A free state-by-state breakdown of firm counts, employment, and wages for forty small-business-typical U.S. industries. The intended audience is operators sizing a market before opening, journalists writing about local economies, researchers studying small-business density, and anyone curious how their industry's footprint differs across states.
Every number on this site comes from the BLS Quarterly Census of Employment and Wages (QCEW), 2024 annual single-file release. QCEW is built from state unemployment-insurance tax records — every employer that pays UI taxes shows up — and covers roughly 95% of U.S. wage-and-salary jobs. Private ownership at the state level, mostly 6-digit NAICS codes with three specialty construction trades (concrete contractors, plumbing/HVAC, roofing) shown at 5-digit aggregation because the 6-digit detail isn't available.
40 industries × 52 states and DC (Puerto Rico where available) = 2,079 populated cells. Total establishments in current snapshot: 2,646,194. For each industry × state cell you get establishments, average annual employment, total wages, and average weekly wage; for each industry you also get a location-quotient (LQ) comparison that shows where the industry concentrates relative to its national footprint.
The base count here is establishments, not firms. An establishment is a single physical location with payroll — one dental office, one restaurant, one barbershop. A multi-location firm (a regional dental group, a restaurant chain) shows up as one establishment per location. That's the right unit for "how many storefronts of this kind are operating," which is what the site is built around; it's the wrong unit for "how many distinct companies."
For an industry in a state, the LQ is the industry's share of state employment divided by the industry's share of national employment. An LQ of 1.0 means the state has its proportional share; 2.0 means twice the national share (a concentration); 0.5 means half. LQ is useful for spotting where an industry is structurally over- or under-represented — drinking places in Wisconsin (LQ 4.0, beer-hall culture), religious organizations in Oregon (LQ 6.8, much higher concentration than the Bible-Belt instinct predicts), parking lots in Hawaii (LQ 4.0, land scarcity) — relative to a generic baseline.
SMB Density is built by Claude (Anthropic's AI model) using public BLS QCEW data. Source code and methodology are open; corrections welcome. See the byclaude lab for context on what's being shipped and why.