1 Introduction

Smart, connected products have been pushing the markets in recent years and promise sustainable competitive advantage. At the same time, they are shifting the ecosystem of companies (Porter and Heppelmann 2014). Especially present is the current change of the automotive industry towards a business model of a mobility provider with concepts like cars and software as upgradable devices supplemented by services and an increased focus on lifecycle management (Kuhnert et al. 2017–2018; Khan and Wuest 2019; Zhang et al. 2019). The literature also refers to the so-called servitization (Kamal et al. 2020; Baines et al. 2017) and the offering of Product-Service Systems (PSS) which “are defined as an integrated bundle of products and services which aims at creating customer utility and generating value” (Boehm and Thomas 2013). At the same time, investments in startups are increasing. According to studies, an increase of around 30% to around 27.5 billion USD worldwide can be observed between 2017 and 2018. Thereby, especially startups focusing on mobility services receive high funding (Oliver Wyman 2019). Furthermore, the sector is forecasted to grow by around $230 billion worldwide between 2019 and 2027 (Reports and Data 2020). This transition from product-oriented towards service-oriented business models is a challenging task for companies. The complexity and uncertainty increase the importance and at the same time the requirements for the strategic planning and evaluation of PSS (Mourtzis et al. 2016; Tenucci and Supino 2019).

Against this background, it can be observed that many researchers examine the new business models in the mobility sector and therefore primarily apply the classification according to Tukker (2004). Thereby, the classification of Tukker (2004) is based on a shift of the ownership and distinguishes the dimensions product-oriented, use-oriented, and result-oriented (see e.g. (Mahut et al. 2015). However, it can be assumed that there have been further evolutions in the business models of the mobility sector since the publication of the classification according to Tukker (2004). Hence, for example, Kohtamäki et al. (2019) and Aas et al. (2020) have derived dimensions of PSS-oriented business models that consider the advancing digitalization. However, these classifications still only consider the influences of the above-described developments towards an ecosystem-oriented business model to a limited extent. Furthermore, Kohtamäki et al. (2019) and Aas et al. (2020) focus on the classification of business models of established companies. For this reason, we see a need to investigate the applicability of the dimensions based on a specific business sector and thus contribute to the advancement of a generally valid classification of PSS.

By analyzing the PSS-business models of startups in the context of the mobility sector, we aim to close the identified research gap. We thereby specifically focus on startups because they indicate and shape the future business models. Furthermore, as described above, the mobility sector is a suitable subject for this investigation because the sector increasingly transforms towards PSS-oriented business models and a mobility ecosystem which is influenced by startups. In order to analyze the business models, we follow the methodical approach of the taxonomy development according to Nickerson et al. (2013). The development of taxonomies has proven to be very useful in information systems research to structure results as well as handle and compare individual cases and objects (Doty and Glick 1994; Glass and Vessey 1995; Nickerson et al. 2013). Consequently, we additionally provide a taxonomy and archetypes in order to reduce the complexity of the business models especially in the mobility sector and support traditional companies by providing a comprehensive overview of the current developments of the startups of the mobility sector. Furthermore, our results should serve as a conceptual framework and starting point for future research and practical application. Against this background, the following research questions arise:

RQ1

What are the elements of PSS-oriented business models of startups in the mobility sector?

RQ2

What business model archetypes can be identified based on an empirical examination of the elements?

To answer these research questions, our paper is structured as follows. The following Sect. 2 provides an overview of the related research. Furthermore, after describing the research method of taxonomy development in Sect. 3, the taxonomy of startups of the mobility sector is described in Sect. 4 in order to answer the first research question. In this context, at first the related research is shortly illustrated following the development of the database of the taxonomy and finally the taxonomy development itself. Based on the taxonomy, Sect. 5 describes the performed cluster analysis in order to answer the second research question. At last, the results are discussed in Sect. 6.

