Digital Servitization Business Model Innovation Practices for Corporate Decarbonization in Manufacturing Enterprises: A Qualitative Meta-Analysis

Abstract

The global shift toward decarbonization and the rise of the digital economy are compelling manufacturing firms to undergo a complex twin transformation across their structures, operations, and value chains. Business model innovation (BMI), especially in digital servitization (DSBMI), emerges as a crucial catalyst in facilitating this change. However, there is a lack of systematic exploration of how DSBMI influences corporate decarbonization (CD). To fill this knowledge gap, a comprehensive qualitative meta-analysis of 27 case studies was conducted, identifying multiple DSBMI practices for CD employed by industrial firms. These practices can be summarized into three main types: efficiency DSBMI, novelty DSBMI, and convergent DSBMI. A system has at least two of these, while all three may coexist. Based on dynamic capabilities theory, this study also introduces six roles for the three types of DSBMI practices, which interact to help firms sense opportunities, seize them through BMI, and transform their operations and ecosystems—collectively enabling decarbonization through internal optimization (efficiency DSBMI), downstream innovation (novelty DSBMI), and value chain-wide cooperation (convergent DSBMI). The findings offer a comprehensive theoretical framework that guides companies to achieve economic benefits while advancing their CD goals through multi-level BMI strategies. Finally, the study discusses its limitations and proposes directions for future research.

1. Introduction

The global shift toward decarbonization and the rise of the digital economy are compelling manufacturing firms to undergo a complex twin transformation across their structures, operations, and value chains [1]. As a result, organizations must pursue business goals while also lowering carbon emissions. Business model innovation (BMI) provides pathways to achieve both objectives simultaneously [2]. Meanwhile, increasing government support and shifting investor priorities mean that corporate decarbonization (CD)—the process by which a company reduces or eliminates its CO₂ emissions from the atmosphere [3]—can open new market opportunities and generate green competitive advantages for early movers [4,5,6]. As a result, firms are rethinking the types of business models best suited for a low-carbon economy [7,8].

CD requires firms to adopt multiple technologies for managing carbon. For example, carbon capture and storage can enable companies to reuse carbon in production [9] or burn it as a fuel, thereby supporting a transition to renewable energy systems and improving energy efficiency [10]. CD also demands zero-waste approaches, such as recycling end-of-life products and materials to recover their embedded value [11]. Implementing measures like renewable energy deployment or product recycling depends on digital technologies and related services [12]. Consequently, achieving CD is closely linked to firms’ servitization and digitalization strategies. Servitization can extend product lifecycles and increase resource use efficiency [13,14]. Digitalization enhances operational efficiency and enables value-chain decarbonization through big-data-driven optimization [15,16,17]. Although digitalization can create environmental paradoxes, some scholars argue that its environmental benefits outweigh the pollution it causes [18]. However, implementing a servitization strategy without digital technologies can lead to a servitization paradox, where costs exceed benefits [19]. In addition, the most persistent and systemic decarbonization challenges often lie beyond a single firm’s direct control, embedded in complex networks of suppliers, customers, and complementors [20]. As a result, recent scholarship often considers servitization and digitalization together, forming the concept of digital servitization (DS) [20,21]. DS is transforming traditional industrial structures and introducing new logics of value creation [20,21]. BMI is central to this shift: the degree to which BMI aligns with a firm’s decarbonization strategy determines how much decarbonization value DS can generate [22]. Accordingly, firms undergoing DS can reconfigure how they create, deliver, and capture value to support their decarbonization goals [2,23]. In this paper, we refer to the BMI that manufacturing firms undertake during DS as the Digital Servitization Business Model Innovation (DSBMI).

Research on DSBMI is an emerging field [15]. Prior studies have focused on methods for implementing BMI [24], the effects of such innovations across different business model types [25], and the benefits of these models [26]. The current study on DSBMI emphasizes three dimensions—solution customization, solution pricing, and solution digitalization—and uses them to classify business models into types such as product-oriented service providers, industrializers, customized integrated solution providers, platform providers, and outcome providers [20]. Sairanen and Aarikka-Stenroos (2024) [7], in their study of low-carbon business models, identify three CD domains: internal (Scope 1 & 2 emissions), downstream (customer-use emissions: a key part of Scope 3), and value chain-wide (along the value chain Scope 3). These findings motivate two questions for manufacturing: which types of DSBMI practices promote CD, and through what mechanisms do they drive CD? A review of the existing research on DSBMI types reveals that some scholars have explored efficiency and integration types [2], highlighting that the efficient type includes a range of digital green innovation activities that enterprises implement using diverse digital technologies to boost the efficiency of green transactions. The integration type involves stakeholders in the digital green business ecosystem to fully incorporate and utilize various digital green resources to maximize system value. Other scholars have proposed two types: efficiency and novelty [27], with some suggesting that efficiency and novelty in BMI can enhance an enterprise’s ability to seek and engage in external activities, strengthen internal and external collaboration, and deepen open innovation [28]. However, these studies lack an in-depth analysis of the DSBMI practices that drive CD and the mechanisms involved [20]. Additionally, most research on BMI during the DS of manufacturing firms relies on single-case studies [29], with few comparative studies across firms. Due to the scale and heterogeneity of Chinese manufacturing, multiple case studies are necessary to improve the generalizability of findings. Furthermore, scholars have suggested that different business models coexist simultaneously in developing DS enterprises. As one study on organizational duality indicates [30], business model duality contains two distinct but interconnected dimensions: the balance dimension and the combined dimension. The former reflects the degree of balance or alignment among BMI types and can be measured as the absolute difference among themes. The latter reflects the level of interaction or integration among BMI types and can be measured as the product of these themes. Nonetheless, further research is needed to examine the relationships between various DSBMI types and their impact on CD.

Building on the previous discussion, this study aims to help manufacturing enterprises drive decarbonization while advancing their CD goals through multi-level BMI strategies. Accordingly, the research questions of this paper are clearly defined as follows: What types of DSBMI practices do manufacturing firms use to promote CD? Additionally, how do these practices interact dynamically to support CD? Particularly in enabling decarbonization efforts that span organizational boundaries? To answer these questions, we perform a qualitative meta-analysis of 122 DSBMI-related cases from 27 studies examining the DS–CD relationship. The analysis identifies three main DSBMI types used by manufacturers to advance CD: efficiency-, novelty-, and convergent-DSBMI. We then analyze how these three types interact, identifying six interaction roles (e.g., promoting, leading, gap-filling) through which they collectively enable CD. Based on dynamic capability theory, we introduce the concept of DSBMI for CD. We define it as the dynamic process of configuration through which firms, enabled by agile interactions supported by digital technologies, sense low-carbon opportunities and seize them through BMI, thereby advancing CD. This process operates across three interconnected domains: enhancing internal operational efficiency, innovating downstream customer solutions, and—through convergent DSBMI—orchestrating collaborative decarbonization along the value chain. The outcome is the fulfillment of the firm’s own decarbonization and competitive objectives.

This paper makes three main contributions. First, by explicitly connecting DSBMI to the specific goals of CD, it provides a more targeted conceptual definition than the broader DSBMI framework for advancing CD. Second, the study introduces a mechanism that clarifies how different DSBMI types dynamically influence CD, addressing the need for a stronger theoretical basis in this area. Finally, on a practical level, the findings offer a structured framework that helps manufacturers diagnose, choose, and coordinate DSBMI practices to achieve CD.

The paper is structured as follows. Section 2 covers the theoretical background, reviews prior research on BMI, DSBMI, and CD, and explains this study’s perspective on dynamic capabilities. Section 3 details the research method and describes the four steps of the qualitative meta-analysis. Section 4 presents the findings, giving a detailed overview of the three types of DSBMI practices. Section 5 offers a thorough analysis of these findings, clarifies how the three DSBMI types interact, and develops the paper’s theoretical framework by introducing six core roles. Section 6, we review our main contributions, practical implications, limitations, and suggestions for future research.

