The incorporation of Industry 4.0 (I4.0) technologies into last-mile delivery (LMD) enhances sustainability and competitiveness in logistics. However, existing studies have not established performance metrics for LMD approaches integrating I4.0 and sustainability. This study aims to develop a sustainability/I4.0-based performance measurement methodology for the LMD methods (S-4.0-LMD-Ms).
This research used a collaborative management research approach with two leading logistics services companies to frame the evaluation factors and LMD-Ms. The scoping study validates and extends the findings from a thorough literature review analysis. Mathematical modelling was employed to analyze and report accurate data, facilitating informed decision-making recommendations to managers regarding the LMD-Ms’ performance.
This study led to the development of a holistic framework for S-4.0-LMD-Ms to guide practitioners and extend theoretical insights. A comprehensive analysis of modern and traditional LMD-Ms was conducted, and a detailed list is presented.
The study extends sustainability performance measurement by integrating I4.0 capabilities into LMD evaluation and demonstrates a configuration-based perspective on performance. Empirical insights from industry experts validate the framework and highlight the coexistence of technological and incremental improvement pathways.
1. Introduction
LMD, the final stage of the supply chain where goods reach end consumers, is widely recognized as one of the most complex and sustainability-intensive logistics activities (Alidaee et al., 2023). The rapid growth of e-commerce has intensified its strategic importance, as LMD significantly influences customer satisfaction, operational efficiency, and overall supply chain performance (Sawik, 2024). At the same time, LMD accounts for a substantial share of logistics costs, estimated between 13 and 75%, while contributing disproportionately to urban congestion and greenhouse gas emissions (Maxner et al., 2022). In response, I4.0 technologies such as automation, real-time analytics, and cyber-physical systems are increasingly deployed to enhance logistics efficiency and sustainability by enabling more responsive, data-driven, and resource-efficient supply chain operations (Mohammed et al., 2025; Zamora Iribarren et al., 2024). Consequently, integrating digital technologies into last-mile logistics has become a key approach to addressing the operational and environmental challenges of urban delivery systems, although their effectiveness remains contingent on infrastructural, regulatory, and operational conditions (Patyal et al., 2022; Huang et al., 2023).
Despite increasing research on sustainable logistics and digital transformation, LMD remains a complex decision context characterised by fragmented operations, diverse delivery methods, and strong dependence on urban infrastructure and regulatory environments. Delivery method choices significantly influence economic efficiency, environmental performance, and social outcomes, particularly in omnichannel and e-commerce supply chains (Hübner et al., 2016). However, existing studies typically examine technological innovation or sustainability aspects separately, rather than analyzing their combined role in shaping LMD performance. Consequently, the literature lacks holistic performance management approaches that systematically evaluate LMD-Ms by jointly considering sustainability and I4.0 technological capabilities, and by capturing the interdependencies between performance objectives and implementation constraints. More critically, existing approaches implicitly assume that sustainability and technological performance can be assessed as independent or additive dimensions, overlooking that, in LMD contexts, performance emerges from their interaction and is conditioned by operational and infrastructural constraints.
From a theoretical perspective, this challenge relates to sustainability performance measurement (SPM), which emphasizes evaluating organisational performance across environmental, economic, and social dimensions (Schaltegger et al., 2014; Morioka and de Carvalho, 2016). Yet these frameworks have rarely been applied to digitally enabled logistics systems and generally do not incorporate I4.0 technological capabilities within their evaluation structures. Moreover, SPM frameworks are predominantly designed for relatively stable organisational settings, and therefore offer limited explanatory power in highly dynamic, technology-enabled environments such as LMD, where performance is contingent, configuration-dependent, and shaped by the alignment between technological capabilities and operational conditions. As a result, conceptual and methodological guidance for systematically assessing LMD methods that integrate sustainability performance with digital technological capabilities remains limited. Addressing this limitation requires developing an integrated evaluation framework capable of assessing LMD-Ms across sustainability and I4.0 dimensions, while accounting for the conditional and context-dependent nature of technological performance.
In response to this gap, the present study addresses the following research questions (RQ):
What sustainability and I4.0-related factors should be considered when evaluating the performance of LMD-Ms?
How can these sustainability and technological factors be integrated into a holistic performance evaluation framework for LMD-Ms?
How do traditional and technology-enabled LMD-Ms perform when evaluated through an integrated sustainability–I4.0 framework?
To address these questions, this study proposes an S-4.0-LMD-Ms. This paper contributes to the literature on sustainable logistics and digital supply chains in three ways. First, it extends SPM research by integrating I4.0 technological capabilities into the evaluation of LMD systems. Thus, the study moves beyond treating sustainability dimensions as static evaluation criteria and reconceptualizes performance as an outcome of the interaction between sustainability priorities and technological capabilities under context-specific constraints. Second, it develops a holistic evaluation framework that systematically assesses LMD-Ms across environmental, economic, social, and technological dimensions, explicitly capturing the interdependencies and trade-offs among these dimensions rather than evaluating them in isolation, and highlighting how performance emerges from the interaction among sustainability priorities, operational efficiency, and technological readiness. Third, it operationalizes this framework through a multi-criteria decision-making approach (MCDMA) applied to both traditional and emerging delivery methods, enabling comparative evaluation of alternative delivery configurations under varying contextual conditions.
From a practical perspective, the study provides decision-support insights to assist logistics managers and policymakers in selecting and designing LMD strategies that balance operational efficiency, technological feasibility, and sustainability performance in increasingly complex urban logistics environments. Specifically, the study enables decision-makers to prioritize delivery configurations not only based on their performance potential but also based on their feasibility under infrastructural and regulatory constraints.
The remainder of this paper is structured as follows. Section 2 reviews the relevant literature on LMD, sustainability, and I4.0. Section 3 presents the research methodology. Section 4 reports the empirical results. Section 5 provides the discussion and research impact. Section 6 concludes with key contributions, limitations, and directions for future research.
