You are on page 1of 20

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/329529109

Berlin Start-ups - The Rise of Data-Driven Business Models

Conference Paper · December 2018

CITATIONS READS

6 2,064

3 authors:

Romy Hilbig Bennet Etsiwah


ESO Education Group, Germany Weizenbaum Institute for the Networked Society
16 PUBLICATIONS   68 CITATIONS    4 PUBLICATIONS   7 CITATIONS   

SEE PROFILE SEE PROFILE

Stefanie Hecht
Fraunhofer Institute for Open Communication Systems FOKUS
12 PUBLICATIONS   16 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Weizenbaum-Institute for the networked Society - The German Internet Insitute View project

Weizenbaum Institut View project

All content following this page was uploaded by Romy Hilbig on 10 December 2018.

The user has requested enhancement of the downloaded file.


This paper was presented at ISPIM Connects Fukuoka – Building on Innovation Tradition,
Fukuoka, Japan on 2-5 December 2018. The publication is available to ISPIM members at
www.ispim.org.

Berlin Start-ups – The Rise of Data-Driven Business


Models

Dr. Romy Hilbig*


University of Arts / Weizenbaum-Institute, Hardenbergstraße 32,
Berlin, 10623, Germany.
E-mail: r.hilbig@udk-berlin.de

Bennet Etsiwah
University of Arts / Weizenbaum-Institute, Hardenbergstraße 32,
Berlin, 10623, Germany.
E-mail: etsiwah@udk-berlin.de

Stefanie Hecht
University of Arts / Weizenbaum-Institute, Hardenbergstraße 32,
Berlin, 10623, Germany.
E-mail: s.hecht@udk-berlin.de

* Corresponding author

Abstract: With the advent of big data technologies such as Cloud Computing, IoT, and
AI data-driven business models have become a new unit of analysis in business model
research. Tracing back to different roots in scholarly business model literature,
researchers have begun to create a new body of work while trying to capture the essence
of data-driven businesses. This article provides an overview of this emerging field,
combining a systematic literature review with a qualitative desk research of recently
founded data-driven and data-enhanced start-ups. To contribute to the discussion, the
paper proposes a definition and a taxonomy of data-driven business models. It then
proceeds to use the taxonomy in order to analyze the business models of 50 Berlin-based
start-ups, focusing on three economic sectors software as a service, fintech, and
education.
Keywords: data-driven business model; data-enhanced business model; data-
driven business model innovation; business model innovation; digitization; big
data; data analytics; start-ups; servitization, deep learning business model

1 Introduction
Over the past years, the computer-enabled production and gathering of digital data has
led to an increased use of data processing techniques. Today, digital enablers such as
Cloud Computing, IoT, Big Data, Automation, and Artificial Intelligence are essential
drivers for a new frontier of business models (BM), the so-called data-driven business
model (DDBM) (Hartmann et al., 2016; Hunke et al., 2017; Luque et al., 2017; Schüritz

1
This paper was presented at ISPIM Connects Fukuoka – Building on Innovation Tradition,
Fukuoka, Japan on 2-5 December 2018. The publication is available to ISPIM members at
www.ispim.org.

et al., 2017). This new class of BMs changes the way organizations communicate,
deliver, create, and capture value, often based on innovative value propositions.
However, “data itself is useless” (Dremel et al., 2017, p.1142) if there is no sense-
making, analysis or interpretation of the generated data. Thus, both the dimension of
value creation and the means of data exploitation are important to understand and
establish a DDBM.
Over 20 years ago, the BM itself became a new unit of analysis and shaped an entire
research field (Massa et al., 2017; Osterwalder et al., 2005; Schneider and Spieth, 2013;
Zott and Amit, 2010). The current stream of literature incorporates various definitions of
the BM concept and, building on that, a multitude of concepts for business model
innovation (BMI) (Casadesus-Masanell and Ricart, 2010; Osterwalder and Pigneur, 2010;
Teece, 2010).
With the advent of digital technologies, a new frontier of BMs arises, the data-driven
business model (DDBM). Both new and established organizations are capturing new
market opportunities by innovating their BMs based on data analytics. In line with these
developments, DDBMs have become a new unit of analysis for BM researchers. In this
paper we aim to contribute to this rather young discussion (Hartmann et al., 2016;
Schüritz et al., 2017; Schüritz and Satzger, 2016). This paper investigates a definition and
a taxonomy of DDBMs based on an in-depth literature review. Furthermore, the proposed
taxonomy will then be used to analyze 50 cases of recently founded start-ups that are
based in Berlin. Today, Berlin is the second largest hotspot for start-ups in Europe (Ernst
& Young GmbH Wirtschaftsprüfungsgesellschaft, 2018b). The entrepreneurial spirit in
the German capital has already been credited for the origination of famous digital and
data related start-ups such as Zalando, Soundcloud, and Delivery Hero. Moreover, many
more start-ups have been established over the last years that are exploiting digital data to
generate new DDBMs. We are thus focusing on the following two research questions:
RQ1: What is a data-driven business model?
RQ2: Which types of data-driven business models have developed in the Berlin start-
up scene in the economic sectors fintech, software as a service, and education?
Building on an overview of the recent BM discourse (section 2) we propose a definition
and a taxonomy of DDBMs based on an in-depth literature review (section 3 and 4). The
proposed taxonomy will then be used to analyze and classify a random sample of 50
recently founded start-ups that are based in Berlin, Germany (section 4). As a result, we
will identify and classify DDBMs in the three economic sectors including education,
fintech and software as a service (SaaS). Finally, we summarize our research results and
highlight further research questions related to DDBMs in section 5.

2
This paper was presented at ISPIM Connects Fukuoka – Building on Innovation Tradition,
Fukuoka, Japan on 2-5 December 2018. The publication is available to ISPIM members at
www.ispim.org.

