Governing the Digital Transition: A Systems Thinking Approach to Analyse Digital Twin Adoption in UK Housing Retrofit

Governing the Digital Transition: A Systems Thinking Approach to Analyse Digital Twin Adoption in UK Housing Retrofit | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Governing the Digital Transition: A Systems Thinking Approach to Analyse Digital Twin Adoption in UK Housing Retrofit Alejandro Zhou Wei, Irene Pluchinotta, Simon Vakeva-Baird This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9227696/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract The decarbonisation of the UK's 29 million homes represents a problem characterised by systemic inertia and fragmented policy. Digital Twin technology offers a niche innovation with the potential to accelerate data-driven retrofitting; however, its adoption is suppressed by complex socio-technical barriers. This article presents a systems thinking analysis to identify and analyse the causal mechanisms shaping DT adoption in the UK housing sector. Adopting a systems thinking perspective, a Causal Loop Diagram is constructed based on a systematic review of contemporary literature (2018–2024), from which 15 core variables are extracted through open coding. Eigenvector centrality and cross-impact matrix analysis are then applied to identify possible deep leverage points for intervention. The analysis reveals five reinforcing loops—including data-quality and simulation-investment cycles—that act as engines for change, currently held in check by institutional capability traps. It is proposed that UK policymakers shift from a technocentric focus to a systemic approach that prioritises interoperability, technical capacity, and stable long-term incentives to unlock the transition toward the Sustainable Development Goals (SDGs 7, 9, 11, and 13). Digital Twin Retrofit Systems Thinking Causal Loop Diagram Socio-Technical Transitions Sustainable Development Goals Figures Figure 1 Figure 2 Figure 3 Figure 4 1. Introduction Upgrading existing buildings is an urgent priority for climate change mitigation, particularly in the United Kingdom (UK), where the built environment accounts for approximately 25% of the UK’s total greenhouse gas emissions ( UKGBC 2023 ) . The UK Government has introduced policies such as the Heat and Buildings Strategy ( HM Government 2021 ) and Minimum Energy Efficiency Standards ( UK Government 2017 ) to drive energy retrofitting in residential and commercial properties. However, despite these initiatives, progress remains slow due to persistent challenges: high upfront investment costs, fragmented building performance data, and slow market uptake of new technologies ( Zhu et al. 2024 ). A recent study by Jradi et al. ( 2023 ) noted that without addressing such systemic barriers, meeting decarbonisation goals in the building sector will be difficult. One promising technological innovation for enhancing retrofit outcomes is Digital Twin (DT) technology, a dynamic, real-time digital replica of a physical asset. By integrating Internet of Things (IoT) sensors, building information modelling (BIM) data, and AI analytics, digital twins create a feedback loop of information that can continuously improve building performance (Attaran and Celi 2023). The potential benefits include reduced operational costs, improved energy efficiency, and verified carbon monitoring ( Sghiri et al. 2025 ), making DTs valuable tools for enabling the scaling of effective retrofit programmes. Beyond its technical advantages, DT technology is closely associated with sustainability objectives. Implementing data-driven retrofits can directly contribute to SDG 7 (Affordable and Clean Energy), Target 7.3 on energy efficiency; SDG 11 (Sustainable Cities and Communities), Target 11.6 on reducing the environmental impact of cities; and SDG 13 (Climate Action), Target 13.2 on integrating climate measures into policy ( United Nations 2023 ) . Despite this promise, adoption remains limited by high costs, technical and privacy challenges, and regulatory gaps, particularly in the context of UK housing retrofits. The persistence and interaction of these barriers indicate that the challenge is fundamentally systemic ( Meadows 1999 ) . High upfront costs shape investment decisions, which influence market demand and slow standardisation ( Zhu et al. 2024 ). Fragmented data infrastructures limit performance transparency, weakening trust among stakeholders and discouraging wider uptake ( Jradi et al. 2023 ). Skills shortages reduce implementation capacity, prolonging uncertainty and constraining diffusion ( Sobowale et al. 2023 ). As a result, interventions that target single issues in isolation, such as cost reduction or technology deployment alone, struggle to produce sustained change. Understanding why this is the case requires situating these barriers within a broader theory of technological change. The Multi-Level Perspective (MLP) on socio-technical transitions offers such a framework, characterising DT adoption as a niche innovation attempting to destabilise an entrenched socio-technical regime ( Geels 2011 ) . A systems thinking approach provides a structured way to analyse this complexity. By focusing on relationships, feedback loops, and system behaviour, systems thinking enables the construction of causal models that make interactions, e.g., between cost, policy incentives, trust, and data availability explicit ( Sterman 2000 ) . This perspective supports the identification of potential leverage points where targeted interventions can influence system behaviour more effectively than incremental or siloed measures. This study aims to apply systems thinking to analyse the factors and interdependencies influencing DT adoption in the UK's domestic retrofit sector. The primary research question is: How do interactions between policy, technology, and market actors create systemic barriers and enablers to the adoption of digital twin technology for domestic building retrofits in the UK? Two sub-questions structure the investigation: (i) What are the key reinforcing and balancing feedback loops governing DT adoption, and how do they correspond to known system archetypes? (ii) Which factors have the greatest influence on the system, and what intervention strategies do they suggest for policymakers and the industry? This study demonstrates how a systems thinking approach can be used to advance sustainability transitions in the built environment. Previous studies have examined the technical aspects of DT or listed generic benefits and barriers, such as reduced operational costs, improved energy efficiency, and data interoperability challenges ( Attaran and Celik 2023 ; Elghaish et al. 2024 ). This paper situates these elements within a holistic causal structure of the system, leveraging a Causal Loop Diagram (CLD) to reveal deeper principles of socio-technical behaviour. The findings are useful for policymakers, particularly the Department for Energy Security and Net Zero (DESNZ) and the Department for Levelling Up, Housing and Communities (DLUHC), and industry stakeholders positioned to design integrated interventions. 2. Theoretical Framework 2.1 The UK Retrofit Sector: System Context and Stakeholders Retrofitting the UK's housing stock involves a diverse network of actors operating across different scales. Central government (particularly DESNZ and DLUHC) sets policy and funding frameworks. Local authorities and housing associations implement retrofit programmes. Technology providers (IoT manufacturers, BIM software firms, DT platforms) supply the digital infrastructure. Contractors and retrofit installers implement physical works. Homeowners and tenants occupy the retrofitted properties and must consent to and engage with the process. Following Reed et al.'s ( 2009 ) typology, which advocates differentiating stakeholders by interest and influence to prioritise engagement, Fig. 1 presents a Power-Interest Grid mapping the principal actors in the UK domestic retrofit system ( Ackermann and Eden 2011 ). Appendix A provides a detailed categorisation of stakeholder groups, outlining their roles, objectives, and the resources they contribute to or require from the system. Each actor has distinct interests, levels of technical expertise, and risk appetite, which shape their response to DT adoption. These actors operate with conflicting priorities that generate structural tensions directly relevant to DT adoption. Table 1 summarises the three principal tensions and their systemic implications, each of which is later reflected in the feedback loops of the CLD. Table 1 Principal stakeholder tensions in UK DT-enabled retrofit and their systemic implications. Tension Conflicting Actors Nature of Conflict Implication for DT Adoption Cost vs. Innovation Retrofit contractors and developers vs. technology firms and consultants Short-term cost minimisation vs. investment in advanced AI/IoT systems with unproven at-scale returns Suppresses DT Adoption Rate by limiting willingness to pay for IoT/sensor integration; directly reflected in the B2 Demand-Driven Cost Inflation loop ( Zhu et al. 2024 ) Privacy vs. Data Utilisation Homeowners and occupants vs. DT vendors and service providers Concerns over intrusive monitoring vs. demand for comprehensive real-time data collection to maximise DT utility Creates a trust deficit (Stakeholder Trust variable) that suppresses Building Data Availability and the R2 Data-Quality loop; requires transparent data governance to resolve ( Jradi et al. 2023 ) Standardisation vs. Flexibility Government agencies and standards bodies (e.g., BSI) vs. technology companies with proprietary platforms Open interoperability requirements vs. proprietary solutions that enable faster iteration but fragment the market Market fragmentation suppresses Interoperability and Technical Capacity, blocking the R4 Standardisation loop; excessive rigidity risks stifling early innovation ( Ammar et al. 2022 ; Sobowale et al. 2023 ) These tensions are not independent: the cost-innovation conflict amplifies the privacy-data conflict (high IoT costs reduce willingness to install comprehensive monitoring), while fragmented standards compound both by preventing cost reductions through scale. Together, they constitute the socio-technical lock-in that the CLD in Section 4 makes explicit. 2.2 Multi-Level Perspective on Socio-Technical Transitions To understand the slow adoption of digital twins, this study integrates two theoretical streams: the Multi-Level Perspective (MLP) on socio-technical transitions and the praxis of systems thinking. The MLP conceptualises transitions as interactions between the socio-technical landscape (e.g., net-zero targets), the socio-technical regime (the stable, traditional construction industry), and niche innovations such as digital twins ( Geels 2011 ) . DT adoption represents a niche-regime interaction where the technology must overcome the lock-in of a regime characterised by low margins and limited appetite for advanced innovation. Systems thinking, particularly causal loop diagramming, provides the analytical apparatus to map these niche-regime interactions, trace the feedback mechanisms that sustain regime lock-in, and identify leverage points for deliberate intervention ( Veldhuis et al. 2025 ). 3. Methodology This study adopts a qualitative systems thinking modelling approach. The inquiry followed a structured four-stage process: (1) systematic literature review and evidence base; (2) variable extraction and conceptual coding; (3) causal loop diagram (CLD) construction and structural mapping; and (4) CLD analysis, including thematic cluster analysis and possible leverage point identification via eigenvector centrality and MICMAC calculation. Each stage is described in turn below. We informed this methodology with prior work on causal mapping from open coding of text data ( Prioreschi et al. 2024 ; Kim and Andersen 2012 ) , CLD analysis ( Pluchinotta et al. 2021 ), and possible leverage point identification ( Coletta et al. 2025 ), adapting these methodologies to our context. Figure 2 illustrates the step-by-step workflow applied. 3.1 Step 1 — Systematic Literature Review and Evidence Base The evidence base for this study was constructed through a systematic review of peer-reviewed literature and pivotal high-quality industry reports. Academic databases were searched—including ScienceDirect, IEEE Xplore, MDPI, Google Scholar, and UCL Explore—using keyword combinations related to 'Digital Twin,' 'Net Zero,' 'Retrofit,' 'UK,' and 'Energy optimisation.' These searches yielded a broad initial set of literature to which inclusion and exclusion criteria were applied (see Table 2 ). Table 2 Inclusion and Exclusion Criteria for the Systematic Literature Review Inclusion Criteria Exclusion Criteria Peer-reviewed academic publications or high-quality industry reports Research papers without any clear application of Digital Twin technology or data-driven retrofit tools Focus on existing building stock (not new construction) Studies focusing only on new building designs or construction phases Discussion of Digital Twin, AI, IoT, or BIM integration within the building sector and retrofitting Research with generic sustainability discussions but no mention of DT, IoT, or AI integration Exploration of barriers, enablers, policy context, or technical implementation in the UK or comparable settings Articles lacking technical or empirical detail Published 2018–2024 to capture the contemporary landscape Purely conceptual studies without any application to retrofits Through database querying and backward/forward citation snowballing, approximately 20 documents were screened at the abstract level before narrowing to a core set of four highly relevant studies: Jradi et al. ( 2023 ), Sobowale et al. ( 2023 ), Elghaish et al. ( 2024 ), and Zhu et al. ( 2024 ). Appendix B provides a summary of each study's scope, contribution, and relevance to the variable extraction process. These sources, along with supporting references, formed the evidence base for extracting causal factors. It should be noted that the relatively small core set of primary studies reflects both the novelty of the topic and the specificity of the inclusion criteria; future work could expand this base, particularly through expert elicitation, to ensure comprehensive variable identification. 3.2 Step 2 — Variable Extraction and Conceptual Coding In the second stage, relevant system variables and causal relationships were extracted from the selected literature using an open coding approach ( adapted from Kim and Andersen 2012 ) . This process translated qualitative insights into structured system components. The following categories of variables were sought: Policy and incentive factors (e.g., availability of government grants and regulatory standards). Technical system properties (e.g., interoperability of platforms, data infrastructure quality, and workforce skills). Economic and financial factors (e.g., upfront cost of IoT sensors and return on investment period). Social and behavioural factors (e.g., stakeholder trust in new technology and homeowner awareness). Key outputs and performance metrics (e.g., energy savings achieved and retrofit rate of homes per year). Each variable was clearly defined and, where possible, associated with a quantifiable measure to maintain conceptual clarity, even though the CLD itself is qualitative. Redundant or overlapping concepts from different sources were merged into a single variable to avoid unnecessary complexity. For instance, various mentions of 'lack of expertise,' 'skills gap,' and 'training needs' were consolidated into a single variable termed 'Technical Capacity of Workforce.' By the end of this coding process, 15 core variables were identified that recurred across sources and were most pertinent to DT adoption in retrofits. Each was classified as a Driver/Enabler, a Barrier, or an Outcome in the system. Appendix C provides full definitions, measurement indicators, and source attributions for all 15 variables. 3.3 Step 3 — Causal Loop Diagram Construction The third stage involved translating the identified variables and relationships into a Causal Loop Diagram (CLD) to represent the system's feedback structure. The CLD was constructed using Vensim PLE (an open source systems dynamics software, Vensim 2025 ) and analysed in Kumu (an online systems mapping tool, Kumu 2025 ). Causal links between variables were drawn based on relationships documented in the literature or on logical inferences from domain knowledge. In a CLD, a '+' (positive) polarity indicates that if the cause increases, the effect increases above what it would otherwise have been; a '–' (negative) polarity indicates the inverse. These polarities describe relative changes, not absolute magnitudes ( Sterman 2000 ). 3.4 Step 4 — Structural and Thematic Analysis and Candidate Leverage Point Identification With the CLD constructed, feedback loops within the diagram were identified and analysed. A feedback loop is a closed chain of cause-and-effect links that 'feeds back' into itself ( Sterman 2000 ) . Seven principal feedback loops were found—five reinforcing (R1–R5) and two balancing (B1–B2), and their causal pathways were traced and recorded (Table 3 ). In identifying loops, established systems archetypes were referenced ( Senge and Sterman 1992 ; Meadows 1999 ) to characterise structural patterns such as virtuous cycles of growth or limit-imposing counteracting forces. Firstly, a thematic cluster analysis (adapted from Pluchinotta et al. 2021 ) was used to group the 15 CLD variables into broader thematic clusters based on their conceptual similarity and shared causal relationships. This approach helps to reveal focal areas of intervention by showing which domains — policy, technology, economics, and social dynamics — contain the most structurally interconnected variables, and where cross-domain dependencies create potential for coordinated action. Beyond visual inspection of loops, a structural analysis based on graph theory was conducted to identify candidate leverage points, the influential variables in Donella Meadows' terminology (1999). Using Kumu, the CLD was mapped as a directed graph, and eigenvector centrality was calculated for each variable. Eigenvector centrality is a network metric that assigns higher scores to nodes that are highly connected to other highly connected nodes, thereby identifying variables located within influential parts of the CLD. Following the methodology and rationale described in Coletta et al. ( 2025 ), a MICMAC (Matrix of Cross-Impact Multiplications Applied to Classification) analysis was applied to all 15 variables to cross-verify and enrich the eigenvector centrality results. This approach generated an influence-dependence score for each variable, quantifying its driving power over and dependence on other variables in the CLD. Plotting these scores on a grid of influence versus dependence allowed the identification of 'deep leverage point candidates' — variables that are both highly influential and relatively autonomous within the network. It is important to note that this analysis quantifies structural connectivity rather than dynamic causal strength; centrality scores, therefore, indicate potential leverage points based on network structure, requiring further validation for their actual impact in a dynamic context. 