2. Literature review
Sustainability pressures and rapid technological advances are reshaping logistics and supply chain management. Sustainable supply chains seek to balance economic viability with environmental stewardship and social responsibility, requiring organizations to reduce waste, optimize transportation, and improve resource utilization (Mohammed, 2026; Zubairu et al., 2025). In parallel, I4.0 technologies, including the Internet of Things (IoT), artificial intelligence (AI), advanced analytics, and cyber-physical systems, enable real-time monitoring, enhanced supply chain visibility, and data-driven decision making (Soori et al., 2023). These technologies also support sustainability improvements through route optimization, fuel efficiency, and improved supply chain transparency (Ghode et al., 2020). Furthermore, IoT monitoring and AI-driven optimization enable more efficient logistics network design and emissions management, while emerging delivery technologies such as autonomous vehicles and drones (UAVs) offer new opportunities to improve delivery efficiency and reduce environmental impacts in urban logistics environments (Tsolakis et al., 2022; Moshref-Javadi et al., 2020; Apruzzese et al., 2023).
Despite these developments, LMD remains one of the most operationally complex and environmentally impactful stages of logistics systems (Alves de Araújo et al., 2022; Chiappetta Jabbour et al., 2020; Huang et al., 2023). As the final interface between logistics providers and consumers, LMD significantly influences delivery costs, service reliability, urban congestion, and emissions (Risberg and Jafari, 2022). Consequently, improving last-mile logistics performance has become a key priority for both researchers and practitioners.
Recent research has increasingly examined the intersection of sustainability and digital transformation in logistics and supply chain management. From a theoretical standpoint, technology adoption in logistics systems is often explained through contingency-based perspectives, which suggest that the performance outcomes of technological innovations depend on contextual factors such as infrastructure readiness, organisational capabilities, and regulatory conditions (Seyedghorban et al., 2020; Huang et al., 2023; Najla et al., 2025). In LMD contexts, these contingencies influence the feasibility and effectiveness of I4.0 technologies, including autonomous vehicles, UAVs, and digital platforms. Consequently, several studies emphasize that logistics technology adoption should be evaluated as a socio-technical system, where operational performance, sustainability objectives, and institutional conditions interact to shape delivery system outcomes.
Within logistics engineering and operations management research, multi-criteria decision-making (MCDM) approaches, particularly fuzzy-AHP and hybrid fuzzy models, have been widely applied to evaluate complex logistics problems characterised by uncertainty and multiple performance criteria. Prior studies have employed fuzzy-AHP in logistics contexts such as supplier selection, sustainable transportation evaluation, and logistics service prioritization (de Araújo et al., 2022; Naseem et al., 2021). Research within the IEOM community has also explored decision-support models for sustainable supply chains and logistics engineering systems, including applications of MCDM techniques to evaluate logistics performance, technology adoption, and supply-chain optimization. These studies highlight the usefulness of structured decision frameworks for analyzing logistics alternatives across economic, environmental, and technological dimensions.
At the same time, the literature on sustainability modelling in logistics has emphasized the need for integrated evaluation frameworks capable of balancing environmental performance, operational efficiency, and social outcomes (Chiappetta Jabbour et al., 2020; Morioka and de Carvalho, 2016). Within the IEOM research stream, recent work has explored sustainable supply chain optimization, green logistics strategies, and technology-enabled logistics transformation, reflecting the growing importance of integrating sustainability considerations into engineering and operations management decision-making (Gebisa and Ram, 2021; Nguyen and Dao, 2019). However, most existing studies evaluate sustainability performance and technological innovation separately, rather than examining their combined role in shaping LMD system performance.
The literature proposes a range of LMD-Ms, broadly classified into traditional and technology-enabled approaches (Mohammad et al., 2023; Karlı and Tanyaş, 2024). Traditional solutions such as home delivery, parcel lockers, and cargo bikes rely on established logistics infrastructures but may face limitations in scalability and environmental performance. In contrast, technology-enabled approaches, including crowdsourced delivery, click-and-collect systems, electric and autonomous vehicles, UAVs, and delivery robots, leverage I4.0 technologies to enhance operational flexibility and sustainability outcomes (Morioka and de Carvalho, 2016; Naseem et al., 2021; Rodriguez-Plesa et al., 2022). Detailed descriptions, advantages, and limitations of each LMD-M are provided in the supplementary file, while the main manuscript focuses on their comparative evaluation and managerial implications.
Given the diversity of these solutions, recent studies increasingly evaluate alternative delivery methods under different operational and sustainability conditions. Quantitative decision-support approaches are frequently used to assess logistics alternatives across economic, environmental, technological, and operational criteria. However, the evolving nature of urban logistics systems requires evaluation frameworks that can adapt to technological developments, regulatory changes, and context-specific operational constraints (Alverhed et al., 2024). As highlighted by Izadkhah et al. (2022), effective evaluation approaches must therefore capture the dynamic characteristics of logistics systems while reflecting the environments in which delivery operations occur.
Although the literature acknowledges the close relationship between technological innovation and sustainability performance within supply chain management (Chiappetta Jabbour et al., 2020; Seyedghorban et al., 2020), important limitations remain. Many studies analyse technological innovation and sustainability considerations separately rather than examining their combined role in logistics decision-making. Moreover, while quantitative evaluation approaches are widely used to assess logistics alternatives, they often focus on limited performance dimensions and rarely integrate technological and sustainability factors within a unified evaluation framework. As a result, decision-support tools capable of jointly assessing both dimensions remain underdeveloped for LMD.
Addressing this limitation is critical, as LMD decisions increasingly require balancing technological feasibility, economic efficiency, environmental performance, and social considerations. In response, this study proposes an evaluation framework that integrates I4.0 technological capabilities with sustainability criteria in assessing LMD-Ms. The study contributes to research on digitalised and sustainable logistics decision-making and provides a systematic basis for comparing alternative LMD solutions in complex logistics environments by combining these dimensions within a structured evaluation approach.
