Supplier evaluation: criteria, scoring, and evidence

Supplier evaluation compares suppliers against the same requirements and scoring rules before a sourcing decision. It turns delivery, quality, and purchasing evidence into a documented comparison, while leaving eligibility, qualitative due diligence, and the final decision with people.

The useful result is a selection record that another reviewer can rebuild from the source rows, including the assumptions that could change the ranking.

Define the purchase and the gates first

Start with the item or service, specification, site, required capacity, and delivery commitment. Decide what makes a supplier eligible before awarding weighted points. A mandatory certification, approved specification, or capacity requirement should be a gate when failure means the company cannot use the supplier.

Then choose the preferences that separate eligible suppliers. Delivery, quality, and comparable cost are a reasonable starting set for repeat industrial purchases, but they are an example rather than universal weights. The procurement owner should agree on the criteria with operations, quality, and finance before seeing the result.

The federal purchasing framework provides a useful illustration: FAR 15.304 calls for evaluation factors tailored to the acquisition and clear disclosure of their relative importance in the solicitation. That is a federal rule; a commercial team can adopt the underlying discipline of agreeing on priorities before scoring proposals.

For an unproven supplier, use supplier trial success criteria to define the evidence the trial must produce. A supplier without comparable history should receive an explicit evidence-gap flag, not a fabricated historical rate.

Prepare a comparable evidence table

An Ohio manufacturer evaluates Lake Components and Pine Industrial for the same part at the same site. Both pass the required specification and capacity gates. Their historical evidence covers the same review period and comparable work; their current delivered quotes cover the same unit, quantity break, and freight scope.

The table below occupies spreadsheet columns A through F, with headers in row 1 and suppliers in rows 2 and 3.

Due lines come from the purchase-order schedule, including lines still outstanding at the close. OTIF lines meet both the committed date and full-quantity requirement. Inspected and accepted units come from the quality file, with reinspection records resolved so the same physical unit is not counted twice.

Quotes are current offer evidence rather than historical invoice averages. If the decision instead concerns renewal pricing, use matched items and current quantities from invoice exports. Keep the pricing basis identical across suppliers.

Use lead-time variability by supplier and site as additional evidence when inconsistent arrival times drive the operational decision. Do not add it mechanically if it duplicates a delivery criterion already carrying substantial weight.

Calculate the rates and the cost score

Add columns G, H, and I for delivery, acceptance, and cost scores. In Excel, enter the delivery formula in G2, then fill down through row 3:

=IF(B2=0,NA(),C2/B2)

Enter the quality formula in H2 and fill down:

=IF(D2=0,NA(),E2/D2)

Enter the cost formula in I2 and fill down:

=IF(F2<=0,NA(),MIN($F$2:$F$3)/F2)

Each score is a fraction; format it as a percentage for display. The cost formula awards the lowest comparable positive quote the full cost score and scales the other quote against it. Review the whole quote column for blanks and invalid values before using its minimum.

These are business scoring choices, not a scientific measure of overall supplier suitability. A critical defect cannot always be balanced against a lower quote. If it invalidates the specification, it belongs in a gate or separate escalation.

Apply the agreed weights and prove the result

Put delivery, quality, and cost weights of 40%, 40%, and 20% in G5:I5. Their sum must be 100%. In J2, calculate the total score and fill down:

=SUMPRODUCT(G2:I2,$G$5:$I$5)*100

Microsoft's SUMPRODUCT documentation explains that it multiplies corresponding entries and adds the products. Covirage is not affiliated with Microsoft.

Keep unrounded scores in the calculation and round displayed points to two decimal places. The full-precision calculation displays as 97.05 for Lake and 96.60 for Pine. The check is that the three weighted contributions equal each supplier's total and no total exceeds 100.

Also verify each numerator against its denominator: OTIF lines cannot exceed due lines, and accepted units cannot exceed inspected units under this definition. Tie due lines to the filtered order register and inspected units to the quality register before trusting a score.

Test what could change the selection

A sensitivity check shows whether the decision depends heavily on the weights. Change only the weights to delivery 30%, quality 30%, and cost 40%, keeping every underlying evidence row unchanged.

The full-precision calculation for Lake displays as 96.60 points after rounding to two decimal places. Pine now ranks first. That reversal tells the decision maker that the commercial-versus-service priority matters. It does not authorize changing weights after the fact to favor a preferred supplier.

Document the chosen weights, evidence limitations, gate results, and approval rationale. FAR 42.1503 also illustrates the importance of performance assessments supported by objective data and a narrative. For a commercial award, use that evidence discipline while retaining your own approved purchasing process.

Move an approved supplier into recurring supplier management reviews after selection. That workflow checks whether actions and commitments hold over time; it is a different job from choosing the supplier.

Where it goes wrong

• Changing weights after viewing results. Agree on them first and show any later sensitivity analysis separately from the approved method.
• Using unlike evidence. A trial batch and years of mixed production do not provide equivalent certainty. Show the population and the evidence gap.
• Treating blanks as success. Missing inspection or quote evidence is unknown. Stop the affected score until it is resolved.
• Letting a total override a failed gate. A weighted average cannot establish eligibility or replace human technical and commercial due diligence.

Evaluate your own purchasing evidence in Covirage

Bring a sample of your own purchasing, receipt, quality, and quote files: Covirage maps the columns, its tools compute and check the evidence, and the external AI model explains the results with every figure traced to its rows. Create a free Covirage account, which takes about a minute, then set up your business and bring the records behind your next evaluation.

See Covirage for procurement teams for the business context.