2. Theoretical Background

2.1. BMI and the Strategic Imperative for CD

Business models allow firms to capitalize on opportunities and generate value for themselves and their stakeholders [31]. They define an organization’s internal logic for creating, delivering, and capturing value [22]. Therefore, BMI involves modifying these core components [32,33,34]. Scholars describe differences in business model change through three main theoretical perspectives, aligning with major schools in strategy research: the rational positioning school, the evolutionary learning school, and the cognitive school [35,36,37,38]. The rational positioning perspective views the business model as the result of managers’ deliberate decisions and operational reasoning [38]. From this perspective, BMI is a rational, design-driven process where managers adapt to external changes to create and capture value. In contrast, the evolutionary learning perspective emphasizes experimentation as the main driver of business model change [22]. It argues that external uncertainty triggers BMI and highlights the importance of conventions, trial-and-error learning, and incremental adjustments in model transformation [39]. Both the rational and evolutionary perspectives see the external environment as the primary force behind BMI. On the other hand, the cognitive perspective focuses inside the firm, examining how managers’ mental models or schemata influence business model change [40]. This view suggests managers can facilitate BMI by changing their cognitive schemas.

In the context of the global push for CD [41], from the perspectives of the government, consumers, and enterprises, it compels BMI not only to adapt to technological changes and market competition [42], but also to adjust to the changes brought about by CD. This requires enterprises to explicitly incorporate the goal of reducing and eliminating CO₂ as a core element of value creation when designing business models (BM) [7]. However, traditional BMI theories lack specific conceptual tools to explain the types of BMI in the DS that drive CD and their interrelationships, thereby facilitating CD [43,44].

Dynamic capabilities theory describes how firms develop and adapt internal and external capabilities to respond to rapidly changing environments [45]. This framework supports our goal of examining the dynamic interactions among different types of DSBMI. Therefore, we use the dynamic capabilities theory to theoretically capture these dynamics. We believe that the core of DSBMI-driven CD is the firm’s demonstration of dynamic capabilities. DSBMI-driven CD involves quickly sensing low-carbon opportunities and regulatory pressures, seizing these opportunities by developing and implementing new business models, and reconfiguring organizational structures and ecosystem relationships to continuously align with CD objectives.

2.2. BMI in DS (DSBMI)

The penetration and derivative effects of the digital economy can be deeply integrated with traditional industries, creating a new form of business involving low pollution and energy consumption [2]. Furthermore, servitization represents a significant avenue for advancing CD [13,14]. Consequently, many scholars have combined these two elements to investigate how their conjunction influences CD.

The BMI of DS is the latest research perspective on DS [15]. Researchers have recognized that digital transformation in manufacturing servitization has changed the methods used for business models [46,47,48]. For instance, most scholars interested in the concept of digitally enabled BMI focus on how BMI is performed [24], its effects on different types of business models [25], and the benefits identified from the model [26]. Research on customer perspectives on DSBMI emphasizes four main points: the opportunities and challenges of DS for industrial customers [49], high levels of customer proximity [50], customer requirements [51], and agile customer co-creation [52]. The stream of DSBMI involving multiple ecosystem actors mainly refers to the types and characteristics of BM within the ecosystem [53,54], as well as approaches to and processes for managing BMI across organizational boundaries [55,56].

Scholars have started classifying DSBMI based on strategic priorities. Some differentiate between efficiency and integration types [2]. The efficiency type includes digital green innovation activities that firms implement—such as cloud computing and artificial intelligence—to enhance the efficiency of green transactions. The integration type brings stakeholders together within a digital green business ecosystem to consolidate and leverage diverse digital green resources for maximum system value. Others suggest an efficiency–novelty distinction [27], arguing that both efficiency and novelty in BMI enhance a firm’s ability to pursue external opportunities, deepen collaboration between internal and external parties, and expand the scope of open innovation [28]. However, the typology of BMI remains fragmented, and the relationships between these types are not well explored. Building on research on organizational duality [30], scholars have observed that different business models often coexist within growing DS firms. Business model duality can be described by two related dimensions: balance and combination. Balance measures the degree of parity or alignment among BMI types and can be quantified as the absolute differences among themes. The combination reflects the level of interaction among BMI types and can be measured by the product of these themes.

Overall, the previous discussion highlights the lack of a clear and comprehensive classification of DSBMI types. Research on how DSBMI practices influence manufacturing CD remains fragmented: most studies provide only broad descriptions of their effects on environmental sustainability [57] and lack a unified, systematic framework that focuses on CD [58,59]. Additionally, the dynamic connections among different DSBMI practice types, and particularly how they interact along the value chain to enable CD, deserve further exploration [7,8].

2.3. Synthesizing the Framework: Low-Carbon Business Models, CD Pathways, and the Role of DSBMI

To fill this gap, we review the emerging literature on low-carbon business models. The most notable contribution is Sairanen and Aarikka Stenroos (2024) [7], who introduce a key typology. They identify 10 types of low-carbon business models that help companies create, propose, and capture value through internal decarbonization, downstream decarbonization, or decarbonization support across the value chain. This typology corresponds with the established Scope 1, 2, and 3 emissions accounting framework [44]. First, corporate internal decarbonization focuses on lowering direct emissions (Scope 1) and indirect emissions from purchased energy (Scope 2) within a company’s own operations, emphasizing process innovations, energy efficiency, and on-site renewable energy generation [41,58,60]. Second, corporate downstream decarbonization targets emissions during the customer use phase of a product or service (a key part of Scope 3), involving business models that help customers lower their carbon footprint through approaches such as product-service systems or efficiency-as-a-service models [61,62]. Lastly, corporate value chain-wide decarbonization promotes emission reductions along the value chain (Scope 3), including business models that create platforms for industrial symbiosis, provide low-carbon technologies, or offer decarbonization-as-a-service [53,63,64].

These works further synthesize that business models achieve CD by employing specific strategies such as improving resource efficiency, using renewable energy and materials, and supporting circular economy practices [59,65,66]. Therefore, this typology of low-carbon business models defines the essential ‘what’ and ‘where’ of CD—clarifying the strategic areas where value is created through emission reductions. However, it leaves open the question of ‘how’ manufacturing firms, especially through the lens of DS, put these models into practice. In this context, DSBMI acts as the key mechanism and enabling process that allows these abstract low-carbon business model archetypes to be implemented, scaled, and made economically viable. Specifically, integrating digital technologies (e.g., IoT, AI, platforms) with service-oriented value propositions provides the tools, data flows, and relational frameworks that enable firms to carry out CD strategies across these domains with unprecedented accuracy and efficiency.

Therefore, this study builds on the synthesized framework by suggesting that the DSBMI practices examined in the literature are not just strategic orientations but are directly connected to specific CD domains through identifiable mechanisms. For example, efficiency DSBMI, with its emphasis on optimization, mainly appears in internal CD. Similarly, novelty DSBMI, by creating new value propositions for customers, facilitates downstream CD. Finally, convergent DSBMI, however, represents a strategic leap: it enables firms to orchestrate and participate in decarbonization along the value chain. This involves integrating operations and data across multiple organizations, often through digital platforms, to achieve system-level efficiencies and emission reductions that individual firms cannot accomplish alone. Importantly, from the focal firm’s perspective, engaging in such ecosystem-level coordination is a means to achieve its own, more ambitious CD goals and to capture value from the broader system. Thus, our framework moves beyond a broad description of DSBMI toward a mechanistic understanding of how it generates environmental value across multiple levels, thereby laying the foundation for our primary research focus.

3. Research Methods

3.1. Qualitative Meta-Analysis Approach

Given that the research area of DSBMI for CD is still emerging and existing case studies show fragmented findings, a qualitative meta-analysis was considered the most suitable methodological approach [67]. It could organize the current case studies within the academic context of implementing DSBMI practices for CD, offer comprehensive explanations for different conclusions [68], and even generate new theories [69]. This approach directly addresses the need to move beyond single-case studies [70] and provides a systematic understanding of DSBMI practices, such as how firms engage in collaboration for decarbonization along the value chain. Therefore, the method used in this study was based on the four-stage qualitative meta-analysis employed by Shen et al. (2023, 2024) [15,71] to explore the research problem. The first two stages, a thorough database search and manual screening, make up the data collection process. The last two stages, an in-depth examination of individual cases and a cross-case analysis of the selected samples, constitute the main data analysis. This structured process ensures both comprehensive literature coverage and rigor in data interpretation, improving the reproducibility and trustworthiness of our results.