3. Research methodology
This research is motivated by the growing need for the LMD sector to integrate I4.0 capabilities to enhance sustainability performance (Accenture, 2021). The study adopts a collaborative management research approach, combining a structured literature review with a scoping study, to develop a performance management methodology for LMD-Ms grounded in sustainability and I4.0 paradigms. This approach is particularly suitable as it enables the co-creation of knowledge with practitioners, ensuring that the evaluation framework reflects real-world operational constraints, emerging technological practices, and sustainability priorities rather than purely theoretical assumptions.
Empirical insights were gathered through collaboration with two large logistics service providers operating in the Middle East, representing both the government and the private sectors. This dual perspective ensured practical relevance and enhanced the robustness of the evaluation framework. The companies were purposively selected based on their active engagement in sustainability initiatives and technological transformation within LMD operations. For confidentiality reasons, company and country names are not disclosed. Such purposive selection strengthens methodological fit by capturing informed expert judgment from organizations already experiencing the challenges and trade-offs associated with I4.0-enabled sustainable LMD.
The research methodology comprises four main stages (S1–S4), as illustrated in Figure 1:
The flowchart outlines the process of developing a study methodology for evaluating last mile delivery methods. It starts with the identification of key performance evaluation factors. The next step involves identifying and describing the most common traditional and modern last mile delivery methods. Following this, a multi-criteria decision-making analysis-based method, specifically Fuzzy AHP and TOPSIS, is proposed. The final step involves proposing recommendations and policies based on the analysis.The overview of the development of the study methodology
The flowchart outlines the process of developing a study methodology for evaluating last mile delivery methods. It starts with the identification of key performance evaluation factors. The next step involves identifying and describing the most common traditional and modern last mile delivery methods. Following this, a multi-criteria decision-making analysis-based method, specifically Fuzzy AHP and TOPSIS, is proposed. The final step involves proposing recommendations and policies based on the analysis.The overview of the development of the study methodology
S1. Identification of key sustainability and I4.0-related performance evaluation factors for LMD-Ms through literature analysis and a scoping study.
S2. Identification and classification of common traditional and modern LMD-Ms relevant to the regional context, informed by both literature and industry input.
S3. Application of an MCDMA to quantify the relative importance of evaluation factors and assess the performance of LMD-Ms.
S4. Development of managerial and policy-oriented recommendations to support the sustainable, technology-enabled transformation of LMD systems.
The staged design of this methodology is well-suited to the complexity of LMD systems, as it allows qualitative expert knowledge to be systematically translated into quantitative evaluation outcomes, thereby improving analytical transparency, comparability across delivery methods, and decision-making relevance.
3.1 Literature and scoping study
In this study, a structured literature analysis was first conducted to identify commonly adopted LMD-Ms and relevant performance evaluation factors aligned with established I4.0 capabilities (hereafter referred to as technological factors) and sustainability dimensions (Bošković et al., 2017). The identified criteria were subsequently consolidated and refined to eliminate duplication and ensure conceptual clarity.
The refined set of factors was then validated through a scoping study involving eight industry professionals working in LMD operations. Participants were purposively nominated by senior contacts within the collaborating organizations following a briefing on the research objectives. The respondents possessed an average of 7 years of professional experience across key LMD functions, including operations management, planning and scheduling, quality management, customer services, and external business development, with representation from both governmental and private-sector organizations.
Data were collected through a structured three-hour interview session conducted at the university campus by two members of the research team. The session combined an initial briefing to ensure shared understanding of the research scope, followed by semi-structured questions addressing LMD methods, sustainability practices, I4.0 implementation challenges, and performance assessment metrics (Haukås and Tishakov, 2024; Wingate and Bourdage, 2024).
Responses confirmed that no additional LMD-Ms emerged beyond those identified in the literature; consequently, detailed method descriptions are not repeated in the main manuscript. Insights related to performance evaluation factors informed Stage S1, while quantitative inputs for the MCDM analysis supported Stage S3 of the methodology. The resulting evaluation factors are presented in Section 4.1.
3.2 MCDMA-based method
Upon identifying S-4.0-LMD-Ms in stages 1 and 2, the research employs two widely accepted MCDM methods, AHP and TOPSIS, to conduct Stage 3 and support Stage 4. Fuzzy-based MCDM methods are applied to account for the uncertainty in professional opinions, particularly regarding modern LMD approaches. Using fuzzy AHP and TOPSIS allows handling subjective and ambiguous assessments using triangular fuzzy numbers (Zhang and Chu, 2009). These methods offer a user-friendly and reliable decision-making process, even when performance evaluations are uncertain (Naseem et al., 2021). Although the study doesn't aim to contribute specifically to MCDM-related research, these techniques facilitate efficient evaluations.
3.2.1 Weighing performance evaluation factors: fuzzy AHP
This method is employed to weigh the sustainability and technological factors. The chosen approach is supported by its applicability to the study setting and its capacity to record stakeholders' preferences and priorities (Kahraman et al., 2003; de Araújo et al., 2022; Mondal et al., 2023). The data required to build the decision matrix based on the evaluation factors were gathered during the three-hour meeting with the eight professionals. The research team asked participants to do this task individually to avoid bias in the pairwise evaluation.
3.2.2 LMD-ms performance measurement: fuzzy TOPSIS
Fuzzy TOPSIS evaluates the LMD-Ms (see section 2.2) using the weighted factors revealed via fuzzy AHP. The sustainability and I4.0 performance of each LMD-M is measured comprehensively against the defined factors. To gather pertinent data on the LMD-Ms, the meeting with the eight professionals included a session in which participants were asked to evaluate each LMD-M vis-à-vis its sustainability performance and I4.0 capabilities. Participants conducted this session individually to avoid any influence from each other (Kruger et al., 2019; Roth et al., 2023). It is worth mentioning that participants were shown a sample of how to fill out initial decision matrices for MCDM methods to ensure proper data collection. Two research team members also observed this stage of the meeting to address any participant inquiries.
