A role for metered energy data in community-led lending

by Blog

This post summarises the potential role for metered energy data in community-led lending products, as being explored by Carbon Co-op and partners on a project funded under the Green Homes Finance Accelerator programme. 

What are Metered Energy Savings and why might they be useful to this kind of project?

Metered Energy Savings (MES) look at the actual metered energy use (metered gas and metered electricity) after retrofit measure(s) are installed, and compare it to what energy would have been consumed in that home during the post-retrofit period, had there not been a retrofit, i.e. a ‘counterfactual’ energy use.

Metered Energy Savings as a methodology originated in the United States, where it is now a requirement on large scale funding programmes, including those linked to the Inflation Reduction Act (IRA). 

This focus on actual metered energy use is very different to the estimates we currently use across programmes in the UK. Currently, schemes such as ECO, allocate savings from tables using a starting and finishing SAP rating band (differentiating between those in the lower part of a band, and those higher) and floor area segments. 

Establishing the SAP rating for a scheme like ECO only requires an EPC type assessment – this is based on a ‘Reduced’ version of the SAP methodology. Importantly, this is based on a number of assumptions about the home (such as when it was built), and limited input data. It only estimates energy use related to ‘fixed’ things such as the heating and hot water, ventilation systems, and fixed lighting. It does not take into account a household’s actual energy use before or after a retrofit – it is a ‘model’ that can be quite far removed from reality. 

Using a method like Metered Energy Savings (where metered data is available) could allow us to more accurately quantify the energy savings achieved in reality. For parties like One Stop Shops and contractors, if savings were much less than anticipated, it could act as a useful trigger for investigating issues with installation quality or occupant understanding and use of the systems. It’s potentially a much more honest and insightful metric, which chimes with a lot of the values we share across partners. 

What data do you need to run a Metered Energy Savings calculation?

At its most basic level, you can run a Metered Energy Saving calculation using: 

  • Smart meter data 
  • Post code (for location).

This is the version Carbon Co-op uses currently for its PowerShaper Tracker tool. 

The location is used to match weather data (external temperature). This is important, as we know that temperatures vary year-to-year, meaning that we very rarely use the same amount of energy to heat our homes as we did in the previous year. 

Using this baseline ‘pre-retrofit’ energy data, combined with weather data, the methodology fits a regression model (either using the ‘hourly’ or ‘daily’ model, depending on best fit). It generates a counterfactual for what the energy use would have been given the weather conditions.

Carbon Co-op’s PowerShaper Tracker currently reports the following metrics:

  • Pre and post retrofit Energy Use Intensity (EUI)
  • Post retrofit Carbon Intensity 

These are based purely on the observed smart meter usage data.

  • Normalised metered energy consumption (the ‘metered energy saving’), per fuel used in the home – by kWh, and as a % reduction or increase. 
  • Carbon savings, in kg C02e.

The tool also provides some comparative data: 

  • Pre-intervention consumption (the year before the retrofit)
  • Post-intervention consumption (from the first 365 days of the reporting period)
  • Postcode area average consumption (from publicly available datasets)
  • LSOA (local super output area) average consumption (from publicly available datasets).

Follow-on innovation projects (such as RetroMeter) are looking to develop the methodology to further reduce bias and increase accuracy. For example, by: 

  • Comparing a home to those similar, and who haven’t had a retrofit (this can help separate out the energy changes due to retrofit from the energy changes happening in society more broadly – COVID and working from home is a good example of this).
  • Using internal temperature data to account for the fact that householders often ‘take back’ some energy savings as they are now able to heat their home more than they could before. 

These two approaches aren’t ready to roll out just yet, but could help provide an even more robust energy saving metric for future lending.

In which situations can you run the calculation?

The availability of smart meter data is the most obvious requirement, and this isn’t always straightforward. This means that in the short to medium term, a metered energy savings metric will never be the ‘only show in town.’ Let’s think of it as a nice to have, or an enhanced metric. 

To train the model you need 12 months of good quality smart meter data from before the retrofit happened. We say ‘good quality’ because this isn’t always the case with smart meters (despite the promises!) We come across cases fairly regularly of smart meters not actually being smart (i.e. they don’t relay the readings automatically), and sometimes the data has lots of gaps in it. 

But 12 months doesn’t mean we have to wait a year before installing a retrofit measure, as the meters themselves save 13 months worth of historical data. So assuming a household has already had smart meters for over a year, once they consent to share their data, we can already access what’s needed. 

It’s most useful where: 

  • A home uses gas for heating, then makes a big(ish) change to the fabric (like adding insulation, or replacing windows and doors). The counterfactual can tell us how much gas the household would have consumed in the post-retrofit period, had the retrofit interventions not taken place.
  • A home uses gas for heating, then switches to electric heating (like a heat pump). The counterfactual gas usage can be compared with the actual electric heating consumption post retrofit, provided that sub-metered data for the electric heating is available. If we’re only interested in the total energy saved due to the heat pump and fabric retrofit, the comparison can be done on a simple energy basis.

This means it will be better suited where there is a more substantial change to a home. In general homes with solar PV or electric vehicles (EVs) can’t use this method, unless sub-metering is in place. The US working group is actively working on the PV problem, but EVs will remain a limitation. We understand around 20% of One Stop Shop clients (such as at PPR) would fall into this category. So again, this is one metric/tool at our disposal, but can’t be the only way to assess energy and carbon savings. 

How accurate is it? 

Carbon Co-op already has a tool which runs the basic methodology. This is called PowerShaper Tracker. This can be run for individual homes, or with several homes aggregated together. 

For an individual home using the basic (and currently available) methodology:

  • Accuracy of around 19% (accuracy means how much the reporting period predicted and metered gas consumption differ, in either direction) 
  • Bias of around 17% (bias means whether the reporting period predicted gas consumption is, on average, higher or lower than the metered consumption). 

Bundling homes together (e.g. looking at savings across a lending portfolio) improves accuracy and bias, as some of the ‘quirks’ you get on individual homes due to behaviour etc effectively ‘smooth out’ in a group. 

Further accuracy and bias improvements are possible when you match homes with comparable properties, but more development is needed before this option is available. 

Are there other reasons to encourage smart meter data access? 

Yes! Even if we can’t run a Metered Energy Saving calculation smart meter data can be very useful:

  • Where a retrofit assessment happens (e.g. Home Retrofit Planner assessment), it can streamline the collection of baseline energy consumption data – which aids context and understanding.
  • Householders can benefit from better access to their smart meter data via Carbon Co-op’s PowerShaper Monitor platform – they can visualise, download and interrogate it, free from the restrictions of an In Home Display or how their energy supplier provides it. They can also see the carbon emissions associated with their energy use. It can simply be offered as a tool that helps people to engage with their energy data, encouraging ‘energy and carbon literacy.’ 
  • Providing a simple ‘before and after’ energy comparison, whilst heavily caveated, is arguably better than nothing. 

What about people that don’t have smart meters yet?

There are still lots of these households. Where we can’t access smart meter data, we can encourage householders to provide their energy use based on the figures given in their bills. Whilst these are static figures and don’t take into account external temperature and other factors, they are useful for ‘at a glance’ before and after comparisons. 

Communication is key

The key thing for all of the options (Metered Energy Savings / smart meter data / bill data) is clearly communicating the limits of what they tell us. This will be considered as part of the reporting framework we develop as part of this project.