Conceptualising domestic energy service business models: a typology and policy recommendations

Energy service business models (ESBMs) are potentially critical to reducing household energy demand and mitigating climate change. These models are predicated on a shift from the ‘throughput’ sale of energy commodities, towards providing ‘useful’ or ‘final’ energy services. However, the conceptual delineation of these models and their different variants remains opaque in the literature. In this paper, we seek to bring clarification to this issue through the identification of a typology of ESBMs. Through a series of 53 interviews and 7 stakeholder workshops we explore contemporary domestic ESBM examples in Europe. We find that while more basic energy supply contracts are commonplace, models which deliver energy saving performance or final energy services are rarer. We subsequently identify barriers to the adoption of these business models, before proposing 13 policy recommendations. We conclude that the ‘energy throughput orthodoxy’ which has governed liberalised energy markets will need to be challenged for these models to have a significant future impact.

1. How are domestic ESBMs conceptualised in terms of their scope and depth? 2. How might ESBMs incentivise the decarbonisation of housing more effectively than the traditional utility model? 3. What policy framework would be conducive to domestic ESBMs?
We address these questions through a review of previous literature (Section 2) and primary data collected via 53 interviews and 7 stakeholder workshops, as outlined in our methodology (Section 3). Based on previous literature and our data, we then develop a typology of ESBMs and explore emerging residential examples in Section 4, before discussing their potential for the key decarbonisation challenges in the residential sector in Section 5. We conclude in Section 6 by providing 13 policy recommendations to promote the increased adoption of ESBMs in the UK and EU member states.
Instead, ESBMs provide 'useful' energy services like hot water, coolant, or the 'final' energy derivedsuch as illumination, or room temperatures. These models shift the responsibility for the performance of equipment or a building into long-term contracts between an ESCO and the household or business . Solar-as-a-service models, for example, help households to become prosumers without the upfront cost. Here, the ESCO leases PV panels and takes responsibility for finance, installation and maintenance: offering a solar tariff and dealing with export agreements (Overholm, 2015). Heat-as-a-service (HaaS) models sign-up consumers to a comfort agreement and can operate with district heat network or by installing heat pumps, often with ESCOs owning or leasing the infrastructure (Hannon and Bolton, 2015). These models may also offer energy performance contracts based on measured and verified energy savings, further incentivising efficiency in building fabric, lighting and appliances (Sorrell, 2007). Sorrell (2007) and Steinberger et al.'s (2009) seminal works on the economics of energy service contracts, and the potential of an energy performance-based economy, provide a theoretical foundation for the study of these business models. Steinberger et al. (2009) argue that a transition to a sustainable energy system must also involve a transition from business models that rely upon increasing 'throughput' sales of energy, towards a 'performance-based energy economy', where profits are decoupled from energy consumption -leading to absolute reductions in demand. Sorrell (2007) outlines how the economic viability of ESBMs is closely related to their associated transaction costs. Consequently, such contracts are only viable when the transaction costs of implementing and negotiating an energy service contract can be outweighed by the energy production cost savings of efficiency measures and outsourcing to an ESCO. "Transaction costs, in turn, will be determined by the complexity of the energy service, the 'specificity' of the investments made by the contractor, the competitiveness of the energy services market and the relevant legal, financial and regulatory rules" (Sorrell, 2007, p. 507). Sorrell (2007) also defines energy service contracts both by their scope -i.e. the number of energy streams covered by the contract (e.g. electricity, heating, cooling, hot water, lighting, etc.) and by their depth -the extent to which the ESCO has control over final conversion equipment. As shown in Figure  1, shallower utility supply contracts involve the delivery of raw energy commodities whilst deeper performance contracts provide final energy services -incentivising ESCOs to seek maximum efficiency from secondary conversion equipment.

ESBMs: empirical examples
Historically, ESBMs have been restricted to large public and industrial sites due to high transaction costs. A 2014/15 UK survey showed that the majority exist in the public sector, usually at larger sites (such as hospitals and universities), often involving a combined heat and power (CHP) unit and the provision of heating, hot water and electricity . More recently, however, residential examples have emerged. RENESCO in Latvia, Lithuania and Estonia, ICF habitat in France, and Bristol Energy in the UK have each trialled 'pay for performance' models in the provision of thermal comfort in homes (Brown, 2018). Further, the Energiesprong initiative has deep-retrofitted several thousand homes in the Netherlands, bundling rooftop solar with multiple energy services into 30 year net-zero energy performance contracts with social housing providers (Brown et al., 2019b). In the UK, Smartklub and OVO energies' aggregator platform, Kaluza, are also trialling sophisticated electricity market participation from rooftop solar systems, grid connected batteries, heat pumps and smart home devices in a move to valorise their flexibility potential (Brown et al., 2019a;Wang, 2018).
Several factors contribute to the growing viability of these business models. Firstly, intermediaries (Kivimaa et al., 2019) are helping to reduce transaction costs, through standardised contracts and procurement frameworks . Secondly, the profusion of smart appliances, monitoring and machine learning (Wang, 2018) is helping ESCOs to reduce the costs of implementing and maintaining such contracts (Mcelroy and Rosenow, 2018). Equally, as electricity markets are decentralised, 'prosumer business models' open up new revenue streams from the production and selfconsumption of renewable energy, allowing ESCOs to 'stack' revenues to improve their underlying Figure 1 Energy service contracts: scope and depth. Adapted from  Utility Supply Contract

