Scoring technology proposals: evaluation criteria, weights, and committees

Reference article · 5 min read

A best value selection is only as defensible as the scoring plan written before the first proposal arrives. When criteria, weights, and rating definitions are fixed in advance and applied the same way to every offer, the file shows why one proposal was stronger than another. Assembled afterward, the same elements can make an honest evaluation look reverse engineered to reach a preferred result.

Most state and local procurement codes let an agency design its own evaluation method for competitive sealed proposals, and most require the solicitation to disclose the criteria and their relative importance. Requirements vary by state, are often narrowed by local ordinance, and can carry extra conditions when federal funds are involved. Confirm what applies with your procurement authority and counsel before publishing.

Build criteria from the statement of work

Every scored criterion should trace to something the statement of work asks the vendor to do, staff, deliver, or guarantee. If a criterion maps to no obligation, either the scope is incomplete or the criterion measures a preference the agency cannot enforce.

A workable set for a software or systems procurement usually covers:

• Functional and technical fit. How the solution meets documented requirements, including configuration versus custom development and integration with existing systems.
• Implementation approach. Work plan, schedule, data conversion, testing, training, and cutover, including assumptions about agency staff time.
• Experience and past performance. Comparable work at agencies of similar size and regulatory setting.
• Key personnel. The named people who will do the work, their availability, and substitution commitments.
• Support and product direction. Service levels, release practices, hosting responsibilities, and end of contract obligations.
• Price or total cost of ownership.

Keep the list short enough that each criterion carries real weight. Pass or fail items such as insurance and submission format belong outside the scoring sheet.

Set weights before proposals arrive

Weights express the agency's judgment about where risk lives, not how many words a requirement took to write. Set them while the solicitation is being drafted, record the reasoning in the file, and publish them in the RFP as points or as a statement of relative importance.

Two failure modes are common. Flat weights across many criteria produce clusters of nearly identical totals. A single criterion carrying most of the points makes the rest of the evaluation decorative. If subcriteria will be scored, assign their points up front, so no hidden second layer of weighting appears later.

Decide how price enters the score

Fix the price method before release. Common approaches:

• Scored price formula. The lowest responsive price receives full price points, others a proportional share.
• Two envelope evaluation. Technical proposals are scored before price is opened, which keeps price from anchoring technical judgment.
• Rank then negotiate. Technical rank sets the order of negotiation. Some professional services must use qualifications based selection under state law; confirm whether your scope does.

Evaluate price across a defined term rather than year one. A common pricing workbook separating subscription, implementation, integration, hosting, support, and escalation lets the committee compare totals instead of formats. Include exit assistance and data extraction so a low entry price does not hide a high cost of leaving.

Write rating definitions evaluators apply consistently

A numeric scale without definitions becomes personal preference. Use a short scale and define each level in terms of evidence and risk. On a four level adjectival scale, the top rating means the proposal demonstrates the capability with specific, verifiable evidence and low risk of unsuccessful performance, and the lowest means it does not demonstrate it.

Require a written narrative with every score naming strengths, weaknesses, and any deficiency. Narratives are what a debriefing and a protest response are built from. Before scoring begins, hold a calibration session where the committee scores one sample section together.

Treat demonstrations and interviews as scored events

Demonstrations reward preparation, so structure them. Use the same script, scenarios, time allotment, and audience for every finalist, built on the agency's own workflows and sample data rather than the vendor's showcase. Decide in advance whether the demonstration is its own weighted criterion or evidence informing the functional fit score, and say so in the solicitation. A demonstration can confirm or contradict what a proposal claimed, but it cannot introduce requirements the RFP never stated.

Run the committee with documented mechanics

• Independent scoring first. Each evaluator scores alone before group discussion, and those initial sheets stay in the file.
• Consensus with a stated method. Decide up front whether the committee produces one consensus score with a shared narrative or keeps individual scores that are averaged. Either is defensible; changing midstream is not.
• A facilitator who does not score. The procurement officer runs the process, keeps discussion on the record, and is the only channel for vendor contact.
• Conflict disclosures in writing. Collect signed conflict of interest and confidentiality statements before proposals are distributed, and again if a relationship surfaces.
• Records that survive scrutiny. Keep every version of every sheet and explain score changes in writing. In many states these records become public after award, so write as though proposers will read them.

How to start

1. Draft criteria from the scope outline, deleting any you cannot tie to a requirement, and assign weights and the price method with a short rationale in the file.
2. Write the rating definitions and scoring sheet, and publish both as RFP attachments.
3. Name the committee, collect disclosures, and hold the calibration session.
4. Score independently, reach consensus on the record, and build the award recommendation from the narratives.

HQT helps agencies structure technology requirements and evaluation criteria, and does not bid on solicitations it helped write.

Frequently asked questions