3.3 Applicability and validity
The MCDMA outcomes in S3 were shared with the eight professionals via email to check and validate the results and present more rigorous results with potential managerial impact and practicality. In this regard, researchers sought to avoid presenting conclusions that might be far from respondents' opinions. To further ensure robustness, consistency checks were applied within the fuzzy AHP process, and the alignment between AHP-derived weights and TOPSIS rankings was examined. While full sensitivity analysis is beyond the scope of this exploratory, expert-driven study, the convergence of expert evaluations and validation feedback provides confidence in the stability and reliability of the results. The research methodology generates suggestions based on the results of S1-S3. The best options for embedding I4.0 to improve sustainability performance are then highlighted in the recommendations. These recommendations align with the study's goals and are supported by the results.
4. Results and application
4.1 Performance evaluation factors
Technological advancements are transforming logistics systems in the I4.0 era, while sustainability performance is increasingly recognized as a core evaluation dimension. This research stage identifies the key sustainability and technological (I4.0-related) factors for evaluating LMD-Ms, as derived from the scoping study. These factors form the foundation for the subsequent evaluation stages (S2 and S3) of the research methodology. Figure 2 presents the full set of identified and agreed-upon factors across economic, environmental, social (societal), and technological dimensions.
The diagram illustrates four main categories of factors: Economic, Environmental, Social/Societal, and Technological. Each category lists specific factors. Economic factors include distance traveled, vehicle fill rate, warehouse fill rate, stopping time, delivery time, and packaging reduction. Environmental factors include greenhouse gas emissions, recycling materials, and weather/climate. Social/Societal factors include customer satisfaction, safety and risk, number of jobs created, road congestion, and noise. Technological factors include accessibility and availability, privacy concerns, shifting expectations, infrastructure limitations, real-time data, regulatory landscape, organizational culture, workforce readiness, tracking system, cybersecurity threats, integration challenges, and maturity and availability of skills.Key sustainability and technological (I4.0) performance factors for LMD-Ms’ performance evaluation
The diagram illustrates four main categories of factors: Economic, Environmental, Social/Societal, and Technological. Each category lists specific factors. Economic factors include distance traveled, vehicle fill rate, warehouse fill rate, stopping time, delivery time, and packaging reduction. Environmental factors include greenhouse gas emissions, recycling materials, and weather/climate. Social/Societal factors include customer satisfaction, safety and risk, number of jobs created, road congestion, and noise. Technological factors include accessibility and availability, privacy concerns, shifting expectations, infrastructure limitations, real-time data, regulatory landscape, organizational culture, workforce readiness, tracking system, cybersecurity threats, integration challenges, and maturity and availability of skills.Key sustainability and technological (I4.0) performance factors for LMD-Ms’ performance evaluation
The sustainability factors encompass economic (distance travelled, vehicle fill rate, warehouse fill rate, stopping time, delivery time, packaging reduction), environmental (greenhouse gas emissions, recycling materials, weather/climate), and social (societal) dimensions (customer satisfaction, safety and risk, number of jobs created, road congestion, noise). The technological factors include accessibility and availability, privacy concerns, shifting expectations, infrastructure limitations, real-time data, regulatory landscape, organisational culture, workforce readiness, tracking system, cybersecurity threats, integration challenges, and maturity and availability of skills.
4.2 LMD-M′ performance
This section presents results from two stages of the research methodology: (1) quantifying the importance of sustainability and technological evaluation factors using fuzzy AHP, and (2) evaluating the performance of LMD-Ms using fuzzy TOPSIS.
In Stage S2, fuzzy AHP was applied using pairwise comparisons from eight industry professionals. The aggregated fuzzy decision matrices across economic, environmental, social, and technological dimensions are presented in Table 1. The resulting quantitative weights and rankings of the evaluation factors are reported in Table 2. Intermediate calculations and individual matrices are omitted for brevity.
The Fuzzy decision matrices for economic, environmental, social, and technological factors
| Factors | A | B | C | D | E | F | G | H | I | J | K | L |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Economic Factors | ||||||||||||
| A | 1.0 | 0.15 | 0.33 | 0.13 | 1.0 | 0.17 | ||||||
| B | 6.6667 | 1.0 | 0.4 | 0.2 | 6.0 | 0.15 | ||||||
| C | 3.0303 | 2.5 | 1.0 | 0.333 | 4.5 | 0.25 | ||||||
| D | 7.6923 | 5.0 | 3.003 | 1.0 | 6.0 | 6.0 | ||||||
| E | 1.0 | 0.1667 | 0.2222 | 0.1667 | 1.0 | 0.15 | ||||||
