Current Practices in the Information Collection for Enterprise
Architecture Management
Robert Ehrensperger, Clemens Sauerwein and Ruth Breu
Institute of Computer Science, University of Innsbruck, Technikerstr. 21A, 6020 Innsbruck, Austria
Keywords: Enterprise Architecture Management, Information Collection, Enterprise-external Information Sources.
Abstract: The digital transformation influences business models, processes, and enterprise IT landscape as a whole.
Therefore, business-IT alignment is becoming more important than ever before. Enterprise architecture
management (EAM) is designed to support and improve this business-IT alignment. The success of EAM
crucially depends on the information available about a company's enterprise architecture, such as
infrastructure components, applications, and business processes. This paper discusses the results of a
qualitative expert survey with 26 experts in the field of EAM. The goal of this survey was to highlight current
practices in the information collection for EAM and identify relevant information from enterprise-external
data sources. The results provide a comprehensive overview of collected and utilized information in the
industry, including an assessment of the relevance of such information. Furthermore, the results highlight
challenges in practice and point out investments that organizations plan in the field of EAM.
1 INTRODUCTION
The ongoing digital transformation affects
longstanding business models and creates
opportunities for new ones (Berman, 2012) that
enable increasing the company’s profits and sales
figures (Amit and Zott, 2012). This digitalization
leads to changing business processes and require-
ments, which force organizations to transform. In the
mid-1970s, the average life cycle of very large
software applications was between 10 and 15 years,
which subsequently decreased to an average value of
5 years by 2005 (Soto-Acosta et al., 2016; Masak,
2006; Beck et al., 2001). This decreasing timespan
underlines the fact that organizations continue to
transform their enterprise architecture (EA) at an
increasingly rapid rate. Business-IT alignment plays
a crucial role in this transformation (Roth et al.,
2013). It is required to provide transparency between
the business requirements and the derived technical
implementations. EAM is designed to support and
improve this alignment (Maes et al., 2000; Farwick et
al., 2016). In particular, it is responsible for
transforming a company’s “as-is” IT landscape to a
“to-be” IT landscape in accordance with an
enterprise’s business strategy. Thus, the importance
of EAM is increasing.
EAM provides a holistic view of the entire
enterprise architecture with the help of EA models.
These models provide a comprehensive overview of
the interrelationships between business processes,
applications, processed information objects (e.g.,
business partner information), and the underlying IT
infrastructure components (e.g., server, firewall,
network) (Roth et al., 2013).
The success of EAM crucially depends on the
amount and quality of available EA information.
Thus, various researchers have focused their work on
relevant information sources for EAM, such as
Farwick et al. (2013) and Buschle et al. (2012).
However, no such studies focus on enterprise-
external information sources. Through the
introduction of social networks, the internet of things,
sensors, and smartphones, the world-wide existing
amount of information is significantly increasing
(Gantz and Reinsel, 2012). Among them,
unstructured information is the fastest-growing type
of digital information (Bakshi, 2012). This
information provides insights about numerous current
changes and events in the real world (Harris and Rea,
2009).
Therefore, this qualitative expert survey
investigates the status quo in the information
collection and questions which enterprise-external
information sources are relevant for EAM and might
Ehrensperger, R., Sauerwein, C. and Breu, R.
Current Practices in the Information Collection for Enterprise Architecture Management.
DOI: 10.5220/0009316907170727
In Proceedings of the 22nd International Conference on Enterprise Information Systems (ICEIS 2020) - Volume 2, pages 717-727
ISBN: 978-989-758-423-7
Copyright
c
2020 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
717
improve EAM. We designed the following research
questions in order to highlight this status quo and to
point out the relevance of enterprise-external
information for EAM:
RQ1: What are the current practices of
collecting information for EAM in
organizations?
RQ2: What is the most relevant information for
EAM?
RQ3: Which enterprise-external information
sources are relevant for EAM in practice?
RQ4: What is the value of collecting enterprise-
external information to EAM?
RQ5: What are the challenges to collect the
identified information?
In order to provide a more detailed understanding
of the topic, based on the research questions we
derived detailed survey questions according to the
recommendation of Gläser and Laudel (2010). In
summary, this led to a semi-structured survey
containing thirteen questions with twelve sub-
questions. In a further step, we selected EAM experts
to answer this survey. The results were evaluated
according to the method of Mayring (2010).
The remainder of this paper is structured as
follows. Section 2 provides a discussion of related
work, before section 3 documents the research
methodology applied. Section 4 outlines the results,
and section 5 discusses the key findings by answering
our research questions. Finally, section 6 concludes
the research at hand and provides an outlook for
future work.
2 RELATED WORK
Different authors have focused on leveraging
enterprise-internal information sources for EAM. The
survey by Farwick et al. (2013) analyzed potential
sources and their appropriateness for EAM.
