Free survey and a written price before any work starts. Book a free survey.

How we get our loft insulation figures

Every insulation figure on this site comes from one public dataset, the Energy Performance of Buildings register. We worked the figures out ourselves. This page is the method, so you can check it. It is written for anyone who wants to quote a number from here and needs to know what it counts. The council figures cover 248,651 houses and bungalows in six council areas, of which 168,519 have a loft depth on record.

We read the register on 16 September 2026. The certificates in it were issued between 14 October 2004 and 9 September 2026.

Source: Energy Performance of Buildings Register, bulk domestic download, analysed September 2026. Depths are as recorded by the assessor, not as measured by us. Licensing: the register's own licensing restrictions, read 22 September 2026.

How a figure on this site is made

Six steps, in this order.

  1. 1

    Download the whole register

    The government publishes every domestic Energy Performance Certificate as a bulk download. The extract we hold carries 3,896,340 certificates across 54 councils.

  2. 2

    Keep the six councils in Cambridgeshire and Peterborough

    Cambridge, South Cambridgeshire, East Cambridgeshire, Huntingdonshire, Fenland and Peterborough. That leaves 451,464 certificates. The council figures on this site describe those six. Five towns we work in sit outside them: Royston in North Hertfordshire, Newmarket, Haverhill and Bury St Edmunds in West Suffolk, and Saffron Walden in Uttlesford. For those we counted the homes in the town's own parish, in the same way, and show them on their own pages only: 4,633 homes in Royston, 4,620 in Newmarket, 7,637 in Haverhill, 12,218 in Bury St Edmunds and 4,573 in Saffron Walden.

  3. 3

    Turn certificates into homes

    A house that has been sold three times has three certificates. Counting certificates would count that house three times, so we keep one row per address and take its most recent certificate. We match on the property's unique reference number (UPRN) where the register has one, and otherwise on postcode, floor area and roof description together.

  4. 4

    Keep houses and bungalows

    Flats and maisonettes are dropped. That leaves 248,651 homes. That is the number of houses and bungalows we read. Shares are worked out from the 168,519 of them with a recorded depth.

  5. 5

    Read the depth out of the assessor's own words

    The register does not hold a tidy number. It holds the assessor's sentence, such as Pitched, 270 mm loft insulation, and we take the figure out of it. Where the assessor found no insulation we record nought. Where the sentence does not give a depth at all, the home has no depth, and we say so rather than guessing one. Where it says (assumed), the assessor could not inspect the loft and the certificate filled in a default from the age of the house. That is not a reading, so we do not count it as one.

  6. 6

    Only then work out the share

    Every percentage on this site was worked out from a counted number of homes. We will not publish a share from fewer than 30 homes. Below that you get the counts and no percentage, because a percentage of a handful of homes reads as authoritative and is not.

What we leave out, and why

None of these exclusions makes lofts look worse than the register shows.

  • 80,132 homes with no depth written down. They are counted as homes and left out of every depth figure. They are not assumed to be badly insulated, and they are not assumed to be well insulated either.
  • 8,978 lofts the certificate assumes have no insulation. The assessor could not inspect them, and the certificate writes "no insulation (assumed)" for an older house. No assessor saw those lofts, so they are in the group above with no depth. Counting them as empty would have made lofts look worse than the register can support.
  • 5,920 homes with a room in the roof. There is no loft to measure and the insulation sits in the slopes.
  • 2,057 homes with a flat roof. No loft at all.
  • 570 thatched homes. A specialist trade, and not comparable to a tiled pitched roof.
  • Any group of fewer than 30 homes. Counts only, never a share.

We also do not publish anything at postcode or street level. The address fields in this register are licensed separately from the rest of it, and a single home's energy data is personal data. Our area figures are aggregates of a town or a council, which is a different thing. The loft checker is different again. It shows one home's certificate to the person who looks it up, and does not store it.

67.8% of the 248,651 homes we count have a loft depth written on the certificate by an assessor. Every share on this site is worked out from those 168,519 homes, never from all 248,651.

