Child Care Center Ownership Data: A Site-to-Operator Protocol

Executive Summary

A licensed child care record is not a company. In child care center ownership data, the same row can be a single-site independent, one site in a multi-state chain, a franchised location whose operator is not the brand owner, or a family home that sits outside a typical acquisition mandate. This report sets out a state-source and unit-of-analysis protocol for private equity (PE) origination teams, using California and Texas as worked examples. It does not rank chains and does not contain a target list.

The two states publish different things. California's licensing data hub reported 44,291 licensed child care facilities with 1,089,536 licensed slots in June 2026, of which 30,842 were family child care homes (^[1]) (^[1]). California's open licensing files carry a licensee field, but its public guidance says small family homes can only be found by exact license number (^[2]). Texas's open dataset, refreshed on 31 October 2026, is keyed on an operation identifier (^[3]) (^[3]), yet its field list has no owner or legal-entity column, only an administrator name (^[4]). On ownership changes, the two states' sources point the same way: California says, per a CDSS-hosted manual with old effective dates that must be checked against current regulation text, that a license is not transferable (^[5]), and Texas revokes the permit when ownership changes (^[6]).

Denominators move the answer. For the United States, a federal survey estimated 121,000 center-based providers in 2019 (^[7]), while Census County Business Patterns counts establishments with paid employees (^[8]). One public operator reported 1,555 centers at the start of 2026 (^[9]); against those two denominators that is roughly 1.3% to 1.9% (analyst arithmetic that mixes 2019, 2023 and 2026 figures and different unit definitions, so illustrative only), and the operator itself estimated the top five providers held about 6% of capacity (^[9]). Within Texas alone, "licensed centers" can be read as 8,293 or 9,859 operations depending on whether school-age programs are included (^[10]).

The protocol therefore requires five habits: declare the universe before counting; keep site, licensee, brand, franchisor and parent as separate fields; keep proof URLs and dates for every link; report coverage as an observed minimum with an unresolved range whenever definitions or dates do not align; and test any vendor chain map against the same state file as a blind sample. Vendor chain maps are not used here as a benchmark: vendor pages are revised and rebranded, and a map address can resolve to a notice rather than a list (^[11]). Any vendor count should be treated as a candidate and blind-sampled against the state file before use. Licensed capacity is a regulatory ceiling, not enrollment, occupancy or revenue (^[12]).

Introduction and Background

Searches such as "child care license lookup" and "childcare chain ownership" return parent-facing state portals, policy analyses and vendor-authored chain maps. None answers the question an origination analyst faces on day one: what, precisely, is being counted? A state license is issued to a legal person for a regulated operation, a brand is a trade name, and a chain parent is a further layer that licensing files rarely record.

The distinction has formal support. Texas defines a permit holder as the person or entity granted the permit (^[13]) and an operation as a person or entity offering a program that may be regulated (^[13]). Government statistics make the same cut: the Census Bureau defines an establishment as a single physical location (^[14]), and a firm as one or more establishments under a single organization (^[15]). A site-level list and a company-level list of the same market therefore differ by construction, not by error.

Two further limits frame the work. First, no national child care provider database exists; one recent economics working paper states that there is no national database of child care providers and had to obtain a proprietary list (^[16]). The same paper is labelled preliminary by its publisher (^[16]). Second, public ownership evidence for private companies is thin: a Congressional Research Service (CRS) analysis of large chains says its findings should be treated as illustrative rather than comprehensive (^[17]), and it states that CRS has not independently verified the source figures (^[17]).

This report is written for PE origination and research teams. It does not rank chains, publish owner contact details or claim a national site count. It specifies how to inventory state sources, reconcile site, licensee, brand and parent, compute coverage honestly, and audit a third-party map. The publisher, Docket, offers research services and is an interested party in the broader market-mapping category; the reconciliation fields described below are the evidence standard any such work, including its own, can be held to. Portal facts reflect the dates shown beside each figure, and portals change without notice, so confirm each portal on the date of use.

Defining the Mandate and the Unit of Analysis

Five questions that decide what a record is

Before touching a state file, an origination team should answer five mandate questions in writing. Each one removes or reclassifies records.