Accordingly, they identified many different
enterprise-internal information sources such as
portfolio management tools, configuration manage-
ment databases (CMDBs), and license management
tools.
Buschle et al. (2012) outlined that an enterprise
service bus (ESB) may be used as an appropriate
information source for EAM. Furthermore, they
showed that leveraging on an ESB leads to an
improved quality of EA models.
Only a few researchers have analyzed enterprise-
external information sources for EAM. For example,
Zimmermann et al. (2017) only highlighted the notion
that gathering enterprise-external information may
further improve EAM. However, they did not
mention any concrete information sources.
The types of enterprise-external information can
be diverse. In general, there are three types of
information, namely structured, semi-structured, or
unstructured (Sint et al., 2009), among which the
latter is the fastest-growing type (Bakshi, 2012).
Research has shown that more than 80 % of useful
business-related information is stored in an
unstructured form (Das and Kumar, 2013). This fact
underlines the potential that lies in exploiting
unstructured information.
The first research steps in gathering unstructured
information for specific EAM requirements have
been undertaken. For example, Johnson et al. (2016)
argued for the use of machine learning techniques to
gather unstructured information and maintain EA
models. These techniques would even enable
handling information with a varying structure. For
analyzing massive amounts of diverse EA
information (e.g. spreadsheets, documents, presen-
tation), Hacks and Saber (2016) evaluated different
big data frameworks. Their goal was to find the best-
in-breed solution for the needs of EAM.
Many new opportunities to gather unstructured
information such as blog postings, log file contents,
or customer reviews are on the rise, although
currently research lacks leverage on them. As a first
step, it is necessary to identify enterprise-external
information sources that may increase the value of
EAM. Within the scientific literature, no research
could be found investigating the relevance of
enterprise-external information sources for EAM.
3 RESEARCH METHODOLOGY
This survey is designed in a semi-structured form
(Wohlin et al., 2012), containing a mix of open and
closed questions. It aims to analyze the current
practice of the information collection for EAM in
industry.
3.1 Participants
In order to identify eligible survey participants, the
following selection criteria were defined.
Accordingly, we applied the following two criteria
for the selection of the survey participants: (1)
employment in the field of EAM and (2) at least three
years of professional experience in the field of EAM.
Moreover, all persons participated voluntarily in this
study without any financial compensation.
ICEIS 2020 - 22nd International Conference on Enterprise Information Systems
718
Table 1: Overview of the participants.
ID Role
Years of
experience
Industry branch Region
1 EAM Lead 4 Automotive GER
2 Deputy Head of Collaborative EA 6 IT Consulting GER
3 Management of IT Infrastructure 18 Automotive GER
4 Enterprise Architect 10 Pharma GER
5 IT-Architect 3 Insurance -
6 Senior Expert EA 10 Oil & Gas AT
7 Head of EAM and Innovation 12 Financial Services GER
8 Senior Technology Architect 8 IT Consulting GER
9 Global Technology Consulting 7 IT Consulting GER
10 Director EAM 15 Construction GER
11 Head of Digitalization, Strategy, Architecture 15 Manufacturing GER
12 Project Lead, Product Owner 9 Financial Services GER
13 Enterprise Architect 4 Financial Services GER
14 EA Technical Lead 20 Defense / Military GER
15 Senior Enterprise Business Architect 5 Industrial Engineering GER
16 Enterprise Architect 5 Automotive GER
17 Digital Architect 4 Financial Services GER
18 Senior Enterprise Architect 20 Financial Services GER
19 Sub Product Owner 6 Automotive GER
20 IT Senior Professional (IT Referent) 10 Automotive GER
21 Enterprise Architect, Product Portfolio Manager 12 Financial Services /
Automotive
GER
22 Principal Enterprise Architect & Account Chief Architect 17 - GER
23 Product Portfolio Manager 4 Automotive GER
24 Enterprise Architect 10 Digital Industries GER
25 Senior Lead IT Consultant 14 Cross-industry AT
26 Consultant for Transformation and Business Development 20 Information
Technology
AT
Table 1 provides a comprehensive overview of the
participants, their roles, years of experience, industry
branch, and region. On average, the participants have
a work experience of 10.3 years. They work for
fifteen different organizations across thirteen
different industry branches such as automotive, IT
consulting, insurance, pharma, military/defense,
financial services, construction, industrial
engineering, digital industries, information
technology, pharma, oil and gas, and cross-industry.
One participant did not disclose the organization's
name and another did not state the industry branch.
Furthermore, all participants are located in central
Europe, although they are working for globally-
operating organizations. In order to protect personal
dates and keep the organization's intellectual
property, the participants are only referenced by an
ID. It is worth mentioning that some of the
participants work for IT consulting companies.
Therefore, they described situations from the
perspectives of multiple companies.
