Keywords

1 Introduction

Organizations have significantly increased the use of cloud computing services, which offer virtualized IT resources in terms of infrastructure, data and applications, over the last years [1]. This development is set to continue: according to a current study of Gartner [2], the worldwide public cloud service market is projected to grow 16.6% in 2018 to total USD 287.82 billion, up from USD 246.84 billion in 2017. The largest segment will remain Software as a Service (SaaS) with USD 55.14 billion.

However, cloud providers face considerable challenges, as they generally require profound expertise with regard to both technical infrastructure concepts and the design and management of service-oriented business models [3]. In practice, it can be particularly observed that several cloud providers encounter difficulties to effectively design suitable business models. This is why many are still experimenting with a variety of business models aiming to put themselves in a sustainable and profitable position within the cloud computing ecosystem [4, 5]. Indeed, market studies have revealed major differences in the level of performance between cloud providers [6, 7]. The literature’s understanding of success factors of cloud providers’ business models and thus, the reasons for this performance discrepancy is, however, still limited [8, 9].

Especially the largest segment, SaaS, has only been covered in a few isolated studies with regard to success factors [10,11,12]. These contributions mainly focus on the value proposition, while other business model components are disregarded, or are solely literature-based, trying to summarize the fragmented literature on SaaS or transferring success factors of related fields. Against this background, the goal of this research paper is to investigate the business models’ success factors of SaaS providers by conducting an exploratory multiple-case study. 21 expert interviews with representatives from 17 cloud providers offering SaaS services serve as main data collection method.

2 Related Work

Cloud computing is defined as “[…] a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction” [13]. The literature distinguishes between three main service models – Infrastructure as a Service (IaaS), Platform as a Service (PaaS) and Software as a Service (SaaS). These three service models form layers which are interrelated, each building upon the former. A further differentiation is made between four deployment models: public, hybrid, private and community [1]. The characteristics of cloud services, including on-demand self-service, broad network access, resource pooling, rapid elasticity and measured service, distinguish it from its traditional counterpart the on premise IT [4].

Whereas research has primarily emphasized the technical aspects of cloud computing, significantly less consideration has been given to the substantial changes within the business perspective [1, 6]. A literature research of Herzfeldt, Floerecke, Ertl and Krcmar [6] revealed that three classes of publications from the cloud provider’s business perspective have been advanced so far: papers that are (1) proposing models or algorithms focusing on the optimization of costs and other resources, (2) dealing with business models and ecosystem models, or (3) discussing fundamental business benefits and challenges of cloud computing from different perspectives and for various industries. Concerning (2), the introduction of cloud computing has radically changed the way IT resources are produced, provided and consumed [1]. Hence, it is considered as a co-evolution of computing technology and business models [14].

A business model, in general, is regarded as a tool for depicting, innovating and evaluating the business logic of a firm [15]. Even if no commonly accepted definition of the term “business model” has been established yet, the component-based view dominates the research [16]. Accordingly, a business model is a system comprising a set of components and the relationships between them [15]. There is, however, no consensus on the specific set of relevant components [17]. Nevertheless, a large number of cross-industry and industry-specific business model frameworks provide possible design options for selected components [16]. One comprehensive and widespread cross-industry framework is the Business Model Canvas [18], which includes nine components: key activities, key resources, partner network, value propositions, customer segments, channels, customer relationships, cost structure and revenue streams.

Specific research on cloud computing business models is nascent [4, 19]. Giessmann and Stanoevska-Slabeva [20] proposed a classification model for PaaS providers’ business models and hypotheses regarding future directions. Giessmann and Legner [21] published a set of possible design principles that guide software providers to define PaaS business models. So far, the only holistic cloud-specific business model framework was developed by Labes, Erek and Zarnekow [22]: their morphological box entails various categories representing the basic components of a business model, each broken down into design features that show design options [22]. Based on this, Labes, Hanner and Zarnekow [8] analyzed the business models of selected cloud providers, matched them with the framework and identified four common patterns of cloud business models. Another research strand is the analysis of the impacts of the shift from an on premise to a cloud computing business model (e.g. [4, 5, 23,24,25,26]). Other scholars have dealt with the process itself of transforming an on premise to a cloud business model [27, 28]. For supporting this business model development process, Ebel, Bretschneider and Leimeister [29] developed and evaluated a software tool. A literature study of Labes, Erek and Zarnekow [19] shows that several further contributions have dealt with one or a small quantity of business model components, such as the revenue [30] or the resource model [6], while a holistic approach remains an exception. Such a strict separation, however, contradicts the logic of business models as the single components are understood as interrelated [15].

