Scoring Methodology — Fed/State Ops

Overview

Opportunity Scout pulls active solicitations and pre-solicitations from multiple jurisdictions, then scores each opportunity on a 0–93% Win Likelihood scale. This is not a statistical win probability — it's a relative priority rank designed to surface the best-fit, lowest-friction opportunities for a small business at an early stage.

Scoring runs entirely server-side in the Cloudflare Worker. The formula is not visible in page source. All scores are computed fresh on each scan, enriched with historical data from USASpending.gov (federal records only), and can be refined further by Claude AI analysis of the actual solicitation document.

Sources & taxonomy

Every opportunity carries a source tag: federal (SAM.gov, pulled live against your 7 NAICS codes), state-md (Maryland eMMA) or state-va (Virginia eVA). Federal records are NAICS-coded; state records are UNSPSC-coded — both are normalized into a shared classificationCode field. State opportunities are loaded via upload (JSON or a fixed-column CSV template) and persist in the Worker's KV store; live state-portal adapters are stubbed pending API access.

Because state records have no NAICS, the profile-code boost and the NAICS exclusion list don't apply to them — scope alignment is judged from the title keyword patterns alone (see Step 3), and the agency-tier penalties (Step 6) are federal-only.

Win Likelihood = Base(25) + Eligibility + Value + Set-aside + Remote + Keyword − No-match penalty + Profile-code signal + Complexity penalty + Competition density − Red flag penalties − Agency tier penalty (federal) / + State buyer credit (state) → clamped 8%–93%

Step 1 — Exclusion Filter

Before scoring, any opportunity matching an exclusion pattern is flagged "off-profile" and given a score of 5% (excluded from ranked view by default). Exclusions are hard filters — they indicate scope that is structurally outside your capabilities. Federal records are also excluded by out-of-scope NAICS code (construction, manufacturing, etc.); state records, having no NAICS, are filtered by the title patterns below only.

Category	Example keywords
Aviation / aerospace	aircraft, aviation, airframe, rotor, helicopter
Weapons / ordnance	weapon, ordnance, ammunition, missile, artillery
CBRN / hazardous materials	nuclear, radiological, hazmat, biological agent
Heavy civil construction	dam, bridge, freeway, dredge, excavate
Medical devices	implant, prosthetic, surgical instrument, diagnostic kit
Food service / agriculture	cafeteria, catering, livestock, crop
Electronics manufacturing	circuit board, PCB, semiconductor, cable assembly
Naval vessels	vessel, ship, submarine, destroyer

Step 2 — Base Score & Adjustments

Base

Every non-excluded opportunity starts at 25% — roughly the average SBA-reported small business win rate on competitive federal contracts.

Eligibility

Condition	Adjustment
Eligible now (SB, open competition, or unrestricted)	+25%
Future eligible (WOSB, SDVOSB, 8(a), HUBZone — cert gap)	+5%
Other / check required	+0%

Contract value

Contract value	Adjustment	Rationale
Unknown / $0	+5%	Unpriced — assume small
≤ $25K	+20%	Micro-purchase range — easiest to win
$25K–$100K	+16%	Simplified acquisition threshold
$100K–$250K	+10%	Small enough to execute easily
$250K–$500K	+4%	Moderate size
$500K–$1M	−4%	Increasing competition and scrutiny
> $1M	−14%	Large contracts attract established contractors

Complexity

Complexity (Low −0% / Medium −5% / High −14%) is decoupled from raw dollar value. Large contract size alone does not imply execution difficulty for a low-friction service.

Category	Complexity rule
Janitorial, Grounds/Lawn, Software Resale, Admin Support	Low up to $1M; Medium above $1M — never High from size alone
Real Estate, Consulting, Mental Health, unclassified	Standard value tiers: Low <$150K, Medium $150K–$1M, High >$1M
Any category with a Systems Engineering, Enterprise Scope, or COOP red flag	+1 complexity level (capped at High)

Complexity also drives sort order — the ranked list is ordered by WinLikelihood² ÷ Complexity², so a High-complexity opportunity ranks well below a Low-complexity one at the same win likelihood.

Set-aside & location

Condition	Adjustment
SB set-aside or unrestricted	+10%
Nationwide / remote performance	+4%

Step 3 — Keyword Boost & No-Match Penalty

The opportunity title is scanned for keywords matching your service profile. The first matching pattern applies its boost and labels the opportunity (e.g. "Janitorial", "Mental Health").

