Method benchmark

Why these figures differ from raw EPC maps

EPC datasets are open, but the method changes the answer. We're independent — on no one's payroll, not affiliated with any installer or agent — so we can state these differences plainly and let you decide how to cite them. This page shows where the Home Index should be cited differently from official statistics or raw certificate-count maps.

Data vintage 2026-Q2 · methodology 0.3.0.

This index

18,149,673

retained domestic dwellings, deduped to latest certificate.

Retrofit gap

48.9%

of rated dwellings below EPC C.

Matched-sale layer

44.1%

national Land Registry match coverage.

Comparison table

ApproachWhat it is best forMain caveat
The Home IndexLatest-dwelling EPC exposure, local rankings, matched-sale exposure and portfolio triage.Covers rated dwellings; matched-sale signals are partial estimates and must be cited with coverage.
Raw EPC certificate mapsFast public summaries of all certificates in the register.Raw certificate counts can double-count homes with multiple certificates unless deduped.
ONS / official statisticsOfficial national and local statistical context.EPC coverage is incomplete because certificates are mainly triggered by construction, sale, conversion or letting.
Property-data platformsBroad property intelligence, comparables, planning and API workflows.Usually broader-scope property intelligence; EPC exposure as a distinct citable metric is rarely the primary focus.

Next: controlled C-or-above price/m² comparison

The current public spread is an observed matched-sale signal, not a causal premium. The next analysis layer will compare C-or-above and below-C homes within tighter cohorts: local area, property type, construction age, floor-area band and sale-year window, with sample size and coverage thresholds shown beside every result.