3.2 Survey Design
The survey contains three different documents. The
preliminary information contains a description of
the research context and the intention of the survey.
The interview protocol asks personal information
like the years of experience in EAM and the industry
branch. The research protocol lists all survey
questions.
In order to highlight the different aspects of the
defined research questions, several survey questions
were derived from these research questions,
according to Gläser and Laudel (2010). Using the
survey questions aims to reach a thematic structure of
the survey. The idea is to organize the sequence of
questions in a way that provides an introduction to the
topic and makes it comprehensible for the participants
(Kaiser, 2014). This approach facilitates gaining an
easier understanding of the survey (Kaiser, 2014).
Table 2 shows the research questions and their unique
ID’s.
Current Practices in the Information Collection for Enterprise Architecture Management
719
Table 2: Mapping of research questions to survey questions.
RQ ID Research question
RQ1 What are the current practices of collecting
information for EAM in organizations?
RQ2 What is the most relevant information for
EAM?
RQ3 Which enterprise-external information
sources are relevant for EAM in practice?
RQ4 What is the value of collecting enterprise-
external information to EAM?
RQ5 What are the challenges to collect the
identified information?
Moreover, table 3 provides an overview of the
survey questions (SQs) and their mapping to the
corresponding research questions (RQs).
The survey was conducted between May and
November 2019. The surveys were evaluated
according to Mayring (2010). This approach was
required to analyze the open questions. Mayring
(2010) provides a method for qualitative text analysis
that offers guidance on how to paraphrase, code
terminologies, generalize to a higher abstraction
level, and reduce to the core gist of the given answers.
4 RESULTS
This section discusses the main results of the survey
by answering the research questions.
4.1 RQ1: Current Practices of
Information Collection for EAM
In order to highlight the current practices of
information collection for EAM (RQ1), we surveyed
the participants about (1) the information that is
currently collected and populated to EA models,
(2) the automation of the information collection
including an explanation of the current realization
approaches and their advantages, and (3) planned
technological improvements in the field of EAM.
Therefore, the survey questions SQ1, SQ3, SQ4,
SQ4.1, SQ4.1.1, SQ5, SQ5.1 are answered
accordingly.
Our survey discovered the overall distribution of
the information that organizations currently collect
for EAM. Figure 1 illustrates this distribution,
whereby 100 % equals the total number of 26
participants.
Figure 1: Distribution of the currently-collected EAM
information.
Table 3: Overview of survey questions.
SQ ID Survey question RQ ID
SQ1 Which information is currently acquired for EAM? RQ1
SQ2 Is all relevant acquired information stored within EA models? RQ2
SQ2.1 In case the answer to SQ2 is “yes,” we asked: What are the most important ones? RQ2
SQ2.2 In case the answer to SQ2 is “no,” we asked: Which ones are not? RQ2
SQ3 How is the EA model information maintenance process designed? RQ1
SQ4 Is there already an automatic EA information gathering in place? RQ1
SQ4.1 In case the answer to SQ4 is “no,” we asked:
Would this bring advantages to enterprise architects in your field?
RQ1
SQ4.1.1 In case the answer to SQ4.1 is “yes,” we asked: What are the advantages? RQ1
SQ5 Are there any technological improvements planned in the field of EAM? RQ1
SQ5.1 In case the answer to SQ5 is “yes,” we asked: What are the examples of this? RQ1
SQ6 In case that you plan the next project portfolio, what additional information would help you? RQ3
SQ7 What additional enterprise-external information may provide added value to EAM? RQ3
SQ7.1 What are the advantages for enterprise architects knowing the aforementioned (in SQ7)
information?
RQ4
SQ7.2 What might be the advantages for other stakeholders knowing the aforementioned (in SQ7)
information?
RQ4
SQ8 What is the original source of this information? RQ3
SQ9 What are the three main reasons why this information is not already leveraged for EAM? RQ5
ICEIS 2020 - 22nd International Conference on Enterprise Information Systems
720
The following three examples are the most
frequently-mentioned answers in the survey, as
shown in figure 1. The majority of the participants
(83.3 %) mentioned information objects such as exis-
ting interfaces, applications, and their interrelations
out of the field application architecture information as
currently collected for EAM. The collection of
business architecture information such as existing
business processes, capabilities, business objects, and
business domains was highlighted by 75 % of the
participants. Whereby, a business object can be a
“customer,” and its attributes can be “name,” “second
name,” “age,” “country.” Furthermore, information
that corresponds to the infrastructure architecture
such as servers, network devices, deployed
technologies was mentioned by 62.5 % of the
participants. Half of the participants (50 %) collect
information out of the field information architecture,
such as business partner information, payment
information, or the delivery time information of
certain products. 20.8 % of the participants mentioned
collecting information for requirement management
such as desired changes in certain parts of the EA. It
is visible that organizations collect a wide range of
information for EAM. 16.7 % of the participants
outlined organizational, project portfolio, and release
management information. Few participants (12.5 %)
stated to collect information about the target EA. A
minority of 8.3 % suggested collecting information
about innovation management, license management,
frameworks, and decision management.