• Center versus family home. California separates family child care homes (small homes up to 8 children, large homes up to 14) from centers (^[2]), and Health and Safety Code defines a day care center as any child day care facility other than a family day care home (^[18]). Texas draws the line at capacity: a licensed center cares for seven or more children (^[19]), while a registered home may not exceed 12 children at once (^[19]) and a listed family home is limited to three unrelated children (^[19]).
• Regulated versus exempt. Exempt classes vary by state. The federal Administration for Children and Families (ACF) says how a state defines "license-exempt" varies (^[20]), and lists public-school programs (^[20]) and faith-based or part-day programs (^[20]) among classes that can be exempt. California names four main exempt groups (^[2]); Texas names four exemption categories (^[21]).
• Site versus operating entity. One licensed site can change licensee without changing address, and one licensee can hold many sites.
• Corporate versus franchised. A franchisor, a franchisee and a brand are separate roles. Item 1 of a franchise disclosure document (FDD) must name the franchisor's parents and affiliates (^[22]) and the trade names it uses (^[22]).
• Geography and size. Size filters depend on capacity, which California defines as the maximum number of children who can receive care at one time (^[12]).

Why the denominator is a decision, not a lookup

A coverage figure is a ratio, and the denominator is a mandate choice. A team that screens for multi-site center operators in two states has a different universe from a team scanning every licensed provider. Federal survey data illustrate the spread of plausible answers. The National Survey of Early Care and Education (NSECE) counted 121,000 center-based providers in 2019 (^[7]) and built its frame from state or national administrative lists (^[7]), while 21,600 of those centers served 25 children or fewer (^[23]). Home-based providers on state lists fell 25 percent between 2012 and 2019, and the survey also reached unlisted home-based providers that no licensing roster contains (^[24]) (^[24]).

A state licensing file is the best available site-level roster, but it is a roster of regulated operations, not of the market. ACF's analysis of national licensing studies reports that more than 80 percent of licensed slots sat in centers across the study years (^[25]), so nationally a center-only mandate gives up a minority of slots but many sites. California is the exception to plan for: its family child care homes held 330,450 of 1,089,536 licensed slots in June 2026 (about 30 percent; analyst arithmetic from the CDSS figures cited in the Executive Summary), so a center-only screen there excludes roughly three in ten licensed slots.

Reading a record: the four layers

Hold four layers apart for every row: the site (location and regulator ID), the licensee (the legal person holding the license or permit), the brand (the trade name on the door), and the parent (the company controlling the licensee). State files usually give the first two, sometimes a DBA (doing-business-as) name, and almost never the fourth.

State Source Inventory: California and Texas

What the inventory records

Table 1 below summarizes the two states already checked. For each state it records the regulator, record unit, public fields, bulk access, refresh evidence, excluded classes and legal-entity evidence. The same rows form the reusable 50-state template described after the table; every other state should be labelled "to be researched" until its row is completed from its own regulator pages, never inferred from these two.

Table 1 shows that neither state hands over a chain field, and that California's open files at least name a licensee string while Texas's open dataset does not. Texas site-to-operator links must therefore come from the permit holder on the live operation record plus corporate filings, while California links can start from the licensee string but still need a corporate lookup. Both require interactive verification before publication.

The reusable 50-state template

The blank template has one row per state and these columns, each filled with a URL and an access date:

• Regulator URL and the public search URL.
• Public fields visible without login, split into site, licensee, capacity, status and DBA.
• Record unit: facility, operation, license or permit, and what the ID is called.
• Provider classes inside and outside the record (centers, school-age, family homes, group homes).
• Restricted or exempt categories, quoted from the state's own page.
• Bulk access: download link, format, and whether the file carries the same fields as the search.
• Update cadence, with both the portal stamp and any as-of date inside the file.
• Compliance-history fields, recorded only as present or absent, since they are not part of an ownership map.
• Legal-entity evidence: licensee or permit holder field, controlling-person disclosures, corporate registry link.
• State aggregate totals by type and capacity, with their as-of date, for coverage reconciliation.

A federal rule helps with the first column. Under 45 CFR 98.33, a state lead agency must run a consumer-friendly website (^[34]) with a localized list of licensed providers, searchable by zip code, differentiating licensed from license-exempt providers (^[34]). A public listing exists in principle everywhere, but its fields, bulk access and exclusions are state-specific.

California in detail

The California data hub reports facilities and licensed capacity each year by facility type beginning June 2017 (^[35]). The licensing statute requires a single childcare center license that serves infant, toddler, preschool and school-age children, with all day care centers licensed as childcare centers (^[36]). That explains why the CCLD tables show a "Single License Child Care Center" type rising from 141 facilities in June 2024 to 1,347 in June 2026 (^[1]), while the legacy "Child Care Center" type fell from 10,361 facilities in June 2021 to 8,502 in June 2026 (^[1]). A time series by type is not comparable across those years; the sum of types is. The public Facility Search page also carried a banner that 860 single-licensed centers were not yet fully reflected in the public search (^[26]), a reminder that the search portal and the aggregate tables can lag each other.