| F | 5.8824 | 6.6667 | 4.0 | 0.1667 | 6.6667 | 1.0 | ||||||
| Environmental Factors | ||||||||||||
| A | 1.0 | 8.0 | 3.0 | |||||||||
| B | 0.125 | 1.0 | 0.33 | |||||||||
| C | 0.3333 | 3.0303 | 1.0 | |||||||||
| Social Factors | ||||||||||||
| A | 1.0 | 2.0 | 3.0 | 6.0 | 5.0 | |||||||
| B | 0.5 | 1.0 | 4.0 | 4.0 | 7.0 | |||||||
| C | 0.3333 | 0.25 | 1.0 | 7.0 | 8.0 | |||||||
| D | 0.1667 | 0.25 | 0.1429 | 1.0 | 3.0 | |||||||
| E | 0.2 | 0.1429 | 0.125 | 0.3333 | 1.0 | |||||||
| Technological Factors | ||||||||||||
| A | 1.0 | 2.0 | 4.0 | 1.0 | 7.0 | 5.0 | 8.0 | 2.0 | 1.0 | 4.0 | 1.0 | 3.0 |
| B | 0.5 | 1.0 | 0.5 | 0.2 | 4.0 | 0.4 | 3.0 | 5.0 | 0.2 | 1.0 | 6.0 | 1.0 |
| C | 0.3 | 2.0 | 1.0 | 0.2 | 0.4 | 0.2 | 1.0 | 0.2 | 0.2 | 0.2 | 1.0 | 0.2 |
| D | 1.0 | 6.7 | 5.0 | 1.0 | 6.0 | 2.0 | 6.0 | 3.0 | 5.0 | 4.0 | 4.0 | 2.0 |
| E | 0.1 | 0.3 | 2.4 | 0.2 | 1.0 | 0.3 | 1.0 | 0.1 | 0.7 | 0.2 | 0.2 | 0.1 |
| F | 0.2 | 2.4 | 5.8 | 0.5 | 3.0 | 1.0 | 6.0 | 5.0 | 5.0 | 5.0 | 3.0 | 3.0 |
| G | 0.1 | 0.3 | 1.0 | 0.2 | 1.0 | 0.2 | 1.0 | 0.2 | 1.0 | 0.1 | 0.2 | 0.2 |
| H | 0.5 | 0.2 | 5.0 | 0.3 | 7.0 | 0.2 | 5.6 | 1.0 | 6.0 | 7.0 | 4.0 | 1.0 |
| I | 1.0 | 5.0 | 5.0 | 0.2 | 1.4 | 0.2 | 1.0 | 0.2 | 1.0 | 0.2 | 0.1 | 0.2 |
| J | 0.3 | 1.0 | 6.7 | 0.3 | 5.0 | 0.2 | 7.0 | 0.1 | 5.0 | 1.0 | 0.2 | 1.0 |
| K | 1.0 | 0.2 | 1.0 | 0.3 | 6.7 | 0.3 | 6.7 | 0.3 | 8.3 | 6.7 | 1.0 | 1.0 |
| L | 0.3 | 1.0 | 5.0 | 0.5 | 7.5 | 0.3 | 6.7 | 1.0 | 5.0 | 1.0 | 1.0 | 1.0 |
The quantitative weight of sustainability and technological factors
| Pillar | Factors | Weights | Rank |
|---|---|---|---|
| Sustainability | |||
| Economic | A. Distance Travelled | 0.0329 | 5 |
| B. Vehicle Fill Rate | 0.1704 | 3 | |
| C. Warehouse Fill Rate | 0.1373 | 4 | |
| D. Stopping Time | 0.3392 | 1 | |
| E. Delivery Time | 0.0320 | 6 | |
| F. Packaging Reduction | 0.2882 | 2 | |
| Environmental | |||
| A. Greenhouse Gas Emissions | 0.6735 | 1 | |
| B. Recycling Materials | 0.0817 | 3 | |
| C. Weather/Climate | 0.2449 | 2 | |
| Social | |||
| A. Customer Satisfaction | 0.3012 | 1 | |
| B. Safety and Risk | 0.2923 | 3 | |
| C. Number of Jobs Created | 0.2938 | 2 | |
| D. Road Congestion | 0.0808 | 4 | |
| E. Noise | 0.0319 | 5 | |
| Technological | |||
| A. Accessibility and Availability | 0.1253 | 3 | |
| B. Privacy Concerns | 0.0732 | 8 | |
| C. Shifting Expectations | 0.0222 | 10 | |
| D. Infrastructure Limitations | 0.1468 | 1 | |
| E. Real-Time Data | 0.0212 | 11 | |
| F. Regulatory Landscape | 0.1282 | 2 | |
| G. Organizational Culture | 0.0177 | 12 | |
| H. Workforce Readiness | 0.1214 | 4 | |
| I. Tracking System | 0.0498 | 9 | |
| J. Cybersecurity Threats | 0.0893 | 7 | |
| K. Integration Challenges | 0.1076 | 5 | |
| L. Maturity and Availability of Skills | 0.0973 | 6 | |
Beyond the pairwise comparison structure (Table 1), Table 2 provides the key decision insights by quantifying factor importance across pillars. Environmentally, greenhouse gas emissions (0.6735) dominate all criteria, indicating that emissions reduction is not only a priority but the central organizing objective in evaluating LMD performance. This suggests that environmental considerations are weighted in a way that structurally prioritizes carbon-related outcomes over other sustainability dimensions.
Economically, stopping time (0.3392) and packaging reduction (0.2882) are the most influential factors, suggesting that economic performance is primarily reflected through operational efficiency mechanisms. This indicates that time-related inefficiencies and resource utilization directly shape perceived economic value in last-mile systems.
Socially, customer satisfaction (0.3012), job creation (0.2938), and safety (0.2923) carry comparable weights, indicating a distributed importance structure rather than dominance by a single factor. This suggests that social performance in LMD is multi-dimensional, reflecting the simultaneous importance of service quality, workforce conditions, and employment impact.
Technologically, infrastructure limitations (0.1468), regulatory landscape (0.1282), and accessibility (0.1253) are the most critical enablers, indicating that I4.0 performance is primarily constrained by foundational system readiness rather than advanced digital capabilities. This implies that technological effectiveness is contingent on enabling conditions, with infrastructure and regulation shaping the feasibility of innovation adoption.
These results indicate that LMD performance is structured around a hierarchy of dependencies, where environmental priorities dominate, but are operationalised through efficiency mechanisms and constrained by technological and institutional readiness. This highlights that sustainability performance in LMD is not achieved in isolation but emerges from the alignment of environmental objectives, operational capabilities, and enabling infrastructure.
The second phase evaluates the performance of LMD-Ms using the weighted criteria derived in Stage S2. Ten alternatives (A1–A10) were assessed, including delivery boxes, attended and unattended home delivery, cargo bikes, crowdsourcing, click-and-collect, innovative/electric vehicles, UAVs, autonomous delivery robots (ADRs), and delivery by droids.
The aggregated fuzzy evaluation matrix is presented in Table 3, based on expert assessments across all criteria. The fuzzy TOPSIS method was then applied to compute the closeness coefficient (Ci) and rank each alternative.