Contract Scope
business case (Brown et al., 2019a). These developments present a prescient moment to study the potential contribution of these business models to the decarbonisation of residential buildings. Fell (2017, p. 137) defines energy services as "those functions performed using energy which are means to obtain or facilitate desired end services or states", highlighting the distinction between sources of energy (i.e. natural gas), energy consuming practices (i.e. showering), end services or states (i.e. being clean), and the energy service of providing hot water itself. Consequently, ESBMs can be described as:

ESBMs: academic studies to date
The provision of useful or final energy services and/or guaranteed savings, how organisations and networks provide these services and the means of capturing revenues from them.
Literature on ESCOs, energy services, and energy performance contracts has grown significantly since the early 2000s. Early work by Sorrell (2005) and Bertoldi et al. (2006) emphasises their potential for a low carbon economy. The notion of a third-party ESCO providing information, finance, installation, plus O&M of energy systems under a long-term contract (Bertoldi et al., 2006) -where users pay for energy services rather than units of fuel (Steinberger et al., 2009) -is now vital for scholars studying low carbon energy transitions (Knoeri et al., 2016;Roelich et al., 2015). Indeed, moving from an energy economy based on increasing throughput sales, towards one based on provision of energy services is today explicitly recognised by multinational bodies including the International Energy Agency (IEA, 2018) and EU Commission(Boza-Kiss and . While much of this literature has been conceptual, studies have sought to map the size and nature of the ESCO market Irrek et al., 2013;Kindström and Ottosson, 2016;Marino et al., 2011;Navigant, 2015;Panev et al., 2014). Boza-Kiss and Bertoldi (2017b) estimated the EU ESCO market at €2.4 billion in 2015, forecasted to grow to €2.8 billion by 2024 with a 1.7% annual growth rate(Boza-Kiss and ). Yet, this still represents a tiny fraction of the EU's overall gas and electricity market (IEA, 2020b).
These studies reveal that ESCO markets remain dominated by large contracts in the public and industrial sectors. There is increasing interest, in the potential of energy service models in the domestic sphere Morris-Marsham and Firth, 2017;Winther and Gurigard, 2017), a sector constituting roughly 2/3 of heat demand in most advanced economies with temperate climates (IEA, 2018). Recent work describes how different energy service models can enable decentralised electricity systems (Brown et al., 2019a;Hall and Roelich, 2016), whole house retrofits (Brown, 2018;Brown et al., 2019b) and reduce the energy performance gap for new build housing (Mcelroy and Rosenow, 2018;Winther and Gurigard, 2017).
A review of the existing literature, however, suggests that the conceptual delineation of these models remains opaque at best and at worst convoluted. Although many papers describe 'ESCO models' (Hannon et al., 2013), this catch-all term conceals significant variability in their nature and purpose. Likewise, the term energy supply contract is used interchangeably with the current utility business model (Morris-Marsham and Firth, 2017), based on the delivery of primary energy commodities. Further, the notion of the energy performance contract is commonly used in finance circles to describe where both financing and energy saving guarantees are outsourced to a third party (Lee et al., 2015;SUSI Partners, 2017), often without the provision of energy services (Fell, 2017). Thus, ESBMs may simply relate to guaranteed energy savings, include the financing of energy savings measures and may further include the upstream utility supply (Kim et al., 2012).
We seek to unpack this complexity, developing a typology of ESBMs, before introducing some emerging examples in the European residential sphere. We subsequently discuss policy options for the expansion of these business models in the context of the existing throughput-based domestic energy market.

Methodology
This research adopted a cross-sectional research design of contemporary domestic energy services business models in Europe, using a mix of qualitative methods. This involved synthesis of multiple data sets, across several research projects. The aim of the research was to map the diversity of domestic ESBMs and identify core features that can aid their categorisation, resulting in the typology described in Section 4. Subsequently, we evaluated the potential of these business models to deliver residential energy demand reduction and climate change mitigation, leading to policy recommendations for their increased adoption. The paper draws on 53 interviews and 7 workshops across the UK and EU, conducted between 2016 and 2021.
3.1 Data collection Interviews were conducted in phases. Phase 1 (2016-2018) focussed on retrofit business and financing models, phase 2 on prosumer business models (2018-2019), and phase 3 (2020-2021) on households, designers and building management practitioners with experience of ESBMs. This included interviews with n=12 ESCOs/finance providers, n=6 energy suppliers/aggregators, n=3 community energy groups, n=14 households, n=10 contractors/housing practitioners, and n=11 consultants and policymakers. Interview data from phases 1 (I#1-17) and phase 2 (I#18-28) was used to understand and better characterise ESBMs and their applicability to different market segments. Phase 3 was divided into n=5 case study buildings. Questions were focussed on the users' perceptions of their existing heat and electricity business model, and then alternatives, including those that provide final energy services. This included two Finnish apartment blocks which previously ran on Helsinki's district heating system, and which had converted to a ground source heat pump (GSHP) (I#29-38). UK household interviews (I#38-53) were focussed on a social housing estate in London, connected to an ageing district heating system which was due to be replaced with a low carbon alternative. Greek interviews were focussed on two apartment blocks with communal heating; one which included solar thermal hot water (I#39-42), and another which was oil and electric heated only (I#43-44).
Interviews were mostly semi-structured, following a set interview protocol, but with some flexibility to pursue relevant lines of enquiry. Interviews were either face-to-face or held online and were subsequently recorded and transcribed. Interviewees were given an information sheet and consent form, with the chance to stay anonymous, but some consented to be named. Further details can be found in Appendix A.
Throughout 2019, seven stakeholder workshops were conducted in the UK (x3), Spain, Portugal, Belgium and the Netherlands. The events brought together local government, community energy organisations, businesses, citizens, academia and non-governmental organisations (NGOs). Four workshops (W#1, #3, #4 & #5) were focused on identifying and characterising decentralised energy business models in each host country, with many examples of ESBMs providing services beyond the basic utility supply model. Three further workshops (W#2, #6 & #7) were held in the city of Bristol (UK), where an energy service offer was being developed at the time of research. The first workshop (W#2) focussed on the challenges of developing decentralised energy systems as direct subsidies are removed. The second (W#6) focussed on financing decentralised energy systems. The third (W#7), delivered jointly with Bristol Energy Company, focussed on the challenges of developing domestic ESBMs. This workshop involved a plenary session with three speakers -Bristol Energy Company, Energiesprong UK and Smartklub from the Nottingham Trent Basin project. Details of these workshops are also found in Appendix A.