2 Literature (Data-Driven) Business Models and Innovation


In this paper we draw on the literature about BMs and BMI. There is a fair body of
research in both of these areas. A fundamental debate in BM literature still revolves
around the definition of BMs as a concept. Thus, throughout the decades many
definitions have been introduced by researchers (Abdelkafi et al., 2018; Abdelrahim et
al., 2013; Amit and Zott, 2001; Casadesus-Masanell and Ricart, 2010; Johnson, 2010;
Massa et al., 2017; Osterwalder et al., 2005; Schneider and Spieth, 2013; Teece, 2010).
However, while the literature on business models provides a multitude of competing
definitions, many understandings agree on three value-based core dimensions as in value
delivery, value creation, and value capture (Täuscher, 2018; Teece, 2010). According to
Massa et al. (2017), different business model concepts can further be distinguished by
their underlying interpretations, resulting in three distinctive categories. While two of
these categories are based on either formal or informal simplifications of real systems, a
third one supports the identification of BM archetypes by representing real attributes of
BM systems which is what we aim to do within this study. A BM concept that builds on
both value dimensions and BM archetypes has been developed by Gassmann et al.
(2017). Their BM framework, visually represented as a triangular template aims to
represent and distinguish BM patterns in order to facilitate processes of BMI (Gassmann
et al., 2017; Täuscher, 2018). In section 4 we describe how we utilized this BM concept
in order to cluster our data on the Berlin start-up scene.
In the course of digitization new BMs are emerging which in turn impacts BM research
accordingly. New concepts such as digital platforms introduce new terminologies and
thus new units of investigation (Cenamor et al., 2017; Parker et al., 2016). In line with
these developments a new BM concept has emerged, the DDBM. Looking at the
mainstream of business model research (Foss and Saebi, 2017; Massa et al., 2017;
Schneider and Spieth, 2013), there are research gaps in regard to definitions of DDBMs
as well as their relevance for organizational performance and the potential of BMI
through data. One of the first working papers we identified that focusses solely on
DDBMs was published by Hartmann et. al (2014). Here, the authors introduce a DDBM
framework which consists of six dimensions related to other common BM frameworks
(Gassmann et al., 2017; Johnson, 2010; Osterwalder and Pigneur, 2010): key resources,
key activities, value proposition, customer segment, revenue model/cost structure
(Hartmann et. 2014, 2016). It turns out that within the DDBM framework the key
resource is data, which can be generated either internally or externally. Moreover, the
core activities are described such as data generation, data processing, visualization, and
interpretation of the collected data (Hartmann et al., 2016, 2014). In a paper with a more
practical focus Bulgur et al. (2014, p.30) introduce the term “business models for big
data”. They identify different data sources and the monetization of data. Furthermore,
Sadovsky et al. (2014) highlight the importance of big data and data analysis for the
emergence of DDBMs. Overall, the scientific discourse on DDBM highlights the
connection to big data (Bulger et al., 2014; Hartmann et al., 2016, 2014; Sadovskyi et al.,
2014).
Gartner (2012), a technology research firm, proposes one of the most commonly cited
definitions for big data by describing it as “high-volume, high-velocity and high-variety
information assets that demand cost-effective, innovative forms of information processing
for enhanced insight and decision marking”. Here, the attribute “big” refers to the three
dimensions volume, velocity (the speed at which data is generated and transferred) and
variety (the bandwidth of data types and sources). This definition can be extended by two
additional V’s – value and veracity – representing the concepts of entrepreneurial value
and the assurance of data quality (Bitkom, 2015; Eckert et al., 2014).These 3-5V

3
This paper was presented at ISPIM Connects Fukuoka – Building on Innovation Tradition,
Fukuoka, Japan on 2-5 December 2018. The publication is available to ISPIM members at
www.ispim.org.

characteristics are fundamental for our analysis in this paper and the development of
DDBMs. Still, their specifications may vary depending on the BM. Due to the
exponential growth of data through digitization, the importance of generating, analyzing
and interpreting data for establishing new BMs or improving existing ones, rises in
economy and research alike (Brownlow et al., 2015). Data only becomes information
when a meaning is assigned to it, i.e. it is aggregated, processed and finally interpreted so
that it can be connected intersubjectively. Data needs a semantic context in order to be
understood as a meaningful symbol for a given object of observation, eventually
generating information and knowledge (Bodendorf, 2006).
Finally, digitization and the increasing use of digital enablers such as Cloud Computing,
IoT, and AI (Luque et al., 2017; Schüritz et al., 2017) that constitute big data can have
two major impacts in the organizational context. They can either facilitate the
incremental optimization of current business processes or even lead to the establishment
of new BMs with products and services innovated by the use of data. As a result,
researchers and practitioners today discuss the occurrence of new BM patterns such as
Data-as-a-Service, Analytics-as-a-Service or Information-as-a-Service (Bitkom, 2015;
Hartmann et al., 2016, 2014). In order to combine these findings and to make the DDBM
concept more tangible we provide a definition and a taxonomy in section 4.

3 Research Design

Overview
The research design is based on two main steps. First, we provide an in-depth systematic
literature review in order to identify definitions of the DDBM concept. The aim of the
literature review is to understand what a DDBM is, following our research question RQ1.
Based on a comparison of current definitions on DDBMs, we introduce a new taxonomy
and propose a new definition of DDBMs that divides current business model types into
three main categories: pure data-driven business models, data-enhanced business models,
and low data business models.
In a second step, we conduct a qualitative desk research, combining our taxonomy with
practical examples from the Berlin start-up scene. Desk Research is an important
instrument of qualitative research and makes material accessible that is not primarily
collected but draws on secondary sources such as on- and offline texts, films, audio files
or objects (Mayring, 2010, 2002). On the internet a lot of information about organizations
is available through websites, databases, social media channels or press articles, allowing
an initial mapping of BMs through document analysis.
In total, we analyzed a random sample of 50 Berlin-based start-ups that were founded
between 2016 - 2018 by exploring three online databases, their websites and online press
articles.