2 Related research

To the best of our knowledge, current research contains four publications focusing on taxonomies regarding PSS-oriented business models (see Remané et al. (2017a, b), Kohtamäki et al. (2019), Lembcke et al. (2020) and Aas et al. (2020)). Remané et al. (2017a, b) for example, provide a taxonomy for digital business models in the mobility sector for a specific use-case of personal mobility and the transportation from one location to another. As well, Lembcke et al. (2020) focus on the specific use-case of ridesharing business models. Hence, Remané et al. (2017a, b) and Lembcke et al. (2020) provide multiple dimensions and characteristics like for example the travel distance and range coverage which are primarily applicable for specific use-cases. In contrast to this, Kohtamäki et al. (2019) and Aas et al. (2020) provide as part of their work more general applicable models. Thereby, the authors focus on the solution and the value in order to support the configuration of digital business models. Furthermore, from a broader focus, additional publications can be identified that describe taxonomies respectively dimensions for the classification of business models in general and with a different focus than on the mobility sector. For example, Eickhoff et al. (2017), Beinke et al. (2018), Gimpel et al. (2018), Tönnissen et al. (2020) and Weking et al. (2020a, b) focus on the financial sector and on blockchain technology. Furthermore, Chesbrough and Rosenbloom (2002), Hedman and Kalling (2003), Osterwalder et al. (2005), Remané et al. (2017b), Urbinati et al. (2017), Täuscher and Laudien (2018), Mengelkamp et al. (2019), Perboli and Rosano (2020) and Weking et al. (2020b) focus on providing taxonomies of business models on a generally valid level without focusing for example on specific industries or use-cases.

In summary, the related research lacks a taxonomy and classification of PSS-oriented business models of startups in the mobility sector which neither focuses on a specific business model of the mobility sector nor provides general dimensions. Although advances have already been made to update taxonomies by considering the digitalization, the shift of business models towards ecosystems and specifically mobility ecosystems have not been considered. Furthermore, primarily established companies and business models have been considered by the related research.

3 Methodical approach of taxonomy development

To answer the research questions, we followed the iterative method of taxonomy development by Nickerson et al. (2013) as illustrated in Fig. 1. Based on this method, a cluster analysis was performed in the second step. The objective of the taxonomy development is the classification of objects in order to better understand a complex domain (Nickerson et al. 2013).

Fig. 1
figure 1

Taxonomy development process based on Nickerson et al. (2013, p. 435)

Full size image

From this background, in a first step, Nickerson et al. (2013) suggest defining a meta-criterion as basis for the identification of the characteristics of the taxonomy. Because our work focuses on startups in the mobility sector that offer PSSs, we define the meta-criteria as “PSS-oriented business model components of startup companies in the mobility sector”. In a second step, ending conditions in order to determine when to terminate the iterative taxonomy development process should be defined. We therefore apply the objective and subjective ending conditions presented in Table 1. At the beginning of every iteration, either an “empirical-to-conceptual” or “conceptual-to-empirical” approach can be followed. The deductive conceptual-to-empirical approach follows the abstraction of elements based on the existing theoretical knowledge and their evaluation by empirical data, whereas the inductive empirical-to-conceptual approach means the opposite.

Table 1 Objective and subjective ending conditions based on Nickerson et al. (2013)
Full size table

The dimensions and characteristics are developed by iteratively examine PSS-oriented business models of startups in the mobility sector based on the information on the firms’ websites considering the previously defined meta-criterion. After an iteration is finished and a dimension or characteristic was added to the taxonomy, the ending conditions are checked. Either the ending conditions are satisfied and the development process end or a new iteration starts (Nickerson et al. 2013).

