The Companies House bulk data download is the free Free Company Data Product: a monthly snapshot of every live company on the register, offered as one ZIP of about 470 MB or as several smaller part files, each holding a CSV. It is too large for an Excel worksheet, so open it with Power Query, a database or a script, keep company numbers as text, and filter by status, SIC code, postcode area and accounts category before you load anything to a sheet.
This guide is part of the Companies House lead generation series. It covers what the file contains, how to open it, five filter recipes and the monthly refresh.
What the product contains
Companies House describes it as "a downloadable data snapshot containing basic company data of live companies on the register", provided "as ZIP files containing data in CSV format". GOV.UK lists the contents as basic company information, SIC codes, company status, filing dates and previous names. Key facts:
- Live companies only. Dissolved companies are not included. A company that dissolves disappears from the next month's file.
- Monthly. "The latest snapshot will be updated within 5 working days of the previous month end", and the contents are compiled up to the end of the previous month.
- Free and unsupported. The page says the snapshot "is provided free of charge and will not be supported".
- Sorted by company name, with URI links to each company's record for up-to-date details.
The main columns you will use (check the header row of your file, because several names have a leading space, such as CompanyNumber):
| Column | Example (fictional) | Notes |
|---|---|---|
CompanyName | NORTHFIELD PLUMBING LIMITED | Upper case |
CompanyNumber | 09876543, SC123456 | Keep as text: leading zeros and prefixes such as SC, NI and OC matter |
RegAddress.AddressLine1, RegAddress.PostTown, RegAddress.PostCode | UNIT 4 MILL LANE, LEEDS, LS12 3AB | Registered office, not trading address |
CompanyCategory | Private Limited Company | Company type |
CompanyStatus | Active | Also values such as Active - Proposal to Strike off and Liquidation |
IncorporationDate | 09/11/2021 | Day first |
Accounts.AccountCategory | DORMANT, MICRO ENTITY, TOTAL EXEMPTION FULL | Type of the last accounts filed |
Accounts.NextDueDate, Accounts.LastMadeUpDate | Filing timing | |
SICCode.SicText_1 to SICCode.SicText_4 | 43220 - Plumbing, heat and air-conditioning installation | Code and description in one field |
ConfStmtNextDueDate | Confirmation statement due date | |
URI | Link to the company record | |
PreviousName_1.CompanyName and so on | Up to ten previous names |
A field description PDF is linked from the download page. Value lists can change, so run a quick count of the distinct values in CompanyStatus and Accounts.AccountCategory in your own copy before you write filters.
File size and splitting
In November 2026 the download page offered:
BasicCompanyDataAsOneFile-2026-11-01.zip: 471 MB- Seven parts,
BasicCompanyData-2026-11-01-part1_7.ziptopart7_7.zip: about 70 MB each, the last about 50 MB
Unzipped, the CSV is several times the ZIP size, so allow several gigabytes of free disk space.
Choose by tool:
- Power Query, a database or a script: the single file is simplest.
- Plain Excel without Power Query, or a low-memory PC: use the parts, but each part may still exceed the worksheet limit, so filtering on import is still the safer route.
Excel's worksheet limit is 1,048,576 rows by 16,384 columns. The full register has more rows than that, which is why opening the CSV directly either fails or cuts off data without much warning.
Opening it in Excel with Power Query
Power Query reads the file without loading every row into a sheet.
- In Excel, choose Data > Get Data > From File > From Text/CSV and pick the unzipped CSV.
- In the preview, set File Origin to 65001: Unicode (UTF-8) and the delimiter to comma, then choose Transform Data (not Load).
- Select
CompanyNumberandRegAddress.PostCodeand set their type to Text. If Power Query has guessed a number type, change it before any other step, or leading zeros are lost. (The same problem affects phone numbers; see why Excel removes leading zeros.) - Set
IncorporationDateto Date using Change Type > Using Locale > English (United Kingdom) so day-first dates are read correctly. - Use Home > Remove Columns to drop columns you do not need (mortgages, limited partnerships, previous names). Fewer columns make every later step faster.
- Apply the filters from the recipes below.
- Choose Close & Load To... and load the filtered result to a table. If it is still too big, choose Only Create Connection and add it to the Data Model, then use a PivotTable.
If company names show odd characters such as é, the file was read with the wrong encoding; see fixing garbled characters in CSV files.
Opening it with a script or database
For repeatable work, a short script is easier to rerun each month. This Python example reads the file in chunks, so memory stays modest:
import pandas as pd, re
path = "BasicCompanyDataAsOneFile-2026-11-01.csv"
sic_wanted = {"43210", "43220", "43320"}
areas = re.compile(r"^(LS|BD|HX|HD|WF)[0-9]") # West Yorkshire postcode areas
keep = []
for chunk in pd.read_csv(path, dtype=str, chunksize=200_000, encoding="utf-8"):
chunk.columns = chunk.columns.str.strip() # some headers have leading spaces
sic = chunk[[f"SICCode.SicText_{i}" for i in range(1, 5)]].apply(
lambda col: col.str[:5])
mask = (
chunk["CompanyStatus"].eq("Active")
& sic.isin(sic_wanted).any(axis=1)
& chunk["RegAddress.PostCode"].fillna("").str.upper().str.match(areas)
& ~chunk["Accounts.AccountCategory"].fillna("").eq("DORMANT")
)
keep.append(chunk[mask])
pd.concat(keep).to_csv("west-yorkshire-trades.csv", index=False, encoding="utf-8")
dtype=str keeps company numbers and postcodes intact. The postcode pattern requires a digit after the letters, so L does not match LS by accident.
