EPC reform… how ‘sensing’ can unlock a world of benefits

By Steven Heath, Technical Director
March 29, 2023

A ‘wide ranging review’ of the UK Energy Performance Certificate (EPC) is to kick off imminently. Much needed, but what needs to happen?

First up – accept a single core principle: we now live in a ‘sensed world’. The average car has 50+ sensors, a phone 14 and a plane 10,000. Emergency hospital entry will see your heart & oxygen levels tracked while sensors will go to work divining signals hidden in various bodily fluids. Granular feedback from sensors & data processing helps make sense of the modern world.

The home EPC currently senses nothing. Smart meters sense energy flows into the home but nothing senses the speed at which heat leaks out or how well heating systems convert gas or electricity to heat. As useful feedback goes, imagine a Doctor waving a doll at a critical care patient asking ‘where does it hurt’? Not completely useless, but hardly deploying all available tools.  

So what energy bill & energy security benefits can sensing homes offer?

The power of sensing & measurement

Through the deployment of sensors collecting real world data, we measured the pre and post retrofit efficiency – how quickly heat leaked out – in 12 homes in Manchester. The retrofit was ‘light’ – taking out existing insulation to re-install with attention on quality. The details are here but key was the works barely changed the dial on the EPC home efficiency rating. The EPC didn’t recognise the improved quality of the replacement measures while the desktop software (SAP) underpinning the EPC didn’t value the only ‘new’ insulation measure – party wall insulation – as having an impact. In short, the EPC said our works wouldn’t materially impact home running costs or comfort.

Before the results, a short description on what’s being sensed. Raw data collected via small sensor boxes is worked on by an algorithm to create multiple real performance metrics: one being kWh pa heat requirement in averaged conditions. We can measure the efficiency of the home and then plug a ‘basket of average weather, thermostat & heating system timer settings’ through that model to give a kWh heat demand over a year. Very similar to an EPC space heating metric – but accurate, reliable, and trustworthy (desired aims in the EPC Acton Plan).  Other metrics can & must flow, but more on that later.

Benefit 1 – Measuring actual home efficiency

Our Manchester light retrofits saw an average 31% drop in heat demand to achieve those standardised comfort settings. The best performing home needed 51% less while the worst performer still saw a 21% improvement. What do these savings figures mean for heating? If we model a gas boiler & energy unit cost at the numbers below, we get the following annual savings derived from retrofit.

Boiler efficiency


Gas unit cost (kWh)



So, ignoring inflation – or hoped for deflation – a £268 a year averaged energy bill saving across the 12 homes.  Or £2,680 over the 10-year validity of the EPC. That’s a collective savings for the occupants of £32,160.

Had the retrofit need for these homes been judged by the current EPC, or the Government retrofit schemes reliant on it, these works would not have taken place. The EPC would have said ‘there are no easy fabric savings to find in these homes! Look elsewhere.’

Benefit 1 – sensing fabric efficiency in the EPC, & key policy tools, tells you which homes don’t need improvement & which do. It also gives an indication by how much they might be improved pre-retrofit and a concrete measurement of how much they have been improved post-retrofit.

But these homes weren’t heated with gas…

Benefit 2 – unmasking heat pump and fabric efficiency dependence

“Heat pumps won’t work in old homes, warns Bosch” - Daily Telegraph, March 14th

“Heat pumps work': Study confirms Air Source Heat Pumps three times more efficient than gas boilers” – Business Green, March 16th

The power of a robust fabric efficiency measurement is to make a known unknown, known. Energy Systems Catapult’s recent heat pump study on 290+ homes mostly replicated previous studies on heat pump efficiency. Some occupants realised the magic trick of 3.5 units of heat for each unit of electricity over the year. This efficiency level will see occupant bills come back down in line with gas boiler running costs. Heat pumps can ‘work’. But other homes in the study, the majority, did not see this efficiency, and the study is not clear on why the differing experience:

‘…whilst the average performance has improved, performance variation remains high…. generally this variation is difficult to explain’

And this is where our sensed world can ride to the rescue. There are only 4 credible suspects on what’s causing that ‘difficult to explain’ variation. The heat pump (product or install), the heat distribution set up, the occupant and the home fabric. The culprit will be one, or a combination, of these 4. The study itself sensed all it could on the heat pump and distribution system but they did not sense fabric efficiency. Being able to measure 3 of the 4 key parameters and tune them correctly should be able to help occupants realise that higher efficiency level of 3.5 units of heat or higher.