4 Taxonomy development

4.1 Data set

Our taxonomy development starts with the identification of related research, which primarily serves as the data basis for the first iteration. We therefore have performed a literature review following the guidelines of Webster and Watson (2002). The databases Elsevier’s Scopus, EBSCO Business Source Complete and Thomson Reuters Web of Science (WOS) were searched in April 2021 by the terms "Taxonomy and ("Product-service system*" OR PSS OR Startup* OR Mobility OR Automotive OR "Digital business*" OR "Business model")". Furthermore, we searched for publications in English and do not limit our search to any year or document type. Excluding the duplicates, 276 papers remain relevant from the database search. After screening the title, abstract and whole text by excluding publications not directly aimed at taxonomy or clustering of the topics product-service systems, startups, mobility, automotive, digital business or business models as well as a backward and forward search, 18 related publications remain (see Table 2). Thereby, we explicitly included publications providing a taxonomy or clustering of screening the title, abstract and whole text by excluding publications not directly aimed at taxonomy or clustering of the topics product-service systems, startups, mobility, automotive, digital business, or business models. The dimensions taken from the related research (see Table 2) are used for the first iteration of the taxonomy development as described above.

Table 2 Data set of the first literature review
Full size table

Next to the related research and as basis for the empirical-to-conceptual iterations of our taxonomy development, we made use of the company information of startups provided by the Crunchbase database (Crunchbase Inc. 2021) in order to identify companies and relevant PSS oriented business models in the mobility sector. At first, we filtered for companies associated to the keywords “Automotive”, “Vehicle”, “Car” and “Mobility”. Accordingly, the industries “Autonomous Vehicles”, “Automotive”, “Car Sharing”, “Electric Vehicle”, “Fleet Management”, “Last Mile Transportation”, “Parking”, “Procurement”, “Public Transportation”, “Ride Sharing”, “Leasing”, “Auto Insurance”, “Financial Services” and “Payments” were selected. Furthermore, as China, the United States, Japan and Europe are the most important regions for the automotive industry (OICA 2021; VDA 2021), we have restricted the location of the headquarters to these regions. At last, we excluded companies with less than 251 employees and a last funding before 2020 in order to consider the most relevant startups. In doing so, we perceived an initial list of 78 companies which are subsequently used for the collection of PSS offered by the startups and elaboration of the taxonomy.

The database search with Crunchbase was performed on 03/06/2021. Since Crunchbase does not provide detailed information about the company’s business models, we used the listed companies as database for a subsequent detailed analysis of the information provided on the company’s website until 07/30/2021. The following analysis of the websites of the company list results in 49 PSS-oriented business models which were analyzed in multiple iterations following the guidelines by Nickerson et al. (2013) as described in Sect. 2. Thereby, we excluded 44 companies due to unavailability of data because the website was not available or in other languages than English or German, the company was founded more than 15 years ago, the company focuses on the development of automotive parts, cars, or vehicles itself, components, only on sale of products or the company’s business models is not essentially mobility oriented. The remaining 34 companies are primarily located in the United States of America (20 companies) followed by Europe (ten companies) and finally China (four companies). The results are illustrated in Table 3.

Table 3 Clustered PSS-oriented business models per company
Full size table

Furthermore, we used the three common clusters “product-oriented”, “use-oriented” and “result-oriented” of PSS provided by Tukker (2004) in order to structure the objects during the iterations and for analysis of our results. As described by Tukker (2004), product-oriented services are services, advice or consultancy offered in addition to a product like maintenance contracts. The business model focuses primarily on the sale of the product, which therefore changes the owner (Mahut et al. 2015, p. 843). Use-oriented services are characterized by the fact, that the ownership does not shift to the customer. In general, the provider is responsible for maintenance, repair and control and the customer pays a fee. However, it can be distinguished between product lease with usually unlimited access, renting/sharing with limited access and pooling with simultaneous use of the product. At last, in case of result-oriented services, the customer and provider generally agree on a defined result without a defined product. This could be for example the outsourcing of activities like catering or pay per service units, where the customer buys the output (Tukker 2004).