If you prefer a database, SQLite's command line .import --csv or PostgreSQL's COPY will load the file into a table; index CompanyNumber and the postcode column before running filters.
Filter recipes
1. Active companies in a niche and area
CompanyStatusequalsActive- First five characters of any
SICCode.SicText_nin your code list RegAddress.PostCodestarts with your postcode areasCompanyCategoryisPrivate Limited CompanyorLimited Liability Partnership
See SIC codes for prospecting for a starter code list by niche.
2. Exclude dormant and non-trading companies
Accounts.AccountCategoryis notDORMANT- No
SICCode.SicText_nstarts with99999(dormant company) or74990(non-trading company)
More in dormant companies: find and exclude.
3. Established companies only
IncorporationDateis more than two years agoAccounts.LastMadeUpDateis not blank (the company has filed accounts)
4. Overdue filers to avoid
Accounts.NextDueDateis in the past, orConfStmtNextDueDateis in the past
Overdue filings can precede strike-off. Hold these for a status check rather than contacting them.
5. Service address clusters to flag
Group by RegAddress.PostCode plus RegAddress.AddressLine1 and count companies. Addresses with dozens or hundreds of companies are accountants, formation agents or virtual offices. Keep the companies, but mark their location as unknown until you find a trading address. Companies House lead generation explains how to keep the two addresses apart.
Refresh cadence
The file is monthly, so build a monthly habit:
- First week of the month: download the new snapshot once it is published (the file name carries the date of the first of the month).
- Rerun your saved query or script against the new file.
- Compare with last month by company number:
- In this month, not last month: new companies.
- In last month, not this month: dissolved or otherwise removed. Suppress them.
- In both, with changed SIC codes, address or accounts category: review.
- Check status live through the public API for anything you plan to contact this week. A month is a long time for a company in difficulty.
For daily new companies, the bulk file is too slow; use the API or stream described in finding newly incorporated companies every day.
How Map Extract helps
Map Extract's Companies House matching workflow, part of the Windows desktop app, can import the snapshot directly as a CSV or ZIP, so you do not need Power Query or a script to get the register into a project. It uses public data from Companies House and is not affiliated with it.
- Snapshot imports and status checks. With your own free Companies House API key, the workspace can check company status. A company missing from a later snapshot is not assumed closed.
- Inactive companies excluded or held without deleting their history, and status history is kept.
- Registered office kept apart from the trading location.
- Company-first website discovery. Candidate websites are found and the link is checked before an accepted site goes to the website data extractor for public business contacts. Unmatched companies stay in staging.
- Related tools. The AI business finder proposes niches and locations for you to review, and the UK property intelligence workspace (beta) reads the Companies House stream with SIC targeting.
What it does not do: it does not make SIC codes more accurate than the companies made them, it does not add email addresses to the register, and it does not replace your own review of filters and matches.
Public business and social results the app collects are uploaded automatically to Map Extract's servers while it is licensed and running; the privacy notice explains this.
Key takeaways
- The free bulk product is a monthly CSV snapshot of live companies only, about 470 MB zipped.
- It exceeds Excel's 1,048,576-row limit: use Power Query, a database or a chunked script.
- Keep company numbers and postcodes as text, and read dates as day first.
- Filter by status, SIC code (first five characters), postcode area and accounts category before loading.
- Refresh monthly, compare files by company number and check status live before contact.
Frequently asked questions
Where do I download the Companies House bulk data?
From the Free Company Data Product page at download.companieshouse.gov.uk. It offers the whole register of live companies as one ZIP file or as several smaller parts, each containing a CSV. It is free, needs no account and is updated monthly, within five working days of the previous month end. Companies House says the snapshot itself is not supported.
Why won't the Companies House CSV open in Excel?
The file has more rows than an Excel worksheet can hold, which is 1,048,576 rows. Excel either refuses or silently cuts the data. Load it through Power Query instead and filter before loading to a sheet, or use the split part files, or import it into a database or a script that reads it in chunks.
Does the bulk file include dissolved companies?
No. Companies House describes it as a snapshot of live companies on the register. Companies that dissolve drop out of the next month's file. To find dissolved companies use advanced search, which shows companies dissolved since 1 January 2010, or the public API. Compare monthly files by company number to see which companies left the register.
How often is the Companies House bulk data updated?
Monthly. The snapshot is compiled up to the end of the previous month and published within five working days of month end, with the date of the first of the month in the file name. For anything fresher, such as today's new companies or a status check before outreach, use the free public data API or the streaming API.
Sources
- Free Company Data Product download page
- Companies House data products (GOV.UK)
- Public data product FAQ (Companies House)
- Excel specifications and limits (Microsoft Support)
- SIC 2007 condensed list (Companies House)
- Advanced company search (Find and update company information)
- Companies bulk data notes (CH Guide, community resource)
Written by the Map Extract editorial team. Found an error? Email [email protected] and we will check and correct it. How we research guides: editorial policy.