Methodology & Sources

The Succession Index is a reference dataset about where small businesses are, which industries concentrate where, and how the population base under them is changing. It is assembled entirely from public federal and state records. This page describes how it is built and, just as importantly, what it deliberately does not do.

Sources

Every page lists the specific sources behind its figures, with the date the data was retrieved. Where a source is a U.S. government work, the underlying data is in the public domain.

The coverage gate

Pages in this index exist in one of two states. A page is published only once a sourced dataset has been loaded against it. Until then it is marked coverage pending: it renders with a noindex directive, it is excluded from the sitemap, and it shows a notice instead of numbers.

This is a deliberate constraint. It would be trivial to fill every page with modeled estimates and let search engines sort it out. We do not, because a statistic that cannot be traced to a source is not reference data. It is decoration. If a number appears on this site, a public record backs it.

County versus city figures

The Census Bureau publishes establishment counts by county and by ZIP code, but never by municipality. City pages therefore report the ZIP codes the Postal Service assigns to that city, using the assignment that ships inside ZIP Code Business Patterns rather than a mapping of our own.

A city’s ZIP codes are not its municipal boundary. ZIP areas can extend well past city limits, and neighbouring places often share them. So a city page says “establishments across the ZIP codes assigned to this city” and never “businesses inside the city limits”, and it lists the ZIP codes it summed. We do not apportion county totals down to cities by population share and present the result as a city count.

ZIP Code Business Patterns carries establishment counts and the size distribution but no employment or payroll, so those two figures appear only at county level.

It also suppresses more than the county file does. Census withholds cells small enough to identify an individual business, and at ZIP level many more cells are that small. For a common industry this barely registers: restaurants reach 99% of the county-level statewide total. For a thin one it is severe: food manufacturing reaches 16%. City pages therefore exist only for industries where enough survived suppression to publish, which is why more than half of the city pages in this index are still held back, and why a city ranking covers fewer places than the county ranking for the same industry.

Two sources, two purposes

County Business Patterns is the structural backbone of this index: it has the fullest coverage and the only establishment size distribution. It also runs about two years behind its reference year, which is a real limitation for anyone trying to read a market today.

The Bureau of Labor Statistics publishes quarterly with roughly a five-month lag, so county industry pages carry a second, more recent reading beside the Census figures. It is shown alongside rather than instead, because the two count different things. Census counts establishments with paid employees during the pay period including March 12; BLS counts worksites covered by unemployment insurance at any point in the quarter. Either figure can be the larger one depending on the industry, and BLS has no size distribution.

They are therefore not directly comparable, and we do not blend them, difference them, or present one as a correction to the other. Each is labeled with its source and period so you can see which question it answers.

Derived comparisons

Counts on their own do not tell you much. Pages therefore also carry ratios and ranks: how much provision a market has relative to its size, position against every other place in the state, the local size distribution against the statewide one, and the change in establishments over the years loaded.

Counties and cities are measured differently, and the reason matters. A county page reports establishments per 10,000 residents, because county establishments and county residents describe the same area. A city page cannot use that ratio. Its establishment counts cover the ZIP codes assigned to the city, not the city limits, so dividing by municipal population compares two different areas, and it does so in a direction that depends on the city. Los Angeles keeps all 3.9 million residents while every business in a ZIP the Postal Service calls Van Nuys or Encino is counted elsewhere, so it reads far emptier than it is. A small town keeps its own few thousand residents while collecting every business in the rural ZIP around it, so it reads far busier than it is.

City pages therefore report the industry’s share of all establishments in the same ZIP codes, indexed against that industry’s share statewide. Numerator and denominator then cover identical ground, so a city that loses establishments to a separately named ZIP loses its denominator along with them and the ratio holds. This is the same construction the Bureau of Labor Statistics publishes as a location quotient. Cities with fewer than 500 establishments in total are left unranked, because on a base that small the share moves on rounding. County and city figures are not comparable to each other and are never mixed in one ranking.

These are arithmetic over the sourced counts, not new data and not estimates. No figure is modeled, smoothed or filled in. If the underlying count is missing, the comparison is left blank rather than guessed.

Two judgment calls are worth stating. Median and percentile calculations exclude places under 2,000 residents, because a town of 400 with one plumber produces a per-capita figure that is arithmetically true and analytically useless, and letting those into the median distorts the benchmark for everywhere else. And the plain-language readings attached to those ratios (“thinly served”, “densely served”) are fixed thresholds applied consistently, not a judgment about whether a market is worth entering.

Location quotients are the exception to all of this: they are computed and published by the Bureau of Labor Statistics and read straight from the source.

Disclosure suppression

Federal law requires the Census Bureau to withhold statistics that would identify an individual establishment. In small counties, employment and payroll for a given industry are frequently suppressed for this reason. Where that happens we display “not disclosed” rather than zero, because zero would be false.

What this index does not publish

  • Assessments, predictions or scores about any individual named business. Nothing here says a particular company is struggling, closing or for sale.
  • Personal information about business owners. No ages, no family or estate events, no inferred personal circumstances.
  • Modeled or generated statistics of any kind presented as observed fact.

The public index reports aggregate statistics only. Business-level signal data is a separate, subscriber-only product governed by its own terms.

Corrections

If a figure here is wrong, or a page misrepresents a place or a business, write to corrections@bizsleuth.com. Include the URL. We will correct the record or remove the page, and we do not require a legal demand to do either.

Update cadence

Population estimates are released annually and refreshed within the index shortly after publication. County Business Patterns runs roughly two years behind the reference year; each page states the reference year of the figures it shows. Pages revalidate every 24 hours.

← Browse the California index