Innovation portfolio management: a systematic review and research agenda in regards to digital service innovations

Abstract

Portfolio Management (PM) for innovation is as relevant as ever before as many incumbent firms undergo massive transformation in response to digitalization and face the challenge to allocate resources for traditional product innovation and service innovation projects more efficiently and effectively. Digital service innovations, regarded as new business fields for many industrial firms are in the forefront of much discussion in practice and academia; nonetheless, it is unclear to what extent existing innovation PM has acknowledged how to manage a portfolio of service and digital service innovations. To address this gap, this work sets out to (1) review and synthesize decades of contributions in the field of innovation PM in a structured way, (2) examine to what extent research has considered and elaborated on innovation PM for services and digital services, and finally (3) provide a research agenda to foster future contributions in this field. We classified relevant findings in innovation PM into four categories (antecedents, consequences, models/frameworks, challenges) and found that literature has acknowledged services more than anticipated, but that still much of today’s innovation PM research is focused on physical products. In more recent years, the attention towards services has resulted in a few publications delving into the differences between service and product innovation PM; however, digital service innovations have been overlooked by the research so far. Lastly, we point out how innovation PM for services and digital services may diverge from traditional products and outline a research agenda.

1 Introduction

Emergent digital technologies are changing the way companies create value, the way they interact with their customers, and the way they operate from within. This technology-induced change, referred to as digitalization or the digital transformation is described by Fitzgerald et al. (2014, p. 2) as the “use of new digital technologies (social media, mobile, analytics or embedded devices) to enable major business improvements such as enhancing customer experience, streamlining operations or creating new business models”. While certain aspects of new technology introduction replicate transformations of the past (e.g. industrial revolution), the speed and scope of impact suggests that this revolution may impact industries, institutions, and societies in a fundamentally new way. Particularly, Moore’s law (Moore 1998) and the concept of zero marginal costs (Rifkin 2014) are among the fundamental forces behind digitalization as the former projects exponential growth of computing capacity and the latter that digital good production often means zero marginal costs without capacity limitations.

In many instances, the advent of these technologies create new business models in the form of novel services or supplement existing product offerings with digital services (Matzner et al. 2018) such as shared vehicle services (Kessler and Buck 2017) or music streaming services (Turetken et al. 2019). Many manufacturing firms (e.g. Apple, General Electric) have moved beyond the simple sale of products, generating an increasing share of revenues from new service offerings (Demirkan and Spohrer 2016; Matzner et al. 2018). This transition from the traditional good-dominant setting where the emphasis is on the delivery of products (Ostrom et al. 2010) to customer-centric service-dominant business models is regarded as servitization (Vandermerwe and Rada 1988). The concept of servitization first appeared in academia by Levitt (1981) but did not find widespread recognition until the article “Servitization of Business: Adding Value by Adding Service” was published by Vandermerwe and Rada (1988). They claim that clear distinctions between traditional manufacturers and service companies will gradually fade and as servitization progresses (Vandermerwe and Rada 1988). Generally, the reasons for servitization are manifold. For some firms, moving towards services is just a natural progression. For others, especially companies in mature industries, have been said to look towards services to extend their product’s lifecycle and to seize new growth opportunities (Vandermerwe and Rada 1988; Wise and Baumgartner 1999; Oliva and Kallenberg 2003).

Given the parallels, the business phenomenon of servitization goes hand-in-hand with digitalization and has recently emerged in a number of publications (e.g. Kowalkowski et al. 2017; Vendrell-Herrero et al. 2017; Skylar et al. 2019). Skylar et al. (2019) extend a frequently cited definition of servitization “the transformational processes whereby a company shifts from a product-centric to a service-centric business model and logic” (Kowalkowski et al. 2017, p. 7) with the “utilization of digital tools” to define digital servitization.

As the number of digital service offerings rise and firms undergo costly transformations in response to digitalization and servitization, there is an inherent pressure to manage product and service innovations efficiently and effectively. Service innovations have been defined as “a new service or such a renewal of an existing service which is put into practice and which provides benefit to the organization that developed it; the benefit usually derives from the added value that the renewal provides the customers.” (Toivonen and Touminen 2009, p. 893). The prioritization and allocation of resources to different service innovations falls under portfolio management (PM). PM is responsible for “maximizing the value of the portfolio, seeking the right balance of projects, ensuring that the portfolio is strategically aligned”, and selecting the right number of projects (Cooper et al. 2001a, p. 3). While PM of innovation, also referred to as new ‘product’ development (NPD) PM, has been extensively studied, the primary focus lies on physical products. The subject of service PM most prominently appears in information technology research (e.g. Peppard 2003; Queiroz 2009; Kim et al. 2014; Comerio et al. 2015; Schmidt and Freeland 1992). Yet, this stream of research has been developed in isolation to the innovation PM literature as pointed out by Kumar et al. (2008) and Frey and Buxman (2012). It delves into how IT departments manage their portfolio of IT assets (e.g. infrastructure services, help desk support) across the entire lifetime (Leliveld and Jeffrey 2004; Comerio et al. 2015). Service innovations, in contrast, are new services that can be marketed to end-customers (Kumar et al. 2008) and often result in a transformation across multiple dimensions of a business model (e.g. value proposition, revenue streams, customer segments) (Stauss et al. 2010). As such, the degree to which service innovations have been acknowledged and examined in PM thus remains unclear.

Given the widespread assertion that the nature of service and digital services significantly differ from non-digital products (Shapiro and Varian 1998; Menor et al. 2002; William et al. 2008; Nylén and Holmström 2015; Wulf et al. 2017; Vendrell-Herroro et al. 2018), practitioners and scholars alike seek a better understanding of how to manage a portfolio of service/digital service innovations. Leimeister et al. (2014) argue that traditional approaches for designing and delivering products are outdated and incapable for addressing the rapid changes and shorter development cycles in the digital era. Hence, prior to further studies in the realm, we argue that a thorough review of innovation PM literature at this point is necessary in order to gain a more holistic understanding of the state-of-the-art for service and digital service innovations and to lay the foundations for further research in this sphere.

We thereby acknowledge and build upon a previous literature review of innovation PM conducted by Meifort (2015) given its quality and the comprehensive collection of peer-reviewed innovation PM publications until the year 2014. The paper has been cited by many renown scholars in the field (e.g. Kock et al. 2016; Nguyen et al. 2018; Cooper and Sommer 2019) and its research scope to “provide a comprehensive review of innovation PM” replicates ours perfectly. Yet, the paper does not make any inferences to differences in product and service innovation. To our knowledge, our study is the first to investigate and synthesize the innovation PM literature in terms of service and digital services. Further, the innovation PM research field has evolved immensely since Meifort (2015). The field witnessed significantly more quantitative studies, validating or negating initial findings (i.e. optimal portfolio size) and newer research streams such as agility and entrepreneurial culture in PM (Kock and Gemünden 2016). A more recent review of the previous years is also necessary, as services and digital services are a recent phenomenon in project-level and portfolio-level research. In this paper, we thus set out to synthesize theoretical and empirical contributions in the field of innovation PM until end of 2019. Unlike Meifort (2015), which conceptualizes innovation PM based on different lines of thought, we classify innovation PM research in a way that eases the application of knowledge in practice and the identification of the main research streams and findings for academics. Meifort (2015) refrained from synthesizing literature according to common conclusions because of the abundance of context-specific case study and conceptual research. Yet, given the advancements in the field including a magnitude of wide-scale quantitative studies (e.g. Kock and Gemünden 2016), we synthesize findings and offer a new classification. More specifically, we conceptualize innovation PM based on our exploration of (1) which antecedents are central to innovation PM, (2) which frameworks and models were developed in the innovation PM context, (3) what are the consequences of effective and strong innovation PM, and lastly (4) what challenges result through the entire innovation PM process. Lastly, the research agenda will address to what extent innovation PM needs to be further developed to fulfill digital service specific characteristics.

