The Characteristics of Data-driven Business Model Development and how to Succeed
The experience of recent years shows that companies often overestimate the value of their own data and their own ability to generate revenue from it. In practice, very few companies can establish a sustainable business model based on data.
In our recently published book chapter Besonderheiten datenbasierter Geschäftsmodellentwicklung in the book Datenwirtschaft und Datentechnologie - wie aus Daten Wert entsteht (Springer, 2022), we go into more detail on this topic. You can access the book for free here.
The chapter, summarized in this blog post, describes the characteristics of data-driven business models, explains which companies are suitable for them, and illustrates the approach to developing them.
To start, we need to detail what a data-driven business model actually is. According to the definition in the Business Model Navigator, a business model provides a holistic picture of how a company creates and captures value (Gassmann et al. 2013). Therefore, data-driven business models can be understood as those models in which digitized data - in various degrees of processing - offers the central added value for customers or consumers.
To give you a better insight into these types of business models, we use the so-called magic triangle of business model innovation, as it provides a high-level overview that is more suitable for analysis. If you follow us closely, you should know it by now, as we use it to describe a business model in the four dimensions (see Fig. 1.1):
For data-driven business models, the value proposition always depends on the data. This is also why it is very important to assess the actual value of your data as early as possible (more about this later in this post). When companies fundamentally change their value proposition based on data, at least one, if not all, other dimensions are affected: to offer data or derived insights, companies usually need to change their processes (the how dimension). In addition, the way these new services are priced is often different (the value dimension). Last but not least, this new business model sometimes targets a new target audience (the who dimension).
The patterns and commonalities in data-driven business models
Business models are rarely completely unique, and research by Prof. Gassmann and Prof. Frankenberger has shown that the same patterns are often used in different industries. It makes sense to observe the essence of the business models of individual companies to recognize these patterns. This helps you to better understand how these work, and, more importantly, allows you to apply their logic to your own business model. In the area of data-driven business models, three different patterns can be fundamentally identified, which differ primarily in the degree of processing of the data and the value proposition that can be realized as a result:
1. Data as a Service (DaaS)
The best-known data-driven business model pattern is Data as a Service. In this model, a central organization collects data from users mostly through a digital platform and offers the raw data anonymized and unprocessed or only slightly processed for commercial purposes via a marketplace. The most common revenue model for this pattern is "pay per data", but "subscription" or "flat rate" models are also conceivable. Furthermore, companies that have a dominant role in an industry and can collect a high volume of data are more likely to be successful with the Data as a Service business model.
Example 1
Facebook, the social network platform, offers a wide variety of user data anonymously to third-party providers and software development companies.
Example 2
Snowflake is a cloud-based platform that unifies data silos for customers and makes corresponding data lakes available on demand. It provides a holistic solution for companies to share data with other units or use it for monetization.
Information as a Service (IaaS) - The Distribution of Analytics
This business model pattern is based on the sale of analyses or reports based on data, which can be self-collected or from third-parties. The customers can range from traditional companies to end consumers. To implement an IaaS model, primary analysis and visualization capabilities are indispensable, as well as industry know-how. Also, an IaaS business model can be structured similarly to a DaaS model, with customers paying for individual analyses or information.
Example 3
Google Maps is an example of an IaaS solution, since it offers a combination of real-time data and existing data, and charges customers per call or per order received via Google Maps.
Example 4
Celonis is a process mining service that offers companies the ability to retrieve process queries based on big data. The service is based on an IaaS model and offers a free version that collects registered information.
Answers as a Service (AaaS) - Concrete answers to questions
Answers as a Service (AaaS) is a data-driven business model that enables companies to provide answers to questions posed by customers. These answers help customers to make better decisions and enable further services and revenue sources based on the answers. The revenue model for the company can include a payment per answer or a subscription model for frequent use. The company can also generate revenue through consulting or implementation support. The AaaS model is particularly suitable for companies with close customer contact that have a deep knowledge of their clientele.
Example 5
Runtastic is a company that offers customized training programs based on body and movement data. The company also offers apparel and wearables to generate additional revenue.
Example 6
Schneider Electric's EcoStruxure Asset Advisor program enables companies in the industrial sector to detect any maintenance work on machines at an early stage. The program uses sensors on industrial equipment and delivers statistics and overviews of the condition of the equipment via cloud services.
But what are the opportunities and risks of these business models?
Regarding the opportunities, data-driven business models allow for rapid scaling (economies of scale) and the addition of new services (add-ons). For example, Airbnb uses customer data to tailor its search function to each customer and shows accommodations in neighborhoods where people also searching for the same city typically book. In addition, data-driven business models often use subscription and flatrate revenue models, where the customer pays a fixed rate for the service at a fixed time interval.
For the challenges, we would like to give you an industry example: Johnson & Johnson improved its data management system after discovering that 70% of its logistical data was incorrect. [2] The main challenge for data-based business models is to ensure high data quality - if this is not possible, the model can’t work. The case of Johnson & Johnson was a succes: data management was improved, regular quality measurements were introduced, and new workflows were defined. This new data management system reduced the error tolerance to zero. [2] To avoid false evaluations of generic data, it is important to know the context, to be aware of the limitations of the analysis, and not to blindly trust the meaningfulness of the data.
In the current environment, finding suitable employees to develop data-based business models is a challenge. Companies such as Google are attracting talent with high starting salaries. Also, data-based business models face the challenge of realizing the desired value proposition with an initially limited amount of data and therefore need to find intermediate solutions that are attractive enough for customers to continue to hand over data. Therefore, we see more and more service providers showing up, helping companies fill that skill gap externally.