Aggregated initial Evaluation matrix for Fuzzy TOPSIS
| C1 | C2 | C3 | C4 | C5 | C6 | C7 | C8 | C9 | C10 | C11 | C12 | C13 | C14 | C15 | C16 | C17 | C18 | C19 | C20 | C21 | C22 | C23 | C24 | C25 | C26 | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| A1 | 5 | 8 | 8 | 6 | 7 | 5 | 6 | 6 | 5 | 6 | 8 | 7 | 5 | 3 | 6 | 6 | 7 | 6 | 4 | 7 | 6 | 8 | 6 | 6 | 8 | 7 |
| A2 | 4 | 5 | 5 | 2 | 3 | 8 | 3 | 3 | 3 | 8 | 4 | 5 | 7 | 1 | 4 | 7 | 7 | 8 | 5 | 7 | 8 | 8 | 6 | 8 | 8 | 9 |
| A3 | 1 | 8 | 7 | 5 | 5 | 3 | 5 | 6 | 5 | 6 | 8 | 7 | 5 | 1 | 6 | 6 | 7 | 8 | 2 | 5 | 6 | 8 | 5 | 8 | 8 | 8 |
| A4 | 4 | 5 | 5 | 5 | 5 | 6 | 4 | 5 | 7 | 5 | 8 | 7 | 7 | 3 | 7 | 6 | 7 | 8 | 6 | 7 | 8 | 8 | 6 | 8 | 6 | 8 |
| A5 | 5 | 6 | 6 | 6 | 6 | 6 | 5 | 5 | 6 | 6 | 8 | 7 | 9 | 4 | 7 | 5 | 6 | 7 | 6 | 5 | 7 | 6 | 6 | 5 | 8 | 6 |
| A6 | 1 | 8 | 8 | 7 | 5 | 8 | 2 | 5 | 2 | 3 | 8 | 4 | 2 | 1 | 5 | 7 | 7 | 8 | 6 | 8 | 8 | 8 | 8 | 4 | 6 | 9 |
| A7 | 5 | 8 | 8 | 7 | 7 | 8 | 8 | 8 | 7 | 6 | 8 | 7 | 5 | 7 | 7 | 8 | 6 | 6 | 8 | 7 | 6 | 6 | 8 | 4 | 4 | 6 |
| A8 | 6 | 6 | 6 | 8 | 8 | 7 | 8 | 9 | 9 | 8 | 9 | 8 | 5 | 7 | 8 | 8 | 6 | 7 | 8 | 3 | 5 | 4 | 8 | 4 | 3 | 6 |
| A9 | 5 | 7 | 7 | 7 | 7 | 8 | 8 | 7 | 8 | 5 | 8 | 7 | 5 | 7 | 6 | 8 | 6 | 4 | 7 | 3 | 5 | 4 | 8 | 2.5 | 4 | 4.5 |
| A10 | 5 | 7 | 7 | 7 | 7 | 8 | 7 | 7 | 8 | 6 | 8 | 7 | 4 | 6 | 7 | 8 | 6 | 3 | 8 | 2 | 5 | 4 | 8 | 4 | 4 | 5.5 |
The results (Table 4) indicate that technology-enabled delivery methods achieve the highest overall performance scores, with UAVs(Ci = 0.797), innovative/electric vehicles (Ci = 0.729), and delivery by droids (Ci = 0.693) ranking highest. This reflects their ability to better align with the dominant evaluation criteria identified in Stage S2, particularly emissions reduction and operational efficiency. In other words, these methods perform well not only because they are technologically advanced, but because they directly address the most heavily weighted sustainability and efficiency drivers.
Sustainability, technological performance, and the rank of LMD-M obtained by Fuzzy TOPSIS
| Alternatives | Si+ | Si− | Ci | Ranking |
|---|---|---|---|---|
| Delivery boxes | 0.020 | 0.040 | 0.660 | 4 |
| Attended home delivery | 0.043 | 0.027 | 0.385 | 10 |
| Unattended home delivery | 0.024 | 0.032 | 0.576 | 7 |
| Cargo Bike | 0.027 | 0.029 | 0.524 | 9 |
| Crowdsourcing | 0.019 | 0.035 | 0.641 | 6 |
| Click and Collect | 0.031 | 0.036 | 0.54 | 8 |
| Innovative Vehicles/Electric Vehicles | 0.015 | 0.041 | 0.729 | 2 |
| Drones/Unmanned Aerial Vehicles (UAV) | 0.013 | 0.049 | 0.797 | 1 |
| Autonomous Delivery Robots (ADRs) | 0.020 | 0.038 | 0.652 | 5 |
| Delivery by Droids | 0.017 | 0.039 | 0.693 | 3 |
However, the results also show that performance differences across alternatives are relatively moderate, with mid-ranking solutions such as autonomous delivery robots (0.652) and crowdsourcing (0.641) indicating incremental performance variation rather than sharp differentiation. This suggests that LMD performance is influenced by multiple interacting factors rather than a single dominant capability.
Traditional methods also remain competitive, as delivery boxes (Ci = 0.660) achieve performance levels comparable to those of advanced alternatives. This suggests that performance is not solely driven by technological capability, but also by the ability to deliver consistent operational efficiency and service reliability under existing infrastructural and regulatory conditions. In this sense, traditional methods benefit from implementation feasibility and system compatibility, which partially offset their lower technological sophistication.
This pattern indicates that LMD performance is shaped by the interaction between performance potential and implementation constraints. While advanced I4.0-enabled solutions offer higher theoretical performance, their effectiveness depends on enabling conditions, such as infrastructure, regulations, and organisational readiness. Accordingly, two complementary pathways emerge: (1) deployment of advanced delivery modes where enabling conditions are sufficiently developed, and (2) incremental optimization of traditional methods, such as routing, electrification, and packaging improvements, where such conditions remain limiting.
Insights from industry professionals (Table 5) reinforce and contextualize the quantitative findings. Organizations reported implementing route optimization, recyclable packaging, renewable energy use, and driver welfare initiatives, aligning with the importance of operational efficiency and sustainability drivers identified in the AHP results. At the same time, pilot deployments of UAVs and energy-efficient infrastructure support the strong performance of advanced delivery methods observed in the TOPSIS ranking, although their implementation remains constrained by regulatory and operational barriers.
