Define Requirements → Select Criteria → Assign Weights → Score Each Vendor → Rank and Select. This structured process replaces gut-feel decisions with data-backed comparisons that stand up to scrutiny.
Step 1 — Define Requirements
Start with mandatory requirements as pass/fail gates — ISO certification, minimum capacity thresholds, geographic presence, financial stability. If a vendor can't check every pass/fail box, don't waste time scoring them. This step alone typically eliminates 40-60% of candidates before formal evaluation begins. Document the rationale for each mandatory requirement — you'll need it when a rejected vendor asks why they weren't considered.
Step 2 — Select Criteria
Choose 5-7 criteria. Research shows that beyond 7 criteria, evaluators experience criteria fatigue and weighting becomes arbitrary. Each criterion must be independently measurable — avoid compound criteria like "quality and delivery" which conflate two different dimensions. The most common trap: including criteria that sound important but you lack data to score objectively. If you can't measure it with evidence, don't include it. The 6 default criteria (Quality, Cost, Delivery, Risk, Capacity, Innovation) cover the universal dimensions of vendor performance.
Step 3 — Assign Weights
This is where procurement strategy meets execution. Weighting forces your team to articulate what actually matters. If Cost is weighted at 20% but you know the CFO will overrule any selection that isn't cheapest, be honest about actual decision drivers — misaligned weights produce reports that get ignored. Weight discrepancies between cross-functional team members are diagnostic — they reveal misalignment about priorities. Resolve these before scoring begins.
Step 4 — Score Each Vendor
Use a standardized 1-5 rubric with clear behavioral anchors for each rating level. Score of 3 = meets requirements, no more, no less. Score of 5 = clearly exceeds requirements with verifiable evidence. The most common scoring error is leniency bias — evaluators defaulting to 4s and 5s, compressing the entire scale into the top two points and producing results where every vendor scores 4.0+. Combat this by defining what a 1 and a 2 look like for each criterion before scoring begins.
Step 5 — Rank and Select
The model informs the decision, it doesn't make it. A weighted score is a decision-support tool, not an oracle. If vendor #2 scored 82 and #1 scored 83 but you suspect #2's quality culture is stronger based on site visit observations, document that judgment call. Good procurement judgment uses the model as a structured input, not a substitute. Every selection decision should be defensible with both quantitative scores and qualitative rationale.