Business Buyer School

Business Buyer School Research · Industry Atlas · Methodology

Industry Atlas methodology

Small Business Acquisition Industry Atlas · Version atlas-2026.2 · Published October 6, 2026. How five studies are joined, what is left blank and why, and the exact rule behind every buyer lens.

By , Founder, Business Buyer School · Published · Last updated · Version atlas-2026.2 · What this does not prove

What the Atlas is

A side by side view of five existing Business Buyer School studies for the same 152 industries. It reads their published files and recomputes no score, rank or share. It has no combined score and no weights.

Sources and versions

Source files read (rows as published)
PartStudyFileRowsUnderlying data
supplyAcquisition Supply Study (Report 4)buyer-school-acquisition-supply-2026.csv167Census SUSB 2022 (NAICS 2017 codes)
recessionRecession Resistance Study (Report 1)buyer-school-recession-resistance-2026.csv167Census BDS 2007 to 2023 and BLS QCEW monthly 2019 to 2021
durabilityEstablishment Durability Study (Report 2)buyer-school-establishment-durability-2026.csv167Census BDS 2014 to 2023
owner ageOwner Age Study (Report 3)buyer-school-owner-age-succession-2026.csv265ACS 2019 to 2023 5-year PUMS (NAICS 2022 groups)
owner sampleAcquisition Supply Study (Report 4) derived filesupply_scored.csv152Report 4 owner age overlay with sample records
aiAI Exposure and AI Opportunity Studies (edition 3.0)AI-INDUSTRY-MASTER.csv211BLS OEWS May 2025 and published occupation level AI exposure measures (NAICS 2022)

The SHA256 of every source file is in the version manifest.

The spine: 152 industries

The supply, recession resistance and durability studies rank the same 152 four digit NAICS 2017 industries (118 High and 34 Medium confidence); 15 Low confidence industries are left unranked in all three and are not in the Atlas. The build stops if the three ranked sets differ, if a confidence tier differs between them, or if the supply study's own copies of the recession and durability ranks disagree with those studies.

Join rules

  1. Recession resistance and durability: exact naics_4. 152 of 152 joined.
  2. Owner age: only the values the supply study already attached through its crosswalk: the NAICS 2017 code must map one to one to NAICS 2022, and the ACS industry groups must cover that code completely. No new mapping is made. 123 of 152 have a value (68 from an exact ACS group, 55 from a pooled group shared with other industries); 29 are not mappable and stay blank.
  3. Owner age confidence: the weakest Report 3 tier among the ACS groups behind the value (High, Medium, Low; Not rated where a group is outside the Report 3 universe). Counts: High 52, Medium 23, Low 36, Not rated 12, not mapped 29. The sample size (unweighted PUMS records) is shown in each record.
  4. AI exposure and operating leverage: joined only where the AI study row has the same four digit code and its NAICS 2017 parent is exactly that code, and only where the AI study ranked the industry (its 85% coverage gate and acquisition relevant universe). 60 ranked, 23 matched but not ranked by the study, 69 not joined.
  5. Never imputed: a blank stays blank. No value is borrowed from a parent sector, a neighbor code or another vintage.

NAICS vintages

Census BDS and SUSB 2022 files use NAICS 2017 codes; ACS PUMS 2019 to 2023 and BLS OEWS May 2025 use NAICS 2022. Codes that changed in 2022 are where joins fail. In the AI study, five retail codes combine a 2017 store code with nonstore retail (4541, 4543), so they are not exact matches and are left blank. The AI study also has one to one recodes (the same industry under a new 2022 code). They are not joined under the exact rule and are listed here for review:

Grouped industries

58 industries carry a grouping note: the official code holds several kinds of business that the source cannot separate, or the owner age value comes from an ACS group shared with other industries.

Quarters and halves

For the three studies with 152 ranks: top quarter = ranks 1 to 38, bottom quarter = 115 to 152, top half = 1 to 76. For AI exposure, quarters are of the 81 ranked industries: 1 to 20, 21 to 40, 41 to 60, 61 to 81 (quarter = rank times 4 divided by the number ranked, rounded up).

Buyer lens rules

Each lens is a filter, applied once in the build and stored in the dataset
LensRuleIndustries
L1 More businesses to investigateAcquisition supply rank 1 to 38 (the top quarter of 152).38
L2 Supply and resilienceAcquisition supply rank 1 to 76 AND recession resistance rank 1 to 76 (top half of 152 on both).42
L3 Supply and durabilityAcquisition supply rank 1 to 76 AND durability rank 1 to 76 (top half of 152 on both).23
L4 Succession opportunityOwner age mapped, owner age confidence High or Medium, AND share of owners 65 or older at or above 12.20%, the median of the 75 industries that meet the first two conditions. Owner age counts people, not firms, and age is not intent to sell.38
L5 Lower AI exposureRanked by the AI Exposure Study (coverage gate met) AND AI exposure rank 61 to 81 (the least exposed quarter of 81).20
L6 Higher AI operating leverage opportunityRanked by the AI Opportunity Study AND quadrant “AI Leverage Opportunity” AND quadrant stable (same quadrant in at least 6 of 8 specifications). This is the study's staffing mix classification, not a finding that these industries will gain from AI.13

Where the measures disagree

An industry qualifies when one of its supply, recession resistance and durability ranks is in the top quarter and another is in the bottom quarter. 52 qualify. The article shows the 12 with the widest spread (highest minus lowest of the three ranks), ties broken by NAICS code.

