A Longitudinal History-Based Review of the Product-Service System: Past, Present, and Future

Abstract

The research field of product-service systems (PSS) has experienced rapid growth since it emerged at the end of the 80s. Consequently, the conceptual structure of the PSS research field has expanded, and this phenomenon has driven novel research topics and future trends. This study analyzes the evolution and conceptual structure of the PSS research field since its emergence. A systematic literature review based on a bibliometric analysis was conducted, followed by a co-keyword analysis and content analysis, aiming to develop a historical overview of the nature of the research. The dataset involved 868 studies published between 1988 and 2022. The results present an analysis of the evolution of the literature in the following periods: (i) from 1988 to 2015, referred to in this study as the ‘past’, and (ii) from 2016 to 2022 (the ‘present’). A comparative analysis of the two periods is carried out, and the main topics and research trends and perspectives for future studies are identified (the ‘future’). Some of the main research topics include digitalization and sustainability, for which discussions on smart PSS, life-cycle management, ecodesign, knowledge management, development of assessment tools, simulation, and modeling technologies, have appeared. Research opportunities for future studies aligned with those topics are also identified, e.g., map key success factors to positive implementation of proposals that reconcile PSS with digitalization.

1. Introduction

The servitization of manufacturing firms and the offer of product-service systems (PSS) have attracted much attention in the academic literature and business context, motivated, among other factors, by the sustainability potential of such solutions [1]. The drivers for companies to start offering PSS solutions are varied, including the potential of improving firms’ performance, increasing customer satisfaction and improving sustainable resource management [2]. The provision of integrated products and services has also led to business model transformation in many manufacturing companies [3]. The transformation of current business models to start offering integrated product-service solutions brings many challenges to companies, such as organizational and process changes, new ecosystems, and the eminent need to incorporate sustainability and digitalization in business proposals [4].

From an academic perspective, the PSS concept has attracted attention from a range of research communities (e.g., engineering design, operations management, service management, service marketing, information science, sustainability, digitization and circular economy), bringing a variety of perspectives in exploring PSS topics [5,6]. Given that multiple research communities have investigated PSS research topics, the number of publications has grown exponentially over the years, which brings about difficulties in navigating through large volumes of literature in the field. Thus, tools and techniques can be used for synthesizing the knowledge that has been developed, and among them is bibliometric mapping [7]. Scientific mapping aims to show the structural and dynamic aspects of research [8]. A study that employs scientific mapping techniques to identify and discuss PSS research can contribute to complementing existing reviews and provide new directions and perspectives for future research.

Previous reviews have examined PSS research and adopted a specific perspective, such as servitization [9], digital technologies in PSS [10], sustainability [11], life-cycle management [12], and marketing [13]. These studies primarily conducted conventional literature reviews, frequently relying on content analysis. In contrast, our comprehensive bibliometric review, conducted from a historical standpoint, provides an overview of the research field on product-service systems and the diverse research topics examined within it. Furthermore, it aims to showcase the evolution (past, present, and future) of topic areas over time. As a result, it is intended to appeal to a wider audience and, optimistically, encourage interdisciplinary reading and research on PSS that congregates several research communities.

Using the approach of a systematic literature review, the paper develops a history-based overview of the nature of PSS research, considering a bibliometric and content analysis since the genesis of discussions on the PSS subject. The results also provide new insights into the conceptual structure of the PSS research field and identify avenues for future research. The contributions of this article are threefold. First, the article considers that PSS and servitization are two interrelated theoretical constructs that stem from the broader movement of strategic incorporation of services in manufacturing firms. Therefore, both broad constructs are considered from a holistic perspective with their related terms, clarifying their interdisciplinary relationships, whereas previous reviews only provided overviews of the research field on a limited number of topics. Second, the paper relies on bibliometric and content analysis supported by visualization techniques, which are scarce in previous reviews. Third, the article allows visualizing the thematic evolution of the PSS research field in a holistic way. To this end, the article considers a comprehensive dataset of peer-reviewed outlets (868 articles) published from 1988 to 2022.

The remainder of this article is structured as follows. Section 2 describes the research method employed. Section 3 presents the findings and discussions by dividing them into ‘past’, ‘present’, and ‘future’, thus, showing the evolution of the literature and outlining opportunities for future research. Finally, Section 4 draws the paper’s conclusions.

2. Research Methods

This section highlights the research methods used to select and analyze the publications from the PSS research field. It was conducted in three stages. The first one was carried out at the end of 2015 by contemplating an analysis from 1988 to 2015, named in this work as ‘the past’, since the analysis envisaged the first publications on servitization from the 1980s to 2015. The second stage consisted of an update of the search by extending the period of analysis from 2016 to 2022. This stage was referred to in this paper as ‘the present’. Perspectives for future studies are also identified, based on the findings in the two periods that were called ‘future’.

2.1. Defining the Dataset

The Web of Science (WoS) and Scopus databases were selected as the source of articles due to their multidisciplinary nature and the availability of full metadata from indexed articles. In addition, these databases are considered the most comprehensive in the literature [14]. A broad range of PSS terms or keywords was used to develop the search strings. The selected keywords included: servitization, servicification, product-service system, total care products, industrial product-service systems, functional sales, functional products, extended product, hybrid product, hybrid offering, service infusion, product-related services, productization, service integration, covalent products, hybrid value, service engineering, services manufacturing companies, product bundling, service-dominant logic, integrated services, and value bundle.

The search was conducted based on the following criteria: (i) articles and reviews written in English; (ii) published from 1988 to 2022; and (iii) in peer-reviewed scientific journals. A pre-selected sample resulted in 6202 articles from WoS and 7665 articles from Scopus. The next step was to perform a filtering process. A careful reading of the articles’ titles, keywords, and abstracts (and, if necessary, the full texts) was done to verify the fit of the selected articles within the PSS literature. Duplicated articles and those not dealing with PSS topics were removed from the pre-selected sample. Eight hundred and twenty articles remained after the filtering process.

A snowball process helped to expand the initial filtered sample. Additional articles were added from the most cited PSS and servitization references, from the Google Scholar database, and from articles that appeared in other literature reviews. Moreover, a further search was conducted by checking the most cited references (with 10 or more citations) extracted from the PSS articles included in the sample. In both cases, new articles related to the scope of this work were also added since they were not identified in the initial search. This consistent snowball process added 48 additional articles, establishing a final sample of 868 articles.