Professionals' insights regarding sustainable I4.0 for LMD
| Question | Professionals feedback |
|---|---|
| 3 |
|
| 4 |
|
| 5 |
|
| 6 |
|
| 7 |
|
Professionals also highlighted key challenges, particularly high initial investment in IoT technologies and workforce resistance, which explains the conditional performance of I4.0-enabled solutions identified in Table 4. To address these barriers, firms are adopting phased implementation strategies and investing in workforce training. Looking forward, emerging technologies such as automation, IoT sensors, AI-driven analytics, and digital twins were identified as critical enablers for improving efficiency, reducing emissions, and enhancing decision-making. Additionally, organizations are increasingly adopting performance measurement practices, such as carbon tracking, energy consumption, and customer satisfaction KPIs, which further support the integrated sustainability and operational performance perspective identified in this study.
5. Discussion
This section interprets the results of the S-4.0-LMD-Ms framework by explaining the drivers of performance and positioning them within the literature on sustainable and I4.0-enabled LMD. More importantly, it shows that performance emerges from the interaction between sustainability priorities, technological capabilities, and contextual constraints, rather than from isolated dimensions.
A first key insight is the dominance of environmental performance, led by greenhouse gas emissions, but realised through operational efficiency mechanisms, particularly stopping time and packaging reduction. This indicates that sustainability in LMD is embedded within efficiency-driven decision logics rather than treated independently. This corroborates research linking sustainability to optimization and resource efficiency (Ghode et al., 2020), while refining perspectives on sustainability performance (Morioka and de Carvalho, 2016). More critically, environmental outcomes are achieved conditionally through operational configurations rather than directly optimized as standalone objectives. This is reflected in the ranking, where UAVs (Ci = 0.797) and electric vehicles (Ci = 0.729) outperform other methods (Tsolakis et al., 2022; Moshref-Javadi et al., 2020; Apruzzese et al., 2023).
Second, the social dimension is balanced, with customer satisfaction, job creation, and safety contributing comparably. While this supports service-centric views of LMD (Risberg and Jafari, 2022), it challenges broader social sustainability constructs (Chiappetta Jabbour et al., 2020) by showing that social performance is primarily operationalised through service quality. This suggests that social sustainability in LMD is functionally embedded within service performance rather than operating as an independent dimension. This aligns with droids' delivery performance (Ci = 0.693) and with the literature on service responsiveness (Hübner et al., 2016).
Third, the findings show that I4.0 performance is constrained by foundational enablers, particularly infrastructure, regulation, and workforce readiness, rather than advanced digital capabilities. This supports contingency-based perspectives (Seyedghorban et al., 2020; Huang et al., 2023) and indicates a sequencing effect, where basic conditions precede technological value creation. This highlights that technological capability alone does not translate into performance but depends on alignment with enabling conditions. This explains variations in rankings and aligns with prior studies (Huang et al., 2023; Alverhed et al., 2024).
Fourth, although technology-enabled methods outperform traditional ones, their superiority is conditional. The gap between potential and implementation reflects infrastructural and regulatory constraints. Lower-ranked methods such as attended home delivery (Ci = 0.385) and cargo bikes (Ci = 0.524) reflect known limitations. This indicates that performance differences are driven by alignment between technological potential and contextual feasibility, rather than technological sophistication alone.
The results also reveal a hybrid transition pathway in which traditional methods remain competitive. The close performance between delivery boxes (Ci = 0.660) and delivery by droids (Ci = 0.693) demonstrates that incremental optimization can rival advanced technologies, consistent with evolutionary logistics transformation (Bjørgen et al., 2021; de Assis et al., 2022). This challenges linear assumptions of technological superiority and suggests multiple coexisting pathways of performance improvement.
LMD performance emerges as context-dependent and configuration-driven, shaped by the alignment between sustainability priorities, technological capabilities, and operational conditions, reinforcing its socio-technical nature (Seyedghorban et al., 2020). Thus, performance should be understood as an emergent outcome of interacting system elements rather than the optimization of individual dimensions.
5.1 Theoretical implications
This study contributes to the literature on sustainable logistics and I4.0 by advancing the conceptualization and evaluation of performance in LMD, with specific relevance to SPM. While SPM frameworks conceptualize performance across environmental, economic, and social dimensions (Schaltegger et al., 2014; Morioka and de Carvalho, 2016), the findings show that these dimensions are not independently operationalised in digitally enabled LMD systems but rather interact within a broader technological and contextual framework.
First, the findings refine SPM by demonstrating that sustainability dimensions are hierarchically and conditionally structured rather than equally weighted. Environmental performance dominates, but only when aligned with operational efficiency drivers such as stopping time and packaging reduction. This indicates that sustainability outcomes are embedded within operational decision logics rather than functioning as standalone evaluation criteria.
Second, the study shows that incorporating I4.0 into performance evaluation requires extending SPM beyond its traditional boundaries. Technological factors, particularly infrastructure, regulation, and workforce readiness, emerge as key conditions shaping performance. This highlights that sustainability performance in LMD is contingent on technological feasibility and system readiness, rather than solely driven by sustainability objectives.
Third, the results demonstrate that LMD performance is configuration-dependent rather than method-dependent. While technology-enabled delivery methods achieve higher performance overall, their effectiveness is conditional, and traditional methods remain competitive when incrementally improved. This indicates that performance evaluation should move beyond static comparisons toward understanding how combinations of sustainability, technology, and operational factors jointly shape outcomes. From a theoretical perspective, the TOPSIS-based ranking reflects this configuration logic, as performance emerges from proximity to an “ideal” profile rather than from isolated factor superiority.
Fourth, this study contributes by demonstrating how SPM can be operationalised within an MCDM framework. The S-4.0-LMD-Ms framework captures trade-offs, priorities, and interdependencies across performance dimensions, positioning SPM as a decision-oriented, context-sensitive evaluation approach by integrating sustainability and I4.0 factors within a single analytical structure.
The study extends SPM in LMD by showing that performance is not determined by isolated dimensions, but by the interaction and alignment among sustainability priorities, technological capabilities, and contextual conditions, requiring a configuration-based evaluation perspective.
5.2 Practical implications
The findings provide direct decision-support guidance for logistics managers and policymakers operating in complex LMD environments.
First, the results indicate that performance improvements should prioritize operational efficiency drivers that directly influence sustainability outcomes, particularly stopping time and packaging reduction. As shown in the analysis, environmental performance is achieved through efficiency mechanisms, suggesting that managers should focus on route optimization, delivery consolidation, and packaging redesign to generate immediate sustainability gains.
