For: Dealer purchasing leaders and commercial equipment procurement teams.
The decision: Whether to maintain, expand, restrict or remediate a supplier relationship.
60-second answer: Score comparable suppliers against evidence and requirements chosen before the review. Show sample sizes and missing data. Use the result to agree specific improvements or purchasing limits; a weighted total must not conceal a critical unresolved requirement.
Define the decision before choosing weights
A review for a new opening may emphasize complete delivery and commissioning. An installed fleet may need parts continuity and service restoration. Set the requirements for that use case, document any weights, and identify conditions that need separate approval regardless of the total score. There is no universal FEX passing percentage or required number of technicians.
Pull purchase orders with original and revised promises, receiving exceptions, open backorders, service tickets, warranty submissions, credits and contract commitments. Use the same period and product category for comparisons. Keep customer-caused delays identifiable without erasing their effect on the actual opening date.
| Measure | Suggested definition | Interpretation guardrail |
|---|---|---|
| Complete, on-time delivery | Orders complete by the agreed date ÷ orders due | State whether the date is original or formally revised |
| Parts availability | Lines filled as promised ÷ parts lines due | Show critical-part backorder age as well as the average |
| Warranty outcome | Approved, denied and credited amounts for a defined claim cohort | Separate incomplete submissions and pending cases |
| Service restoration | Outage duration and repeat visits by comparable incident type | Show sample count and long unresolved cases |
| Support continuity | Evidence against the written support commitment | Missing evidence is uncertainty, not proof of failure |
Make the review fair enough to be useful
A supplier with two deliveries should not be ranked confidently above one with two hundred because both achieved the same percentage. Note small samples. Distinguish evidence that is unavailable from evidence of poor performance. A refusal to disclose employee names can be addressed with service-company coverage and training confirmation rather than an automatic zero.
Do not reward approvals of ineligible claims as the definition of good warranty behavior. Assess whether the supplier applies the agreed terms, explains decisions and pays approved claims within its commitment. Similarly, a lower invoice price can disappear into change orders, emergency freight or an unsupported configuration; compare delivered scope before scoring cost.
End the meeting with an operating decision
Ask the supplier which causes it accepts, what corrective action it proposes, who owns it and when the result can be measured. Share its own evidence and scores without exposing competitors’ confidential information. A purchasing restriction should reflect the risk, available alternatives and customer commitments. Abruptly switching every order can create a different service problem.
Watch for weights changed after a preferred vendor scores poorly, averages that hide old backorders, and attractive national market figures used to excuse local failures. Market growth does not establish the supplier’s performance on your orders.
Run one category review
- Choose the purchasing decision, review period, definitions and critical conditions before scoring.
- Attach records to each measure and invite corrections to factual errors.
- Agree allocation, remedial actions and a dated review of the specific weak measures.
Continue with FEX
- Assess continuity and financial exposure separately
- Improve the quality of the warranty evidence
- Score a service provider’s visit reports before renewal
- Separate a stalled service call from a capacity problem
FEX editorial guidance: the decision tools and operating recommendations in this guide are FEX analysis. They are not manufacturer specifications, contractual entitlements or industry benchmarks.
Editorial update, October 7, 2026: Removed arbitrary passing thresholds, unsupported supplier-failure predictions and claims about vendors’ motives. Replaced automatic zeroes for missing confidential data with documented uncertainty and alternative evidence.
