How much of each certified small business population is actually capturing federal work, and how widely the money in each lane is spread. Every rate is shown with the numerator and denominator that produced it, so the arithmetic is inspectable rather than asserted.
Loading certification supply and award window...
Notice flow: what the federal market posted, and what we hold
Every contract notice posted government-wide in the trailing 365 days, counted from the full SAM.gov notice corpus (5,488,500 notices), set beside the rows we actually hold. This is the denominator the rest of the portal is measured against. It is the notice axis; the certification panel below it is the entity axis, and the two do not combine.
Loading notice flow...
Loading notice flow...
The gap is footprint, not a collection failure
The denominator is every federal notice of that type, across every industry the government buys in. Our collection is scoped to the NAICS codes on the active company profile, on purpose. So a coverage figure of 28 percent does not mean seventy-two percent went missing: it means the work we chase is roughly a quarter of what the government posts. The figure that would actually grade collection is coverage inside our own NAICS footprint, and it is not computable yet: the corpus carries a NAICS code on zero of its 5,488,500 rows. That is the enrichment pass, and until it lands this panel reads as footprint share.
Two denominators, never one
Solicitation-side notices and award notices are held in different tables and answer different questions: one is work you can still bid, the other is work already placed. A single coverage number spanning both would average a live-pipeline figure against a history figure and mean nothing. They are counted separately here and no combined rate is shown.
Participation and concentration by certification
One lane per certification. The headline is the share of firms holding that certification today that captured at least one award coded to that lane in the trailing window. Beneath it, the same lane measured by money: how much of the obligated dollars the largest awardees hold. Counts are always shown beside the rate.
Loading participation and concentration...
What the concentration numbers say
On a dollar-weighted basis, no lane here is dominated by a handful of firms. Every certification reads as unconcentrated, and by a wide margin: the most clustered lane on the page sits far below the threshold where a market is even called moderately concentrated. Whatever is limiting these certifications, it is not a few incumbents holding the money.
Participation and concentration answer different questions
Participation counts firms. Concentration counts money. A lane can have very few winners and still spread its dollars evenly among them, and a lane with many winners can still funnel most of its dollars to a short list. Read the two together and neither one alone: a low participation rate is a statement about how many certified firms are getting in, not a statement about whether the work is being captured by a clique.
Why there is no market total
No total row is shown, by design. A single firm can hold several of these certifications at once, so the lanes overlap. Adding them together would count the same firms more than once and produce a denominator that does not exist in the market.
Reading the 8(a) rate
8(a) is a nine-year fixed-term program. Firms that graduate leave the certified population, so they drop out of both the numerator and the denominator of the same window. The remaining cohort is, by construction, currently enrolled and actively developing business, which is why the 8(a) participation rate sits structurally far above the open-ended certifications. It is not evidence that 8(a) firms are better at capture, and the lanes should not be ranked against each other on this number alone.
How this is computed
Glass-box by default. If a number on this page cannot be traced to a count, it does not belong here.
This is not a win rate. There is no bid or offer data anywhere in the corpus. Nothing in SAM.gov or the award archive records who competed and lost, so a wins-per-bid conversion is not computable and no figure on this page should be read as one.
Denominator: firms currently holding the certification, counted from the entity registration corpus as of the supply date shown above.
Numerator: distinct certified firms with at least one real award coded to that set-aside lane inside the trailing window. Awards are matched to the firm by UEI and the firm must actually hold the certification for the award to count.
Real awards only. IDV parent vehicles are excluded, so a firm is counted for work actually placed on contract, not for holding a vehicle that may never be ordered against.
Momentum is the same calculation over the trailing twelve months. It is a subset of the thirty-six month figure and is read as the current pace, not as an independent rate.
Concentration is dollar-weighted, computed over the same thirty-six month window and the same cert-holding population as participation. Top-5 and top-10 shares are those firms' portion of every obligated dollar in the lane. HHI is the sum of each firm's squared percent share, on the standard zero to ten thousand scale.
The HHI bands are a borrowed yardstick, not a legal finding. Under 1500 is unconcentrated, 1500 to 2500 is moderately concentrated, above 2500 is highly concentrated. Those cut points come from the DOJ and FTC merger guidelines and are used here only because they are familiar. A set-aside lane is not an antitrust market, and nothing on this page is a claim about anyone's conduct.
Concentration excludes more rows than participation does. Same IDV exclusion, plus any award row with a null or non-positive obligated dollar value, since a dollar share cannot be computed from an absent dollar. That is why the awardee firm count on the concentration block runs slightly below the active awardee count on the participation block above it: a firm can hold a real award that carries no positive obligation on the record. The gap is a handful of firms per lane, and it is a reporting artifact, not a disagreement between the two figures.
Known approximation. Certification status is read as of today rather than as of the award date, because point-in-time certification dates are not yet loaded. A firm that certified last month sits in the denominator against a three-year award window, which biases participation rates downward. Treat these as a floor, pending point-in-time supply.
Bounds check. The views are bounded by construction, and the page independently flags any share that returns at or above 100 percent, or any HHI outside zero to ten thousand, rather than rendering it as a normal value.
Not yet available: certification churn. The fourth market health signal is the rate at which firms enter and exit a certification, which is what separates a lane that is genuinely growing from one that is only recycling its membership. It needs certification entrance and exit dates, which are not loaded yet. It is left off the page rather than approximated.
LLC · SAM.GOV EXPLOITATION · 2026