evenEco Commercial
Commercial›Methodology
Transparency

How we produce the data on this site

Every figure here comes from the public register. This page says exactly how we get from the raw extract to the numbers on an area page, including the decisions that could reasonably have been made differently.

The source

The GOV.UK Energy Performance of Buildings Register bulk extract, published under the Open Government Licence v3.0. We hold the full non-domestic dataset locally, which is why lookups on this site return instantly rather than depending on a third-party API.

At the date of this analysis the extract contains 1,297,123 non-domestic certificates.

Deduplication – certificates to properties

A building can hold many certificates over time. Counting certificates would overstate the stock and double-count buildings that have been re-assessed, so we reduce to one row per property, keeping the most recent certificate.

Properties are identified by UPRN where one is recorded (about 90% of rows), falling back to the recorded address plus postcode. Where two certificates share a key, the one with the later lodgement date wins.

That reduces 1,297,123 certificates to 998,164 unique commercial properties.

Where this could be done differently The address-plus-postcode fallback is imperfect. An assessor who typed the address differently on two visits produces two rows for one building, which slightly overstates the property count. We prefer that to the alternative – collapsing genuinely separate units that share an address – because on a register dominated by multi-let commercial buildings, over-merging would understate the problem materially.

Scope

  • England and Wales only. MEES is an England and Wales regime; Scotland and Northern Ireland operate separately. Rows without a local authority code, and the handful of Scottish rows in the extract, are excluded from area pages.
  • Non-domestic certificates only. Display Energy Certificates and domestic certificates are excluded entirely.
  • 319 local authorities have non-domestic certificates recorded against them.

Definitions used throughout

Definitions
Term as we use itDefinition
Below EPC BCurrent band C, D, E, F or G
Below EPC ECurrent band F or G – below the minimum that has applied to lettings since April 2023
The 2031 cohortFloor area over 1,000 m² and band below B
ExpiredCertificate expiry earlier than the analysis date of 11 September 2026. The published register carries no expiry column, so we calculate it as ten years from the lodgement date
National rankPosition among the 319 local authorities by share of stock below EPC B, rank 1 being the highest share
Average floor areaMean of the recorded total floor area across the deduplicated set

Sector normalisation

The register carries both pre-2020 and post-2020 use class naming for the same categories – "A1/A2 Retail and Financial/Professional services" and "Retail/Financial and Professional Services" are the same thing recorded under two conventions. Reporting them separately would split every retail figure roughly in half.

We normalise 45 recorded values into 24 sectors by matching on the substantive part of the label. Anything unmatched falls into "Other non-domestic" rather than being dropped.

What this data cannot tell you

  • Whether a building is let. MEES bites on letting, and the register records no tenure. A sub-E building may be owner-occupied and entirely lawful.
  • Whether an exemption is registered. The Exemptions Register is separate.
  • Whether the rating is right. A certificate records what was evidenced on the day. See why your score is what it is.
  • Anything about buildings with no certificate at all – which, for a building never sold or let since 2008, may be a substantial and invisible population.
We publish the caveats because they are the interesting part Anyone can put a percentage on a page. The question a journalist, a lender or an asset manager should ask is what the number excludes – so we have written it down rather than waiting to be asked.

Updates and corrections

The analysis is refreshed when a new bulk extract is published. Each page carries the date of the analysis it was built from. Corrections policy is on the about page; if you think a figure is wrong, tell us and we will check it and publish the correction.

Written and reviewed by the evenEco assessment team Last updated Figures from the GOV.UK Energy Performance of Buildings Register, non-domestic extract for England & Wales, analysed 11 September 2026 How we produce this