For a long time, business models have played a central role in explaining a firm’s performance and deriving success factors [16, 17]. According to Rockart [31], success factors are defined as “[…] the limited number of areas in which results, if they are satisfactory, will ensure successful competitive performance for the organization”. Success factors are applicable to all companies in an industry with similar objectives and strategies [31, 32]. A distinction can be made between generic success factors, which are valid for all kind of companies, and domain-specific success factors, in this case cloud-specific success factors [8]. Therefore, it is difficult to transfer the success factors from adjacent research areas to the cloud computing ecosystem without prior examination. The literature’s understanding of success factors of cloud providers’ business models is limited [8, 9]. Beside a recent study focusing on success factors that relate to the providers’ relationship with the consumer in the end consumer market [9], research has only provided one extensive analysis of the characteristics of cloud computing business model components and their link to business success: Labes, Hanner and Zarnekow [8] derived abstract success factors by relating publicly available characteristics of the business model components to a firm’s web visibility and profit. However, both studies disregard that the cloud computing ecosystem contains providers fulfilling various roles and thus, is characterized by a high degree of heterogeneity [33]. Focusing on the ecosystem role of SaaS providers, Ernst and Rothlauf [12] proposed potential success factors by transferring success factors from related research fields. Walther, Plank, Eymann, Singh and Phadke [11] summarized single success drivers based on a review of general SaaS literature. Wieneke, Walther, Eichin and Eymann [10] conducted expert interviews with employees of a SaaS provider with particular focus on the value proposition, while neglecting the other business model components.

3 Research Design

3.1 Research Methodology

To investigate the success factors of SaaS providers’ business models, a multiple-case study approach was selected [34, 35]. Yin [35] defines a case study as “[…] an empirical inquiry that investigates a contemporary phenomenon (the “case”) in depth within its real-life context, especially when the boundaries between phenomenon and context may not be clearly evident”. Case study research is suitable (1) to undertake research in a field in which few previous studies have been carried out, (2) to answer “how”, “why” and “what” questions in order to understand the nature and complexity of the processes taking place and (3) to learn about the state of the art and generate theories from practice [34]. All of these three factors apply to the study at hand.

This case study follows the positivistic tradition [36]. More precisely, an inductive research approach was applied with the aim to reach predominantly exploratory conclusions [35]. In line with that, this research did not start with a specific hypothesis being tested. However, Eisenhardt [37] argues that a priori specification of constructs can be a helpful tool in shaping the initial design of a study. Therefore, this research used existing concepts from the literature on cloud computing business models and success factors for initial framing. But “[…] no construct is guaranteed a place in the resultant theory” [37], because of the study’s exploratory nature.

3.2 Site Selection

Case studies can be categorized as single and multiple-case studies [36]. For this research, a multiple-case study approach was chosen as the intent was hypothesis and theory building. In addition, a multiple-case design allows cross-case analysis which yields more general research results and ensures internal validity [34, 35].

Multiple-case designs depend on careful case selection to maximize insights [36]. The units of analysis are individual SaaS providers. These were chosen for enabling literal (conditions of the case lead to predicting the same results) and theoretical replication logics (conditions of the case lead to predicting contrasting results) [35]. As a prerequisite, the respective provider had to be listed in a ranking as a successful provider by a leading research and advisory company like Gartner or Forrester. In this way, it was ensured that the basic settings of the providers were fundamentally the same (literal replication). For theoretical replication, the aim was to include providers with different experience, size, geographic coverage, number of occupied ecosystem roles, target markets, served industries, and assessment of the importance of cloud compared to on premise. Due to the heterogeneity of the cloud providers, it was possible to draw cross-case results, as it replicated findings across all cases and helped to detect similar and contrasting results, which lead to generalizable conclusions [35]. The selection procedure resulted in seven SaaS providers. This is in line with Eisenhardt [37] who recommends four to ten cases as a reasonable number to reach external validity. Besides, ten successful IaaS and PaaS providers that additionally offer SaaS services were chosen using the same case selection procedure.