You want to offer a data-driven business model?
A new business model is developed by addressing target groups outside the core market of the respective companies, implementing new value propositions and/or new payment models, and finally validating the new business model. As you have seen, moving to a data-driven business model usually touches nearly all of these dimensions. Since assumption-based business model development has proven itself in practice to systematically reduce the inherent uncertainty in five phases, it offers a verified way to also move forward with data-driven business models. At the end of the overall process, there is a validated new business model that can be launched on the (pilot) market (see Fig. 1.2).
Tip: It is important that the company obtains external technical expertise as early as possible to increase flexibility and build up know-how internally.
The starting point for assumption-based business model development is an initial concept for a business model. A proof of concept (PoC) should be performed to validate the value generated from the data at an early stage. This is where the process derives a bit from the usual approach, as a PoC can be needed early on to prove that the value proposition is achievable. Luckily, as knowledge for around data analyis increases, the basic feasibility of an idea can often be derived from an increasing amount of similar projects. Then, in a second phase, interviews with potential customers are used to validate whether the problem or need targeted by their business model exists and whether their business model really solves or satisfies it.
Tip: Be open to the fact that the business model can also develop away from data at this point. Only the problems and needs of the customers are decisive at this point, otherwise, a business model can easily be developed that will not be successful in the end.
The next step in testing and validation: the product-market-fit.
In this phase (and explained in more detail here) you need to validate whether the target group is interested in the specific business model in its entirety, including the mix of service, product and its characteristics. The goal is to design a product that meets a clear demand among the addressed target groups and delivers great value to the target customer. At this stage, it is crucial to understand your customers’ process in detail, as your data input is used in their context. Failure to input your solution into their process usually means failure for your whole model.
In order to implement the value proposition properly, it is also crucial to select the right partners. This can be achieved by finding a suitable implementation partner that complements the company's own capabilities accordingly. The partner should be regarded as a strategic partner who can accompany the further development of the company in the best case. This can range from training in the relevant interface skills to setting up in-house digital departments.
Tip: A partner should be consulted to assess whether the envisaged value proposition is technically feasible and to ensure that data protection regulations are respected. Also, given that he has UI/UX competencies, he can take over interviews and validations, but make sure that the general approval of the product is also reviewed.
Now let’s talk money – how do we check for willingness to pay?
When a solution is mature enough, it can be clearly communicated what exactly is being offered. This allows a reliable assessment of willingness to pay to be made and allows the company to receive more honest and manageable feedback. A good way to give it a little more strength is a letter of intent – a contract that spells out the key aspects of a solution and the customer's promise to purchase it. It is used to obtain written agreement from a potential customer without them directly purchasing the product. For digital components, it can be beneficial to immediately agree on a demo/beta or pre-release usage. This establishes a concrete next step that clearly involves the potential customer, since the customer must make resources available for use. After this, a detailed business case must be developed or detailed, paired with detailed market estimations.
And what about the value delivery?
Data-based business models require a different technological architecture than product-oriented business models, and may require adapting sales structures or even establishing an alternative sales channel. At this point, the necessary technical capabilities should be built up internally, the business model should fit strategically with the company's orientation, and feedback from the market should clearly show that there is interest in the solution. To find qualified employees with technical skills such as data analytics, cloud engineering and software system architectures, companies should choose a trustworthy partner.
Don’t forget to scale and measure your model correctly!
In the last step before the (pilot) market entry, it is determined how the success of the business model can look like and be measured at all. For this, it is crucial to define meaningful key performance indicators (KPIs) that can measure the goals of the new business model. This new business model should measure learning success and growth, rather than profit and efficiency. Also, close feedback from the users is crucial to validate the new business model.
.svg)
.svg)
Conclusion: data-driven business models are not for everyone - but if they are, the benefits are big.
Data-based business models offer particular incentives for companies, but the skills required to handle the data correctly are not as easy to overcome as is often assumed. This gives technical implementation partners increased and therefore strategic importance, and selecting the correct partners becomes a critical success factor. The leap from correct handling of data to profitable use in new business models is another hurdle. A systematic approach to validation is therefore indispensable, and an early proof of concept is particularly important. Only when companies are aware of both the technical and business model-related changes and actively manage them, can they benefit from the opportunities offered by data-based business models.
Getting more in-depth views on this topic? Read our chapter Besonderheiten datenbasierter Geschäftsmodellentwicklung (by Richard Stechow, Leonie Schäfer & Peter Brugger) in the book Datenwirtschaft und Datentechnologie - wie aus Daten Wert entsteht (Springer, 2022).
Open Access link: https://link.springer.com/book/10.1007/978-3-662-65232-9
[1] O. Gassmann, K. Frankenberger und M. Csik, Geschäftsmodelle entwickeln: 55 innovative Konzepte mit dem St. Galler Business Model Navigator, München: Hanser, 2013.
[2] B. Otto, Stammdatenqualität: Das Rückgrat moderner logistischer Systeme, Fraunhofer Institut für Materialfluss und Logistik IML Dortmund, 2014.
We as the BMI Lab prepare companies for tomorrow by helping them to develop innovation strategies, design new business models, and build up internal innovation capabilities for long-term success. Do you need a sparring partner for your business model innovation journey? With our proven methodology, broad tool set, and comprehensive experience, we will master the challenges together.
Leonie Schäfer
.svg)
We share what shapes our work and creates pivotal seconds. Mostly through our scientific and entrepreneurial perspective.
.webp)
Science, entrepreneurship or consulting? Every innovation journey starts with a conversation. Pick the perspective that resonates - and let’s begin.