Category label	Trigger keywords (sample)	Boost
Mental Health	mental health, counseling, therapy, behavioral health, clinical, psycholog	+20%
Janitorial	janitorial, custodial, cleaning, housekeeping, sanitation	+18%
Grounds/Lawn	lawn, landscape, grounds, mowing, turf, vegetation	+18%
Consulting	consulting, advisory, management support, program support, acquisition support	+15%
Software Resale	software, license, subscription, SaaS, technology resell	+14%
Real Estate	real estate, property, space search, facility finder, lease	+14%
Admin Support	administrative, admin support, office support, clerical, records management	+12%

No-match penalty: −12%
If no keyword pattern matches the title, the opportunity scores a −12% penalty. Vague titles (e.g. "Support Services — Fort Bragg") are likely misclassified or out-of-profile even if the NAICS code matches. Unknown scope defaults to skepticism.

Profile-code signal

Federal records get a NAICS boost when the code is in the profile:

NAICS	Description	Boost
621330	Mental Health	+12%
561720	Janitorial Services	+10%
561730	Landscaping / Lawn Care	+10%
541611	Admin & Mgmt Consulting	+8%
541618	Other Mgmt Consulting	+8%
511210	Software Reseller	+7%
531390	Real Estate	+6%

State / regional records carry a UNSPSC code, not a NAICS, so there is no profile-code table to match. Instead, a state record whose title matches one of the keyword patterns above receives a flat +8% — the median of the NAICS boost values — standing in for the profile hit. This replaces the NAICS boost; it is not added on top of it.

Step 4 — Competition Density (USASpending)

USASpending.gov is queried for recent awards under the same NAICS code. The query runs in up to three passes: (1) small-business set-aside awards only (FPDS codes SBA + SBP) in a 3-year window; (2) if that returns fewer than 5 results, all awards in the 3-year window; (3) if still sparse, a 6-year window. Every pass is value-capped per NAICS. This is federal-only — state records have no NAICS to query. The number of unique awardees serves as a proxy for market competition.

Unique awardees	Adjustment
≤ 5	+8% (niche market — less-contested)
6–20	+0%
21–50	−4%
51–100	−8%
> 100	−12% (crowded market)

USASpending also computes incumbent concentration: if the top 2 recipients hold >60% of award value under this NAICS, the market is flagged as high-concentration and a warning is shown in the detail panel. This is informational and does not change the score directly (competition density captures it indirectly).

Step 5 — Red Flag Deductions

Title patterns that signal structural disadvantage are scanned. Total deduction is capped at −30% regardless of how many flags trigger. Up to 2 flag labels are shown in the opportunity row.

Flag	Trigger keywords (sample)	Deduction
Clearance Req	classified, secret clearance, top secret, TS/SCI	−8%
O&M	operations and maintenance, O&M	−15%
Incumbent Signal	incumbent, recompete, follow-on contract, option exercise	−12%
IDIQ/Vehicle	IDIQ, MATOC, GWAC, BPA, blanket purchase agreement	−12%
Sys Engineering	systems engineering, systems integration, systems architect	−12%
COOP	continuity of operations, COOP	−10%
Enterprise Scope	enterprise-wide, enterprise solution, portfolio management	−10%
Health IT	medical records, EHR, electronic health record, clinical data	−8%
Logistics	logistics support, supply chain, inventory management	−8%

Step 6 — Agency Tier Penalty (federal) / State Buyer Credit (state)

State and regional records skip this step's penalties entirely and instead receive a flat +4% state buyer credit, reflecting the generally lighter clearance, compliance, and audit burden of state and municipal procurement.

For federal records, the agency/department path returned by SAM.gov (and the title as fallback) is checked against known high-friction agency categories. These agencies tend to require prior experience, clearances, or have procurement cultures that favor large primes and established incumbents.

Label	Agencies	Deduction
IC Agency	NSA, CIA, DIA, NRO, NGA, Intelligence Community	−18%
DoD	Army, Navy, Air Force, Marines, DARPA, DLA, DHA, Pentagon	−6%
DOE	Department of Energy, National Nuclear Security Administration	−8%
NASA	National Aeronautics and Space Administration	−8%
Health Research	NIH, National Cancer Institute, CDC	−4%

Civilian agencies (GSA, VA, USDA, HUD, SBA, etc.) receive no agency penalty — these are better-fit customers for early-stage small businesses.