3.3 Data Collection and Analysis

As data triangulation is highly recommended in case study research [34, 36], data collection relied on more than one source. The study started with a screening of the websites of the respective providers to gain background information. This initial data collection was directly incorporated in the following expert interviews with one or two representatives from each organization. The interview partners were selected based on the following criteria: the person should hold a managerial position and have responsibility for overseeing the organization’s business models and strategies. Seven interviews across seven SaaS providers were conducted. Besides, fourteen interviews with IaaS and PaaS experts from ten cloud providers which also offer SaaS were performed. The 21 experts stemmed from twelve large and five medium-sized cloud providers, had between three and ten years’ experience in the cloud field and held leading positions within their companies (board members, portfolio, product, sales, marketing and IT managers, and senior consultants). This research paper presents the integrated results as the discussions were not limited to the chosen cloud layer. Interviewees often drew on their experience gathered in SaaS. This may be due to the fact that the three cloud layers build up on one another. That proved to be a strong advantage as it increased the data pool and thus, the generalizability of the findings.

The interviews were based on a pre-tested interview guide, encompassing open-ended questions. When designing the interview guide, it had to be considered, whether to ask with a framework such as the Business Model Canvas [18] in mind or without a given set of business model components concerning success factors. In line with the study’s exploratory character, it was decided not to push the experts into a given framework. Instead, they should be given the chance to think and answer freely, independent of predefined components. This approach was also deemed more appropriate by the two pre-test sessions. The resulting interview guide (available upon request from the author) focused on identifying the success factors of SaaS business models from various perspectives (Fig. 1). These perspectives were derived from the literature on collecting success factors [32] and on the characteristics of business models.

Fig. 1.
figure 1

Selected perspectives on success factors of cloud computing business models

Full size image

The interview guide was not sent to the experts in advance as a spontaneous response was desired. The 21 interview sessions took place from June to November 2017 in a face-to-face manner or via telephone in German. The duration of the interviews ranged from 30 to 100 min. Whenever appropriate, the laddering technique was applied, which follows a process of digging deeper by asking further questions [38]. Thus, the interviews were conducted as guided conversations rather than structured interviews. In order to facilitate the data analysis, all interviews were recorded with the permission of the participants. Each interview was subsequently transcribed and proof read. As the experts were ensured anonymity, the data acquired was sanitized so that no individual person or organization can be identified. The data collection was undertaken until theoretical saturation was reached [39]. This was the case, when the answers of the experts were repeated many times and no new aspects were added, so that further data would have not provided additional insights. This was achieved relatively quickly as the fourteen experts for IaaS and PaaS often also referred to SaaS.

The data analysis process was conducted in two phases based on the recommendations of Corbin and Strauss [39] using MAXQDA. The first phase consisted of open coding – a line-by-line analysis of the transcribed data. The components of the Business Model Canvas [18] served as the basis categories of the coding scheme. The resulting codes were discussed iteratively with colleagues of the chair until consensus was obtained. Within this step, the number of codes was reduced. In the second phase, the codes were further consolidated via axial coding technique. The whole data analysis was an iterative process of coding data, splitting and combining categories, and generating new or dropping existing categories. A factor was classified as “success”, if it was expressed as such by various experts or if, in case of a single notion, convincing arguments were provided, or if it emerged as an important aspect during analysis of an individual case or if it was identified as such when conducting cross-case comparisons. In line with Corbin and Strauss [39], no attempt was made to statistically evaluate the importance of the results. Theoretical relevance of the concepts was established by their repeated presence or notable absence. As a SaaS provider is dependent on the IaaS and PaaS layers, only success factors that can be influenced actively by a SaaS provider were incorporated. Every success factor was further classified by the author as generic (g), similar to other domains, or SaaS-specific (s). In the following, the classification is noted in brackets after the designation of each factor.

