Mandatory gates, weighted comparison and unresolved risk belong in separate layers.
The evaluation architecture
Mandatory gates
Does the offer satisfy genuine conditions for proceeding?
Weighted evidence
How do qualified options perform against the agreed criteria?
Economics + residual risk
Which tradeoff can the authorized owner accept and operate?
Agree the model before reviewing proposals. Mandatory gates answer whether an offer can proceed. Weighted criteria compare qualified offers. A risk register captures unresolved issues that a score cannot safely absorb. The decision record explains why the selected option is acceptable across all three layers.
Calibrate the scoring scale
| Score | Meaning | Evaluator action |
|---|---|---|
| 0 | Requirement is not met. | Record the demonstrated gap; distinguish it from missing information. |
| 1 | Major gaps or material workarounds. | Explain the operational consequence and remediation dependency. |
| 2 | Partially meets; limitations remain. | Identify the conditions necessary to make the offer acceptable. |
| 3 | Meets the defined requirement with adequate evidence. | Reference the response, demo or test result. |
| 4 | Exceeds the requirement in a way that has verified business value. | Explain the incremental value; do not reward irrelevant features. |
An unanswered question is not automatically a zero. Mark it as unverified and request clarification under the event rules. Do not replace missing evidence with evaluator confidence. Set scale anchors for the particular criterion: a good implementation response means something different from a good functional demonstration.
Test whether the weights represent the decision
Test the effect of your weights
PRIVATE · IN YOUR BROWSERIllustrative suppliers A and B have fixed evidence scores out of 4: capability 4/3, delivery 2/4, economics 3/2. Change weights to see how priorities change the ranking. All gates are assumed passed for this example.
The example weights above are adjustable teaching inputs, not recommended universal percentages. If a modest change in a plausible weight changes the leader, inspect the underlying tradeoff. Do not manipulate weights after seeing bids to justify a preferred supplier. If the model was flawed, document and govern the correction.
Avoid counting the same advantage repeatedly in features, usability, implementation and “overall impression.” Decide whether price is a score or a separate economic comparison. If scoring price, document the formula and inspect its behavior at unusually low or high offers.
Independent review before consensus
Have reviewers score independently and attach evidence before discussing results. Compare the largest differences, resolve interpretation gaps and preserve the reason for the consensus score. Seniority should not determine a score. Demonstrations and references should test specific uncertainties rather than merely reinforce an initial impression.
Normalize implementation, subscriptions, growth and exit costs on the same basis. Then write the decision: what made the preferred option better, which disadvantages were accepted, what must be resolved before signature and who owns each risk. A 0.2-point lead does not cancel an unworkable contract position.
Write the recommendation behind the score
An evaluation record should be intelligible to an approver who did not attend the demonstrations. Separate the observed difference from the business consequence and the proposed treatment. Use precise statements such as “requires a nightly batch interface” rather than “integration is weaker.”
| Observation | Consequence | Treatment before award |
|---|---|---|
| Supplier A performs the core workflow with less operator effort. | Potential operating benefit, subject to representative-volume validation. | Confirm the time and error assumptions in a bounded test. |
| Supplier A requires more buyer integration work. | The implementation timetable depends on internal capacity. | Obtain an accountable resource commitment and price the internal work separately. |
| Supplier B has a stronger delivery plan but a higher normalized cost. | A higher-cost option may reduce a material execution risk. | Present both scenarios and identify the risk the sponsor is being asked to accept. |
The final record may select A, select B or defer the choice. It should explain which evidence supports that outcome. Do not write a new criterion after the event simply because a preferred supplier performed well on it. Where a genuine new fact changes the requirement, follow the applicable process for an amendment, re-evaluation or restart.
Before closing, reconcile evaluator notes, consensus results, risk decisions, pricing assumptions and the negotiated offer. If a supplier changes a decisive feature, team or commercial assumption during negotiation, revisit the affected evaluation rather than carrying the original score forward uncritically.
RFP evaluation matrix
Gates, criterion weights, evidence, individual scores and consensus. CSV · Opens in Excel or Google Sheets.
Sources & context
- U.S. FAR 15.304 · Evaluation factors ↗U.S. federal procurement requirements, cited for context. They are not presented as rules governing private purchases.
- CIPS · Supplier evaluation ↗Supplier assessment is an ongoing activity and can use different forms of evidence.
The decision frameworks and illustrative examples are original editorial guidance. Sources support the stated context; they do not endorse this guide.