Weighted evaluation criteria that hold up

Make the scoring method understandable before suppliers respond and defensible when the team chooses an award.

Start with the decision you need to make

Weighted evaluation combines several preferences into one comparison. It is useful when price alone does not describe the purchase. Start by asking which differences between eligible offers would affect the decision. Cost, delivery, quality and ESG may matter, but each needs a reason tied to the scope. Do not add a criterion only because it appears in a standard template.

A weight is an explicit statement of relative importance. It is not an objective truth supplied by software. The team should discuss the trade-off before collecting offers. If delivery is critical, ask whether a late offer should be ineligible or merely score lower. That distinction changes the method. Write the answer down so a reviewer can understand why the scoring model has its particular shape.

Separate gates from preferences

Mandatory requirements define eligibility. A required specification or acceptance condition belongs in that gate when failure cannot be compensated. Weighted preferences compare the offers that pass. Mixing the two can allow an otherwise attractive offer to offset a failure that the buyer cannot accept. Make the pass conditions and the evidence needed for them visible to suppliers.

Have a process for incomplete evidence. Decide who checks it and how clarifications work. A missing document is different from evidence that a requirement is not met. Apply the same process across offers and record the result. Do not let a reviewer grant informal exceptions to one supplier while applying the published rule to another. Consistency depends on the workflow as well as the formula.

Give each criterion a measurable basis

Define what a score means. For delivery, specify the starting point and the commitment being assessed. For quality, identify the evidence and the scoring rubric. For ESG, decide which relevant evidence can be verified. Avoid overlapping criteria that reward the same fact twice. A delivery promise should not accidentally gain points under both lead time and a loosely defined service score.

Use a scale that reviewers can apply without inventing intermediate judgments. Explain any normalization used to compare numerical values. A low raw number may be favorable for cost and unfavorable for capacity, so the direction must be clear. Test the method with contrasting example offers before publishing it. Examples are a check on the method, not evidence of supplier performance or a customer result.

Review weights and sensitivity

Check that the full set of weights is complete and consistent. Ask what happens when two offers are close. A small change in a subjective score may reverse the order. That is a reason to improve evidence and review assumptions, not a reason to adjust weights until a preferred supplier wins. Retain the published method and explain how the evidence supports the score.

Use sensitivity analysis to understand the decision where the process permits it. Compare reasonable alternative weight sets without rewriting the official evaluation after submission. The discussion can reveal which criterion drives the result and which data deserves closer review. If capacity or contract commitments constrain the award, inspect those separately. A strong weighted score does not guarantee a feasible supplier allocation.

Keep the reasoning with the score

A score without evidence is difficult to review. Record the source, the evaluator and the reason for each non-price judgment. If several people evaluate offers, agree the rubric before they start and resolve differences through a defined review. Keep original supplier responses accessible. A final summary should point to the evidence instead of substituting a polished narrative for it.

After the decision, review how well the method supported the actual purchase. Ask whether a criterion was hard to assess, whether evidence arrived in a useful format and whether the scoring basis was understood. Improve the next event using those lessons. Do not change the record of the completed decision to make it look cleaner. A defensible method is one that another buyer can follow and challenge.

Checklist

• Tie every criterion to the purchasing need.
• Separate mandatory requirements from weighted preferences.
• Define evidence, scales and normalization.
• Check for overlap between criteria.
• Test the method before asking suppliers to respond.
• Retain the scoring reasons and source evidence.