4
This paper was presented at ISPIM Connects Fukuoka – Building on Innovation Tradition,
Fukuoka, Japan on 2-5 December 2018. The publication is available to ISPIM members at
www.ispim.org.

Systematic Literature Review on DDBM


The initial search for the literature review was conducted in May 2018. Based on a list of
20 keywords, we used the Science Direct database 1 to discover any publication type
published between 2013 and 2018 that contained at least one of the keywords in full text
or metadata. The keywords ranged from directly related search terms (e.g. “Data-Driven
Business Model”) and synonyms (e.g. “Data-Based Business Model”) to terms from
related fields (e.g. “Digital Platform”). This led to an initial collection of 1765
publications, including duplicates (see figure 1).
In order to clean the corpus and focus on the most recent publications, we proceeded to
remove all duplicates and focused solely on peer-reviewed academic journal papers that
were published between 2017 and 2018. We have chosen this period to identify the latest
and most up-to-date papers. This reduced the number of results to 817 articles.
In the next step, three researchers from the team performed a manual screening of the
remaining articles and indexed them by keyword in regard to three categories – relevance
for our research question(s), research method used, and explored industry. These
categories were defined prior to the screening process. Out of the 169 articles that were
deemed relevant after the screening, 25 focused on definitions for at least one or more of
the following concepts: Business Model (BM), Digital Business Model (DBM), Business
Model Innovation (BMI), Digital Business Model Innovation (DBMI).
However, none of the papers provided a substantial definition that was exclusive for
DDBMs. We thus decided to conduct an additional search in a different database with a
focus on conference papers published between 2016 - 2018, as they tend to be published
faster than regular peer-reviewed journal articles. Using the IEEE database 2, we were
able to identify 3 conference papers that were added to the corpus. In a final step, we
performed a backward and forward reference search which led to the identification of
another 5 papers that were, again, added to the corpus, leading to a total count of 8
publications in our second search that provided definitions for DDBMs (see figure 1;
table 1).

1 https://www.sciencedirect.com/ (accessed May 2018)


2 https://ieeexplore.ieee.org/Xplore/home.jsp (accessed July 2018)

5
This paper was presented at ISPIM Connects Fukuoka – Building on Innovation Tradition,
Fukuoka, Japan on 2-5 December 2018. The publication is available to ISPIM members at
www.ispim.org.

Figure 1 Literature Review: Keywords and Process

Desk Research Berlin Start-ups


The second step of our analysis focuses on the identification of DDBMs in the Berlin
start-up scene in three different sectors, using qualitative desk research (see figure 2).
To gather data about German start-ups we identified three relevant online databases
including Bundesverband Deutscher Start-ups e.V. (734 data sets) 1, deutsche-start-ups.de
(1078 data sets) 2, and gruenderszene.de (2550 data sets) 3.
Two web scraping tools were used to download the data and a refine tool was used to
clean the corpus. Each data set consists of a number of attributes such as website, legal
form, year of foundation, foundation place, number of employees, short description etc.
During the automated refinement process as well as the qualitative analysis of the
individual sectors, multiple duplicates were identified and eliminated accordingly.

1 https://deutschestartups.org (accessed July – August 2018)


2 http://deutsche-start-ups.de (accessed July – August 2018)
3 https://www.gruenderszene.de (accessed July – August 2018)

6
This paper was presented at ISPIM Connects Fukuoka – Building on Innovation Tradition,
Fukuoka, Japan on 2-5 December 2018. The publication is available to ISPIM members at
www.ispim.org.

The final corpus contained of 4362 data sets that were classified by economic sectors,
following the definition of individual sectors as provided by Bundesverband Deutsche
Start-ups e.V. – a German federal association for start-up promotion. Data sets of start-
ups that had not already been assigned to a sector in the original database were assigned
manually by the research team. For better accuracy, additional categories for economic
sectors have been added to the classifications of the Bundesverband für deutsche Start-
ups e.V. such as travel and mobility. If categorizations were unclear, two researchers
evaluated the respective cases independently and reached an agreement later on in a joint
discussion.
In a second step, we focused solely on data sets of start-ups in Berlin with the aim of
investigating the Berlin start-up scene. According to the “EY Start-up Barometer“ (Ernst
& Young GmbH Wirtschaftsprüfungsgesellschaft, 2018a, 2018b), in 2017 Berlin was
once again the hotspot of the German start-up scene with local start-ups raising 2,969
million euros in investments. This puts Berlin in first place in a national comparison,
ahead of Bavaria (407 million euros) and Hamburg (230 million euros). In a European
comparison, Berlin falls behind London in 2017 (4,878 million euros), but replaces Paris
(1,973 million euros) in second place. In the third step, we conducted a qualitative
analysis of all start-ups that were founded in Berlin within the last three years (2016-
2018). 169 start-ups could be identified in this way.
We proceeded to selected three sectors that promised to be particularly relevant for
DDBMs and our future research: software as a service (n=29), fintech (n=14), and
education (n=7) start-ups, resulting in a data set of 50 start-ups.
The first two sectors were selected because Berlin is making high investments in these
areas (Ernst & Young GmbH Wirtschaftsprüfungsgesellschaft, 2018a). The education
sector was selected in order to gather insights on the impact of learning analytics and new
digital and data-based learning approaches which is becoming increasingly relevant for
future research due to changes in work processes (work 4.0) in companies. In a desk
research we reviewed the start-up’s websites as well as articles, and press interviews to
gather information about the business relation they are addressing (e.g. B2B or B2C),
underlying BM patterns for the revenue stream, and the data used as a key activity
(Gassmann et al., 2017; Hartmann et al., 2016, 2014). Furthermore, the start-ups were
clustered along our proposed taxonomy consisting of pure data-driven business model,
data-enhanced business model or low data business model (see figure 2). In order to
ensure the data triangulation of the desk research analysis the business models of the
start-ups were examined by at least two researchers (Denzin, 2017; Flick, 2011) . The
results of the analysis will be presented in the next chapter (see figure 4 and 5).

