Consensus Scoring in RFP Evaluation: How It Works and What to Record

Why one evaluator's 5 and another's 2 is the start of the meeting

A committee chair opens the scores after the independent round closes. On "technical approach" for the second-ranked proposal, one evaluator gave a 5 and another gave a 2. The other three sit at 3 and 4. The average is 3.6, and it describes nobody on the committee.

If the chair simply averages and moves on, the file shows a number but not the reason for it. A monitor, auditor or unsuccessful offeror who later asks why that proposal ranked where it did will find no explanation, only arithmetic.

Consensus scoring exists for this situation. This article explains what it is, how it differs from averaging, how to run the meeting so that disagreement gets discussed first, and what to write down so the file shows how each final score was reached.

What consensus scoring is, and what it is not

Consensus scoring is a two-step method. Each evaluator first scores every proposal alone, against the same criteria and the same rating definitions. The committee then meets, discusses the proposals criterion by criterion, and agrees on one final score for each criterion for each proposal, with a written reason.

It is not a vote. It is not a negotiation in which the loudest person prevails. It is also not the same as a group that scores together from the start, because the independent round is what gives each evaluator's own reading a place in the record. We cover that first step in evaluator scoring independence.

Two features define the method:

• The final score is agreed, not computed. The committee arrives at it through discussion of the proposal text against the rating definitions.
• The reason is recorded. A short written rationale explains why that score, citing what the proposal said or did not say.

People searching for what consensus scoring is in procurement are often really asking whether the final number must equal the average of the individual scores. It does not have to. A consensus score can differ from the average, provided the file explains why.

Consensus scoring versus average scoring

Both are legitimate ways to turn several individual scoresheets into one score per criterion. Neither is required by the federal rules as such. Which one a committee uses is a method choice that belongs in the issuer's own written evaluation procedures and in the solicitation, and it should be fixed before proposals arrive.

	Average scoring	Consensus scoring
How the final score is reached	Arithmetic mean of independent scores	Discussion, then an agreed score
Handling of a large spread	Absorbed into the mean	Discussed and explained
What the file shows	Individual scores and a formula	Individual scores, the agreed score and a rationale
Main risk	A mean that reflects no one's reading; an outlier goes unexamined	Dominant voices; thin rationales written after the fact
Time required	Low	Higher

Averaging is quick and transparent about the arithmetic. Its weakness is that a spread of 2 and 5 produces the same 3.5 as two evaluators who both said 3.5, and the file cannot tell those cases apart.

Consensus takes longer. Its strength is that a disagreement becomes a recorded conversation about the proposal rather than a number that nobody examined. Its weakness is that a rationale can be vague or written after the fact. Some committees use a hybrid: average where evaluators agree closely, and meet only on the criteria with a wide spread. That is also a method choice, and it should be written down in advance.

Under 2 CFR 200.320(b)(2)(ii), an entity using the proposals method must have written procedures for conducting technical evaluations and making selections. Whichever approach you choose, those procedures are the natural place to state it. Confirm the current text on eCFR, and check whether your state or local rules say more.

What 2 CFR 200.318(i) asks the file to show

For a procurement under a federal award, 2 CFR 200.318(i) requires records detailed enough to show the history of each procurement. Those records include four elements: the rationale for the procurement method, the contract type selection, the contractor selection or rejection, and the basis for the contract price. Confirm the current text on eCFR.

The third element is where consensus rationale does its work. A selection or rejection that the file explains only with a total score leaves the reader to guess how the total came about. A consensus rationale for each final score gives the record a sentence or two of reasoning at each step on the way to the ranking.

The rule does not say "write a rationale for every criterion." That is a practice, not a quotation from the regulation. It is a practice that follows from what the section asks the file to show, and it helps the issuer document what it did.

Two related provisions frame the meeting itself:

• Under 2 CFR 200.320(b)(2)(i), a request for proposals must identify all evaluation factors and their relative importance. The consensus discussion should be conducted against those published factors and nothing else.
• Under 2 CFR 200.318(c), no employee, officer, agent or board member with a real or apparent conflict of interest may take part in the selection, award or administration of a contract supported by the federal award. Anyone in the meeting should already have disclosed or attested.

Under 2 CFR 200.318(k), the recipient settles source-evaluation issues, protests, disputes and claims. That means the issuer's own file is its principal explanation of how the selection was made. For the broader picture of what that file holds, see RFP evaluation documentation.

How to run the consensus meeting from the widest spreads

The most common failure in a consensus meeting is that the committee works through criteria in the order of the scoresheet, runs out of time, and rushes the hardest items. A better approach is to let the data set the agenda.

Step 1: Lock the independent round

Do not begin until every evaluator has submitted and the individual scores are final. If anyone can see another's scores before submitting, the independence of the round is open to challenge. Anchoring, or an accusation of it, is a familiar way for an evaluation to be questioned.

Step 2: Calculate the spread for each criterion

For each proposal and each criterion, find the gap between the highest and lowest individual score. This can be done by hand in a spreadsheet. Sort the list so the widest spreads come first.

Step 3: Set a threshold in advance

Decide before scoring what counts as a spread worth discussing. On a five-point scale, a committee might flag any criterion where scores differ by two or more points. The number is a judgment for the committee and should be recorded in the evaluation procedures. It is an attention signal, not a rule about what the score must be.