4.2 Iterations

In order to define the dimensions of the taxonomy, our first iteration follows the conceptual-to-empirical approach, building up on the identified related research and the literature on business models. We aimed to select dimensions for which information on the business models of the considered companies are accessible on their websites. Since Eickhoff et al. (2017) and Beinke et al. (2018) as well-developed a taxonomy based on information from companies' websites, we used the dimensions focused by the authors as orientation for our first iteration (see Table 2). On closer observation, it can be noted that the dimensions derived by Eickhoff et al. (2017) and Beinke et al. (2018) are also based on further related research of Table 2 such as Chesbrough and Rosenbloom (2002) and Osterwalder et al. (2005). Furthermore, the dimensions can be found primarily in the close related research of Remané et al. (2017a, b) and Lembcke et al. (2020) with focus on carsharing business models. Their taxonomy as well contain the dimension “value proposition” and “revenue model”. In further detail, several intersections can be recognized. For example, Remané et al. (2017a, b) list the dimension “sales channels” and Lembcke et al. (2020) the dimension “payment options” under the main dimension “interface” with which intersections can be drawn with the dimension sales channel in this paper. A further example is the main dimension “service platform” of Lembcke et al. (2020) and the corresponding dimension “service scope” which is similar to the dimension “product/service offering”. Nevertheless, as mentioned above, a dimension like “organization model” (see Lembcke et al. (2020)) does not fit for the research aim and method of this paper since the authors focus on carsharing business models in particular and also pursue a different research approach. At the same time, the dimensions are comparable with the dimensions applied in more general valid taxonomies like, e.g., by Kohtamäki et al. (2019).

This paper does not aim to build on this work, but to give a new view for PSS in the automotive industry. Finally, we selected six dimensions which can be defined after Eickhoff et al. (2017, p. 10) as follows:

  • D1 Value Proposition = Describes the value the company creates for its ecosystem (customers, partners etc.)

  • D2 Sales Channel = Describes how the products and services are ordered by the customers

  • D3 Market Segment = Describes to whom the company intends to offer its products and services

  • D4 Revenue Stream = Describes how the company generates revenue from its products or services

  • D5 Product/Service Offering = Describes what the company offers to its customers

  • D6 Technology = Dominant IT artifact that is the driver for the IT-based business model

Since these rough dimensions do not yet satisfy the ending conditions, the second iteration followed an empirical-to-conceptual approach based on the 20 business models that are assigned to the product-oriented cluster. Thus, we were able to identify 33 characteristics according to the dimensions of the first iteration. Since we added several characteristics and considered only a part of the objects, the taxonomy has not yet reached all ending conditions. The third iteration as well follows an empirical-to-conceptual approach. Thereby, the 18 use-oriented business models were added to the iteration. We identified the eleven characteristics C1,13 = Integrate mobility in daily routine, C1,14 = Earnings C3,3 = P2P, C4,4 = Abonnement, C4,5 = Leasing, C5,13 = Vehicle sharing, C5,14 = Vehicle renting, C5,15 = Parking, C5,16 = Autonomous Driving, C5,18 = Logistics and C6,11 = Augmented Reality. Furthermore, three characteristics were specified C2,1 = Product as Embedded device, C4,2 = Fee as Fee per use and C5,1 = Charging as Power. Similar to the second iteration, the taxonomy is still changing significantly and does not meet all ending conditions. In the fourth iteration, we add the remaining 11 result-oriented business models to the scope. Thereby, one characteristic C5,18 = Ride-hailing is added as well as C1,2 = Cost reduction and C1,14 = Earning is combined to C1,15 = Monetary and C1,9 = Lower entry barrier to electrification and C1,13 = Integrate mobility in daily routine to C1,16 = Integration of sustainable mobility. Furthermore, we specify C1,7 = Vehicle availability as PSS availability. After this iteration, all identified PSS-oriented business models are considered. Because of the minor changes in this iteration, we consider all ending conditions as met. The resulting final taxonomy, illustrated in Table 4, contains six dimensions and 52 characteristics. The corresponding crosstab analysis is illustrated in Table 5.

Table 4 Taxonomy of mobility-oriented PSS
Full size table
Table 5 Results of Crosstab Analysis based on Ward’s method
Full size table

5 Cluster analysis

Based on the previously developed taxonomy, we have performed a cluster analysis following the approaches of Beinke et al. (2018, p. 5) and Tönnissen et al. (2020, p. 313) with support of SPSS (version 26) in order to review and identify clusters respectively archetypes of the business models in the mobility sector. Therefore, the assignment of the PSS-oriented business models to the characteristics of the previously developed taxonomy served as the data basis for the cluster analysis.