Statutory exclusions matter for scoping: California's Child Day Care Act does not apply to, among others, a State Preschool program run by a local educational agency under contract with the State Department of Education (^[37]). Large family homes are searchable by area (^[2]).

Texas in detail

Texas HHSC regulates under Human Resources Code Chapter 42 (^[29]); minimum standards chapters map to provider type, 746 for centers (^[38]) and 747 for homes (^[38]). Two features matter. First, a "licensed child care center" in the HHSC data book includes before-and-after-school and school-age programs, although the handbook says a school-age program is not a licensed child care center (^[19]). Second, the open dataset includes residential and placement-agency records, and the handbook states that CCR does not issue a permit to a foster home (^[19]), so a day-care denominator must filter by operation type.

On access, the portal front page describes its search types, but record-level fields in this report were taken from HHSC's March 2026 user guide, which says Operation Details include permit status and issue date (^[27]). The interactive search is JavaScript-driven, so record-level fields are easiest to confirm by opening the portal in a browser. Any team using Texas for a deliverable should still open sample operations interactively and record what is displayed on the date of use.

Reconciling Sites, Licensees, Brands and Parents

Normalize into separate fields

The reconciliation worksheet has one row per site and keeps these fields separate: state, regulator ID, address, public name, licensee, DBA, franchisor, corporate parent, proof URL and date, and status. The main failure modes are name matching and address matching done on a single concatenated string. A brand name is never a parent: a research team that attributes ownership by brand alone will misattribute franchised and multi-brand cases.

Parents are layered in public sources. One economics paper notes that Learning Care Group is the name of a parent company that owns 11 separate brands (^[16]), and a Colorado newsroom analysis found that a center's true owner is sometimes hidden by layers of parent companies (^[39]). Both are observations about method difficulty.

Join to corporate records

A defensible join uses three corporate lookups.

• State of formation or registration. California entity numbers are a "C" plus seven digits for corporations (^[31]) and 12 digits for LLCs and limited partnerships (^[31]). The entity address is the executive office (^[31]). Texas's Comptroller search accepts an entity name of 2 to 50 characters, among other keys (Source: mycpa.cpa.state.tx.us)), and officer and director data come from the most recent Public Information Reports (^[40]), which can lag actual changes (analyst inference).
• DBA and assumed-name filings. California requires a fictitious business name statement within 40 days of commencing business (^[41]), filed with the county clerk, and the statement generally expires after five years (^[42]). Texas entities file an assumed name certificate with the Secretary of State (Source: www.sos.state.tx.us)) for a term not to exceed ten years (Source: www.sos.state.tx.us)), and entities that file there are not required to also file with the county clerk (Source: www.sos.state.tx.us)). A sole proprietor's DBA is therefore not found in the same place as a corporation's.
• Controlling persons and licensee disclosures. Texas defines a controlling person by the ability to influence management, expenditures or policies (^[43]). California's regulations require a corporate applicant to list each person owning more than 10 percent of stock (^[5]); that source is a CDSS-hosted manual with old effective dates, so it needs checking against current regulation text before reliance.

Beneficial ownership data will not fill the gap. FinCEN removed the Corporate Transparency Act reporting requirement for US companies in 2025 (^[44]) and made that permanent in a final rule on 11 August 2026 (^[45]); even when collected, that information was confidential (^[46]).

Franchise locations and operator-reported counts

A franchised school is typically operated by an independent owner. Kiddie Academy's site states that each Academy is independently owned and operated (^[47]), as does Goddard School's (^[48]). Brand-level counts are dated claims: Kiddie Academy's locations page says 375+ centers (^[47]), while a June 2025 release gave 350 (^[49]). Goddard's October 2025 release reported more than 660 schools (^[50]) and named a franchisor entity separate from the brand (^[50]). Primrose Schools' franchise site distinguishes 520 active franchises from 558 franchised schools open at the end of 2025 (^[51]), tying its figures to an FDD dated 24 April 2026 (^[51]). The unit "franchise" is therefore not the unit "school".

FDD Item 20 is the structured source for this split. It requires a table of franchised and company-owned outlets (^[22]), state-level tables of outlet counts and ownership changes (^[22]), and it defines a transfer as acquisition of a controlling interest in a franchised outlet (^[22]). The FTC describes the FDD as 23 items (^[52]). State registries differ: California requires a franchisor offering a franchise there to file an application (^[53]), and Wisconsin's search is keyed to the legal or trade name entered (^[54]).

