Home energy ratings, honestly: what the EPC record shows and how to read yours

There is a number stapled to almost every home in England and Wales, and most people glance at it once — during a purchase, buried in a solicitor's pack — and never think about it again. It is a single letter, A to G, on the Energy Performance Certificate. It looks like a school report. It is treated like a verdict.

It is neither. The EPC band is a model of how a building should perform, not a meter reading of how it does. Understanding the difference is the whole game — because that gap is where the money is, and where the marketing isn't.

We analysed the full England and Wales register: 22,518,549 certificates. Here is what it says, read plainly, and here is how to read the one certificate that matters to you.

What the band actually measures

An EPC is produced by an assessor walking a property and feeding its dimensions, construction, glazing, heating and insulation into a standardised model called RdSAP. The model returns a score from 1 to 100 — the SAP score — and that score is bucketed into a band: A (92+), B (81–91), C (69–80), D (55–68), E (39–54), F (21–38), G (1–20). The EPC band you see is just that bucket.

Two things follow immediately, and both matter.

First, the score is modelled, not metered. It estimates the cost to heat and light the home under standard occupancy — a notional family, notional thermostat settings — not your bills. A frugal household in a band-D house can spend less than a profligate one in a band-B flat. The certificate is a property fact, not a household fact.

Second, every certificate carries two scores: the current rating and a potential rating, the band the assessor thinks the home could reach if you did the recommended work. Almost nobody reads the second number. It is the most useful figure on the page.

The national picture: we are a nation of Ds

Run the whole register by band and the shape is unambiguous:

Band D is the single most common rating in the country — more than a third of all certificates. Add E, F and G and you find that more than half of homes (54.9%) sit below a C. At the other end, fewer than one in eight reach A or B, and genuine A-rated homes are a rounding error: 0.6%.

So when a listing brags "EPC C", it is describing a home that is, at best, average. And when it says "EPC D", it is describing the most ordinary home in Britain. The band is a poor bragging right. It is a much better to-do list.

The gap nobody reads: potential vs current

This is the figure that reframes everything. Across the 20.8 million certificates that record both scores:

• The average current score is 64.8 — a middling D.
• The average potential score is 79.6 — a solid C, nearly a B.
• That is an average headroom of roughly 15 SAP points sitting unclaimed on the typical certificate.

Push on the homes that are actually underperforming and it gets starker. Of the 11.6 million certificates currently rated below a C, 81.6% — nearly 9.5 million homes — could reach C or better using only the improvements their own certificate already lists. Loft insulation, cavity fill, a hot-water cylinder jacket, better heating controls, in many cases a heat pump. Not speculative work. The measures the assessor already wrote down.

The energy conversation in this country is usually framed as what a home is. The data says the more honest frame is what a home could be, and hasn't bothered to become yet. If you own one of those 9.5 million homes, the certificate is quietly telling you where two bands of value are hiding. The playbook for closing that gap is retrofit — and the order you do the work in matters more than most homeowners realise.

Age tells you more than the band does

If you want to predict a home's energy performance before you've seen its certificate, ask one question: when was it built? Group the register by construction age and the gradient is almost perfectly monotonic:

• Before 1900: average 51.2 — an E.
• 1900–1929: 56.8 — low D.
• 1930–1949: 59.5 — the classic cold semi, mid-D.
• 1950–1966: 62.7.
• 1976–1982: 66.3.
• 1996–2002: 69.5 — the point where the average home finally crosses into C.
• 2012 onward: around 78 — a comfortable C.

A Victorian terrace and a new-build estate house are, on average, 27 SAP points and two whole bands apart before you account for anything the current owner has done. This is why the "period charm vs. new-build box" argument is never only about aesthetics — it's a running-cost decision dressed as a taste decision. We break the full head-to-head down in the studies section, and the same age gradient is why a beautiful pre-1919 home can be the most expensive one on the street to keep warm.

None of this means old homes are doomed. It means they start further back, and the recommended-measures list on a period home is usually longer, more expensive, and more constrained — solid walls can't be cavity-filled, and conservation rules can veto external insulation and the obvious solar PV roof.

The off-grid penalty

There is one more split in the data worth internalising before you buy, and it is about fuel, not fabric. Average current score by main heating fuel:

• Mains gas: 64.7 — the national baseline.
• Electricity: 58.6.
• Oil: 51.2 — an E.
• LPG: 42.0 — deep into E, edging toward F.

