Picture the person who scores your proposal. Usually not the contracting officer. An engineer or a logistics specialist borrowed from the program office, who signed a nondisclosure agreement to sit on the board and still has a day job.
In front of that person is a worksheet built from the evaluation factors, written before your proposal existed. They are not hunting for the best firm. They are hunting for sentences they can turn into findings.
Nobody publishes a number for how many losing proposals were fully compliant, and any consultant who quotes you one invented it. But in GAO’s protest report for fiscal year 2025, the top ground for sustaining a protest was an unreasonable technical evaluation.
Who reads it, and what they write down
The FAR says almost nothing about how a board is built. FAR 15.303 names the source selection authority and the contracting officer, then stops. The rest comes from agency supplements. In the Army’s, AFARS Appendix AA, that authority “cannot further delegate,” and on buys over $100 million it cannot be the procuring contracting officer.
The two sections that decide everything
You will hear that the FAR rewrite abolished Section L and Section M. It did not. FAR 15.204-5 still lists Section L, “Instructions, conditions, and notices to offerors or respondents,” and Section M, “Evaluation factors for award.” The rewritten Part 15 keeps both at 15.109-4, same letters, same titles. The labels survived. The citation moved.
Both texts are live. The new numbering applies only where an agency adopted it by class deviation: GSA in November 2025, DOE in January 2026, NARA in February, DoD in March.
Most small business awards never involve a Section L or M at all. FAR 15.204 exempts construction, subsistence, and anything using a special format prescribed elsewhere, commercial buys included, where the criteria sit inside FAR 52.212-2. On a Schedule order, FAR 8.405-2 puts them in the quote request.
The finding is the unit of work
FAR 15.305(a) sets the duty. An agency assesses proposals “solely on the factors and subfactors specified in the solicitation,” and “the relative strengths, deficiencies, significant weaknesses, and risks supporting proposal evaluation shall be documented in the contract file.”
Read that list twice. Ordinary weaknesses are not on it. Under FAR 15.306(d)(3) the contracting officer only has to raise deficiencies, significant weaknesses and adverse past performance, and only inside the competitive range. An ordinary weakness can cost you the award with nobody telling you it existed.
Army evaluation guidance says ratings “must be supported by evaluation findings and narrative statements.” In Torden, LLC, a strength for keeping incumbent management read that experienced people “familiar with the operational tempo, end users, and mission requirements would reduce mission risk.”
What you proposed, why it helps, which risk it reduces. And under FAR 52.215-1(f)(4) and FAR 52.212-1(g), the government intends to award without discussions.
Compliance is a floor, and floors don’t win
Four things get a proposal eliminated rather than downgraded. Arriving late, under FAR 52.215-1 and FAR 52.212-1. Ignoring a format instruction. Blowing a stated page limit. And a deficiency: codified FAR 15.001 calls that a material failure to meet a government requirement, the rewritten text any nonconformance with a material requirement.
Format is not a suggestion. In Supplying Demand, LLC, decided in November 2025, the FBI threw out a price proposal that did not use the required spreadsheet. Offerors who exceed a page limit likewise “assume the risk that the agency will not consider the excess pages,” per Tipping Point Solutions.
Content in the wrong volume is quieter about it. In Flairsoft, Ltd., the solicitation told offerors to assume the agency had “no knowledge of the offeror’s facilities and experience,” and GAO refused to make technical evaluators go read the past performance volume.
One myth to retire: no FAR section requires a compliance matrix. Build one anyway, tracking each solicitation paragraph to the page where you answered.
Mirror the evaluation factors, not your outline
The two sections were written as a matched pair. Army guidance on developing the request for proposals requires “a direct linkage” between the requirements, each evaluation factor and subfactor, and the proposal instructions, and orders the team to “never ask for information that will not be evaluated.”
So build the volume from the government’s headings, verbatim. Open each one by naming the criterion, then answering it. Never rename a factor. “Our Approach” where the solicitation says “Technical Capability and Management Approach” makes a tired reader translate before scoring, and some won’t.
“Everything I wrote about a proposal could end up in front of agency counsel, so I wrote only what the proposal handed me. More than once I watched a firm lose on a requirement I was fairly sure they could meet, because nobody on my team could find where they said so. I never got to tell them that.”