We pose our research objective in the following questions:

• (i) Which antecedents, models/frameworks, consequences, and challenges of innovation PM are discussed in the realm of existing empirical and theoretical research?

• (ii) To what extent does existing research investigate innovation PM of services and digital services?

• (iii) What future fields of research can be identified in terms of service and digital service innovation PM?

The remainder of the paper is structured as follows. Before presenting the selected methodology, the underlying concepts surrounding PM and digital service innovations will be introduced in Sect. 2. The methodology including the review scope and the search process is then outlined in Sect. 3. Next, the main contribution of our study-the analysis and synthesis of the identified literature-follows in Sect. 4. Thereafter, we suggest an agenda for future research in Sect. 5. Section 6 concludes this paper with a summary of our research contributions and an outline of limitations.

2 Fundamental concepts and their interrelations

This section provides a conceptual background to digital service innovation and portfolio management. More specifically, Sect. 2.1 provides historical context to the development of PM in research and differentiates innovation PM to other streams of PM research. Section 2.2 defines digital services and discusses how they distinguish from physical products and non-digital services.

2.1 Portfolio management

PM is simply put as the management of a portfolio of artefacts. The artefacts can be financial assets, products, projects, business units, customers and more. The concept of PM, however, originated in the financial domain through the seminal work of Nobel Laureate Markowitz (1952). He proved that the risk and return dimensions of a diversified investment portfolio are preferable to a homogenous portfolio. The years following, a few authors examined the idea of applying Markowitz’s PM theory to the business domain (Wind 1974; Mahanjan and Wind 1982; Cardozo and Smith 1983; Cardozo and Wind 1985); however, some fellow scholars (Devinney et al. 1985; Bettis and Mahajan 1985) quickly criticized these thoughts. They argue that “the product portfolio problem is of an entirely different sort and requires a theory that makes quite different assumptions about the nature of investments” (e.g. limited to only current investments) (Devinney et al. 1985, p. 108).

Distancing itself from the financial portfolio theory, the concept of selecting and prioritizing items in a portfolio did find application in the business domain and experienced tremendous growth in the 1970s (Henderson 1970; Ansoff and Leontiades 1976; Day 1977). The first stream of PM research in the business domain focused on the development of sophisticated mathematical models for optimizing the selection of businesses from a corporate level perspective. Nevertheless, these studies failed to find wide application in practice as models were too complex and time consuming to use (Schmidt and Freeland 1992; Cooper et al. 1999). In the 1970s, the second stream of corporate level portfolio models were introduced to substitute the former. Among the most well-known portfolio models is the Boston Consulting Group Matrix developed by Bruce Henderson (Nippa et al. 2011). While these portfolio analysis tools found widespread application in practice, the development of the research stream has stalled in academia. Untiedt et al. (2012) identify two possible causes for the scholarly domain: (1) criticism concerning inherent flaws or (2) proof that corporate diversification is inferior to market diversification may have made the research stream obsolete.

More recently, a new stream of PM research focused on managing and optimizing innovation portfolios has developed (Cooper et al. 1999, 2001a, b; Chao et al. 2009; McNally et al. 2009, 2013). The rise of innovation PM is also attributed to the increasing relevance of technological innovations and the acknowledgement that effective selection and management is critical to innovation success (Killen et al. 2007). In times of rapidly advancing technology and changing customer demands, firms increasingly rely on PM to help them navigate where to invest in the future. However, Bharadwaj et al. (2013) point out that it is uncertain to what extent today’s understanding of innovation PM can be applied to the management of a portfolio of digital services and therefore urges researchers to examine the impact of digital business strategy on innovation PM. Prior research has also already pointed out that most innovation PM literature focuses on tangible products, rather than encompassing services or intangible products (Cooper et al. 2000). The established usage of the term new ‘product’ development as synonym to innovation may in part have attributed to misconceptions in the research stream. We question whether the product-focus in innovation PM still remains two decades later, or whether Servitization has triggered greater interest in services, even if in disguise under NPD portfolio management. Our paper plans to address this query in innovation PM as outlined before.

From here on out, we will refer to innovation PM as PM in line with fellow scholars (e.g. Cooper, Edgett, and Kleinschmidt).

2.2 Digital service

Despite the widespread usage of the term digital service, few explicit definitions are found in literature. Beverungen et al. (2017, p. 784) derived a definition for digital services based on Vargo and Lusch’s (2007) definition of services: “A digital service is the application of digital competencies through deeds, processes, and performance for the benefit of another entity or the entity itself”. Thereby, Bevergungen et al. (2017) refer to digital competencies as assets and capabilities, which are made available to other parties utilizing information technology. Williams et al. (2008, p. 506) have proposed a slightly different definition for digital services: “services, which are obtained/or arranged through a digital transaction (information, software models, or consumer goods) over internet protocol”. The two proposed definitions vary as the latter suggests that for a service to be classified as digital, it must be transacted digitally (e.g. digital platforms Amazon or Facebook), whereas the former definition is broader as it circumferences any service that applies digital competencies to provide a benefit (e.g. digital signage in a travel agency). To gather more acumen, we turn to the digital business model literature, as digital services are a type of digital business model. Bärenfänger and Otto (2015, p. 18) classifies business models as digital if the “underlying business logic deliberately acknowledges the characteristics of digitalization and takes advantage of them; both in interaction with customers and business partners, and in its internal operations”.