Data analysis
Interviews were analysed with the NVIVo 12 qualitative analysis software, using common themes to code the data and structure the analysis. Each interview has a unique signifier (I#X), which is referred to in Section 4.
During workshops W#1, #3, #4 & #5 we adopted a 'Business Model Co-production Method' (Hall et al., 2020), with participants producing diagrams of ESBMs that were at, or close to, market in their host country. This method captures the flows of energy, payments, services, and system interactions by creating component diagrams showing how each business model works. By creating such diagrams, a typology of different business model 'archetypes' emerged. We next compared how each archetype addressed different problems faced by the energy system. An example of these component diagrams is shown in Appendix B.
In W#7, participants were asked to complete a 'Business Model Canvas' (Osterwalder and Pigneur, 2010) outlining the potential features of ESBMs for three key sectors: new build housing, low carbon heat and whole house retrofit. The Business Model Canvas provides a template for developing and documenting business models using nine building blocks. Subsequently, participants were asked to focus on specific opportunities and challenges for ESBMs in their respective sector. Respondents reflected on the challenges and opportunities of these business models from the perspectives of five key actor groups. The three completed business model canvases from W#7 are shown in Appendix C.
The data collection and analysis methodology is summarised in Figure 2.

Residential ESBMs in Europe
Our analysis identified a typology of six ESBMs: energy supply contracts (ESC); energy service financing (ESF); energy performance contracts (EPC); energy services agreements (ESA); energy as a service (EaaS); and managed energy services agreements (MESA). In the following section we explore these models using illustrated empirical examples to show a simplified breakdown of the payment structure of a typical household bill for each Figures (3-8). For simplicity we have focussed on the single energy service of space heating. We have used commonly understood definitions where possible, building on the work of Black (2020), who describes a range of 'heat as a service' models, Brown (2018), who identifies six energy efficiency business model archetypes, and Kim et al., (2012) who explore alternative financing mechanisms for energy efficiency. This literature and definitions are combined with our empirical data to elaborate the typology in the following section.

Energy Supply Contract (ESC)
Under an ESC, instead of supplying primary energy, an ESCO provides a useful energy supply, such as hot water, directly to users . The most common example of these business models is in district heating systems, where ESCOs deliver hot water for space heating and sanitation. Users either pay a fixed price or are metered volumetrically for their hot water usage (kWh), as shown in Figure 3. These business models incentivise ESCOs to deliver this 'useful energy' as efficiently as possible, shifting responsibility for primary conversion efficiency from users to the ESCO. As these models do not cover the secondary conversion equipment, however, they do not incentivise ESCOs to seek demand reductions or building fabric improvements. By taking control of primary conversion equipment, these models do enable ESCOs to access additional sources of revenue from the flexibility of decentralised energy systems, such as batteries, heat pumps and CHP generators.
Domestic ESCs were found to be commonplace and are the typical business model for district heat networks, which are particularly common in northern European countries. Several of our household and building manager case studies -in the UK (London) (I#47,48,52,53), Finland (Helsinki) (I#29-38) and Greece (Athens) (I#39-44) -paid a single charge based on the size of their apartment. In Finland, heating is commonly paid as part of a monthly building service charge, which is set annually based on the overall heating usage of the apartment block. Several building designers we interviewed (I#37, 49-51) commented that ageing district heating systems are locking in high carbon heat provision from coal fired power stations and other fossil fuel sources, with the high sunk costs of these networks a barrier to change. Indeed, even in more modern district heat systems, energy planners had built systems around natural gas CHP systems, only to find that their carbon emissions compared unfavourably to grid electricity.
There are also increasing examples of these models being used for district heating provision from low carbon sources, such as GSHP (I#29-37), usually serving new housing developments (I#20, I#21). In the UK, private developers have tended to avoid installing district heat systems themselves as they disfavour ongoing contracts, where high up font cost might be recovered through service payments. That said, community-owned and financed examples, such as Smartklub (I#19) and BHESCo (I#3) in the UK, are looking to buck this trend.
District heating systems usually require centralised governance and planning, necessitating the involvement of a municipal ESCO or city authority, such as Helsinki in Finland (I#37) or Bristol City Council in the UK (I#28). In the UK, these models serve only 446,517 domestic customers, or <0.2% of homes (ADE, 2020), compared to 63% in Denmark and 93% in Iceland (Euroheat, 2015). Our interview and workshop data suggest that the differing municipal energy, planning, governance, and funding regimes in large part explains this difference; with Nordic countries having a strong tradition of municipal energy provision compared to the UK.
An alternative is ESCs from systems located within or close to the building. Here the ESCO has responsibility for the installation, and O&M of primary conversion systems such as boilers, CHP units or heat pumps within the client's premises (such as our Finnish GSHP examples I#29-37). Residents pay a service charge that covers both the useful energy supply and the O&M costs, with the ESCO ensuring efficient operation. In the Greek examples, residents felt the flat rate for heating based on the size of their apartment (I#39-44) was unfair and penalised them for the energy profligacy of their neighbours. These models are also seen as an important route to the adoption of low carbon heat systems, particularly for multi-family buildings, where the ESCO is incentivised to deliver useful energy at lowest cost. For example, several UK municipal housing providers we interviewed (I#17, I#28) were considering setting up their own ESCOs to replace individual fossil fuel boilers with highly efficient GHSPs, with residents paying for hot water instead of kWh's of natural gas.
Because these models allow the ESCO to control the operation of the primary conversion equipment, they are also an important way of creating value from flexible assets such as heat pumps or li-ion batteries. Interviews with an energy company (I#22), their independent aggregator arm (I#23) and a community ESCO (I#20), indicated that in future ESCOs will increasingly look to provide grid flexibility services that can be aggregated for electricity network operators and earn revenues from arbitragei.e., purchasing electricity at cheaper periods, storing as heat or electricity. The domestic flexibility aggregator commented that the key to unlocking this value is ensuring that end users are not required to control these systems themselves or be directly exposed to time of use pricing (I#23).