7
This paper was presented at ISPIM Connects Fukuoka – Building on Innovation Tradition,
Fukuoka, Japan on 2-5 December 2018. The publication is available to ISPIM members at
www.ispim.org.

Figure 2 Desk Research: Process and Focus

4 Research Findings and Results

Literature Review and Definition of DDBM


In the first part of our literature review we identified 25 papers with a focus on
definitions for concepts such as BM, DBM, BMI, and DBMI. While none of these
papers provided a definition of DDBMs some mentioned data or data analytics as key
process to create e.g. DBMs. For instance, Krämer and Wohlfarth (2018) indicate a
potential of DBMs by using data analytics for targeted advertisement. In the context of
industry 4.0, the concept digitized manufacturing and a model for data-driven
manufacturing get introduced, referring to predictive tools or service improvements
based on data analysis (Bokrantz et al., 2017; Paritala et al., 2017; Weber et al., 2017).

8
This paper was presented at ISPIM Connects Fukuoka – Building on Innovation Tradition,
Fukuoka, Japan on 2-5 December 2018. The publication is available to ISPIM members at
www.ispim.org.

The researchers and their teams referred to new business models in order to describe the
emergence of new BMs based on the implementation of data analytics and similar
digital technologies such as Cloud Computing, intelligent senorsisation, and embedded
systems (Ibarra et al., 2018; Luque et al., 2017; Vendrell-Herrero et al., 2017).However,
the initial purpose of the literature analysis was to understand how current research
defines DDBMs. Because of that, we extended our literature review (see chapter 3) with
conference papers and the backward and forward reference search method to detect
more exclusive definitions on DDBM (see figure 1). The final results of this second
search are presented in table 1.
However, table presents that Hartman et al. (2016) as well as other authors identify data
as key resource for any DDBM (Schüritz et al., 2017; Zolnowski et al., 2016). Exner et
al. (2017) focus on the role of data integration for core products, technologies or digital
platforms, thereby contributing to the idea of data-driven value creation consisting of
three core processes: data acquisition, analytics and visualization. Additionally, Benta et
al. (2017) highlight the use of user data in order to enrich products or services which
ultimately creates a new value proposition while Zolnowski et al. (2017) consider data
more as an enabler to create new and innovative services based on sensor or IoT
technologies. Exploiting big data and analytics seems to be the key process for data-
driven business model innovation (DDBMI) (Hartmann et al., 2016; Hunke et al.,
2017). Hartmann et al. (2016, p. 1385) observe two main forms of BMI – incremental
and radical in nature – as a result of data analytics: “first, (big) data is used for the
incremental improvement and optimization of current business practices, processes and
services. Second, new products and business models can be innovated based on data
use“. Schüritz & Satzger (2016, p. 135) introduce the term data-infused business
models as a means to define existing BMs that evolve by incrementally implementing
digital data and data analytics: “[E]very existing business model in the market offering
products or services will need to be infused with data at some point in order to stay or
become competitive. Every business will sooner or later incorporate some amount of
data - and analytics on top of it - into its core business model which then may gradually
become more “data driven”. Zolnowski et. al. (2017, p.3) describe BMI through data
analytics as data-driven innovation that “include[s] the acquisition of data, its
subsequent aggregation, the analysis of data and possible automated assignments, and
actions that are triggered “. In addition to that, in a paper from our initial search that
provides a literature review on big data, Günther et al. (2017) identify, data and data
analytics as potential element for BMI “data being a potential key resource of
organizations’ business models […]Organizations also innovate their business models
when big data leads them to inter alia develop whole new value propositions, target
different customers, or interact with customers in different ways” (Günther et al., 2017,
p. 197). In summary, all these definitions demonstrate that data is the key resource of
DDBMs. Furthermore, the key processes of data aggregation, data analysis, data
interpretation, and visualization are understood as means to optimize existing business
models or create new ones.

9
This paper was presented at ISPIM Connects Fukuoka – Building on Innovation Tradition,
Fukuoka, Japan on 2-5 December 2018. The publication is available to ISPIM members at
www.ispim.org.

Table 1 Definitions of DDBMs

Author Definition

Exner et al. (2017, p.149) “The data-driven business model needs to fulfil
application oriented requirements and content-
related requirements [...] The main part is described
as Data-driven value creation and represents the core
processes, resources, abilities and partners to enable
the individual customer solution. The processes focus
on data processes including phases such as data
acquisition, data analysis and visualization as it will
be needed for the customer solution. The resources add
the core products, digital platforms or other
technical infrastructure.”
Hunke et al. (2017, p.150) “Nevertheless, exploiting data as a key resource for
the business model innovation is gaining moment.
Hence, leveraging data opens up a continuum of
transformation opportunities, from optimizing internal
processes to developing entirely new, data-driven
business models.”
Schüritz et al. (2017, p.5349) “As the central feature of these business models is that
data is considered to be their key resource […]”

Zolnowksi et al. (2017, p.183-184) "Within data-driven business models, data act as
enabler for the development of innovative
services“[…]„we consider the data-driven business
model that might consume data from sensors and the
Internet of Things as well.“
Benta et al. (2017, p.350) „Big Data Analysis is required for realizing a data-
driven Business Model, because the insights from
gathered data are the main resource for the value
proposition which is derived from data “[…]"Data-
driven business model shows the potential of
implementing collected user data from the
companies’ products and services as well as the
needed data resources for offering data-driven value
proposition.“
Zolnowski et al. (2016, p.2) „When data are exploited as main resource for
innovative service business models, they are called
data-driven business models. “
Schüritz & Satzger (2016, p.153) “Other authors claim that data and analytics bring to
bear entirely new “data-based” or “data-driven”
business models and define them broadly by stating
that every business that uses data as key resources
can be considered applying a data-driven business
model [...]”
Hartmann et al. (2016, p.6) “[…] business model relying on data as a key
resource […]”

10
This paper was presented at ISPIM Connects Fukuoka – Building on Innovation Tradition,
Fukuoka, Japan on 2-5 December 2018. The publication is available to ISPIM members at
www.ispim.org.