Public company filings give operator-reported counts that serve as blind checks. KinderCare reported 1,555 centers in its main brand and 46 schools of a second brand in 15 states at 3 January 2026 (^[9]) (^[9]), 77 of them employer-sponsored (^[9]), and also reported 1,153 before-and-after-school sites that are counted separately (^[9]). Bright Horizons reported 1,010 centers, of which 597 were in North America (including Puerto Rico) (^[55]) (^[55]); about 25 percent operate under a cost-plus model in which an employer sponsor bears the financial risk (^[55]), yet the operator retains responsibility for running the center (^[55]). Neither filing gives state-by-state counts, so they can test national or brand totals only.

Implementation Considerations and Process Changes

Build the site-to-operator graph

Treat the worksheet as a graph. Nodes are sites, licensees, brands, franchisors and parents; edges are typed links (licensed-to, trades-as, franchised-by, controlled-by) with a proof URL, an observation date and a link status. Three statuses are enough: verified (two independent public sources agree), single-source (one source), and unresolved (name or address similarity only). Unresolved clusters are reported, not forced into a parent.

Table 2 below shows the worksheet. The rows are a labelled Hypothetical Example using fictional names and an invented identifier format; they illustrate the structure, not any real operator.

Table 2 makes three points: a shared brand (rows 2 and 3) does not imply a shared licensee; the franchisor field can be filled where the parent is unresolved; and each row carries its own evidence date, so a stale link is visible. Never count a center and its licensee as two companies: distinct operating groups are the graph's connected components.

Preserve history

California states that a license is not transferable (^[5]), that a license is forfeited by operation of law on listed events (^[5]), and, per a CDSS-hosted manual with old effective dates that should be checked against current regulation text, that listed changes, including sale or transfer of a majority of stock, require a new application (^[5]). Texas states that if a change in ownership occurs the permit is revoked (^[6]), that the new owner must apply for a new permit (^[6]), and that sale of a permit holder that is a business entity counts (^[6]). HHSC announced a change-of-ownership expedited licensing process on 6 March 2026 (^[56]). In both states a closed or renamed record may be a successor to an acquired site, so the worksheet keeps closed, renamed and licensee-change history with dates and never overwrites a prior row. Whether an acquisition appears as a new facility number or a changed licensee string must be tested per state.

Compute coverage by declared universe

Coverage is unique verified in-scope sites divided by the official in-scope site count, and only when the definitions and dates align. The calculation steps are:

1. Declare the universe: state, provider classes, status (licensed only), and exclusions such as family homes.
2. Take the official count for the same classes and the same date (for example, a regulator's annual table).
3. Count unique, in-scope, verified sites in the worksheet, deduplicated by regulator ID.
4. If definitions or dates differ, report an observed minimum (verified over official) and an unresolved range (adding single-source and unresolved sites).
5. Report facility count, licensed slots and unique operating groups as three separate measures.

A Hypothetical Example: suppose a regulator's table lists 1,000 licensed centers on a date. The worksheet holds 700 sites, of which 640 are verified and 60 single-source, so 300 official sites are not yet reached. Observed minimum is 64%; counting single-source rows gives 70% as the upper bound for sites reached so far; the other 30% of the official count is unreached and unknown. Capacity coverage is computed the same way on slots. The figures are invented.

Audit a vendor map as a blind sample

A vendor's national chain map can be tested without trusting it. Draw a random sample of sites from the same state file, stratified by size or region, and check whether the vendor map includes each site and assigns the same parent. Publish error categories and unknowns:

• Missing site: licensed in the state file, absent from the map.
• Extra site: on the map, closed or not found in the state file.
• Wrong parent: brand-level attribution of a franchised or multi-brand site.
• Stale parent: an ownership label that predates a transaction.
• Unit mismatch: slots, sites and operating groups mixed in one count.
• Unknown: cannot be resolved from public sources.

With a sample of 200 sites and an observed agreement of 90%, the 95% margin is about plus or minus 4.2 percentage points (analyst arithmetic); a state-level claim needs its own sample. Vendor map URLs can redirect to rebrand notices, so any vendor map should be treated as a candidate and re-fetched on the date of use; on 2 November 2026 one vendor-authored chain-map address resolved to a rebrand notice, "A new name. Our next chapter." (^[11]), and not a chain list. This report does not reproduce vendor counts or sponsor labels, and the blind-sample test above is presented as a method, not a completed test.