3.2. Data Sources and Collection

3.2.1. Stage One: Search in Databases

The first stage was data retrieval. As shown in Figure 1, in September 2024, all the articles in both the Web of Science and Scopus databases were searched for the two subject areas using keywords related to DS and sustainability. We employed “sustainability” as a broad search term to encompass environmental dimensions, including decarbonization. This strategy ensured the inclusion of foundational studies on low-carbon strategies within a “sustainability” or “green” framework, even when they did not explicitly use the term “decarbonization.” Using overly narrow search criteria might have excluded such relevant works. To ensure the quality of the selected papers, we concentrated on searching the Expanded Science Citation Index and Social Science Citation Index versions of the World Library of Science. The keywords used were summaries of those used in literature reviews on the topics of DS and sustainability in high-quality journals. The key search strings combined terms related to four main concept groups: Digitalization, Servitization, Digital Servitization, and Sustainability. For the concept of Digitalization, the search included the following keywords: “remote control” OR “Industry 4.0” OR “automation and industrial robots” OR “additive manufacturing” OR “digiti*ation” OR “Internet of Things” OR “virtual reality” OR “cloud computing” OR “ICT” OR “digital manufacturing” OR “network” OR “artificial intelligence” OR “simulation” OR “cyber-physical system*” OR “digital technology*” OR “digital” OR “remote” OR “digital transformation” OR “smart solution” OR “smart product” OR “autonomous solution*” OR digitali*ation OR “smartization” OR “emerging technologies” OR “IoT” OR “remote monitoring” OR “predictive analytic*” OR “advanced manufacturing” OR “augmented reality” OR “cyber-security” OR “RFID” OR “big data” OR “3D printing” OR “smart data” OR “smart manufacturing” OR “smart factory” OR “AI” OR “digital twin” OR “platform” OR “wearables” [15,21,72,73]. For the concept of Servitization, the keywords were: “service infusion*” OR “IPSS” OR “product-service system*” OR “integrated solution*” OR “servitization*” OR “service transition*” OR “service transformation*” OR serviti*ation OR “PSS” OR “servitisation*” OR “smart service*” OR “advanced service*” [15,21,72,73]. Specific to Digital Servitization, the keywords used were: “smart product-service system*” OR “digital servitization*” OR “digital serviti*ation” OR “digital servitisation*” OR “smart PSS” OR “smart serviti*ation” OR “digital PSS” [15,21,72,73]. Finally, for the Sustainability dimension, which encompasses environmental and decarbonization aspects, the search included: “environmental perf*” OR “social sustainab*” OR “sustainab*” OR “social responsib*” OR “business ethic*” OR “environ*” OR “Circular Economy” OR “Triple Bottom Line” OR “social perf*”OR “CSR” OR “corporate social responsibility” OR “environmental manag*” OR “ESG” OR “environment, social & governance” OR “carbon” OR “climate” OR “emission*” OR “circular*” OR “green*” OR “sustainab* transition” OR “socio-ecologic*” OR “corporate responsibility” OR “corporate environmentalism” OR “responsib*” OR “environmental sustainab*” OR “corporate social respons*” OR “sustainab* development” OR “CO₂” OR “ecol*” OR “environmental innovation” OR “natural environment*” OR “green* innovation” [74,75,76,77,78,79,80,81,82,83,84,85,86,87].

Initially, the keywords were applied to the Web of Science and Scopus databases, from which 2938 and 4750 articles were retrieved, respectively. However, the results contained unnecessary articles in areas such as computer science, so they were filtered by the field of study category. This left only articles in business, management, and economics/accounting. This search process removed a large number of articles, leaving only 179 and 523 articles from the Web of Science and Scopus databases, respectively. The new results contained different article categories, such as books and conference papers, whereas only articles and reviews related to the research topic were needed. After further filtering, the two databases were left with 322 (Scopus) and 162 (Web of Science) articles. Next, articles in languages other than English were removed, leaving 312 articles in the Scopus database and 162 articles in the Web of Science database (this remained unchanged). Summarizing the results in the two databases revealed 89 duplicate papers. Thus, the first step of the search produced 385 articles. These articles formed the basis for the second part of the process, manual screening.

3.2.2. Stage Two: Manual Filtering Samples

According to Shen et al. (2023, 2024) [15,71], the second stage of data collection in qualitative meta-analysis involves manual screening, which includes three sub-stages: evaluating research methodologies, selecting research topics, and adding supplementary literature. In the first sub-stage, the requirement was to keep only articles that used a case study approach. After screening, 143 of the 385 articles met this criterion. In the second sub-stage, the focus was exclusively on articles featuring case studies in the manufacturing context. After screening, papers were removed if they discussed smart cities or other topics outside this research area. Additionally, in the manufacturing context, only articles discussing DSBMI for CD realization were retained. After screening, 22 articles met the requirements and were kept. During the third sub-stage, a Google Scholar search was conducted for papers on the research topic, which yielded five more articles that met the criteria. In total, 27 eligible case study articles were identified for the research sample; this number aligns with the recommended range for qualitative meta-analyses, which is 12 to 100 papers [67,88]. These 27 articles included 122 business cases discussing the realization of BMI for DS in DS (see

Table 1. Case studies are presented in this paper.

