How Self-Scored Federal Proposals Work, and Where Offerors Lose Points

A growing number of large federal contract vehicles don't ask you to write a traditional proposal. Instead, they hand you a scorecard. You score your own company against it, attach the evidence, and the government checks your math.

GSA's OASIS+ works this way. So do the Army's MAPS competition and USSOCOM's SOF-GSD. If you've never done one, the format feels easier than a written proposal. In our experience, it's actually less forgiving. We've prepared self-scored submissions across all three vehicles, and the offers that struggle rarely lack experience. They lose points in the documentation.

How self-scoring works

Every self-scored solicitation is different, but most follow the same pattern:

1. The government publishes a scorecard. Each element is worth a set number of points or credits: things like project size, relevance to a domain, number of positions filled, certifications held, or past performance ratings.
2. You score yourself. You pick the projects you'll use, usually called qualifying projects or work samples, and claim the points you believe they earn.
3. You submit evidence for every claim. Contracts, statements of work, modifications, performance ratings, verification forms, rosters or other records that prove each point.
4. The government validates. Evaluators check each claim against your documents. Points that aren't clearly supported are removed.
5. Awards follow the validated scores. Depending on the vehicle, that means either everyone who clears a minimum score (as on OASIS+) or only the top scorers (SOF-GSD planned 15 awards to its highest-rated offerors).

The key shift is in step 4. The evaluator isn't asking whether your proposal is persuasive. They're asking whether a specific page of a specific document proves a specific claim.

Why "just fill out the scorecard" is misleading

Three things catch first-time offerors off guard.

Unsupported points disappear. On the solicitations we've worked, scores could be adjusted downward, but not upward. SOF-GSD stated this explicitly. If you undersell a project, the government won't add points for you. If you oversell one, it takes them away.

Misleading documentation can cost you the whole offer. Some solicitations go further than removing points. SOF-GSD, for example, warned that an inaccurate self-score based on ambiguous or misleading documentation could eliminate the offer with no chance to fix it. That turns every borderline claim into a risk to the entire proposal.

The evidence rules are strict and vary by vehicle. One solicitation we worked prohibited offerors from creating any new documentation, including project narratives, and accepted only official contract records like the statement of work, deliverables lists and signed contract pages. Another relied on tagged contract documents plus a verification form signed by the customer. What counts as proof in one competition may be rejected in the next.

Where offerors lose points

These are the problems we see most often when we review self-scores.

Codes that look relevant but don't qualify

Some vehicles automatically treat projects as relevant if they carry certain NAICS or product service codes. It's easy to assume a code is on that list when it isn't. We've caught more than one case where a project's code looked like an obvious fit but wasn't actually on the solicitation's list, which meant the project needed a separate, harder form of relevance proof. Check every code against the actual list, not your memory of it.

Documents that don't say what you need them to say

A claim is only as good as the sentence that proves it. If a scoring element counts work performed at specific locations, the location needs to appear in the contract documents. If it counts labor categories, the documents need to show those roles. "We did that work" doesn't count if the paperwork doesn't say so.

Projects that miss a minimum

Qualifying projects usually have floors: a minimum dollar value, a minimum period of performance, a recency window, or a minimum number of staff. We once reviewed a contract that looked like a strong fit until we checked its period of performance: roughly five weeks, against a six-month minimum. Screen every project against every minimum before you build around it.

Subcontract work claimed at the prime's scope

When you use a subcontract, typically only the work in your subcontract counts, not everything the prime contract covered. It's tempting to cite the prime's full statement of work. Evaluators will score your piece, so tag your piece.

Scoring exactly at the threshold

If the vehicle awards everyone who meets a minimum, landing exactly on the line leaves no room for a single point to be disallowed. We plan for a cushion above the threshold, and when a submission sits right at the line, we go looking for additional supportable points before it goes in.

Misreading eligibility questions

Many scorecards open with yes-or-no gate questions. Some are worded so the correct answer feels counterintuitive, such as a question about meeting a threshold for negative performance ratings, where "yes" actually confirms you're clear of it. Answer one wrong and the evaluation can stop before scoring begins.

Points left on the table

The opposite problem is just as common. Certifications, clearances, past performance ratings, or recent contract modifications that could earn points go unclaimed because nobody checked every scorecard element against everything the company has.

How we build a defensible self-score

Our process is built around one question for every point: could an evaluator who has never heard of this company find the proof in under a minute? We start from a client's full history, find the strongest defensible combination of projects, get outside signatures moving early, and have someone who didn't build the score review it as an evaluator would before anything is submitted.

Verify it yourself before the government does

Self-scoring puts the burden of proof on you. The government doesn't go looking for evidence on your behalf; it checks what you point it to. So before anything is submitted, we verify our own claims the same way an evaluator would.

Our rule is simple: every point we claim must be proven in the documentation at least once. If we can't put a finger on the exact page, paragraph and sentence that supports a claim, we don't claim it. Where the documents allow, we go further and support each claim in more than one place, so a single passage that an evaluator reads differently doesn't cost us the point. That discipline is what lets us, and our clients, submit with confidence instead of hoping the score holds up.

How we prepare the documents

A defensible score also has to be easy to verify. Evaluators may be reviewing dozens or hundreds of submissions, and a package where the proof is buried in a 200-page statement of work works against you. So we prepare every document to lead the evaluator straight to the evidence:

• Highlight the supporting passages. Every sentence that proves a claim is highlighted directly in the contract documents, so it stands out on the page.
• Label each highlight. Each one carries a comment naming the exact scorecard element, domain or functional area it supports, using the solicitation's own terminology, never our paraphrase.
• Keep the documents official. We mark up the actual contract records rather than creating summaries, since some solicitations won't accept new documentation at all.
• Cross-reference the package. The scorecard, the verification forms and the highlighted documents all point to each other, so an evaluator can move from any claimed point to its proof and back without searching.
• Organize it in the order it's scored. Files are named and sequenced to follow the scorecard, so the review flows from one element to the next.

The result is a submission package that reads cleanly from start to finish: detailed enough to prove every point, and organized so a reviewer can confirm it quickly. Some solicitations require this kind of tagging for certain projects, but we do it for every project, because an evaluator who can verify your score easily has no reason to reduce it.

Which vehicles use self-scoring?

Self-scoring has become common on large multiple-award vehicles, including:

• GSA OASIS+, which scores qualifying projects, federal experience, systems and certifications against a 50-point scorecard for each domain. Small business pools require 36 of 50 and the unrestricted pool 42 of 50 for most domains. Phase II expanded OASIS+ to 13 domains, and submissions have been open on a rolling basis since January 2026, so there's no fixed deadline.
• Army MAPS, a $50 billion professional services vehicle scored with qualifying projects, performance data, certifications and gate questions.
• USSOCOM SOF-GSD, which scored work samples across categories such as positions filled, performance locations and labor categories, and planned awards to the 15 highest-scoring offerors.

Each solicitation defines its own elements, points and evidence rules, and amendments can change them mid-competition. MAPS is the clearest example: its eighth amendment, in May 2026, required every offeror to rebuild and fully resubmit, changed how CPARS ratings were scored, and expanded what counted as a qualifying project. Always work from the current version.

Where to go from here

See our federal proposal services or browse more proposal guides. If you're preparing a self-scored submission, a second set of eyes on your scorecard, before the government's, is the cheapest insurance you can buy.