Data Analysis and Evidence

California: aggregate tables versus open files

CDSS's June 2026 table reports 44,291 child care facilities and 1,089,536 slots, up from 43,569 and 1,088,012 a year earlier (^[1]). Family child care homes account for 30,842 facilities and 330,450 slots (^[1]). By analyst arithmetic, homes are 69.6% of facilities but 30.3% of slots, and the non-home types total 13,449 facilities and 759,086 slots. A CDSS fact sheet, last updated 4 February 2026, rounds this to approximately 13,600 centers (^[12]). A July 2026 year-over-year capacity PDF gives a slightly different statewide total of 44,315 facilities and 1,087,475 slots (^[57]), so two CDSS products of different dates already disagree by 24 facilities.

The open center file, retrieved on 2 November 2026, contained 19,426 rows across five statuses; 14,072 were licensed with distinct facility numbers, and their capacity summed to 778,657 (analyst row count of ^[28]). Compared with the June 2026 aggregate of 13,449 non-home facilities, that is 623 more facilities (4.6%) and 2.6% more slots, but the rows carry a file date of 25 May 2025 while the aggregate is June 2026, so the difference is not a coverage error and cannot be read as one. Record both the portal date and the in-file date.

The family home file is the sharper limit: 13,986 licensed rows against 30,842 homes in the annual table, every address field reading "Unavailable". The shortfall (45% of the annual count present) is consistent with the statutory non-disclosure for small homes, though CDSS does not say so for the file (analyst inference). The same file lists 19 licensed homes in Del Norte and Inyo counties, which conflicts with the parents guidance stating those counties' family homes are not in the database (^[2]). Both statements are preserved; the conflict is unresolved. Among licensed center rows, about 7,187 distinct normalized licensee strings appear, 2,190 of them on two or more sites and about 160 on ten or more: a raw string count that overstates distinct companies because spelling variants remain, and therefore is only a starting list for entity resolution.

Texas: dataset versus data book

HHSC's data book, with data as of 7 November 2025 (^[10]), has a table of operations on 31 August by fiscal year (^[10]). Summing its county rows gives, for August 2025, 15,587 operations and 1,214,825 slots; 9,859 licensed centers with 1,166,457 slots; 1,623 licensed homes, 2,051 registered homes and 2,014 listed family homes (analyst sums of county rows). Within licensed centers, child care programs were 8,293, before-and-after-school programs 872 and school-age programs 694. The "licensed center" count is therefore 8,293 or 9,859 depending on whether school-age programs are in scope, a spread of 18.9%.

The open dataset held about 14,990 rows when its metadata was read, with cached column statistics showing about 9,610 licensed centers, 1,731 registered homes, 1,598 licensed homes, 1,380 listed family homes, 431 general residential operations and 240 child-placing agencies (^[4]). Summing the four day care types gives 14,319, which is 8.1% below the data book's 15,587; the dates differ and the cached statistics are undated, so no error is asserted. The dataset's care type field is populated for only about 10,041 rows (^[4]), so center sub-types cannot be separated reliably from the open file alone. A 2021 compilation by Child Care Aware of America listed 6,119 Texas licensed centers (^[58]) and 10,654 in California (^[58]); different years and scopes make these non-comparable with current files, so older compilations should not be mixed in.

National denominators and operator counts

Table 3 below sets out the evidence layers that a team can combine, what each is able to prove, and where it stops.

Table 3 shows that every evidence layer is limited by what regulators and registries publish.

Census data show why unit definitions matter in California and Texas. CBP 2023 lists 9,670 employer establishments for NAICS 624410 in California (^[61]) and 6,242 in Texas (^[62]). Both are lower than the licensed center counts reported above (13,449 non-home facilities in California in June 2026 and 9,859 licensed centers in Texas in August 2025). Possible reasons include nonemployer businesses, which have no paid employees (^[63]), classification of school-based programs in other industries, and differing dates; none is verified here. Firm counts add a further wrinkle: in the Census Statistics of US Businesses definition a firm is geography- and industry-specific, so a multi-state operator can appear as several firms (^[14]).