Digital services co-exist with other offering forms such as physical products, e.g. books, cars, furniture, and non-digital services, e.g. theatre, maintenance, consulting (Vendrell-Herroro et al. 2018). Current trends indicate that physical products are being increasingly replaced by digital services because technology enhances their value respective to products (Parry et al. 2012; Peltoniemi 2015). Drawing upon the servitization and digitalization literature, we argue that the nature of digital services can be distinguished to other offering forms-physical products and/or non-digital services-in terms of several unique traits of digital services. The first differential characteristic refers to the concept of zero marginal costs. Digital services can be reproduced infinite times at practically zero marginal cost (e.g. smartphone apps) (Ekeldo and Sivakumar 2004; Remane et al. 2017; Vendrell-Herroro et al. 2018). Second, while for physical products the value is created within a firm and then sold to a customer, for digital services, similarly to non-digital services, the value is realized in the process of consumption (Vargo and Lusch 2004, 2008; Vendrell-Herroro et al. 2018). This captures upon the idea of “value in use”. Remane et al. (2017) provides an example of a smartphone, which provides little value to users without its interface to access services (e.g. navigation, mobile payment). Third, traditionally, the provision of services and manufacturing of physical products largely remained within a firm’s boundaries. Digital services, on the other hand, are often a result of co-creation. They rely on an ecosystem of partners, competitors and users (e.g. Apple’s Appstore) (Iansiti and Levien 2004; Eaton et al. 2015; Akram 2016; Remane et al. 2017). Fourth, digital services are non-excludable (Barwise and Picard 2015; Vendrell-Herroro et al. 2018), meaning that they can be shared endlessly without reducing the availability of service to others (Vendrell-Herroro et al. 2018). Even more so, many digital services (e.g. Facebook, Airbnb) became exponentially more valuable the more users joined (Shapiro and Varian 1999; Remane et al. 2017). This phenomenon is referred to as network effects. Lastly, digital services have shorter lifecycles and operate in more dynamic environments. As indicated by Moore’s law, technologies constantly evolve, generating higher processing capabilities at lower costs. Updates also occur much more frequently as they can be immediately deployed. Yet, these characteristics accumulate to low entry barriers, enabling new competitors to enter the market and reproduce offerings more easily (Menor et al. 2002).

For the sake of this paper, we derive our own definition for digital services based on the ideas discussed above. We define digital services as integrated solutions that deliver other parties value in use by utilizing digital technologies for at least a portion of the interaction. Uber’s ride hailing service, for example, is classified as a digital service, as it provides riders value in use by transporting them from point A to B and utilizes digital technologies via transactions among drivers and riders on a digital platform (i.e. Uber’s application). In relation to innovation, the introduction of digital services is often inherently classified as service innovations given their newness to the firm and impact on multiple dimensions of their business model. The overwhelming majority of scholars are of the notion that service innovations must only be ‘new to a firm’, but not necessarily also ‘new to the industry or market’ (Snyder et al. 2016). The following section, the methodology, outlines the scope of our review and literature search process in detail.

3 Methodology

This study shall use a systematic literature review following the guidelines of Webster and Watson (2002) to identify peer-reviewed articles that have contributed to the PM research, and hence evaluate and discuss the application to services and digital service innovations. The scope of the review will be condensed based on the taxonomy of literature reviews suggested by Cooper (1988) in Sect. 3.1. Next, the search process will be outlined in detail in Sect. 3.2.

3.1 Review scope

In order to specify the research aim of this paper, we outline the scope in accordance to a taxonomy for literature reviews presented by Cooper (1988) and more recently highlighted by vom Brocke et al. (2019). The proposed framework comprises six characteristics, each consisting of further categories, which present the focus areas of this paper (see Fig. 1). The categories displayed gray in the figure present the main focus points of this paper.

Fig. 1

Taxonomy of literature reviews (following Cooper 1988, p.109)

The main focus (1) of our contribution lies in examining both research outcomes and theories regarding PM in the innovation domain in order to draw deductions for the management of a portfolio of services/digital service innovations. The overarching goal (2) of the literature review is thereby to first integrate the existing body of knowledge and embody a slightly critical stance of whether the existing contributions remain valid in the context of digital services. Further, the organization (3) of this paper will be concept-centric. The perspective (4) of this paper aims to remain as neutral as possible, although certain degree of subjectivism in the representation of results is nearly inevitable (vom Brocke et al. 2019). As aforementioned, the targeted audience (5) for this paper are scholars in the intersection of PM, servitization, and digital business research communities, as well as practitioners, challenged by the new role of digital services in their organizations. Lastly, the coverage (6) can be regarded exhaustive but selective, as we try to gain a comprehensive overview of all relevant research, but only detail a selection of the contributions in our paper. After providing clarification of our research scope, the next section will outline our literature search process in detail.

3.2 Literature search process

Our literature search process is structured as follows: (1) selection of databases, (2) identification of search terms, (3) definition of inclusion and exclusion criteria and (4) database search and refinement, and lastly (5) backward and forward search. It shall ensure the reliability and reproducibility of our review. Additionally, we will compare and contrast our methodology with that of Meifort (2015) as we plan to incorporate the papers in her literature review to address the second research question. Selection of databases. This paper has selected the following three bibliographical databases to facilitate the most exhaustive and reliable collection of publications: Scopus, EBSCO Business Source Complete, and Web of Science. Firstly, Scopus was selected because it is the largest citation database of peer-reviewed literature-scientific journals, books and conference proceedings-with over 23,500 journals that are rigorously vetted by an independent review board (Scopus 2019; IJRVET 2019). As a second source, EBSCO Business Source Complete was chosen as it is regarded as a world-leading scholarly source for research in the business and economic context, providing access to over 3700 peer-reviewed academic journals (EBSCO 2019). Due to the scope of this paper, the business focus of this database deemed appropriate. Lastly, Web of Science was selected to supplement the previous two databases, covering nearly 21,100 journals in a wide array of disciplines (Web of Science Group 2019). Google Scholar was explicitly eliminated as a primary database due to its usability and scope (Bosman et al. 2006; Herrera 2011). While it offers a large variety of search output, many publications are working papers or other non-peer-reviewed articles (Bosman et al. 2006; Bar-Ilan 2008). Meifort (2015), in contrast, only utilized one database-the EBSCO Source Premier-to extract an initial set of relevant papers.

Identification of search terms. Prior to identifying any keywords for the search process, Baker (2000) suggests to consult with publications that provide a thorough overview of the main concepts planned to discuss in the paper in order to uncover relevant search terms (i.e. synonyms and homonyms). In accordance to his recommendation, we identified relevant keywords and further validated these in an iterative process of testing search expressions in different databases to evaluate the resulting collection of publications. We found that there was an inconsistent use of terminology in titles as well as within the text itself for innovation PM (e.g. Hunt et al. 2008; Baptestone and Rabechini 2018; Cooper et al. 2000). Next to innovation PM, the variation included PM for NPD, project PM and product PM. Against this background, we concluded that simplifying the keyword search to one term-portfolio management-in the abstract or title of the article was broad enough to avoid excluding potentially relevant publications from the onset and specific enough as the term is well-established in research. Meifort (2015) also conducts a keyword search in the abstract and titles of the articles, yet unfortunately does not specify which keywords were applied in her search. Definition of inclusion and exclusion criteria. To further specify and condense our search output, we applied a set of inclusion and exclusion criteria, as shown in the table below. The inclusion criteria will be applied from the onset in the search string, whereas many of the exclusion criteria will be utilized later to further refine the initial search output. For the same reasoning as noted above, this process was chosen to forgo eliminating relevant publications ex-ante. Next to the main keyword “portfolio management”, we supplemented the search string with the inclusion criteria. For example, in Scopus the inclusion criteria were depicted as follows: (LIMIT-TO (SRCTYPE, “j”) AND (LIMIT-TO (PUBYEAR, 2014–2019) AND (LIMIT-TO (LANGUAGE “German”, “English”) AND (LIMIT-TO(SUBJAREA, “BUSI”) AND (EXCLUDE (KEYWORD “finance”). To secure the quality of findings, we applied the inclusion criteria peer-reviewed journal following Webster and Watson (2002). Dissertations, books, book reviews, and working papers were excluded. The timeframe from 2014 until 2019 was chosen based on the previous literature review by Meifort (2015), which reviewed publications from 1964 until November 2014. The keyword “finance” was excluded from the onset as the sample otherwise consists largely of investment or asset portfolio management publications. In terms of exclusion criteria, we disregard all papers that stem from different subject areas (e.g. economics, arts and humanities) and were not available in full-text. Further, for quality reasons, journals were excluded that were not listed or did not steam from high-quality sources according to Scimago Journal Rank (www.scimagojr.com). The inclusion and exclusion criteria applied are summarized in the Table 1.