Step 4: Discuss the widest spreads first

Open with the criterion where the evaluators are furthest apart. Ask each of the outliers to point to the passage in the proposal that supports the score. Then compare that passage to the rating definition for each point on the scale. Often the disagreement turns out to be a misread page, a missed attachment, or two different interpretations of an ambiguous rating definition.

Step 5: Agree the score and write the reason

When the committee reaches a score, one person records it with a rationale before the group moves on. Do not defer the writing to the end of the day. By then, the reasoning is gone and what remains is a summary of the outcome.

Step 6: Handle the criteria with little spread

Where all evaluators scored within a point of each other, the committee may confirm the score quickly, but the rationale still needs a line. Even a short one, such as "all five evaluators found the stated approach met the definition of a 4 for named staffing and a schedule," is better than nothing.

Software can calculate the spread and flag it. RFPManagement is designed with variance flags against a spread threshold the issuer sets for each criterion. The flags are informational only: they point the committee at the criteria that need discussion and do not set or change any score.

What to record for each final score

A usable consensus record is short but complete. For each criterion for each proposal, the file should hold:

• the individual scores, each attributed to an evaluator;
• the final agreed score;
• a rationale, in plain sentences, that cites the proposal content or the rating definition behind the score;
• who attended the meeting, and any evaluator who was absent;
• whether the final score differed from the average, and why;
• the date and time, in order, so the sequence can be reconstructed.

What a good rationale looks like

A weak rationale restates the score: "Committee agreed on a 3." A stronger one connects the score to the proposal and the definition. For example: "The proposal describes a transition plan but names no responsible staff or dates. Under our definition, a 3 requires a plan with milestones; named owners are required for a 4. Two evaluators who had scored a 4 agreed after rereading section 3.2."

The second version tells a reader who was not in the room what the committee looked at and why the score landed where it did. It does not claim the committee was right. It shows what the committee did.

What to avoid in a rationale

• Relying on facts outside the proposal. If a committee member says, "I know this firm from a prior project," that is not in the proposal and not in the published factors. Record the discussion honestly, and do not score on it.
• Introducing unpublished criteria. A rationale that turns on something the RFP never listed as a factor is a problem under 2 CFR 200.320(b)(2)(i).
• Language about the offeror's size or name. Stay with what the proposal said.
• Silent changes. If an evaluator changes an individual score after the meeting, the original should remain in the record, with the change and its reason alongside it.

Connecting consensus scores to weights and ties

Once each criterion has a final score, the committee applies the weights from the solicitation to produce a weighted total for each proposal. The weights should total 100 and should be the ones published in the RFP. For how the arithmetic works, see weighted scoring for RFPs.

If two proposals end up with the same weighted total, the file should already say how that will be resolved. A tie rule chosen after the totals are known looks like a choice made to reach a result. See RFP tie-breaking procedure for ways to set one in advance.

Consensus applies to the technical and other non-price factors. Price points are normally calculated by formula from the bids received, not agreed in discussion. The committee should not "discuss" its way to a different price score.

Where a procurement uses the proposals method, 2 CFR 200.320(b)(2)(iii) says the award goes to the responsible offeror whose proposal is most advantageous to the recipient or subrecipient, considering price and other factors. The consensus scores feed that judgment. They do not replace it. The award decision remains the issuer's.

Common problems and how to prevent them

A dominant evaluator sets the outcome. If the most senior person speaks first, others tend to follow. A facilitator can ask each evaluator to state their individual score and the evidence for it before anyone proposes a number. Roles and conduct for the committee are covered in RFP evaluation committee.

The meeting skips the hard items. Working from widest spread to narrowest, as above, makes sure the contested scores get the time.

Rationales are written after the award. A file with all rationales dated the same afternoon, long after the meeting, invites questions. Record as you go.

Evaluators feel pressure to move toward the group. A consensus score is the committee's agreed reading. An evaluator who still disagrees should have a way to record that. The note belongs in the file, not in a hallway conversation.

Individual scores get overwritten. The independent scores are evidence of what each evaluator thought before the meeting. Keep them, and add the consensus score alongside them.

A practical checklist before your next consensus meeting

1. State in the written evaluation procedures whether the committee will average, use consensus, or use a hybrid, and what spread triggers discussion.
2. Confirm every evaluator has submitted, and that scores are fixed before the meeting opens.
3. Prepare a list of criteria sorted by spread, widest first.
4. Assign someone to record each final score and rationale during the meeting.
5. Record who attended and who did not.
6. Keep original individual scores and any changes, with the reason for each.
7. Confirm retention with the awarding agency or pass-through entity. Under 2 CFR 200.334, the general rule is three years from the final financial report unless an exception applies, and local public-records rules may require longer.

The Evaluator Scoring and Consensus Workbook is designed to calculate the spread for each proposal and criterion, flag the pairs above the threshold you set, and hold a written rationale beside each final consensus score. It is available now on its store page. If you would like to hear when the evaluation workspace opens, you can join the waitlist, and you can see plans for details.

RFPManagement records and computes. It does not give legal or procurement advice. The award decision stays with your organization.