4. Results 4.1 The Causal Loop Diagram and Feedback Loop Structure The Causal Loop Diagram (Fig. 3 ) visualises the complex interdependencies between the 15 identified variables, spanning policy, technology, economics, and social dimensions, that describe the causal mechanisms governing DT adoption in the UK domestic retrofit sector. Seven feedback loops were identified and are summarised in Table 3 below. Table 3 Overview of CLD Feedback Loops Loop Name Label Causal Pathway Policy-Driven Adoption Loop R1 Government Incentives → (+) DT Adoption Rate → (+) Energy Optimisation Capability → (+) Energy Renovation Rate → (+) Stakeholder Trust → (+) Government Incentives Data-Quality Loop R2 Building Data Availability → (+) Quality of Retrofit Decision-Making → (+) DT ROI Period → (+) Stakeholder Trust → (+) DT Adoption Rate → (+) Building Data Availability Economies of Scale and Innovation Loop R3 Cost of IoT/Sensor Integration → (–) Integrated AI-IoT-DT Solutions → (+) Synchronous Energy Prediction → (+) Energy Optimisation Capability → (+) Energy Renovation Rate → (+) Stakeholder Trust → (+) DT Adoption Rate → (+) Government Incentives → (–) Cost of IoT/Sensor Integration Standardisation Loop R4 DT Process Standardisation → (+) Interoperability → (+) Technical Capacity → (+) DT Adoption Rate → (+) Stakeholder Trust → (+) Government Incentives → (+) DT Process Standardisation Simulation Investment Tool Loop R5 Availability of Simulation-Based Retrofit Tools → (+) Quality of Retrofit Decision-Making → (+) DT ROI Period → (+) Stakeholder Trust → (+) DT Adoption Rate → (+) Government Incentives → (+) Availability of Simulation-Based Retrofit Tools Policy Exhaustion Loop B1 Government Incentives → (+) DT Adoption Rate → (+) Energy Renovation Rate → (–) Government Incentives Demand-Driven Cost Inflation Loop B2 Stakeholder Trust → (+) DT Adoption Rate → (+) Cost of IoT/Sensor Integration → (–) Integrated AI-IoT-DT Solutions → (–) Stakeholder Trust 4.1.1 Reinforcing Loops (R1–R5) The five reinforcing loops constitute the potential engines of DT adoption in the system. R1 — Policy-Driven Adoption Loop : This loop centres on the role of government support. Government Incentives stimulate the DT Adoption Rate, which improves Energy Optimisation Capability and, consequently, the Energy Renovation Rate. A higher renovation rate builds Stakeholder Trust, which in turn generates political and market pressure to sustain Government Incentives, closing a virtuous circle. This loop exemplifies a 'Success to the Successful' archetype ( Meadows 1999 ) : consistent policy incentives can kick-start a self-sustaining market transformation. R2 — Data-Quality Loop : Building Data Availability (from IoT sensors) enhances the Quality of Retrofit Decision-Making, which shortens the DT Return on Investment (ROI) period. Demonstrated success increases Stakeholder Trust, leading to higher DT Adoption and further expansion of the monitored building data pool. This reinforcing loop indicates that early investment in data infrastructure creates a knowledge snowball effect: more data leads to better results, which leads to more adoption and, in turn, more data. R3 — Economies of Scale and Innovation Loop As the DT Adoption Rate rises, it stimulates demand for Integrated AI-IoT-DT Solutions. Market growth enables technology providers to achieve economies of scale, reducing the Cost of IoT/Sensor Integration. Lower costs facilitate wider adoption, reinforcing the cycle. This loop resembles a 'Growth and Underinvestment' archetype, where scaling up the market lowers costs and lower costs further scale up the market. R4 — Standardisation and Capacity Loop DT Process Standardisation improves Interoperability, which enhances the Technical Capacity of the workforce. A more capable workforce leads to more successful projects, boosting Stakeholder Trust and adoption. Greater adoption motivates Government Incentives and further standardisation efforts. R4 demonstrates how creating a virtuous cycle of standards, skills, and adoption can rapidly mainstream a technology, generating network effects that benefit each new adopter. R5 — Simulation Investment Tool Loop Greater Availability of Simulation-Based Retrofit Tools improves the Quality of Retrofit Decision-Making and shortens the DT ROI Period. Improved economic outcomes boost Stakeholder Trust and DT Adoption, which encourages governments and investors to develop more sophisticated simulation tools. This loop indicates that investment in good planning tools pays off through better projects, which justifies further tool investment. 4.1.2 Balancing Loops (B1–B2) The two balancing loops represent the braking mechanisms that currently suppress the reinforcing cycles and stabilise low adoption levels. B1 — Policy Exhaustion Loop As Government Incentives drive higher DT Adoption and Energy Renovation Rate, the initial political will for generous incentives may diminish—whether through perceptions that targets have been met, budget constraints, or policy shifts. Reduced incentives slow the adoption rate, counteracting the very gains produced by reinforcing loops. This 'Limits to Growth' archetype ( Kim 2000 ) has been observed empirically in schemes such as the UK Green Homes Grant (2020–2021) ( GOV.UK 2021 ) , where a premature withdrawal of funding reversed early market momentum and eroded stakeholder trust. B2 — Demand-Driven Cost Inflation Loop : If DT adoption rises rapidly but Technical Capacity (skilled labour, supply chain for sensors) does not keep pace, the surge in demand produces bottlenecks. These manifest as price increases, higher costs for hiring qualified installers or sourcing IoT equipment, which erode Stakeholder Trust and dampen further adoption. This loop captures the 'immaturity of the supply chain' problem: without concurrently scaling up capacity, any demand-side push can become self-defeating, explaining why past efforts that simply injected public funds without a holistic strategy repeatedly failed to sustain retrofit markets. 4.2 Thematic Cluster Examining the CLD, the 15 variables naturally grouped into four thematic clusters corresponding to different aspects of the system. This clustering confirms that any proposed intervention must consider all domains simultaneously, not just the technical or financial dimensions. Policy and Governance Cluster: Government Incentives and Funding and Digital Twin Process Standardisation. These shape the regulatory and support environment for innovation, underscoring the role of institutions in guiding transitions. Technology and Infrastructure Cluster: Integrated AI-IoT-DT Solutions, Interoperability, Synchronous Energy Prediction, and related variables that determine the technical feasibility and performance of DT systems. Economic and Market Viability Cluster: DT ROI Period, Cost of IoT/Sensor Integration, and Availability of Simulation Tools, key determinants of the cost-benefit calculus for potential adopters. Social and Human Dynamics Cluster: Stakeholder Trust and Technical Capacity. This cluster represents the human dimension: professional readiness, acceptance of technology, and confidence in new approaches. These clusters are interdependent and not silos. The tension between Clusters 2 and 3—investing in robust technology raises short-term costs, whilst neglecting it damages long-term performance and trust—illustrates the multidimensionality of the problem and the need for integrated decision-making tools. Interpreted through the MLP lens introduced in Section 2.2 , these clusters map onto the three analytical levels that structure socio-technical transitions. The Policy and Governance cluster and the Economic and Market Viability cluster operate predominantly at the regime level — they describe the institutional rules, financial logics, and market expectations that currently reproduce low adoption. The Technology and Infrastructure cluster represents the niche level — the emerging technical capabilities seeking to achieve sufficient scale and legitimacy to challenge regime practices. The Social and Human Dynamics cluster operates across all three levels: Stakeholder Trust is the currency through which niche actors gain regime acceptance, while Technical Capacity determines whether niche innovations can be absorbed into mainstream practice. In this reading, the CLD's reinforcing loops describe niche-level growth dynamics, and the balancing loops describe regime-level resistance mechanisms. 4.3 Eigenvector Centrality Analysis Eigenvector centrality results from Kumu identified Government Incentives and Funding (0.168) and DT Adoption Rate (0.115) as the most structurally central nodes, confirming that government action is the primary catalyst for the system (Table 4 ). Stakeholder Trust (0.111) also holds high centrality, suggesting that financial incentives alone are insufficient without accompanying social legitimacy. Cost of IoT/Sensor Integration (0.112) featured prominently due to its presence in multiple loops, both as a barrier suppressed by economies of scale and as an amplifier of demand-side pressures. Table 4 Eigenvector Centrality Analysis Results (Kumu, 2025 ) Variable Eigenvector Centrality Score Government Incentives and Funding 0.168 Cost of IoT/Sensor Integration 0.112 Digital Twin Adoption Rate 0.115 Stakeholder Trust 0.111 Energy Renovation Rate 0.087 Availability of Simulation-Based Retrofit Tools 0.066 Digital Twin Process Standardisation 0.066 Energy Optimisation Capability 0.052 Building Data Availability 0.045 Integrated AI-IoT-DT Solutions 0.044 Quality of Retrofit Decision-Making 0.044 Digital Twin Return on Investment Period 0.035 Interoperability 0.026 Synchronous Energy Prediction 0.018 Technical Capacity 0.010 Note: Yellow-highlighted rows indicate variables identified as high-leverage nodes based on combined eigenvector centrality and MICMAC analysis. 4.4 MICMAC Analysis: Deep Leverage Points The MICMAC grid (Fig. 4 ) classifies variables by their driving power (influence on others) and dependence (being influenced by others). Variables with high driving power and lower dependence represent autonomous drivers; those with both high driving power and high dependence are relay variables—responsive to upstream changes but also capable of transmitting change downstream ( Coletta et al. 2025 ). Table 5 MICMAC Analysis Results. Deep Leverage Points are highlighted. Variable MICMAC Dependence MICMAC Driving Power (Exposure) Government Incentives and Funding 0.608 0.806 Digital Twin Adoption Rate 0.527 1.000 Integrated AI-IoT-DT Solutions 1.000 0.659 Quality of Retrofit Decision-Making 1.000 0.662 Interoperability 0.534 0.625 Technical Capacity 0.984 0.548 Stakeholder Trust 0.543 0.208 Energy Renovation Rate 0.485 0.206 Building Data Availability 0.458 0.520 Digital Twin Process Standardisation 0.020 0.830 Availability of Simulation-Based Retrofit Tools 0.020 0.520 Cost of IoT/Sensor Integration 0.233 0.276 Synchronous Energy Prediction 0.406 0.716 Digital Twin Return on Investment Period 0.406 0.000 Energy Optimisation Capability 0.000 0.401 Six candidate deep leverage points were identified, combining high driving power with significant dependence on upstream drivers (Table 5 ): Government Incentives and Funding, DT Adoption Rate, Integrated AI–IoT–DT Solutions, Quality of Retrofit Decision-Making, Interoperability, and Technical Capacity. Based on their structural position in the network, these variables sit at critical junctions where change could propagate through multiple feedback pathways simultaneously. They are therefore identified here as candidates for prioritised intervention; whether they function as genuine leverage points in practice would require confirmation through quantitative system dynamics modelling, which is identified as a direction for future research in Section 5.4 . 5. Discussion 5.1 Interpreting the Causal Loop Diagram: Why the System Is Stuck The CLD reveals a domestic retrofit sector caught in a low-adoption equilibrium, shaped by the interplay of reinforcing and balancing feedback loops. On paper, the system hosts powerful engines for change: R1 (Policy-Driven Adoption), R2 (Data-Quality), R3 (Economies of Scale and Innovation), R4 (Standardisation and Capacity), and R5 (Simulation Investment) all describe pathways through which early DT successes could, in principle, generate self-reinforcing growth in adoption, performance, and cost reduction. In practice, these virtuous cycles have not yet dominated system behaviour, because they are systematically countered by B1 (Policy Exhaustion) and B2 (Demand-Driven Cost Inflation). High upfront costs for IoT integration keep R3 (Economies of Scale and Innovation) from reaching the scale required for significant learning-by-doing effects, even under existing incentive schemes. This mirrors findings from Jradi et al. ( 2023 ), who demonstrate that without sustained policy support, retrofit-oriented DTs remain confined to pilot projects rather than scaling across the building stock. Simultaneously, weak interoperability and immature standards mean that R4 (Standardisation and Capacity) remains underdeveloped: each new project incurs bespoke integration effort, reinforcing perceptions of DTs as risky and one-off solutions rather than components of an emerging socio-technical regime ( Sobowale et al. 2023 ; Ammar et al. 2022 ). This structural immaturity is not unique to the UK; within a European multi-country context, Doukari and Suliman ( 2024 ) document similar challenges in aligning software tools, organisational processes, and data governance in their renovation digital twin for public buildings. Across all loops, Stakeholder Trust functions as a linchpin variable. As evidenced by its eigenvector centrality score (0.111) and its structural position at the intersection of R1 (Policy-Driven Adoption), R2 (Data-Quality), R3 (Economies of Scale and Innovation), R4 (Standardisation and Capacity), R5 (Simulation Investment), and B2 (Demand-Driven Cost Inflation), trust mediates whether reinforcing feedback loops can take hold or whether balancing forces prevail. Every high-profile policy reversal, such as the early termination of the Green Homes Grant ( GOV.UK 2021 ) , and every problematic retrofit outcome erodes trust among homeowners, contractors, and investors, weakening the reinforcing loops and strengthening the balancing ones. The DanRETwin project ( Jradi et al. 2023 ) illustrates the converse: in the Danish context, transparent performance data underpinned decision-making and supported more confident investment in data-driven retrofits, demonstrating how R2 (Data-Quality Loop) and R5 (Simulation Investment) can shorten perceived ROI periods when trust is established. This cross-national evidence reinforces the argument advanced in Section 2.1 that the three structural tensions identified in Table 1 , cost versus innovation, privacy versus data utilisation, and standardisation versus flexibility, are not independent problems but mutually reinforcing dimensions of a single trust deficit, each of which suppresses a different subset of the reinforcing loops. This configuration is consistent with Kim’s ( 2000 ) “Growth and Underinvestment” archetype. The system contains structural potential for rapid diffusion of DT-enabled retrofits, but repeated underinvestment in enabling conditions—stable policy incentives, interoperable standards, and socio-technical capacity—has produced a capability trap. Short-term funding pulses activate R1 (Policy-Driven Adoption) and R3 (Economies of Scale and Innovation) briefly, but insufficient attention to R4 (Standardisation and Capacity) and the B2 (Demand-Driven Cost Inflation) constraint (labour and supply-chain bottlenecks) leads to cost inflation, disappointing outcomes, and policy fatigue. The result, in MLP terms, is regime lock-in: the incumbent socio-technical configuration — characterised by low-margin contracting, fragmented standards, and risk-averse procurement — reproduces itself through the very feedback loops that should enable transition, consuming the resources that would otherwise build the niche capabilities required to destabilise it ( Geels 2011 ) . Breaking this impasse requires the co-activation of multiple reinforcing loops while deliberately relaxing the balancing constraints, a task for which, as the leverage-point analysis in Section 4 demonstrates, coordinated government action on incentive design, standardisation governance, and workforce development constitutes the structurally indicated entry point. 5.2 Leverage Points and Possible Implications for Policy-Makers The eigenvector centrality and MICMAC results identify six candidate deep leverage points: Government Incentives and Funding, Digital Twin Adoption Rate, Integrated AI–IoT–DT Solutions, Interoperability, Technical Capacity, and Quality of Retrofit Decision-Making. These variables are not simply the most important in isolation; rather, they occupy structurally central positions from which change propagates through multiple feedback pathways simultaneously. From an MLP perspective, they represent the points at which niche-level momentum can most effectively interface with regime-level structures — policy frameworks, standards bodies, and procurement norms — to produce a durable transition rather than temporary market stimulus. The following three clusters of intervention map directly onto these leverage points and are corroborated by the wider empirical literature. 5.2.1 Designing Stable, Outcome-Linked Incentives Government Incentives and Funding are the most central variable in the network (eigenvector centrality: 0.168; MICMAC Driving Power: 0.806), and the CLD makes clear that not all incentive designs are equivalent in their systemic effects. To activate R1 (Policy-Driven Adoption) and R2 (Data-Quality) sustainably while avoiding the B1 Policy Exhaustion loop, incentives must be both multi-year and predictable, and tied to measured performance rather than inputs. The specific instrument recommended here is a performance-based rebate scheme administered by DESNZ, structured to pay out per verified kilowatt-hour saved or tonne of CO₂ avoided — as independently measured by DTs — rather than per installation completed. This design would directly reinforce the Data-Quality (R2) and Simulation Investment (R5) loops by rewarding accurate modelling and installation practice, while creating a market incentive for DT vendors to improve measurement precision. Empirical evidence confirms that where robust monitoring and verification are in place, perceived ROI periods shorten and investor confidence increases ( Jradi et al. 2023 ; Elghaish et al. 2024 ). The Green Homes Grant illustrates the converse: its abrupt withdrawal in 2021, less than a year after launch, triggered B1 (Policy Exhaustion) and generated a boom–bust cycle that eroded the Stakeholder Trust variable on which all five reinforcing loops depend. Critically, the credibility of this instrument depends on its duration. Long-term feed-in tariffs in the UK renewables sector demonstrate that stable, well-signalled policy commitments — rather than competitive short-cycle grant rounds — create the investment certainty that enables supply-side actors to scale. The Warm Homes Plan ( DESNZ 2026 ) , with its £13.2–15 billion five-year commitment, represents an initial step in this direction. For it to activate R3 (Economies of Scale) and R4 (Standardisation and Capacity) as well as R1 (Policy-Driven Adoption), the scheme must include an explicit commitment to a ten-year trajectory with pre-announced review points, rather than an improvised rebuild. DESNZ and HM Treasury should co-publish this trajectory as a long-term signal to the market that the transition is irreversible — reframing the government's role from grant-giver to socio-technical transition manager. 