Table 1 Inclusion and exclusion criteria

In comparison, Meifort (2015, pp. 2–3) restricted the results of her keyword search to 20 journals “considered as leading general management journals and leading in the particular field of technology and innovation management” (e.g. Journal of Product Innovation, Management Science, R&D Management, Research Policy, Strategic Management Journal), resulting in 65 publications.

Database search and refinements. Our initial keyword search conducted on October 7–10, 2019 in all of the selected databases yielded 993 articles (see Table 2). Next, we read the titles and abstracts of our initial search output and eliminated publications that did not meet our research scope. This includes remaining papers that do not stem from the business and management field. Further, many eliminated articles thematized financial/asset PM or brand PM. This resulted in the number of publications being reduced to 213. Next, we scrutinized the full texts of the remaining publications and further eliminated publications that proved irrelevant to our research scope. The reasons for elimination varied extensively including an overly strong emphasis on a specific country and/or industry (e.g. electric power utilities in Brazil) or lack of portfolio level view in project management publications. In this step, we also applied the quality criteria by consulting Scimago Journal Rank as stated above. From the resulting 115 papers, we eliminated 75 publications due to duplicates that appeared in multiple databases or already appeared in the literature sample of Meifort (2015). In regards to the review strategy of Meifort (2015), the papers did not undergo further refinement aside from the journal restrictions.

Backward and forward search. After three rounds of refinements, a set of 40 publications remained. Nonetheless, we decided to follow the recommendations of Webster and Watson (2002) and additionally conduct a forward and backward search. The backward search analyzes the references of our set of relevant articles in order to identify further publications that were missed in the database search. The forward search, on the other hand, identifies further articles that cited the set relevant articles in this study. For the forward search, we utilized Google Scholar, which offers a tool to identify publications that have been cited in other publications. In line with our review strategy, Meifort (2015) also conducted a backward search to uncover 47 additional publications, yet did not pursue a forward search.

Table 2 Literature search process

Our final sample consists of 52 publications (see Table 2). In the next section, we discuss and synthesize the findings of our literature review as well as integrate the previous findings of Meifort (2015). Despite the use of different review methods (see Table 3 for summary), the results show (see Sect. 4.1) that the journals largely replicated those in the sample of Meifort (2015) with the International Journal of Project Management also being the journal with the most selected publications. This hence provides further basis for the inclusion of Meifort’s (2015).

Table 3 Research methodology in comparison to Meifort (2015)

4 Literature analysis and synthesis

This section provides a descriptive overview of the literature research results, followed by a qualitative content analysis in line with (Mayring 2014). Mayring (2014) proposes a standardized method for qualitative content analysis that supports the evaluation and classification of large amounts of text. In our paper, we have selected a specific inductive technique of content analysis referred to as summary (Mayring 2014, p. 65ff) in order to reduce the content to its core. We utilized excel to analyze and store information extracted from the databases (e.g. author, title, journal), and supplemented it with further descriptive elements (e.g. methodology, sample) and the key findings for each article.

4.1 Descriptive analysis

As indicated in Fig. 2, the publications of the review stem from a wide variety of academic journals. The most common journals include the International Journal of Project Management (12), the Project Management Journal (5), IEEE Transactions on Engineering Management (4), Journal of Business Research (4), and Journal of Product Innovation Management (4). The wide spread of journals can be explained by the nature of the topic as well by the magnitude of journals in the scope of project and innovation management.

Fig. 2

Distribution of literature across journals

Fig. 3

Distribution of literature across years and research approaches

Turning to the number of publications, our review has identified 52 relevant publications in a timespan of only 5 years. This shows that the field has experienced tremendous growth in the last years. The number of publications reached a peak in 2015 with 14 publications, whereas the last two years (2017, 2018) witnessed only half as many (see Fig. 3). This trend may indicate increasing saturation and a maturing state of the art. This is in part also reflected in the research approach. The literature samples indicate a high share of both qualitative and quantitative empirical research. In comparison to the sample of Meifort (2015), ranging from 1969 until 2014, the number of conceptual papers was still much higher.

4.2 Content-based analysis

In the following, the main findings in regards to (1) antecedents for PM, (2) PM frameworks and models, (3) consequences of PM, and (4) challenges of PM are presented and discussed. Given the amount of quantitative studies in recent years, we chose to classify and synthesize the common findings in innovation PM rather than view the publications from different theoretical lenses as in the previous literature review by Meifort (2015). As commonly proposed in literature reviews (e.g. Raisch and Birkinshaw 2008), we chose to categorize findings into “antecedents” and “consequences”. The category “PM framework and models” was supplemented as we found many publications that offered frameworks with the intention to support portfolio managers in different processes, yet did not provide direct evidence for improving portfolio success and as such do not fall into the “antecedent” category. These publications are most commonly in the form of case studies and concord largely with Meifort’s (2015) “decision-making perspective”. Lastly, “challenges” was selected as an individual category as they occur throughout all stages and dimensions of innovation PM and offer central opportunities for improvements for practitioners and academics alike. The findings are summarized in the Tables 4, 5, 6, 7, 8, 9, 10, 11.

Table 4 Organizational and managerial antecedents for PM

Table 5 Portfolio content and structure antecedents for PM

Table 6 External environmental antecedents for PM

Table 7 Procedural aspects

Table 8 Methods, tools and metrics

Table 9 Models and frameworks in PM

Table 10 Consequences of PM

Table 11 Challenges of PM

4.2.1 Antecedents for PM

Antecedents for effective and successful PM have been extensively investigated by the selected literature and categorized into five dimensions in this paper: (1) organization & management, (2) portfolio content & structure, (3) external environment, (4) processes and (5) models, tools, and metrics. The dimensions were developed in retrospect to reviewing the publications and their main findings and grouped based on commonalities.

