Triggers analysis of an agile transformation: the case of a central bank

Abstract

Agile transformation is one of the central topics for many IT departments in large organizations. Convinced by the benefits, large companies have entered a generalization phase of these approaches at the level of all information system projects. But little research informs practitioners about how a company goes from an experimentation phase to the generalization. Knowing that agile methods generate several changes in roles, processes and culture, our paper aims to answer the following question: what are the triggers of a decision that will enable an organization moving from experimentation to a wide generalization of an agile method? Based on a qualitative research design via a case study seeking to explain this complex phenomenon. We investigated the adoption process of agile methods during the last 13 years in the IT department of the French central bank. Contributions and practical implications highlight successful ingredients allowing to organize the implementation of agile methods in a large public organization.

Full text

ScienceDirect
Available online at www.sciencedirect.com
Procedia Computer Science 164 (2019) 449–456
1877-0509 © 2019 The Authors. Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Peer-review under responsibility of the scientific committee of the CENTERIS -International Conference on ENTERprise Information
Systems / ProjMAN - International Conference on Project MANagement / HCist - International Conference on Health and Social Care
Information Systems and Technologies.
10.1016/j.procs.2019.12.205
10.1016/j.procs.2019.12.205 1877-0509
© 2019 The Authors. Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Peer-review under responsibility of the scientific committee of the CENTERIS -International Conference on ENTERprise
Information Systems / ProjMAN - International Conference on Project MANagement / HCist - International Conference on
Health and Social Care Information Systems and Technologies.
Available online at www.sciencedirect.com
ScienceDirect
Procedia Computer Science 00 (2019) 000–000
www.elsevier.com/locate/procedia
1877-0509 © 2019 The Authors. Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Peer-review under responsibility of the scientific committee of the CENTERIS - International Conference on ENTERprise Information Systems /
ProjMAN – International Conference on Project MANagement / HCist - International Conference on Health and Social Care Information
Systems and Technologies
CENTERIS - International Conference on ENTERprise Information Systems / ProjMAN -
International Conference on Project MANagement / HCist - International Conference on Health
and Social Care Information Systems and Technologies
Triggers analysis of an agile transformation: the case of a central bank
Akim Berkania,*, Dominique Causseb, Laurent Thomasc,1
aParis-Dauphine University, PSL Research University, Place du Maréchal de Lattre de Tassigny, Paris 75016, France
bProgramme de recherche Aurore, Daylight Consulting, 1 Parvis de la Défense, Puteaux 92044, France
cBanque de France, 31 rue Croix des petits-Champs. 75049 Paris, France
Abstract
Agile transformation is one of the central topics for many IT departments in large organizations. Convinced by the benefits, large
companies have entered a generalization phase of these approaches at the level of all information system projects. But little research
informs practitioners about how a company goes from an experimentation phase to the generalization. Knowing that agile methods
generate several changes in roles, processes and culture, our paper aims to answer the following question: what are the triggers of
a decision that will enable an organization moving from experimentation to a wide generalization of an agile method? Based on a
qualitative research design via a case study seeking to explain this complex phenomenon. We investigated the adoption process of
agile methods during the last 13 years in the IT department of the French central bank. Contributions and practical implications
highlight successful ingredients allowing to organize the implementation of agile methods in a large public organization.
© 2019 The Authors. Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Peer-review under responsibility of the scientific committee of the CENTERIS - International Conference on ENTERprise
Information Systems / ProjMAN – International Conference on Project MANagement / HCist - International Conference on Health
and Social Care Information Systems and Technologies
Keywords: Agile transformation; information system department; project management; management innovation; case study.
* Corresponding author.
Email address: akim.berkani@dauphine.eu
1 The views, thoughts, and opinions expressed in this paper are those of the author and do not necessarily reflect those of the Banque de France.
Available online at www.sciencedirect.com
ScienceDirect
Procedia Computer Science 00 (2019) 000–000
www.elsevier.com/locate/procedia
1877-0509 © 2019 The Authors. Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Peer-review under responsibility of the scientific committee of the CENTERIS - International Conference on ENTERprise Information Systems /
ProjMAN – International Conference on Project MANagement / HCist - International Conference on Health and Social Care Information
Systems and Technologies
CENTERIS - International Conference on ENTERprise Information Systems / ProjMAN -
International Conference on Project MANagement / HCist - International Conference on Health
and Social Care Information Systems and Technologies
Triggers analysis of an agile transformation: the case of a central bank
Akim Berkania,*, Dominique Causseb, Laurent Thomasc,1
aParis-Dauphine University, PSL Research University, Place du Maréchal de Lattre de Tassigny, Paris 75016, France
bProgramme de recherche Aurore, Daylight Consulting, 1 Parvis de la Défense, Puteaux 92044, France
cBanque de France, 31 rue Croix des petits-Champs. 75049 Paris, France
Abstract
Agile transformation is one of the central topics for many IT departments in large organizations. Convinced by the benefits, large
companies have entered a generalization phase of these approaches at the level of all information system projects. But little research
informs practitioners about how a company goes from an experimentation phase to the generalization. Knowing that agile methods
generate several changes in roles, processes and culture, our paper aims to answer the following question: what are the triggers of
a decision that will enable an organization moving from experimentation to a wide generalization of an agile method? Based on a
qualitative research design via a case study seeking to explain this complex phenomenon. We investigated the adoption process of
agile methods during the last 13 years in the IT department of the French central bank. Contributions and practical implications
highlight successful ingredients allowing to organize the implementation of agile methods in a large public organization.
© 2019 The Authors. Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Peer-review under responsibility of the scientific committee of the CENTERIS - International Conference on ENTERprise
Information Systems / ProjMAN – International Conference on Project MANagement / HCist - International Conference on Health
and Social Care Information Systems and Technologies
Keywords: Agile transformation; information system department; project management; management innovation; case study.
* Corresponding author.
Email address: akim.berkani@dauphine.eu
1 The views, thoughts, and opinions expressed in this paper are those of the author and do not necessarily reflect those of the Banque de France.

