The notion of software ecosystems has been popular both in research and industry for more than a decade, but how software ecosystems are created still remains unclear. This becomes more of a challenge if one examines the “creation” of ecosystems that have high probability in surviving in the future, i.e. with respect to ecosystem health. In this paper, we focus on the creation of software ecosystems and propose a process for designing, developing, and establishing software ecosystems based on three basic steps and a set of activities for each step. We note that software ecosystem research identifies that ecosystems typically emerge from either a company deciding to allow development on their product platform or from a successful open source project. In our study we add to this knowledge by demonstrating, through two case studies, that ecosystems can emerge from more than a technological infrastructure (platform). We identify that ecosystems can emerge out of two more distinct types of environments and thus the design should be based on the characteristics of this categorization. Moreover, we follow the approach that design, development, and establishment are not three distinct phases but rather aspects of a single re-iterating phase and thus propose the view of design, development, and establishment as a continous process, running in parallel with and interrelated to the monitoring of the ecosystem evolution.
The notion of software ecosystems is argued to provide clear advantages
compared to traditional software development and distribution as it, among other,
accelerates software development, reduces time to market, and increases user and
customer segment reachability. It is not a surprise that within the recent years
we have experienced an increasing popularity of software ecosystems both as a
topic of study and as a means of developing and distributing software (products).
Despite the popularity, it is still very challenging to create software ecosystems,
especially if one should take into consideration aspects of ecosystem
survivability, productivity, or health. Few studies have been investigating the conditions of
establishment of a software ecosystems and even fewer propose ways of designing
software ecosystems. However, this kind of studies tend to either be too specific
for a type of ecosystem and thus hard to generalize, or too generic and thus hard
to apply. Remarks that are already identified in the most recent and extensive
systematic literature review [
Contemporary public discussion on software ecosystems is much driven by the
most visible players in the digital economy, the platforms and app stores of Apple
and Google being the usual examples in the discussion. Among the practitioners,
this has lead to a platform-centric view of ecosystem thinking where a platform
provider is needed to orchestrate an ecosystem. Further, the terms platform and
ecosystem are closely connected if not treated almost as synonyms. However, the
literature has presented a variety of ecosystems and value networks beyond the
platform-centric approach, such as ecosystems build around standards, common
business and commonly adopted infrastructure [
The limitations of platform-centric ecosystem thinking are, to some extent, visible also in our common thinking on how to build ecosystems. That is, we tend to think that the only way to build an ecosystem is to build a platform and attract participants to it by some means, typically by providing financial benefits to the participants. This underlying assumption may lead to ignorance of a wider view on how to build ecosystems as the viewpoints of actors in the value network and the value creation in the business domain are overlooked if not excluded totally from our thinking.
In this paper, we take the wider view to building ecosystems. We start our journey towards a method for building ecosystems from the observation that ecosystem can emerge out of three distinct types of environments and thus the design should be based on the characteristics of this categorization. We study two cases presenting an actor-rooted and a business-rooted approach to ecosystem building. Adding findings from the two cases to the infrastructure-rooted approach (including platform-centric approach) we propose a process for designing, developing, and establishing software ecosystems based on three basic steps and a set of activities for each step. Moreover, we follow the approach that design, development, and establishment are not three distinct phases, but rather aspects of a single re-iterating phase and thus propose the view of design, development, and establishment as a continuous process, running in parallel with and interrelated to the monitoring of the ecosystem evolution.
The field of software ecosystems has an activity that spreads through several
years. From the first reference in the book of Messerschmitt and Szyperski [
Campbel and Ahmed [
In this section we discuss and analyze two cases of designing and building a software ecosystem. The first case is the telemedicine ecosystem established around the telemedicine services of the Danish healthcare and the second cases is the smart city ecosystem established around the smart city and Internet of Things (IoT) infrastructure and services in an area of one of the most populated cities in Finland.