Organizational and managerial aspects. Organizational and managerial aspects in regards to antecedents refer to the characteristics and attributes of an organization (e.g. structures, culture, capabilities and resources) and management (e.g. hierarchy, management involvement). First, managerial aspects were found to cover primarily two stakeholders-the top management and the project managers. In regards to the former, a high transparency and involvement of top management supported portfolio success (Too and Weaver 2014; Meifort 2015; Patanakul 2015; Yang and Xu 2017). If the top management is highly involved in PM, it is more likely that the firm will conduct regular reviews to align the portfolio with the firm’s strategy (Yang and Xu 2017; Cooper et al. 1999). Project visibility of the status and potential problems enables management to react quicker and lead necessary measures in the way (Patanakul 2015). Padovani and Carvalho (2016) further found that centralization in approval processes increases the effectiveness of PM and hence portfolio performance. In regards to project managers, the review found that in contrast to Carbonell and Rodríguez-Escudero (2016) that decentralization, giving greater authority at project manager level, increases the portfolio success until a certain threshold is reached (inverted-U relationship). From there on out, further decentralization, decreases portfolio success. The latter research indicates that a balanced authority among top management and project managers leads to the best results. Petro and Gardiner (2015) also point out that project managers with strong authority and responsibility in the organization lead to better results in terms of portfolio success. The relationship was further positively moderated by involving a steering committee (Petro and Gardiner 2015). Generally, good project managers have strong capabilities, knowledge, and take responsibility in terms of PM (Yang and Xu 2017). Further, portfolio success heightened when project managers incentives were aligned with the portfolio’s objectives. Another quite central organizational aspect are the multiple levels (e.g. corporate, business-unit, portfolio, project, and technology level, etc.) involved in PM as addressed in Meifort (2015). PM in a large part is about managing these multi-organizational levels and achieving consensus about the portfolio’s future direction. Cultural aspects of the firm such as extensive communication, organizational learning, innovation power, and adverseness to risk played a key role in portfolio success throughout a series of research papers (Meifort 2015; Kock and Georg Gemünden 2016; Kock et al. 2016). A strong innovation climate is characterized through managerial support and autonomy to pursue creative ideas, whereas risk culture is demonstrated by venturing and committing resources to projects uncertain of the outcome or environment (Kock and Gemünden 2016; Kock et al. 2016). Nonetheless, the key towards quality decision-making is that measures for containing risk are implemented and risk is openly communicated (Kock and Gemünden 2016). Lastly, in regards to internal capabilities and resources, the research indicated that information availability belongs to the key drivers to PM performance. PM effectiveness and success are further dependent on (1) maintaining an overview of the firm’s constraints, capabilities, uncertainties, and interdependencies (Padovani and Carvalho 2016) as well as (2) the ability to adapt quickly to these type of internal/external changes (Meifort 2015; Patanakul 2015). To conclude, our review demonstrates that the maturity of organizational and managerial antecedents is already quite advanced.

Portfolio Content & Structure. Next to organizational and managerial aspects, antecedents in terms of the portfolio itself are discussed next. Selecting the right number of projects is part of Cooper et al. (2002) objectives for portfolio management. Yet, there is academic discourse to what the right number of projects is. Is the wider breadth of resource allocation, consequently larger portfolios, more successful than smaller, more focused portfolios? Kang and Montoya (2014) and Klingelbiel and Rammer (2014) agree that larger portfolios (or greater resource allocation breadth respectively) are more successful in terms of financial and portfolio performance. They argue that the greater the number of projects, the greater the chance of one of the projects becoming successful. Kock et al. (2016), Kopmann et al. (2015) and Teller et al. (2012), on the other hand, find that an increasing number of projects also inflicts more complexity. Also in line with Cooper, further research found that portfolios should be balanced (e.g. different levels of risk and innovation) and strategically aligned with the firm’s strategy (Kang and Montoya 2014; Too and Weaver 2014; Meifort 2015; Patanakul 2015; Padovani and Carvalho 2016). Moreover, Kopmann et al. (2015) analyzed the effect of strategy formulation on portfolio success. They found that both the implementation of (1) deliberate strategies, strategy that cascades from corporate strategy to project level, as well as (2) emerging strategies, strategy that emerges from PM conditions, significantly improves portfolio success. The implementation of both further enhances this relationship. Further, the strategic decision to terminate projects that are no longer in line with a firm’s strategy or no longer contribute to business performance is also regarded central to managing successful portfolios (Too and Weaver 2014; Yang and Xu 2017). In terms of portfolio interdependencies, Kock et al. (2016) and Kopmann et al. (2015) found that just as with portfolio size, the greater the interdependency among projects, the greater the complexity. Jugend et al. (2016), however, suggested that greater integration of projects may even improve decision-making when knowledge sharing takes place. Korhonen et al. (2016) also suggested that component commonality can act as a source of innovation in the beginning-trying to optimize product design. In later stages, when subsequent projects use already established design, it may hinder innovation. In regards to the type of projects in a portfolio, Kang and Montoya (2014) advocate to focus more on first-to-market projects as they result in the best financial results in the short and long-term, and Jugend et al. (2017) suggest to adopt green products for stronger portfolio performance. We find that the type of projects is still in early development in terms of its impact on portfolio success. Further research could for example investigate the influence of cooperation projects or projects with shorter versus longer development cycles on portfolio success.

External environmental aspects. Research regarding the influence of the external environment on PM proved to be limited and where mostly considered only as moderators for PM success. The external environments under investigation were defined as turbulent with frequent technological and customer-induced changes. In environments with high turbulence, the value of business controlling becomes more imperative, whereas a deliberate strategy and process formalization becomes less important as a firm needs to constantly adapt its strategy to the changing environment (Kopmann et al. 2015; Kock and Georg Gemünden 2016a). Moreover, Kock and Georg Gemünden (2016) also found that in the context of turbulent environments, a firm’s innovation climate becomes more impactful to making quality decisions and agility in PM. They reason that under highly turbulent environments, the motivation to innovate becomes stronger among project managers. The external environment is still in an early stage of development and further research could expand along the lines of Floricel and Ibanescu (2008) and examine the direct influence of external factors on portfolio success rather than being utilized as a moderator.

Procedural aspects. In this paper, antecedents of processes refer to best practices in terms of ways of doing, managing, coordinating innovation portfolios. Padovani and Carvahlo (2016) identify eleven procedural tasks that are relevant to PM and in turn positively influence a portfolio’s performance (see Table 7). Yang and Xu (2017), Jugend et al. (2016), Kock and Georg Gemünden (2016), and Meifort (2015) explored the formalization of PM processes and found that greater formalization-clear rules, criteria and procedures-lead to higher portfolio decision-making quality and performance. Aas et al. (2017), who explore the distinction between service and product portfolios, find that PM for services, however, requires greater flexibility to account for high degree of heterogeneity in service portfolios. Stettina and Hörz (2015) also suggest implementing agility into PM. Regardless of formalization or agility, controlling and reviews are critical to both portfolio management approaches. The controlling intensity of PM, demonstrated through continuous monitoring and reporting, displays a positive relationship with decision-making and portfolio success as it enables a firm/portfolio manager to respond quicker to changes (Kopmann et al. 2015; Meifort 2015; Kock and Gemünden 2016; Kopmann et al. 2015). The review demonstrates that procedural aspects, especially formalization and review have been extensively studied. Yet, agile portfolio management and question of how firms can best implement agility in their processes is still in an early development stage and calls for future investigation (Tables 8, 9, 10, 11).