450 Akim Berkani et al. / Procedia Computer Science 164 (2019) 449–456
2 Akim Berkani et al. / Procedia Computer Science 00 (2019) 000–000
1. Introduction
Almost 20 years after the agile manifesto, information systems (IS) projects are still under an increasing influence
of agile approaches [1]. Agile methods are known to be slightly formalized methodologies allowing to consider
requirements and solutions evolving throughout a project. Relying on multidisciplinary teams with self-management
principles, they are based on an iterative, incremental and adaptive development process [2].
Since the release of the agile manifesto in 2001, many methodologies appeared and gained legitimacy from
practitioner, mainly due to benefits involved. Thus, large organizations are embarking on the deployment of agile
approaches on a large scale [3]. If Information Technologies (IT) departments are usually the first entities to adopt
these methods, there are still many interrogations about this topic. One of them concerns the way an organization will
manage the adoption, adaptation and implementation of agile methods that include other departments than IT. Many
authors in information system literature characterize the adoption process as an agile transformation [4]. However,
agile transformation goes beyond the simple adoption of agile practices [5] and some organizations clearly state that
their aim is to become an "agile enterprise". Studies about agile transformation are still scarce and the term “agile
transformation” seems ambivalent.
When an agile method is adopted by an organization, the literature in management helps researchers to understand
that this adoption could be explained under the lens of management innovation adoption [ 6]. The adoption of a
management innovation has been theorized and includes several frameworks that describes different sequences such
as initiation, decision and implementation. Thus, the research question that we seek to answer is formulated as
follows: when considering the different sequences of the adoption process of a management innovation in a large
organization, what are the triggers of a decision that will enable moving from an experimentation (initiation) to a
wide generalization (implementation) of an agile method?
This paper is exploratory and seeks to investigate the events that trigger an agile transformation in a large
organization. To investigate this phenomenon, we launched a single case study of an organization rarely studied in
this context. This research deals with the IT department of the French central bank that is a public service organization.
Results contributes to understand the different decisions levels triggering a large implementation of an agile method
creating a new way of organizing a large IT department. Furthermore, the study allows to identify the phases of the
adoption process of an agile method allowing practitioner and researcher to know more about how new management
practices are adopted.
We present in next part a theoretical background about the emergence of the agile transformation topic among
researchers and we mobilize the management innovation literature to explain how we can investigate this
phenomenon. We continue by presenting the qualitative research design based on an in-depth case study, part four
presents an analysis of the case, and we finish with the practical implications and a conclusion dealing with future
works.
2. Theoretical Background
In 2016, Dikert et al., [4] proposed on their literature review to characterize the adoption of agile methods on the
large as a large-scale agile transformation. If few practitioners used the term to characterize the implementation of a
scaled framework at the project level, the agile transformation is a process where a large organization will change the
information systems development methods from a plan-driven process to an agile development on a large perspective
including changes in how different teams involved in projects are organized [7].
2.1. Agile transformation in large IT departments
Referring to Moe and Dingsoyr [8], the main questions to be addressed about large-scale agile transformation are:
Inter-team coordination, business agility, knowledge sharing and knowledge networks, and agile transformation, that
means according to them, how the deployment of agile methods can be organized? To succeed in implementing global
site adoption, organizations needs to define a process to deploy this transformation.
Paasivara and co-authors [5], analysed in a recent investigation, a large-scale agile transformation of an Ericsson
product development program developing a new information technology platform. They presented the steps taken, the
challenges faced, and the mitigating actions taken. They gave four lessons learned about an international