2.2. Data Analysis

Bibliometry has been used for (i) performance analysis of publications and (ii) scientific mapping to reveal the structure and dynamics of research fields. The bibliometric approach was used in this study to identify the most prolific journals, and most cited keywords (citation analysis). In addition, co-keyword analysis was employed, which is a content analysis technique that assumes authors use keywords to represent the research themes and topics [15]. Other studies have used co-keyword analysis to map the evolution of a research field and uncover its conceptual structure [16].

Approximately 94% of the selected articles contained author-provided keywords, which offered a reasonable sample for this study. Whenever keywords were not provided by the authors (due to the journals’ policy or database conditions), we assigned those (limited to a single count) to these articles based on their titles and abstract [17]. After the extraction of metadata, keywords were standardized using cleaning tools in Vantage Point™ software (v. 15.0) and manually checked by two authors to refine the dataset. For instance, keywords denoting the same meaning were changed into the most frequent keyword occurring in the data set (e.g., PSS, product/service system and product-service system were considered synonymous and were represented by product service system).

Considering the analyzed periods, 83 keywords were cited in the first period (1988–2005), 313 keywords were cited in the second one (2006–2010), 689 in the third period (2011–2015), and 628 in the last one (2016–2022). The power law distribution of keywords indicates that an investigation of the most frequent keywords is sufficient to gather information about the most important research topics [15,18]. Based on this, the most frequent keywords were taken into consideration during the total period from 1988 to 2022 (those with no less than five citations).

Subsequently, a keyword co-occurrence matrix was generated using the most cited keywords to demonstrate the frequency with which each keyword pair appears simultaneously in the articles. The co-occurrence correlation matrix indicates both the similarity and dissimilarity of each keyword pair and serves as a solid foundation for studying the degree of correlation between research topics [18]. Bibliometric networks and maps were then constructed based on a co-occurrence matrix.

In a network, when co-occurrence data was used, a transformation was first applied to the data to derive similarities from the data or, more specifically, to normalize the data. The aim of such a transformation was to enable the application of mapping and clustering techniques. VOSviewer™ software (v.1.6.19) was used in this step to visualize the landscape, applying the association strength normalization [19]. After the normalized network has been constructed, the next step was to position the nodes (keywords) in a network based on a two-dimensional space in such a way that strongly related nodes are located close to each other while weakly related nodes are located further apart. For this, a technique named VOS mapping technique (an alternative to the traditional multidimensional scaling technique) was employed [19], which made it possible to provide information regarding interrelationships between different topics. Of note, the same procedures we applied to generate the co-keyword networks were used in this work.

Lastly, to detect emerging topics, the keyword cumulative frequencies and average growth rates were checked during the final period (2016–2022). Afterward, the content analysis supported the identification of the main research topics and trends in the PSS research subfields in two ways. In the first way, keywords were extracted from the articles and presented in the form of frequencies, which is aligned with the co-keyword technique. In the second one, the papers were categorized regarding the research topics explored, leading to the identification of the main contributions and interpretations of the findings.

2.3. Sample Features

For the analysis of the top journals, the 868 articles were divided into four consecutive periods to analyze the evolution of PSS research over time: P1—1988–2005 (48 articles); P2—2006–2010 (151 articles), P3—2011–2015 (377 articles); and P4—2016–2022 (292 articles). The first period was the longest to establish a representative number of published articles. For this reason, it covered 18 years, and the other three periods involved a shorter period. Although they encompass a larger number of articles, this was essential to better identify the trends, as recommended elsewhere [20]. In addition to showing the number of articles per period,

Table 1. Top 15 journals based on the number of articles.

Journals	Area/Sub	JCR *	Periods **
			P1	P2	P3	P4	T	%
Journal of Cleaner Production	Environmental Engineering	11.072	10	13	36	62	121	13.72
Sustainability	Environmental Engineering	3.889	0	0	3	49	49	5.90
Industrial Marketing Management	Business Management	8.890	2	10	15	5	32	3.63
Journal of Manuf. Tech. Management	Management Operation	8.144	0	11	12	2	25	2.83
Journal of Service Management	Management Service	9.450	2	11	9	1	23	2.61
Int. Journal of Production Research	Engineering Manufacturing	9.018	0	1	18	9	28	3.17
CIRP J. of Manuf. Sci. and Technology	Engineering Manufacturing	3.560	0	14	4	12	30	3.40
CIRP Annals—Manuf. Technology	Engineering Manufacturing	4.482	4	4	8	1	17	1.93
Int. J. of Op. and Production Management	Management Operation	9.360	0	6	7	1	14	1.59
Computers in Industry	Computer Science	11.245	0	1	11	6	18	2.04
Int. J. of Advanced Manufacturing Technology	Engineering Manufacturing	3.563	0	1	11	4	16	1.81
Advanced Engineering Informatics	Computer Engineering	7.862	0	0	1	13	14	1.59
J. of Business and Industrial Marketing	Business Management	3.319	2	0	9	3	14	1.59
Int. J. of Computer Integrated Manufacturing	Computer Engineering	4.420	0	2	8	1	11	1.25
P. In. Mech. Eng. Part B-J. L Eng. Manuf.	Engineering Manufacturing	0.182	0	4	6	1	11	1.25
Total			20	78	154	170	422	47.85

* 2021 JCR indexes (searched on 7 June 2023). ** P1: 1988–2005; P2: 2006–2010; P3: 2011–2015; P4: 2016–2022; T: total.

presents the top 15 journals, classified according to subject area and impact factor as per the Journal Citation Report [21]. As can be seen in Table 1, contributions to the PSS research field were provided by outlets from different disciplines, which enabled the research areas to be approached from multiple perspectives.

3. Results and Discussion

The findings are divided into the aforementioned periods: 1988–2015, which represents ‘the past’ of PSS research, and 2016–2022, which refers to ‘the present’ (it includes the recent past and more recent PSS developments in the past three years). Then, this work provides perspectives for ‘the future’, by providing further work in PSS.

