[NewSearch millions of jobs from your AI agent with MCP→](/blog/jobspipe-mcp-server)

[All posts](/blog)

![](/listly/card-field.png)

ReferenceISIC

Reference·Aug 9, 2026·7 min read

# ISIC codes: the industry classification behind global labour market data

ISCO-08 tells you what the worker does; ISIC tells you what the employer does. This is the industry half of that pair: how the Rev.4 hierarchy of 21 sections and 88 divisions works, the full section list, how ISIC relates to NACE, NAICS and SIC, what live job postings look like ranked by division - staffing agencies at number three is the interesting one - and how to filter postings by employer industry with a single query parameter.

![Dvir Atias](/authors/dvir-atias.jpg)

Dvir Atias

Founder, JobsPipe

An ISIC code is a numeric identifier from the International Standard Industrial Classification of All Economic Activities, the UN Statistics Division’s system for classifying what a business does. It is the industry counterpart to [ISCO-08](/blog/isco-08): ISCO classifies the work a person does, ISIC classifies the economic activity of the organisation employing them. You need both axes to answer the questions labour data gets asked - “are banks hiring more software engineers than software companies?” is an ISCO filter crossed with an ISIC filter, and neither axis alone can answer it.

This page covers what an ISIC code actually classifies, how the Rev.4 hierarchy works, the full section list with division ranges, how ISIC relates to NACE, NAICS and SIC, what two million live job postings look like ranked by division, and how to query postings by employer industry without maintaining a classification pipeline yourself.

## ISIC classifies the employer, not the job

The single most common mistake with ISIC is reading it as a job category. It is not. An ISIC code describes the principal economic activity of the establishment - a software developer employed by a bank sits in ISCO unit group `2512` and ISIC division `64` (financial services), while the same developer at a SaaS company is `2512` in division `62`. The job is identical; the industry is not. That distinction is exactly what makes the second axis useful: it separates who is doing the hiring from what they are hiring for.

## How the Rev.4 hierarchy works

ISIC Rev.4, published in 2008 and still the revision almost all national data is keyed to, has four levels: 21 sections lettered A through U, 88 two-digit divisions, 238 three-digit groups and 419 four-digit classes. The section letter never appears in the code itself - the division digits imply it. So `6201` reads: section J (Information and communication), division `62` (Computer programming, consultancy and related activities), group `620`, class `6201` (Computer programming activities). In practice the two-digit division is the grain most analysis runs at - fine enough to separate banks from insurers, coarse enough that classification stays reliable.

| Section | Activity | Divisions |
| --- | --- | --- |
| A | Agriculture, forestry and fishing | 01-03 |
| B | Mining and quarrying | 05-09 |
| C | Manufacturing | 10-33 |
| D | Electricity, gas, steam and air conditioning supply | 35 |
| E | Water supply, sewerage and waste management | 36-39 |
| F | Construction | 41-43 |
| G | Wholesale and retail trade, vehicle repair | 45-47 |
| H | Transportation and storage | 49-53 |
| I | Accommodation and food service activities | 55-56 |
| J | Information and communication | 58-63 |
| K | Financial and insurance activities | 64-66 |
| L | Real estate activities | 68 |
| M | Professional, scientific and technical activities | 69-75 |
| N | Administrative and support service activities | 77-82 |
| O | Public administration, defence, social security | 84 |
| P | Education | 85 |
| Q | Human health and social work activities | 86-88 |
| R | Arts, entertainment and recreation | 90-93 |
| S | Other service activities | 94-96 |
| T | Households as employers | 97-98 |
| U | Extraterritorial organisations and bodies | 99 |

The division numbers run 01 to 99 with deliberate gaps (04, 34, 40, 44 and others are unassigned), which is how 99 slots hold 88 divisions. The gaps are room for future revisions - and a small trap for anyone validating codes against a dense range.

## ISIC vs NACE vs NAICS vs SIC

The alphabet soup resolves into one family tree. NACE Rev.2 is the EU’s implementation of ISIC Rev.4 - identical at the division level, more detailed below it - so European labour data joins against ISIC almost for free. The UK’s SIC 2007, the code every UK company files at Companies House, is built on NACE, which makes it ISIC-compatible at two digits too. NAICS, the North American system, is a genuinely different structure - 20 sectors, six-digit codes, revised on its own cycle - and only official concordance tables connect it to ISIC. The legacy US SIC of 1987 still appears in SEC filings and nowhere else that matters. ISIC Rev.5 was adopted by the UN Statistical Commission in 2023, but national implementations move slowly and virtually all published labour data remains keyed to Rev.4-aligned systems today.

The practical rule mirrors the [occupation-side advice](/blog/isco-08): if your data spans more than one country, standardise on ISIC divisions and crosswalk outward. It is the level everything else already maps to.