We also followed the two-step method of Punj and Stewart (1983) to identify the appropriate number of clusters. In a first step, Ward´s method is applied with support of SPSS (version 26). Thereby, the most similar objects, according to the number of identical characteristics, are grouped iteratively until all objects belong to the same group. Based on the agglomeration schedule as one output of clustering with Ward's method which also represents the increase in heterogeneity in each stage through a coefficient and the “squared Euclidean distance”, we were able to select a suitable number of clusters. The scree plot of the coefficient indicates that a three- respectively four-cluster solution is suitable.

In a second step, these possible cluster solutions were compared using the K-Means method. ”In a nutshell, K-means is a prototype-based, simple partitional clustering algorithm that attempts to find K non-overlapping clusters. These clusters are represented by their centroids (a cluster centroid is typically the mean of the points in that cluster).” Wu (2012) As a result, the three-cluster solution required five iterations and the four-cluster solution three iterations until the sum of the squared deviations from the cluster focal points is zero. Against this background, we select a four-cluster solution.

Based on the resulting crosstab analysis (see Table 5), the clusters are visualized in Fig. 2. In order to develop Fig. 2 one characteristic was assigned to a maximum of two clusters in which the dominance of this characteristic is highest. An exception was made for the characteristics System-/Application Software and customer experience, due to the high dominance in all clusters. At last, characteristics with a low dominance in any cluster (< 15%), are not considered in Fig. 2. By the following description of each cluster and its typical characteristics, archetypes of PSS-oriented business models and the answer to the second research question are provided.

Fig. 2
figure 2

Clusters of PSS-oriented business models of startups in the mobility sector

Full size image

5.1 Cluster 1: vehicle operation and availability

The first cluster contains companies that support the operation and availability of a vehicle by services like maintenance, warranty, driver assistance systems and power. Since maintenance and warranty are long-established services, the technological shift of the mobility sector towards electrification and autonomous driving are supported by providing driver assistance systems and power (currently especially electric power) to the customer. Against this background, the revenue is not generated through fees per use, but through subscription and freemium services in a B2C (Business-to-Customer) and B2B (Business-to-Business) market segment. Customers can order the products and services traditionally by physical contact or via App, whereby in general IT-technologies are not very dominant in this cluster. Only the share of the system—and application software is significant but still below the values in the other clusters. The companies assigned to this cluster promise to place a high value on customer experience, time efficiency and effectivity, as well as customizability. This is expressed for example by the ease of use of vehicle charging and customer specific solutions in B2B environment. The following companies are assigned to cluster 1: Aurora, CAR Inc., ChargePoint, CollisionRight, Go Insurance, Joby Aviation, Lilium, Mapbox, NIO, Tesla, Wallbox, WM Motor and Xpeng Motors. Cluster 2: Platform operation.

5.2 Cluster 2: platform operation

The PSS-oriented business models of the second cluster focus on the B2B market segment by supporting fleet management and providing digital service—and retail platforms to the customers. As well as in cluster 3 and 4, system—and application software are very dominant IT-technologies. Furthermore, companies use digital platforms, analytics, and application programming interfaces (APIs) which support the business management of the customer in a B2B environment by providing information. The primarily sales channel is the website, respectively the contact through the website. Since the objective of the companies of this cluster is to support the business of other companies, it stands to reason that the value propositions are efficiency, transparency, and information, customizable as well as monetary like the reduction of cost and increase of earnings. Furthermore, it is characterized by patterning and taking over activities of the other company to provide customer satisfaction and all in one solution. The following companies are assigned to cluster 2: Arrive, Auto1 Group, Bird, Bolt, Helbiz, MotorK, Surve Mobility, Tekion and Via.