5.2.2 Governing for Interoperability and Shared Data The R4 (Standardisation and Capacity) loop, in combination with the MICMAC results for Interoperability (Driving Power: 0.625) and Digital Twin Process Standardisation (Driving Power: 0.830), positions standardisation governance as the pivotal mechanism for shifting the system from fragmented pilots to a coherent regime. Without common data models and open application programming interfaces, each additional DT project reproduces integration costs and vendor lock-in risks documented across the built-environment literature ( Ammar et al. 2022 ; Sobowale et al. 2023 ; Zhu et al. 2024 ). By contrast, Doukari and Suliman's (2024) renovation digital twin, developed across a European multi-country context, demonstrates how a more standardised organisational ecosystem can support continuous performance monitoring across multiple buildings — precisely the scalable, trust-generating outcome that activates R2 (Data-Quality) and R5 (Simulation Investment). Two complementary instruments follow from this analysis. First, DESNZ and the British Standards Institution (BSI) should jointly convene a National Digital Retrofit Standards Forum, comprising technology providers, retrofit coordinators, housing associations, and local authorities, tasked with co-developing open data standards for retrofit-oriented DTs within 24 months. Building on existing BSI PAS 2035 and PAS 2038 retrofit frameworks, compliance with these standards should become an eligibility condition for public subsidy under the scheme proposed in Section 5.2.1 . This directly reinforces R4 (Standardisation and Capacity), reduces perceived technical risk for SMEs, and promotes competition on service quality rather than proprietary data formats. A critical risk applies: premature or overly rigid standardisation could freeze sub-optimal designs or entrench incumbents at the expense of smaller innovators. Governance arrangements should therefore emphasise modular standards and iterative two-year revision cycles, maintaining agreed interface boundaries while allowing technical evolution within them. Second, and closely linked to both R2 (Data-Quality) and the Quality of Retrofit Decision-Making leverage point, the same Forum should be mandated to oversee a National Measured Performance Database (NMPD): a centralised, open-access repository of verified pre- and post-retrofit energy performance data, DT simulation accuracy metrics, and cost benchmarks, drawn from all publicly funded retrofit projects. Unlike proprietary vendor databases, the NMPD would be operated under a public-interest governance model — administered through a dedicated DESNZ agency — with data submitted in anonymised, standardised formats as a condition of receiving public subsidy. This operationalises the R2 Data-Quality loop at a national scale: growing empirical evidence improves decision-making quality, shortens the DT ROI period, and rebuilds the stakeholder trust that the B1 (Policy Exhaustion) cycle has historically eroded. Key operational challenges must be acknowledged: sustained public funding for the database infrastructure, GDPR compliance in managing building-level data, and the risk of low data quality if submission requirements are not enforced. These challenges are not insurmountable — analogous databases operate successfully in the energy performance certificate (EPC) register and the Display Energy Certificate (DEC) scheme — but they require deliberate governance design from inception rather than as an afterthought. 5.2.3 Building Socio-Technical Capacity Ahead of Demand Loop B2 (Demand-Driven Cost Inflation) highlights a structural risk: a successful demand-side stimulus, whether through R1 (Policy-Driven Adoption) or R3 (Economies of Scale), will outstrip supply-chain capacity if Technical Capacity is not treated as a precondition rather than an afterthought. The MICMAC results confirm this quantitatively — Technical Capacity exhibits the highest Dependence score in the network (0.984) alongside a Driving Power of 0.548, classifying it as a deeply embedded relay variable that amplifies the effect of every other intervention. Ammar et al. ( 2022 ) and Sobowale et al. ( 2023 ) both report that skills deficits and fragmented professional responsibilities are among the most persistent obstacles to realising the performance benefits of DTs in the built environment, corroborating the structural centrality of this variable in the CLD. The specific intervention recommended here is the establishment of a National Digital Retrofit Skills Academy (NDRSA), jointly funded by DESNZ and the Department for Education, and delivered in partnership with further education colleges, universities, and industry bodies such as the Retrofit Academy and CIBSE. The NDRSA would deliver: national accreditation pathways for "digital retrofit specialists" trained in DT operation, IoT integration, and data-driven building performance analysis; integration of DT-related competencies into existing retrofit coordinator curricula as mandated under PAS 2035; and regional SME hubs providing access to shared tools, testbeds, and advisory support to reduce the cost barrier for smaller contractors. When aligned with the stable long-term demand signals recommended in Section 5.2.1 , these measures can shift the cost trajectory of IoT and integration from scarcity-driven inflation to learning-driven decline, enabling R3 (Economies of Scale) to dominate over B2 (Demand-Driven Cost Inflation). Evidence from DanRETwin and comparable cross-disciplinary initiatives suggests that where institutionalised team structures are established, DTs transition from exceptional project interventions to standard elements of retrofit planning and monitoring ( Jradi et al. 2023 ; Sobowale et al. 2023 ). Finally, the three interventions proposed across Sections 5.2.1 to 5.2.3 are designed as a coordinated package, not isolated measures. Multi-actor governance arrangements — specifically, a cross-departmental Retrofit Transition Taskforce comprising DESNZ, DLUHC, BSI, industry representatives, and academic partners — should be established to monitor the CLD's key indicators (Stakeholder Trust levels, DT adoption rates, IoT cost trajectories, and training capacity utilisation) and iteratively adjust the policy mix as the system evolves. This adaptive governance function, consistent with the use of systems models as living learning tools rather than one-off analytical outputs ( Senge and Sterman 1992 ) , is essential to detecting and pre-empting the emergence of new balancing loops before they stabilise the system at a new, equally suboptimal equilibrium. 5.3 Contribution to SDGs and Systems Thinking in Practice By reframing DT adoption as a socio-technical transition rather than a technology deployment problem, this study advances the argument that progress towards SDGs 7 (Affordable and Clean Energy), 9 (Industry, Innovation and Infrastructure), 11 (Sustainable Cities and Communities), and 13 (Climate Action) in the built environment cannot be achieved through siloed technical or financial interventions. As established in Section 1 , retrofitting the UK housing stock is a structurally complex challenge that has resisted conventional policy solutions; the CLD makes the structural reasons for this resistance visible. Each SDG target addressed by DT-enabled retrofit, energy efficiency (SDG 7.3), reduced urban environmental impact (SDG 11.6), and integrated climate policy (SDG 13.2), requires not merely the deployment of technology but the sustained alignment of policy incentives, market structures, and socio-technical capacity that the reinforcing loops describe. The three-cluster intervention framework developed in Section 5.2 constitutes, in effect, a theory of change for SDG delivery in the built environment: one that is grounded in the causal structure of the system rather than aspirational targets alone. Methodologically, this study demonstrates the analytical utility of Causal Loop Diagrams for SDG-oriented research in contested policy domains. Unlike prescriptive barrier lists or static stakeholder maps — which characterise much of the existing DT literature ( Attaran and Celik 2023 ; Elghaish et al. 2024 ) — the CLD renders feedback structure visible, enabling researchers and policymakers to reason about second-order consequences, systemic trade-offs, and the conditions under which well-intentioned interventions may backfire. As noted in Section 5.4 , the model has not yet been validated through direct stakeholder engagement; its communicative potential — as a tool for surfacing and negotiating competing mental models among homeowners, contractors, and government actors — therefore represents a concrete direction for future research rather than a demonstrated outcome of this study. The analytical framework itself is transferable: analogous CLD-based analyses could be developed for other SDG-aligned socio-technical transitions, such as green hydrogen adoption or circular-economy procurement, where similar archetypes of growth and underinvestment are likely to operate. 5.4 Limitations and Directions for Future Research This study’s contribution lies in rendering the feedback structure of the UK’s DT retrofit transition explicit and actionable. However, several limitations must be acknowledged, each of which points to a concrete research priority. First, the CLD is a qualitative model. It identifies variables, link polarities, and feedback loops, but does not quantify causal magnitudes or time delays, and therefore cannot predict the pace of change under different policy scenarios or the thresholds at which reinforcing loops would dominate over balancing ones. A natural and necessary next step is to translate the CLD into a parameterised system dynamics model, informed by emerging data from pilot programmes and national schemes. This would enable rigorous testing of policy packages, combinations of incentive design, training investment, and standardisation measures, and exploration of tipping points and unintended consequences in a manner that qualitative diagramming cannot achieve ( Senge and Sterman 1992 ). Second, the model is constructed primarily from secondary literature and conceptual reasoning rather than direct stakeholder engagement. The four primary studies provide a robust starting point, but they inevitably reflect the emphases and blind spots of existing scholarship, and cannot substitute for the situated knowledge that practitioners, policymakers, and building occupants hold about how the UK retrofit system actually operates. Future validation work should proceed in two stages. In the first stage, semi-structured interviews with 15–20 purposively sampled UK-based experts — drawn from DESNZ, DLUHC, retrofit technology providers, PAS 2035 retrofit coordinators, housing associations, and tenant representatives — would test whether the 15 variables and their causal polarities reflect practitioners' understanding of the system, and surface mechanisms and relationships that the literature does not capture, such as political economy pressures, procurement culture, and the lived experience of retrofit disruption. It is essential that this validation sample is grounded in the UK context: engaging practitioners from other national systems, such as Denmark or the Netherlands, would risk importing assumptions about policy environments, market structures, and building typologies that do not transfer directly, potentially undermining the contextual validity of the recommendations made in Section 5.2 . In the second stage, group model-building workshops — structured around the CLD as a shared artefact — would bring together mixed stakeholder groups to negotiate contested causal links, test the recommendations against their operational realities, and build the collective ownership that the model's policy implications ultimately require ( Hovmand 2014 ). Beyond validation, these workshops would simultaneously serve the communicative purpose identified in Section 5.3 : surfacing where different stakeholder groups' mental models diverge from the structural analysis and from each other, which is itself a valuable finding for understanding why the system has resisted change. Third, the analysis bounds the system at the level of the UK domestic retrofit sector, treating exogenous influences—global supply-chain shocks, macroeconomic cycles, and diffusion of DT standards from other countries—as background conditions. This is defensible for an initial scoping model, but it limits transferability under substantially different external contexts. Future research could extend the analysis in two directions: comparative CLDs across countries with different policy regimes (for example, Denmark, the Netherlands, and selected EU member states); and integrated models coupling the domestic retrofit system to upstream supply chains and international standardisation processes. Finally, the study concentrates on energy and carbon outcomes (SDGs 7 and 13) and, to a lesser extent, urban sustainability (SDG 11). Distributional impacts, co-benefits, and potential rebound effects are not explicitly modelled, and this represents a meaningful gap as DT-enabled retrofit programmes begin to scale. Mixed-methods research integrating system dynamics modelling with empirical evaluation of social and economic outcomes is needed to assess whether DT adoption genuinely advances a broader SDG portfolio or risks generating new trade-offs. Three specific directions are proposed. First, quantitative impact assessment: future work could calibrate the CLD against emerging pilot programme data to estimate concrete outcomes — for example, tonnes of CO₂ avoided annually per percentage-point increase in DT adoption, or the threshold adoption rate at which R3 (Economies of Scale) begins to reduce IoT costs measurably. Second, rebound effect analysis: if smart home systems make occupants more comfortable, increasing their energy use, the net carbon benefit of retrofit may be partially offset; this behavioural dynamic is absent from the current model and warrants explicit investigation. Third, inter-SDG spillover mapping: improvements driven by DT-enabled retrofit through SDGs 7 and 13 may generate positive or negative spillovers to SDG 8 (Decent Work and Economic Growth) through green job creation in the retrofit and digital sectors, and to SDG 9 (Industry, Innovation and Infrastructure) through the development of a domestic DT supply chain. These connections are structurally implied by the CLD but not modelled. Embedding pre-agreed measurement frameworks within pilot programmes — such as the regional retrofit trials recommended in Section 5.2.1 — from their inception would provide the empirical foundation needed to progressively enrich the model as the transition unfolds. 6. Conclusions This study has transformed a sectoral challenge, accelerating the adoption of digital twin technology for UK building retrofits, into a systemic narrative and strategic roadmap. Using a systems thinking lens, a causal loop model was constructed that illuminates how policy, technology, economic, and social factors interact to shape the trajectory of innovation in the UK's domestic retrofitting sector. The analysis revealed five reinforcing loops representing pathways to rapid progress, and two balancing loops explaining past stagnation. Government Incentives and Funding emerged as the highest-leverage node, but the analysis cautions that financial investment alone is insufficient without parallel commitments to interoperability, workforce capacity, and data transparency. The system is currently held in a low-adoption equilibrium by the B1 Policy Exhaustion and B2 Demand-Driven Cost Inflation loops, which neutralise the reinforcing engines before they can reach self-sustaining momentum. In terms of the Multi-Level Perspective, the UK's DT retrofit transition remains in a pre-breakthrough phase: the niche has demonstrated technical viability in isolated cases, but the regime-level conditions — policy stability, interoperability infrastructure, and workforce capacity — have not yet shifted sufficiently to enable the niche to scale ( Geels 2011 ) . By embracing the four systemic recommendations — a stable, outcome-linked funding framework, open interoperability standards, a national measured performance database, and investment in socio-technical capacity — the UK can transform its built environment from a carbon liability into a digitally enabled asset. Such an outcome directly advances SDGs 7, 9, 11, and 13 while providing a replicable, systems-informed blueprint for other national contexts. As the 2025 SDG Progress Report confirms that only 18% of global targets are on track ( United Nations 2025 ) , approaches that identify and activate systemic leverage points rather than addressing barriers in isolation are urgently needed. Declarations Competing Interests: The authors declare no competing interests. Funding: No funding was received for conducting this study. The authors did not receive support from any organisation for the submitted work. Author Contribution Conceptualization: A.Z.W, I.P, S.VB; Methodology: A.Z.W, I.P; Formal analysis and investigation: A.Z.W; Writing – original draft preparation: A.Z.W; Writing – review and editing: I.P, S.VB; Supervision: I.P, S.VB. Data Availability The Causal Loop Diagram constructed in this study is publicly accessible via Kumu at: https://kumu.io/alejandrozhou/cld-analysis. No other datasets were generated or analysed during this study References Ackermann, F., Eden, C. (2011). Strategic Management of Stakeholders: Theory and Practice. Long Range Planning , 44 (3), 179–196. https://doi.org/10.1016/j.lrp.2010.08.001 Ammar, A., Nassereddine, H., AbdulBaky, N., AbouKansour, A., Tannoury, J., Urban, H., Schranz, C. (2022) Digital twins in the construction industry: a perspective of practitioners and building authority. Front Built Environ 8:834671. https://doi.org/10.3389/fbuil.2022.834671 Attaran, M., Celik, B.G. (2023). Digital twin: benefits, use cases, challenges, and opportunities. Decis Anal J 6:100165. https://doi.org/10.1016/j.dajour.2023.100165 Coletta, V. R., Pluchinotta, I., Pisinaras, V., Panagopoulos, A., Giordano, R., Pagano, A., Fratino, U., Montanari, A. (2025). Leverage points and cascading impacts analysis in nexus systems using system dynamics modeling. Earth's Future 13(7). https://doi.org/10.1029/2025ef006190 DESNZ (2026). Warm Homes Plan. GOV.UK. https://www.gov.uk/government/publications/warm-homes-plan Doukari, O., Suliman, A. (2024). Renovation digital twin for building retrofit monitoring: a software product and an organizational ecosystem. In: Proc Eur Conf Comput Constr (EC3 2024). https://ec-3.org/publications/conferences/EC32024/papers/EC32024_225.pdf. Accessed 24 Apr 2025 Elghaish, F., Matarneh, S., Hosseini, M.R., Tezel, A., Mahamadu, A.M., Taghikhah, F. (2024). Predictive digital twin technologies for achieving net zero carbon emissions: a critical review and future research agenda. Smart Sustain Built Environ. https://doi.org/10.1108/sasbe-03-2024-0096 