Returning to our 12 homes in Manchester, what might be the heating cost in three different heat pump scenarios? The first, where the homes weren’t retrofitted – because the EPC said no – and received heat pumps that were then only able to realise the lower efficiency threshold in the study - 2 units of heat per unit of electricity (200% efficiency). The second, where we model our retrofitted homes and their lower heat demand with the same 200% heat pump efficiency (the red bar in the graph). Finally, the third bar shows the heat pump efficiency figure improved to 350% as now ‘less work’ is asked of it to achieve desired temperatures as its supported by a newly efficient fabric. The improved fabric efficiency enables an improved heat pump efficiency.


Heat pump base case efficiency


Heat pump efficiency improvement linked to retrofit


Electricity unit cost (kWh)



This graph shows a dramatic drop in home heating costs. The crucial multiplier here is not just the lower energy demand driven by the more efficient fabric, but the modelled improvement to heat pump efficiency linked to home fabric efficiency improvement. While the fabric measurements in each home are real measured numbers, these homes weren’t heated with heat pumps.

We don’t know the impact robust fabric measurement can have on reliably improving heat pump efficiency. Neither does the Catapult study. No-one’s yet described the link at scale. Yet heat pump and fabric clearly work collectively as a ‘whole system’; without understanding one we can’t understand the performance of the other. At the heart of the problem this EPC review must solve, is the current desktop calculated EPC would have described predicted heating demand in all three scenarios as the same. It wouldn’t have recognised the improvement offered by the fabric, and it would have plugged in a generic, unmeasured, ‘un-sensed’, heat pump efficiency number.

Whether I’m exaggerating an uplift or underplaying it, something is driving heat pump efficiency variation in the Catapult study between the desired 350%+ and the unwanted 200% figures. Sensing all key parameters including the fabric allows us to chase down the relationship. Once the relationship is understood, the supply chain can shift to assuring heat pump performance nearer the desired 350%. The EPC, and all policy it underpins, has to drive the market to solve these questions if we want large scale heat electrification.

Benefit 2 - sensing fabric efficiency in the EPC, & key policy tools, will flush out the relationship between heat pump sizing, operation and efficiency and fabric efficiency. Divining that relationship relies on driving sensed measurement at scale. Once the relationship is clear, householders can be offered guarantees around both fabric and heat pump operation; the much discussed ‘heat as a service’ better expressed as ‘warm, affordable homes as a service’.

Benefit 3 – home flexibility is next

I mentioned multiple ‘real performance’ metrics can flow from sensed data. A key final parameter to learn if we are to offer genuinely low bills is home ‘heat up and cool down’ times in different weather. Once done, a smart algorithm can manage the heating system in line with occupant comfort settings but also half-hourly electricity grid price signals.

A crude model below considers our measured un-retrofitted house scenario with an inefficient heat pump (200%) vs our measured retrofitted home with newly efficient heat pump (modelled at 350%) but also where 50% of the home heating is shifted to an off-peak kWh price.

Heating efficiency COP / SPF


Heat efficiency improvement linked to retrofit


Heat peak unit cost


Heat off peak unit cost



Benefit 3 – once the first two benefits are realised, the final ‘sensed’ benefit is smart home operation. Much like the heat pump & fabric efficiency interdependency, we can’t know exactly what heat load can be shifting in what home in what external temperatures. But we know the grid could benefit hugely from peak heat load shifting as we move to rely on intermittent renewable energy. To understand & ultimately manage this relationship, we must deploy sensors at scale.

Sensing the world allows control

Our Manchester case study is a robust in-use fabric efficiency performance measurement for 12 homes before and after retrofit. We use this sensed fabric performance pre-and post-retrofit as a platform to model different heating system efficiencies taken from the large-scale Energy Systems Catapult study.  The modelling shows a virtuous cycle of beneficial multiplier effects to reduce heating costs are possible, but only sensing can make these a reality. It is for the EPC review to allow these ‘bolt-on’ sensed measurements to inform revitalised EPC metrics.

How? The laboured smart meter roll-out experience says incentivise, validate, & accredit commercial metrics rather than allow policy to dictate the ‘right’ metric. Government should focus on validating & auditing the outputs of the sensed world, not attempt to extensively police the inputs to derive a metric it thinks is the right one. 

In our real-world case study, the whole estate had direct electric heating. More expensive than any scenario considered above. Occupants faced huge bills to achieve standardised comfort settings prior to the retrofit, becoming more affordable but still expensive post-retrofit.

To note, while we only had time to pre-retrofit measure 12 homes on the estate, we managed to retrofit and then measure fabric efficiency for 31 estate homes. All showed significant improvement relative to a similar architype where a pre-retrofit measurement existed. This estate at least was provided a buffer against recent energy-price inflation, but it was not thanks to the current EPC design. This review needs to deliver a modular EPC that stops guessing and starts sensing.