EPCs only tell part of the story, smart meter data tells the other

If you are working in social housing or retrofit, you already know that EPCs have limitations. It helps to be specific about those limitations when making decisions about real properties.

What EPCs are and what they are not

An EPC is legally required to have in England, Wales and Scotland every time a property is bult, sold or rented. It rates a building's energy costs on a scale from A to G and is valid for ten years.

The certificate is based on a standardised model and uses data about a building’s construction, insulation, heating system and other physical features to predict its energy efficiency based on standard assumptions (such as average occupancy, typical heating behaviour, etc). Standard assumptions are used so that buildings can be compared to each other, but don’t reflect the infinite variations in the way real people live in their homes.

They can be a useful benchmark for understanding a property but are just predictions and not based on actual measurements.

UK housing 🏘️

According to the Office for National Statistics, homes in England now have a median EPC score at the lower end of band C, but analysis of government EPC data published in April 2026 found that 55% of all properties in England and Wales still sit at band D or below.

The government has set a deadline for all privately rented properties to reach band C by October 2030, with fines of up to £30,000 per property if this deadline is not met. This means there is a significant amount of work to be done across many older, hard to improve properties!

The performance gap 📊

Research published in the Energy and Buildings journal (led by UCL’s Smart Energy Research Lab) comparing EPC predictions with real smart meter data, across 1,374 gas-heated homes, found that EPC’s consistently overestimate energy usage for properties rated C to G. For Band C homes the gap was 8%, for F and G bands EPCs were over predicting by up to 48%.

Researchers at UCL were able to rule out occupant behaviour as the cause for these overestimations, as the gap persisted even in homes where modelled assumptions around occupancy, thermostat setting and heating patterns matched actual occupant behaviour.

Real-world heat loss measurements back this up. Research by our partners at Build Test Solutions revealed that EPC heat loss predictions were flat-out wrong for 57% of homes, with an average inaccuracy of 30%. In fact, nearly one in five homes had errors wider than 50%.

Physical tests on airtightness and insulation across walls, floors, and roofs reveal exactly why this happens. In the world of EPCs, two walls built with similar materials are assumed to perform identically - in the real world, they simply don’t. Due to unknown materials, hidden air gaps, and wear-and-tear; it is common to see performance variations of 50% or even 100% between properties that look identical.

Simply put, EPCs are a guessing game. A property flagged as inefficient' might actually be a solid heat-retainer, while a home with a glowing EPC rating could be harbouring major, hidden thermal defects

Inaccurate assumptions on energy use are contributing to inaccurate measurements of home efficiency. This means a property flagged as “inefficient” according to its EPC may actually be using significantly less energy than the certificate suggests. On the other hand, a home that has a high EPC score may have real performance issues that EPCs cannot pick up.

What smart meter data changes 🎯

Smart meter data gives you a view of actual energy consumption patterns at half hour granularity for electricity and gas, not a modelled prediction. And when you do need to model, to determine carbon and cost savings, you can build that model on facts not assumption.

When combined with technology such as indoor environmental sensors and/or algorithms for determining home efficiency scores*, this data becomes even more useful. You can see not just how much energy is being consumed, but whether that energy is translating into a warm, comfortable home that is effective at retaining heat.

Two properties can run at similar consumption levels, but have very different indoor temperatures, telling completely different stories of home efficiency and resident comfort – exactly the kind of insight a certificate can’t give you.

For housing providers and retrofit teams, having access to half hourly smart meter data, as well as historical data can support:

• Identifying properties that are underperforming before major issues arise and choosing the priority properties for intervention
• Build strong evidence for retrofit investment decisions
• Get real quality assurance on retrofit improvements to make sure they deliver value for money based on immutable data
• Protect vulnerable tenants: Spot hidden, underperforming properties where residents are at a high risk of fuel poverty, underheating, damp, and mould

EPCs aren’t going away and can be useful step in understanding a home, but it is important to understand their strengths and limitations.

At Glow, we work with housing providers, retrofit teams and researchers to help them understand the truth of what's happening across their properties using smart meter data through our Glow Data Service.

We've developed a range of ways to streamline the energy consumer verification and consent process, helping you quickly onboard customers with ease. Access to smart meter data has never been easier!

We also offer a range of hardware products, such as our Consumer Access Devices (CADs) for real-time data access, and our temperature and humidity sensors for monitoring environmental conditions in a home.