Original analysis, with the method attached.

We turn public datasets into clear comparisons and design experiments for questions the existing data cannot answer. Published reports include sources, limits, and update dates; planned work is labeled before results exist.

What qualifies as BizFacts research

A published analysis must identify the source vintage, unit of analysis, formulas, exclusions, and revision policy. An experiment must freeze its sample and analysis rules before results are interpreted. A plan is labeled as a plan; it never borrows the authority of findings that do not yet exist.

WorkStatusEvidence available
July 2026 business applications reportPublishedCensus monthly file, calculations, downloadable extract, and limitations.
Local-business AI visibility pilotPlannedFixed design and blank observation, mention, and citation schemas; no results.

The shared rules for source vintages, transformations, comparisons, and experiments are in the BizFacts data and methodology policy.

Data report

Business Applications Hit a Record in July 2026—Driven by Retail

U.S. business applications reached a monthly record in July 2026, but retail drove the surge while measures tied more closely to employer formation changed little.

published · 8 min read
Research plan

Can AI Assistants Find Your Business? Study Plan

How BizFacts plans to measure local-business recommendations, citations, overlap, and repeatability across AI-assisted search products.

planned · 10 min read