Moreover, information examples were given that
are collected only occasionally. Among these, the
collection of information about gained EA knowledge
(4.2 % of the participants) was stated. In the daily
work with EAs, employees gain knowledge such as
detailed interdependencies of long-term applications
used and their overarching processes. Architects want
to access this knowledge gained at the EA models.
Moreover, a small minority of participants (4.2 %)
expressed the need to collect cost information about
(1) business processes, (2) applications, and (3)
infrastructure components. Some architects strive to
know the total cost of ownership for operating a
business process or an application.
In a second step, we focused on the automation
of information collection processes. Therefore, we
asked the participants whether there is already an
automatic information collection in place. The vast
majority of 73.1 % mentioned that there is no
automatic information collection in place. Only 26.9
% stated that they use automatic information
collection for EAM.
If there was automatic information collection in
place, we continued asking the participants how this
automatic information collection is realized. The
most frequent answers were by using automated data
imports out of the (1) CMDB (Configuration
Management Database), with automated (2) asset
scanning tools, with data imports of (3) middleware,
and (4) security tools. The participants highlighted
that these imports mainly provide basic master
information about technical details such as server
names, installation date, or deployed software
versions. We asked the participants that do not
currently use any automatic information gathering for
EAM to assess the potential of automation. 77.3 %
of the participants who do not have automation in
place stated that this would bring advantages to their
business. The following advantages of automatic
information collection were stated:
Time savings
Timeliness of information
Improved collaboration
Better transparency of the entire IT landscape
Correctness of information
Consistency of information
Accuracy of information
The distribution of the advantages is illustrated in
figure 2.
Figure 2: Advantages of an automatic information
collection.
When talking about planned technological
improvements within EAM, more than 62.5 % of
the participants stated that they plan technological
improvements. Additionally, we asked the remaining
participants who plan technological improvements
to list these. In doing so, investing in the automation
of the information collection for EAM was most
frequently named by 42.9 % of the remaining partici-
pants, whereby some of the participants focused more
on the collection of application documents while
others more on the collection of business information
objects such as business partners, contracts, or bill of
materials that are relevant for their business.
Current Practices in the Information Collection for Enterprise Architecture Management
721
4.2 RQ2: Relevance of Information
The aim of this study is also to identify the most
relevant information that is acquired and stored
within EA models (RQ2). Therefore, we discuss the
answers related to SQ2, SQ2.1, SQ2.2 within this
section.
We asked whether every relevant acquired
information is stored within EA models, whereby
61.5 % of the surveyed participants responded that
not all relevant information is currently acquired and
stored within EA models.
Subsequently, we surveyed this remaining 61.5 %
of the participants concerning relevant information
that is not acquired and stored within EA models.
The following figure 3 depicts the distribution of the
given answers.
Figure 3: Overview of relevant information that is not
acquired and stored within EA models.
Information concerning software architecture
such as the application design or information about
existing interfaces was named by 42.9 % as missing
information. Furthermore, information details out of
the business architecture such as business processes,
business models and capabilities and the overall
business strategy were highlighted by 42.9 % as
relevant but not collected. Furthermore, 35.7 %
outlined the project portfolio of an organization as
often not being collected. The project portfolio
contains information about all projects of an
organization, with its interrelated dependencies.
Having this information linked to different layers of
an EA (e.g. conceptual and logical layers) was seen
as often missing. Moreover, information about the
current infrastructure architecture (21.4 %) and the
target EA (14.3 %) was highlighted. Only a few
participants (7.1 %) stated operation information,
financial information and information about deci-
sions taken as relevant but not acquired and stored
within EA models.
In addition, we surveyed the participants who
agreed (38.5 %) that all relevant information is
already acquired and stored within EA models
about the most important information. The
following figure 4 provides an overview of the given
answers.
Figure 4: Overview of the most important information.
64.3 % of the remaining participants who agreed
that all relevant information is already acquired and
stored within EA models outlined examples from the
application area of architecture information. For
instance, an organization’s established interfaces and
the need to have an asset inventory that acts as a
single point of truth within the EA were expressed as
most important. A significant number of participants
(42.9 %) also highlighted that business architecture
information is most important. As examples of
business architecture information, a process
inventory, business capabilities, and corresponding
IT capabilities were described. A minority of 26.6 %
of the participants see organizational information,
such as responsibilities, resource allocation, and
ongoing activities as most important. Furthermore,
infrastructure architecture (21.4 %) and project
portfolio information (14.3 %) were expressed as the
most important information for some participants.
Finally, only a few participants (7.1 %) mentioned
examples like legal information, information
architecture, and financial information.
