Four in ten repeat assessed homes recorded a higher energy band

Our analysis of the official Energy Performance of Buildings data found 7,656,746 homes with at least two usable certificate dates linked by the same property reference. The latest recorded energy band was higher for 3,161,142 homes, or 41.29% of the matched group.

Another 3,406,595 homes, or 44.49%, had the same band at their earliest and latest usable assessments. The latest band was lower for 1,089,009 homes, or 14.22%. The source was updated on 1 November 2026 and the usable assessment dates extend to 31 October.

Follow the recorded transitions

The energy changes explorer shows the first and latest band for each matched home as an aggregate. Choose a place and inspect a transition in the chart. Small groups are withheld rather than displayed as a precise result.

The national group includes 209 repeat assessed homes whose latest usable certificate has no area. They remain in the national result and are excluded from local comparisons. We excluded 52,075 properties with conflicting grades on the same certificate date.

A higher recorded band has several possible explanations

This comparison records assessment outcomes. It does not establish that a renovation caused the change. Assessment methods, the interval between certificates and the information available to an assessor can differ. The time between first and latest assessments also varies between homes.

Homes with repeated certificates are a selected group. Their results do not estimate the proportion of all homes that improved. A change within one energy band is also outside this comparison.

Heating and insulation assessments

We also compared the recorded assessments of main heating, walls and roofs at the same first and latest certificate dates. Both feature ratings were known for 7,602,978 heating pairs, 7,625,995 wall pairs and 5,930,339 roof pairs. The latest recorded assessment was higher for 14.62%, 20.70% and 32.77% of those respective groups.

Unknown or not applicable ratings remain outside each paired feature comparison. The counts with at least one unknown endpoint were 53,768 for heating, 30,751 for walls and 1,726,407 for roofs. These groups differ, so the percentages do not describe the same complete set of homes.

EPC feature ratings can use assumptions about a building. A higher recorded rating does not prove that new heating or insulation was installed. Download the feature comparison or inspect it in the interactive view.

Compare the current assessed stock

The separate home energy atlas uses the government annual tables published on 29 October. Those tables retain the latest valid assessment within a ten year window and provide a different view of energy performance. They should not be substituted for a matched comparison of the same homes.

The source is Energy Performance of Buildings Data. The analysis uses licensed non address fields and publishes aggregates. Download the supporting evidence.

The download was updated on 9 November to correct local authority names. Its evidence cutoff remains 8 November 2026.