In order to structure our findings, we used the literature corpus from our second search in
the IEEE database to extract key concepts and definitions revolving around our first
research question (RQ1). Based on our findings, we developed three categories that help
distinguish between different BMs according to their usage of data. Our taxonomy builds
on two different variables: the degree of data exploitation within an organization (low to
high), and the degree of digitization of an organization (low to high). The latter variable
indicates if and how key processes of the BM are enriched or optimized by means of new
technologies such as cloud computing, IoT or AI (see figure 3). Furthermore, these
variables are based on the two notions from our literature review regarding the key
resources and key processes of DDBMs.

Figure 3 DDBM Taxonomy

Our first category, the low data business model (LDBM), is characterized by a low level
of data exploitation as well as a low degree of process digitization. We define it as
follows:
Companies with low data business models don’t or just barely depend on digital
technologies and thus don’t or hardly produce any digital data to exploit. The
core of their business model is thus analogue and digitization, if detectable at
all, is restricted to isolated parts of the value chain.
Examples for LDBMs are daily and local services such as shoemakers, hairdresser or
bakeries. Their value proposition is a physical product or service that is not embedded in
the digital world. Most of their BM dimensions are not enhanced by digital data
exploitation but their value creation processes are starting to implement digital tools in
order to improve the value chain in terms of logistics, production or marketing.

11
This paper was presented at ISPIM Connects Fukuoka – Building on Innovation Tradition,
Fukuoka, Japan on 2-5 December 2018. The publication is available to ISPIM members at
www.ispim.org.

The second category builds on the work of (Schüritz et al., 2017) who propose different
patterns on data-infused BMs. After scanning the literature and the Berlin start-up scene
we are convinced that these types of BMs are rather data-enhanced by using digital
technologies which is why we call this category data-enhanced business model
(DEBM) and define it as follows:
Companies with data-enhanced business models offer physical products and/or
services while enhancing given aspects of their business model by means of
digital technologies and the exploitation of digital data in order to generate
competitive advantages.
For example, the bike sharing provider Nextbike 1 offers bikes and a digital rental system,
all combined in a mobility app. Their app enables users to geographically track free bikes
in their vicinity and rent them. Here, a physical product gets enhanced by using digital
technologies and internal data exploitation.
Finally, the third category is the pure data-driven business model (PDDBM):
A pure data-driven business model uses data as a key resource to generate any
type of digital services by means of key processes such as data aggregation,
data generation, data analytics, data exchange, data processing, data
interpretation, data distribution and data visualization in order to create value
for customers, users or stakeholders and to capture revenue. PDDBMs always
rely on digital data and digital technology as fundamental enablers.
Examples of PDDBMs are digital platforms such as Airbnb, Netflix or Uber. They offer
digital services such as renting a flat, video streaming or transportation by using digital
technologies that are able to aggregate, analyze, interpret or visualize data to create new
values for its customers. These BMs would not be possible without digitization and the
exploitation of big data. However, there might be a next level of PDDBMs, the so-called
deep-learning business model that uses data to create a self-learning, smart BM based on
AI techniques without human interaction to fulfill customer needs.
On a second layer, figure 3 presents how the exploitation of data can lead to BMI.
Businesses can grow from a LDBM to DEBM to PDDBM by integrating digital
technologies that enable them to generate, analyze or interpret data within their respective
BM or to create completely new ones. It follows, that the RQ1 “What is a data-driven
business model?” can be answered with the presented arguments drawn from scientific
literature as well as the creation of the three categories on DDBMs and the derived
definitions.

Berlin Start-ups – The Rise of DDBMs


Following our second research question, RQ2, we analyze DDBMs in the Berlin start-up
scene. To do so, a random sample of 50 start-ups, recently founded between 2016-2018
from three industries software as a service, fintech and education has been analyzed.
Their BM patterns were analyzed based on the triangular BM framework developed by
Gassmann et al. (2017). In addition, we based our analysis on a taxonomy of DDBM key
activities as proposed by Hartmann et al. (2014, 2016) and which became an essential

1 https://www.nextbike.de/en/ (accessed 06.11.18)

12
This paper was presented at ISPIM Connects Fukuoka – Building on Innovation Tradition,
Fukuoka, Japan on 2-5 December 2018. The publication is available to ISPIM members at
www.ispim.org.