No.	Paper	Research Question	Industry/Firms	Examples of CD Content
1	Pal (2016) [89]	How do firms extend their responsibilities through servitization in used-clothing PSS?	The textile-fashion industry/7 cases	“Case 7, towards a more sustainable consumption, …dispose their old clothes, based upon a GPS system, along with relevant discount information and vouchers on dropping off their old clothes.” (p. 10)
2	Spring and Araujo (2017) [90]	How does the product life cycle affect its value and opportunities for reuse in the servitization and CE?	Manufacturing industry, automotive industry, IT hardware industry, etc./2 cases	“Products get repaired, refurbished, upgraded, tinkered with, dismantled, reassembled and discarded.” (p. 27). “…modifying product designs with re-use and remanufacturing in mind at the outset.” (p. 17) …
3	Grubic and Jennions (2018) [91]	What are the characteristics of the organizational application of Remote Monitoring Technology (RMT) in the context of a service-oriented strategy?	Manufacturer/4 cases	“The RMT functionality required here is more than prognostic functionality since the latter aims to determine the remaining useful life of a product and not to foresee how its effectiveness will change.” (p. 15) …
4	Cedeño et al. (2018) [92]	How can manufacturing companies leverage the IoT from a customer-oriented perspective for developing smart services?		“Optimization of equipment operations: Implement operation optimized based on historic operation data” (p. 5) “Predictive services: Trigger service activities based on current component condition” (p. 5) “Anticipated spare part orders by forecasting real-time demands” (p. 5) …
5	Holtström et al. (2019) [93]	What are the key aspects of developing sustainable clothing consumption business models?	The apparel retailer Houdini Sportswear/1 case	“The re-projects consist of four interlinked service areas: recycle, reuse, repair and rental. […] Together, these re-projects aim to reduce garment consumption, but not necessarily the use of products which can achieve higher usage” (p. 10) …
6	Frishammar and Parida (2019) [94]	How does the transformation of circular business models actually occur in existing enterprises?	Manufacturing/8 cases	“Case 2: Digital fleet management system… The circular economy outcomes include reduced fuel consumption (5–6%) and prolonged life of tires.” (p. 10) …
7	Yang and Evans. (2019) [95]	What is the sustainable value created in different archetypes of PSS business models?	Manufacturing/3 cases	“Increased resource and energy efficiency and reduced carbon emission … In most situations, both customers and manufacturers have the incentive to increase resource and energy efficiency in the use phase of products. (p. 6) Table 3. Reduced total emission (C) (p. 13) …
8	Jabbour et al. (2020) [96]	How do manufacturing companies transition from linear manufacturing to the sharing economy?	Manufacturing/2 cases	“Company A has increased the level of services in its products… it has expressed interest in adopting SE as a means to achieve its environmental sustainability objectives, such as zero carbon emissions.” (p. 8) …
9	Marić (2020) [97]	How can entrepreneurs develop business strategies to tackle the challenging 3D printing market?	Manufacturing/1 case	“...our customers teach us new things every day! And it is unbelievable how creative they are with the machines [3D printers]...” (p. 10) …
10	Chen et al. (2021) [98]	How is the DS business model of traditional product manufacturers changing?	Manufacturing/1 cases	“Gree entered into partnership with Yingli Solar, a manufacturer of solar panels, with the objective of jointly developing solar-powered air conditioning solutions. This solution labeled ‘Zero Carbon Health Home’, was based on Yingli’s solar panel, Gree’s solar-panel-compatible air conditioner and the digital management module specific to photovoltaic air-conditioning solutions, called ‘Integrated Management System’...” (p. 17) …
11	Haftor and Climent. (2021) [99]	How can industrial organizations reduce the products that have a negative impact on the natural environment?	Manufacturing/1 cases	“The fuel volume for the measured trucks was reduced by 22% to 26% per truck and kilometer driven. The direct effect was a corresponding reduction in CO2 emissions of between 76 and 86 CO2 tons per year and a corresponding reduction in fuel costs.” (p. 5) …
12	Paiola et al. (2021) [100]	How DS leads to sustainability?	Packaging machines and retail equipment/4 cases	“In other words, sustainability benefits have to become an inherent constitutive element of new value propositions in BMI, founded on advanced digital services that allow unprecedented performance in machines and equipment energy efficiency, material consumption, and waste reduction.” (p. 8) …
13	Reim et al. (2021) [101]	What are the capability development and challenges of large manufacturing firms in implementing CBMs?	Manufacturing/3 cases	“Service agreements are closely connected to the reuse, remanufacture, and recycle logic of the circular economy. A researcher at Company B emphasizes the point: ‘We are very much working with sensor and measuring projects to offer service applications such as predictive maintenance by building on machine learning to do high quality maintenance.’” (p. 7) …
14	Ding et al. (2021) [102]	How to develop a cyber-physical production monitoring service system?	manufacturing/1 case	“In addition, the energy efficiency of an individualized product can be computed. As shown in Figure 7 (d), the total energy consumption of a customer’s product reached 41 kWh, the final carbon footprint 0.5 kgCO2e, and its average economic carbon efficiency 33.61 yuan/(W·h). The information above is presented to the customer to inform them of the environmental impact of the individualized product.” (p. 12) …
15	Fani et al. (2021) [103]	How to use hybrid simulation design for PSS in the fashion industry?	the fashion industry/1 case	“The phenomenon of Fashion renting is opposed to the widely established business model of fast Fashion, ensuring a longer life cycle of clothing, reducing the consumption of raw materials, the pollution generated in the production phases and at the end of the life cycle of the articles, in order to promote environmental sustainability.” (p. 6) …
16	Moro and Cauchick-Miguel. (2022) [104]	How to analyze the bike sharing system from the perspective of business model?	the bike-sharing industry/1 case	“PSS could favor obtaining circular results in the value configuration because the suppliers own the product rather than the customers. In circular economy research, PSS often appears as an enabler; allowing multiple life cycles to the product involved in the solution…” (p. 6) …
17	Thomson et al. (2022) [105]	How do industrial equipment manufacturers coordinate the development of technology, business models, and ecosystem relationships?	Manufacturing/4 cases	“Digitalization is recognized as a key factor in the distributed nature of specialized knowledge… enabling a system-of-systems approach that optimizes resource use and reduces waste.” (p. 9) … “Implement and refine standard interfaces to increase the potential for platform consolidation and collaboration… enabling smarter analytics and reducing ‘islands of data’.” (p. 20) …
18	Sjödin et al. (2023) [106]	How can dynamic capabilities enable industrial enterprises to commercialize AI circular business model innovation (CBMI) in DS?	Manufacturing, transport solutions, shipping, construction, and mining/5 cases	“Autonomous vehicles can be programmed to take the most efficient routes, which can increase efficiency by reducing fuel consumption and emissions… The combination of autonomous and electrified vehicles has reduced CO2 emissions by 95% on an industrial customer site.” (p. 8) …
19	Laur and Berntzen (2023) [107]	Why and how have the business models of the energy industry changed?	Digital utility provider, Information services, Electricity distribution and service, Energy services, Electro installations, Software services, Energy flexibility provider, Energy producer, It and computer-based services, IT services, Measurement equipment and service, Data and IT services, Electricity service and installations, Data and IT services, Electricity production and service/19 cases	“Energy entrepreneurs lifted the most benefits from service provision rather than energy sales due to strict regulations ... Instead, add-on services focused on energy flexibility (monitoring, control and effectivisation) overrun traditional businesses... We get an almost 70% increase in intakes due to the delivery of innovative solutions.” (p. 8) …
20	Wu and Pi. (2023) [108]	How synergistic are the PSS and digital technology in manufacturing firms in a CE?	Manufacturing/3 cases	“Through the mapping of P-oriented PSS, manufacturing firms can meet the reduction principle by improving product quality to reduce the elimination rate for better controlling carbon pollution ... Through the mapping of U-oriented PSS, manufacturing firms can increase their product utilization through the sharing of services to reduce raw material waste.” (p. 6) …
21	Asi et al. (2024) [109]	How can the DS model bridge the relational asymmetry between providers and customers?	Manufacturing/1 case	“Xerox adopts a Value-Based Pricing strategy, where prices are determined based on the value customers attach to the products and services rather than the production costs.” (p. 12) …
22	Chang et al. (2024) [12]	How is the current literature studying smart PSS, and why is customization crucial for it, especially in the field of sustainability?	Civil aviation industry, Automobile industry, Electronics industry, Marine equipment industry, Special equipment industry, Steel industry, Telecom and networking industry/7 cases	“Company 3 realized energy savings through the energy efficiency management of the smart campus service solutions. Company 4 improved resource allocations for customers by using a smart port management service solution, which realized energy savings and emission reductions.” (p. 17) …
23	Rong and Luo (2023) [110]	What is the evolutionary path of the sharing economy?	Sharing economy/19 cases	“In the adapted sharing model, people rent their excess resources that are adapted to be shared—like cars, beds, boats, and other assets—directly from each other (p. 8) ... By contrast, in the born sharing model, firms design and produce products that are able to be shared and sold as part of this access to services (p. 2) ... The ‘born sharing’ products are designed and manufactured for sharing with certain attributes that are feasible for an access-based sharing model, and will be efficiently consumed in a shared mode throughout the whole lifecycle.” (p. 8) …
24	Miehé et al. (2023) [111]	How can new complements use connectivity technology to align with the value proposition of existing ecosystems?	Autonomous transportation, digital mobility platforms/4 cases	“Second, the new complementors increased the environmental sustainability of the ecosystem value proposition in every case. In case ALPHA, the drones... reducing energy consumption... reduce noise emissions and air pollution... In case BRAVO, autonomous vehicles... tend to consume less energy than vehicles operated by a human driver... In case CHARLIE...” (p. 8) …
25	Mattos et al. (2024) [112]	How can organizations develop new business models from the Internet of Things?	Manufacturing/3 cases	“The intelligence generated by the platform allows, among other advantages, to increase sales by up to 15%, reduce maintenance expenses by 10%, and reduce energy costs by 15%. Therefore, there is no doubt that innovations like these, based on the IoT and which gather a gigantic and powerful volume of relevant data, can also be widely used by brands to generate surprising and differentiated experiences for consumers.“ (p. 9) …
26	Alcayaga and Hansen (2024) [113]	What SCS activities have companies adopted to implement maintenance/repair, reuse, remanufacturing, and recycling strategies?	Machine building and heavy-duty vehicles, Components and equipment manufacturing/17 cases	“Figure 3. Smart circular system funnel … spanning smart cross-strategy, smart use, and smart circular strategies (maintenance/repair, reuse, remanufacturing, recycling).“ (p. 14) …
27	Turienzo et al. (2024) [114]	Where, how, and by whom will the creation, capture, and delivery of value perceived by the customer be concentrated?	Telecommunications, oil, logistics, electrical suppliers, repair workshops, road infrastructure, insurance companies, consultancy/69 candidates	Cooperation through digital platforms can achieve cost savings that make it possible to meet environmental requirements (electrification, hydrogen, etc.) and improve service by maximizing cost containment through increased efficiency and maximized use of resources. (p. 18) …

).