Geels, F.W. (2011). The multi-level perspective on sustainability transitions: responses to seven criticisms. Environ Innov Soc Transit 1:24–40. https://doi.org/10.1016/j.eist.2011.02.002 GOV.UK (2021). Apply for the Green Homes Grant scheme. https://www.gov.uk/guidance/apply-for-the-green-homes-grant-scheme. Accessed 2021 HM Government (2021). Heat and Buildings Strategy. https://assets.publishing.service.gov.uk/media/61d450eb8fa8f54c14eb14e4/6.7408_BEIS_Clean_Heat_Heat___Buildings_Strategy_Stage_2_v5_WEB.pdf Hovmand, P.S. (2014). Community based system dynamics. Springer, New York. https://doi.org/10.1007/978-1-4614-8763-0 Jradi, M., Madsen, H., Kaiser, M. (2023). DanRETwin: a digital twin solution for optimal energy retrofit decision-making and decarbonization of the Danish building stock. Appl Sci 13:9778. https://doi.org/10.3390/app13179778 Kaewunruen, S., Rungskunroch, P., Welsh, J. (2019). A digital-twin evaluation of net zero energy building for existing buildings. Sustainability 11:159. https://doi.org/10.3390/su11010159 Kim, D.H. (2000). Diagnosing systemic issues and designing high-leverage interventions. Syst Thinker 11(1) Kim, H., Andersen, D.F. (2012). Building confidence in causal maps generated from purposive text data: mapping transcripts of the Federal Reserve. Syst Dyn Rev 28:311–328. https://doi.org/10.1002/sdr.1480 Kumu (2025). CLD analysis. https://kumu.io/alejandrozhou/cld-analysis#cld-leverage-points. Accessed 24 April 2025 Meadows, D. (1999). Leverage points: places to intervene in a system. Sustainability Institute, Hartland Okimi, T.O., Sarhan, S., Awodele, I.A., Olaniran, T.O., Adetola, O.D., Olagunju, O.A. (2024). Towards adopting digital twins for enhancing circular economy in the UK construction industry: benefits and enablers. https://doi.org/10.46254/na09.20240022 Pluchinotta, I., Salvia, G., Zimmermann, N. (2021). The importance of eliciting stakeholders' system boundary perceptions for problem structuring and decision-making. Eur J Oper Res 302:280–293. https://doi.org/10.1016/j.ejor.2021.12.029 Prioreschi, E., Zimmermann, N., Davies, M., Pluchinotta, I. (2024). Interrelationships and trade-offs between urban natural space use and biodiversity. Sustainability 16:4051. https://doi.org/10.3390/su16104051 Reed, M.S., Graves, A., Dandy, N., Posthumus, H., Hubacek, K., Morris, J., Prell, C., Quinn, C.H., Stringer, L.C. (2009). Who's in and why? A typology of stakeholder analysis methods for natural resource management. J Environ Manage 90:1933–1949. https://doi.org/10.1016/j.jenvman.2009.01.001 Saback, V., Popescu, C., Täljsten, B., Blanksvärd, T. (2023). Analysis of digital twins in the construction industry: current trends and applications. Lect Notes Civ Eng, pp 1080–1088. https://doi.org/10.1007/978-3-031-32511-3_110 Senge, P.M., Sterman, J.D. (1992) Systems thinking and organizational learning: acting locally and thinking globally in the organization of the future. Eur J Oper Res 59:137–150. https://doi.org/10.1016/0377-2217(92)90011-W Sobowale, M., Elghaish, F., Brooks, T. (2023). A systematic review of digital twin as a predictive maintenance approach for existing buildings in the UK. Proc e-report, pp 1206–1218. https://doi.org/10.36253/979-12-215-0289-3.119 Sghiri, A., Gallab, M., Merzouk, S., Assoul, S. (2025). Leveraging digital twins for enhancing building energy efficiency: a literature review of applications, technologies, and challenges. Buildings 15:498. https://doi.org/10.3390/buildings15030498 Sterman, J.D. (2000). Business dynamics: systems thinking and modeling for a complex world. McGraw-Hill, Boston UK Government (2017). Domestic private rented property: minimum energy efficiency standard — landlord guidance. https://www.gov.uk/guidance/domestic-private-rented-property-minimum-energy-efficiency-standard-landlord-guidance UKGBC (2023). Climate change mitigation. https://ukgbc.org/our-work/climate-change-mitigation/ United Nations (2023). The Sustainable Development Goals. https://sdgs.un.org/goals United Nations (2025). The Sustainable Development Goals Report 2025. https://unstats.un.org/sdgs/report/2025/The-Sustainable-Development-Goals-Report-2025.pdf Veldhuis, G.A., Smits-Clijsen, E.M., van Waas, R.P.M., Hof, T., Maccatrozzo, V., Rouwette, E.A.J.A., Kerstholt, J.H. (2025). The influence of causal loop diagrams on systems thinking and information utilization in complex problem-solving. Comput Hum Behav Rep 17:100613. https://doi.org/10.1016/j.chbr.2025.100613 Vensim (2025). Vensim Personal Learning Edition. https://vensim.com/vensim-personal-learning-edition/. Accessed 24 April 2025 Zdunek, R. (2024). How digital twin technology is transforming real estate sales. Chameleon Interactive. https://chameleon-interactive.com/2024/10/14/how-digital-twin-technology-is-transforming-real-estate-sales/ Zhu, H., Ye, G., Wu, Z., Han, Y. (2024). Building on digital twin: overcoming barriers and unlocking success in the construction industry. J Constr Eng Manage 150(10). https://doi.org/10.1061/JCEMD4.COENG-14234 Additional Declarations No competing interests reported. Supplementary Files Appendices.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 28 Apr, 2026 Reviewers agreed at journal 28 Apr, 2026 Reviewers agreed at journal 14 Apr, 2026 Reviewers invited by journal 10 Apr, 2026 Editor assigned by journal 31 Mar, 2026 Submission checks completed at journal 26 Mar, 2026 First submitted to journal 25 Mar, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9227696","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":622919126,"identity":"784c5031-6eac-4226-95cc-d0429607997f","order_by":0,"name":"Alejandro Zhou Wei","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAw0lEQVRIiWNgGAWjYJACZiCWA1INYJ4EsVqMGRgYSdSS2EC0Fvn2w4c/F1TYpW84frCB4UcNQ+LMBgJaGHvS0qRnnEnO3XAmsYGx5xhD4myCjpLgMWPmbWPO3XAD6DDeBobEeYS0sEnwf/7M21afbgDUwviXGC08EjwM0rxthxNAWphBthB0mARPmpk0z5njhjOBfjksc0zCmKD3gSH2+DNPRbU83/HDBx++qbGRnXGAkDXI4ACRETkKRsEoGAWjgBAAAK4MOtG3XZO+AAAAAElFTkSuQmCC","orcid":"","institution":"University College London","correspondingAuthor":true,"prefix":"","firstName":"Alejandro","middleName":"Zhou","lastName":"Wei","suffix":""},{"id":622919127,"identity":"4c269af7-8c76-42ec-a1ee-d2b8663160b6","order_by":1,"name":"Irene Pluchinotta","email":"","orcid":"","institution":"University College London","correspondingAuthor":false,"prefix":"","firstName":"Irene","middleName":"","lastName":"Pluchinotta","suffix":""},{"id":622919128,"identity":"ffa7ac98-1d7d-416c-857d-99fbe8b324f7","order_by":2,"name":"Simon Vakeva-Baird","email":"","orcid":"","institution":"University College London","correspondingAuthor":false,"prefix":"","firstName":"Simon","middleName":"","lastName":"Vakeva-Baird","suffix":""}],"badges":[],"createdAt":"2026-03-26 00:38:28","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9227696/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9227696/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":107305239,"identity":"8c26e988-1158-4a1e-b8c0-9dd324738463","added_by":"auto","created_at":"2026-04-20 08:13:26","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":84252,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003ePower–Interest Grid mapping stakeholders in the UK domestic retrofit system across four quadrants — Context Setters, Players, Crowd, and Subjects — classified by their level of power and interest in DT adoption outcomes. Players (including UK Government, DT software developers, and property developers) hold high power and high interest, while tenants and the general public occupy the Crowd quadrant with low power and low interest.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9227696/v1/a3d0f7f17c7ebbed6ba6a0fd.jpg"},{"id":107305236,"identity":"f044fc05-1d3f-464f-b46c-f73138d83e00","added_by":"auto","created_at":"2026-04-20 08:13:23","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":146421,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eFour-stage methodological workflow used to construct and analyse the Causal Loop Diagram, showing the sequential progression from systematic literature review and open coding, through CLD construction and feedback loop identification, to thematic cluster analysis and candidate leverage point identification via eigenvector centrality and MICMAC analysis\u003c/em\u003e\u003c/p\u003e","description":"","filename":"2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9227696/v1/c5ba1836d5a94050716c2582.jpg"},{"id":107305101,"identity":"d8eaea35-4bff-4844-b825-5413db4fc7a3","added_by":"auto","created_at":"2026-04-20 08:13:07","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":126072,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eCausal Loop Diagram of DT adoption mechanisms in UK housing retrofit, visualising 15 variables and seven feedback loops — five reinforcing (R1–R5, shown as closed self-amplifying pathways) and two balancing (B1–B2, shown as constraining pathways). Positive (+) links indicate a relationship in the same direction; negative (–) links indicate an inverse relationship. Constructed in Vensim PLE and analysed in Kumu (Vensim 2025; Kumu 2025)\u003c/em\u003e\u003c/p\u003e","description":"","filename":"3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9227696/v1/fd1795c1cc3eb990b0474767.jpg"},{"id":107305168,"identity":"135be039-1762-48dd-8c4c-85c8269c5518","added_by":"auto","created_at":"2026-04-20 08:13:13","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":101651,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eMICMAC grid classifying all 15 CLD variables by their driving power (vertical axis, influence on other variables) and dependence (horizontal axis, susceptibility to influence from other variables). Six candidate deep leverage points — identified as variables combining high driving power with significant dependence — are highlighted in the upper-right quadrant. Variables in the lower-left quadrant (low driving power and low dependence) have limited systemic influence. Constructed using a cross-impact matrix derived from the CLD.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9227696/v1/fa4d25096814d6bc84503f20.jpg"},{"id":107484619,"identity":"c672f5ea-0a62-4db1-aefa-3f440d4cf30f","added_by":"auto","created_at":"2026-04-22 02:32:32","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1009811,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9227696/v1/46cf1155-08a9-4e38-ba1a-2ba4db4ba780.pdf"},{"id":107305238,"identity":"f5e5d0e2-e033-4a3b-9993-baa817e3a201","added_by":"auto","created_at":"2026-04-20 08:13:25","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":30089,"visible":true,"origin":"","legend":"","description":"","filename":"Appendices.docx","url":"https://assets-eu.researchsquare.com/files/rs-9227696/v1/05d3508c922b350f051dbd43.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Governing the Digital Transition: A Systems Thinking Approach to Analyse Digital Twin Adoption in UK Housing Retrofit","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eUpgrading existing buildings is an urgent priority for climate change mitigation, particularly in the United Kingdom (UK), where the built environment accounts for approximately 25% of the UK\u0026rsquo;s total greenhouse gas emissions \u003cem\u003e(\u003c/em\u003eUKGBC \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2023\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e. The UK Government has introduced policies such as the Heat and Buildings Strategy \u003cem\u003e(\u003c/em\u003eHM Government \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2021\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e and Minimum Energy Efficiency Standards \u003cem\u003e(\u003c/em\u003eUK Government \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2017\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e to drive energy retrofitting in residential and commercial properties. However, despite these initiatives, progress remains slow due to persistent challenges: high upfront investment costs, fragmented building performance data, and slow market uptake of new technologies \u003cem\u003e(\u003c/em\u003eZhu et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). A recent study by Jradi et al. (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) noted that without addressing such systemic barriers, meeting decarbonisation goals in the building sector will be difficult.\u003c/p\u003e \u003cp\u003eOne promising technological innovation for enhancing retrofit outcomes is Digital Twin (DT) technology, a dynamic, real-time digital replica of a physical asset. By integrating Internet of Things (IoT) sensors, building information modelling (BIM) data, and AI analytics, digital twins create a feedback loop of information that can continuously improve building performance \u003cem\u003e(Attaran and Celi 2023).\u003c/em\u003e The potential benefits include reduced operational costs, improved energy efficiency, and verified carbon monitoring \u003cem\u003e(\u003c/em\u003eSghiri et al. \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), making DTs valuable tools for enabling the scaling of effective retrofit programmes.\u003c/p\u003e \u003cp\u003eBeyond its technical advantages, DT technology is closely associated with sustainability objectives. Implementing data-driven retrofits can directly contribute to SDG 7 (Affordable and Clean Energy), Target 7.3 on energy efficiency; SDG 11 (Sustainable Cities and Communities), Target 11.6 on reducing the environmental impact of cities; and SDG 13 (Climate Action), Target 13.2 on integrating climate measures into policy \u003cem\u003e(\u003c/em\u003eUnited Nations \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2023\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e. Despite this promise, adoption remains limited by high costs, technical and privacy challenges, and regulatory gaps, particularly in the context of UK housing retrofits.\u003c/p\u003e \u003cp\u003eThe persistence and interaction of these barriers indicate that the challenge is fundamentally systemic \u003cem\u003e(\u003c/em\u003eMeadows \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e1999\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e. High upfront costs shape investment decisions, which influence market demand and slow standardisation \u003cem\u003e(\u003c/em\u003eZhu et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Fragmented data infrastructures limit performance transparency, weakening trust among stakeholders and discouraging wider uptake \u003cem\u003e(\u003c/em\u003eJradi et al. \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Skills shortages reduce implementation capacity, prolonging uncertainty and constraining diffusion \u003cem\u003e(\u003c/em\u003eSobowale et al. \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). As a result, interventions that target single issues in isolation, such as cost reduction or technology deployment alone, struggle to produce sustained change. Understanding why this is the case requires situating these barriers within a broader theory of technological change. The Multi-Level Perspective (MLP) on socio-technical transitions offers such a framework, characterising DT adoption as a niche innovation attempting to destabilise an entrenched socio-technical regime \u003cem\u003e(\u003c/em\u003eGeels \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2011\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e.\u003c/p\u003e \u003cp\u003eA systems thinking approach provides a structured way to analyse this complexity. By focusing on relationships, feedback loops, and system behaviour, systems thinking enables the construction of causal models that make interactions, e.g., between cost, policy incentives, trust, and data availability explicit \u003cem\u003e(\u003c/em\u003eSterman \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2000\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e. This perspective supports the identification of potential leverage points where targeted interventions can influence system behaviour more effectively than incremental or siloed measures.\u003c/p\u003e \u003cp\u003eThis study aims to apply systems thinking to analyse the factors and interdependencies influencing DT adoption in the UK's domestic retrofit sector. The primary research question is: How do interactions between policy, technology, and market actors create systemic barriers and enablers to the adoption of digital twin technology for domestic building retrofits in the UK? Two sub-questions structure the investigation: (i) What are the key reinforcing and balancing feedback loops governing DT adoption, and how do they correspond to known system archetypes? (ii) Which factors have the greatest influence on the system, and what intervention strategies do they suggest for policymakers and the industry?\u003c/p\u003e \u003cp\u003eThis study demonstrates how a systems thinking approach can be used to advance sustainability transitions in the built environment. Previous studies have examined the technical aspects of DT or listed generic benefits and barriers, such as reduced operational costs, improved energy efficiency, and data interoperability challenges \u003cem\u003e(\u003c/em\u003eAttaran and Celik \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Elghaish et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). This paper situates these elements within a holistic causal structure of the system, leveraging a Causal Loop Diagram (CLD) to reveal deeper principles of socio-technical behaviour. The findings are useful for policymakers, particularly the Department for Energy Security and Net Zero (DESNZ) and the Department for Levelling Up, Housing and Communities (DLUHC), and industry stakeholders positioned to design integrated interventions.\u003c/p\u003e"},{"header":"2. Theoretical Framework","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 The UK Retrofit Sector: System Context and Stakeholders\u003c/h2\u003e \u003cp\u003eRetrofitting the UK's housing stock involves a diverse network of actors operating across different scales. Central government (particularly DESNZ and DLUHC) sets policy and funding frameworks. Local authorities and housing associations implement retrofit programmes. Technology providers (IoT manufacturers, BIM software firms, DT platforms) supply the digital infrastructure. Contractors and retrofit installers implement physical works. Homeowners and tenants occupy the retrofitted properties and must consent to and engage with the process. Following Reed et al.'s (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2009\u003c/span\u003e) typology, which advocates differentiating stakeholders by interest and influence to prioritise engagement, Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e presents a Power-Interest Grid mapping the principal actors in the UK domestic retrofit system \u003cem\u003e(\u003c/em\u003eAckermann and Eden \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2011\u003c/span\u003e\u003cem\u003e).\u003c/em\u003e Appendix A provides a detailed categorisation of stakeholder groups, outlining their roles, objectives, and the resources they contribute to or require from the system. Each actor has distinct interests, levels of technical expertise, and risk appetite, which shape their response to DT adoption.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThese actors operate with conflicting priorities that generate structural tensions directly relevant to DT adoption. Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e summarises the three principal tensions and their systemic implications, each of which is later reflected in the feedback loops of the CLD.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePrincipal stakeholder tensions in UK DT-enabled retrofit and their systemic implications.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTension\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eConflicting Actors\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNature of Conflict\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eImplication for DT Adoption\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCost vs. Innovation\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRetrofit contractors and developers vs. technology firms and consultants\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eShort-term cost minimisation vs. investment in advanced AI/IoT systems with unproven at-scale returns\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSuppresses DT Adoption Rate by limiting willingness to pay for IoT/sensor integration; directly reflected in the B2 Demand-Driven Cost Inflation loop \u003cem\u003e(\u003c/em\u003eZhu et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2024\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePrivacy vs. Data Utilisation\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHomeowners and occupants vs. DT vendors and service providers\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eConcerns over intrusive monitoring vs. demand for comprehensive real-time data collection to maximise DT utility\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCreates a trust deficit (Stakeholder Trust variable) that suppresses Building Data Availability and the R2 Data-Quality loop; requires transparent data governance to resolve \u003cem\u003e(\u003c/em\u003eJradi et al. \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2023\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eStandardisation vs. Flexibility\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGovernment agencies and standards bodies (e.g., BSI) vs. technology companies with proprietary platforms\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eOpen interoperability requirements vs. proprietary solutions that enable faster iteration but fragment the market\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMarket fragmentation suppresses Interoperability and Technical Capacity, blocking the R4 Standardisation loop; excessive rigidity risks stifling early innovation \u003cem\u003e(\u003c/em\u003eAmmar et al. \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Sobowale et al. \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2023\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThese tensions are not independent: the cost-innovation conflict amplifies the privacy-data conflict (high IoT costs reduce willingness to install comprehensive monitoring), while fragmented standards compound both by preventing cost reductions through scale. Together, they constitute the socio-technical lock-in that the CLD in Section \u003cspan refid=\"Sec10\" class=\"InternalRef\"\u003e4\u003c/span\u003e makes explicit.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Multi-Level Perspective on Socio-Technical Transitions\u003c/h2\u003e \u003cp\u003eTo understand the slow adoption of digital twins, this study integrates two theoretical streams: the Multi-Level Perspective (MLP) on socio-technical transitions and the praxis of systems thinking. The MLP conceptualises transitions as interactions between the socio-technical landscape (e.g., net-zero targets), the socio-technical regime (the stable, traditional construction industry), and niche innovations such as digital twins \u003cem\u003e(\u003c/em\u003eGeels \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2011\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e. DT adoption represents a niche-regime interaction where the technology must overcome the lock-in of a regime characterised by low margins and limited appetite for advanced innovation. Systems thinking, particularly causal loop diagramming, provides the analytical apparatus to map these niche-regime interactions, trace the feedback mechanisms that sustain regime lock-in, and identify leverage points for deliberate intervention \u003cem\u003e(\u003c/em\u003eVeldhuis et al. \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e"},{"header":"3. Methodology","content":"\u003cp\u003eThis study adopts a qualitative systems thinking modelling approach. The inquiry followed a structured four-stage process: (1) systematic literature review and evidence base; (2) variable extraction and conceptual coding; (3) causal loop diagram (CLD) construction and structural mapping; and (4) CLD analysis, including thematic cluster analysis and possible leverage point identification via eigenvector centrality and MICMAC calculation. Each stage is described in turn below. We informed this methodology with prior work on causal mapping from open coding of text data \u003cem\u003e(\u003c/em\u003ePrioreschi et al. \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Kim and Andersen \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2012\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e, CLD analysis \u003cem\u003e(\u003c/em\u003ePluchinotta et al. \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), and possible leverage point identification \u003cem\u003e(\u003c/em\u003eColetta et al. \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), adapting these methodologies to our context. Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e illustrates the step-by-step workflow applied.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Step 1 \u0026mdash; Systematic Literature Review and Evidence Base\u003c/h2\u003e \u003cp\u003eThe evidence base for this study was constructed through a systematic review of peer-reviewed literature and pivotal high-quality industry reports. Academic databases were searched\u0026mdash;including ScienceDirect, IEEE Xplore, MDPI, Google Scholar, and UCL Explore\u0026mdash;using keyword combinations related to 'Digital Twin,' 'Net Zero,' 'Retrofit,' 'UK,' and 'Energy optimisation.' These searches yielded a broad initial set of literature to which inclusion and exclusion criteria were applied (see Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eInclusion and Exclusion Criteria for the Systematic Literature Review\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"2\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInclusion Criteria\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eExclusion Criteria\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePeer-reviewed academic publications or high-quality industry reports\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eResearch papers without any clear application of Digital Twin technology or data-driven retrofit tools\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFocus on existing building stock (not new construction)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eStudies focusing only on new building designs or construction phases\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDiscussion of Digital Twin, AI, IoT, or BIM integration within the building sector and retrofitting\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eResearch with generic sustainability discussions but no mention of DT, IoT, or AI integration\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eExploration of barriers, enablers, policy context, or technical implementation in the UK or comparable settings\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eArticles lacking technical or empirical detail\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePublished 2018\u0026ndash;2024 to capture the contemporary landscape\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePurely conceptual studies without any application to retrofits\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThrough database querying and backward/forward citation snowballing, approximately 20 documents were screened at the abstract level before narrowing to a core set of four highly relevant studies: Jradi et al. (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), Sobowale et al. (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), Elghaish et al. (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), and Zhu et al. (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Appendix B provides a summary of each study's scope, contribution, and relevance to the variable extraction process. These sources, along with supporting references, formed the evidence base for extracting causal factors. It should be noted that the relatively small core set of primary studies reflects both the novelty of the topic and the specificity of the inclusion criteria; future work could expand this base, particularly through expert elicitation, to ensure comprehensive variable identification.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Step 2 \u0026mdash; Variable Extraction and Conceptual Coding\u003c/h2\u003e \u003cp\u003eIn the second stage, relevant system variables and causal relationships were extracted from the selected literature using an open coding approach \u003cem\u003e(\u003c/em\u003eadapted from Kim and Andersen \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2012\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e. This process translated qualitative insights into structured system components. The following categories of variables were sought:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003ePolicy and incentive factors (e.g., availability of government grants and regulatory standards).\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eTechnical system properties (e.g., interoperability of platforms, data infrastructure quality, and workforce skills).\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eEconomic and financial factors (e.g., upfront cost of IoT sensors and return on investment period).\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eSocial and behavioural factors (e.g., stakeholder trust in new technology and homeowner awareness).\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eKey outputs and performance metrics (e.g., energy savings achieved and retrofit rate of homes per year).\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eEach variable was clearly defined and, where possible, associated with a quantifiable measure to maintain conceptual clarity, even though the CLD itself is qualitative. Redundant or overlapping concepts from different sources were merged into a single variable to avoid unnecessary complexity. For instance, various mentions of 'lack of expertise,' 'skills gap,' and 'training needs' were consolidated into a single variable termed 'Technical Capacity of Workforce.' By the end of this coding process, 15 core variables were identified that recurred across sources and were most pertinent to DT adoption in retrofits. Each was classified as a Driver/Enabler, a Barrier, or an Outcome in the system. Appendix C provides full definitions, measurement indicators, and source attributions for all 15 variables.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Step 3 \u0026mdash; Causal Loop Diagram Construction\u003c/h2\u003e \u003cp\u003eThe third stage involved translating the identified variables and relationships into a Causal Loop Diagram (CLD) to represent the system's feedback structure. The CLD was constructed using Vensim PLE (an open source systems dynamics software, Vensim \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2025\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e and analysed in Kumu (an online systems mapping tool, Kumu \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Causal links between variables were drawn based on relationships documented in the literature or on logical inferences from domain knowledge. In a CLD, a '+' (positive) polarity indicates that if the cause increases, the effect increases above what it would otherwise have been; a '\u0026ndash;' (negative) polarity indicates the inverse. These polarities describe relative changes, not absolute magnitudes \u003cem\u003e(\u003c/em\u003eSterman \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2000\u003c/span\u003e\u003cem\u003e).\u003c/em\u003e\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e3.4 Step 4 \u0026mdash; Structural and Thematic Analysis and Candidate Leverage Point Identification\u003c/h2\u003e \u003cp\u003eWith the CLD constructed, feedback loops within the diagram were identified and analysed. A feedback loop is a closed chain of cause-and-effect links that 'feeds back' into itself \u003cem\u003e(\u003c/em\u003eSterman \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2000\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e. Seven principal feedback loops were found\u0026mdash;five reinforcing (R1\u0026ndash;R5) and two balancing (B1\u0026ndash;B2), and their causal pathways were traced and recorded (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). In identifying loops, established systems archetypes were referenced \u003cem\u003e(\u003c/em\u003eSenge and Sterman \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e1992\u003c/span\u003e; Meadows \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e1999\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e to characterise structural patterns such as virtuous cycles of growth or limit-imposing counteracting forces.\u003c/p\u003e \u003cp\u003eFirstly, a thematic cluster analysis (adapted from Pluchinotta et al. \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) was used to group the 15 CLD variables into broader thematic clusters based on their conceptual similarity and shared causal relationships. This approach helps to reveal focal areas of intervention by showing which domains \u0026mdash; policy, technology, economics, and social dynamics \u0026mdash; contain the most structurally interconnected variables, and where cross-domain dependencies create potential for coordinated action. Beyond visual inspection of loops, a structural analysis based on graph theory was conducted to identify candidate leverage points, the influential variables in Donella Meadows' terminology (1999). Using Kumu, the CLD was mapped as a directed graph, and eigenvector centrality was calculated for each variable. Eigenvector centrality is a network metric that assigns higher scores to nodes that are highly connected to other highly connected nodes, thereby identifying variables located within influential parts of the CLD.\u003c/p\u003e \u003cp\u003eFollowing the methodology and rationale described in Coletta et al. (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), a MICMAC (Matrix of Cross-Impact Multiplications Applied to Classification) analysis was applied to all 15 variables to cross-verify and enrich the eigenvector centrality results. This approach generated an influence-dependence score for each variable, quantifying its driving power over and dependence on other variables in the CLD. Plotting these scores on a grid of influence versus dependence allowed the identification of 'deep leverage point candidates' \u0026mdash; variables that are both highly influential and relatively autonomous within the network.\u003c/p\u003e \u003cp\u003eIt is important to note that this analysis quantifies structural connectivity rather than dynamic causal strength; centrality scores, therefore, indicate potential leverage points based on network structure, requiring further validation for their actual impact in a dynamic context.\u003c/p\u003e \u003c/div\u003e"},{"header":"4. Results","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e4.1 The Causal Loop Diagram and Feedback Loop Structure\u003c/h2\u003e \u003cp\u003eThe Causal Loop Diagram (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e) visualises the complex interdependencies between the 15 identified variables, spanning policy, technology, economics, and social dimensions, that describe the causal mechanisms governing DT adoption in the UK domestic retrofit sector. Seven feedback loops were identified and are summarised in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e below.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eOverview of CLD Feedback Loops\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLoop Name\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLabel\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCausal Pathway\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePolicy-Driven Adoption Loop\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eR1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGovernment Incentives \u0026rarr; (+) DT Adoption Rate \u0026rarr; (+) Energy Optimisation Capability \u0026rarr; (+) Energy Renovation Rate \u0026rarr; (+) Stakeholder Trust \u0026rarr; (+) Government Incentives\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eData-Quality Loop\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eR2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBuilding Data Availability \u0026rarr; (+) Quality of Retrofit Decision-Making \u0026rarr; (+) DT ROI Period \u0026rarr; (+) Stakeholder Trust \u0026rarr; (+) DT Adoption Rate \u0026rarr; (+) Building Data Availability\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEconomies of Scale and Innovation Loop\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eR3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCost of IoT/Sensor Integration \u0026rarr; (\u0026ndash;) Integrated AI-IoT-DT Solutions \u0026rarr; (+) Synchronous Energy Prediction \u0026rarr; (+) Energy Optimisation Capability \u0026rarr; (+) Energy Renovation Rate \u0026rarr; (+) Stakeholder Trust \u0026rarr; (+) DT Adoption Rate \u0026rarr; (+) Government Incentives \u0026rarr; (\u0026ndash;) Cost of IoT/Sensor Integration\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStandardisation Loop\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eR4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDT Process Standardisation \u0026rarr; (+) Interoperability \u0026rarr; (+) Technical Capacity \u0026rarr; (+) DT Adoption Rate \u0026rarr; (+) Stakeholder Trust \u0026rarr; (+) Government Incentives \u0026rarr; (+) DT Process Standardisation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSimulation Investment Tool Loop\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eR5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAvailability of Simulation-Based Retrofit Tools \u0026rarr; (+) Quality of Retrofit Decision-Making \u0026rarr; (+) DT ROI Period \u0026rarr; (+) Stakeholder Trust \u0026rarr; (+) DT Adoption Rate \u0026rarr; (+) Government Incentives \u0026rarr; (+) Availability of Simulation-Based Retrofit Tools\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePolicy Exhaustion Loop\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eB1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGovernment Incentives \u0026rarr; (+) DT Adoption Rate \u0026rarr; (+) Energy Renovation Rate \u0026rarr; (\u0026ndash;) Government Incentives\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDemand-Driven Cost Inflation Loop\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eB2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eStakeholder Trust \u0026rarr; (+) DT Adoption Rate \u0026rarr; (+) Cost of IoT/Sensor Integration \u0026rarr; (\u0026ndash;) Integrated AI-IoT-DT Solutions \u0026rarr; (\u0026ndash;) Stakeholder Trust\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cdiv id=\"Sec12\" class=\"Section3\"\u003e \u003ch2\u003e4.1.1 Reinforcing Loops (R1\u0026ndash;R5)\u003c/h2\u003e \u003cp\u003eThe five reinforcing loops constitute the potential engines of DT adoption in the system.