Energy Service Financing (ESF)
In ESF, the ESCO will also act as project developer, financing the primary conversion systems as well as taking responsibility for their O&M (Sorrell, 2005). These options can be attractive to end users as they can upgrade a building's heating/cooling plant at no up-front cost, paying off their cost through energy service payments. For example, in one of the cases in Finland (#29-31) a move from district heating to a GSHP was financed by a loan taken by the apartment block, which households then pay back in their monthly building service charge (with the option to pay back earlier) -meaning little financial risk to the householder. These models tend to suit situations where the building owner has limited access to capital or wishes to take the project off their balance sheet. More common in the commercial and industrial sector, these models require ESCOs to mobilise significant capital. As shown in Figure 4, financing can provide more efficient primary conversion equipment, such as a new boiler, reducing the cost of the energy supplied (hot water) once finance repayments are complete. The inclusion of financing, however, can place constraints on the balance sheets of ESCOs with multiple projects (I#2). This has created demand for "third party financing" (Bleyl-Androschin et al., 2009) for projects with proven cashflows, allowing ESCOs to move these projects off their balance sheets, while continuing to receive service payments. These types of projects remain dominated by commercial and industrial buildings, although actors such as the European Regional Development Bank (I#11), Swiss Susi Partners (I#10) and the UK's Green Investment Group are now developing these financial products for larger residential projects.
One specific area where this model is showing potential is through 'solar-as-a service' business models.
Here the ESCO installs, maintains and finances a solar PV system, with the end user receiving the useful electricity in the form of a power purchase agreement (PPA) from the solar array (I#18). These models have been prevalent in the USA and may be partnered with a battery to maximise the onsite consumption of power. Tesla are the most high-profile provider, although several smaller actors are also active in this market. Ideally, the customer would pay a single electricity bill to an ESCO who would also provide electricity when the system was not generating (see MESA, discussed later). As outlined in several of our workshops (W#1,#3, #4, #5 & #7), however, EU 28-day electricity switching requirements, which would allow the consumer to defect from on bill finance agreements, mean that the ESCO risks being unable to recover the capital investment on the system, and thus the solar-as-a-service payments tend to be disaggregated from the electricity bill.
In the UK, SmartKlub (I#20) are trialling this model in the Trent Basin project, featuring a large PV array, 2.1MW battery, rooftop solar and GSHP connected to a district heat network. Smartklub have created an ESCO to manage the system, with limited involvement from residents. Residents receive a reliable power and heat supply with the ESCO optimising the system to secure the best revenues and balance between import and export, using the large battery to contract into flexibility markets through an aggregator for additional revenues. Profits from the ESCO are recycled into a community fund, whilst the ESCO itself is designed to pass into community ownership at the end of the trial phase.

Energy Performance Contracts (EPC)
Under an EPC, an ESCO provides guarantees for measured and verified performance savings from one or more final energy services such as heating or illumination . To ensure that performance is delivered, the ESCO controls secondary conversion equipment such as lighting, heat emitters and controls, and is also incentivised to ensure that the building fabric is efficient (Sorrell, 2005). Because ESCOs are obligated to deliver measured performance, these models create incentives for the maximum efficiency from both primary and secondary conversion equipment. Under a basic EPC, the customer pays the ESCO for measured performance plus O&M. Here the customer secures their own finance and retains a separate energy supply contract with a utility company. In our example ( Figure 5) the user saves €500/year once finance repayments are complete.
EPCs have historically been rarer in the domestic sphere due to the high transaction costs of delivering verified performance savings on individual dwellings. The examples we found have almost exclusively been in social housing, where sufficient scale and standardisation enable these costs to be manageable relative to energy savings. Despite these difficulties, these models promise to exploit the vast fabric energy efficiency potential in the existing housing stock. Proponents argue that these models may also be important in driving household's uptake, who's lack of trust in predicted energy savings has been a barrier to the update of low energy retrofits (I#12-17). Indeed, a UK interviewee (I#52) viewed guarantees on quality and energy saving performance favourably, given their negative experience from previous renovations on their property. Our interviewees and workshop participants described how the technical and economic challenges of implementing EPCs in the domestic sphere remain a major barrier to uptake. These include the performance risk that the measures will not deliver the modelled savings (I#8); the behavioural risk that users will engage in energy profligacy (I#28); the high costs of monitoring and controlling the energy systems in individual homes (I#1); and the risk for the client that the ESCO will cease trading -voiding the performance guarantee (I#1,I#12). Many of the initiatives interviewed were investigating the behavioural and physical characteristics of homes, to mitigate the aforementioned risks (I#1, I#3, I#12-17, I#22-28). Indeed, practitioners described how insurance-backed performance guarantees, low-cost 'smart home' and remote monitoring devices, machine-learning techniques and big data should help to make performance-based business models increasingly viable.