Methods, tools & metrics. This section covers methods, tools and metrics that have yielded to greater decision-making and portfolio performance. In regards to methods, the research displays mixed results. Jugend et al. (2016), Killen (2017), and Meifort (2015) find that greater usage of methods such as scoring models, checklists, or network mapping supports formalization or decision-making quality, respectively, and hence improves portfolio performance. Spieth and Lerch (2014), on the other hand, examined the direct effects of method usage on PM performance and found no significant effect. Yet, they also found certain methods (e.g. strategic buckets, team decision making, checklists) to have significant effects on management perception and thereby indirectly foster PM performance (Spieth and Lerch 2014). De Villiers et al. (2016) tested the efficacy of a range of tools (e.g. BGC Matrix, devil’s advocate) and found that it leads to better product portfolio decision making only in certain contexts, and that over reliance can lead to decision incompetence. Turning to metrics and selection criteria, Kaiser et al. (2015) outline that metrics should be derived from the firm’s strategy and access to high quality information should be secured for achieving PM success. Further, Patanakul (2015), Baptestone and Rabechini (2018), Relich and Pawlewski (2017) and Behrens (2016) explore which type of selection criteria firms should consider for effective PM. Patanakul (2015) found that firms should emphasize more on the expected value (financial and non-financial) of projects in portfolio selection. Baptestone and Rabechini (2018) found strategic fit as well as profitability and return on investment are the most central selection criteria. Lastly, Behrens (2016) finds that managers find four selection attributes most relevant: (1) financial returns, (2) demand uncertainty, (3) competitive advantage, and (4) risk. In terms of services again, Aas et al. (2017) find that the role of non-financial criteria is much higher than for products. To conclude, despite extensive research and development of many different tools and methods, we still find discrepancies among scholars regarding their influence on portfolio performance. Metrics, for portfolio selection and performance measurement, should also be further developed to determine which key performance indicators forecast portfolio success best.

4.2.2 PM models and frameworks

Next to the aforementioned antecedents for portfolio success, the research sample consisted of a number of models and frameworks that were developed to support managers in (1) the portfolio selection process, (2) the continuous portfolio optimization, and (3) the implementation of sustainability. Models and frameworks are widely addressed in PM research and as such a review of all would push the boundaries of this paper. In this section, we thus supplement key contributions since 2014 (see Meifort (2015) for prior publications). In contrast to similar aspects (i.e. tools and decision criteria) discussed in the antecedents, research in this section does not provide direct evidence for improving decision-making or portfolio success. The papers in this category are primarily conceptual by nature and illustrated through the use of case studies. They are also more specific; detailing individual components and the usage of the proposed frameworks. Findings in the antecedents are broader and, for example, delve into whether certain criteria (i.e. non-financial criteria) or the greater usage of tools supports the performance of PM.

To support managers in the portfolio selection process, Lin and Yang (2015) and Relich and Pawlewski (2017) have developed ‘fuzzy’ weighted portfolio selection models (based on estimates/imprecise data) that help rank projects and estimate portfolio performance. Sampath et al. (2015) combine different modeling, simulation, and optimization techniques for portfolio selection, illustrated in a case study of Intel Corporation. Wang and Thiele (2017) propose two mathematical decision-making models for R&D projects that take the large upside risk of radical innovations into account. Baker and Bourne (2014) extend NPD stage gate processes by proposing a governance framework-feedforward anticipatory control (FAC)-for improving decision-making in complex product portfolios with short life-cycles. The FAC is derived from a combination of feedforward controls such as planning and forecasting for go/kill decisions (Cooper 1990) with management’s validation of targets. Based on case studies in the sports and apparel industry, Baker and Bourne (2014) demonstrate that FAC help balance control and creativity in the portfolio as well as drive productivity and profit growth. Lastly, Santiago and Soares (2018) propose a framework that helps decision-makers in the design of strategic buckets. The next set of models and frameworks concern the continuous optimization of portfolios. Dash et al. (2018) propose a framework for product optimization and scheduling of activities across various lifecycle stages. Marcondes et al. (2018) propose a framework that goes beyond the realm of portfolio selection and specifies how to adjust existing portfolios, and Abrantes and Figueiredo (2015) provide a framework for optimizing the allocation of human resources across projects in dynamic environments. Lastly, Brook, and Pagnanelli (2014) provide managers with a step-by-step framework of how best to incorporate sustainability practices (e.g. ecological, social, and economic) into the PM process.

4.2.3 Consequences of PM

The consequences of PM are referred to as the output resulting from effective PM. The research in our literature sample, indicates that successful PM leads to (1) greater innovation performance (Spieth and Lerch 2014; Meifort 2015; Sicotte et al. 2015; Yang and Xu 2017), (2) innovation speed (Meifort 2015), (3) greater firm innovativeness (Spieth and Lerch 2014; Tidd and Thuriaux-Alemán 2016), and (4) better firm performance (Spieth and Lerch 2014; Meifort 2015; Hermano and Martín-Cruz 2016). The items utilized to explore these relationships varied slightly; however, the number of studies exploring the relationship of PM with innovation and firm performance within the relatively short period is quite large and thus solidifies the importance of PM.

4.2.4 Challenges of PM

The challenges of PM are much less explored than the antecedents for PM. Challenges are, however, often inherently related to antecedents. For example, while having a balanced portfolio is an antecedent for successful portfolios, this objective is often times quite challenging for firms to fulfill because they tend to overly focus on incremental innovation (less risk, less effort), and consequently focus too little on long-term, radical innovation (Jugend et al. 2016). However, the most commonly identified challenge among the publications occurs during the decision-making process. Gutiérrez (2014) identifies ambiguity in the decision-making, which results in portfolio managers to allow ideas to be further developed until they grasp a better understanding of the purpose/benefits and can construct better judgements. This prolonged decision, however, requires additional resources that could be assigned to another project. Meyer (2014) identifies a further challenge in decision-making-optimism bias. This occurs when managers believe that the predicted cost or time of a project can be reduced through their own managerial efforts. Bias in decision-making was also identified by Lechler and Thomas (2015), who examined challenges related to termination decisions. They argue that managers often times display bias towards personal goals (or so-called “pet projects”) and thereby potentially sacrificing organizational goals. In Meifort (2015), Loch and Bode-Greul (2001) also specify the challenges involved with resource allocation decisions-lack of consistent and reliable information and continuous adaption required. Lastly, optimization models were formerly dismantled because they were unable to grasp the complexities of PM e.g. project interdependencies (Meifort 2015). With today’s technology such complexities can be incorporated, yet these models still have strong information requirements and decrease transparency for project managers (Loch and Pich 2001 cited in Meifort 2015).

To conclude this section of literature synthesis, our classification of PM research is summarized and visualized in the Fig. 4.

Fig. 4

Classification framework of PM research

4.3 Exploration of services and digital services in current innovation PM research

In this section, we address the second research question and explore to what extent the current PM research has acknowledged services and digital services. We cross-examined the papers in our sample [52 papers] as well as the totality of papers presented in Meifort (2015) [112 papers] to create a complete picture. We classified empirical papers based on their research sample, and conceptual papers based on the degree to which the paper acknowledged and thematized services. While some authors consider new product development to consist of both products and services, many studies, especially early PM papers, have a strong focus on R&D in manufacturing intensive industries and as such their research is not applicable to services. The findings of the papers categorized as ‘universal’ are applicable to both services and products, whereas the findings of product-focused papers do not necessarily remain valid for services. Papers that did not specify their scope sufficiently (e.g. whether NPD includes services), and/or the research sample provided no further indication, were classified as ‘not specified enough’ (Fig. 5).