3.1. Stage 1: The Past (1988–2015)

A topological network analysis was carried out to map the past of PSS research. It is worth noting that cluster membership is mutually exclusive, and the position of each node (keyword) displays the similarities or differences among them. Hence, keywords with a high degree of similarity are positioned closer. Figure 1 shows the network of keywords (co-keyword) used by the articles published between 1988 and 2015. The network shows the partition of keywords in five clusters, represented by different colors. In addition, the size of each node (keyword) reflects its normalized weight and, consequently, its importance in the normalized keyword co-occurrence matrix. Each cluster is summarized next.

3.1.1. Cluster 1—The Business Perspective

Cluster 1 (red nodes located at the top-left quadrant of Figure 1) included 28 keywords strongly connected to servitization. The main research topics were service-dominant logic, business model, value, supply chain management, service strategy, service transition, capabilities, service infusion, and performance management. Thus, this cluster involved the discussion about tactical factors concerning the alignment of service strategy, offerings, business model, organizational structure, capabilities, and resources. This discussion was grounded by strategic management theories, such as resource-based view (RBV), dynamic capability, and relational view [22]. While earlier studies [23,24] in this period had strict assumptions about the shift along to a product-service continuum, recent studies have proposed a more contingent approach, which influences the configuration of strategy–structure–resources [25].

The impact of servitization on manufacturers’ performance was a strong topic in this cluster, showing a robust interconnection between servitization and PSS research. Neely [26], for example, provided empirical evidence that servitized firms often generate lower net profits as a percentage of revenues compared to pure manufacturing firms. In contrast, Kohtamaki et al. [27] stated that manufacturing companies can profit from PSS, but it is necessary to engage in the creation of a strong service orientation linked to their PSS offerings. Thus, the servitization–performance relationship attracted attention from researchers in this period. Another important research topic concerned the supply chain, focusing on the effects of the relationship between the subsystem provider–supplier, the outsourcing system, and the creation of collaborative networks [28].

A group of keywords (i.e., service-dominant logic, value, co-creation, and capabilities) demonstrated studies focused on the involvement between PSS and service-dominant logic (S-D Logic) research in terms of problems, concepts, and theories. Seminal works (e.g., refs. [29,30]) analyzed the shift from product-dominant logic towards service-dominant logic, proving a strategic orientation. In particular, S-D Logic in PSS deals with the definition and evaluation of a value proposition [13], the design of integrated solutions and systems to provide such value [31], and the network of actors involved in delivering value [32]. At last, there was a general assumption that servitized companies need new capabilities to make the shift to services [33]. Thus, capabilities were another hot research topic in the past of the PSS research field.

3.1.2. Cluster 2—The Design Perspective

Cluster 2 (blue nodes located at the bottom right quadrant of Figure 1) included 27 keywords. The cluster covered topics related to PSS design, such as service engineering, conceptual design, product life-cycle management, knowledge management, requirement analysis, and service design (and service development), as well as the methods and tools applied in PSS design. For instance, service engineering appeared in the first years of PSS research as an engineering-scientific perspective for developing services and integrated solutions [34]. Arai and Shimomura [35], for instance, presented the fundamentals of service engineering and created a computer-aided design tool (service explorer/service model) for designing services and analyzing the roles of those involved in servicing, especially the receiver. During this period, several PSS design methodologies were proposed, such as the methodology for product-service systems or MePSS [36], fast-track total care [37], and integrated product and service design processes [38]. Some of these PSS design methodologies tended to be service-centric (i.e., service model and MePSS), while others were more product-centric, incorporating the service design activities into product design, i.e., the integrated product and service design processes). Overall, these methodologies considered new activities, actors, relationships, artifacts, values, and other elements when compared to the traditional reference models for new products and new service development.

Another topic that received special attention was the application of sophisticated design and statistical techniques in conceptual design and other PSS design phases [39]. Another topic was the requirement analysis. Although customers, stakeholders, business, and sustainability requirements had been addressed in PSS design literature, it was challenging to consider all these requirements together, and they were considered at varying levels of abstraction [40]. For instance, Berkovich et al. [41] concluded that requirements management was more developed in product and software engineering but neglected other important requirements for developing PSS. In this case, service approaches hardly dealt with requirements management. Addressing this problem, Berkovich et al. [42] proposed a requirement data model (RDMod) that considers different types of requirements and the relations among them. Therefore, topics related to the development of PSS (including requirements, methodologies, and tools) appeared as the main topic in the PSS past research.

3.1.3. Cluster 3—The Sustainable Perspective

The third cluster (green nodes located at the bottom left quadrant of Figure 1) involved 21 keywords and focused on the link between PSS and sustainability. Sustainable PSS, eco-design, and life-cycle were some of the most representative topics in Cluster 3. It is remarkable that there were connections with Cluster 2 (the design perspective) related to the development of sustainable PSS and with Cluster 1 (the business perspective) regarding the transformation process involved in adopting a sustainable business model. This proximity may explain the inclusion of keywords that seem to belong to other clusters (e.g., transformation and design for X). Thus, seminal works in the first period (e.g., refs. [43,44,45]) aimed to clarify the PSS concept and its contributions to the reduction of environmental impacts. Moreover, the conceptual structure of PSS has evolved to provide insights into developing a sustainable PSS. Initiatives in this theme were related to eco-design [46], enhanced ecoefficiency during product and service life cycle [47], and design for sustainability [43].

Sustainability evaluation also played a vital role in developing sustainable PSS. However, it is not a simple task due to the involvement of a vast set of variables at various stages of PSS design. Despite these challenges, sustainability evaluation has also gained the attention of researchers who investigate this topic [48]. Hence, sustainable topics related to PSS continue to be explored in the period, although their scope has expanded. Examples of the emergent topics included the comprehension of PSS acceptance for customers and other stakeholders [49] and how to implement PSS with both environmental and social benefits [50]. To sum up, in the period from 1988 to 2015, the development of sustainable PSS was a challenge for researchers, which raised different lines of investigation.

3.1.4. Cluster 4—The Operational Perspective

Cluster 4 (pink nodes located at the top right quadrant of Figure 1) comprised 17 keywords. This cluster focused mainly on topics related to PSS implementation and operationalization, including the proposal of service operation guidelines, the use of simulation and modeling tools, and changes in the delivery system. As seen in Figure 1, it is noteworthy that some keywords in this cluster were positioned close to Cluster 1 (the business perspective), which means there was a close relationship among these topics. For instance, this was the case for change management and organizational design topics [51].

