Our Methodology

Rigorous, transparent, and reproducible. Here's exactly how we build every report.

Industry Classification

Every report is organized around NAICS (North American Industry Classification System) codes, the standard used by all federal statistical agencies. We support 4-, 5-, and 6-digit codes for varying levels of industry specificity.

Data Collection

We collect data exclusively from authoritative government and regulatory sources. See our primer on market research methodology for the broader context on how rigorous research is structured:

  • Census Bureau County Business Patterns (CBP) — establishment counts, employment, payroll by NAICS and geography
  • Census Bureau Statistics of US Businesses (SUSB) — firm size distributions, receipts/revenue data
  • Bureau of Labor Statistics (QCEW) — quarterly wages, employment trends, industry growth rates
  • Federal Reserve Economic Data (FRED) — GDP, inflation, interest rates, macroeconomic indicators
  • SEC EDGAR — public company filings, financial data, competitive intelligence
  • Census Foreign Trade — import/export statistics by industry

Market Sizing

Market size is derived from Census SUSB revenue/receipts data when available. When direct revenue data is unavailable, we use total industry payroll as a proxy metric, clearly labeled as such. All figures cite their source year.

Forecasting

Five-year industry forecasts are built using compound annual growth rate (CAGR) projections derived from historical data trends, adjusted for the industry life cycle stage. Forecast charts include upper and lower confidence bands to represent uncertainty ranges. We do not present projections as certainties.

Analysis & Validation

Raw data is structured into narrative analysis. Every section undergoes a validation pass that cross-checks numerical claims against the source data and auto-corrects discrepancies. Pre-computed derived metrics (CAGRs, ratios, year-over-year changes) ensure accuracy and prevent errors.

Data Currency

Data freshness varies by source. Census CBP typically has a ~2 year lag, BLS data lags by 6-9 months, and FRED provides near real-time macroeconomic data. Every report includes a data freshness section showing actual year ranges per source so readers know exactly how current the data is.

How We Use AI

AI makes us faster. It does not make the numbers.

Every figure in our reports and Data Packs is official U.S. government data: public, free, and checkable at the source (Census Bureau, Bureau of Labor Statistics, Federal Reserve). Where an insights page cites an industry source instead, it names that source. We use AI to turn the figures into readable analysis, which is what lets us be fast and affordable. Anyone can pull the same data and check our work.

What AI writes

  • The analysis and commentary in reports
  • The text of insights pages
  • Overviews on industry hub pages
  • Scope notes on industry pages

What AI does not touch

  • The federal figures and their citations
  • Growth rates, ratios and projections, computed in code
  • Charts, drawn directly from the data
  • Data Packs: no AI-written text at all

How it is checked

  • Every numerical claim in a report is checked against the source data by an automated validation pass
  • Every page is reviewed by us before it goes live

DATS Digital LLC publishes vantainsights.com and is responsible for everything on it. Found an error? Write to support@vantainsights.com.

Common questions about our methodology, validation, and data freshness are answered in the data sources and methodology section of our FAQ.

The published federal figures behind every report are also queryable directly: our MCP connector serves them to Claude and other AI assistants, free and without an account, each figure carrying its source agency and year.