Why this matters: DoD represents a large share of federal contracting, including significant mental health, consulting, janitorial, and grounds work — so a blanket heavy penalty would exclude too many viable opportunities. The small −6% reflects modest additional friction (procurement complexity, past performance expectations) without deprioritizing DoD as a customer category. Clearance-required solicitations receive a separate −8% red flag rather than being screened out, since team members with prior military/government service may already hold or be able to reactivate clearances.

Score Signal

Each score is accompanied by a Signal indicator (High / Med / Low) that tells you how much data supports the score — not whether the score is good or bad.

● High signal Strong keyword match (category boost triggered) AND NAICS boost AND USASpending competition data available. Score reflects real information about scope alignment and market density.

● Med signal NAICS boost applies, or USASpending data available, but keyword match is absent or weak. Score is reasonable but title is ambiguous.

● Low signal No keyword match, no external data. Score is a conservative structural estimate — treat it as a flag for closer review, not a reliable likelihood.

USASpending Enrichment

After the top 30 results are scored, each is enriched asynchronously with data from USASpending.gov. The enrichment runs in the background — columns fill in as results return.

Est. Value: Median award amount across comparable contracts (same NAICS, small-business set-aside awards preferred, 3-year window widening to all awards / 6 years when sparse)
Confidence: How well historical awards match this specific solicitation — higher when title keywords appear in award descriptions
Competitors: Number of unique awardees under this NAICS in the dataset
Incumbent concentration: Share of total award value held by top 2 recipients

USP data is cached per NAICS code for 7 days. All opportunities sharing a NAICS see the same median value — this is intentional to avoid excessive API calls.

Why median, not mean or mode?

Federal contracting data is heavily right-skewed. A single large enterprise award — a $50M IDIQ to a major integrator — can pull the mean far above what any small business would realistically bid. Even with per-NAICS value caps filtering the worst outliers, the mean remains sensitive to extremes.

Mode doesn't apply cleanly here because award amounts are continuous: no two contracts land on exactly the same dollar figure. Binning the data into ranges would require an arbitrary choice of bin width that heavily influences the result, and with only 25 data points the mode is essentially meaningless.

The median — the middle value when awards are sorted — is resistant to outliers by definition. It answers the question "what does a typical award under this NAICS look like?" which is the right benchmark for sizing a bid. The confidence score layered on top tells you how well that median applies to any specific opportunity.

Still a heuristic. The 25 results fetched are sorted by award amount descending, so the sample skews toward larger contracts within the cap — it is not a true random sample. NAICS codes are also broad: 541611 (Management Consulting) spans everything from a 2-person policy review to a 200-person transformation program. Treat Est. Value as a useful order-of-magnitude reference, not a precise bid target.

Claude AI Analysis

On demand (click "Analyze ✦"), Claude reads the actual solicitation document from SAM.gov and returns a structured assessment including:

Overall fit: strong / moderate / weak
Staffing estimate and key roles
Required certifications or clearances
Period of performance
Incumbent risk assessment
Win factors and red flags

AI analyses are cached for 90 days. After a scan, any previously analyzed opportunities have their results automatically restored so you don't re-pay for Claude on the same solicitation. All analyzed opportunities are stored in the Analysis Archive.

AI vs. algorithmic score: The algorithmic score is fast and free — it's designed to filter out noise and surface candidates worth closer inspection. The AI analysis is the deeper read. The goal is to close the gap between the two by continuously improving the algorithm based on patterns seen in AI outputs.

Known Gaps & Limitations

What the score does not capture:

Modification history: Heavily amended solicitations (many modifiedDate updates) may indicate incumbent negotiation or scope changes — not yet scored.
Solicitation number format: RFQ vs. RFP vs. Sources Sought indicate different procurement vehicles and effort levels — not yet extracted.
Agency SB utilization rate: Some agencies award 50%+ to small businesses; others rarely do. Per-agency SB rate from USASpending is not yet incorporated.
Geographic mismatch: Specific non-remote locations that require local presence are not penalized unless the title flags it.
Contract type (FFP vs. T&M vs. CPFF): Fixed-price contracts are generally more favorable for small businesses. Contract type is only visible in the solicitation document, not the SAM.gov listing.
Solicitation age: Very recently posted solicitations with tight deadlines may indicate a wired bid or last-minute add. Not scored.
Past performance requirements: Many solicitations require 3–5 years of directly relevant past performance. Captured in AI analysis but not in algorithmic scoring.