# State of Main Street Ownership, Q4 2026: methodology

DealFlow OS · Data as of November 6, 2026 · Report: https://dealflow-os.com/reports/main-street-q4-2026

## Sources and as-of dates

| Source | What it gives | As of |
|---|---|---|
| Overture Maps Places, release `2026-09-23.1` | The business universe: name, category, address, city, state, phone, website, brand, operating status | Release of 2026-09-23 |
| DealFlow OS website crawl | Emails and owner names found on the business's own website | Oct 31 – Nov 6, 2026; ~1.05M sites crawled so far |
| SBA PPP FOIA loan data | Jobs reported; owner names (sole proprietors) | File of 2024-09-30; loans approved 2020-04-03 to 2021-06-30 |
| CMS NPPES (NPI registry), organizations only | Authorized official (owner name) | September 2026 dissemination |
| FMCSA carrier census | Trucking and transport registrations | Records through 2026-08-28 |

Business, website, email and owner counts come from DealFlow OS's nightly refresh of November 5, 2026.

## Universe and exclusions

Applied when businesses are loaded:
- **Dropped:** permanently closed places.
- **Dropped:** non-business categories: public schools, colleges, government, places of worship, geographic/cultural entities.
- **Dropped:** a blocklist of large company-owned chains. Franchises are kept and flagged.
- **Dropped:** places outside the 50 states + DC.

Applied in the counts:
- **Excluded:** places marked permanently closed or removed from Overture.
- **Excluded:** unreachable listings (no phone and no website).
- **Excluded:** individuals (a person's listing, not a business).
- **Excluded:** businesses that hit an industry's negative terms. For example, a "Dental" category record whose name says "dental lab" is not counted as a dental practice.
- **Included:** temporarily closed businesses.

**Result:** 13,625,907 businesses in 51 jurisdictions (50 states + DC). 11,453,743 of them (84.1%) map to a named industry. State figures on the report's state pages count these classified businesses.

## Dedupe

- **One row per Overture place.** A multi-location business counts once per location, which is how Overture lists it.
- **No merge across listings:** locations that share a website are counted separately.
- **Matching other sources** (PPP / NPPES / FMCSA → Overture) is tiered. Each source record matches at most one business:
  1. Exact phone.
  2. Exact normalized name + same ZIP.
  3. Fuzzy name (Jaro-Winkler ≥ 0.92) + same ZIP.
  4. Fuzzy name (≥ 0.97) + same city/state.
- **Owner names** use only tiers 1–3. Tier 4 had too many false positives.

## Industries

- **Taxonomy:** each Overture category belongs to one DealFlow OS industry. There are 431 industries; 390 have businesses.
- **Overlaps:** a category claimed by more than one industry goes to the most specific one.
- **Umbrella industries excluded** (Construction, Home Services, Retail, Financial Services). Each is a union of other industries and would double count.
- **Not counted in any industry:** 274 catch-all categories, 15.9% of businesses. The largest are uncategorized (628,641), contractor (169,230), professional_service (148,706), doctors_office (144,581), health_care (135,170) and financial_service (101,197). These businesses are real but can't be assigned to an industry from the category alone.

## Files

All files are aggregates by industry × state. None contains a business name, address, email or owner name.

| File | Contents |
|---|---|
| `counts_industry_state.csv` | Industry × state: business count and counts with website, phone, email, owner name (any source), and owner name from public records (PPP / NPPES) |
| `fragmentation.csv` | Industry × state: count, branded and franchise locations, distinct brands, top-4 brand share, HHI, independent share |
| `metros.csv` | Top 25 cities by count for the 20 largest industries |
| `size_bands.csv` | Industry × state: count by PPP jobs reported (0–1, 2–4, 5–9, 10–19, 20–49, 50–99, 100+), median, coverage %. Band counts and the median are left blank where fewer than 5 PPP-matched businesses report |

**Phone counts** come from the same Overture release with the same exclusions applied, capped at the business count. They match the business counts within 0.14% per state × category.

## Known gaps and biases

1. **No founding year.** None of the sources publishes a business founding date, so this edition reports no business-age statistics.
2. **Email and owner coverage reflect the crawl's order, not the industry.**
   - Emails come only from websites crawled so far. The crawl started with HVAC, plumbing, electrical, roofing, pest control, landscaping, septic, auto, insurance and dental, plus waste and grease-trap services.
   - Owner names from public records come from NPPES (health care) and PPP (sole proprietors).
   - Industries the crawl hasn't reached yet (restaurants, salons, real estate, attorneys…) show near-zero email rates. That's a coverage gap, not a property of the industry. Don't compare email or owner rates across industries.
3. **Fragmentation understates chain concentration.**
   - Large company-owned chains were removed at load; franchises are kept.
   - Brand coverage in Overture is partial, so unbranded locations count as independent.
   - Read top-4 share and HHI as concentration among the businesses DealFlow OS tracks, not the whole market.
4. **No population data.** There are no per-resident figures.
5. **Cities, not metros.** `metros.csv` and the state pages rank cities by the city name as Overture lists it, not metropolitan areas. Suburbs aren't rolled into their metro.
6. **Size bands are PPP-era and self-selected.** These are 2020–21 jobs reported by businesses that took a PPP loan, matched to 2026 listings. Very small and very large employers are under-represented, and headcounts may have changed since.
7. **Overture category quality.** Categories are assigned by Overture's sources and are sometimes wrong or too broad. The catch-all categories above are the clearest case.
8. **Temporarily closed businesses are counted** and not flagged in these files.

These are counts of operating businesses, not businesses for sale, and say nothing about any owner's intent to sell or any buyer's returns.