Second, the TOPSIS ranking highlights that technology-enabled delivery methods, particularly UAVs (Ci = 0.797) and electric vehicles (Ci = 0.729), offer the highest overall performance, but their implementation is conditional. Managers should therefore adopt a phased investment approach, prioritizing foundational enablers such as infrastructure readiness, regulatory compliance, and workforce capability before scaling advanced I4.0 solutions.
Third, the results demonstrate that traditional delivery methods remain viable when optimized, as reflected in the strong performance of delivery boxes (Ci = 0.660). This indicates that firms operating under resource or regulatory constraints can pursue incremental improvement strategies, including routing optimization, partial electrification, and digital tracking, rather than relying solely on full technological transformation.
Fourth, the findings highlight that technology adoption should be treated as a system-level transformation rather than a standalone investment. Constraints related to infrastructure, regulation, and workforce readiness indicate that successful implementation requires alignment between operational processes, organisational capabilities, and external conditions, rather than isolated technology deployment.
From a policy perspective, the results suggest that regulatory frameworks and infrastructure development are critical enablers of sustainable LMD transformation. Policymakers can accelerate adoption by supporting urban logistics infrastructure, regulatory clarity for emerging technologies (e.g. UAVs and autonomous systems), and workforce development initiatives, thereby reducing implementation barriers and enabling scalable, sustainable delivery systems.
Thus, the findings support a dual pathway for LMD transformation: (1) targeted deployment of advanced I4.0 delivery solutions where conditions permit, and (2) systematic optimization of existing delivery systems where constraints persist.
5.3 Social implications
The findings highlight that social sustainability in LMD is operationally driven rather than broadly conceptual, with customer satisfaction, job creation, and safety emerging as the most influential social factors.
First, the prominence of customer satisfaction (0.3012) indicates that social performance in LMD is primarily realised through service quality and reliability. This is reflected in the strong performance of high-ranking delivery methods, such as UAVs and droid delivery, which enhance delivery speed and flexibility. However, their adoption depends on user acceptance and trust, suggesting that socially sustainable LMD systems must align technological innovation with customer expectations rather than prioritizing efficiency alone.
Second, the comparable importance of job creation (0.2938) and safety (0.2923) highlights a balanced workforce-related impact. While technology-enabled delivery methods improve efficiency, they also imply shifts in labor requirements. The results suggest that LMD transformation should consider workforce adaptation and safety conditions, particularly as operational models evolve with increasing automation.
Third, the findings indicate that social outcomes are closely linked to environmental and operational performance, rather than independent objectives. High-performing methods, such as UAVs and electric vehicles, reduce emissions and improve delivery efficiency, thereby indirectly enhancing urban conditions, such as congestion and service accessibility. However, as identified in the evaluation factors (e.g. privacy concerns, cybersecurity), these benefits are conditional on responsible and context-sensitive deployment.
Finally, the continued competitiveness of traditional methods, such as delivery boxes (Ci = 0.660), suggests that socially sustainable transitions in LMD can follow incremental, inclusive pathways. This enables organizations with limited technological or financial capacity to improve service quality and sustainability outcomes without full reliance on advanced technologies.
Arguably, the results show that social sustainability in LMD is embedded in service performance, workforce considerations, and implementation conditions rather than driven by standalone social objectives.
6. Conclusion remarks
This study develops and applies an S-4.0-LMD-Ms to address the need for integrated decision-support approaches in increasingly complex urban logistics systems. Using a collaborative research design and a hybrid fuzzy AHP–TOPSIS framework, the study systematically identifies key evaluation factors and assesses the performance of both traditional and technology-enabled delivery methods.
The study contributes to the literature by advancing the conceptualization and evaluation of LMD performance. First, it extends SPM by integrating I4.0 technological capabilities into the evaluation of LMD systems, addressing the common separation of these domains in prior research. Second, it develops a holistic, multi-dimensional evaluation framework that captures environmental, economic, social, and technological factors within a unified analytical structure. Third, it demonstrates how MCDMA can be operationalised to evaluate configuration-based performance, in which outcomes emerge from the interaction among sustainability priorities, operational efficiency, and technological readiness rather than from isolated factors.
From a practical perspective, the study provides a structured decision-support approach for logistics managers and policymakers to evaluate and design LMD strategies. The findings show that while advanced I4.0-enabled solutions, such as UAVs and electric vehicles, achieve higher overall performance, their effectiveness remains conditional on infrastructure, regulatory frameworks, and workforce readiness. At the same time, traditional delivery methods, particularly delivery boxes, remain competitive when supported by operational and technological improvements, highlighting a dual pathway for last-mile transformation through both technological deployment and incremental optimization.
Despite these contributions, the study has several limitations. The empirical analysis is based on expert judgments within a specific regional context, which may limit generalisability across different environments characterised by varying regulatory conditions, infrastructure maturity, and technological readiness. In addition, the reliance on fuzzy MCDM methods introduces a degree of subjectivity, despite efforts to ensure consistency and validation. The limited availability of real operational data further constrains the extent of quantitative validation of the proposed framework.
Future research directions are outlined in Table 6, which identifies opportunities to extend the framework across different contexts, incorporate richer empirical data, and explore emerging I4.0 -enabled delivery models.
Suggested future research topics
| Theme | Further research topics | Article section |
|---|---|---|
| Performance measurement frameworks |
| 4 |
| Sustainable LMD practices |
| 5 |
| Industry 4.0 technologies for LMD |
| 6 |
| MCDM for LMD |
| 3 |
| Long-term considerations |
| 6 |
Declaration of using AI-related tools
We declare that the Grammarly tool embedded in MS Word has been used to improve the writing of some paragraphs throughout the paper using the “improve it” option.
The supplementary material for this article can be found online.
We would like to sincerely thank the Editor, the Assistant Editor, and the anonymous reviewers for their valuable time, thoughtful comments, and constructive efforts, which have substantially improved the quality of the manuscript.