5.3 Cluster 3: sharing concepts

Companies of the third cluster contain companies with vehicle sharing business models and digital service platforms to support vehicle sharing with focus on B2C, B2B but as well P2P (Peer-to-Peer) market segment. The customer can acquire the service via App or the embedded device. The revenue is generated through fees per use or subscription. In contrast to cluster 1 and 2, the companies of this cluster promise environmentally sustainable PSS as well as the integration of sustainable mobility in the society through a low entry barrier. Furthermore, the customer experience and satisfaction, flexibility as well as safety and security of the PSS are key value propositions of this cluster. The following companies are assigned to cluster 3: Bird, Bolt, Cityscoot, Getaround, Helbiz, Lime, TIER Mobility, Turo, Urent and VOI Technology.

5.4 Cluster 4: autonomous mobility concepts

The fourth cluster contains PSS-oriented business models focusing on autonomous driving and ride-hailing, as well as logistics (transportation of goods). Customers of primarily the B2C environment pay a fee per use and can acquire the service via App. In contrast to the other clusters, machine learning, GPS, and Internet of Things (IOT) are, next to the system—and application software, the dominant IT-technologies. Since this cluster focuses on autonomous driving, safety and security is a central characteristic of the value propositions. Like in cluster three, the environmental sustainability, customer experience and flexibility of the PSS are as well addressed. But since the customer must not drive by himself, time efficiency and effectivity are a key value of the companies of this cluster. The following companies are assigned to cluster 4: Cruise, Local Motors, Nuro, Pony.ai and Waymo.

6 Discussion

Based on the developed taxonomy, the results of the cluster analysis show clear differences between the PSS-oriented business models of startups in the mobility sector, with several theoretical and practical implications. From a theoretical perspective, to a certain extent, correlations between the resulting archetypes of our analysis and the cluster provided by Tukker (2004) can be drawn. Our results show that the majority of the investigated PSS-oriented business models of our first cluster were assigned to product-oriented services of Tukker (2004) like maintenance, warranty, and power services. Nevertheless, in the context of the mobility sector, our analysis indicates, that regardless of whether the ownership is held by the provider or the customer, the operation and availability of the vehicle and mobility is a relevant topic of startups. Hence, our results from the perspective of the mobility sector indicate that the distinction based on the ownership of Tukker (2004) is not a stand-alone characteristic for differentiating product-oriented business models. Thus, our results support the approaches of the enhancement and combination of the dimensions of PSS of Aas et al. (2020) and Kohtamäki et al. (2019). Against this background, our results show first of all similarities between the product-oriented and performance-oriented contracts of the value capture dimension of Aas et al. (2020) and the revenue streams of fee per use and subscription of the taxonomy provided in this work. However, our results contribute to the literature by pointing out, that as well freemium offers should be considered, especially in the growing mobility ecosystem. Furthermore, in addition to the dimensions of Aas et al. (2020), traditional sales or leasing offers are as well dominant in the second cluster. Second, in line with the differentiation of value delivery by Aas et al. (2020) into smart digital services and non-smart services, our results confirm the distinction of an embedded device, website, APP and physical of the dimension sales channel. However, our results show the need to distinguish between applications and physical contact since business models in the B2C environment are primarily based on APPs, which require a corresponding end device, whereas the B2B business relies on websites and physical contact.

