Reviewer | Rationale | Open_issue | Retest_due
```

Do not upload confidential trial material to an external tool merely to populate the workbook. Use the club’s approved processing environment.