3.3. Data Analysis

3.3.1. Stage Three: Within-Case Analysis of Samples

Based on the third stage of the qualitative meta-analysis approach described by Shen et al. (2024) [71], a within-case analysis was conducted on the 122 identified business cases. A hybrid deductive-inductive coding approach was used to ensure the analysis was both theoretically informed and grounded in the empirical data [115]. First, drawing on the three theoretical foundations of BMI—rational positioning, evolutionary learning, and cognitive theory [35,36,37,38]—and combined with the four common BM dimensions [116,117]—value proposition, value creation, value delivery, and value capture—the deductive coding scheme was developed for the within-case analysis (see

Table 2. Deductive Coding Scheme.

Theoretical Background	Central Themes	1st Order Concepts (Deductive Codes)	Exemplary Papers
Rational positioning: BMI is an optimal design process for managers to respond to external changes and achieve value creation and capture.	Source of change: External shocks have altered interdependence Concept: Optimization based on rational economic analysis of exchange relationships Design: Redesign to reflect the analysis of the interdependence of changes BMI outcome: repositioning to achieve the best fit with specific backgrounds and strategies	Repositioning of new offerings (customer) to fit the given environment and strategy. Repositioning of new customer segments/markets (customers) to fit the given environment and strategy. Repositioning of new capabilities (value creation logic) to fit the given environment and strategy. Repositioning of new technologies/equipment (value creation logic) to fit the given environment and strategy. Repositioning of new processes/structures (value creation logic) to fit the given environment and strategy. Repositioning of new partnerships (value creation logic) to fit the given environment and strategy. Repositioning of new channels (customer) to fit the given environment and strategy. Repositioning of new customer relationships (customer) to fit the given environment and strategy. Repositioning of new revenue models to fit the given environment and strategy. Repositioning of new cost structures to fit the given environment and strategy.	(Abdelkafi et al., 2013 [35]; Ammar and Ouakouak, 2015 [36]; Clauss, 2017 [37]; Martins et al., 2015 [38])
Evolutionary learning: Business model change arises from external uncertainty, values the role of routine, and emphasizes trial-and-error learning and incremental revision of BMI.	Source of change: Uncertainty caused by exogenous changes Concept: Unspecified entrepreneurial cognition and insight, generating preliminary hypotheses Design: Trial and error learning to reduce uncertainty BMI outcome: Constantly evolving designs adapt to small changes	Evolutionary design adapted to small changes in new offerings (Customer) Evolutionary design adapted to small changes in new customer segments/markets (Customer) Evolutionary design adapted to small changes in new capabilities (Value creation logic) Evolutionary design adapted to small changes in new technologies/equipment (Value creation logic) Evolutionary design adapted to small changes in new processes/structures (Value creation logic) Evolutionary design adapted to small changes in new Partnerships (Value creation logic) Evolutionary design adapted to small changes in new channels (Customer) Evolutionary design adapted to small changes in new customer relationships (Customer) Evolutionary design adapted to small changes in new revenue models Evolutionary design adapted to small changes in new cost structures	(Abdelkafi et al., 2013 [35]; Ammar and Ouakouak, 2015 [36]; Clauss, 2017 [37]; Martins et al., 2015 [38])
Cognitive: It starts from within the enterprise, focuses on the role of managers’ mental models or schemas in business model changes, and points out that managers can help achieve BMI by changing their cognitive schemas.	Source of change: Process of schema change—Analogical reasoning—Concept combination Concept: Utilize concept transfer knowledge to restructure or modify existing business models—analogical knowledge transfer—combinatorial knowledge restructuring Design: Modify and integrate transferred knowledge into existing business models BMI outcome: A reconfigured business model; A variant of the existing model.	New Offerings (customer) innovation resulting from managers changing their cognitive schemas (analogical reasoning, conceptual combination) New customer segments/markets (customer) innovation resulting from managers changing their cognitive schemas (analogical reasoning, conceptual combination) New capabilities (value creation logic) innovation resulting from managers changing their cognitive schemas (analogical reasoning, conceptual combination) New technologies/equipment (value creation logic) innovation resulting from managers changing their cognitive schemas (analogical reasoning, conceptual combination) New processes/structures (value creation logic) innovation resulting from managers changing their cognitive schemas (analogical reasoning, conceptual combination) New partnerships (value creation logic) innovation resulting from managers changing their cognitive schemas (analogical reasoning, conceptual combination) New channels (Customer) innovation resulting from managers changing their cognitive schemas (analogical reasoning, conceptual combination) New customer relationships (Customer) innovation resulting from managers changing their cognitive schemas (analogical reasoning, conceptual combination) New revenue models innovation resulting from managers changing their cognitive schemas (analogical reasoning, conceptual combination) New cost structures innovation resulting from managers changing their cognitive schemas (analogical reasoning, conceptual combination)	(Abdelkafi et al., 2013 [35]; Ammar and Ouakouak, 2015 [36]; Clauss, 2017 [37]; Martins et al., 2015 [38])

). Using this coding scheme, first-order coding was performed on DSBMI practices across the 122 business cases. During this process, many DSBMI practices were not identified within the deductive coding scheme; therefore, following the recommendations of Gioia et al. (2013) [115], this information was explored through inductive or open coding, resulting in inductive first-order coding. After the third step of the within-case analysis, 420 first-order concepts were identified, including 30 deductive and 390 inductive codes.

3.3.2. Stage Four: Cross-Case Analysis of Samples

In the last stage of the qualitative meta-analysis approach, a cross-case analysis was conducted of the 420 first-order codes revealed in stage 3. More specifically, these first-order codes were compared and simplified, for example, by merging some of the 30 first-order codes that had similar meanings. Upon further analysis, these 30 codes (sub-DSBMI activities) could be condensed into eight second-order themes (key DSBMI behaviors). The connections and differences among these eight second-order themes were further analyzed, revealing three aggregate dimensions (types of DSBMI practices) after the interrelated themes were aggregated. Their data structure is shown in Figure 2.

4. Findings

Our qualitative meta-analysis shows that manufacturing firms employ three distinct DSBMI practices (efficiency, novelty, and convergent DSBMI) to enhance CD. These types are differentiated along three key dimensions: (1) the primary locus of value creation (internal, customer-facing, or ecosystem-level); (2) the degree of interorganizational coordination required (internal optimization, bilateral collaboration, or multilateral orchestration); and (3) the principal decarbonization scope addressed (internal Scopes 1 & 2, downstream Scope 3, or along the value chain Scope 3). The overall data structure leading to these dimensions is shown in Figure 2.

Table 3. DSBMI Practices for CD: Typology and Literature Evidence.

DSBMI Type	Specific Practices	Representative Literature Evidence
Efficiency DSBMI	Adding digital technology into physical products to facilitate the development of digital business models	Grubic & Jennions (2018) [91]; Rong & Luo (2023) [110]; Chen et al. (2021) [98]; Reim et al. (2021) [101]
	The incorporation of application software into enterprises to enhance agile interaction	Cedeño et al. (2018) [92]; Sjödin et al. (2023) [63]; Holtström et al. (2019) [93]; Thomson et al. (2022) [105]
	The construction of digital platforms to facilitate efficient collaboration within ecosystems	Sjödin et al. (2023) [63]; Haftor & Climent (2021) [99]; Chen et al. (2021) [98]; Miehé et al. (2023) [111]
Novelty DSBMI	Personalized customization services that leverage digital technology	Chang et al. (2024) [12]; Mattos & Novais Filho (2024) [112]; Paiola et al. (2021) [100]; Ding et al. (2021) [102]
	Services that are informed by AI learning capabilities	Sjödin et al. (2023) [63]; Haftor & Climent (2021) [99]; Reim et al. (2021) [101]
	Solution provider practices	Frishammar & Parida (2019) [94]; Jabbour et al. (2020) [96]; Asi et al. (2024) [109]; Haftor & Climent (2021) [99]
Convergent DSBMI	The establishment of interconnectivity between front- and back-end company operations through the utilization of smart connected products	Chen et al. (2021) [98]; Reim et al. (2021) [101]; Ding et al. (2021) [102]; Holtström et al. (2019) [93]
	Convergence of operations across organizational boundaries via digital platforms	Sjödin et al. (2023) [63]; Mattos & Novais Filho (2024) [112]; Miehé et al. (2023) [111]

summarizes the three types of DSBMI practices and some supporting literature.