Considering a comprehensive perspective of PSS operationalization, Baines et al. [33] provided guidelines for manufacturing companies to configure their operations to enable the effective delivery of PSS. In summary, the delivery system should consider (i) testing and repair processes near customers; (ii) the development of relationships in the supply chain; and (iii) high levels of product knowledge associated with customer management and relationship skills. In the same way, Datta and Roy [52] highlighted key elements of an operation strategy (e.g., flexibility, contract incentives, and performance indicators) that may assist manufacturing companies in supporting the delivery of performance-based service contracts.

Another notable topic was PSS simulation, which was relevant for both PSS design and PSS operation and evaluation. Kuo [53] proposed a simulation model for PSS procurement processes that consider renting or buying decisions in a reverse logistics context. In this work, the total system cost was found to be better for the procurement model than for the rental model. Although this result was restricted to a specific case, the paper demonstrated how simulation techniques can be helpful for buying or renting decisions in a PSS operation. Teixeira et al. [54] used ‘prognostics and health management simulations’ to monitor the asset conditions in PSS contracts. They proposed a model with an online simulation framework, which can be used to support short-term operational maintenance decisions. Dimache and Roche [55] developed a dynamic modeling approach using discrete event simulation to assess maintenance systems and help decision-makers to select the appropriate asset monitoring levels.

3.1.5. Cluster 5—The Decision Perspective

Cluster 5 (yellow nodes located at the top right quadrant of Figure 1) was closely connected to Cluster 4, being also influenced by operation and information system disciplines. It was a diffuse cluster dealing with several research topics connected to other clusters, yet there was a general focus on industrial PSS and particularly decision-making, both of which were more mature than the other topics. Industrial PSS covered context-specific and transversal research topics related to the design, implementation, and use of a PSS in a B2B context. The industrial PSS literature covered several other topics, including business models, sustainability, risk management, availability, knowledge management, and PSS processes [56]. Decision-making was another comprehensive topic, covering different problems in PSS implementation, including consequences of behavioral processes in the transition to services [57], the development of customization procedures in PSS offers [58], and how to overcome the barriers of a PSS transition. In addition to these two major topics, the research in this cluster also covered collaboration, cost estimation, uncertainties, complexity, and information sharing. Although they represented peripheral topics, some of them (e.g., cost estimation, collaboration, and availability) gained importance in the last years of the 1988–2015 period.

3.2. Stage 2: The Present (2016–2022)

Figure 2 depicts the co-keywords network based on the most cited keywords in the articles published between 2016 and 2022. It represents the recent past and, in recent years, the present of PSS research. By using the same procedures employed in the previous network, five clusters were identified, which are also represented in different colors.

3.2.1. Cluster 1—The Digitalization Perspective

Cluster 1 (red nodes in Figure 2) includes 27 keywords strongly linked to servitization and digitalization. Important topics in this cluster are servitization, smart products, smart product-service systems, digitalization, and Industry 4.0. Thus, the 1988–2015 Cluster 1 (also focused on servitization) evolved to incorporate the development of digital technologies in the provision of services and product-service systems. For instance, discussions about smart PSS already emerged in 2015, when the term was coined by Valencia Cardona et al. [59], but it has been gaining notoriety in recent years, being a major rising research topic of this cluster.

The acceleration of technological innovation, driven by digitalization and Industry 4.0, has triggered an emerging IT-based business paradigm called digital servitization or smart PSS, depending on the work context. More than 12% of the articles published in this period address smart PSS or have related to digitalization, proving the ascendancy of the theme. The offering of smart solutions and the implementation of digital servitization involve novel digital technologies such as big data, artificial intelligence (AI), digital twin, Internet of Things (IoT), and cyber-physical systems (CPS) [60]. These technologies create a digital transformation by providing interactions between machines, devices, products, and services, accelerating the adaptability of companies to market changes [60]. The link between PSS and digitalization was also driven by the COVID-19 pandemic and also connected to several advantages associated with its implementation, such as customization of requirements, products and services, and optimizations in production processes, such as planning, production and monitoring [61].

A critical discussion in this cluster refers to the provision of smart PSS. In this case, Zheng et al. [62] classify key elements of smart PSS into three levels: (i) product-service level; (ii) system level; and (iii) system-of-systems level. The product-service level refers to connected products, their services, and the supporting infrastructure to serve the customer. The system level considers the systematic integration of a family of solutions, the involvement of multiple stakeholders, and intelligent platform-based systems for performance optimization. Finally, the system-of-system extends the industry boundaries to a set of associated smart PSS as well as external ancillary systems (e.g., data management systems) to further impact the total ecosystem, such as a smart building or a smart home. In addition, Carrera-Rivera, Larrinaga and Lasa [63] highlight that due to its adaptive capacity, smart PSS is able to also generate a transition to sustainable business models, considering environmental, social, and economic perspectives. Thus, Pirola et al. [10] conclude that the research on smart PSS is still recent, and it can take advantage of the scientific maturity of other related areas, such as digitalization and servitization, to develop conceptual models to support both academics and practitioners in the rapid deployment of scientific research advances on smart PSS.

3.2.2. Cluster 2—The Circular Economy Perspective

Cluster 2 (green nodes in Figure 2) includes 23 keywords strongly linked to sustainability and circular economy. In addition to those previous words, the main topics are innovation, life-cycle analysis, and economic growth. Interestingly, the 1988–2015 Cluster 2 (also focused on sustainability) evolved to incorporate topics related to circular economy, which still was an emergent theme in the previous observed period. However, increasing growth is observed between the interrelationship between PSS and circular economy in this cluster. In many studies [6,64], PSS is a type of strategy or business model to implement a circular economy. Both concepts emerged from sustainability [11], and the main advantages of merging them are resource and energy use efficiency, longer product lifespan, and waste reduction [65].

