Skovgaard Systems

What 46,402 planning permissions say about who buys M&E work

28 August 2026 · Magnus Skovgaard · 46,402 permissions, January 2024 to September 2026

Every commercial building in Britain is a mechanical and electrical package waiting to be let. The permission is public two years before the package is. We pulled every one granted since January 2024, classified them, and traced who applied — to find out how much of that pipeline is actually visible in advance.

46,402permissions
11,080commercial
5,418substantial enough to matter
944with a named developer

Method

We took every planning application recorded across British local authorities between January 2024 and September 2026 that was granted, classified medium or large, and lodged as a full or outline application rather than a variation. Conditions and amendments were excluded: they are paperwork on permissions that already exist, and they account for roughly two thirds of the raw feed.

Each description was then classified by sector from its own wording, and scored for substance. A permission scores on construction verbs (“erection of”, “change of use”), on size, and on dwelling count; it loses points for the vocabulary of domestic work — porches, dormers, garages, replacement windows. Only permissions scoring two or more were kept.

Finally we traced the applicant. Planning aggregators do not publish it, so each one was read from the council's own case page.

What is actually out there

Of 46,402 permissions, 11,080 fall in sectors that buy mechanical and electrical work. After the substance filter, 5,418 remain.

SectorPermissionsShare
Industrial and warehousing1,83434%
Retail81115%
Energy — solar, storage, substations71713%
Offices69513%
Schools and colleges54010%
Hotels and leisure4218%
Healthcare4007%

The energy figure is the one worth pausing on. Solar arrays, battery storage, substations and vehicle charging are not a niche within construction — at 717 permissions they are the third largest category of commercial work in the country, and they are almost entirely electrical. Anyone classifying construction work by building type will miss them, because most of them are not buildings.

The applicant is usually not a developer

We traced the applicant on every commercial permission that passed the substance filter. Names were recovered on about two thirds of them. Of those:

ApplicantCount
A company or public body944
A private individual1,286

Private individuals outnumber organisations. That is not a data problem, it is what the register looks like: a great deal of commercial-sector permission is applied for by the person who owns the building rather than by a developer. For anyone trying to identify buyers from planning data, it means the applicant name is a weak signal on its own and a strong one once filtered.

A caution we had to learn the expensive way. Searching a company register for an applicant name will nearly always return a match — Britain has over five million registered companies. We checked, and the matches were frequently the wrong entity: a dissolved software firm sharing a name with an industrial developer, a defunct restaurant, a surname that happens to be a company. A name match is not an identification.

The timing question, and why we cannot answer it properly

A permission granted last week is not useful to a subcontractor. The design is not finished, the main contractor is not appointed, and nobody is pricing anything. A permission granted a year ago is a different matter.

We tried to measure the lag by matching permissions to subsequent public contract awards at the same postcode. Across 21 matched pairs the median gap between permission and contract award was 9.1 months, with quartiles at 3.0 and 13.7.

Twenty-one pairs is an indication, not a measurement. The match is on postcode alone, so some pairs are the same site but different projects. And public contract awards are main contracts — the M&E package is let later still. If anything the true figure for subcontract packages sits further out than nine months. We publish it because a weak number stated honestly is more useful than a confident one, not because it settles anything.

What this is for

We build this dataset because we sell meetings to mechanical and electrical subcontractors, and it tells us which buyers near a given firm have work coming. Publishing it costs us nothing: the value was never in knowing the permissions existed, it was in the fortnight of work required to make them usable.

If you run an M&E firm and want to know what is approved within an hour of you, ask us and we will run your postcode.

Sources: local authority planning registers via PlanIt, individual council case pages, UK Contracts Finder and Find a Tender.
Figures current at 28 August 2026 · Skovgaard Systems · Earlier research: half of UK construction firms have no website