4 Success Factors of SaaS Providers’ Business Models

4.1 Revenue Streams

Setting Prices According to the Generated Customer Business Benefit (s)

For SaaS providers, a monthly subscription fee is the common pricing practice. A traditional basis for price calculation is the provider’s incurring costs. A much more promising approach is pricing according to the generated customer benefit: the provider tries to tip into the additional value his service yields for the customer’s business – direct cost savings or additional revenues. This has the benefit of being very transparent and comprehensive for clients which increases their willingness to pay. A prerequisite is that the provider maintains a close and long-term relationship with the client and understands his business processes. This approach has proved most suitable for applications related to production processes, but less for standard office software.

Offering Flexibility Within the Pricing Model (s)

Besides offering predefined pricing models, an open attitude towards individual requests from customers concerning the design of the price model can have a strong effect on customer satisfaction. According to the experts, a SaaS provider should not insist on a standard pricing model. Instead, a price model should take the characteristics of the business model and the situation of different customers into account. Therefore, a SaaS provider should, in certain cases, be capable to adapt the pricing model according to specific customer preferences.

Supporting of “Bring Your Own License” (s)

In order to attract new customers or to move long-term on premise customers into the cloud, it is crucial to allow bringing previously purchased licenses of on premise software for a corresponding cloud service. This is challenging in case the license stems from a third-party provider. An on premise license has a specific calculation basis, most commonly the customer’s CPU cores. But due to virtualization of hardware, it is difficult to ensure that the service solely runs on the allowed number of CPUs in the cloud. Nevertheless, it is vital to comply with the license conditions.

4.2 Key Resources

Building Up Domain Knowledge and Industry Expertise (g)

SaaS providers must aim to accumulate a broad domain knowledge and industry expertise in the fields they operate in. Despite the importance of technical aspects of SaaS service implementation, it is a success factor to have a deep understanding of the customers in their situation within their industry. If providers are not able to build this up in-house, they must rely on partners that possess such knowledge. This is a prerequisite to ensure that the developed SaaS service delivers added value to customers.

Possessing the Leading Certificates (s)

A SaaS provider has to obtain the relevant certificates. These are commonly demanded within tendering processes. The importance of certificates varies in relation to client size and type of SaaS service: whereas small firms may partly consider certificates as dispensable, they are vital in an enterprise environment. SaaS services working with insensitive data commonly do not demand certifications. As the procedure of obtaining a certificate is time-consuming and expensive, smaller providers are unable to compete in this aspect. Providers offering SaaS exclusively are strongly advised to choose a certified IaaS provider.

Having a Multitude of Highly Qualified Employees (g)

The experts stressed the importance of having a multitude of highly qualified employees. This particularly includes software developers who implement the cloud service portfolio. Besides that, employees with technical know-how are needed to support the customers in case of uncertainties and problems over the whole cloud service lifecycle. A high volume of qualified employees generates speed and innovation which is required in the rapidly growing and changing cloud computing ecosystem. However, the interviewees reported that it is becoming increasingly difficult to attract and retain skilled developers as the market for developers is running dry.

Owning a Large (Pre-cloud) Customer Base (g)

A large customer base that a provider established before the emergence of cloud computing is considered a high-valued resource. A successful, established business relationship provides easy and fast access to potential cloud customers: First of all, long-standing and satisfied customers have confidence in the qualities of the provider. Second, the customers know the provider’s support processes and do not want to contact another provider when problems occur. Third, customers are used to the provider’s applications, so they are commonly reluctant to change the provider.

4.3 Value Propositions

Offering Cloud Native Applications (Microservices) (s)

Traditional on premise software is not cloud ready, per se. This means, it is unable to meet the high expectations in terms of cost advantages, performance and scalability. Hence, it is mandatory to transform the architecture or to rebuild the whole application for cloud purposes. Specifically, SaaS services should be built as a system comprising modular microservices. These microservices are characterized by their ability to manage autonomously which and how much IaaS resources they require for their current operation. Deployed into an open source container, they can be directly run at different IaaS providers. In addition, these single modules can be combined according to the individual requirements of a specific customer.