According to Ref. [66], an approach used in the circular economy, like Design for Excellence (DfX), can support the incorporation of services during the design of PSS proposals while assisting designers with knowledge related to the life-cycle phases of a product-service system. DfX is considered a systematic way to design products, aiming at optimizing benefits during the whole PSS lifecycle [67], i.e., besides DfX being linked to Product Lifecycle Management (PLM). Sassanelli et al. [68] also point out its contribution towards Product Service Lifecycle Management (PSLM) by demonstrating its contribution towards PSS proposals. Nevertheless, although the publications previously cited have addressed DfX, PLM, and PSLM, the discussions on these topics are rather superficial and not well-explored in the literature analyzed. This opens up the possibility of future in-depth analysis.

A research stream that remains important is the life-cycle analysis (LCA), and several studies have been conducted on this topic (e.g., refs. [69,70]). These works demonstrate that PSS have superior results in reducing environmental impacts in comparison with other traditional ownership models. Nevertheless, depending on the type of product, resources are required to perform maintenance and upgrading [71]. To evaluate the sustainable potential of PSS proposals, only analyzing the environmental perspective is not enough since sustainability is composed of social, economic, and environmental perspectives. In view of this, other assessment methods have gained recognition, such as life-cycle costing (LCC) and social life-cycle assessment (SLCA). For instance, Kambanou and Sakao [72] point out that LCC is used to assess profitability and provide information about the circularity of materials. On the other hand, SLCA is very little addressed in the literature on PSS, where the only work found is that of Sousa-Zomer and Cauchick-Miguel [73]. However, for the PSS to be considered sustainable, social issues such as safety, welfare and labor need to be evaluated, considering all the stakeholders involved in the life cycle of a product-service system [73].

Among the most recent topics in this cluster, innovation and business model innovation stand out. However, both are naturally driven by a sustainability perspective [74]. As a result, recent topics contemplate an approach focused on advances in information technology (IT) and dynamic forms of production, covering modularization and innovation ecosystems. Furthermore, the association between sustainability and product-service systems occurs in parallel, where environmental aspects are treated in publications about circular economy and economic analyses through a strategy-focused view [75].

3.2.3. Cluster 3—The Performance Perspective

Cluster 3 (blue nodes in Figure 2) includes 20 keywords strongly linked to performance perspective. The main topics are manufacturing, supply chain, life cycle, and product design. This cluster covers topics related to operations and marketing management, considering the entire lifecycle of a product [13]. For this, PSS requires a value network perspective. Hence, collaboration among different stakeholders throughout the supply chain to effectively deliver integrated offerings that meet customer needs and create competitive advantages [76]. Thus, this cluster represents a merge of previous Cluster 4 and 5 in Figure 1 (the operational and decision perspectives, respectively) since it integrates operational aspects in PSS research.

Several studies in this cluster discuss manufacturing services. Chen et al. [77] point out that in the face of high competitiveness, differentiation through high quality is paramount. In addition, supply chain management (SCM) is used to evaluate the flexibility, where mixed integer linear programming models and meta-heuristic algorithms show promising results for solving SCM problems [78], for example, by contributing to maximum resource utilization. This contribution of mixed integer programming was studied by Ghomi et al. [79], aiming to solve service load balancing and transportation optimization problems. An integrated resource flow and load balancing algorithm was proposed to implement task allocation using cloud computing, and the results point to an improvement in resource flow efficiency by 7% and a reduction in response time by 35.5% [80].

Ferreira Junior, Scur, and Nunes [81] emphasize that manufacturing companies, in general, are product-oriented, i.e., tangible attributes. According to Baines et al. [9], simple service strategies aggregated to products have been practiced since 1980, and today they no longer represent a competitive differential. Therefore, several authors (e.g., refs. [59,82]) argue that one way to obtain market advantage is through the incorporation of IoT and big data in manufacturing services. Furthermore, industries are increasingly pressured to find alternatives to sustainable development. In this context, the automotive sector stands out, given the direction in the face of electric vehicles and the demand for batteries and recycling of end-of-life vehicles [83]. The automotive sector already presents significant changes since the advances of Tesla, Waymo (Google’s autonomous car project), and Uber have emerged and challenged dominant business models and the design process of the mobility sector [84].

3.2.4. Cluster 4—The Sustainable Design Perspective

Cluster 2 (yellow nodes in Figure 2) includes 13 keywords strongly linked to sustainable design. The main topics are environmental impact, eco-design and sustainable development. In the 1988–2015 network, the design perspective was also relevant (see Cluster 2), although its emphasis was on the technical and methodological integration of products and services. Thus, an engineering focus overcame in that first period. However, the PSS design perspective conveys a sustainable focus in 2016–2022 due to the strong influence of circular economy in PSS research. The proximity between the keywords of Clusters 1 and 2 in Figure 2 ratifies this observation. In addition, the relationship between the sustainability of product-service systems is dispersed and multidisciplinary in the literature, given the complexity and scope of this area since there are different types of PSS (product-, use- and result-oriented), different assumptions (type of energy used in production, transport modal selected, segment worked, among others) and different methodologies used to measure the sustainable potential [85].

Among the most recent topics related to this cluster stands out the use of AI to develop sustainable PSS and allow cleaner production [85]. Li et al. [86] proposed a data-driven reversible framework to achieve sustainable smart PSS, and an illustrative example is portrayed in the sustainable development of a smart 3D printer. This work contributes by providing empirical evidence of the contribution of artificial intelligence in the face of the sustainable potential of product-service systems. The discussion about eco-design still remains strong, but they have expanded to derivative approaches such as lean PSS design [87] and smart PSS design [82]. Among the strategies of these approaches, the use of rapid prototyping to support the design process [88] stands out. Moreover, the co-creation of value [89] is highlighted, where several stakeholders have the power of decision in the design of products and services. In addition, to identify customer needs and better serve them, neuromarketing is used, where neuroscience is used to support the location of the brain region according to the type of emotion provoked by stimuli and behavioral tendencies of customers [89]. To conclude, in this period, the interplay between sustainability and PSS has gained more relevance in PSS research due to the consolidation of new concepts, mainly circular economy.