4.3 RQ3: Relevance of
Enterprise-external Information
Besides the relevance of information in general, we
also focused on which enterprise-external informa-
tion may provide added value to EAM (RQ3). We
set this focus by addressing the survey questions SQ6,
SQ7, SQ8.
The identified enterprise-external information is
depicted in figure 5.
Figure 5: Enterprise-external information that may provide
added value to EAM.
ICEIS 2020 - 22nd International Conference on Enterprise Information Systems
722
The majority of the participants (46.2 %) state
product information such as (1) available product
versions, (2) product lifecycle information, (3) the
functional scope of products, and (4) emerging trends
as relevant enterprise-external information. The
second most frequent answer (34.6 %) was
benchmark information, such as (1) best practice
approaches, (2) EA patterns, (3) reference software
installations, and (4) touchpoints to external value
streams. Few participants (23.1 %) outlined market
information, such as (1) buying trends, (2) opinions
related to products, (3) financial risk ratings of
solution providers.
Moreover, enterprise-external information such as
(1) customer feedback about the (2) usage of
processes and applications, and information about (3)
the unforeseen usage of products is seen as relevant
by 23.1 % of the participants. Moreover, 15.4 % of
the participants mentioned end-to-end information
that involves the entire value chain across company
borders such as supplier and customer information
about their business processes, business information
models, and EA models. A small minority of
participants (7.7 %) highlighted cost information,
legal and regulatory information, and no enterprise-
external information at all as value-adding for EAM.
Furthermore, we investigated the origin of the
identified enterprise-external information. There-
fore, we asked the participants about the origin of
their mentioned information examples. Figure 6
illustrates the given answers.
Figure 6: Sources of EAM-relevant enterprise-external
information.
The origin is seen mainly with the vendors of
certain software products (34.6 %). A significant
number of participants responded that either the
origin is unknown (19.2 %) or they did not answer
this question (19.2 %) within the survey. Further
origins were identified at governmental institutions
(11.5 %), benchmark providers (11.5 %), and users of
the applications (11.5 %). Only a few participants
(7.7 %) mentioned social media, professional
journals, customers, or competitors as the origin of
the given information example.
Moreover, our investigations identified use cases
whereby the link between the previously identified
origin and the value of information was outlined. For
example, Participant (ID 6) stated that information
about the product lifecycle underpins enterprise
architects to plan the timing for the substitution of
products. Further information about the functional
scope of available software solution also supports by
making fit/gap analysis for this substitution. Both
pieces of information have their origin enterprise-
externally at vendors.
Furthermore, the participants were also asked
which additional information would help for the
project portfolio planning in organizations.
The functional scope of processes and
applications was named by 30.8 % of the participants.
Moreover, information about the business models and
strategy was stated by 23.1 % of the participants.
Additionally, 23.1 % of the participants highlighted
data models and flows as relevant additional
information for the project portfolio planning.
Furthermore, 19.2 % of the participants stated either
information about the project organization or did not
answer at all. 15.4 % of the participants expressed
information about the usage of software solutions and
financial information as helpful. Only 7.7 % of the
participants mentioned security and compliance
information. A small minority of 3.8 % stated product
lifecycle information, risk information, information
about proof of concepts, information about
interrelations between EA layers, and no additional
information at all as being helpful for the project
portfolio planning.
The stated information examples reveal that for
the task of project portfolio planning, information
with enterprise-external origin such as the functional
scope of applications also plays an essential role
besides the enterprise-internal examples.
4.4 RQ4: Value of Collecting
Enterprise-external Information
In order to assess the value of collecting enterprise-
external information (RQ4), we addressed the
survey questions SQ7.1, SQ7.2.
Approximately half of the participants (46.2 %)
stated that an accurate understanding of future EA
needs by being informed about new trends regarding
products, technologies, and customer needs would
allow enterprise architects to achieve a more active
role in designing the EA instead of reacting to
demands from top management.
Moreover, the opportunity to conduct compari-
son-based functional evaluations with other available
software products was highlighted as an advantage by
34.6 % of the participants. A comparison-based
Current Practices in the Information Collection for Enterprise Architecture Management
723
evaluation may help enterprise architects to identify
the best-in-breed software solutions for required EA
changes. This value may be leveraged by collecting
enterprise-external information about software
products (functional scope, supported business
processes, best practices) from software vendors.
19 % of the participant reported the benefit of
collecting cost information about business processes
and the IT systems, whereby the aim is to gain an
overview of the return on investment (ROI) of an EA
at a glance, allowing enterprise architects to improve
their evaluation and decision-making on a target EA.
In order to view the ROI, cost information is required
that comprises enterprise-external information such
as license costs and enterprise-internal information
such as the cost of operation.
Furthermore, the survey asked about advantages
for other stakeholders by collecting enterprise-
external information.