\u003c/p\u003e \u003cp\u003e \u003cb\u003eR1 \u0026mdash; Policy-Driven Adoption Loop\u003c/b\u003e: This loop centres on the role of government support. Government Incentives stimulate the DT Adoption Rate, which improves Energy Optimisation Capability and, consequently, the Energy Renovation Rate. A higher renovation rate builds Stakeholder Trust, which in turn generates political and market pressure to sustain Government Incentives, closing a virtuous circle. This loop exemplifies a 'Success to the Successful' archetype \u003cem\u003e(\u003c/em\u003eMeadows \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e1999\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e: consistent policy incentives can kick-start a self-sustaining market transformation.\u003c/p\u003e \u003cp\u003e \u003cb\u003eR2 \u0026mdash; Data-Quality Loop\u003c/b\u003e: Building Data Availability (from IoT sensors) enhances the Quality of Retrofit Decision-Making, which shortens the DT Return on Investment (ROI) period. Demonstrated success increases Stakeholder Trust, leading to higher DT Adoption and further expansion of the monitored building data pool. This reinforcing loop indicates that early investment in data infrastructure creates a knowledge snowball effect: more data leads to better results, which leads to more adoption and, in turn, more data.\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eR3 \u0026mdash; Economies of Scale and Innovation Loop\u003c/strong\u003e \u003cp\u003eAs the DT Adoption Rate rises, it stimulates demand for Integrated AI-IoT-DT Solutions. Market growth enables technology providers to achieve economies of scale, reducing the Cost of IoT/Sensor Integration. Lower costs facilitate wider adoption, reinforcing the cycle. This loop resembles a 'Growth and Underinvestment' archetype, where scaling up the market lowers costs and lower costs further scale up the market.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eR4 \u0026mdash; Standardisation and Capacity Loop\u003c/strong\u003e \u003cp\u003eDT Process Standardisation improves Interoperability, which enhances the Technical Capacity of the workforce. A more capable workforce leads to more successful projects, boosting Stakeholder Trust and adoption. Greater adoption motivates Government Incentives and further standardisation efforts. R4 demonstrates how creating a virtuous cycle of standards, skills, and adoption can rapidly mainstream a technology, generating network effects that benefit each new adopter.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eR5 \u0026mdash; Simulation Investment Tool Loop\u003c/strong\u003e \u003cp\u003eGreater Availability of Simulation-Based Retrofit Tools improves the Quality of Retrofit Decision-Making and shortens the DT ROI Period. Improved economic outcomes boost Stakeholder Trust and DT Adoption, which encourages governments and investors to develop more sophisticated simulation tools. This loop indicates that investment in good planning tools pays off through better projects, which justifies further tool investment.\u003c/p\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section3\"\u003e \u003ch2\u003e4.1.2 Balancing Loops (B1\u0026ndash;B2)\u003c/h2\u003e \u003cp\u003eThe two balancing loops represent the braking mechanisms that currently suppress the reinforcing cycles and stabilise low adoption levels.\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eB1 \u0026mdash; Policy Exhaustion Loop\u003c/strong\u003e \u003cp\u003eAs Government Incentives drive higher DT Adoption and Energy Renovation Rate, the initial political will for generous incentives may diminish\u0026mdash;whether through perceptions that targets have been met, budget constraints, or policy shifts. Reduced incentives slow the adoption rate, counteracting the very gains produced by reinforcing loops. This 'Limits to Growth' archetype \u003cem\u003e(\u003c/em\u003eKim \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2000\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e has been observed empirically in schemes such as the UK Green Homes Grant (2020\u0026ndash;2021) \u003cem\u003e(\u003c/em\u003eGOV.UK \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2021\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e, where a premature withdrawal of funding reversed early market momentum and eroded stakeholder trust.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eB2 \u0026mdash; Demand-Driven Cost Inflation Loop\u003c/b\u003e: If DT adoption rises rapidly but Technical Capacity (skilled labour, supply chain for sensors) does not keep pace, the surge in demand produces bottlenecks. These manifest as price increases, higher costs for hiring qualified installers or sourcing IoT equipment, which erode Stakeholder Trust and dampen further adoption. This loop captures the 'immaturity of the supply chain' problem: without concurrently scaling up capacity, any demand-side push can become self-defeating, explaining why past efforts that simply injected public funds without a holistic strategy repeatedly failed to sustain retrofit markets.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e4.2 Thematic Cluster\u003c/h2\u003e \u003cp\u003eExamining the CLD, the 15 variables naturally grouped into four thematic clusters corresponding to different aspects of the system. This clustering confirms that any proposed intervention must consider all domains simultaneously, not just the technical or financial dimensions.\u003c/p\u003e \u003cp\u003e \u003col\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003ePolicy and Governance Cluster: Government Incentives and Funding and Digital Twin Process Standardisation. These shape the regulatory and support environment for innovation, underscoring the role of institutions in guiding transitions.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eTechnology and Infrastructure Cluster: Integrated AI-IoT-DT Solutions, Interoperability, Synchronous Energy Prediction, and related variables that determine the technical feasibility and performance of DT systems.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eEconomic and Market Viability Cluster: DT ROI Period, Cost of IoT/Sensor Integration, and Availability of Simulation Tools, key determinants of the cost-benefit calculus for potential adopters.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eSocial and Human Dynamics Cluster: Stakeholder Trust and Technical Capacity. This cluster represents the human dimension: professional readiness, acceptance of technology, and confidence in new approaches.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003c/ol\u003e \u003c/p\u003e \u003cp\u003eThese clusters are interdependent and not silos. The tension between Clusters 2 and 3\u0026mdash;investing in robust technology raises short-term costs, whilst neglecting it damages long-term performance and trust\u0026mdash;illustrates the multidimensionality of the problem and the need for integrated decision-making tools.\u003c/p\u003e \u003cp\u003eInterpreted through the MLP lens introduced in Section \u003cspan refid=\"Sec4\" class=\"InternalRef\"\u003e2.2\u003c/span\u003e, these clusters map onto the three analytical levels that structure socio-technical transitions. The Policy and Governance cluster and the Economic and Market Viability cluster operate predominantly at the regime level \u0026mdash; they describe the institutional rules, financial logics, and market expectations that currently reproduce low adoption. The Technology and Infrastructure cluster represents the niche level \u0026mdash; the emerging technical capabilities seeking to achieve sufficient scale and legitimacy to challenge regime practices. The Social and Human Dynamics cluster operates across all three levels: Stakeholder Trust is the currency through which niche actors gain regime acceptance, while Technical Capacity determines whether niche innovations can be absorbed into mainstream practice. In this reading, the CLD's reinforcing loops describe niche-level growth dynamics, and the balancing loops describe regime-level resistance mechanisms.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e4.3 Eigenvector Centrality Analysis\u003c/h2\u003e \u003cp\u003eEigenvector centrality results from Kumu identified Government Incentives and Funding (0.168) and DT Adoption Rate (0.115) as the most structurally central nodes, confirming that government action is the primary catalyst for the system (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Stakeholder Trust (0.111) also holds high centrality, suggesting that financial incentives alone are insufficient without accompanying social legitimacy. Cost of IoT/Sensor Integration (0.112) featured prominently due to its presence in multiple loops, both as a barrier suppressed by economies of scale and as an amplifier of demand-side pressures.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eEigenvector Centrality Analysis Results (Kumu, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2025\u003c/span\u003e)\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"2\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEigenvector Centrality Score\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGovernment Incentives and Funding\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.168\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCost of IoT/Sensor Integration\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.112\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDigital Twin Adoption Rate\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.115\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStakeholder Trust\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.111\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEnergy Renovation Rate\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.087\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAvailability of Simulation-Based Retrofit Tools\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.066\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDigital Twin Process Standardisation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.066\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEnergy Optimisation Capability\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.052\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBuilding Data Availability\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.045\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIntegrated AI-IoT-DT Solutions\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.044\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eQuality of Retrofit Decision-Making\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.044\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDigital Twin Return on Investment Period\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.035\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInteroperability\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.026\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSynchronous Energy Prediction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.018\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTechnical Capacity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.010\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eNote: Yellow-highlighted rows indicate variables identified as high-leverage nodes based on combined eigenvector centrality and MICMAC analysis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e4.4 MICMAC Analysis: Deep Leverage Points\u003c/h2\u003e \u003cp\u003eThe MICMAC grid (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e) classifies variables by their driving power (influence on others) and dependence (being influenced by others). Variables with high driving power and lower dependence represent autonomous drivers; those with both high driving power and high dependence are relay variables\u0026mdash;responsive to upstream changes but also capable of transmitting change downstream \u003cem\u003e(\u003c/em\u003eColetta et al. \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMICMAC Analysis Results. Deep Leverage Points are highlighted.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMICMAC Dependence\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMICMAC Driving Power (Exposure)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGovernment Incentives and Funding\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.608\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.806\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDigital Twin Adoption Rate\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.527\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIntegrated AI-IoT-DT Solutions\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.659\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eQuality of Retrofit Decision-Making\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.662\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInteroperability\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.534\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.625\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTechnical Capacity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.984\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.548\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStakeholder Trust\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.543\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.208\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEnergy Renovation Rate\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.485\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.206\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBuilding Data Availability\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.458\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.520\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDigital Twin Process Standardisation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.020\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.830\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAvailability of Simulation-Based Retrofit Tools\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.020\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.520\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCost of IoT/Sensor Integration\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.233\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.276\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSynchronous Energy Prediction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.406\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.716\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDigital Twin Return on Investment Period\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.406\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEnergy Optimisation Capability\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.401\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eSix candidate deep leverage points were identified, combining high driving power with significant dependence on upstream drivers (Table\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e): Government Incentives and Funding, DT Adoption Rate, Integrated AI\u0026ndash;IoT\u0026ndash;DT Solutions, Quality of Retrofit Decision-Making, Interoperability, and Technical Capacity. Based on their structural position in the network, these variables sit at critical junctions where change could propagate through multiple feedback pathways simultaneously. They are therefore identified here as candidates for prioritised intervention; whether they function as genuine leverage points in practice would require confirmation through quantitative system dynamics modelling, which is identified as a direction for future research in Section \u003cspan refid=\"Sec24\" class=\"InternalRef\"\u003e5.4\u003c/span\u003e.\u003c/p\u003e \u003c/div\u003e"},{"header":"5. Discussion","content":"\u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003e5.1 Interpreting the Causal Loop Diagram: Why the System Is Stuck\u003c/h2\u003e \u003cp\u003eThe CLD reveals a domestic retrofit sector caught in a low-adoption equilibrium, shaped by the interplay of reinforcing and balancing feedback loops. On paper, the system hosts powerful engines for change: R1 (Policy-Driven Adoption), R2 (Data-Quality), R3 (Economies of Scale and Innovation), R4 (Standardisation and Capacity), and R5 (Simulation Investment) all describe pathways through which early DT successes could, in principle, generate self-reinforcing growth in adoption, performance, and cost reduction. In practice, these virtuous cycles have not yet dominated system behaviour, because they are systematically countered by B1 (Policy Exhaustion) and B2 (Demand-Driven Cost Inflation). High upfront costs for IoT integration keep R3 (Economies of Scale and Innovation) from reaching the scale required for significant learning-by-doing effects, even under existing incentive schemes. This mirrors findings from Jradi et al. (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), who demonstrate that without sustained policy support, retrofit-oriented DTs remain confined to pilot projects rather than scaling across the building stock. Simultaneously, weak interoperability and immature standards mean that R4 (Standardisation and Capacity) remains underdeveloped: each new project incurs bespoke integration effort, reinforcing perceptions of DTs as risky and one-off solutions rather than components of an emerging socio-technical regime \u003cem\u003e(\u003c/em\u003eSobowale et al. \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Ammar et al. \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). This structural immaturity is not unique to the UK; within a European multi-country context, Doukari and Suliman (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) document similar challenges in aligning software tools, organisational processes, and data governance in their renovation digital twin for public buildings.