American Staffing Association - Staffing Statistics by State

The American Staffing Association is the trade body of the US staffing industry, and its Staffing Statistics by State series is the industry's own accounting of itself, one fact sheet per state. Every sheet answers the same six questions: average temporary help workers each week, annual staffing employment, annual payroll, estimated number of offices, the state's top temporary-help occupation groups, and national benchmarks to read them against.

As published, the record is fifty-one separate one-page documents formatted for human reading. As delivered here, it is one comparable panel - fifty-one rows on a shared six-column shape, keyed by state, with reference years carried as data. Within Datadory's diversified-reits shelf this is the tenant-economy layer for the flexible-workforce sector: who works an assignment, who pays them, and from how many offices. It scores 6/10 on our quality rubric against a catalog mean of 7.81 - narrow by design, and useful because of it. Get a sample of this dataset cut to your states before building anything on top of it.

Each row is one state. The California row, exactly as the fields arrive:

state                       : California
avg_temp_workers_weekly     : 334100
annual_staffing_employment  : 1723800
annual_payroll_usd          : 50600000000   # 2023 reference year
estimated_offices           : 3940          # 2023 reference year
top_occupation_groups       : <top five temporary-help occupation groups, ranked>
national_benchmarks         : <US staffing industry comparison figures>

Read it as three layers of one state's staffing economy. The flow layer - 334,100 people on assignment in a typical week - is the number that moves first when hiring plans change. The stock layer - 1,723,800 employed across the year and $50.6 billion in payroll - sizes the industry in people and dollars. The footprint layer - 3,940 offices - maps the physical distribution network of the sector, which is the part a landlord reads differently than an economist.

Swap the state and the identical shape returns for every other jurisdiction, all fifty states plus the District of Columbia. Name your states when you request a sample and Datadory returns exactly those rows, benchmarks attached.

Six documented measures, verified against the published sheets during the August 2026 research pass - nothing inferred. Two identify the scale of the weekly and annual workforce (avg_temp_workers_weekly, annual_staffing_employment), two price and locate it (annual_payroll_usd, estimated_offices), and two give the row its texture and context (top_occupation_groups, national_benchmarks).

The detail worth noticing is the reference-year stamps: headline employment runs on the 2025 edition while payroll and office counts cite 2023. Delivered rows keep those stamps visible so nobody accidentally divides a 2023 payroll by a 2025 headcount and calls the result productivity.

State identifier join keys, computed ratios and cross-edition alignment fold under additional fields on request - built on the confirmed core, confirmed in full alongside your sample.

Geography - all 50 US states plus the District of Columbia, one row each, with no holes in the grid. The panel is deliberately a state-level instrument; it is the grain at which policy, tax and site-selection conversations actually happen.

Temporal - the 2025 edition of the series, with component vintages disclosed per measure: headline weekly and annual employment on the 2025 edition, payroll and office counts carrying 2023 reference years. Mixed-vintage panels punish silent aggregation, so the stamps ship on the rows.

Granularity - six measures per state plus national benchmarks for scale. Occupation detail arrives as the ranked top-five temporary-help groups per state rather than a full taxonomy, which is enough to tell an industrial-heavy market from an office-clerical one at a glance.

API, files, or your warehouse. Daily, weekly, or hourly.

A fifty-one-row panel loads anywhere in seconds, which is precisely why it should arrive shaped like your other tables rather than like a folder of formatted pages. Deliveries come typed to the field dictionary above, keyed to state identifiers so the panel joins against your property, market or vendor rosters on the first query, and refreshed on whatever cadence you set - load it once, or keep a standing feed so new editions diff cleanly onto the same keys.

• Office-demand signal by state - staffing office counts and employment size tomorrow's demand for the shallow-floor suites staffing firms occupy, before leases are signed.
• Leading-indicator tenant screens - weekly temp headcount by state turns early; a rollover flags local-service stress ahead of earnings calls and rent rolls.
• Labor-market sizing for site selection - rank states on realized flexible-labor capacity and occupation mix rather than population.
• Workforce-services market sizing - annual payroll by state prices the addressable market for vendors selling into the staffing industry.
• National benchmarking - benchmarks shipped beside each state row turn every figure into a spread against the country.

Investors and quants treat state-level temp employment as the leading indicator it is - the labor series that turns before the headline ones. Market researchers and consultants get fifty state markets sized in one pull instead of fifty separate document hunts. Sales and growth teams get the account-density map: where staffing offices cluster, buyers cluster. Data scientists and ML engineers get a small, definitionally stable panel that joins on FIPS keys without cleaning. Journalists, academics and students get the industry's own state accounting, citable to a named publisher. Persona workflows sit at investors & quants x diversified REITs and market researchers x diversified REITs.

Against the rest of the diversified-reits shelf: the LEHD Quarterly Workforce Indicators - NAICS 56 observes the same administrative-and-support workforce at quarterly frequency with hires, separations and earnings - far deeper on dynamics, silent on industry payroll and office footprint. The BLS NAICS 56 employment series reports the monthly level of sector employment, not the industry's own payroll and office accounting. Business Dynamics Statistics counts firm births and deaths economy-wide. Only this record answers how large is the staffing industry in this state, in dollars and offices, and what kind of work does it place - the association's own testimony about its own market.

Four honest caveats. First, mixed reference years: payroll and office counts cite 2023 while employment runs on the 2025 edition - align vintages before computing any ratio, and the delivered rows carry the stamps to make that easy. Second, grain: this is a state-level instrument, so metro or county questions want the quarterly workforce indicators or the BLS series on the same shelf paired alongside. Third, methodology opacity: these are the industry association's published estimates, and the derivation behind each state figure is not documented on the face of the sheet - treat them as industry accounting, not statistical-agency tabulation, and cite accordingly. Fourth, size: fifty-one rows by six measures is a benchmarking layer, not a modeling corpus - its job is to anchor other panels, not replace them. None of these bite unexpectedly; they ship flagged against the analysis you plan to run.

Because the raw artifact is fifty-one separately published one-page documents, each formatted for human reading, each mixing reference years inside a single page - and most questions want one keyed panel and three columns of it. Datadory normalizes the fifty-one into one comparable shape, keeps reference-year stamps as data, adds state identifier joins on request, and ships the result typed for warehouse loading so the next edition diffs cleanly onto the last. Browse the rest of the industry on the diversified REITs data hub, the best diversified-reits datasets ranking, or the full catalog.