part of our introduced definition of PDDBM: data generation, data acquisition,


processing, aggregation, analytics, visualization and distribution. Finally, we assigned
each start-up from our data pool to one of the categories of our own DDBM taxonomy:
LDBM, DEBM and PDDBM.
Start-ups in the segment software as a service focused primarily on the B2B market
(76%; n=22). However, there are also start-ups that address a two-sided market: B2B and
B2C (17%; n=5). The analysis of the sector fintech is almost similar except that there are
additional business relations involved such as B2B2X, referring to white label products
and services. For example, the start-up ELEMENT 1 (*2016), a BaFin licensed risk
carrier, offers insurance as well as complete white label or API based modular products.
Portal operators (B2B) can thus expand their service portfolio and offer their customers
(B2BX) insurance products at the point of sale with the look and feel of their own brand.
In comparison, start-ups in the educational sector have significantly more B2C relations
by percentage (43%, n=3), and often represent a two-sided market by offering
educational platforms (57%, n=4).
Looking at the category of data key activities, we could identify differences between the
sectors. For example, start-ups in the sector education focus primarily on the distribution
of data and information (71,4%; n=5). The start-up Stackfuel 2 (*2017) teaches companies
and their employees data skills such as machine learning. With the help of online training
in all relevant data technologies and classroom training, they distribute their information
and knowledge as part of a blended learning program. Organizations from the sector
software as a service are mainly concerned with data acquisition (13,8%; n=4),
processing (17,2%; n=5), and analytics(58,6%; n=17). A start-up called i2x 3 (*2017)
wants to use machine learning to automatically analyze telephone calls in order to
improve customer communication. During or immediately after calls, users receive
personalized feedback on the conversation. The software recommends company-specific
terminology while highlighting personal habits such as filler words. The voice
recognition technology intends to make telephone data transparent and usable. The
fintech sector focusses especially on analytics (57,1%; n=8), followed by distribution
(21,4%; n=3) and processing (14,2%; n=2), as shown in figure 4. For example, the start-
up PAIR Finance 4 wants to sensitize customers to outstanding payments with behavior-
oriented, data-driven strategies and a combination of machine-learning procedures within
the context of order management. To this end, customers are contacted via various
communication channels with tailor-made content in order to increase the realization rate
and customer satisfaction.
Reflecting these findings on the basis of the work of Gassmann et al. (2017) on
distinctive BM patterns related to the revenue streams, in fintech the most commonly
used revenue pattern is performance-based contracting (n=6) while in education it is
rather pay per use (n=3). Revenue patterns of start-ups in the software as a service
industry are very diverse including freemium (n=2), subscription (n=7), and pay per use
(n=7) to performance based contracting (n=2) and rent instead of buy (n=2). Start-ups
often combine these revenue patterns with each other. The analysis of our data supports
the research of Schüritz et al. (2017) who studied the revenue models of data-driven

1 https://www.element.in/de/ (accessed 30.10.18)


2 https://stackfuel.com/ (accessed 05.11.18)
3 https://i2x.ai/de/ (accessed 05.11.18)
4 https://www.pairfinance.com/ (accessed 05.11.18)

13
This paper was presented at ISPIM Connects Fukuoka – Building on Innovation Tradition,
Fukuoka, Japan on 2-5 December 2018. The publication is available to ISPIM members at
www.ispim.org.

services start-ups and identified well known revenues models such as subscription, pay-
per-use, freemium or rent instead of buy. We came up with the same results in our sample
and could not identify any additional revenue models with the desk research analysis, yet.
Further research is needed to understand how revenue models might change if data
becomes the key resource within a business model and if “pay per data” might be a new
equated business model pattern.

Figure 4 Key activities sorted by sector

In addition, figure 5 presents that 39 out of 50 start-ups can be classified as PDDBM. 10


Start-ups can be classified as DEBM and none as LDBM. In the fintech sector only one
DEBM could be classified. The start-up Fincompare 1 (*2016) offers a digital brokerage
platform with over 200 banks and other financing partners. In addition to the platform,
professional financial consultants are available – from the application stage until the
conclusion of the contract. In the educational sector the aforementioned start-up Stackfuel
was categorized as a DEBM. The start-up CAYA 2 links the physical and digital world
and was categorized as DEBM for SaaS sector. It digitizes daily letter mail in compliance
with data protection regulations. Customers can opt to be notified via push message or e-
mail about new mail that can be viewed online while the physical mail can be archived on
request. An example for PDDBM from fintech is the already introduced start-up
ELEMENT. In SaaS the start-up Polyteia 3 provides an intelligent control platform for
cities and municipalities. They offer to merge and process data from previous data silos
and isolated specialist processes to make them available to decision-makers on a daily
basis.

1 https://fincompare.de/ (accessed 04.11.18)


2 https://www.getcaya.com/ (accessed 04.11.18)
3 https://www.polyteia.de/ (accessed 04.11.18)

14
This paper was presented at ISPIM Connects Fukuoka – Building on Innovation Tradition,
Fukuoka, Japan on 2-5 December 2018. The publication is available to ISPIM members at
www.ispim.org.

It follows, that the RQ2 “Which types of data-driven business models have developed in
the Berlin start-up scene in the economic sectors fintech, software as a service, and
education?” can be answered with the presented arguments drawn from scientific
literature as well as our analysis of the BMs of 50 Berlin-based start-ups.

Figure 5: Categorization of the 50 startups into LDBM, DEBM, PDDBM

5 Conclusion and Directions for Future Research


Our in-depth literature review provides an overview of the current DDBM literature
stream and related research fields such as industry 4.0 and digital platforms. Our
proposed taxonomy presents and defines the current state of the art and categorizes three
types of BMs: data-driven, data-enhanced, and low data BMs. However, the proposed
taxonomy needs further empirical testing. There are only very few publications that focus
exclusively on DDBM or DDBMI (Benta et al., 2017; Brownlow et al., 2015; Bulger et
al., 2014; Exner et al., 2017; Hartmann et al., 2016, 2014; Hunke et al., 2017; Schüritz et
al., 2017; Schüritz and Satzger, 2016; Zolnowski et al., 2017, 2016).
All of these publications tend to be rather young and come from a small community of
researchers which might explain why they sometimes resemble each other in terms of
their methodological frameworks and findings (Kühne and Böhmann, 2018).
Our analysis of Berlin-based start-ups shows that 39 out of 50 start-ups can be classified
as PDDBM. 10 Start-ups can be classified as DEBM and none as LDBM. It further
indicates that data analytics is the key process in Berlins SaaS and fintech sectors.
Interestingly, the processes of data visualization and data distribution proved to be
important in the educational sector. However, data generation and aggregation were not
monitored in the current sample.
Further research is needed to understand which digital technologies e.g. Cloud
Computing, IoT or AI are implemented within respective BMs and how they determine
the usage of data within key processes. Another important issue is that the DDBM