Strengths are the product
The FAR does not define “strength.” FAR 15.001 defines deficiency, weakness and significant weakness, and that is the whole list. The working definition comes from agency procedures and, more usefully, from the solicitation in front of you.
A real one, quoted by GAO from an Army solicitation in Inalab Consulting, Inc.: a strength is “an aspect of an Offeror’s proposal with merit or will exceed the Government’s specified performance or capability requirements to the advantage of the Government during task order performance.”
Meeting a requirement well earns nothing. In that same case the Army praised the protester’s handling of classified information at the debriefing but assigned no strength, because the proposal met the requirement without exceeding it. GAO agreed: an evaluation record need not “prove a negative.”
The line is between a claim and a mechanism. A GSA solicitation quoted in Sev1Tech, Inc. put it plainly. The government wanted “a coherent discussion of how the offeror proposes to meet its requirements, rather than a mere restatement of the requirements.”
The same paragraph, rewritten
Say the criterion reads: the government will evaluate whether the quality control plan is comprehensive, verifiable, and self-implementing. Here is what most firms submit.
Our team brings extensive quality management experience and a proven, industry-standard quality control plan that ensures consistently high-quality deliverables. Our processes are fully compliant with the performance work statement.
Nothing there can become a finding. No criterion named, no mechanism, nothing exceeded, no benefit attributed. Now the same content positioned, with brackets you fill from your own records.
Comprehensive: the plan inspects [the deliverable classes in the work statement] against [the quality levels in the surveillance plan]. Verifiable: [named role] records every inspection in [named tool], available to the contracting officer’s representative within [interval]. Self-implementing: when a metric falls below its level, [named role] opens and closes a corrective action within [interval], unprompted. On [prior contract] this cut [defect measure] from [X] to [Y]. The benefit to the Government is that quality problems get fixed before they reach a deliverable.
Every criterion word appears verbatim. A mechanism is named, a person owns it, a tool holds the record. The last sentence states the benefit in risk terms, which is what an evaluator lifts into a finding.
One requirement, two responses
Take the dullest line in a work statement: the contractor shall provide a monthly status report covering work performed, metrics against the performance standards, issues, and planned activities. Most firms answer that they will submit it as required, and that their reports will be accurate, timely and comprehensive.
That restates the requirement and promises compliance. The same shape drew a weakness in the Inalab record, where evaluators wrote that the proposal “made a few brief references to providing information on a monthly basis in other sections, but failed to provide a clear approach.”
The positioned version names the delivery date, puts a table of each performance standard, its result and a red, amber or green status on page one, says where the numbers come from, and closes with the benefit: the review meeting is spent on exceptions instead of reconstructing status. Same requirement, two outcomes. One firm mentioned it in passing across several sections. The other answered it where the evaluator was looking.
Past performance is an argument about relevance
FAR 15.305(a)(2) tells evaluators to weigh “the currency and relevance of the information, source of the information, context of the data, and general trends in contractor’s performance.” Relevance turns on scope, magnitude, complexity and recency.
Magnitude is the one firms misread. It is measured against this buy, not against your company: the dollar value and staffing size of the prior job, beside the dollar value and staffing size of the work being competed.
The record is not yours to write. Under FAR 42.1503 agencies prepare evaluations in CPARS, rating technical quality, cost control, schedule, management, and small business subcontracting from exceptional to unsatisfactory. Yours was written by a different contracting officer and is visible to this one.
Neutral is not safe either. FAR 15.305(a)(2)(iv) says an offeror with no relevant record “may not be evaluated favorably or unfavorably,” and against a firm carrying strong ratings, neutral loses.
So manufacture relevance. 13 CFR 125.11 gives a small business two routes almost nobody uses: a joint venture’s contracts, if you identify the venture to the contracting officer and state what duties your firm carried out, or a first-tier subcontract under a prime contract with a subcontracting plan, if you request a written rating from the prime within 30 days after performance ends.
That deadline is one reason what a teaming agreement says about performance records is worth negotiating early. Then write the narrative as a mapping, not a memoir: the element of the current requirement, then the matching element of the prior contract, with dollars, people and dates attached. Same discipline as a capability statement buyers actually read, at higher resolution.
Risk, and the question nobody prints
Under every technical factor sits one unwritten question. Can this firm actually do it.
It is in the regulation, just not in the evaluation section. FAR 9.104-1 makes the contracting officer determine before award that you have adequate financial resources, that you can meet the schedule “taking into consideration all existing commercial and governmental business commitments,” and that you have the organization, experience and facilities the job needs.