\u003c/p\u003e \u003cp\u003eAcross all loops, Stakeholder Trust functions as a linchpin variable. As evidenced by its eigenvector centrality score (0.111) and its structural position at the intersection of R1 (Policy-Driven Adoption), R2 (Data-Quality), R3 (Economies of Scale and Innovation), R4 (Standardisation and Capacity), R5 (Simulation Investment), and B2 (Demand-Driven Cost Inflation), trust mediates whether reinforcing feedback loops can take hold or whether balancing forces prevail. Every high-profile policy reversal, such as the early termination of the Green Homes Grant \u003cem\u003e(\u003c/em\u003eGOV.UK \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2021\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e, and every problematic retrofit outcome erodes trust among homeowners, contractors, and investors, weakening the reinforcing loops and strengthening the balancing ones. The DanRETwin project \u003cem\u003e(\u003c/em\u003eJradi et al. \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) illustrates the converse: in the Danish context, transparent performance data underpinned decision-making and supported more confident investment in data-driven retrofits, demonstrating how R2 (Data-Quality Loop) and R5 (Simulation Investment) can shorten perceived ROI periods when trust is established. This cross-national evidence reinforces the argument advanced in Section \u003cspan refid=\"Sec3\" class=\"InternalRef\"\u003e2.1\u003c/span\u003e that the three structural tensions identified in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, cost versus innovation, privacy versus data utilisation, and standardisation versus flexibility, are not independent problems but mutually reinforcing dimensions of a single trust deficit, each of which suppresses a different subset of the reinforcing loops.\u003c/p\u003e \u003cp\u003eThis configuration is consistent with Kim\u0026rsquo;s (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2000\u003c/span\u003e) \u0026ldquo;Growth and Underinvestment\u0026rdquo; archetype. The system contains structural potential for rapid diffusion of DT-enabled retrofits, but repeated underinvestment in enabling conditions\u0026mdash;stable policy incentives, interoperable standards, and socio-technical capacity\u0026mdash;has produced a capability trap. Short-term funding pulses activate R1 (Policy-Driven Adoption) and R3 (Economies of Scale and Innovation) briefly, but insufficient attention to R4 (Standardisation and Capacity) and the B2 (Demand-Driven Cost Inflation) constraint (labour and supply-chain bottlenecks) leads to cost inflation, disappointing outcomes, and policy fatigue. The result, in MLP terms, is regime lock-in: the incumbent socio-technical configuration \u0026mdash; characterised by low-margin contracting, fragmented standards, and risk-averse procurement \u0026mdash; reproduces itself through the very feedback loops that should enable transition, consuming the resources that would otherwise build the niche capabilities required to destabilise it \u003cem\u003e(\u003c/em\u003eGeels \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2011\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e. Breaking this impasse requires the co-activation of multiple reinforcing loops while deliberately relaxing the balancing constraints, a task for which, as the leverage-point analysis in Section \u003cspan refid=\"Sec10\" class=\"InternalRef\"\u003e4\u003c/span\u003e demonstrates, coordinated government action on incentive design, standardisation governance, and workforce development constitutes the structurally indicated entry point.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003e5.2 Leverage Points and Possible Implications for Policy-Makers\u003c/h2\u003e \u003cp\u003eThe eigenvector centrality and MICMAC results identify six candidate deep leverage points: Government Incentives and Funding, Digital Twin Adoption Rate, Integrated AI\u0026ndash;IoT\u0026ndash;DT Solutions, Interoperability, Technical Capacity, and Quality of Retrofit Decision-Making. These variables are not simply the most important in isolation; rather, they occupy structurally central positions from which change propagates through multiple feedback pathways simultaneously. From an MLP perspective, they represent the points at which niche-level momentum can most effectively interface with regime-level structures \u0026mdash; policy frameworks, standards bodies, and procurement norms \u0026mdash; to produce a durable transition rather than temporary market stimulus. The following three clusters of intervention map directly onto these leverage points and are corroborated by the wider empirical literature.\u003c/p\u003e \u003cdiv id=\"Sec20\" class=\"Section3\"\u003e \u003ch2\u003e5.2.1 Designing Stable, Outcome-Linked Incentives\u003c/h2\u003e \u003cp\u003eGovernment Incentives and Funding are the most central variable in the network (eigenvector centrality: 0.168; MICMAC Driving Power: 0.806), and the CLD makes clear that not all incentive designs are equivalent in their systemic effects. To activate R1 (Policy-Driven Adoption) and R2 (Data-Quality) sustainably while avoiding the B1 Policy Exhaustion loop, incentives must be both multi-year and predictable, and tied to measured performance rather than inputs. The specific instrument recommended here is a performance-based rebate scheme administered by DESNZ, structured to pay out per verified kilowatt-hour saved or tonne of CO₂ avoided \u0026mdash; as independently measured by DTs \u0026mdash; rather than per installation completed. This design would directly reinforce the Data-Quality (R2) and Simulation Investment (R5) loops by rewarding accurate modelling and installation practice, while creating a market incentive for DT vendors to improve measurement precision. Empirical evidence confirms that where robust monitoring and verification are in place, perceived ROI periods shorten and investor confidence increases \u003cem\u003e(\u003c/em\u003eJradi et al. \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Elghaish et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The Green Homes Grant illustrates the converse: its abrupt withdrawal in 2021, less than a year after launch, triggered B1 (Policy Exhaustion) and generated a boom\u0026ndash;bust cycle that eroded the Stakeholder Trust variable on which all five reinforcing loops depend.\u003c/p\u003e \u003cp\u003eCritically, the credibility of this instrument depends on its duration. Long-term feed-in tariffs in the UK renewables sector demonstrate that stable, well-signalled policy commitments \u0026mdash; rather than competitive short-cycle grant rounds \u0026mdash; create the investment certainty that enables supply-side actors to scale. The Warm Homes Plan \u003cem\u003e(\u003c/em\u003eDESNZ \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2026\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e, with its \u0026pound;13.2\u0026ndash;15\u0026nbsp;billion five-year commitment, represents an initial step in this direction. For it to activate R3 (Economies of Scale) and R4 (Standardisation and Capacity) as well as R1 (Policy-Driven Adoption), the scheme must include an explicit commitment to a ten-year trajectory with pre-announced review points, rather than an improvised rebuild. DESNZ and HM Treasury should co-publish this trajectory as a long-term signal to the market that the transition is irreversible \u0026mdash; reframing the government's role from grant-giver to socio-technical transition manager.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section3\"\u003e \u003ch2\u003e5.2.2 Governing for Interoperability and Shared Data\u003c/h2\u003e \u003cp\u003eThe R4 (Standardisation and Capacity) loop, in combination with the MICMAC results for Interoperability (Driving Power: 0.625) and Digital Twin Process Standardisation (Driving Power: 0.830), positions standardisation governance as the pivotal mechanism for shifting the system from fragmented pilots to a coherent regime. Without common data models and open application programming interfaces, each additional DT project reproduces integration costs and vendor lock-in risks documented across the built-environment literature \u003cem\u003e(\u003c/em\u003eAmmar et al. \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Sobowale et al. \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Zhu et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). By contrast, Doukari and Suliman's (2024) renovation digital twin, developed across a European multi-country context, demonstrates how a more standardised organisational ecosystem can support continuous performance monitoring across multiple buildings \u0026mdash; precisely the scalable, trust-generating outcome that activates R2 (Data-Quality) and R5 (Simulation Investment).\u003c/p\u003e \u003cp\u003eTwo complementary instruments follow from this analysis. First, DESNZ and the British Standards Institution (BSI) should jointly convene a National Digital Retrofit Standards Forum, comprising technology providers, retrofit coordinators, housing associations, and local authorities, tasked with co-developing open data standards for retrofit-oriented DTs within 24 months. Building on existing BSI PAS 2035 and PAS 2038 retrofit frameworks, compliance with these standards should become an eligibility condition for public subsidy under the scheme proposed in Section \u003cspan refid=\"Sec20\" class=\"InternalRef\"\u003e5.2.1\u003c/span\u003e. This directly reinforces R4 (Standardisation and Capacity), reduces perceived technical risk for SMEs, and promotes competition on service quality rather than proprietary data formats. A critical risk applies: premature or overly rigid standardisation could freeze sub-optimal designs or entrench incumbents at the expense of smaller innovators. Governance arrangements should therefore emphasise modular standards and iterative two-year revision cycles, maintaining agreed interface boundaries while allowing technical evolution within them.\u003c/p\u003e \u003cp\u003eSecond, and closely linked to both R2 (Data-Quality) and the Quality of Retrofit Decision-Making leverage point, the same Forum should be mandated to oversee a National Measured Performance Database (NMPD): a centralised, open-access repository of verified pre- and post-retrofit energy performance data, DT simulation accuracy metrics, and cost benchmarks, drawn from all publicly funded retrofit projects. Unlike proprietary vendor databases, the NMPD would be operated under a public-interest governance model \u0026mdash; administered through a dedicated DESNZ agency \u0026mdash; with data submitted in anonymised, standardised formats as a condition of receiving public subsidy. This operationalises the R2 Data-Quality loop at a national scale: growing empirical evidence improves decision-making quality, shortens the DT ROI period, and rebuilds the stakeholder trust that the B1 (Policy Exhaustion) cycle has historically eroded. Key operational challenges must be acknowledged: sustained public funding for the database infrastructure, GDPR compliance in managing building-level data, and the risk of low data quality if submission requirements are not enforced. These challenges are not insurmountable \u0026mdash; analogous databases operate successfully in the energy performance certificate (EPC) register and the Display Energy Certificate (DEC) scheme \u0026mdash; but they require deliberate governance design from inception rather than as an afterthought.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section3\"\u003e \u003ch2\u003e5.2.3 Building Socio-Technical Capacity Ahead of Demand\u003c/h2\u003e \u003cp\u003eLoop B2 (Demand-Driven Cost Inflation) highlights a structural risk: a successful demand-side stimulus, whether through R1 (Policy-Driven Adoption) or R3 (Economies of Scale), will outstrip supply-chain capacity if Technical Capacity is not treated as a precondition rather than an afterthought. The MICMAC results confirm this quantitatively \u0026mdash; Technical Capacity exhibits the highest Dependence score in the network (0.984) alongside a Driving Power of 0.548, classifying it as a deeply embedded relay variable that amplifies the effect of every other intervention. Ammar et al. (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) and Sobowale et al. (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) both report that skills deficits and fragmented professional responsibilities are among the most persistent obstacles to realising the performance benefits of DTs in the built environment, corroborating the structural centrality of this variable in the CLD.\u003c/p\u003e \u003cp\u003eThe specific intervention recommended here is the establishment of a National Digital Retrofit Skills Academy (NDRSA), jointly funded by DESNZ and the Department for Education, and delivered in partnership with further education colleges, universities, and industry bodies such as the Retrofit Academy and CIBSE. The NDRSA would deliver: national accreditation pathways for \"digital retrofit specialists\" trained in DT operation, IoT integration, and data-driven building performance analysis; integration of DT-related competencies into existing retrofit coordinator curricula as mandated under PAS 2035; and regional SME hubs providing access to shared tools, testbeds, and advisory support to reduce the cost barrier for smaller contractors. When aligned with the stable long-term demand signals recommended in Section \u003cspan refid=\"Sec20\" class=\"InternalRef\"\u003e5.2.1\u003c/span\u003e, these measures can shift the cost trajectory of IoT and integration from scarcity-driven inflation to learning-driven decline, enabling R3 (Economies of Scale) to dominate over B2 (Demand-Driven Cost Inflation). Evidence from DanRETwin and comparable cross-disciplinary initiatives suggests that where institutionalised team structures are established, DTs transition from exceptional project interventions to standard elements of retrofit planning and monitoring \u003cem\u003e(\u003c/em\u003eJradi et al. \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Sobowale et al. \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFinally, the three interventions proposed across Sections \u003cspan refid=\"Sec20\" class=\"InternalRef\"\u003e5.2.1\u003c/span\u003e to \u003cspan refid=\"Sec22\" class=\"InternalRef\"\u003e5.2.3\u003c/span\u003e are designed as a coordinated package, not isolated measures. Multi-actor governance arrangements \u0026mdash; specifically, a cross-departmental Retrofit Transition Taskforce comprising DESNZ, DLUHC, BSI, industry representatives, and academic partners \u0026mdash; should be established to monitor the CLD's key indicators (Stakeholder Trust levels, DT adoption rates, IoT cost trajectories, and training capacity utilisation) and iteratively adjust the policy mix as the system evolves. This adaptive governance function, consistent with the use of systems models as living learning tools rather than one-off analytical outputs \u003cem\u003e(\u003c/em\u003eSenge and Sterman \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e1992\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e, is essential to detecting and pre-empting the emergence of new balancing loops before they stabilise the system at a new, equally suboptimal equilibrium.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec23\" class=\"Section2\"\u003e \u003ch2\u003e5.3 Contribution to SDGs and Systems Thinking in Practice\u003c/h2\u003e \u003cp\u003eBy reframing DT adoption as a socio-technical transition rather than a technology deployment problem, this study advances the argument that progress towards SDGs 7 (Affordable and Clean Energy), 9 (Industry, Innovation and Infrastructure), 11 (Sustainable Cities and Communities), and 13 (Climate Action) in the built environment cannot be achieved through siloed technical or financial interventions. As established in Section \u003cspan refid=\"Sec1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, retrofitting the UK housing stock is a structurally complex challenge that has resisted conventional policy solutions; the CLD makes the structural reasons for this resistance visible. Each SDG target addressed by DT-enabled retrofit, energy efficiency (SDG 7.3), reduced urban environmental impact (SDG 11.6), and integrated climate policy (SDG 13.2), requires not merely the deployment of technology but the sustained alignment of policy incentives, market structures, and socio-technical capacity that the reinforcing loops describe. The three-cluster intervention framework developed in Section \u003cspan refid=\"Sec19\" class=\"InternalRef\"\u003e5.2\u003c/span\u003e constitutes, in effect, a theory of change for SDG delivery in the built environment: one that is grounded in the causal structure of the system rather than aspirational targets alone.\u003c/p\u003e \u003cp\u003eMethodologically, this study demonstrates the analytical utility of Causal Loop Diagrams for SDG-oriented research in contested policy domains. Unlike prescriptive barrier lists or static stakeholder maps \u0026mdash; which characterise much of the existing DT literature \u003cem\u003e(\u003c/em\u003eAttaran and Celik \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Elghaish et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) \u0026mdash; the CLD renders feedback structure visible, enabling researchers and policymakers to reason about second-order consequences, systemic trade-offs, and the conditions under which well-intentioned interventions may backfire. As noted in Section \u003cspan refid=\"Sec24\" class=\"InternalRef\"\u003e5.4\u003c/span\u003e, the model has not yet been validated through direct stakeholder engagement; its communicative potential \u0026mdash; as a tool for surfacing and negotiating competing mental models among homeowners, contractors, and government actors \u0026mdash; therefore represents a concrete direction for future research rather than a demonstrated outcome of this study. The analytical framework itself is transferable: analogous CLD-based analyses could be developed for other SDG-aligned socio-technical transitions, such as green hydrogen adoption or circular-economy procurement, where similar archetypes of growth and underinvestment are likely to operate.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003e5.4 Limitations and Directions for Future Research\u003c/h2\u003e \u003cp\u003eThis study\u0026rsquo;s contribution lies in rendering the feedback structure of the UK\u0026rsquo;s DT retrofit transition explicit and actionable. However, several limitations must be acknowledged, each of which points to a concrete research priority.\u003c/p\u003e \u003cp\u003eFirst, the CLD is a qualitative model. It identifies variables, link polarities, and feedback loops, but does not quantify causal magnitudes or time delays, and therefore cannot predict the pace of change under different policy scenarios or the thresholds at which reinforcing loops would dominate over balancing ones. A natural and necessary next step is to translate the CLD into a parameterised system dynamics model, informed by emerging data from pilot programmes and national schemes. This would enable rigorous testing of policy packages, combinations of incentive design, training investment, and standardisation measures, and exploration of tipping points and unintended consequences in a manner that qualitative diagramming cannot achieve \u003cem\u003e(\u003c/em\u003eSenge and Sterman \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e1992\u003c/span\u003e\u003cem\u003e).