## What live postings look like by division

Official industry statistics are published annually from survey data. Job postings give the same axis a live reading. Ranked by deduplicated active postings in the JobsPipe index (measured 9 August 2026, across the 304,669 active postings that carry an ISIC division - classification is still rolling out across the rest of the two-million-posting index):

| Division | Activity | Active postings |
| --- | --- | --- |
| 62 | Computer programming and consultancy | 43,548 |
| 86 | Human health activities | 26,959 |
| 78 | Employment activities (staffing agencies) | 20,921 |
| 64 | Financial services, except insurance | 15,927 |
| 47 | Retail trade | 14,975 |
| 85 | Education | 11,094 |
| 26 | Computer, electronic and optical manufacturing | 8,899 |
| 28 | Machinery and equipment manufacturing | 8,582 |

Division 62 leading is no surprise given where postings concentrate online. The interesting row is `78`, Employment activities, at number three: those are staffing and recruitment agencies, and a posting from an agency is classified under the agency’s industry, not the client’s. If you are building sales intelligence on top of [hiring signals](/blog/hiring-signals-api), that is the division to handle explicitly - an agency posting a warehouse job tells you a logistics company somewhere is hiring, but the ISIC code on the posting says “staffing”.

## Querying live postings by ISIC division

Every classified posting in the JobsPipe index carries `isic_division` and `isic_division_label` fields alongside its ISCO-08 occupation code - [how the classification pipeline works](/blog/classified-job-postings-api) is its own post. The search filter takes division codes directly:

```
curl -X POST https://api.jobspipe.dev/v1/jobs/search \
  -H "Authorization: Bearer $JOBSPIPE_API_KEY" \
  -d '{
    "isic_division_or": ["62", "63"],
    "job_country_code_or": ["de"]
  }'
```

The two-axis queries are where it earns its keep. Software developers being hired by financial firms rather than software firms is one occupation filter crossed with one industry filter:

```
curl -X POST https://api.jobspipe.dev/v1/jobs/search \
  -H "Authorization: Bearer $JOBSPIPE_API_KEY" \
  -d '{
    "occupation_code_or": ["2512"],
    "isic_division_or": ["64", "65", "66"]
  }'
```

And the aggregate endpoints answer the ranking questions without paging through postings: the first call below ranks every division by live posting volume, the second returns which skills are trending inside one industry.

```
curl "https://api.jobspipe.dev/v1/insights/industries?window_days=90" \
  -H "Authorization: Bearer $JOBSPIPE_API_KEY"

curl "https://api.jobspipe.dev/v1/insights/industries/62/skills/trending" \
  -H "Authorization: Bearer $JOBSPIPE_API_KEY"
```

## Practical traps

**Divisions are strings.** Codes 01 through 09 have leading zeros, and an integer cast turns division `01` (crop and animal production) into `1`, which is not a valid division at all. Same trap as ISCO codes, same fix: text everywhere.

**The employer axis is not the job.** A nurse hired through a staffing agency is division `78`, not `86`. Inferring industry from the job description reinvents the misclassification ISIC exists to avoid - filter on the occupation axis when you mean the work, the industry axis when you mean the employer.

**Watch the revision.** Datasets from before roughly 2010 are usually keyed to ISIC Rev.3.1, and several divisions were renumbered in Rev.4. Concordance tables exist; mixing revisions in one time series without converting produces quiet nonsense.

**Coverage under 100% is honest.** Some employers genuinely cannot be classified from public information. A pipeline that leaves `isic_division` empty for those beats one that guesses - if a vendor claims every record has an industry code, ask how the ambiguous ones were decided.

## Where to go next

The official Rev.4 structure with full definitions is free from the UN Statistics Division, and it is the right reference for the taxonomy itself. For the occupation axis, start with [ISCO-08](/blog/isco-08) and the [skills layer](/blog/skills-taxonomy) that sits under both. For the broader landscape of sources this data comes from, the [labour market data sources](/blog/labor-market-data-sources) comparison covers what is free and what is not. And if you want the codes on live postings today, the free tier includes 1,000 jobs a month - enough to pull the full industry ranking and cross it with a few occupations every day.

Filter two million live postings by employer industry - free API key in 30 seconds.

[

← Previous

ESCO API: the EU's free skills and occupations API, explained for developers

](/blog/esco-api)[

Next →

Introducing Signals: stop checking whether a company started hiring

](/blog/introducing-signals)

---
Canonical URL: https://jobspipe.dev/blog/isic-codes
Title: ISIC codes: the industry classification behind global labour market data
Description: ISCO-08 tells you what the worker does; ISIC tells you what the employer does. This is the industry half of that pair: how the Rev.4 hierarchy of 21 sections and 88 divisions works, the full section list, how ISIC relates to NACE, NAICS and SIC, what live job postings look like ranked by division - staffing agencies at number three is the interesting one - and how to filter postings by employer industry with a single query parameter.

---
AI agents: the full machine-readable index of this site is https://jobspipe.dev/llms.txt?src=md-twin - API quickstart, no-key sandbox (POST https://api.jobspipe.dev/v1/sandbox/jobs/search), MCP server, pricing. Free key: https://jobspipe.dev/signup