Danish healthcare, following the tendency in many other western countries, is
facing a number of challenges due to changes in the demographics. The increase
in life expectancy and decrease of birth-rate in combination with a rapid increase
of lifestyle conditions and the continuously improving healthcare diagnosis and
treatment are putting a pressure on the economics of a welfare-based1 and
position the continuous care of the elderly and the chronically ill in even more central
focus [
technologies are faced with severe integration and interoperability issues caused
by the increasing need to interact with other medical system characterized as
”silo” solutions and organizationally complex systems [
Thus, the establishment of the telemedical ecosystem deviates from the
typical view of ecosystem emergence (i.e. from a successful platform or product).
In this ecosystem, the design was motivated from a set of clear incentives. The
state and healthcare authorities have been part of shaping and clarifying the
incentives, however this kind of actors have not been otherwise active in the
design and establishment of the ecosystem. Therefore, the ecosystem was, during
design, characterized form the lack of orchestration. The steps taken to establish
the ecosystem was2:
– Identify and map the existing (and future) actors, (software) systems and
their relationships [
The second case is the establishment of an ecosystem in the smart city domain.
The contribution of the digital technologies is considered to form a foundation for
so called smart cities. Smart cities are complex systems and consist of multiple
domains like transportation, energy, living, and governance. Smart city domains
utilize digital technologies by collecting and storing both private and public
data. They increasingly release the public information and data sets for external
parties. The idea behind releasing the public data sets is to provide a possibility
for external stakeholders to develop and create smart applications and services
for citizens. Naturally, an ecosystem would support and facilitate the actor and
smart city service interaction. An example is the environment for agile software
and internet of things product and service development and experimentation
with real users (citizens) in real-world settings [
In this context, our case, an urban area in one of the ten most populated
cities in Finland is on the process of establishing a smart city ecosystem. The
ecosystem establishment process was initiated by a set of actors interested in the
smart city domain. These actors created a consortium that aimed at promoting
the interaction of digital and software services in collaboration with
independent business models, i.e. an ecosystem. Purpose of the smart city ecosystem is
to develop new applications and internet of things service solutions in
collaboration with construction companies, smart grid providers, nursing houses, city
governance, and citizens. The initial actors in the smart city ecosystem included
representatives from universities and city as well as the stakeholders from private
2 A more detail description on this work can be found in [
– Build technological infrastructure (e.g. platform) to support the ecosystem.
As noted, the two cases studies are examples of ecosystem established by other
than a common technological infrastructure (or platform). The telemedicine
ecosystem is a business-rooted3, while the smart city ecosystems is an actor
rooted4. These two cases contribute with di↵erent perspectives on how
ecosystem are established. They add more parameters to the up-to-now knowledge of
ecosystems being created by a successful or popular technological infrastructure
(platform) [
Up to the current point and to the best of our knowledge of the field, there is no previous work suggesting an applicable and holistic or generic (i.e. applicable to most or all types of ecosystems) way of creating a software ecosystem. This is the gap that we are trying to address with this approach, as we argue that a method for designing ecosystems that is easy to apply and mature enough would support the maturity of the field both theoretically and empirically.
In our approach, we propose the view of ecosystem design, development, and establishment as one continuous and re-iterative phase rather than three distinct phases. In order to initiate this process, the basic information needs to be collected and the first initial designs need to be drawn. Thus, we identify three main steps in our process to conduct the necessary work for the iterative design. Figure 1 shows the proposed steps and the tasks included in each step. Our approach includes three main steps: pre-analysis, design, and evaluate & monitor. In the subsections bellow we describe these steps. Our approach has a strong focus on the ecosystem health, thus apart from the design, we support the view of continuous monitoring and evolution of the ecosystem making the separation between design and establishment unclear. This is reflected in step 3.
Furthermore, taking the approach demonstrated from our cases, we identify that ecosystem design can occur based on three di↵erent ecosystem types: infrastructure-rooted, where the ecosystem is established around a technological
4 I.e. initiated by a set of actors to drive the ecosystem development. 1.4 Existing ecosystem aspects
Technological infrastructure
Incentives Value proposition for the actor for the ecosystem Customer segment
Revenue steam 2.2 Map existing & new actors Actors Roles Contributions Interaction
Among actors
Actor to software 2.3 Define/Identify orchestrator strategy 2.4 Build technological infrastructure Step 1: Pre-analysis 1.1 Domain 1.2 Scope 1.3 General principles (a) Infrastructure (platform) rooted 2.1 Identify extension
possibilities 2.2 Define business
Incentives Value proposition for the actor for the ecosystem Customer segment
Revenue steam 2.3 Map existing & new actors Actors Roles Contributions Interaction
Among actors
Actor to software 2.4 (Re) Define orchestrator strategy 2.5 Open - extend
infrastructure 2.4 Involve actors Step 3: Evaluate and monitor 3.1 Desired behavior 3.2 Define ecosystem measures 3.3. Define iteration/observation intervals 3.4 Iterate ecosystem evaluation
Measure Evaluate
Act infrastructure5;actor-rooted, where the establishment is around a strong actor consortium; and business-rooted, where the ecosystem is established around a strong business (or incentives).