3.2.5. Cluster 5—Diffuse Perspective

Cluster 5 (purple nodes in Figure 2) includes only 11 keywords strongly related to topics of other clusters, which is proven by their intermediate position between these clusters. Thus, some publications in the cluster have focused on customer acceptance, business modeling, and change to PSS proposals. Social and financial pressures can create a sense of urgency needed for companies to recognize the need for this change. Nevertheless, it requires skills in areas such as marketing, design, sustainability, human resource management, networking and partnerships. It also requires companies developing a PSS business model to think about providing opportunities for customers to consume sustainably. To achieve greater engagement with customers of a product-service system, some authors (e.g., Ref. [76]) highlight the importance of co-creation, interacting with the customer in the business model development process. In addition to this interaction with the customer, the concept of ecosystem emerges, which has been restructuring the way value is created. An ecosystem is defined as an economic community where a group of companies interacts beyond their own borders, adapt to each other, and develop a shared vision, defining governance rules and achieving sustainable competitive advantage [90]. The ecosystem context is directly related to PSS, as exemplified by John Deere, Siemens, and Electrolux, which have integrated services into their product offerings to develop smart farms, smart cities, and smart home ecosystems, respectively [91].

This cluster is also related to customer experience. Publications on this subject have been increasing since companies have been trying to improve customer experiences through data collection, analysis and interpretation, aiming to better meet their needs [92]. This data allows the customization of products and services, according to the customer’s preferences, by meeting different preferences, ages, and purchasing power, among others [93]. Customer experience is a multidimensional construct that encompasses both during and after consumption. It corresponds to a continuous iterative process that analyzes positive and negative customer experiences. Thus, customer experience can be considered a marketing approach applied to achieve differentiation in competitive markets, used to achieve and sustain ‘long-term customer loyalty’ [94].

Cluster 5 integrates concepts from other clusters, e.g., the work of Tokarz et al. [95] analyzes trends and opportunities that digitalization (addressed in Cluster 1) and sustainability-oriented alternatives (addressed in Clusters 2 and 4) have brought to addressing the COVID-19 pandemic. Technologies, such as 3D printing and IoT, to connect professionals in shared telecom laboratories are examples cited by the authors previously cited. Thus, the cluster extols the importance of multidisciplinary approaches in order to mutually contribute to the success of business models.

3.3. Stage 3: The Future (Research Perspectives)

Analysis of keyword citations over the last few periods provided a more accurate indication of future research trends in each cluster.

Table 2. Trends in research topics (2016–2022).

Clusters	Accumulative Frequency							Average Growth
	2016	2017	2018	2019	2020	2021	2022	2016–2022
Cluster 1—The Digitalization Perspective
Digital servitization	4	6	9	7	13	21	13	50%
Digital technologies	4	9	5	11	20	16	20	53%
Digitalization	2	4	7	11	16	14	23	61%
Industry 4.0	2	4	3	5	10	11	15	52%
Product and services	1	5	8	9	15	9	13	52%
Service industry	2	3	7	5	9	14	12	53%
Smart product-service system	1	3	6	12	15	21	30	71%
Smart products	1	2	3	5	11	9	11	58%
Value co-creation	4	7	3	9	17	14	18	52%
Cluster 2—The Circular Economy Perspective
Business	0	1	2	3	1	3	5	58%
Innovation	1	1	3	5	5	9	12	50%
Life cycle analysis	0	1	2	2	3	5	9	58%
Product development	1	2	3	4	6	6	11	50%
Resource use	1	1	2	3	7	6	9	50%
Strategic approach	0	1	3	5	7	5	9	53%
Sustainable business models	1	2	1	3	4	4	7	54%
Sustainability	1	2	4	8	9	12	13	67%
Cluster 3—The Performance Perspective
Big Data	2	4	5	10	8	13	20	58%
Competition	1	5	3	9	15	8	12	58%
Life cycle	2	5	9	6	15	12	16	57%
Customer satisfaction	1	3	5	8	6	8	13	61%
Manufacture	2	5	4	8	16	12	12	50%
Product design	2	5	9	15	12	11	16	56%
PSS design	3	11	5	11	20	17	21	53%
Quality function deployment	4	2	7	4	8	15	18	54%
Cluster 4—The Sustainable Design Perspective
Design	1	3	5	8	5	9	11	63%
Ecodesign	1	2	3	5	5	4	7	58%
Environmental impact	1	2	2	4	3	6	6	50%
Sustainable development	1	3	3	5	8	8	14	63%
Cluster 5—Diffuse Perspective
Business development	1	4	8	14	14	15	21	58%
Business model	1	2	3	7	11	9	15	62%
Internet of Things	3	5	10	15	14	25	27	58%

shows the keywords by clusters, their cumulative frequencies, and their respective average growth rates that correspond to (Year 2 − Year 1)/Year 1). The keywords that were considered were based on the following criteria: (i) reached at least five citations, and (ii) had average growth rates of 50% or more. This demonstrates the association between the keywords and identifies them with promising research topics. Once again, this work excluded keywords meaning general descriptors (e.g., product-service systems and servitization), research objects, or research methods.

The research trends were outlined based on Table 2, i.e., the most recent research approaches in the literature, and are presented next according to the clusters shown in Figure 2.

3.3.1. Cluster 1—The Digitalization Perspective

The topics in Cluster 4 contemplate recent research subjects with a wide rise in recent years, in particular smart PSS (the topic with the highest rise, as can be seen in Table 2). The smart PSS was formally introduced into the literature in 2015 [59], but it can take advantage of the scientific maturity of related areas, e.g., servitization and digitalization, making it possible to develop conceptual models that support both academics and professionals in the rapid and effective implementation of smart PSS. Thus, future research can focus on integrating these research areas to mutually contribute to technological advancement tied to innovation in the subject of product-service systems.

In this context, many publications that integrate engineering and computer science methods and techniques in PSS design are also related to service engineering, a research topic that is expected to maintain its relevance in future years. However, the transition to a digitized model requires new knowledge, methodologies, and tools. The study by Cong et al. [82] indicates that no methodology exists to fully address the design characteristics of smart PSS proposals. Thus, it is expected that future studies will focus on developing suitable approaches for the evaluation and diagnosis of digital technologies.