What is not in the Atlas

Quality checks

Before the files are written the build checks: one row per code in every source; the same ranked set and confidence in the three spine studies; every owner age group found in the Report 3 file; the Report 4 derived owner share equal to its published value; no Atlas field outside the sources; and every lens count reproduced from the CSV. Fifteen rows were also checked by hand against the source files.

State supply: establishments by state

Source: U.S. Census Bureau, County Business Patterns 2022, state file and U.S. file (state record layout 2020 to 2023). Unit: establishments, not firms. County Business Patterns counts establishments: each physical location with paid employees. The Atlas supply study counts firms, and one firm can own several establishments. This page reports establishments only and never adds the two together.

  1. Codes: County Business Patterns 2022 uses NAICS 2017 codes (its file carries 4471 and 4541, which NAICS 2022 replaced), the same vintage as the supply study. All 152 Atlas codes are in the U.S. file; the build stops otherwise.
  2. Rows: the all legal forms row (LFO dash) of each exact four digit code. 7,670 state rows joined of 7,752 possible.
  3. Value: establishments with 5 to 99 employees = n5_9 + n10_19 + n20_49 + n50_99. The 5 to 99 band matches the supply study's band, but the supply study counts firms and this counts establishments.
  4. Blanks: a cell is blank when any of the four size classes carries the Census flag N (not available or not comparable: 2,698 cells) or when Census published no row (82 cells). Size classes flagged N: 5 to 9 467, 10 to 19 652, 20 to 49 982, 50 to 99 2,573. Nothing is estimated, borrowed from a neighbor or set to zero. 4,972 of 7,752 cells are published.
  5. Share of national: state value divided by the U.S. file value for the same code and band. Per 100,000 residents: state value divided by the Census Vintage 2023 population estimate for July 1, 2022, joined on state FIPS.

Reconciliation. For the 6 industries with no blank state (5617, 6211, 7139, 7225, 4451, 7211), the 51 state counts add up exactly to the U.S. count. Total establishments (all sizes) add up exactly for 129 of 152 industries; in the other 23, Census published no row for one or more states and the published states fall short of the U.S. total by 1 to 8: 7212 (7,968 of 7,969), 3327 (22,079 of 22,081), 4412 (14,049 of 14,050), 3152 (3,822 of 3,824), 2379 (4,672 of 4,673), 4245 (6,211 of 6,212), 3323 (14,345 of 14,347), 3335 (5,895 of 5,898), 3328 (5,405 of 5,407), 3372 (3,685 of 3,687), 3279 (3,305 of 3,307), 5622 (3,212 of 3,214), 3141 (1,366 of 1,371), 3211 (3,064 of 3,067), 3256 (2,555 of 2,558), 3332 (2,906 of 2,910), 3333 (1,725 of 1,728), 3259 (2,261 of 2,263), 3272 (1,566 of 1,569), 3111 (1,864 of 1,869), 3255 (1,678 of 1,681), 3353 (1,832 of 1,840), 3321 (1,974 of 1,981).

Version history

Atlas editions
VersionDateChange
atlas-2026.12026-10-06First edition.
atlas-2026.2this editionAdds establishments with 5 to 99 employees by state from County Business Patterns 2022 for all 152 industries: a state view page, state CSV and JSON, and a state filter in the buy-box matcher. Industry values of atlas-2026.1 are unchanged.

Suggested citation

Cite as: Business Buyer School, Small Business Acquisition Industry Atlas (methodology), https://businessbuyerschool.com/research/small-business-industry-atlas/methodology/. Version atlas-2026.2. Published October 6, 2026.

Reusing this research

You may reuse reasonable excerpts, individual charts, individual tables, selected findings and quotations for media coverage, with credit to Business Buyer School, the name of the study, and a link to the report or dataset page where practical, as long as the meaning is not materially changed.

Separate written permission is needed to republish a dataset wholesale, redistribute it in bulk, resell it, republish substantially the entire report, remove methodology or source notes from a reproduced table, present the analysis as your own original research, or alter a chart or table in a way that changes its conclusion.

The underlying government source data remains subject to its own terms. This policy covers only Business Buyer School's own analysis.

Permission requests, corrections and methodology questions: dave@businessbuyerschool.com.

Take this into your search

The Atlas narrows the industries to look at. These parts of the school turn that into a search and a first screen of a real business.