Homes off the gas grid — the oil tank in the garden, the LPG cylinder round the back — sit a band or more below their mains-gas equivalents on the register, and their running costs behave accordingly. If you're viewing a rural property, "no mains gas" is not a footnote; it's one of the biggest energy facts about the place. It's also the strongest case for a heat pump, which turns that off-grid liability into the cheapest-to-run heating a rural home can have — the maths of which we work through in "is a heat pump worth it?".

How to read your home's certificate

Enough about the nation. Here is how to read the one certificate that matters, in the order that actually helps:

1. Read the potential score first, then the current one. The gap between them is your opportunity budget. A D that could be a B is a very different proposition from a D that tops out at a low C.
2. Read the recommended-measures table, not the letter. It's ordered roughly by value. "Increase loft insulation" and "hot water cylinder insulation" are cheap and near-certain; "external wall insulation" is expensive and often planning-constrained. The measures tell you the character of the work, not just the score.
3. Check the age band and the wall type. Solid walls (typically pre-1919) rule out cavity insulation and change everything downstream. Our note on improving an EPC rating walks through which measures move the number and which barely nudge it.
4. Sanity-check the fuel. Oil or LPG heating? Assume higher running costs than the band alone implies, and price a heat pump into your thinking early.
5. Treat the band as a floor, not a ceiling. More than 8 in 10 sub-C homes can reach a C on paper. The question is never "is this a good band?" — it's "how far, how expensively, and how constrained is the path to a better one?"

You can pull the certificate, construction age, wall type and recommended measures for any address — alongside its flood, planning and price history — straight from the record on a property page, or just ask about a specific home in plain English.

What the EPC will not tell you

Honesty is the point of this whole exercise, so here is what the certificate can't do, no matter how carefully you read it:

• It won't tell you your bills. It models a standard household, not yours. Metered consumption routinely diverges from the modelled figure in both directions.
• It won't catch damp, draughts or a failing boiler that a survey would. It is a desktop-ish model, not an inspection.
• It ages. A certificate is valid for ten years, and a home improved (or degraded) since lodgement carries a stale letter until it's reassessed.
• Assessors disagree. Two RdSAP assessments of the same home can land in different bands at the margins — a genuine reliability limit near band boundaries, not a scandal.

An EPC is a useful, free, standardised first read of a home's energy story. It is not the last word, and anyone who treats a single letter as gospel — buyer, seller or agent — is over-reading it. We'd rather say where the record stops than sell you certainty it can't support.

Explore the energy topic

This hub sits over a set of shorter, single-question pages. Start here:

• What does an EPC band actually mean? — the bands, the score, and the difference between them.
• How do I improve an EPC rating? — which measures move the number, in cost order.
• Is a heat pump worth it? — the running-cost case, especially off the gas grid.
• Do solar panels add value? — what solar PV does to the band, the bill and the sale price.
• Glossary: EPC band · SAP score · retrofit · heat pump · U-value

And two deeper reads from the blog: the EPC is lying to you (a little) — how two near-identical homes land a band apart — and new build vs. 1930s semi, where we ran both archetypes head-to-head on our own energy and sold-price data.

Methodology

Source and scope. The England and Wales Energy Performance Certificate register: 22,518,549 certificates, analysed on 18 July 2026.

• Band distribution: percentages use all valid A–G certificates. Ninety-two non-standard or blank ratings are excluded from the A–G narrative.
• Current vs potential gap: 20,776,504 certificates carry both scores. Their averages are 64.8 current and 79.6 potential.
• Reachability: of 11,604,194 certificates currently below C, 9,469,080 (81.6%) have a potential score at C or above.
• Age gradient: the quoted figures use the register's standard national construction-age cohorts and exclude one-off, free-text and unusable age values.
• Fuel: legacy duplicate labels were consolidated into their underlying fuel type before comparison.

Limitations, stated plainly. These are certificate counts, not unique-dwelling counts — a home reassessed several times appears several times, slightly over-weighting recently transacted and improved stock. The register covers England and Wales only; Scotland runs a separate register. About 1.46 million certificates carry no usable construction age and are excluded from the age figures. Every SAP score is a modelled RdSAP estimate, not measured consumption.