Parallel Offering of Cloud and on Premise Applications (s)

Many customers have their core IT systems, e.g., manufacturing systems, internal and want to keep them as is over a longer period. This is mainly due to security, but also performance concerns. In total, the demand for on premise applications remains strong. Therefore, providers that have a high amount of clients within the on premise segment should in no case stop offering these systems and their support. Providing both SaaS and on premise applications is especially important for global players as they can thereby address the varying cloud adoption rates between countries.

Providing Adaptability of the Application on the Customer’s Side (s)

SaaS customers are by definition expected to give up their desire for individual adaptions. Nevertheless, clients want to be able to parameterize and configure the SaaS service to their requirements and flavor. Otherwise, the probability that the SaaS service is used is low. A SaaS service should consequently cover a broad spectrum of best practices for the clients to choose from. This must be possible without any programming skills as SaaS is usually utilized by business users. For customers, it is not tolerable that each small adaption requests a change project.

Achieving a High User Experience of Cloud Services (g)

The experts agreed that the topic of user experience is highly significant in the SaaS field. The handling of SaaS services must be self-explanatory and simple, to enable immediate use. Traditional concepts which have been utilized within large on premise software systems are less suitable. Customers desire an uncomplicated user experience, which they are accustomed to in their private environment. This has the additional advantage to minimize training course expenditures for customers’ employees.

Offering Private Cloud Deployment Models (s)

Currently, the customer demand is significantly higher for private than for public cloud deployment models. Many clients consider private clouds as an interim solution on their way to public cloud. Private clouds are preferred due to data protection, security, regulation and compliance reasons. In addition, a private cloud allows a higher degree of customization which customers often demand. Ultimately, it is mandatory for SaaS providers to include private cloud offerings in their portfolio.

Guaranteeing a High Availability (g)

It is absolutely imperative to ensure a high availability of the offered SaaS services. However, negative incidents in the past prove how challenging this undertaking is. A long-lasting service failure comes along with a significant loss of customer confidence, leading to adverse effects on the provider’s economic situation. Therefore, the specification and compliance of the service level agreements is very relevant.

Ensuring the SaaS Service to Run on All Leading Platforms (s)

A SaaS service has to be developed to be compatible with all leading IaaS/PaaS platforms. For this purpose, a SaaS provider has to be careful in utilizing proprietary PaaS services as these vary between providers. Otherwise, the SaaS service is tied to that specific PaaS platform making it hardly portable. Hence, the experts recommended to rely on standard open source services and to use platform-specific services only when absolutely necessary. This way, SaaS providers can ensure that their services are, if necessary with slight adaptions, compatible with most IaaS/PaaS platforms.

Offering Customer-Specific Service Individualization (g)

SaaS services are characterized by a high degree of standardization. Nonetheless, it is important to preserve a certain degree of flexibility. Some customers have more individual requirements that cannot be entirely met with the standard offering. Of course, each customization leads to additional costs. But customers are willing to pay for the considerable added value. Offering customer-specific adaptions is a way to differentiate from competitors. Smaller providers have an advantage as their organization structure is more flexible allowing them to address individual demands more easily.

Offering a Broad SaaS Service Portfolio (g)

The experts found offering a multitude of SaaS services important to succeed. The reasons for this are the following: First, a broad service portfolio achieves increased attractiveness of a provider as customers feel more prepared for the future. Second, customers always look for an answer to their individual requirements. By means of a broad portfolio a provider can better respond to that demand. Third, taking a broad approach offers a way to distinguish oneself from competitors.

Providing Extensive Customer Support (g)

Providing extensive support over the whole SaaS service lifecycle is essential. This includes support services related to selection and composition, as well as usage and operation of SaaS services. SaaS providers should commit to work very closely with customers. A lot of customers appreciate a personal contact partner and are willing to pay extra for qualitative support. They want to call if a problem should occur. The customer support process has to be integrated, meaning that customers should not have to talk to more than one employee to identify and solve a problem.

4.4 Customer Relationships and Channels

Offering Personal Sale Beside Self-service (g)

Some SaaS providers have the illusion that each type of cloud service can be sold by self-service without considering the size of the client company. However, the acceptance of self-services decreases from IaaS toward SaaS. A reason for this is that SaaS services often have to be integrated in the customers’ business processes and IT landscapes. Whereas small firms intensively use the self-service option, medium-sized and large enterprises commonly insist on personal contact to the provider combined with individual contract negotiations. Therefore, depending on the target group, it is mandatory to offer the additional possibility of engaging in negotiations.