Pirola et al. [10] highlight the need to concurrently consider the technical and socioeconomic perspectives of a smart PSS. Since it corresponds to a multidisciplinary research area, several factors should be considered before taking the proposal to the market, where future studies can focus on mapping key success factors to ensure the successful implementation of business proposals that reconcile PSS with digitalization. To this end, factors such as (i) structural characteristics such as company size [10,96]; (ii) difference between product, usage, and outcome-oriented PSS proposals; and (iii) difference between B2B and B2C proposals should be taken into consideration.

Another research topic on the rise is the sustainable smart product service system [86], since studies have been proving the relationship between digitalization and sustainable development, contributing to the closing of the cycle before the circular economy [97]. Thus, the concept of a smart circular system arises, which has several competitive advantages over smart PSS, such as cost reduction and new revenue potentials, besides facilitating maintenance, reconfiguration, and recycling during the life cycle [98]. Furthermore, uncertainties in the return flow, such as product quality, location, and turnaround time, can be reduced through digitization [10]. However, an integrated framework linking PSS with sustainability by means is still a research question, where it is suggested to utilize the transdisciplinary domain knowledge of these areas to develop a more consolidated literature and greater autonomy in the development of sustainable smart PSS.

3.3.2. Cluster 2—The Circular Economy Perspective

Although PSS research has emerged in the field of sustainability, Cluster 1 (focused on sustainability and circular economy) has been classified as a cross-cutting theme as it is present in other clusters. Due to the diverse nature of the cluster, it is expected that the development of sustainable PSS and circular PSS are going to continue to attract researchers. Among the most emerging topics in this cluster, life-cycle analysis stands out. This approach is complex due to the integrated design of products, services, information, skills, infrastructure, value chain, and stakeholder integration throughout the life-cycle phases. This will require new solutions by researchers and practitioners. For example, the simulation of life-cycle scenarios through screening life-cycle modeling (SLCM) [99] or the involvement of different stakeholders in PSS design [100].

Specifically, researchers should focus on the role of information and communication technologies (ICT) and circular economy as key topics related to PSS design. In addition, the impact of the circular economy, IoT, smart products and services, Industry 4.0, and big data on PSS design topics should be considered. Design for Excellence (DfX) approaches can be incorporated not only to contribute to the circular economy but also to act in quality improvement, cost reduction, and time optimization throughout the life cycle of PSS proposals. Nevertheless, although works citing the contributions of DfX vis-à-vis PSS and lifecycle management are present in the literature (e.g., Sassanelli et al. [66] and Ringen et al. [67]), there could be a research call for further exploration in order to broaden the literature on those research topics.

As previously highlighted, more integrative approaches are also expected, considering the impacts of a PSS from all scopes of the triple bottom line: (i) environmental through life-cycle analysis; (ii) economic through life-cycle costing; and (iii) social through social life-cycle assessment. An approach that measures impact across the three scopes is still lacking in the literature, then represents a research opportunity. Another important area on the rise in the literature is the interface of circular economy with Industry 4.0 technologies, recognized as an important contribution to achieving sustainability. Several authors highlight that these can result in social, environmental and economic opportunities, while their scarcity can have reverse effects [64]. However, Dev, Shankar and Qaiser [101] discuss the lack of knowledge of smart management and there are financial challenges that prevent companies from using technologies for the circular economy. Research on how to overcome these barriers represents an important research opportunity.

3.3.3. Cluster 3—The Performance Perspective

Traditional research topics such as service infusion, service transition, and service strategies may continue to gain the attention of researchers in the coming years. However, as opposed to the general perspective (i.e., drivers, barriers, and organizational challenges) currently observed, future research is expected to focus on addressing them by demonstrating greater sophistication in dealing with these topics. Examples are risk management [102], desertification pathways [103], and context-specific small and medium-sized enterprises [104]. Emerging complementary research topics focus on business model innovation, service-dominant logic, capabilities, and value system networks.

The link between servitization and S-D logic is evident. Consequently, more attention to research related to S-D logic is expected, especially in relation to value and co-creation. Continuing this trend, future work may further explore the premise of going beyond simple product/service integration in a value-oriented perspective [105]. Furthermore, the roles of customers as co-creators of value [90], the use of skills and expertise as a unit of exchange [9], and innovation ecosystems [91] will also be emphasized. Therefore, further studies on the execution of the service-dominant logic in practice are of paramount importance.

Therefore, the importance of the research topics discussed above may continue in the coming years. In addition, the impact of new external cross-cutting trends (e.g., circular economy, IoT, smart products and services, Industry 4.0, and big data) on service firms should be considered. More specifically, the business model, capabilities, and value system networks will have another resurgence, driven by the servitization relationship and these cross-cutting trends. In addition, new cross-cutting trends linked to sustainability, such as the sharing economy, collaborative consumption, and especially the circular economy, will renew the interest of researchers. So far, these cross-cutting trends are still marginally cited but should become more robust in the near future.

3.3.4. Cluster 4—The Sustainable Design Perspective

PSS design methods represent an important research field as they continue to be explored, as shown in Figure 2. There is a tendency for PSS designers to select a specific aspect and apply different information and communication technologies and for data-driven design [106]. Sustainable design-based methods and tools gain notoriety in the literature, such as eco-design, modularization, biomimicry, and others [107,108]. Another emerging trend is the incorporation of knowledge management construction into PSS design, especially in the conceptual phase of the PSS. Akasaka et al. [109] proposed a system that supports the acquisition of new PSS design solutions by integrating the accumulated information into a knowledge base supported by computer-aided design (CAD). Nemoto et al. [110] focused on using a knowledge base for creating PSS concepts through PSS design catalogs, an idea borrowed from the product design research area. The catalogs organized various design elements, such as core product, needs, function, entity, and actor, and identified them with classification labels to facilitate knowledge search and application. In this topic, knowledge management is applied in the conceptual phase of PSS design.

Digital design technologies gain notoriety in the literature, such as CAD and additive manufacturing. Chaney, Gardan, and Freyman [88] point out that additive manufacturing is able to improve the level of servitization through customized products and the provision of more sustainable value propositions. However, sustainability is composed of environmental, social, and economic perspectives, and sustainable design should stick to this. Maxwell and van der Vorst [111] proposed a sustainable product and service development method based on the triple bottom line (TBL). However, the authors’ view is narrow regarding social impact. Thus, future studies need a more holistic perspective without neglecting any TBL perspective.