In addition to the dimensions of Aas et al. (2020), further correlations between the results of this work and the dimensions provided by Kohtamäki et al. (2019) can be noticed. First, similarities can be identified between the first cluster of vehicle operation and availability as well as the business models product-oriented service and integrated solution provider of Kohtamäki et al. (2019). Furthermore, the second cluster of platform operation and the business model of the platform provider of Kohtamäki et al. (2019) are very similar. However, for example it can be recognized that Kohtamäki et al. (2019) assign the business model of platform provider to the outcome-oriented solution pricing whereby the results of this paper show that the platform operation primarily takes place in a B2B environment and the services are usually sold to the partner instead of concluding outcome-oriented agreements. For example, companies like Bird and Bolt offer a vehicle sharing platform and their vehicles like e-scooters to a local sharing provider for a fee, enabling the provider to develop a business on its own. Furthermore, regarding the dimension of solution customization of Kohtamäki et al. (2019), for example, the products and services related to charging of vehicles of the first cluster of our results show that customizable solutions are required in the B2B environment. Against this background, our results indicate correlations between the first cluster in this paper and the business model of an integrated solution provider of Kohtamäki et al. (2019). Additionally, from the perspective of the dimension of the solution digitalization of Kohtamäki et al. (2019), for example, the autonomous function is described as a characteristic of an outcome provider. This description is consistent with the business models of a robotaxi service of the fourth cluster of our results. However, considering for example the PSS-oriented business model of the autopilot of Tesla assigned to the first cluster, our results show business models characterized by autonomous function which are not assigned to the business model of an outcome provider of Kohtamäki et al. (2019). Moreover, the characteristic of the autonomous function of the first cluster, which has as previously described correlations to the product-oriented service provider of Kohtamäki et al. (2019), is a differentiation to the description of the level of digitalization of a product-oriented service provider of Kohtamäki et al. (2019), which is described by Kohtamäki et al. (2019, p. 388) as "some smart features based on remote diagnostics". Furthermore, parallels can be identified between third and fourth cluster of this paper and the characteristics of the outcome provider according to Kohtamäki et al. (2019). These include, for example, the revenue stream of fee per use, which, as described by Gebauer et al. (2017), is similar to the outcome respectively performance pricing model described by Kohtamäki et al. (2019). However, the characteristics of third and fourth cluster, for example, suggest that, in contrast to the description of Kohtamäki et al. (2019, p. 388) on outcome providers, it is not necessarily the manufacturer of the product that offers the PSS-oriented business model but as well other companies.

In summary, the characteristics of the business models of a product-oriented service provider and industrializer according to Kohtamäki et al. (2019) are minorly represented in the startups of the mobility sector. Rather, the results of the cluster analysis in this paper show correlations to the characteristics of the business models of an integrated solution, platform, and outcome provider according to Kohtamäki et al. (2019). This is as well indicated by the advanced level of digitalization of the startups in the mobility sector. Especially against the background of the ecosystem concept, which is as well focused by Kohtamäki et al. (2019), it can be discussed if the distinction between a product-oriented, industrializer and integrate solution provider will support research and practice in the future. We rather note that these business models are similar to the degrees of servitization rather than business models in an ecosystem (Weiller and Neely 2013, p. 19; Zheng et al. 2017). As the results of this paper show, the product-oriented, industrializer, and integrate solution provider could be named product operation and availability provider. In this way, it is expressed that the business model is part of the value creation of an ecosystem. Furthermore, the discussion above indicates on the one hand that the distinction between B2C, B2B and P2P is another characteristic between the business models such as platform provider and outcome provider respectively sharing concepts. On the other hand, it shows that due to the advances in the digitalization, a distinction of third and fourth cluster, which can be assigned to the outcome providers according to Kohtamäki et al. (2019), is reasonable. The distinction of autonomous and sharing concepts is also shown by other industries respectively products like the sharing of tools, apartments, and parking spaces as well as autonomous robots. Additionally and in contrast to Kohtamäki et al. (2019) and Aas et al. (2020), the characteristics of the third and fourth cluster show that the environmental sustainability and customer experience in particular should be taken into account when characterizing PSS-oriented business models. Hence, the results and discussion contribute to the theory by indicating that in total, one model is not sufficient in order to characterize PSS-oriented business models. Rather, it can be derived, that several models should be built.

Nevertheless, since previous research either focus on the specific use-case of ridesharing respectively personal transportation or in general on PSS-oriented business models without a specific focus on the mobility sector, our study addresses this gap and provides a taxonomy focusing on PSS-oriented business models of the mobility sector. Furthermore, in the context of PSS in the mobility sector, our study is, to the best of our knowledge, the first which bases on a comprehensive database of companies. Like all taxonomies, our taxonomy contributes to the theory by providing a comprehensive overview and clustering of existing PSS-oriented business models in the mobility sector and a starting point for future research.