Energy Services Agreement (ESA)
Energy service agreements ( Figure 6) are a variant of EPCs that involve integrated financing of energy saving measures, backed by a long-term performance guarantee. These models are viewed as an important way for housing providers to undertake energy saving improvements 'off balance sheet', with the ESCO taking both the financial and performance risk of the project.   Whether the ESCO or the client should finance the EPC is determined by both the cost of capital (interest rates) and the preferred balance sheet treatment. In the ICF habitat example, after adopting an ESA structure with third party private capital on the Schiltigheim project, I#4 described how in future they would self-finance EPC projects due to their access to low-cost debt. By contrast, RENESCO saw the financing provision of an ESA as a crucial feature of their offer, where the client -low-income owner occupiers -cannot easily take on debt to cover the cost of the works. Several Greek household interviewees (I#43-44) viewed integrated finance with suspicion, however, given their lack of trust in the finance sector more broadly. All three European examples of domestic ESAs reverted to some form of public financing in the delivery of their retrofit interventions. This suggests public financial institutions such as development banks or government loan guarantees have an important role in developing the domestic EPC/ESA market.

Energy as a Service (EasS)
EaaS models, bundle the upstream energy supply into a single final service payment. For example, households may pay a comfort charge relating to room temperatures, lighting lumens, or pay for access to certain services such as showering or clothes washing. In our illustrated example (Figure 7), users would pay for a guaranteed room temperature i.e., 21°C for a period of 'warm hours' (€/21°c/hr). Bristol Energy in the UK have explored such a 'heat-as-a-service' (HaaS) business model to overcome the dual challenges of high energy bills and climate change (I#28). During their 100 home trial, households were offered either a flat rate HaaS tariff, or a variable tariff based on their consumption of warm hours (Energy Systems Catapult, 2019). Using household energy monitoring, the trial aimed to generate data on building fabric performance, occupant behaviour and heating system operation, for future refinement of the offering.
In the HaaS model trialled by Bristol Energy, the ESCO does not provide any energy saving measures or hardware, but instead operates the existing heating appliance and controls / bears the performance risk of the secondary conversion equipment -i.e., the building fabric of the home. The key challenge identified for Bristol Energy (W#7) was developing a comfort plan that accurately represents the energy demands of the home and can accommodate changes in behaviours and occupancy. Through advanced modelling and remote monitoring, these challenges, and the cost of implementing HaaS contracts are expected to reduce significantly in the coming years.
The promise of HaaS is its potential to improve the value proposition in the electrification of heatperhaps the single biggest challenge for reducing carbon emissions. When combined with dynamic or time of use tariffs, HaaS models allow ESCOs to arbitrage electricity prices and match periods of high renewable electricity generation and low-cost on the grid with heat demand, with the ESCO controlling an appliance such as a heat pump. Customers pay for preferred room temperatures and can maintain improved levels of comfort through zoned smart thermostats. For now, this potential remains unrealised, however, with Bristol Energy recently sold to the UK energy giant Centrica. There are no future plans for expansion of the HaaS trial (I#28). Further, I#39 highlighted the subjectivity of 'comfort' and described conflicts with family members relating to preferred room temperatures in the winter and summer months.

Managed Energy Services Agreement (MESA)
A MESA (Figure 8) integrates the ESA model for energy saving improvements with the EaaS model for final energy service provision. Savings, financing, and upstream energy supply are bundled into a single performance contract based on final energy services. Because this model combines a financed energy savings guarantee, and the value that can be created by optimising both the primary and secondary conversion systems, it has the greatest potential to deliver energy and carbon savings. Because of this complexity, however, MESA models have thus far only been adopted in the industrial and commercial sector, where the high transaction costs can be outweighed by substantial savings. In recent years, the Energiesprong initiative has been seeking to deploy a variant of the MESA model in new build and social housing retrofits in several countries (#1, #12-17). The Energiesprong model involves a 'net zero energy' retrofit that includes a new exterior façade, renewable microgeneration, and new heating systems and controls. Over a 30-year contract, the ESCO guarantees the performance of the building, energy bills stay constant, and the householder benefits from an improved exterior appearance and internal building health. Savings on the energy bill pay back a large proportion of the investment and PV panels contribute to any residual electricity bill that accrues through the year. Thus far, these models have been tried in higher density and social housing units. As discussed, the high transaction costs of these contracts may be a greater barrier to their adoption in single dwellings and for owner occupiers.
The original MESA model was centred around the Dutch net metering system, where no residual energy bill would be required provided the 'net-zero electricity' objective was delivered. With requirements for 28-day supplier switching, as stipulated in the EU Electricity Market Directives, there is a significant risk that the customer could switch energy supplier away from the ESCO/solution provider, meaning the full cost of the retrofit could not be recovered from a single service charge. In the first UK trial, this Consequently, our typology of ESBMs is shown in Table 1 and Figure 9 using a synthesis of the multiple data sets described above and published literature. Developing the typology was a cumulative process throughout the study period with the typology and individual model archetypes delineated by developing Sorrell's (2007) notions of 'scope' and 'depth'. For business model scope, we articulate three key components: the nature and number of energy service(s) included (e.g., heat, light etc.), the inclusion of energy efficiency measures, and the inclusion of financing within the offer. The depth of the ESBM is a product of whether primary, useful, or final energy services are provisioned. This determines the nature of the energy efficiency measures involved; whether at the point of extraction, primary or secondary conversion, which in turn directs the type of financing required.