Fig. 5

Distribution of PM literature with product and/or service focus

The findings show that the majority of papers are either ‘universal’ [61] or ‘product-specific’ [67]. Only one paper was classified as ‘service-specific’ (Aas et al. 2017) and one more as ‘digital-specific’ (Echterfeld and Gausmeier 2018). The remaining papers [34] are not specified enough to classify, but are likely to fall into the product-specific category, if detailed more.

The number of ‘universal’ papers is surprising given the assumption that services have been widely ignored by innovation PM literature up to date. Yet, we found that many of the ‘universal’ papers were categorized based on whether their sample included services (e.g. financial services, consultancy) and did not necessarily extensively address services. Still, we present and briefly discuss these publications as they offer valuable insights in regards to what extent services have been acknowledged in innovation PM. The ‘universal’ papers that explored both product and service innovation PM are mostly found in the research stream antecedents for PM. The findings in terms of (1) organization and management (e.g. top management involvement, risk culture) and the (3) external environment are particularly well explored with only few exceptions (Kraiczy et al. 2014). In terms of (2) portfolio content and structure, “balanced, strategically aligned, and interdependent” service portfolios are found to have a positive relationship to portfolio performance (Spieth and Lerch 2014; Too and Weaver 2014; Kopmann et al. 2015; Patankul 2015; Kock et al. 2016; Padovani and Carvalho 2016). Further, likewise to products, services require further research to resolve contrary findings in regards to portfolio size. The remaining antecedents (4) PM processes and (5) methods, tools and metrics are far less explored in terms of services. In a case study including the examination of service firms (i.e. insurance, telecommunication services), Patanakul (2015) finds that greater attention to the expected value both financially and non-financially associates with greater PM effectiveness. Kasier et al. (2015), who included construction service firms in their sample found that metrics derived from the firm’s strategy and high quality information associate with greater PM success. Unlike solely product-focused research (Jugend et al. 2016; Killen 2017), Spieth and Lerch (2014), who included services in their definition of innovation projects investigated the method usage (e.g. strategic buckets, checklists, etc.) found no significant impact on IPM performance.

Next to antecedents, service providers are also widely considered in research regarding PM consequences. The literature shows that successful PM in services leads to greater innovation performance, greater firm innovativeness, and better firm performance (e.g. Spieth and Lerch 2014; Sicotte et al. 2015; Tidd and Thuriaux-Alemán 2016). Yet, innovation speed in services remains unexplored. Lastly, the categories ‘models and frameworks’ and ‘challenges’ are uniformly product-focused and as such reveal no insights in regards to services.

In Meifort (2015) and our literature review, four papers explored the distinction between services and products in detail (Killen et al. 2008; Koh and Crawford 2012; Sicotte et al. 2015; Aas et al. 2017). In retrospect, we identified three additional papers (Storey and Hughes 2013, Calantone et al. 2010; Killen and Hunt 2010) that display a high relevancy for our paper but were disregarded in Meifort (2015). Their findings will also be discussed here. Killen et al. (2008) found that PM practices among services and products are similar to a large extent, but have specific challenges. Service-firms operate in environments with dynamic markets and technologies, shorter lifecycles, and greater ease of service imitation. Manufacturing-intensive firms, on the other hand, are less flexible in dynamic competitive environments and are increasingly challenged to move towards greater service integration (Killen and Hunt 2010). Further differences were found: services have shorter development and business planning processes, require greater flexibility and collaborative approaches to account for the high degree of heterogeneity in service portfolios, and focus on more non-financial criteria and less on specific tools (Killen et al. 2008; Calatone et al. 2010; Koh and Crawford 2012; Storey and Hughes 2013; Aas et al. 2017). Similar to products, Storey and Hughes (2013) found that superior performing service firms do not spread their resources across too many projects, too thinly. They also ‘kill’ projects that are no longer relevant, strive for a well-balanced portfolio (not too risk adverse) and conduct periodic reviews (monthly or quarterly) so that it does not become too late to correct potential problems (Storey and Hughes 2013).

Although the concept of PM was applied to services much later, Killen et al. (2008) found that the use of portfolio methods among service organizations are just as mature (measured in terms of formality, decision-making, performance measurement, and methods used) as for product organizations. Storey and Hughes (2013) in contrast finds that only 8% of service firms (financial services) have well-executed PM, in comparison to 21% found in studies of product-centered firms (Cooper et al. 2001b). Storey and Hughes (2013) states that service services do not yet have PM in their strategic focus. Methods originally developed for products were utilized for services. Yet, the findings that service firms invest a higher percentage of their revenues in the development of services, but product-focused firms report significantly more profit from new products, may also suggest that PM processes are not yet optimized for services.

5 Research agenda

Our literature review demonstrates that the PM of services and digital services are under-researched. While there has been research regarding service innovation (e.g. de Brentani 1991; Kelly and Storey 2000; Storey and Kelly 2001; Storey and Hughes 2013; Storey et al. 2016; den Hertog et al. 2010) and digital service innovation (e.g. Eaton et al. 2015; Chowdhury 2015; Nylén and Holmström 2015; Akram 2016) on project level, the research on portfolio level remains narrow for services and non-existent for digital services. Given the growing significance of services, in particularly of digital services, to developed economies and the inherently different nature of digital services, we argue that there is a high relevance to examine how the design and execution of PM should optimally look like. In the following, we outline potential areas for future research in the context of services/digital services.

• 1. In product PM, research has found that PM processes with greater formalization in terms of clear rules, criteria and procedures improve decision-making quality and portfolio performance. A case study paper from Aas et al. (2017) suggests that service portfolio processes are often more flexible but they do not investigate the influence of formalization or agility on performance. Given that digital services compete in much more dynamic environments with shorter lifecycles, formal processes may need to shorten and allow for greater flexibility. We urge researchers to investigate the role of formalization and agility on the performance of services and digital services.

• 2. Similar to the previous research suggestion, the utilization of portfolio methods and tools for service portfolios is also an interesting and under-researched area. The literature sample indicates that greater usage of tools and methods yields higher performance for product portfolios. We suggest investigating the utilization on service/digital service portfolio performance as well. Along these lines, it would also be essential to know which tools or methods work in the context of services best and to what extent established tools may need to be adapted. Digital services inherently offer new possibilities of collecting and analyzing data, thus potentially improving the accuracy and usefulness of tools applied in portfolio selection.

• 3. New selection criteria and performance indicators are needed as most have originated on the basis of product-orientated business models (Baines et al. 2006, 2009; Martinez et al. 2010). In the context of digitalization, traditional metrics are no longer fundamentally effective. This is because a large majority of assets in the digital world are intangible (e.g. user experience, algorithms) and thus more difficult to identify and quantify (Bughin and Manyika 2013). Digital traction metrics (e.g. daily active users, engagement rates) and metrics capturing the financial impact of acquiring, retaining and monetizing customers become critical (Fleming 2015) and as such should also be considered when allocating resources to different digital services and measuring their performance.