15
This paper was presented at ISPIM Connects Fukuoka – Building on Innovation Tradition,
Fukuoka, Japan on 2-5 December 2018. The publication is available to ISPIM members at
www.ispim.org.

discourse still seems to miss out on the role of data protection and data security.
Additional topics for consideration would be data culture and data ethics and their
respective influence on DDBMs. Because of that further research is needed to understand
the processes and relevance of data protection and data security for start-ups in general.
Furthermore, all definitions on DDBM define data as a key resource but we suppose that
data just becomes the key resource because of the interaction with human capital in
organization, which also needs further verification.
It is critical to note that our analysis is based on a desk research which has no general
validity (Mayring, 2010, 2002) However, freely accessible databases, corporate websites,
press articles, videos or podcasts provide researchers a new way of desk research. In
order to deepen our understanding of data-driven start-ups in Berlin we aim to conduct
interviews and build case studies based on our sample of start-ups. It seems like scientific
discourse needs more empirical work in the sense of what Massa et al. (2017) identified
as the interpretation of BMs as attributes of real organizations. Researchers need a better
understanding of the key activities in data-driven organizations. Only then can we
develop new frameworks and tools that will eventually aid researchers and practitioners
in understanding and implementing innovation processes in data-driven or data-enhanced
companies. We thus aim to explore PDDBMs and DEBMs in more detail.
In summary, the output of this paper addresses the following issues:
• Identification of research on DDBMs based on the in-depth literature review.
• Theoretical development of new DDBM-related definitions by introducing the
taxonomy of pure data-driven business ´models (PDDBM), data-enhanced business
models (DEBM), and low data business models (LDBM).
• First Analysis and classification of 50 Berlin-based start-ups from three different
sectors (fintech, education, SaaS)
• Identification of further research topics and gaps related to DDBM and DDBMI.
Our analysis emphasizes the meaning of DDBM for both BM researchers and
practitioners. It describes new data and digitization-driven services while encouraging
new best practices. It also demonstrates the spirit and the innovative culture of the city of
Berlin and its significance for Europe.

16
This paper was presented at ISPIM Connects Fukuoka – Building on Innovation Tradition,
Fukuoka, Japan on 2-5 December 2018. The publication is available to ISPIM members at
www.ispim.org.

Literature Review

Abdelkafi, N., Hilbig, R., Laudien, S.M., 2018. Business models of entrepreneurial
universities in the area of vocational education–an exploratory analysis. Int. J.
Technol. Manag. 77, 86–108.
Abdelrahim, M., Postoyan, R., Daafouz, J., 2013. Event-triggered control of nonlinear
singularly perturbed systems based only on the slow dynamics. IFAC Proc. 46,
347–352.
Amit, R., Zott, C., 2001. Value creation in e‐business. Strateg. Manag. J. 22, 493–520.
Benta, C., Wilberg, J., Hollauer, C., Omer, M., 2017. Process model for data-driven
business model generation, Proceedings of the 21st International Conference on
Engineering Design (ICED17), 347-356.
Bitkom, 2015. Big Data und Geschäftsmodell-Innovationen in der Praxis: 40+ Beispiele,
Berlin.
Bokrantz, J., Skoogh, A., Berlin, C., Stahre, J., 2017. Maintenance in digitalised
manufacturing: Delphi-based scenarios for 2030. Int. J. Prod. Econ. 191, 154–
169.
Brownlow, J., Zaki, M., Neely, A., Urmetzer, F., 2015. Data and Analytics - Data-Driven
Business Models: A Blueprint for Innovation 16, Cambridge Service Alliance.
Bulger, M., Taylor, G., Schroeder, R., 2014. Data-driven business models: challenges and
opportunities of big data, Oxford Internet Institute.
Casadesus-Masanell, R., Ricart, J.E., 2010. From strategy to business models and onto
tactics. Long Range Plann. 43, 195–215.
Cenamor, J., Sjödin, D.R., Parida, V., 2017. Adopting a platform approach in
servitization: Leveraging the value of digitalization. Int. J. Prod. Econ. 192, 54–
65.
Denzin, N.K., 2017. The research act: A theoretical introduction to sociological methods.
Routledge.
Dremel, C., Overhage, S., Schlauderer, S., Wulf, J., 2017. Towards a Capability Model
for Big Data Analytics, Proceedings Der 13. Internationalen Tagung
Wirtschaftsinformatik (WI 2017), St. Gallen, 1141-1155.
Eckert, K.-P., Henckel, L., Hoepner, P., 2014. Big Data – ungehobene Schätze oder
digitaler Albtraum 19, Berlin.
Ernst & Young GmbH Wirtschaftsprüfungsgesellschaft, 2018a. Start-up Barometer
Deutschland, URL https://start-up-initiative.ey.com/wp-
content/uploads/2018/01/EY_Start-up_Barometer_Januar_2018.pdf (accessed
07.11.18)
Ernst & Young GmbH Wirtschaftsprüfungsgesellschaft, 2018b. EY Start-up-Barometer
Europa, URL https://start-up-initiative.ey.com/wp-content/uploads/2018/03/EY-
start-up-barometer-europa-maerz-2018.pdf (accessed 07.11.18)
Exner, K., Stark, R., Kim, J.Y., Stark, R., 2017. Data-driven business model a
methodology to develop smart services. IEEE, 146–154.
Flick, U., 2011. Triangulation, in: Empirische Forschung Und Soziale Arbeit. Springer,
Wiesbaden.
Foss, N.J., Saebi, T., 2017. Fifteen Years of Research on Business Model Innovation:
How Far Have We Come, and Where Should We Go? J. Manag. 43, 200–227.
Gartner, 2012. What Is Big Data? - Gartner IT Glossary - Big Data, URL
https://www.gartner.com/it-glossary/big-data/ (accessed 11.5.18).
Gassmann, O., Frankenberger, K., Csik, M., 2017. Geschäftsmodelle entwickeln: 55
innovative Konzepte mit dem St. Galler Business Model Navigator, 2.,
überarbeite und erweiterte Auflage. ed. Hanser, München.

