NOMIS – UK Business Register & Employment Survey by SIC

The Business Register and Employment Survey is the Office for National Statistics' annual head-count of who works where, and Nomis — the labour-market data service Durham University runs on the ONS's behalf — is where its SIC-coded tables live. Together they answer the question most go-to-market plans start with: how many people work in industry X, in place Y, and across how many businesses?

The survey reaches roughly 1,171 SIC 2007 industries at five-digit subclass depth — so 63110 data processing, hosting and related activities stays distinct from 63120 web portals and 63990 other information services — across hundreds of geographies a year, from nations and regions down to local authority districts, parliamentary constituencies and wards. A companion family, UK Business Counts, adds enterprise and local-unit tallies by employment size band and turnover band. One verified cut: England carried 43,000 employees in SIC 63110 in 2024, 39,000 of them full-time, inside 2,960 enterprises counted the following year.

Datadory delivers Data Processing & Outsourced Services data covering every one of those cuts, shaped and typed into your environment. Get a sample of this dataset to see the shape before anything ships.

Three rows exactly as they land in an extract, so you can judge identifiers, splits and values before requesting anything:

date | geography_name | geography_code | industry_code | employment_status   | measure | obs_value
2024 | England        | E92000001      | 63110         | Employees           | Count   | 43000
2024 | England        | E92000001      | 63110         | Full-time employees | Count   | 39000
2025 | England        | E92000001      | 63110         | Enterprises         | Count   | 2960

The first two rows come from the employment survey, where the unit is persons and the status dimension splits employees into full-time, part-time and self-employed. The third comes from the business-counts companion, where the unit flips to businesses — 2,960 enterprises classified to data processing and hosting in England, splittable further by employment size band.

Nine fields come standard in every extract we cut from this family. They carry the full spine of each observation — when, where, which industry, which slice of the workforce, and the number itself — which is enough to join straight onto company lists, sales territories or your own CRM geography without reshaping anything on your side.

Additional fields on request: rows also carry a value-versus-percent selector, an observation-quality flag, a confidentiality marker, and a URN that encodes the full selection key behind the record. Tell us the analysis you are running and we map those onto your sample rather than shipping columns nobody reads.

• Geography — the whole United Kingdom: local authority and district, combined authorities, regions and nations, 2024 parliamentary constituencies, 2011 wards in England and Wales, with postcode and coordinate lookups on top.
• Time — the survey runs annually from 2009 onward; the business-counts series runs through 2025. One seam to respect: from 2015 the survey folds in PAYE-only businesses, so long-run series need that break handled deliberately.
• Grain — area × SIC 2007 subclass × employment status × measure, the finest public grain available for British industry-by-place employment.

That constituency-and-ward reach is the quiet superpower here: territory design and political-boundary analysis draw on exactly the same numbers national planners use.

API, files, or your warehouse. Daily, weekly, or hourly.

You pick the channel and the cadence; extraction, cleaning, and schema stability are our problem. When a survey year gets revised or classification labels shift, the feed you consume stays normalized — same columns, same types, same join keys. Bulk pulls land as files, continuous consumption runs through the API, and warehouse-native loads write straight into your own storage. A sample ships first either way, sized to test in your pipelines the same day.

Three of our persona groups lean on it hardest:

1. Sales & Growth Teams — rank every district and constituency by head-count in SIC 63110 or 63120 to find where the buyers actually cluster, then size each territory before rep time goes anywhere near it. More in the sales & growth teams hub.
2. Data Scientists & ML Engineers — train market-sizing and site-selection models on more than fifteen years of industry-by-area employment, with the size-band splits arriving as ready-made features. Details on the data scientists page.
3. Competitive Intelligence & Product Teams — watch whether data-processing employment is concentrating in your hubs or leaking to new ones, and benchmark your head-count footprint against the industry's actual distribution. See competitive intelligence & product teams.

The recurring use cases: market-sizing decks built from area × SIC counts; territory design balanced on full-time versus part-time mix; vendor and supplier due diligence that wants real employment weight behind a counterparty's address.

Cards worth reading next: the neighboring datasets in this same industry, the comparison that pits this survey against the European structural-business-statistics branch, and the primers that decode the classification behind every row.