The initial step for the design of a software ecosystem is to identify the general
information and characteristics of the future ecosystem. This includes identifying
the applied domain of the ecosystem, i.e. how is the domain defined and what are
the general characteristics of this domain. Further, this step includes defining the
scope of the ecosystem and marking the borders of what is considered part of the
ecosystem. Moreover, this step includes identifying the general principles of the
ecosystem, i.e. core values and characteristics of the ecosystem that essential for
the ecosystem [
If we examine how ecosystems are created, the most common way appearing
in the literature is from a (software) company opening their platform to
external actors or an open source software (OSS) project that is gaining popularity.
Examining the existing ecosystems in the industry (or in the literature e.g. the
list in [
Finally, as already explained, in our approach we propose the view of the design and development as a continuous and iterative process where software ecosystem design, development, and establishment are not distinct phases but rather part of one continuous and re-iterative phase. In order to achieve that, the ecosystem should be constantly monitored on its evolution and reaction to changes and potential deterioration should initiate new actions on the ecosystem architecture or orchestration. Thus, this step includes activities that focus on identifying what should be measured in the ecosystem to identify evolution and change in ecosystem health. After the measures are identified monitoring and evaluation activities will focus on (i) intervening in the operation of the ecosystem with changes and (ii) evaluating the e↵ect of potential changes (as much as the whole design). It is essential to underline that identification of measures is an essential step as it defines the scope of action within the ecosystem. Too narrow measure might result in lack of overview of the whole ecosystem while not accurate or poorly defined measures might guide to wrong conclusions on the ecosystem activity and evolution.
This paper aims at bringing focus to a central issue in the field of software ecosystem by proposing a method on designing ecosystems. Although generic and applicable, our method does not cover all the possible and potentially essential aspects in ecosystem design and evolution. One relevant aspect not adequately discussed is the orchestration of software ecosystems. The orchestration is central aspect in the health and evolution of an ecosystem and eventually the design of an ecosystem should include concrete considerations on the orchestration, in order to support the di↵erent characteristics of the ecosystem, its domain, and scope. Another relevant aspect is the establishment of the proper interfaces both technical and organizational. The di↵erent interfaces between the software components (e.g. in the common technological infrastructure) and between the di↵erent actors, should reflect the orchestration strategy of the ecosystem and respect the domain, boarders, and roles of the ecosystem and its actors.
Finally, as already discussed, the choice of the proper measures for monitoring the ecosystem is central to the evolution of the ecosystem towards the right direction. The monitored measures should also be influences as much as influence the ecosystem orchestration.
In this work, we try to put focus on the gap in research and industry on how to “create” software ecosystems. Using our deep knowledge on software ecosystem literature and industry and experience from designing software ecosystem, we propose a method for designing software ecosystems that is easy to use and applicable. Our method is consisted of three steps and a set of activities for each step.
We are currently empirically validating and improving this method. Further work includes the empirical evaluation and improvement with cases of each different type of design. Moreover, we plan to identify characteristics of the method for specific domains, i.e. how this “generic” method changes when applied to a domain with specific characteristics. It is our hope that this will be a first step towards a better informed and explicit design of software ecosystems and eventually further maturity in the field.
This work was partially conducted under the 5K6 project, co-funded from the TEKES foundation (all authors), and the SCAUT7 project, co-funded by Innovation Fund Denmark, grant #72-2014-1 (Manikas). 6 https://agoracenter.jyu.fi/projects/5k 7 http://www.scaut.dk/