The participants most frequently mentioned
(38.5 %) an improved decision-making process as an
advantage. This advantage is associated with an
extended base of information. For example, for
business managers this could improve decision-
making about investments, the make or buy question,
or the prioritization of projects by having information
on the end-user perception and utilization of an EA.
Moreover, IT managers can more easily re-assess and
re-evaluate past decisions about selected standard
software products with the help of best practice
information concerning applications, business
processes, and the re-use of functionalities.
4.5 RQ5: Challenges of Collecting the
Identified Information
The paper at hand also investigates the challenges to
collect the identified information (RQ5). This
investigation includes all previously-described
examples coming from an enterprise-internal and an
enterprise-external environment. We addressed the
survey question SQ9 to investigate these challenges.
The overall findings are illustrated in the following
figure 7.
Approximately one-quarter of the participants
(26.9 %) stated that resources play the most
challenging role. It was stated that the manual effort
to collect the required information is time-consuming
that they primarily focus on the collection of the most
important information. This survey outlined that
organizations perceive the information collection for
EAM, not as the most crucial task for their business
success. This lack of priority and existing legacy
architectures were highlighted as challenging by 23.1
% of the participants. Moreover, we discovered that
EA experts suggest missing EA knowledge (19.2 %)
and the difficulty of identifying the relevant EAM
information (15.4 %) within organizations as a
significant challenge. These two points play an
essential role in the selection and usage of informa-
tion.
Figure 7: Main challenges why relevant information is not
leveraged for EAM.
Furthermore, 11.5 % of the participants stated
challenges like unclear profitability of EAM
investments, the interpretation of information, the
stakeholder alignment, and market competition as
challenging. Only a few participants (7.7 %) men-
tioned examples like concerns about security issues
and the quality of acquired information or they did not
answer at all.
5 DISCUSSION
Within this section, the key findings and potential
limitations of our study will be discussed.
5.1 Key Findings
Key Finding 1: Industry Does Not Collect and
Utilize All Relevant Enterprise-internal and
Enterprise-external Information for EAM.
This survey discovered that 61.5 % of the participants
stated that not every relevant information is collected
for EAM.
We identified a list of not-collected but relevant
information for EAM: first, the lack of available
information about software architectures (42.9 %) of
deployed software solutions as the most important
one; second, information about the business strategy,
models, and processes is often (42.9 %) not collected
and stored within EA models; and third, information
about the project portfolio is described by 35.7 % of
the participants.
Furthermore, this work discovered an apparent
mismatch in assessing the relevance of information.
It was revealed that some organizations see
information about business processes and capabilities
ICEIS 2020 - 22nd International Conference on Enterprise Information Systems
724
as highly relevant, while others do not collect this
information at all (cf. Figure 3). However, this
information is essential for achieving alignment
between business and IT, which is the primary
responsibility of EAM.
This finding emphasizes a gap of relevant
information and its collection in practice.
Key Finding 2: Practitioners Highlighted the
Relevance of Enterprise-external Information.
This study discovered the relevance of enterprise-
external information for EAM in practice. The
participants mentioned several examples of relevant
enterprise-external information, relating to available
software products and its versions, product lifecycle
information, and the functional scope of the products.
For instance, Participant (ID 19) described the
relevance of acquiring enterprise-external
information by the following use case. Enterprise-
external information enables architects to conduct a
comparison-based evaluation with other available
standard software solutions that focus on the same
functional scope. Enterprise architects may focus on
the weak parts of the business processes, and try to
optimize them by selecting the best-fitting on the
market available software solution for this process
part.
Besides a list of relevant enterprise-external
information, this study also identified the origins of
the information. These were seen mainly at vendors
of software products (34.6 %), users of applications
(11.5 %), benchmarking providers (11.5 %),
governmental institutions (11.5 %).
Key Finding 3: Practice Plans to Invest in the
Automation of Information Collection for EAM.
The paper at hand identified that industry is planning
to direct investments in the automation of information
collection for EAM. The automation of the
information collection is already a longstanding
research topic, as highlighted by many researchers
(e.g. Farwick et al., 2011; Moser et al., 2009; Grunow
et al., 2013; Buschle et al., 2011). However, the
current state in practice is not that far yet. Participants
of this study described mostly manual processes to
collect and maintain the EA models.
However, the first automated EA information
collection processes already exist. Seven participants
described automated data imports out of the CMDB,
asset scanning, middleware, and security tools. These
automated imports collect information from mainly
technical layers, such as server names, hardware
configurations, and software installations. Thus, this
work also identified that no automated collection of
business processes and the flow of information
objects between IT systems are established within the
surveyed industry. Furthermore, our work identified
no automated collection of enterprise-external
information among the participant's organizations.