In addition to these theoretical implications, our study also reveals implications for practice. First, by focusing on startups, our study shows the current state of startups and new PSS-oriented business models in the mobility sector, which is changing rapidly. Based on this initial analysis, practitioners can draw first conclusions like for example that round about 50% of our initial identified startups of the mobility sector show a PSS-oriented business model.

Second, our developed taxonomy and archetypes can be used by practitioners to a certain extent as guidance for business model innovations in order to manage and expand their business model portfolio, the classification of competitors, potential business partners or business models. For example, like identified by Lembcke et al. (2020, p. 13), who examined ridesharing business models, companies also offer their ridesharing platform to other companies instead of operating them by their own since this business model should be more efficient. As well, the second cluster of our analysis contains business models offering ridesharing platforms. In this context, like pointed out by for example Moazed and Johnson (2016) and Weiß et al. (2018), that digital platforms and their proper marketing will play a central role in the business models of the automotive industry. Furthermore, innovative charging solutions like Battery as a service or battery swap by NIO to support the integration of electro mobility and to increase the availability of the vehicle of the customer and the company’s own fleet for sharing models are entering the market and increase the pressure on established companies.

Third, the results of our work could support strategic planning and evaluation of PSS-oriented business models, since the different clusters could have individual requirements regarding evaluation methods and KPI. For example, it could be discussed if on the one hand sharing and ride-hailing models (see cluster 3 and 4) require the evaluation of for example active users and operation costs while on the other hand cluster 1 and 2 require the evaluation of running contracts as one measure for the performance. Furthermore, a closer look at the individual clusters of our analysis reveals that business models focusing on the transportation of people and goods (see cluster 3 and 4) are, as well like the results of Lembcke et al. (2020, p. 13) reveal, characterized by the value propositions flexibility, safety and environmental sustainability as well as monetary factors. In addition, we found that especially our fourth cluster is characterized by IT-technologies like IOT, machine learning and that the system—and application software are of central significance in all clusters. In general, the taxonomy and clusters can serve as a basis for practitioners for further additions and adjustments according to their individual needs and processes as well to better meet the interest of the customer. The main findings and implications are illustrated in Table 6.

Table 6 Main findings and implications
Full size table

7 Conclusion

Motivated by the increasing importance of startups in the mobility sector, we developed a taxonomy from their PSS-oriented business models for, based on data from the companies' websites. Furthermore, the provided taxonomy was used to identify archetypes of these business models. Both, the taxonomy, and the archetypes offer researchers and practitioners insights into the business models of startups and provide a starting point for further investigations and applications. In combination with the established theoretical dimension of business models, our research is robust and at the same time not as generic as other taxonomies and clusters. Nevertheless, as any research, limitations must be taken into account and future research should be pointed out. Our database for the development of the taxonomy is based on the data provided by Crunchbase. Furthermore, we analysed the companies only based on the data of their websites, which are available in English and German. In this context, it should be noted that these data and business models of startups in particular are constantly and dynamically changing. Furthermore, as with the use of other data sources, information may therefore be disregarded. Hence, future research could take into account these changes over time and review the provided taxonomy and clusters. In this context, information from additional data sources like information provided for investors would support the analysis. Since we initially analysed 78 companies and finally focused on 34, an extension of the database and an investigation beyond that of the websites (for example qualitative research such as mixed-method cross-case studies) would lead to further interesting results. Also, due to the sole availability of information for example in the Chinese languages, individual companies could not be considered so that a consideration of these in future research would be meaningful. Considering and pointing out local differences in the taxonomy and clusters could also provide further interesting insights. As for all taxonomy developments, it also applies for this paper that subjective perceptions and errors in the analysis could influence the results. Furthermore, the detailing of the taxonomy and clusters is challenging. Against this background, future research may apply and expand the results of this paper as well as focus on a single dimension like for example the technology dimension, which could only be analysed superficially but could lead to interesting insights. At last, our research can serve future research to analyse and compare companies and business models both internally and externally, as well as to structure and compare research results.