4.1. Efficiency DSBMI Practices

Efficiency DSBMI can use digital technologies to enhance internal operations, improve resource use, and cut waste. It targets three key areas: first, integrating digital technology into physical products to facilitate the development of digital business models; second, incorporating application software into enterprises to enhance agile interaction; and third, constructing digital platforms to facilitate efficient collaboration within ecosystems. These practices contribute to internal CD (Scope 1 and Scope 2 emissions) by streamlining processes, increasing resource efficiency, and minimizing waste.

4.1.1. Adding Digital Technology into Physical Products to Facilitate the Development of Digital Business Models

The use of digital technologies enables firms to enter the digital economy by adopting digital business models, which include both digital transformation products and services [91,101,110]. Through digital technology adoption, firms create new value propositions, for example, by positioning technological solutions as unique products [93] or offering shared cloud-computing capabilities [110]. Organizations generate value by pursuing intelligent transformation, gradually automating and digitizing production to address inefficiencies and enhance product quality. They use automated equipment, expand digital infrastructure, and connect factory devices to networks to gather real-time data. These actions combine automated manufacturing with connectivity and digitalization, leveraging digital control and analytics to promote industrial growth and reduce resource waste. To deliver value, firms have opened new sales channels [104] and introduced digitally enabled after-sales services [98], thereby lowering costs and boosting delivery efficiency. To capture value, challenges mainly involve developing new revenue models, providing better after-sales support, and adapting to service-driven revenue changes, such as pay-per-use models for disposables [94].

4.1.2. The Incorporation of Application Software into Enterprises to Enhance Agile Interaction

Embedding internal development platforms and mobile applications within enterprises fosters agile interactions across both back-end and front-end domains [114]. From a value proposition perspective, firms focus on increasing efficiency in front- and back-end production and services—such as ERP systems—to enable integrated analysis of production, sales, and after-sales activities [98]. In terms of value creation, companies are shifting toward higher levels of intelligence, reconfiguring critical resources and streamlining specific processes [112,113]. Enterprise back-end data self-service systems facilitate intelligent analysis and use of production data, including operational status monitoring [92,99]. They also support autonomous management of the smart shop floor for tasks such as maintaining operations [102], predicting equipment failures [63], and providing intelligent visualization [100,112]. Conversely, front-end systems gather data from customer relationship management platforms and other sources [108] to improve service efficiency. To deliver value, firms have implemented transparent systems to facilitate supplier collaboration [93] and have adopted digital marketing and sales approaches [89]. These advancements can streamline production and enable the distribution of new, more sustainable fabrics to other manufacturers [93]. Regarding value capture, companies are developing smart payment solutions that remain accessible without reliable local Wi-Fi [26] and designing operational processes aligned with their equipment architecture [105].

4.1.3. The Construction of Digital Platforms to Facilitate Efficient Collaboration Within Ecosystems

The adoption of digital technologies, particularly through digital platforms, has enabled enterprises to connect front- and back-office functions [102], making the study of enterprise ecosystems a vital research focus [63]. Building such platforms facilitates more efficient cross-firm boundary collaboration [63]. In terms of value proposition, firms now focus on improved cross-boundary coordination and comprehensive solutions that meet customer needs [63]. Concerning value creation, platform-based firms can more effectively integrate actors’ resources and capabilities, such as recruiting technical talent [96] and fostering communities based on professional experience and skills—collective assets difficult for a single firm to replicate [99]. Value delivery has shifted from individual firms to inter-actor collaboration [101], with channels transitioning from single-channel approaches to coordinated multi-channel strategies, significantly improving delivery efficiency [98]. Lastly, firms diversify revenue streams. For example, platform membership fees and training charges for individual actors now contribute to income growth [111].

In summary, efficiency DSBMI practices enhance a company’s operational efficiency throughout the industry chain, lowering emissions, optimizing resource use, and providing benefits to both the implementing firm and its broader ecosystem.

4.2. Novelty DSBMI Practices

Novelty DSBMI is characterized by its customer-centric locus of value creation. It focuses on innovating value propositions and delivery mechanisms to meet emerging or latent customer needs, often through bilateral coordination between the firm and its customers. Its primary decarbonization impact is downstream (a key part of Scope 3), achieved by enabling customers to reduce their carbon footprint through more efficient, tailored, or alternative product-use patterns. Deploying novelty DSBMI can foster effective alignment between technology and consumer needs, creating conditions for successful commercialization [27]. Novelty DSBMI can be broken down into three main aspects: personalized customization services that leverage digital technology, services informed by AI learning capabilities, and solution provider practices.

4.2.1. Personalized Customization Services That Leverage Digital Technology

Novelty-oriented firms enable customers’ co-creation in manufacturing [102,104], offer personalized products via agile manufacturing [12], and design products as experiences [112]. They also use scenario-based innovation to craft distinctive, attractive brand images. Consequently, firms must identify and meet consumer preferences [63,96], pursue sustainable production, and integrate resources rapidly to enable real-time responses [98]. This is realized through interactive customer platforms [12] and extensive data analysis to capture personalized demand, stimulate engagement, and develop targeted solutions [100,105,112], thereby enabling precise resource allocation [63,99]. Scenario-based innovation is also applied offline, incorporating multiple experiential elements to encourage consumption [109,112]. This strategy strengthens online–offline consumer interactions [105] and aligns brand concepts with consumer values. Such practices reduce the firm–customer gap, deepen relationships, improve the accuracy of resource allocation, and enhance efficiency by preventing resource mismatches and backlogs [118,119].

4.2.2. Services That Are Informed by AI Learning Capabilities

Artificial intelligence (AI) has enabled numerous novel business models, with ChatGPT as a notable example, and has sparked a new wave of AI-driven innovation. Many companies and individuals now provide AI training, generating additional income from consultancy fees. For instance, a virtual production business model focused on sustainability can be achieved through digitalization, although some models are not viable if they lack sustainable potential [101]. Similarly, digitizing a platform can transform the delivery of a machine into a service system built around it, allowing for new payment models [99]. AI systems also enhance customer processes to utilize resources more efficiently. As a result, AI’s learning ability supports the development of advanced, customer-specific solutions [63].

4.2.3. Solution Provider Practices

Firms adopt a range of solution–provider practices. First, they strengthen R&D in digital technologies and build patent portfolios [95,99,110], consolidating core technological capabilities. Second, firms explore higher-value markets, formalize market innovations, and develop new approaches to target customer segments [94,99]. Third, they expand product–service offerings [94,96] to provide greater customer value, such as through process-optimizing solutions [109]. Finally, companies build digital platforms [12,98] to accelerate service delivery, supporting intelligent operations [63] or AI solutions [99]. Together, these practices shift firms from product manufacturers to solution providers, combining technological and market innovation. This transition enhances competitive advantage and sustainability, removes product-resource constraints through servitization, and encourages healthier industry competition driven by technological and service innovation.

To improve their value proposition, firms grow sustainable-value businesses and develop higher-value offerings for customers. They also adopt strategies to boost customer loyalty and build unique brands. AI-powered services assist in delivering personalized offerings, reducing ineffective products and services and thus preventing waste of resources.

4.3. Convergent DSBMI Practices

Convergent DSBMI differs fundamentally from the novelty type by shifting the locus of value creation to the ecosystem level. It is defined by multilateral coordination and integration along the value chain, involving partners, complementors, and sometimes even competitors [28]. Its core objective is to orchestrate a platform ecosystem to enable value chain-wide decarbonization [7]. This is achieved not merely by serving customers better, but by reconfiguring roles, relationships, and resource flows across the network. For the focal firm, successfully implementing convergent DSBMI is a strategic means to achieve deeper decarbonization and a competitive advantage that would be unattainable in isolation. Convergent DSBMI emphasizes two main areas: first, the establishment of interconnectivity between front- and back-end company operations through the use of smart connected products; and second, the convergence of these operations across organizational boundaries via digital platforms.