Figure 9 Typology of ESBMs
These models may be applied to a single energy service such as space heating, with different business models potentially applying across multiple energy services of heat, light, motive power etc. For example, a household may have an ESA for building fabric improvements but retain a conventional utility contract for electricity provision. As we discuss later, the choice of business model has significant implications for the adoption of low carbon energy systems and demand saving measures.

Barriers for key stakeholders
Workshop (W#1-6) participants were asked to identify stakeholder groups critical for the adoption of ESBMs. These included property owners, developers / contractors, energy suppliers, and market regulators. During the final workshop (W#7), participants were divided into three groups to focus on the barriers for these stakeholders across three areas where ESBMs may play an important role: new build housing, low carbon heat, and whole house retrofit. Each group was asked to contemplate the opportunities and challenges of ESBMs from the perspective of these four stakeholder groups, summarised below and in Table 2.

Property owners
Key challenges identified were concerns of higher purchase prices of homes and the costs of low carbon heat and whole house retrofit measures; creating a need for specialist low interest finance, and preferential tax treatment. There were also concerns surrounding insufficient understanding of the energy service concept, requirements for behaviour change, and altered expectations. Issues of disruption for tenants, loss of landlord rental income, ease of access and the space needed for new equipment were also raised. These were viewed as generic to retrofit projects, however, and the offer of guaranteed performance improvements from ESBMs was viewed as a key benefit. Finally, there were concerns surrounding consumer protection for long-term service contracts and the potential to be "locked in" to a poorly performing ESCO.

Developer/contractors
The primary risk of ESBMs identified was the performance of measures and potential losses from underperformance. Relatedly, the skills gap for delivering energy savings measures is pervasive across Europe, with the ongoing energy performance requirements of ESBMs and the lack of accreditation and standards especially challenging. SMEs were viewed as reluctant to innovate and invest in new capacity when demand for these services is inconsistent and uncertain. Housing developers also have little incentive to deliver ongoing performance under current building regulations, with the lack of standardised ESCO contracts and ESBM providers adding uncertainty, cost and complexity.

Energy utilities
Utilities were viewed as threatened by ESBMs, which represent a fundamental challenge to their busines model. Current electricity market regulation was viewed as antithetical to ESBMs, however, with long-term contracts hampered by supplier switching rules. The rollout of smart meters in many European countries was also viewed as a missed opportunity, with a gap between the specification of most meters and the data required for detailed performance monitoring. A further issue is the barriers for distributed energy systems in accessing electricity flexibility markets, which exist to varying degrees in Europe.

Regulators
Many of the previously identified barriers related to the current regulatory regime surrounding energy supply, installer standards and building codes and regulations. Specific areas included: the barriers to offering long-term contracts in electricity supply markets; the lack of specific regulation governing the ESCO market and the accreditation of ESCO contractors; the lack of clear consumer protections and standards for ESCO customers; the absence of performance-based compliance in building regulations for new homes; and electricity wholesale and balancing market designs which currently favour traditional utilities and centralised power generation.
This exercise demonstrated that many of these issues were commonly held between the different sectors, although in general fewer barriers were thought to exist for new build housing and the most for whole house retrofit models -suggesting that ESBMs would be easier to adopt in a new housing context. Workshop W#7 findings are summarised in Table 2 and inform the policy recommendations in Section 6, with crosses representing where challenges were relevant for the sector.

Discussion
This paper has developed a novel typology of ESBMs and identified emerging examples in the domestic sphere. While this builds on previous studies, (Black, 2020;Brown, 2018;Kim et al., 2012) our typology in Table 1 and Figure 9 provides future researchers with a novel analytical framework to rationalise their study and characterisation. Our interviews, workshops and review of previous literature  identifies that while energy supply contracts for district heating and CHP systems are common, deeper performance contracts and models focussed on the final energy services remain rare in Europe. While our research findings support Sorrell's (2007) analysis that the high transaction costs of implementing and enforcing these contracts remain a barrier, we find anecdotal evidence that these costs are reducing through standardised procurement frameworks , smart home devices and remote monitoring systems (Brown et al., 2019b;Wang, 2018) and the paring of these data with machine learning algorithms (Amayri et al., 2016). Indeed, future research should aim to quantify the contribution that such developments are having on these transaction costs.
Alongside these technical and financial challenges, we also identified institutional, legal and cultural barriers to the adoption of domestic energy service models. These findings are consistent with the work of Hannon et al., (2013) and Bolton and Hannon (2016) who describe the co-evolutionary relationship between energy systems and the dominant energy supply paradigm. Our research highlights how the rules governing electricity markets, codified under EU directives, are proving a direct impediment to the adoption of these business models, often reifying consumer choice and liquid markets at the expense of long-term efficiency investments (Brand-Correa and Steinberger, 2017). Further, the emphasis on throughput is also present in the housebuilding and construction sectors, with a clear reticence to develop long-term relationships and performance-based compliance (Winther and Gurigard, 2017). Indeed, we find that many of the impediments to an energy service economy go beyond the regulation of energy markets, and relate to issues of municipal governance (Bale et al., 2012;Hannon and Bolton, 2015;Roelich et al., 2018), the nature of national financial institutions Mikler and Harrison, 2012) and the increasing financialization of housing provision (Blakeley, 2020).
In the following sections, we discuss these issues in the context of three core areas where energy service models can contribute to decarbonising the European energy system: new build housing; low carbon heat and whole house retrofit.