Answer it before someone else does, because an unaddressed staffing gap becomes a weakness written by a person who does not work for you. Transition surfaces it first, and the criterion names its own targets: the Inalab solicitation promised to review “a detailed schedule and required staffing, on-boarding, badging, and knowledge transfer approach.” Five items, answered with a table plus narrative, earned the winner two strengths.
The Torden solicitation demanded at least 98 percent of the required linguists by day 90. A number like that can be checked, which is why it outscores an adjective.
Price positioning without a race to the bottom
Fix one piece of vocabulary, because it trips up experienced people. Lowest price technically acceptable is not the opposite of best value. FAR 15.101 describes a best value continuum and both approaches sit on it.
They behave nothing alike. Under FAR 15.101-1 the solicitation must say whether the non-cost factors combined are significantly more important than, approximately equal to, or significantly less important than price, and the benefits of a higher priced proposal “shall merit the additional cost.” Under FAR 15.101-2, “tradeoffs are not permitted,” so exceeding a requirement buys nothing and costs money.
The rewritten Part 15 adds two more approaches at 15.103-3 and 15.103-4: highest technically rated with a fair and reasonable price, and phased acquisition. On defense work, DFARS 215.101-2-70 sets eight conditions on lowest price technically acceptable and tells contracting officers to avoid it for knowledge-based services.
FAR 15.404-1 separates price analysis, which compares your price against the other prices received, from cost realism analysis, which tests whether your estimated costs are realistic. Realism is required on cost-reimbursement contracts and permitted on competitive fixed-price work only in exceptional cases. So before you discount, look in the evaluation section for language about a price so low it shows you misunderstood the requirement.
FAR 15.308 then makes the source selection authority document the rationale for tradeoffs, including benefits associated with additional costs. Nobody pays a premium for something no evaluator wrote down. In Inalab the protester was rated Good, 10 strengths and 2 weaknesses, at $22,092,936. The awardee was rated Outstanding, 13 strengths and none, at $22,359,148, and won at 1.2 percent higher. Findings beat a discount.
What people get wrong, and why it feels right
Organizing the proposal by your own logic feels like good writing, because everywhere else it is. The instinct comes from a market where the buyer reads a story, and this reader is filling in a form. Burying the answer in paragraph four is the same habit.
Writing to the work statement instead of the evaluation criteria feels rigorous. The work statement says what the job is. The evaluation section says what gets scored. In Inalab, the protester argued the awardee had ignored a work statement paragraph, and GAO answered that the solicitation never required offerors to address it.
Treating the small business participation factor as a formality is common. It is not universal: FAR 15.304(c)(4) requires it only for consolidated or bundled requirements offering significant subcontracting opportunities. Where it appears it carries its own rating, and the Army’s participation commitment form wants named subcontractors with CAGE and NAICS codes and dollars that become “an enforceable contractual requirement.”
Submitting without a second reader feels acceptable at two in the afternoon on the due date. The check is not proofreading. Hand someone who has never seen the draft the evaluation section and the proposal, ask them to find each criterion’s answer in order, and time them.
What to do next
Then lose well. FAR 15.506(a)(1) gives you 3 days after notice of award to request a debriefing in writing, and it should happen within 5 days. If you were excluded from the competitive range instead, FAR 15.505(a)(1) gives you the same 3 days.
FAR 15.506(d) sets what you are owed: the evaluation of significant weaknesses or deficiencies in your proposal, the evaluated price and technical rating of both you and the winner, and the rationale for award. On defense awards over $15 million and not over $150 million, DFARS 215.506 also lets a small business or nontraditional defense contractor request the redacted source selection decision document, and above $150 million the agency provides it unasked. A DoD class deviation renumbered that rule to DFARS 215.301-170, so name both.
Two limits. Under FAR 16.505(b)(6) a debriefing attaches to a task order only when the order exceeds $7.5 million, and the protest stay in FAR 33.104(c) runs off a debriefing that was required.
When our bench works a response, the first deliverable is never prose. It is the evaluation section retyped as headings, a strength target under each, and the verified numbers we may use. That is most of what our proposal work really is.
An evaluator can only credit what is written down. Everything true about your firm that never reached the page was, as far as that file is concerned, never true at all.