4.3.1. The Establishment of Interconnectivity Between Front- and Back-End Company Operations Through the Use of Smart Connected Products

In developing intelligent products, companies adopt unified IoT chips to generate standardized data and facilitate customer data sharing [98], which guides product interaction and connectivity. Transparent supplier relationships also simplify production processes and support the creation of environmentally friendly products [93]. These practices collectively promote the development of business ecosystems—exemplified by Xiaomi’s smart home system. Intelligent interconnection components not only improve the functions of smart connected products but also transform isolated items into tailored, integrated system solutions [101], thereby meeting broader customer needs. For instance, customers can interact with manufacturers and suppliers during production through modular digital platforms like the Gree product platform [98,102]. As a result, companies increasingly redefine internal product boundaries through interaction links, enabling interconnectivity between front- and back-end operations and providing more advanced customization and system solutions. Simultaneously, these practices strengthen collaboration among platform actors, merge the boundaries between different ecosystems, and create meta-ecosystems for value delivery [111].

4.3.2. Convergence of Operations Across Organizational Boundaries via Digital Platforms

To expand their boundaries, firms engage in ecological collaboration [120], such as convergent DSBMI, where companies use digital platforms to enable cross-departmental or cross-firm collaboration, thereby forming platform ecosystems [63]. Informatization-focused firms can share information via public media [89], build industry-related ecosystems, and support product suppliers through information-sharing platforms [111,112]. These platforms facilitate embedded [94] and value-adding [121] supplier relationships, allow product diversification, strengthen control over critical resources, and expand firm boundaries. As intelligent products become more widely and tightly connected, firms can broaden their scope to include a variety of related smart, connected products and services [111]. Establishing multi-scenario linkages further creates smart, connected ecosystems by integrating previously isolated resources with external stakeholders, thereby generating greater resource value and functionality [122].

In conclusion, convergent DSBMI, by integrating resources within and across companies, facilitates more efficient resource utilization through ecosystem collaboration. This integration enhances CD along the value chain, benefiting all actors. For the focal firm, this translates into achieving its own decarbonization targets through leveraging ecosystem synergies, thereby contributing to environmental protection by conserving resources and extending product life cycles.

5. Discussion

This study found that manufacturing firms do not implement DSBMI practices in isolation but deploy multiple practices simultaneously [63], forming a complex interactive system that drives CD. Drawing on dynamic capability theory, we propose a theoretical framework that explains the dynamic interactions among efficiency, novelty, and convergent DSBMI practices (see Figure 3). The framework details how these three types of practices work together through six roles to drive a firm’s CD across internal, downstream, and value-chain levels. Below, we outline the six core interaction roles and then explain how DSBMI, through these roles, promotes CD.

5.1. The Promoting and Driving Roles Between Efficiency and Novelty DSBMI

The study shows that efficiency and novelty in DSBMI practices interact to drive CD jointly. Efficiency DSBMI practices can promote the development of novelty DSBMI practices in two ways. First, by improving back-end production and optimizing front-end customer interactions, efficiency practices establish a stable digital-technology foundation that enables firms to pursue new development opportunities [108]. Second, the operational data produced on that foundation helps firms identify those opportunities, thereby promoting novelty DSBMI practices [100,115]. In turn, the emergence of novelty DSBMI practices—supported by efficiency measures—enhances resource-use efficiency, yielding greater value with fewer inputs and reducing resource intensity and waste, thereby contributing to decarbonization [58,59,104]. This pattern aligns with transformational dynamic capabilities, since it requires reconfiguring internal operations to pursue new strategic directions.

Conversely, novelty DSBMI practices can drive the development of efficiency DSBMI practices. Emerging customer needs drive these novelty practices, compelling enterprises to continuously improve their operational efficiency to meet these demands and seize new business opportunities [28,39]. For example, implementing new AI-based optimization services may lead companies to enhance the energy efficiency of their data centers and computing infrastructure to stay cost-competitive. This feedback loop helps prevent the waste of ineffective resources, further decreasing waste and encouraging ongoing improvements in internal CD [93,94]. This process exemplifies the sensing and seizing capabilities, which involve recognizing and capitalizing on new market opportunities to boost internal efficiency.

This dynamic interaction continually enhances operational efficiency and fosters new innovations, forming a self-reinforcing cycle that strengthens the firm’s dynamic capabilities and supports decarbonization across both internal and downstream operations.

5.2. The Foundational Support and Guiding Roles Towards Convergent DSBMI

Convergent DSBMI practices rely on both efficiency and novelty DSBMI practices. Efficiency DSBMI practices are fundamental because they embed digital technologies into products and organizational processes, improving back-end production and streamlining interactions between the firm and customers [89,92,99,105]. As firms develop, these improvements help connect front-end and back-end functions so that customer needs identified at the front are quickly communicated to and addressed by the back end [102]. This timely connection enables firms to coordinate with partners, including those from other industries, supporting the development of convergent DSBMI practices [63]. Without digital interconnectivity between front- and back-end operations, creating a seamless cross-enterprise ecosystem—and thus advancing convergent DSBMI—is not feasible [53]. Efficiency DSBMI practices provide the critical digital technology foundation for implementing convergent DSBMI. This foundation helps firms effectively combine resources and capabilities across ecosystem actors [101]. It also enables them to respond adaptively to service issues, adjust strategies in real time, and reduce resource waste and inefficiencies. Therefore, these practices promote CD by demonstrating firms’ transformative potential and establishing the internal infrastructure needed for advanced collaboration.

Novelty DSBMI practices guide convergent DSBMI practices by identifying new customer needs [96] and uncovering business opportunities [100,112]. They also provide reference information that supports convergent actions, such as helping firms find suitable partners [63]. By clarifying the direction of corporate BMI, novelty practices reduce resource waste caused by ambiguous strategic objectives [63,99]. Additionally, they enable more targeted and effective ecosystem interventions that promote environmental protection. This guiding role aligns with sensing capabilities since firms actively scan their environment to identify opportunities for value creation at the ecosystem level.

Implementing efficiency and novelty DSBMI practices separately lays the groundwork for convergent DSBMI efforts. These two types of DSBMI practices provide the capabilities and strategic guidance needed for convergent DSBMI initiatives [28]. From a dynamic capabilities perspective, efficiency DSBMI practices serve as the operational base and align with the transformative capability dimension. Conversely, novelty DSBMI practices offer directional guidance and align with the sensing capability dimension. Together, they enable CD across the firm’s entire value chain.

5.3. The Gap-Filling and Quality-Enhancing Roles of Convergent DSBMI

Convergent DSBMI practices are essential for both efficiency DSBMI practices and novelty DSBMI practices [7]. First, these convergent DSBMI practices address a key gap in efficiency DSBMI. By integrating operations across organizational boundaries, convergent DSBMI practices can identify inefficiencies within a company’s ecosystem [28]. This reveals areas needing improvement and helps firms find sources of resource waste that were previously overlooked [100,112]. Moreover, this integration enables companies to learn better practices and methods from their partners [63]. Such insights can result in cost savings, resource conservation, and improved environmental outcomes. This process demonstrates a higher-level view in which the ecosystem itself becomes a vital source of intelligence for spotting internal gaps.

Second, convergent DSBMI practices can enhance the quality of novelty DSBMI practices. These practices encourage customers to express new and more complex needs that were previously unrecognized [101]. Additionally, they motivate companies to identify new opportunities [111,112]. This continuous flow of enriched market intelligence enables firms to adjust their innovation strategies in real-time. As a result, they can more effectively fulfill their guiding role in convergent DSBMI practices and improve their internal efficiency [50,52]. This dynamic helps create greater value for the enterprise [111], generates economic benefits [102], and promotes environmental improvement [93]. Consequently, this role enhances the company’s ability to sense and act on more advanced and valuable opportunities, driven by comprehensive ecosystem intelligence.