\u003c/em\u003e\u003c/p\u003e \u003cp\u003eSecond, the model is constructed primarily from secondary literature and conceptual reasoning rather than direct stakeholder engagement. The four primary studies provide a robust starting point, but they inevitably reflect the emphases and blind spots of existing scholarship, and cannot substitute for the situated knowledge that practitioners, policymakers, and building occupants hold about how the UK retrofit system actually operates. Future validation work should proceed in two stages. In the first stage, semi-structured interviews with 15\u0026ndash;20 purposively sampled UK-based experts \u0026mdash; drawn from DESNZ, DLUHC, retrofit technology providers, PAS 2035 retrofit coordinators, housing associations, and tenant representatives \u0026mdash; would test whether the 15 variables and their causal polarities reflect practitioners' understanding of the system, and surface mechanisms and relationships that the literature does not capture, such as political economy pressures, procurement culture, and the lived experience of retrofit disruption. It is essential that this validation sample is grounded in the UK context: engaging practitioners from other national systems, such as Denmark or the Netherlands, would risk importing assumptions about policy environments, market structures, and building typologies that do not transfer directly, potentially undermining the contextual validity of the recommendations made in Section \u003cspan refid=\"Sec19\" class=\"InternalRef\"\u003e5.2\u003c/span\u003e. In the second stage, group model-building workshops \u0026mdash; structured around the CLD as a shared artefact \u0026mdash; would bring together mixed stakeholder groups to negotiate contested causal links, test the recommendations against their operational realities, and build the collective ownership that the model's policy implications ultimately require \u003cem\u003e(\u003c/em\u003eHovmand \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2014\u003c/span\u003e\u003cem\u003e).\u003c/em\u003e Beyond validation, these workshops would simultaneously serve the communicative purpose identified in Section \u003cspan refid=\"Sec23\" class=\"InternalRef\"\u003e5.3\u003c/span\u003e: surfacing where different stakeholder groups' mental models diverge from the structural analysis and from each other, which is itself a valuable finding for understanding why the system has resisted change.\u003c/p\u003e \u003cp\u003eThird, the analysis bounds the system at the level of the UK domestic retrofit sector, treating exogenous influences\u0026mdash;global supply-chain shocks, macroeconomic cycles, and diffusion of DT standards from other countries\u0026mdash;as background conditions. This is defensible for an initial scoping model, but it limits transferability under substantially different external contexts. Future research could extend the analysis in two directions: comparative CLDs across countries with different policy regimes (for example, Denmark, the Netherlands, and selected EU member states); and integrated models coupling the domestic retrofit system to upstream supply chains and international standardisation processes.\u003c/p\u003e \u003cp\u003eFinally, the study concentrates on energy and carbon outcomes (SDGs 7 and 13) and, to a lesser extent, urban sustainability (SDG 11). Distributional impacts, co-benefits, and potential rebound effects are not explicitly modelled, and this represents a meaningful gap as DT-enabled retrofit programmes begin to scale. Mixed-methods research integrating system dynamics modelling with empirical evaluation of social and economic outcomes is needed to assess whether DT adoption genuinely advances a broader SDG portfolio or risks generating new trade-offs. Three specific directions are proposed. First, quantitative impact assessment: future work could calibrate the CLD against emerging pilot programme data to estimate concrete outcomes \u0026mdash; for example, tonnes of CO₂ avoided annually per percentage-point increase in DT adoption, or the threshold adoption rate at which R3 (Economies of Scale) begins to reduce IoT costs measurably. Second, rebound effect analysis: if smart home systems make occupants more comfortable, increasing their energy use, the net carbon benefit of retrofit may be partially offset; this behavioural dynamic is absent from the current model and warrants explicit investigation. Third, inter-SDG spillover mapping: improvements driven by DT-enabled retrofit through SDGs 7 and 13 may generate positive or negative spillovers to SDG 8 (Decent Work and Economic Growth) through green job creation in the retrofit and digital sectors, and to SDG 9 (Industry, Innovation and Infrastructure) through the development of a domestic DT supply chain. These connections are structurally implied by the CLD but not modelled. Embedding pre-agreed measurement frameworks within pilot programmes \u0026mdash; such as the regional retrofit trials recommended in Section \u003cspan refid=\"Sec20\" class=\"InternalRef\"\u003e5.2.1\u003c/span\u003e \u0026mdash; from their inception would provide the empirical foundation needed to progressively enrich the model as the transition unfolds.\u003c/p\u003e \u003c/div\u003e"},{"header":"6. Conclusions","content":"\u003cp\u003eThis study has transformed a sectoral challenge, accelerating the adoption of digital twin technology for UK building retrofits, into a systemic narrative and strategic roadmap. Using a systems thinking lens, a causal loop model was constructed that illuminates how policy, technology, economic, and social factors interact to shape the trajectory of innovation in the UK's domestic retrofitting sector.\u003c/p\u003e \u003cp\u003eThe analysis revealed five reinforcing loops representing pathways to rapid progress, and two balancing loops explaining past stagnation. Government Incentives and Funding emerged as the highest-leverage node, but the analysis cautions that financial investment alone is insufficient without parallel commitments to interoperability, workforce capacity, and data transparency. The system is currently held in a low-adoption equilibrium by the B1 Policy Exhaustion and B2 Demand-Driven Cost Inflation loops, which neutralise the reinforcing engines before they can reach self-sustaining momentum. In terms of the Multi-Level Perspective, the UK's DT retrofit transition remains in a pre-breakthrough phase: the niche has demonstrated technical viability in isolated cases, but the regime-level conditions \u0026mdash; policy stability, interoperability infrastructure, and workforce capacity \u0026mdash; have not yet shifted sufficiently to enable the niche to scale \u003cem\u003e(\u003c/em\u003eGeels \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2011\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e.\u003c/p\u003e \u003cp\u003eBy embracing the four systemic recommendations \u0026mdash; a stable, outcome-linked funding framework, open interoperability standards, a national measured performance database, and investment in socio-technical capacity \u0026mdash; the UK can transform its built environment from a carbon liability into a digitally enabled asset. Such an outcome directly advances SDGs 7, 9, 11, and 13 while providing a replicable, systems-informed blueprint for other national contexts. As the 2025 SDG Progress Report confirms that only 18% of global targets are on track \u003cem\u003e(\u003c/em\u003eUnited Nations \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2025\u003c/span\u003e\u003cem\u003e)\u003c/em\u003e, approaches that identify and activate systemic leverage points rather than addressing barriers in isolation are urgently needed.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e \u003cstrong\u003eCompeting Interests:\u003c/strong\u003e \u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eFunding:\u003c/h2\u003e \u003cp\u003eNo funding was received for conducting this study. The authors did not receive support from any organisation for the submitted work.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eConceptualization: A.Z.W, I.P, S.VB; Methodology: A.Z.W, I.P; Formal analysis and investigation: A.Z.W; Writing \u0026ndash; original draft preparation: A.Z.W; Writing \u0026ndash; review and editing: I.P, S.VB; Supervision: I.P, S.VB.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe Causal Loop Diagram constructed in this study is publicly accessible via Kumu at: https://kumu.io/alejandrozhou/cld-analysis. No other datasets were generated or analysed during this study\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAckermann, F., Eden, C. (2011). Strategic Management of Stakeholders: Theory and Practice. \u003cem\u003eLong Range Planning\u003c/em\u003e, \u003cem\u003e44\u003c/em\u003e(3), 179\u0026ndash;196. https://doi.org/10.1016/j.lrp.2010.08.001 \u003c/li\u003e\n\u003cli\u003eAmmar, A., Nassereddine, H., AbdulBaky, N., AbouKansour, A., Tannoury, J., Urban, H., Schranz, C. (2022) Digital twins in the construction industry: a perspective of practitioners and building authority. Front Built Environ 8:834671. https://doi.org/10.3389/fbuil.2022.834671\u003c/li\u003e\n\u003cli\u003eAttaran, M., Celik, B.G. (2023). Digital twin: benefits, use cases, challenges, and opportunities. Decis Anal J 6:100165. https://doi.org/10.1016/j.dajour.2023.100165\u003c/li\u003e\n\u003cli\u003eColetta, V. R., Pluchinotta, I., Pisinaras, V., Panagopoulos, A., Giordano, R., Pagano, A., Fratino, U., Montanari, A. (2025). Leverage points and cascading impacts analysis in nexus systems using system dynamics modeling. Earth\u0026apos;s Future 13(7). https://doi.org/10.1029/2025ef006190\u003c/li\u003e\n\u003cli\u003eDESNZ (2026). Warm Homes Plan. GOV.UK. https://www.gov.uk/government/publications/warm-homes-plan\u003c/li\u003e\n\u003cli\u003eDoukari, O., Suliman, A. (2024). Renovation digital twin for building retrofit monitoring: a software product and an organizational ecosystem. In: Proc Eur Conf Comput Constr (EC3 2024). https://ec-3.org/publications/conferences/EC32024/papers/EC32024_225.pdf. Accessed 24 Apr 2025\u003c/li\u003e\n\u003cli\u003eElghaish, F., Matarneh, S., Hosseini, M.R., Tezel, A., Mahamadu, A.M., Taghikhah, F. (2024). Predictive digital twin technologies for achieving net zero carbon emissions: a critical review and future research agenda. Smart Sustain Built Environ. https://doi.org/10.1108/sasbe-03-2024-0096\u003c/li\u003e\n\u003cli\u003eGeels, F.W. (2011). The multi-level perspective on sustainability transitions: responses to seven criticisms. Environ Innov Soc Transit 1:24\u0026ndash;40. https://doi.org/10.1016/j.eist.2011.02.002\u003c/li\u003e\n\u003cli\u003eGOV.UK (2021). Apply for the Green Homes Grant scheme. https://www.gov.uk/guidance/apply-for-the-green-homes-grant-scheme. Accessed 2021\u003c/li\u003e\n\u003cli\u003eHM Government (2021). Heat and Buildings Strategy. https://assets.publishing.service.gov.uk/media/61d450eb8fa8f54c14eb14e4/6.7408_BEIS_Clean_Heat_Heat___Buildings_Strategy_Stage_2_v5_WEB.pdf\u003c/li\u003e\n\u003cli\u003eHovmand, P.S. (2014). Community based system dynamics. Springer, New York. https://doi.org/10.1007/978-1-4614-8763-0\u003c/li\u003e\n\u003cli\u003eJradi, M., Madsen, H., Kaiser, M. (2023). DanRETwin: a digital twin solution for optimal energy retrofit decision-making and decarbonization of the Danish building stock. Appl Sci 13:9778. https://doi.org/10.3390/app13179778\u003c/li\u003e\n\u003cli\u003eKaewunruen, S., Rungskunroch, P., Welsh, J. (2019). A digital-twin evaluation of net zero energy building for existing buildings. Sustainability 11:159. https://doi.org/10.3390/su11010159\u003c/li\u003e\n\u003cli\u003eKim, D.H. (2000). Diagnosing systemic issues and designing high-leverage interventions. Syst Thinker 11(1)\u003c/li\u003e\n\u003cli\u003eKim, H., Andersen, D.F. (2012). Building confidence in causal maps generated from purposive text data: mapping transcripts of the Federal Reserve. Syst Dyn Rev 28:311\u0026ndash;328. https://doi.org/10.1002/sdr.1480\u003c/li\u003e\n\u003cli\u003eKumu (2025). CLD analysis. https://kumu.io/alejandrozhou/cld-analysis#cld-leverage-points. Accessed 24 April 2025\u003c/li\u003e\n\u003cli\u003eMeadows, D. (1999). Leverage points: places to intervene in a system. Sustainability Institute, Hartland\u003c/li\u003e\n\u003cli\u003eOkimi, T.O., Sarhan, S., Awodele, I.A., Olaniran, T.O., Adetola, O.D., Olagunju, O.A. (2024). Towards adopting digital twins for enhancing circular economy in the UK construction industry: benefits and enablers. https://doi.org/10.46254/na09.20240022\u003c/li\u003e\n\u003cli\u003ePluchinotta, I., Salvia, G., Zimmermann, N. (2021). The importance of eliciting stakeholders\u0026apos; system boundary perceptions for problem structuring and decision-making. Eur J Oper Res 302:280\u0026ndash;293. https://doi.org/10.1016/j.ejor.2021.12.029\u003c/li\u003e\n\u003cli\u003ePrioreschi, E., Zimmermann, N., Davies, M., Pluchinotta, I. (2024). Interrelationships and trade-offs between urban natural space use and biodiversity. Sustainability 16:4051. https://doi.org/10.3390/su16104051\u003c/li\u003e\n\u003cli\u003eReed, M.S., Graves, A., Dandy, N., Posthumus, H., Hubacek, K., Morris, J., Prell, C., Quinn, C.H., Stringer, L.C. (2009). Who\u0026apos;s in and why? A typology of stakeholder analysis methods for natural resource management. J Environ Manage 90:1933\u0026ndash;1949. https://doi.org/10.1016/j.jenvman.2009.01.001\u003c/li\u003e\n\u003cli\u003eSaback, V., Popescu, C., T\u0026auml;ljsten, B., Blanksv\u0026auml;rd, T. (2023). Analysis of digital twins in the construction industry: current trends and applications. Lect Notes Civ Eng, pp 1080\u0026ndash;1088. https://doi.org/10.1007/978-3-031-32511-3_110\u003c/li\u003e\n\u003cli\u003eSenge, P.M., Sterman, J.D. (1992) Systems thinking and organizational learning: acting locally and thinking globally in the organization of the future. Eur J Oper Res 59:137\u0026ndash;150. https://doi.org/10.1016/0377-2217(92)90011-W\u003c/li\u003e\n\u003cli\u003eSobowale, M., Elghaish, F., Brooks, T. (2023). A systematic review of digital twin as a predictive maintenance approach for existing buildings in the UK. Proc e-report, pp 1206\u0026ndash;1218. https://doi.org/10.36253/979-12-215-0289-3.119\u003c/li\u003e\n\u003cli\u003eSghiri, A., Gallab, M., Merzouk, S., Assoul, S. (2025). Leveraging digital twins for enhancing building energy efficiency: a literature review of applications, technologies, and challenges. Buildings 15:498. https://doi.org/10.3390/buildings15030498\u003c/li\u003e\n\u003cli\u003eSterman, J.D. (2000). Business dynamics: systems thinking and modeling for a complex world. McGraw-Hill, Boston\u003c/li\u003e\n\u003cli\u003eUK Government (2017). Domestic private rented property: minimum energy efficiency standard \u0026mdash; landlord guidance. https://www.gov.uk/guidance/domestic-private-rented-property-minimum-energy-efficiency-standard-landlord-guidance\u003c/li\u003e\n\u003cli\u003eUKGBC (2023). Climate change mitigation. https://ukgbc.org/our-work/climate-change-mitigation/\u003c/li\u003e\n\u003cli\u003eUnited Nations (2023). The Sustainable Development Goals. https://sdgs.un.org/goals\u003c/li\u003e\n\u003cli\u003eUnited Nations (2025). The Sustainable Development Goals Report 2025. https://unstats.un.org/sdgs/report/2025/The-Sustainable-Development-Goals-Report-2025.pdf\u003c/li\u003e\n\u003cli\u003eVeldhuis, G.A., Smits-Clijsen, E.M., van Waas, R.P.M., Hof, T., Maccatrozzo, V., Rouwette, E.A.J.A., Kerstholt, J.H. (2025). The influence of causal loop diagrams on systems thinking and information utilization in complex problem-solving. Comput Hum Behav Rep 17:100613. https://doi.org/10.1016/j.chbr.2025.100613\u003c/li\u003e\n\u003cli\u003eVensim (2025). Vensim Personal Learning Edition. https://vensim.com/vensim-personal-learning-edition/. Accessed 24 April 2025\u003c/li\u003e\n\u003cli\u003eZdunek, R. (2024). How digital twin technology is transforming real estate sales. Chameleon Interactive. https://chameleon-interactive.com/2024/10/14/how-digital-twin-technology-is-transforming-real-estate-sales/\u003c/li\u003e\n\u003cli\u003eZhu, H., Ye, G., Wu, Z., Han, Y. (2024). Building on digital twin: overcoming barriers and unlocking success in the construction industry. J Constr Eng Manage 150(10). https://doi.org/10.1061/JCEMD4.COENG-14234\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"systemic-practice-and-action-research","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"spaa","sideBox":"Learn more about [Systemic Practice and Action Research](http://link.springer.com/journal/11213)","snPcode":"11213","submissionUrl":"https://submission.nature.com/new-submission/11213/3","title":"Systemic Practice and Action Research","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Digital Twin, Retrofit, Systems Thinking, Causal Loop Diagram, Socio-Technical Transitions, Sustainable Development Goals","lastPublishedDoi":"10.21203/rs.3.rs-9227696/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9227696/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe decarbonisation of the UK's 29\u0026nbsp;million homes represents a problem characterised by systemic inertia and fragmented policy. Digital Twin technology offers a niche innovation with the potential to accelerate data-driven retrofitting; however, its adoption is suppressed by complex socio-technical barriers. This article presents a systems thinking analysis to identify and analyse the causal mechanisms shaping DT adoption in the UK housing sector. Adopting a systems thinking perspective, a Causal Loop Diagram is constructed based on a systematic review of contemporary literature (2018\u0026ndash;2024), from which 15 core variables are extracted through open coding. Eigenvector centrality and cross-impact matrix analysis are then applied to identify possible deep leverage points for intervention. The analysis reveals five reinforcing loops\u0026mdash;including data-quality and simulation-investment cycles\u0026mdash;that act as engines for change, currently held in check by institutional capability traps. It is proposed that UK policymakers shift from a technocentric focus to a systemic approach that prioritises interoperability, technical capacity, and stable long-term incentives to unlock the transition toward the Sustainable Development Goals (SDGs 7, 9, 11, and 13).\u003c/p\u003e","manuscriptTitle":"Governing the Digital Transition: A Systems Thinking Approach to Analyse Digital Twin Adoption in UK Housing Retrofit","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-20 08:12:44","doi":"10.21203/rs.3.rs-9227696/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-04-28T07:33:53+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"307338164328182988716575533334723128761","date":"2026-04-28T06:13:41+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"272795190800437402675508241408729470479","date":"2026-04-14T08:59:01+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-10T10:02:48+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-31T08:17:49+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-27T02:41:27+00:00","index":"","fulltext":""},{"type":"submitted","content":"Systemic Practice and Action Research","date":"2026-03-26T00:30:56+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"systemic-practice-and-action-research","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"spaa","sideBox":"Learn more about [Systemic Practice and Action Research](http://link.springer.com/journal/11213)","snPcode":"11213","submissionUrl":"https://submission.nature.com/new-submission/11213/3","title":"Systemic Practice and Action Research","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"7833c484-7517-4eb0-a39c-a939b0c70321","owner":[],"postedDate":"April 20th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-20T08:12:45+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-20 08:12:44","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9227696","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9227696","identity":"rs-9227696","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}