• 4. Kopmann et al. (2015), Korhonen et al. (2016), and Jugend et al. (2016) investigate the relationship between product portfolio interdependencies and performance, yet further research is required for a coherent stance. We reckon it would be interesting to examine this relationship also for services because cost savings from synergies can be immense, especially for digital services, where back-ends can be more easily standardized.

• 5. The next potential area of research covers models and frameworks for PM. The existing research is mostly of conceptual nature (illustrated through case studies) and developed under product-intensive contexts. Further research could illustrate and validate these frameworks in the context of services/digital services as well as develop models and frameworks that consider the unique nature of services/digital services. Such a contribution would be highly relevant for practitioners in service fields.

• 6. In our literature review, we identified challenges across the dimensions decision-making, termination, and the use of optimization models and found that these papers were also developed with a strong product-centricity. Future research could investigate not only further service-specific challenges but also examine to what extent these challenges apply to services. For example, investigating termination decisions for digital services would be interesting, as we expect digital service-specific factors (e.g. shorter lifecycles, co-creation with partners, marginal-zero costs, etc.) to influence the decision-making process significantly. If the cost of offering an additional unit does not result in nearly any costs, the decision to terminate a digital service may be deferred. Also, digital services may alleviate the challenges in optimization models as they inherently capture data and as such can deliver sufficient information for complex models.

• 7. As addressed in the introduction, the subject of service PM has mostly appeared in the information technology domain and has been developed in isolation to innovation PM. Until recently, IT service PM primarily focused on prioritizing infrastructural IT investments (e.g. networks, servers) and gave little attention to IT projects that offer business value (Trastour and Christodoulou 2009). Yet, as IT organizations seek to grasp a more complete picture of IT-related investments, they have extended existing IT portfolios with IT services that generate more and more value towards the end-customer. Even though the scope of digital services can often only be partially reflected in IT services, future research could explore whether and how to merge these two research streams.

6 Conclusion

In this paper, we conducted a systematic literature review to understand the state of innovation PM research in general as well as for services and digital services specifically, and highlight potential areas for future research. The literature review covered 52 empirical and conceptual papers in peer-reviewed journals and incorporated further 112 papers from a prior literature review of innovation PM conducted by Meifort (2015), covering in total a timeframe from 1969 until 2019. We classified the literature along the categories (1) antecedents, (2) models & frameworks, (3) consequences, and (4) challenges with the objective to create a structured and exhaustive overview of existing PM research areas. Our review revealed that while the product-centricity is still high in innovation PM research, in more recent years, services have been considered more frequently. Digital service innovations, however, have not been investigated in any research papers up to date. Given that the characteristics of digital services significantly distinguish from traditional services and products and its increasing significance in today’s economy, we suggest research concerning the management and steering of a portfolio of digital services.

Our paper results in several contributions to theory and practice. In academia, our paper addresses scholars in the intersection of innovation PM, digitalization and servitization and offers relevant theory contributions. First, by supplementing and incorporating the work of Meifort (2015), our review offers a complete and coherent synthesis of innovation PM research up to date. Our categorization including antecedents and consequences of successful innovation PM is in alignment with many relevant conceptual contributions. Unlike Meifort (2015), who explores PM along different theoretical lenses (e.g. strategic, decision-making perspectives), our categorization lays out the core developments such as what factors lead to successful PM, how differences in PM effects firm performance, and how the successful implementation of PM can be supported. Additionally, in contrast to Meifort (2015), we further grouped findings (i.e. portfolio size, formalization, top management) within a category (i.e. antecedent) and presented the main finding of a study in regards to the category. Meifort (2015) structured papers according to the author and published year in a given category and presented a brief summary rather than the main findings of each paper. Our structure intends to enable researchers to quickly identify existing research streams and for future research, especially in terms of services and digital services, to incorporate findings across the identified categories. The second main contribution of this study includes the investigation of the service and digital service perspective in existing innovation PM research. Due to the established usage of the term new ‘product’ development, it remained unclear to what extent services and digital services have been investigated without immediate recognition. Our study explored the scope and research sample of all our selected papers including the papers gathered in Meifort (2015) to identify potential contributions in terms of services/digital services. We found that significantly more studies have acknowledged services in innovation PM than anticipated. In the last years, publications on the antecedents and consequences of PM have largely acknowledged both products and services, but only few explicitly dive into specific differences between managing a portfolio of services and products (Killen et al. 2008; Calatone et al. 2010; Killen and Hunt 2010; Koh and Crawford 2012; Storey and Hughes 2013; Sicotte et al. 2015; Aas et al. 2017). Other research streams such as PM models and frameworks or challenges of service PM also remain widely unaddressed. Further, we found that digital services, despite their imminent importance to firms and in research, have not been taken up in PM research so far.

Thirdly, we identified potential areas of discrepancy between product-focused and service or digital service focused PM, which call for further examination. We propose that several aspects such as formalization and standardized procedures that resulted in superior performance in the product-centered world, do not necessarily remain valid for services and digital services given their unique characteristics (i.e. shorter lifecycles).

In regards to practitioners, we address portfolio managers and product strategists firstly by offering state-of-the-art insights to PM that shall support them in the design and optimization of their portfolios. As mentioned above, we specifically chose to classify PM research differently than Meifort (2015), to ease translation and utilization of research by practitioners. Our conceptual framework is structured in a way that allows managers to quickly identify how PM processes and the adjacent environment (e.g. culture, management, external environment) should to be organized to successfully manage portfolios and how successful PM can benefit the firm. It also provides an overview of frameworks developed by academia to support portfolio management activities (e.g. prioritization) and points out key challenges to look out for. Secondly, it provides first insights to what extent the PM of services/digital services differ to products, which is especially relevant for historically product-orientated firms transitioning more towards services. Yet, further empirical investigations that offer guidelines for service providers, in particularly, for digital service providers are still outstanding.

Yet, our research contains certain limitations. Firstly, the keyword-based search may have excluded relevant papers that were not identified through our search term in the title, abstract or keywords. Especially, given the lack of a consistent terminology to describe PM of innovation. Researchers referred to it as new product development PM (e.g. Easingwood 1986; Chao et al. 2009), project PM (e.g. McDonough III and Spital 2003) or product PM (e.g. Kang and Montoya 2014; Dash et al. 2018) at times. Yet, we tried to overcome this limitation as much as possible through a broad search term and by supplementing the database search with a backward and forward search. Secondly, the selected time frame of the search yields some limitation. We selected papers from the year 2014 onwards to extend Meifort’s (2015) review, which cut off in 2014. Yet, we found that in Meifort’s review sample, some relevant papers were missing (e.g. Calantone et al. 2010; Killen and Hunt 2010; Storey and Hughes 2013). Thirdly, the development of the categories (e.g. antecedents-organizational and managerial aspects) and the classification of papers into these categories was completed by the authors and thus inherently creates limitations. Lastly, our research only selected papers written in English or German, which entails the risk of disregarding relevant papers in other languages.

To conclude, while our research identified a few papers that investigate how to manage service portfolios, further research regarding how to design PM for firms with a portfolio of digital services is strongly recommended to support managers in the context of the digital economy. We believe that the value contribution of digital services in firms and the economy will continuously growth through advancements in technology and growing customer expectations for on-demand and individualized offerings.