Firmographic data is the set of attributes that describe a company, the business equivalent of demographics: industry, employee count, headquarters and offices, legal entity, founding date, ownership and revenue. B2B teams use it to define an ideal customer profile and to filter, score and route accounts. Most of it is public.
Guides on the topic list the fields. This one fills them in. On 20 November 2026 we built a firmographic record for one public company, Allbirds, from four public sources with four collector runs, then merged them in Python. The table at the end shows which source gave which field, and the three places where the sources disagree.
The firmographic fields that matter
| Field | Why sales and marketing use it | Typical public source |
|---|---|---|
| Industry | Segment the market, pick the message | Company profile pages, the company's own site |
| Employee count | Size band, deal size, who signs | Company profile pages, job boards, annual reports |
| Headquarters and offices | Territory, language, time zone, tax rules | Company profile pages, legal registers |
| Legal entity and identifier | Contracts, invoicing, matching records across systems | Business and LEI registers |
| Founding date and age | Startup or established buyer | Registers, company profile pages |
| Ownership and revenue | Budget and buying process | Annual reports and filings of listed companies |
| Technology | Fit with what you sell | The website itself (see technographic data) |
The last row is technographic rather than firmographic, but the two are almost always used together, and both join on the same key: the company's domain.
Four sources, four collector runs
We picked one company and asked four questions, each answered by a different public source. All four runs were started from Claude through our MCP server and returned the same day.
| Collector | Source | What it adds | Price per row |
|---|---|---|---|
| LinkedIn company collector | The public LinkedIn company page | Industry, employee count, founding year, HQ, office list | $0.005 |
| GLEIF LEI collector | The GLEIF register of legal entities | Legal name, other names, LEI, jurisdiction, entity creation date | $0.0004 |
| company profile collector | The company's own website | Description, industry in its own words, social profiles | $0.03 |
| tech stack collector | The company's homepage and headers | Store platform, CDN, analytics | $0.005 |
The record costs $0.0404 in total. The calls themselves are short; for example, the legal-entity lookup is one line of input:
run_collector { "slug": "gleif_lei", "input": { "query": "Allbirds", "max_results": 3 }, "async": true }
Merging the record in Python
Each run has an id, so a script can collect the four results over the REST API and merge them. This is the script we ran:
import os, json, requests
API = "https://api.quanticdata.io/v1/scraper/collectors/runs/"
headers = {"Authorization": f"Bearer {os.environ['QUANTICDATA_API_KEY']}"}
RUNS = { # four collector runs on 20 Nov 2026, one public source each
"linkedin": "LINKEDIN_RUN_ID", # linkedin_company
"gleif": "GLEIF_RUN_ID", # gleif_lei
"site": "PROFILE_RUN_ID", # company_profile
"stack": "STACK_RUN_ID", # tech_stack
}
def rows(run_id):
return requests.get(API + run_id, headers=headers, timeout=60).json()["payload"]["results"]
li = next(r for r in rows(RUNS["linkedin"]) if r["slug"] == "allbirds")
lei = rows(RUNS["gleif"])[0]
site = rows(RUNS["site"])[0]
stack = next(r for r in rows(RUNS["stack"]) if r["domain"] == "allbirds.com")
record = {
"name": li["name"],
"legal_name": lei["legal_name"],
"other_legal_names": lei["other_names"],
"lei": lei["lei"],
"jurisdiction": lei["jurisdiction"],
"entity_created": lei["entity_created"][:10],
"founded": li["founded"],
"industry": li["industry"],
"industry_site": site["industry"],
"employees": li["employees"],
"hq": f'{lei["hq_city"]}, {lei["hq_country"]}',
"offices": [l["address"] for l in li["locations"]],
"website": site["url"],
"ecommerce_platform": stack["ecommerce_platform"],
"socials": sorted(site["socials"]),
}
print(json.dumps(record, indent=2))
filled = sum(v not in (None, [], "") for v in record.values())
print(f"\n{filled} of {len(record)} fields filled from 4 sources")
The output, unedited:
{
"name": "Allbirds",
"legal_name": "SMARTBIRD, INC.",
"other_legal_names": [
"ALLBIRDS, INC."
],
"lei": "5299005VZ0KFXF1GL494",
"jurisdiction": "US-DE",
"entity_created": "2015-05-05",
"founded": 2016,
"industry": "Retail Apparel and Fashion",
"industry_site": "Footwear",
"employees": 383,
"hq": "San Francisco, US",
"offices": [
"San Francisco, California, US",
"London, GB",
"Shanghai, CN"
],
"website": "https://allbirds.com/",
"ecommerce_platform": "Shopify",
"socials": [
"facebook",
"instagram",
"tiktok",
"twitter",
"youtube"
]
}
15 of 15 fields filled from 4 sources
Where the sources disagree, and why it matters
A firmographic record is a set of claims, each from a source with its own definition. Three of ours do not match, and all three differences are useful.
- Brand name and legal name. The brand is Allbirds. The register lists the legal entity as SMARTBIRD, INC., with ALLBIRDS, INC. as a previous legal name, incorporated in Delaware (US-DE) with its headquarters in San Francisco. Contracts, invoices and credit checks need the legal name and the LEI; outreach needs the brand. Before a contract, check in the register which entity trades under the brand today. A record that keeps only one of them breaks either the CRM match or the email.
- Two founding dates. The LinkedIn page says founded 2016; the legal entity was created on 5 May 2015. Neither is wrong: one is when the brand launched, the other when the entity that used to carry the Allbirds name was formed. Keep both, labelled, instead of picking one.
- Two industries. LinkedIn files the company under "Retail Apparel and Fashion"; its own site says "Footwear". Map both to your own segment list rather than trusting one taxonomy.
The same rule applies to employee counts. 383 is the number the LinkedIn page publishes on the day we read it; an annual report or a job-board count can differ. Store the source and the date next to each value.
Firmographic data sources compared
| Source | Strength | Limit |
|---|---|---|
| Legal registers (GLEIF, national business registers) | Authoritative legal name, identifiers, addresses | No headcount, no industry in sales terms |
| Company profile pages | Industry, size band, offices, founding year | Self-reported by the company, updated when it chooses |
| The company's website | How the company describes itself, links to every social profile | Wording varies from site to site |
| Annual reports and filings | Revenue, ownership, audited headcount | Listed or filing companies only, once a year |
| Resold B2B databases | Millions of companies in one file | Collection date and source per field rarely shown |
Firmographic data enrichment at scale
The script above works for one company. For a list, run each collector once with all the companies in its input (the LinkedIn and tech stack collectors take up to 50 per run), then merge on the domain. Three habits keep the merged table usable:
- Normalise the join key first. Lowercase domains, strip "www." and paths, and keep a mapping table from brand to legal entity once you find it.
- Keep provenance columns. For each field, keep which source gave it and when. When two sources disagree you can decide by rule, not by memory.
- Refresh on a schedule. Headcount and offices change every quarter. A re-run costs the same per row as the first one.
If you would rather start from a finished list, our ready-made business datasets cover businesses by sector and country, and company data scraping builds a company list from a prompt. For the next step, turning firmographics into a target list, read how to build a B2B lead list.