Operator counts can then be placed against denominators. KinderCare's 1,555 centers are about 1.3% of 121,000 and about 1.9% of 82,162 (analyst arithmetic; the operator figure is January 2026, the denominators 2019 and 2023, and the NSECE count includes public and Head Start providers in its frame). KinderCare estimated that the top five providers held approximately 6% of total capacity (^[9]). A CRS report using 2019 survey data said a relatively small share of center-based providers were for-profit franchises or parts of chains (^[64]), and cautioned that the 2019 data should be treated as illustrative, not predictive (^[64]). Its large-chain table covers 16 organizations with reported licensed capacity of 15,000 or more (^[17]). These sources agree on direction (concentration is low) but not on level, because each counts a different unit.

Implications and Future Directions

For target screening. Mapping should start from state rosters and end at the operating group, not start from brand lists. In California, repeated licensee strings across licensed centers are the first screen, resolved against Secretary of State records. In Texas, where the open dataset has no owner column, the permit holder on each live record is the first link and must be captured with a date.

For due diligence in a roll-up. The non-transferable license rule in California (per a CDSS-hosted manual to be checked against current regulation text) and the ownership-change permit rule in Texas mean that license continuity is a process and a timeline, not an asset. A diligence request list for a multi-site operator should ask for the permit or license number of each site, the legal holder of each, any pending change-of-ownership applications, and the franchise status of each location, then reconcile that list to the public record. Separately, a historical parent label should not be carried forward: the CRS analysis itself says its findings are illustrative (^[17]) and this report does not reuse its owner labels.

Open items before a 50-state build:

• Other states. The other 48 states' rows are not yet researched.
• California search fields. The public Facility Search field list has not been inventoried here; confirm it interactively on the date of use.
• Texas statute text. Statutory provisions are not quoted here; confirm them against the current official code before reliance.
• National counts. No federal count of licensed centers more recent than the 2017 licensing study was found; ACF's database still cites it (^[65]).
• Refresh rate. Licensed small family homes fell 20 percent between 2014 and 2017 (^[25]), so home counts shift faster than center counts and need more frequent refresh.

No national denominator is endorsed here. Legal and licensing questions in a specific transaction need review by counsel and the relevant regulator; the sources cited are public guidance, not advice.

Frequently Asked Questions (FAQs)

How do I find child care center owners from public data?

Start with the license or permit holder, not the brand: California's open center file has a licensee field, and Texas names a permit holder (^[13]). Then look up that legal name in the Secretary of State registry (^[31]) and, in Texas, the Comptroller search (^[33]). This report does not extract owner contact details.

What is the difference between a licensee, a center and a chain?

The center is the physical site; the licensee is the legal person holding the license; the chain is a brand or parent that may control many licensees. Franchised brands add the franchisor, separate from the franchisee operating the site (^[22]).

Can state licensing data identify chain ownership by itself?

Not by itself. The California files and the Texas dataset reviewed here have no parent or chain field (see Table 1). Chain attribution needs corporate records, franchise disclosure documents or dated company evidence, and a stated rule for unresolved cases. Academic work on this problem uses fuzzy name and address matching and then manual name searches (^[16]) (^[16]).

How should a child care acquisition target list and roll-up diligence be built?

Define the mandate and denominator first, screen the state roster by class, status and capacity, reconcile sites to operating groups, and keep every link dated. Diligence adds the ownership-change rules (California's non-transferable license, per a CDSS-hosted manual to be checked against current regulation text, ^[5]; Texas's permit revocation on change of ownership, ^[6]) and franchise structure by outlet (^[22]). This report provides no list of targets or owners.

How do I identify multi-site operators?

Group licensed sites by licensee string and corporate entity number, then merge variants. In the California center file, licensee strings repeating on two or more licensed sites number in the thousands before cleaning (analyst parse), so repeated strings are candidates, not conclusions.

Conclusion

The central finding is methodological. A licensed child care record is a regulated operation at a place, held by a legal person, often under a brand, sometimes inside a franchise, and occasionally within a larger parent. State sources supply the first two layers and, unevenly, a licensee string; they supply the franchise and parent layers hardly at all. Neither California nor Texas gives a chain field.

For that reason, the protocol asks for a declared universe, separate fields for site, licensee, DBA, franchisor and parent, typed and dated links, a coverage ratio that is reported as a minimum plus an unresolved range when dates or definitions differ, and a blind sample of any external map against the same state file. Denominator choices alone shift ratios by tens of percent, as the Texas 8,293 versus 9,859 spread shows, and ownership-change rules in both states mean license continuity cannot be assumed across a transaction. Public evidence on private operators is incomplete, and conflicts are kept visible rather than resolved by assumption.

Teams that want this worksheet populated for a defined mandate, with proof links retained, can contact Docket, the publisher, which is an interested party as disclosed in the Introduction.