Conducting Marketing Activities (g)

To stand out from the large number of SaaS providers, marketing is of great significance. In addition, due to the short contractual periods in the SaaS field, the customer royalty is decoupled. Further, the decision on the client’s side in favor of a specific provider is often not based on performance features. On the contrary, the provider’s image is a decisive factor. The importance of marketing is expected to grow further as the decision-making power moves increasingly towards business units. Examples for promising marketing channels are presentations at conferences, publication of articles, cooperation with universities, use of social media and buildup of communities. Moreover, the value of reference customers for newly developed services was emphasized.

Initial Explaining of the Cloud Computing Concept (s)

Many potential cloud customers have considerable doubts whether they should move into the cloud or not. This is especially the case for customers from German-speaking countries. The central concern is data security. Oftentimes the source of these doubts is a lack of knowledge. For SaaS providers, it is crucial to address these fears and unknowingness at the beginning of the customer relationship by explaining the general conditions of cloud computing in detail. This is a valuable contribution to establish trust.

Establishing a SaaS-Specific Incentive System for the Sales Division (s)

The traditional incentive system for the sales division which has worked well for on premise solutions is not directly transferable to SaaS. A long time it has been possible for a salesman to make deals at regular intervals within one customer relationship whereby he received his commissions regularly. In the case of SaaS, clients usually pay a subscription fee for a longer period of time. But from then on, no salesman is needed any more. By selling SaaS the salesman takes away his future up-selling options as customers get an all-inclusive deal. This makes it mandatory to develop a SaaS-specific incentive system to promote motivation within the sales division.

4.5 Key Activities

Utilizing Agile Software Development Models (s)

Traditional, sequential software development models such as the Waterfall Model [40] are not appropriate for the development of SaaS services. Instead, agile methods like Scrum [40] are substantial to realize short innovation cycles. Development speed is an important factor: new services or additional features have to be delivered continuously and fast on the platform in order to improve the portfolio. In the context of agile development, DevOps plays an important role, meaning there has to be a close connection between development and operation of any cloud service.

Conducting Research and Development (g)

The interviewed experts pointed out that research and development at a high level is a deciding factor in the rapidly changing cloud computing ecosystem. The risk of lagging behind in technological development is high, which is why many cloud providers are not only carrying out in-house research and development, but also acquire smaller cloud providers to increase their service portfolio and knowledge.

Involving Customers in the Development of New SaaS Services (g)

It is considered a serious mistake to develop a new SaaS service internally for an anonymous market. Instead, one should work closely together with customers. A promising source for new SaaS services is direct customer feedback. When customers request a feature or suggest a new SaaS service, it should be analyzed whether other customers might be interested, too. Another source for new SaaS services is to scale from customer-specific projects to other customers. Further, the establishment of customer workgroups for constant exchange with regard to customers’ problems and expectations and to forecast future trends is seen as a valuable asset.

4.6 Partner Network

Building Up a Partner Ecosystem (s)

A thriving partner ecosystem firstly serves as a sales and marketing channel: cloud services are highly scalable, but it is impractical to interact personally with each client as the sheer volume of sales staff for this cannot be supported. Hence, partners can act as resellers of a provider’s SaaS services. Thereby, a closer proximity to clients, further geographic coverage and outreach customer segments that do not exclusively fall into the provider’s scope can all be achieved. Secondly, a SaaS provider cannot perform all additional services outside his core business himself. These services include training and support which are fundamental to enable optimum use of a SaaS service.

4.7 Customer Segments

Focusing on Ambitious Medium-Sized and Large Companies (s)

The target customer segment is mainly determined by the type of application: small firms commonly search for small and isolated SaaS services, larger firms require applications to be well integrated in their processes. The experts recommended focusing on a specific segment. The more sophisticated medium-sized and large companies are considered as the most valuable one: (1) Providers can further sell on premise solutions. By focusing on small start-ups, one would miss this opportunity as they tend to obtain their whole IT from the cloud. (2) Many small firms lag behind in technological advances and business trends. (3) Larger firms have more financial resources.