  • Depth recorded168,519
  • No depth recorded80,132

What the words on our town pages mean

  • A typical loft. The middle one. Line every measured loft in the town up from thinnest to deepest and take the one halfway along. Assessors write depths down in steps, such as 100mm, 150mm, 200mm and 250mm, so the typical depth is one of those steps.
  • What you would expect for the age of the homes. We take the average depth for homes of each age across the six councils. Then we work out what a town would average if its homes matched that, given when they were built. A town below that figure has thinner lofts than homes of the same age elsewhere.
  • Measured twice. Homes with two or more certificates where an assessor recorded a depth each time. We compare the first with the latest. The register does not say what happened in between.
  • Told to insulate. The latest certificate for the home carries the recommendation "Increase loft insulation to 270 mm".
  • New homes. A certificate for a newly built home gives a heat loss figure for the roof and no depth. Those homes count in how old a town's housing is. They are not in the depth figures.

What this data cannot tell you

Four limits.

NOT A CENSUS

A home only has a certificate if it has been sold, rented out or newly built. The certificates we read were issued between 14 October 2004 and 9 September 2026. Homes that have stayed in the same hands for thirty years are missing, and those are not a random sample of the housing stock.

AS RECORDED, NOT AS IT IS NOW

The depth is what an assessor wrote on the day, which may be years ago. On 27,797 of these homes the assessor's own wording says the roof was assumed rather than inspected.

DIFFERENT HOUSES

Certificates issued in 2026 are not the same homes as certificates issued in 2014, so a change between two years is not the same homes improving. Where we wanted to ask whether homes improved, we compared homes against themselves.

NOT YOUR HOUSE

A town figure describes a town. It says nothing about one address, and it is not a prediction about yours. The checker reads your own certificate instead.

All non-address fields are published under the Open Government Licence v3.0. The address fields are not, and carry Ordnance Survey and Royal Mail rights.

Energy Performance of Buildings data: licensing restrictions, read 22 September 2026

Questions people ask about these figures

Can I check any of this myself?

Yes, and that is the point of publishing the method. The register is a free public download from the service linked above. Take the six councils named in step two, keep one certificate per address, keep houses and bungalows, and read the depth out of the roof description. You should land on 248,651 homes.

One of your numbers is slightly different in two places. Why?

Because we will not quietly change a published figure to make it tidy. The count of homes formally told to increase their insulation to 270mm reads 29,189 on our pages. Re-running the same work from a later copy of the same data gives 29,207, which is 18 more, a difference of 0.06%. We publish the lower, original figure until we have established which copy is authoritative, and we would rather tell you the gap exists.

How do you work out what loft insulation does to an energy rating?

We take homes that have more than one certificate on the register, lodged between 2008 and September 2026, and compare the first with the latest. Where the loft went from 100mm or less to 270mm or more, and the recorded wording for the walls, the main heating and the windows is the same on both, there are 2,149 homes and the rating rose by a typical 8 points. In 10,254 homes where the loft depth and those three things were all unchanged, it rose by 2. In the first group 58.7% moved up at least one band, against 33.9% in the second. Of the homes that started in band D, 57.6% in the first group ended in a higher band, against 42.8% in the second. Where the loft went from between 150mm and 200mm up to 270mm or more, with the same three things unchanged, there are 2,249 homes and the typical rise was 4 points. It is a comparison of records, not a controlled test: lighting, hot water and the assessor can differ, and the method itself changed in June 2025.

What does a loft recorded as "assumed" mean?

It means the certificate does not record a measured depth. 23,072 houses and bungalows here have a pitched roof described on their most recent certificate with the word "assumed", as insulated, as having limited insulation, or as having none. They are part of the 27,797 roofs of every kind described that way; the rest are flat roofs and rooms in the roof. Of the 23,072, 172 carry the recommendation to increase loft insulation to 270mm, which is fewer than 1 in 100.

How much loft insulation do newer homes here have?

Of 44,442 houses and bungalows built from 2012, only 2,709 have a loft depth written on their certificate, which is 6.1%. A new home's certificate usually records a heat loss figure for the roof instead. Among the 2,709, the typical depth is 300mm. They are not a random sample of new homes.

How many homes here are rated D to G?

115,323 of the 248,651 houses and bungalows, on each home's most recent certificate. Some of those certificates are years old, so a home may have been improved since. It is a count of ratings and says nothing about who lives there or what they earn.

Do you keep my postcode when I use the loft checker?

No. The checker looks the address up and shows you the answer. The postcode and the address are not logged and not stored, and there is no analytics of any kind on that page.

How often is this updated?

The register is updated by the government through the year. We last read it on 16 September 2026. We update the figures when we next read the register.

What does your own certificate say about your loft?

The checker reads your certificate back to you. If you would rather have the loft measured, the survey is free. Book a free survey.