New build housing
ESBMs in new build housing would fundamentally alter the dominant housebuilding model. Currently, the majority of European homes are built by large speculative developers (Eurostat, 2020) who take little or no interest once homes are built. This disincentivises investment in energy efficiency and other low carbon measures, as developers seek to reduce capital costs (Heffernan et al., 2015). This lack of accountability for energy performance is also a major driver of the pervasive 'energy performance gap' that plagues modern housebuilding (Gupta and Dantsiou, 2013). Mcelroy and Rosenow (2018) argue that ESBMs could contractually oblige developers to meet the standards to which homes were designed. Further, by involving an ESCO at an early stage, developers would be compelled by the ESCO to ensure maximum efficiency and carbon reductions through their design, fabric specification, HVAC and onsite renewables. Equally, developers and ESCOs could merge their activities to offer design, build and operate contracts and could offer a single service charge for comfort, appliance use and lighting.
Our review of barriers to adoption indicated that new build housing is perhaps the easiest place to trial ESBMs, providing an opportunity to design buildings that deliver measured performance in practice. This includes the use of rigorous design standards (e.g. the Passivhaus approach), proven to deliver realised energy performance outcomes (Mitchell and Natarajan, 2020), the adoption of modern methods of manufacture and offsite construction (Jin et al., 2020), and early stage integration of decentralised energy systems (Bolton and Hannon, 2016). Moreover, as argued by Burman et al., (2014), the energy performance component of national building regulations should increasingly move towards 'performance based compliance', with contractual penalties for underperformance.

Low carbon heat
ESBMs also have the potential to address the major challenge of heat decarbonisation in homes. Most of Europe's homes are heated using fossil fuels, where households purchase units of fuel (i.e. gas, oil, electricity) taking responsibility for the efficiency and maintenance of the primary conversion equipment themselves (i.e. boilers, heat controls, radiators). The need to decarbonise domestic heating means these high carbon systems must be converted to heat pumps and heat networks in the coming decades. This has presented a problem in many existing homes, where the high cost of infrastructure and equipment combined with different features such as lower flow temperatures and new controls has presented a barrier to uptake (Watson, 2016).
ESC models are already more common in Nordic countries (Euroheat, 2015), often involving heat networks and the delivery of centrally-produced hot water through an ESCO. Although these models will need to make a significant contribution to heat decarbonisation, they tend to be restricted to dense population centres and require the active involvement of municipal authorities (Bale et al., 2012;Hannon and Bolton, 2015). We found evidence of HaaS models emerging where customers adopt a 'comfort tariff', paying for temperature levels in specific rooms in the home. As outlined by Marques et al., (2019), HaaS models may present a solution by providing households with the useful end servicei.e. a thermally comfortable home -and bundling the control, optimisation and financing of these higher capital cost but lower operational cost systems upstream into a single service payment.
Our research did identify highly subjective and personalised attitudes to comfort, however, reinforcing the notion that a single room temperature may not be agreed within households (Sovacool et al., 2020). Indeed, following Fell's (2017) distinction between sources of energy, energy consuming practices, end services or states and energy services, arguably these HaaS models provide the final energy service of room temperature rather than the subjective end state of being 'thermally comfortable', suggesting this framing is a misnomer. Further, Shove (2017) argues the prescription of a universal room temperature (e.g., 21°c) may perpetuate a higher level of service expectation and energy consumption, which is reified and reproduced through building performance standards.
As heat pumps become more prevalent, ESCOs may also introduce smart heating control and storage -taking advantage of variations in the daily electricity price -to maximise revenues whilst ensuring the same comfort and service (Brown et al., 2019a;Richter and Pollitt, 2018). Proponents argue these customer-centric 'servitised' heat offerings will reduce barriers to the adoption of low carbon heat (Energy Systems Catapult, 2019; Marques et al., 2019). Our research suggests that improvement in occupant health and wellbeing are seen on an equal footing with energy savings by households (Brown et al., 2019b;Knoeri et al., 2016;Wilson et al., 2015). Consequently, local authorities and not-for-profit housing providers may be key actors in developing HaaS business models due to their greater emphasis on distributional benefits, social value and public health outcomes (Hall and Foxon, 2014;Roelich et al., 2015).
The MESA example is more challenging under existing energy market regulations. The desire to finance a heat pump and recover the costs through energy savings (including the arbitrage of daily electricity price fluctuations) would represent an optimal route to decarbonising heat at no up-front cost to the user. Currently, unless the upstream electricity supply can be protected from the supplier switching rules, these models represent an unpalatable proposition to the ESCO (Littlechild, 2006). Thus, the client would either need to self-finance their own equipment under an EaaS model or, if connected to a heat network, an ESCO would provide an ESC/ESF offering for volumes of hot water.