Nevertheless, there is a joint agreement on the
potential of automation for EAM. 77.3 % of the
participants who do not yet have automation in place
agreed on the potential of automation of the
information collection processes. Furthermore, this
paper has outlined the advantages observed, such as
time savings, better timeliness of information, and
improved collaboration.
Moreover, the majority of the described
organizations (62.5 %) plan technological improve-
ments in the field of EAM. Many of them (30 %)
focus on the automation of the information collection
processes. Participants highlighted this investment as
the most pressing one in the field of EAM. This
finding also reveals that the potential of automation is
not yet fully leveraged.
Key Finding 4: Current Practice Has only Limited
Resources for the Information Collecting of EAM.
Our work identified a list of challenges why
EAM-relevant information is not collected. The first
three challenges are a lack of resources, missing
priority, and EA knowledge. These challenges are
closely linked to each other and may have the same
origin.
Organizations may tackle these three challenges
by assigning more budget to EAM projects and EA
employees. Accordingly, it is essential to focus on the
return of their investments. In the case of collecting
information for EAM, it is difficult to provide a
method that calculates reliable the expected ROI.
However, organizations need to be able to calculate
upfront a clear business case for investing in EAM.
As a result of this, research can provide guidance on
the assessment of business cases by providing
statistical information on the return of investments
from other EAM projects.
5.2 Limitations
Our survey might be limited by certain threats to
validity, namely the (i) selection of eligible
participants, (ii) the missing reproducibility of the
results, and (iii) false categorization and analysis.
In order to overcome (i), we applied the following
two participant selection criteria: (1) employment in
the field of EAM and (2) at least three years of
professional experience in the field of EAM. A
Current Practices in the Information Collection for Enterprise Architecture Management
725
detailed description of the participants can be found
in section 3.1.
In order to overcome (ii), we noted the personal
information, including contact details of each
participant and we, used the software tool MAXQDA
(Rädiker and Kuckartz, 2019) for the data analysis.
This tool provides traceability from given survey
answers to the analysis results and the conclusions
that we draw.
We evaluated the survey according to a method
for qualitative text analysis introduced by Mayring
(2010). This method provides systematic guidance on
how to paraphrase, code terminologies, generalize to
a higher abstraction level and reduce to the core gist.
Moreover, each instance of the paraphrasing and
coding was reviewed by at least two authors of this
publication. As a result, the risk of (iii) is at an
acceptable level.
6 CONCLUSION AND OUTLOOK
EAM’s principal objective is to optimize the strategic
IT alignment of organizations. A thriving EAM
crucially depends on available information within the
EA models. Therefore, the information selection and
collection is a pivotal issue.
In this paper, we analyzed the current practices of
the information collection for EAM in the industry
within Europe. Initially, we looked at the related work
and discovered that (1) the automation of information
collection for EAM is already a longstanding dis-
cussed topic within research, although current
practices are not investigated at all, and (2) only little
research has taken place in the field of collecting
enterprise-external information for EAM. Subse-
quently, we conducted a qualitative expert survey
among EAM practitioners to address the research
gaps (1) and (2).
Our survey reveals that the industry within Europe
does not collect all relevant information, while EA
practitioners underline the utility value of this
information for their organizations. Furthermore, we
discovered that EA practitioners also express the
relevance of enterprise-external information for
EAM. Moreover, we could outline an emerging trend
since most organizations lack but plan to invest in the
automation of information collection for EAM.
Finally, we also identified the main challenges of
leveraging all relevant information for EAM. Our
results provide researchers with a detailed view of the
current practices in information collection for EAM.
The findings of this survey rise to several
directions for further research. The lack of
automation of the collection of information, such as
business processes, business information objects.
Future research could highlight how to automate a
semantical integration into EA models of these
information examples. In terms of the challenges
identified, further research could give guidance on the
assessment of investments within EAM concerning
the ROI. Finally, regarding the collection of
enterprise-external information, further research may
investigate frameworks that enable integrating
external sources into an EA model.
REFERENCES
Amit, R., & Zott, C. (2012). Creating value through
business model innovation. 2012.
Bakshi, K. (2012, March). Considerations for big data:
Architecture and approach. In 2012 IEEE Aerospace
Conference (pp. 1-7). IEEE.
Beck, K., Beedle, M., Van Bennekum, A., Cockburn, A.,
Cunningham, W., Fowler, M., ... & Kern, J. (2001). The
agile manifesto.
Berman, S. J. (2012). Digital transformation: opportunities
to create new business models. Strategy & Leadership,
40(2), 16-24.
Buschle, M., Ekstedt, M., Grunow, S., Hauder, M.,
Matthes, F., & Roth, S. (2012). Automating enterprise
architecture documentation using an enterprise service
bus.
Buschle, M., Holm, H., Sommestad, T., Ekstedt, M., &
Shahzad, K. (2011, June). A Tool for automatic
Enterprise Architecture modeling. In International
Conference on Advanced Information Systems
Engineering (pp. 1-15). Springer, Berlin, Heidelberg.