The DSBMI framework helps companies adapt their business models in real-time to meet changing internal and external conditions. This process improves their CD performance and strategic flexibility throughout the entire value chain.

5.4. Defining DSBMI for CD

The dynamic interactions among the three DSBMI types collectively enable manufacturing firms to pursue decarbonization at multiple, interconnected levels. Efficiency DSBMI lays the foundation by optimizing internal operations, directly reducing Scopes 1 and 2 emissions. Novelty DSBMI extends this impact downstream by innovating value propositions to help customers reduce their Scope 3 emissions. Convergent DSBMI amplifies and integrates these efforts by orchestrating collaborative networks, enabling the focal firm to address emissions embedded in the broader ecosystem and thus achieve a more systemic and profound decarbonization impact.

Therefore, we define DSBMI for CD as a dynamic, capability-driven process wherein manufacturing firms, leveraging digital technologies and service logic, innovate their business models to sense and seize low-carbon opportunities. This process strategically coordinates three domains: (1) internal optimization (efficiency), (2) downstream engagement (novelty), and (3) value chain-wide orchestration (convergent), to advance the firm’s own decarbonization and competitive goals. The convergent form, in particular, represents a strategic choice to transcend organizational boundaries, not as an end in itself, but as a means to unlock collective efficiencies and circularity that directly contribute to the firm’s carbon reduction performance and green competitiveness.

This conceptualization deepens our understanding of DSBMI by viewing it as a dynamic capability rather than a collection of separate practices. It helps organizations systematically address CD challenges by coordinating various types of BMI. Our theoretical framework (Figure 3) offers a tool for understanding how different types of DSBMI practices interact through specific mechanisms to promote decarbonization in a synergistic way. This framework not only clarifies the different types of DSBMI practices but also shows how they work together in complex, dynamic systems to drive CD, providing both theoretical and practical benefits.

6. Conclusions

Using a qualitative meta-analysis approach, this study examines 122 BMI-related cases from 27 studies exploring the relationship between DS and CD. The research identifies three significant types of DSBMI that drive CD in manufacturing enterprises: efficiency DSBMI, novelty DSBMI, and convergent DSBMI. We further explain how these three types of DSBMI work synergistically through six roles—promoting, driving, foundational support, guiding, gap-filling, and quality-enhancing—to collectively advance CD across internal operations, downstream customer use, and along the value chain collaboration. Moreover, grounded in dynamic capabilities theory, this study introduces the concept of DSBMI for CD, defined as a dynamic innovation process in which manufacturers, enabled by digital technologies and services, configure and reconfigure their business models to sense and seize low-carbon opportunities. This process strategically integrates internal optimization, downstream innovation, and ecosystem orchestration to achieve the firm’s decarbonization objectives. The convergent form, in particular, enables firms to transcend their boundaries and orchestrate collaborative networks, thereby achieving systemic, reciprocal decarbonization impacts that ultimately enhance the firm’s carbon reduction performance and green competitiveness. The theoretical contributions, practical implications, limitations, and future research directions of this study are outlined below.

6.1. Theoretical Contributions

The findings of this study offer critical theoretical contributions to the literature on manufacturing CD, dynamic capabilities, and BMI. First, by clearly defining the concept of DSBMI, this study enhances the understanding of the specific goal of CD in manufacturing. This concept goes beyond the broad idea of environmental sustainability, grounding it in the practical aspect of enterprise decarbonization and framing it as a dynamic innovation process. This conceptualization provides a crucial focus for future research and addresses the earlier-mentioned fragmentation in the literature on integrating DSBMI to promote CD.

Second, this study advances BMI theory by developing an integrated framework that elucidates how three DSBMI types—efficiency, novelty, and convergent—function as an interconnected system. This framework details six interaction roles (promoting, driving, foundational support, guiding, gap-filling, and quality-enhancing) through which these types collectively enable CD. Rooted in dynamic capabilities theory, it demonstrates that the sensing, seizing, and transforming capabilities are manifested through the synergistic interplay of different DSBMI practices. Crucially, it clarifies that ecosystem orchestration (via convergent DSBMI) is a strategic capability firms develop to achieve their own decarbonization goals through collaborative networks, thereby bridging firm-level strategy with ecosystem-level outcomes.

Finally, our research integrates and extends existing theoretical streams. It connects the three core perspectives of BMI (rational positioning, evolutionary learning, and cognitive) with the CD pathways identified in low-carbon business model literature and the dynamic capabilities perspective. This integration provides a more granular model for understanding how DS, through BMI, enables manufacturing firms to navigate decarbonization challenges from internal operations to value chain-wide collaboration.

6.2. Managerial Implications

The findings of this study can effectively help managers in manufacturing companies that are undergoing simultaneous digitalization, servitization, and green transformation. First, the proposed DSBMI practice types and their dynamic theoretical framework offer manufacturing managers a strategic roadmap and diagnostic tool. Executives can use this to evaluate whether their current BMI portfolio is suitable and then adjust their strategic deployment accordingly. For example, a company that is strong in efficiency DSBMI but weak in novelty DSBMI might miss out on new market opportunities and the driving force of more radical ecosystem-level innovation. Similarly, managers should evaluate whether their convergent DSBMI initiatives are effectively designed to orchestrate ecosystem actors, thereby unlocking ecosystem-level decarbonization benefits that directly enhance the firm’s carbon reduction performance.

Second, the specific DSBMI practices identified in this study provide manufacturing executives with a practical menu of options. Using this menu, companies can effectively avoid costly strategic missteps and detours. Specifically, by comparing their key DSBMI practices to this menu, executives can identify weaknesses and leverage digital technologies to improve in those specific areas. This makes resource allocation for digital-green transformation more targeted and effective, helping managers prioritize interventions based on their current capabilities and strategic CD objectives.

Finally, the three dimensions of manufacturing CD identified in this paper—internal, downstream, and value chain-wide—can help corporate executives clarify the value proposition of their DSBMI investments. Using these dimensions, they can communicate more effectively to stakeholders how specific innovations contribute not only to economic value but also to concrete emission reductions: within operations (Scope 1 & 2), in customer use (download Scope 3), and through collaborative action across the platform ecosystem (value chain-wide decarbonization). This aligns strategic initiatives with both corporate sustainability reporting and the more ambitious goal of driving net-zero transitions within their industrial ecosystems.

6.3. Limitations and Avenues for Future Research

Every study has its limitations, which also suggest areas for future research. The main limitations of this study are as follows. First, our reliance on published case studies may introduce a success bias, as unsuccessful DSBMI attempts or implementations with neutral environmental outcomes are less likely to be documented. This may present an overly optimistic picture of the DSBMI-CD relationship. Second, while qualitative meta-analysis is excellent for theory building and integrating insights across contexts, it has inherent constraints. It does not allow for statistical generalization or the testing of causal hypotheses. Furthermore, synthesizing findings from cases with unique contextual factors inevitably entails some interpretive simplification. Finally, our focus is limited to the manufacturing sector, which may affect the transferability of findings to service or hybrid industries. Although our sample includes a variety of manufacturing sub-sectors, which aids robustness, the specific dynamics we identify may manifest differently elsewhere.

Therefore, future research could follow these directions. First, to address methodological limitations, future work should focus on empirical testing and quantitative methods. Developing and validating scales to measure the three DSBMI types and their six roles would enable large-scale hypothesis testing to strengthen the evidence for the links we propose. Second, future research should delve deeper into the contextual factors and barriers affecting DSBMI’s efficacy for CD. Studies could examine how organizational structure, leadership, culture, or national policy environments enable or constrain the dynamic interactions within our framework. Third, a crucial avenue is to investigate the governance, value-capture, and incentive alignment mechanisms across value chain-wide decarbonization enabled by convergent DSBMI. Research is needed on how environmental and economic value is distributed among ecosystem actors to ensure collaborative efforts are sustainable and equitable. Finally, future research should include in-depth discussions from an ecosystem perspective. Studies need to explore how DSBMI for CD co-evolves with the roles of policymakers, standards, financial institutions, and civil society in enabling or limiting the synergistic interactions within our framework. Research on how policy instruments differently affect various DSBMI types and their interactions would be especially helpful for developing effective policy mixes.