Writing selection criteria after supplier meetings makes it easy to overweight what your favourite presenter did well. A structured RFP defines intended outcomes, acceptable evidence and comparison methods before promises and totals arrive.
This proposed model is for a private technology purchase, not tender regulations or legal wording. Adapt it to your project and policies and send the same rules to each company. Its purpose is an explainable evidence-based choice rather than a persuasive meeting alone.
A short request suppliers can answer
A fictional field-service company wants a portal for creating, allocating and closing service requests. Its brief identifies users, existing systems, first-phase boundaries, exclusions and inputs. It need not describe dozens of speculative features it will not buy.
Give important outcomes identifiers. Requirement A might let an employee track their own request without seeing unauthorized ones. Ask the supplier to map solution, limits, assumptions, evidence and associated effort to that identifier, instead of making you search a sales presentation.
Essential gates before weighted comparison
Identify requirements whose absence blocks selection, such as agreed access and handover requirements or a core function. Request suitable evidence without making every preference mandatory; treating a cosmetic detail as essential can unnecessarily narrow options.
World Bank guidance distinguishes mandatory requirements from weighted differentiators in its context. That clarity is useful without transferring international tender or funding rules to your contract. Where your organization has binding procedures, an appropriate specialist checks compatibility.
A fictional weight matrix
| Criterion | Illustrative weight out of 100 | Evidence requested |
|---|---|---|
| Process and scope fit | 30 | Outcomes mapped to the solution, boundaries and exclusions. |
| Delivery and verification | 25 | Phases, acceptance cases and evidence suited to complexity. |
| Handover and continuity | 15 | Defined files, accounts, permissions, documentation and knowledge transfer. |
| Operations and risk handling | 10 | Responsibilities, dependencies and an exception or change route. |
| Comparable-scope cost | 20 | Build, operations and dependencies over an agreed horizon. |
The illustrative weights total one hundred. They are neither market recommendations nor universally suitable. Operational continuity or a particular integration may deserve greater weight. Decide before receiving responses and avoid repeatedly scoring the same requirement without justification.
Define what scores mean
For the first four criteria, an original simple scale could assign zero to no answer, one to a generic claim, two to a specific bounded answer with assumptions, and three to that answer plus appropriate evidence addressing the important risk. Define evidence examples so reviewers interpret suitability consistently.
Divide the score by three and multiply by its weight. A score of two under a thirty-point criterion contributes twenty points. Arithmetic cannot repair weak judgement: record reasons, evidence locations and unproven aspects beside each result.
Normalize cost scope before scoring it
Agree period, usage, currency, build and operating items, inputs and exclusions. An offer including integration and training is not directly comparable with one excluding them. Define treatment of conditional or unpriced work instead of silently treating it as free.
If using relative cost points, publish the method beforehand. For positive comparable costs, one possible formula divides the lowest acceptable cost by the candidate's cost and multiplies by cost weight. This is optional and does not establish that cheapest is best. Exclude failed essential gates or unresolved scope from that calculation.
Review evidence after individual assessments
In the example, operations reviews request flow, technology reviews feasibility and purchasing reviews commitment and cost. Each records observations first, then the group discusses differences. GCA guidance supports consistent criteria and recorded score reasons; this borrows the principle rather than its local legal procedures.
Record missing-evidence questions, answers and changed commitments. If the requirement materially changes, clarify its new version and comparison approach instead of changing rules for one supplier. Do not share another bidder's documents; request evidence against your own needs.
Keep the decision understandable afterward
Record the chosen option, fit, accepted differences, conditions to close and deferred work. Scores organize comparison but do not erase a critical concern or turn assumptions into facts. A bounded experiment may be preferable when important uncertainty remains.
Use the software company selection guide for the wider framework and scope comparison worksheet for existing offers. Discuss technology consulting with Nextwave when preparing a request with outcomes, evidence and decision ownership.
Method references
World Bank rated-criteria guidance and GCA bid-evaluation guidance, updated May 2026, were opened on 3 November 2026. The original weights, scale and example are neither accreditation standards, legal advice nor market rankings. The text was prepared with AI assistance and review of the content and sources.
