What each segment of a Chinese company code reveals before you search.

1. The 18-character map

The USCI is not a random serial number. Each segment sits at fixed spots and encodes one official fact. The standard is GB 32100-2015, titled “Coding Rules for the Unified Social Credit Identifier for Legal Entities and Other Organizations” (法人和其他组织统一社会信用代码编码规则).

The character set is not the full alphanumeric range. The USCI uses 31 characters: the digits 0–9 plus uppercase letters A–Z excluding I, O, Z, S and V. Five letters are left out to prevent mix-ups (I/1, O/0, Z/2, S/5) and to keep the set at exactly 31 characters, a prime needed by the check-character formula.

2. Position 1: who registered the entity

The first character shows the registrar. For a buyer checking a trade supplier, the expected value is 9.

If a supplier gives you a USCI starting with 1 or 5, the entity is a government body or a social group — not a trade firm. That does not mean the code is wrong. It means the entity behind it is not what you would expect from a factory or trading company.

3. Position 2: what kind of organisation

Under the market-regulation authority (position 1 = 9), position 2 divides entities into three categories:

The combination of positions 1 and 2 is the first filter. A code starting with 91 is an enterprise registered with market regulation — the most common type for a trade supplier. A code starting with 92 is an individual business. It can legally trade, but it is a sole proprietor, which affects contracts and risk.

4. Positions 3–8: where the file sits

Positions 3 through 8 are the six-digit administrative division code defined by GB/T 2260. China gives a layered code to every local unit. The first two digits identify the province or municipality, the next two the prefecture-level city, and the last two the county or district.

This tells you where the entity was filed, not always where it works. A factory in a zone outside the city may be filed at the district office downtown. But the filing area narrows your search to one local bureau. It also tells you which province’s records to look for in a search.

The common province codes a buyer encounters in cross-border trade: 33 Zhejiang, 44 Guangdong, 32 Jiangsu, 35 Fujian, 37 Shandong, 31 Shanghai. If the code points to a province you did not expect, it is worth asking the supplier why.

5. Positions 9–17: the unique identifier

This nine-character segment is the entity’s organisation code, once issued under GB 11714 as the “organisation code cert” (组织机构代码证). When the unified coding reform merged five old codes into the USCI, this segment was embedded as positions 9 through 17.

For a buyer, this segment is an opaque identifier. Unlike positions 1–8 it does not encode readable facts. Its value is uniqueness. No two entities under the same authority, category and jurisdiction share the same nine-character code. Together with the first eight characters, it produces a nationally unique 17-character string whose integrity is guarded by position 18.

6. Position 18: the check character

The last character is computed from the preceding 17 using a weighted-sum method from ISO 7064, MOD 31-3. Our separate measurement confirmed it catches 100% of single-character swaps and 100% of flips between spots across 1,080,000 full tests.

Before you type a USCI into any registry, run it through the check-character calculator. A failure means the string has a typo. Any search with it will either fail or return the wrong entity. Fix the code first.

7. Worked example

Take a sample USCI starting with 91370102. Reading it segment by segment:

Before searching any registry, you already know: this is a trade firm, filed at the market regulation bureau in Lixia District, Jinan, Shandong. If your supplier told you they are a factory in Guangdong, this code says otherwise — and that gap is worth fixing before money moves.

8. Method and sources

The segment maps on this page come from GB 32100-2015, the national standard from the SAC. The area codes follow GB/T 2260. The check-character method follows ISO 7064, MOD 31-3, as the standard states.

Each mapping was cross-checked against live GSXT (国家企业信用信息公示系统) records: codes starting with 91 always match enterprises under market regulation, codes starting with 92 match sole traders. The check character was tested with this site’s own USCI checker, which runs the same MOD 31-3 method and was tested on its own against 1,080,000 full mutations.

What this page does not cover

• Sub-codes under authorities other than market regulation. Position 2 codes under authorities 1, 5 and Y follow separate schemes not shown here. This page covers the three types (enterprise, sole trader, co-op) that appear under authority 9.
• GB/T 2260 division codes in full. There are thousands of county-level divisions. This page explains the structure; a full lookup requires the current edition of GB/T 2260 or the NBS statistical division database.
• What the code proves about the entity. The USCI encodes official grouping, not business quality. Active status, registered capital, business scope and legal representative live in the registration record, not the code. A valid USCI from a real authority still needs a dated filing check.

Questions about this study

Citing this study

You may quote or reuse these segment maps, even for profit, as long as the standard reference (GB 32100-2015) and the date travel with them. The check-character measurement data is published separately in the check-digit protection study.

A citation with everything it needs:

Currawong, “What each segment of a Chinese USCI tells before you search”, segment-by-segment anatomy of the 18-character Unified Social Credit Code per GB 32100-2015, cross-checked against live GSXT records and the site’s own MOD 31-3 checker, 5 September 2026. https://currawongweb.com/verify/usci-code-anatomy-study/

This study maps code structure, not a service claim. Currawong verifies public records on request and states sources, query dates and limits with every result; it does not certify suppliers.