3.3.5. Cluster 5—Diffuse Perspective

There is a relative trend toward using refined techniques such as the analytic hierarchy process (AHP), quality function deployment (QFD), failure mode and effects analysis (FMEA), analytic network process (ANP), and data envelopment analysis (DEA) to generate and evaluate PSS concepts [39,112,113,114]. Thus, future developments should cover how to better define, integrate, and evaluate different types of requirements while considering different levels of abstraction; how to balance economic, environmental, and social perspectives in PSS concepts; and how to consider requirements derived from the value chain in PSS design. Although some works (e.g., [40,64]) have already explored these topics, further studies are still needed.

Moreover, large opportunities in Cluster 5 are in the application of simulation, modeling, and system dynamics methods to operation problems supported by technologies such as IoT. For instance, Mashhadi, Vedantam, and Behdad [115] developed a decision model to develop an agent-based simulation framework aimed at modeling the effects of heterogeneous personal traits, social influences, and prior decisions on consumers’ decisions to choose product-service systems. The study by Bortoluzzi et al. [104] corroborates by highlighting the need for small and medium manufacturing enterprises need to align servitization with technological investments. This occurs through a synergistic relationship of Industry 4.0 technologies (such as IoT, advanced simulation, cloud computing and big data analytics—BDA) with the context of production and implementation of PSS proposals.

Furthermore, future studies can focus on measures to promote the sustainable production of PSS proposals, given the increasing pressure for production methods aimed at achieving the Sustainable Development Goals of the UN Agenda 2030 [116]. For example, sustainable and biodegradable products are being produced with biomass-related inputs and can be produced sustainably by adopting technologically innovative processes aimed at decreasing greenhouse gas emissions [117]. Alternatives are to investigate (energy) technological innovation, such as carbon capture and storage technologies and digital mineral processing [117], to invest in renewable energy and other alternatives in the face of cleaner production [118].

4. Conclusions

This study consolidated the PSS literature from over three decades (1988–2022) by identifying its progress and providing future directions. The identification of future trends after decades of research in the field is essential for the progress of the field and for attracting new researchers. The results of the study aim to lead to a comprehensive understanding of the field and provide insights into the future development of PSS research.

The paper presents a historical evolution of the discussions about PSS, complementing previous literature reviews. To this end, the evolution of the period called the past (1988 to 2015) was presented where five clusters were formed: a business perspective, design perspective, sustainable perspective, operational perspective, and decision perspective. In the present period, five clusters were also identified: digitalization perspective, circular economy perspective, performance perspective, sustainable design perspective, and diffuse perspective. The study confirms that PSS has become a more complex, mature, and sophisticated research area. The results also show an increase in the number of research explored topics and ratify that the theoretical foundations of PSS are multidisciplinary. Specific topics have emerged over time and have become influential in the conceptual framework of PSS. Moreover, traditional topics have evolved with new perspectives, then representing a cumulative depth and rigor as the body of knowledge has advanced.

The findings also reveal the division of the field into five main groups of topics. While this review also revealed different ‘schools of thought’ within PSS research (reflecting different specializations, research traditions, and communities), there is no clear evidence of fragmentation of PSS research. Instead, there is interconnectivity between topics (e.g., business model, capabilities, sustainable design, PSS simulation, value proposition, and others). In the present analysis, Cluster 1 (digitalization perspective) and Cluster 2 (circular economy perspective) acted as “motor themes”, leading research in the field and exemplifying the cumulative tradition of PSS/servitization research. Clusters 3 (performance perspective) and 4 (sustainable design perspective) are up-and-coming themes, moving forward alongside Clusters 1 and 2, using concepts from both digitalization and the circular economy to advance. Finally, Cluster 5 (diffuse perspective) has a transversal role, i.e., its topics maintain relevant links with a set of core topics in other clusters.

The discussions on product-service systems are moving towards a theoretical consolidation, where seminal and traditional topics are being complemented with new research perspectives. The biggest example of this is the incorporation of digital technology into the PSS context since this is a transversal approach that is present in practically all the clusters of the present period analysis (Figure 2). For example, Clusters 3 and 4 of the last period’s analysis (performance perspective and sustainable design perspective) have evolved in the literature through a strong interaction with sustainable development and digital technologies (such as big data, artificial intelligence, digital twin, Internet of Things, and cyber-physical systems). Thus, technologies such as CAD and life-cycle simulations (e.g., LCA, SLCA, and LCC) represent a research perspective on the rise and an alternative in the face of sustainable development and proposals for the circular economy. Although derivative approaches to PSS have emerged over the years, and these do not necessarily present a relationship with sustainability, this subject continues to be a strong strand in the literature, driven by the emerging integration between PSS and circular economy, concepts increasingly fused. The discussion about sustainability, previously more generic (in the analysis of the past period—Figure 1), gains greater discussion in the present period (Figure 2), encompassing areas such as design and production, among others. Traditional topics have been complemented with new perspectives, representing accumulated depth and rigor as the body of knowledge advances.

The paper contributes to the existing literature in three ways. First, the article considers that PSS and servitization are two wide theoretical constructs that arise from the broader movement of strategic incorporation of services in manufacturing firms. Therefore, both constructs are considered from a holistic perspective with their related terms, clarifying their interdisciplinary relationships, whereas previous reviews only provided overviews of the research field on a limited number of topics. Second, the article relies largely on analyses supported by visualization techniques, which are also somewhat missing from past reviews. Third, the article allows for quantifying and visualizing the thematic evolution of the PSS research field in a holistic way. To this end, the article considers a comprehensive dataset of peer-reviewed articles (868 articles) published from 1988 to 2022.

The main limitations of this study resulting from the methodological choices that were made: (i) subjectivity in the assignment of keywords in the context of the co-keyword analysis technique; (ii) subjectivity in merging keywords; (iii) subjectivity due to the selected clustering technique employed; and (iv) some keywords can be associated to more than a single cluster. Nevertheless, these limitations do not significantly affect the results presented. For future research, it would be of interest to complement the developed review by employing other techniques, such as co-citation, bibliographic coupling, and others. In addition, it is suggested to carry out an in-depth content analysis for each cluster presented, develop theoretical framings, and analyze the interdependence of topics.