Offering the Opportunity for Firms of All Sizes to Become a Customer (g)

Although it is primarily important to focus on a special target group, SaaS providers should also cater other possible clients. Small firms, in particular, should be given the opportunity to order a SaaS service and pay with credit card. Another option is to distribute special SaaS services for smaller companies through partner firms. To reach larger companies, a SaaS provider must mandatorily offer personal support as well. The overall target should be to reach as many customers as possible.

5 Discussion

This study revealed that several SaaS providers simply transfer their traditional on premise applications to a virtual server without modification and offer it as SaaS. However, this way, the requirements in terms of cost advantages, performance and scalability cannot be met. Instead, a SaaS application should be developed as a system comprising modular microservices that are characterized by the ability to autonomously manage their need of IaaS resources for operating a specific task. As a consequence, the main concern of IaaS providers will no longer be to solely offer a virtual server because it is becoming less common for customers to order a pre-defined amount of compute, storage and network resources. The current IaaS business model is therefore expected to change drastically in the near future.

Moreover, it turned out that demand for on premise solutions will remain strong over the next years. However, many providers have emerged on the basis of the cloud concept and thus do not have any existing on premise offering. On the other hand, a lot of established IT firms have just begun developing a cloud version of their applications. Thus, providers who are already able to offer applications for both worlds have a great advantage. A promising strategy is to develop each new application as a cloud native version because it must only be slightly adapted to be sold as on premise.

In addition, private clouds are currently very popular. They are mostly seen as an interim solution on the way to public clouds. The experts hence predict a significant rise in demand for public clouds in the coming years. Nonetheless, clients from certain sectors such as the public or the banking sector are expected to remain skeptical. Also, only a few firms are likely to transfer sensitive data into public clouds.

Furthermore, this study showed the major advantage of openness towards client-specific requests, e.g., regarding the pricing model or the design of the SaaS service. Related to that, ordering a SaaS service via self-service is hardly accepted particularly among medium-sized and large companies. Personal contact is preferred. However, addressing client-specific requests contradicts the SaaS providers’ goal of achieving economies of scale.

Concerning the business model development and innovation process as such, it became obvious that a large majority of the considered cloud providers does, contrary to the recommendation of scholars [15, 16], not undergo a systematic, phase-oriented process. It is rather conducted within occasional workshops in which new ideas regarding the business models are collected and then released for implementation.

6 Summary, Limitations and Outlook for Future Work

In order to investigate the success factors of SaaS providers’ business models, an exploratory multiple-case study was conducted. This resulted in a catalogue of 27 success factors. Most success factors relate to the value proposition, whereas the cost structure was not addressed by the experts. Besides being a promising starting point for further research, the results are particularly useful for practitioners. Established SaaS providers get a reference framework to compare, rethink and innovate their present business models. Companies that are planning to offer SaaS in future gain valuable insights that should directly feed into their business model design process.

The limitations of this study include the relatively small number of SaaS providers. However, through the integration of the insights of fourteen interviews with IaaS and PaaS experts who also gathered experience in the SaaS field, the data pool and hence, the generalizability of the findings was significantly increased. A second limitation may be the geographic scope being centered on Germany. But as the majority of the selected cases consisted of internationally operating providers, this influence is regarded as low. Despite the valuable results achieved, there remains a considerable need for further research: First, as not all derived success factors are of equal importance, their relevance has to be ranked. Second, it should be investigated to which extent the success factors are currently covered in practice. Third, the focus of this study was laid upon the isolated effects of individual success-driving business model characteristics. As a business model is a system comprising a set of components and the relationships between them, the combined effect of specific characteristics has to be examined. Fourth, as success factors may change over time, they have to be reassessed in future regarding their ongoing relevance. Thereby, researchers might use a given business model framework to overcome the unequal distribution of success factors. Finally, due to the applied research design, the direct comparability of the study’s results with the existing, isolated contributions on success factors of SaaS business models is only possible to a limited extent. Against this background, there is the necessity to conduct a systematic, comparative and integrative meta-study.