Das, T. K., & Kumar, P. M. (2013). Big data analytics: A
framework for unstructured data analysis. International
Journal of Engineering Science & Technology, 5(1),
153.
Farwick, M., Agreiter, B., Breu, R., Ryll, S., Voges, K., &
Hanschke, I. (2011). Requirements for automated
enterprise architecture model maintenance. In 13th
International Conference on Enterprise Information
Systems (ICEIS), Beijing.
Farwick, M., Breu, R., Hauder, M., Roth, S., & Matthes, F.
(2013, January). Enterprise architecture document-
tation: Empirical analysis of information sources for
automation. In 2013 46th Hawaii International
Conference on System Sciences (pp. 3868-3877). IEEE.
Farwick, M., Schweda, C. M., Breu, R., & Hanschke, I.
(2016). A situational method for semi-automated
Enterprise Architecture Documentation. Software &
Systems Modeling, 15(2), 397-426.
Gantz, J., & Reinsel, D. (2012). The digital universe in
2020: Big data, bigger digital shadows, and biggest
growth in the far east. IDC iView: IDC Analyze the
future, 2007(2012), 1-16.
ICEIS 2020 - 22nd International Conference on Enterprise Information Systems
726
Gläser, J., & Laudel, G. (2010). Experteninterviews und
qualitative Inhaltsanalyse. Springer-Verlag.
Grunow, S., Matthes, F., & Roth, S. (2013). Towards
automated enterprise architecture documentation: Data
quality aspects of SAP PI. In Advances in Databases
and Information Systems (pp. 103-113). Springer,
Berlin, Heidelberg.
Hacks, S., & Saber, M. (2016). Supportive Role of Big Data
Technologies in Enterprise Architecture Management.
Full-scale Software Engineering/Current Trends in
Release Engineering, 13, 37.
Haren, V. (2011). TOGAF Version 9.1 A Pocket Guide.
Harris, A. L., & Rea, A. (2009). Web 2.0 and virtual world
technologies: A growing impact on IS education.
Journal of Information Systems Education, 20(2), 137.
Johnson, P., Ekstedt, M., & Lagerstrom, R. (2016,
September). Automatic probabilistic enterprise IT
architecture modeling: a dynamic bayesian networks
approach. In 2016 IEEE 20th International Enterprise
Distributed Object Computing Workshop (EDOCW)
(pp. 1-8). IEEE.
Kaiser, R. (2014). Qualitative Experteninterviews: Kon-
zeptionelle Grundlagen und praktische Durchführung.
Springer-Verlag.
Maes, R., Rijsenbrij, D., Truijens, O., & Goedvolk, H.
(2000). Redefining business-IT alignment through a
unified framework. Universiteit Van Amsterdam/Cap
Gemini White Paper.
Masak, D. (2006). Moderne Enterprise Architekturen.
Springer-Verlag.
Mayring, P. (2010). Qualitative Inhaltsanalyse. In
Handbuch qualitative Forschung in der Psychologie
(pp. 601-613). VS Verlag für Sozialwissenschaften.
Moser, C., Junginger, S., Brückmann, M., & Schöne, K. M.
(2009). Some process patterns for enterprise
architecture management. Software Engineering 2009.
Rädiker, S., & Kuckartz, U. (2019). Ein MAXQDA-Projekt
vorbereiten und Daten importieren. In Analyse
qualitativer Daten mit MAXQDA (pp. 23-42). Springer
VS, Wiesbaden.
Roth, S., Hauder, M., Farwick, M., Breu, R., & Matthes, F.
(2013). Enterprise Architecture Documentation:
Current Practices and Future Directions. Wirtschafts-
informatik, 58.
Sint, R., Schaffert, S., Stroka, S., & Ferstl, R. (2009, June).
Combining unstructured, fully structured and semi-
structured information in semantic wikis. In CEUR
Workshop Proceedings (Vol. 464, pp. 73-87).
Soto-Acosta, P., Placer-Maruri, E., & Perez-Gonzalez, D.
(2016). A case analysis of a product lifecycle infor-
mation management framework for SMEs. Interna-
tional Journal of Information Management, 36(2), 240-
244.
Wohlin, C., Runeson, P., Höst, M., Ohlsson, M. C.,
Regnell, B., & Wesslén, A. (2012). Experimentation in
software engineering. Springer Science & Business
Media.
Zimmermann, A., Schmidt, R., Sandkuhl, K., Jugel, D.,
Bogner, J., & Möhring, M. (2017, October). Decision
management for micro-granular digital architecture. In
2017 IEEE 21st International Enterprise Distributed
Object Computing Workshop (EDOCW) (pp. 29-38).
IEEE.
Current Practices in the Information Collection for Enterprise Architecture Management
727