17
This paper was presented at ISPIM Connects Fukuoka – Building on Innovation Tradition,
Fukuoka, Japan on 2-5 December 2018. The publication is available to ISPIM members at
www.ispim.org.

Günther, W.A., Mehrizi, M.H.R., Huysman, M., Feldberg, F., 2017. Debating big data: A
literature review on realizing value from big data. J. Strateg. Inf. Syst. 26, 191–
209.
Hartmann, P.M., Zaki, M., Feldmann, N., Neely, A., 2016. Capturing value from big data
– a taxonomy of data-driven business models used by start-up firms. Int. J.
Oper. Prod. Manag. 36, 1382–1406.
Hartmann, P.M., Zaki, M., Feldmann, N., Neely, A., 2014. Big Data for Big Business? A
Taxonomy of Data-driven Business Models used by Start-up Firms. Taxon.
Data-Driven Bus. Models Used Start- Firms, Cambridge Service Alliance.
Hunke, F., Seebacher, S., Schuritz, R., Illi, A., 2017. Towards a Process Model for Data-
Driven Business Model Innovation. IEEE, 150–157.
Ibarra, D., Ganzarain, J., Igartua, J.I., 2018. Business model innovation through Industry
4.0: A review. Procedia Manuf. 22, 4–10.
Johnson, M.W., 2010. Seizing the white space: Business model innovation for growth
and renewal. Harvard Business Press.
Krämer, J., Wohlfarth, M., 2018. Market power, regulatory convergence, and the role of
data in digital markets. Telecommun. Policy 42, 154–171.
Kühne, B., Böhmann, T., 2018. Requirements for Representing Data-Driven Business
Models-Towards Extending the Business Model Canvas. Presented at the
Twenty-fourth Americas Conference on Information Systems, New Orleans,
2018.
Luque, A., Peralta, M.E., Heras, A. de las, Córdoba, A., 2017. State of the Industry 4.0 in
the Andalusian food sector. Procedia Manuf. 13, 1199–1205.
Massa, L., Tucci, C.L., Afuah, A., 2017. A Critical Assessment of Business Model
Research. Acad. Manag. Ann. 11, 73–104.
Mayring, P., 2010. Qualitative Inhaltsanaylse. Grundlagen und Techniken. Beltz Verlag,
Weinheim und Basel.
Mayring, P., 2002. Einführung in die Qualitative Sozialforschung. Beltz Verlag,
Weinheim und Basel.
Osterwalder, A., Pigneur, Y., 2010. Business model generation: a handbook for
visionaries, game changers, and challengers. John Wiley & Sons.
Osterwalder, A., Pigneur, Y., Tucci, C.L., 2005. Clarifying Business Models: Origins,
Present, and Future of the Concept. Commun. Assoc. Inf. Syst. 16, 1–25.
Paritala, P.K., Manchikatla, S., Yarlagadda, P.K.D.V., 2017. Digital Manufacturing-
Applications Past, Current, and Future Trends. Procedia Eng. 174, 982–991.
Parker, G.G., Van Alstyne, M., Choudary, S.P., Foster, J., 2016. Platform revolution:
How networked markets are transforming the economy and how to make them
work for you. WW Norton New York.
Sadovskyi, O., Engel, T., Heininger, R., Böhm, M., Krcmar, H., 2014. Analysis of Big
Data enabled Business Models using a Value Chain Perspective. Presented at
the Multikonferenz Wirtschaftsinformatik 2014, 1126-1137.
Schneider, S., Spieth, P., 2013. Business model innovation: Towards an integrated future
research agenda. Int. J. Innov. Manag. 17, 1340001.
Schüritz, R., Satzger, G., 2016. Patterns of data-infused business model innovation,
Business Informatics (CBI) - IEEE, 133–142.
Schüritz, R., Seebacher, S., Dorner, R., 2017. Capturing Value from Data: Revenue
Models for Data-Driven Services, Proceedings of the 50th Hawaii International
Conference on System Sciences, 5348-5357.
Täuscher, K., 2018. Platform Business Models: Conceptual Development and Design of a
Simulation-Based Assessment Method (Dissertation). Universität Leipzig,
Leipzig.

18
This paper was presented at ISPIM Connects Fukuoka – Building on Innovation Tradition,
Fukuoka, Japan on 2-5 December 2018. The publication is available to ISPIM members at
www.ispim.org.

Teece, D.J., 2010. Business Models, Business Strategy and Innovation. Long Range
Plann. 43, 172–194.
Vendrell-Herrero, F., Bustinza, O.F., Parry, G., Georgantzis, N., 2017. Servitization,
digitization and supply chain interdependency. Ind. Mark. Manag. 60, 69–81.
Weber, C., Königsberger, J., Kassner, L., Mitschang, B., 2017. M2DDM – A Maturity
Model for Data-Driven Manufacturing. Procedia CIRP 63, 173–178.
Zolnowski, A., Anke, J., Gudat, J., 2017. Towards a Cost-Benefit-Analysis of Data-
Driven Business Models, Proceedings Der 13. Internationalen Tagung
Wirtschaftsinformatik (WI 2017), St. Gallen, 181–195.
Zolnowski, A., Christiansen, T., Gudat, J., 2016. Business Model Transformation
Patterns of Data-Driven Innovations, ECIS 2016 Proceedings, 1-16
Zott, C., Amit, R., 2010. Business model design: an activity system perspective. Long
Range Plann. 43, 216–226.

19
View publication stats

You might also like