Whole house retrofit
Many homes may also require invasive energy efficiency measures, alongside low carbon heat, smart controls electricity microgeneration and batteries. Indeed, the scale of domestic heat consumption necessitates absolute reductions in demand if domestic heat is to be electrified (CCC, 2019b), with many homes not suitable for lower temperature heat pumps, requiring efficiency improvements as a pre-requisite (Barnes and Bhagavathy, 2020). Although some incremental improvements to the European housing stock have been implemented, many of these low hanging fruit have now been exploited . What remains are millions of un-insulated walls, floors and singleglazed windows. This 'whole house retrofit' challenge is therefore among the most beguiling of all decarbonisation goals. Increasingly, policymakers and practitioners are looking to ESBMs as a potential route to overcoming this challenge (Green Alliance, 2019).
require public engagement programs, minimum energy performance standards, investment in the skills gap and a range of grant funding and financing packages.

Conclusion and policy implications
This paper has sought to clarify the conceptual understanding of ESBMs, introducing a new typology and a review of emerging domestic examples in Europe. We characterise ESBMs as 'The provision of useful or final energy services and/or guaranteed savings, how organisations and networks provide these services and the means of capturing revenues from them'. Through a novel typology, we differentiate ESBMs by whether they deliver useful energy supply, performance or final energy services and the inclusion or exclusion of financing upstream of the customer. It is hoped that this typology will be useful to scholars and practitioners in the future study and evaluation of ESBMs.
Following our interviews and workshops, we propose 13 policy recommendations for local and national governments in Europe, which could facilitate the increased adoption of domestic ESBMs, shown in Table 3 below. , and switching times are to reduce to as little as 24 hours. This legislation currently disincentivises long term service contracts and hampers ESCOs offering these contracts. The EU should review these rules for the provision of electrified heat which would put it on an equal footing with thermal/fossil heat supply contracts. The UK's exit from the EU could also present an opportunity to review this logic.

2.
New regulation is needed to manage the domestic ESCO market and to protect consumers from lock into poorly performing ESCOs. This could include developing a 'supplier of last resort' to take on failed ESCOs' contracts.

3.
National governments should develop a training and skills program for the low carbon housing sector -emphasising measured performance outcomes and a move away from 'fit and forget' construction practices.

4.
Electricity network charges should be made increasingly dynamic and cost reflective and move away from static and volumetric charging. This will incentivise business models which provide energy services to homes, whilst also providing flexibility to electricity system operators.

5.
Access to low-cost capital is critical for the financial viability of energy service models which include finance. As part of Green Stimulus and COVID recovery packages governments and public banks should facilitate low cost, patient investment to ESCOs and grant funding to projects which deliver clear public good outcomes. New Build Housing 6. The EU should introduce legislation for performance-based compliance for new buildings, to ensure developers are required to deliver measured energy performance outcomes. Alternatively, developers could be required to meet similar standards through municipal planning policy.

7.
In many European countries, the responsibility for ensuring energy performance and building regulations compliance has been outsourced to the private sector, with the result that much of the industry is now self-regulating. This is creating perverse incentives and a lack of enforcement of regulations. These functions should be brought back under public purview with strict enforcement of energy and environmental standards. Low Carbon Heat 8. Domestic heat decarbonisation will require electrification, especially through the adoption of heat pumps. Current EU electricity market regulations mean that customers must have a single electricity supplier. Some national regulators are trailing the potential for multiple suppliers under a single meter point, potentially allowing an ESCO to enter a long-term contract for heat service provision, whilst allowing the customer to switch their main electricity supplier. This should be encouraged across Europe. 9.
The share of policy costs on electricity bills unfairly prejudice electrification -these costs should be shifted into general taxation to level the playing field away from gas and oil. Whole House Retrofit 10. In several EU countries Value Added Tax (VAT) is added at full rate to building materials and products which promote energy saving, while household energy supply receives a VAT discount. These fiscal policies should be reviewed considering the imperative to reduce energy consumption. 11. Minimum energy efficiency standards (MEES) across all housing tenures are likely to be necessary to drive the uptake of whole-house retrofits in Europe. 12. Performance based retrofit in local authority and social housing could be first step to wider performance contracts becoming mainstream. 13. In several national contexts the offer of a combined rent and service charge is prohibited. Following the example of the Netherlands, legislators should enable this 'whole cost of living approach' in social housing, to encourage ESMBs for retrofit.
Domestic ESBMs represent a huge opportunity to decouple the profits of energy supply from everincreasing throughput sales of energy. These business models may prove instrumental to building homes that deliver the performance outcomes for which they were designed, enable the decarbonisation of domestic heat, and provide a route to delivering energy efficiency measures to millions of homes in the coming decades. Such approaches may be a crucial tool as governments seek to build back better from the COVID-19 pandemic and meet decarbonisation targets. Our study has identified a series of technical, economic, institutional, and cultural barriers that constrain their increased adoption in the domestic sphere. Ultimately, we believe that, for these models to play a significant role, a paradigm shift will be required to alter the throughput energy policy orthodoxy that has governed liberalised energy markets for the past 30 years.