Akim Berkani et al. / Procedia Computer Science 164 (2019) 449–456 451
Akim Berkani et al. / Procedia Computer Science 00 (2019) 000–000 3
implementation dealing with: Experimental Transformation Approach, a Stepwise Transformation, a lack of a
Common Agile Framework is highlighted, and the Limited Team Interchangeability.
Agile Transformation touches many IT departments of large companies in many industries. An exploratory study
analysing twelve global cases shows the adopted agile structures at different large enterprises [9]. The study reveal
that companies start their agile transformation by adopting a template for a generic fully agile unit regardless of
whether the initial setting is bimodal or not. With increased agile maturity stages, companies enhance this template to
their needs by incorporating shared service tribes from the Spotify model [10]. Their case study reveals that the
adoption of agile structures currently takes place in enterprises at large scale regardless industry or size. However,
their works doesn’t give insights about public services organizations.
When agile approaches are introduced within an organization, we hypothesize that this introduction can be likened
to the internal adoption process of a managerial innovation (MI) [11]. Thus, to meet our ambition to characterize the
agile transformation sequences across an entire IT department, we mobilize a theoretical framework consistent with
the analysis of the introduction of a new method, technique or practice.
2.2. Management innovation adoption process
The adoption of a management innovation in organizations is defined as a process that delineates how the
organization selects and uses a practice for the first time [12]. The adoption process can be divided into three general
phases of initiation, decision and implementation [11]. The adoption process can therefore be considered as a
progression of orderly sequential phases. These three phases of an adoption process can be characterized as follows:
The first is relative to initiation. It consists of all activities related to the perception of problems or needs, the search
for solutions, the collection of information on these solutions, the training of attitudes towards these solutions and
their evaluation. It then culminates in the second phase, decision-making [13]. In other words, the members of the
organization discover the existence of an innovation, evaluate its relevance for the organization, communicate and
share about it until making the decision to adopt it [11]. These two phases have been the subject of frequent
fragmentation, some authors having made the choice to distinguish the different stages of a rational decision-making:
the awareness of the problems and needs, the formation of attitudes internally faced with a potential adoption,
evaluation of possible choices, and selection of the best solution [6].
The third phase called "implementation" is composed of all events and actions related to the preparation of the
implementation, the exploratory implementation and the initial use of the innovation, which involves adaptation and
modification of both innovation and organization [13]. In this phase, we can assess whether members of the
organization agree to use this innovation or show resistance to its use [11]. Finally, the continuation phase of the use
corresponds to the fact that the innovation is used in a usual way, even generalized and that it becomes even a current
characteristic of the organization. It can be understood as the culmination of the adoption process and a way to judge
its success.
This literature review on the adoption process of a management innovation has two main findings relating to our
research object. When an agile method is introduced to an organization and if it's new to internal practices, we can
assume that the adoption process could be explained under the analytical framework of management innovation
adoption. Furthermore, the literature review, gives few insights about the different decisions made during the adoption
process of an agile method during the different sequences of the adoption.
3. Methodology
Among different qualitative research strategies, we decided to choose a case study approach. It is not attached to
an epistemological paradigm and may be used to understand, explain, test or generate a theory [14]. Yin [15], proposes
a typology that distinguishes four designs based on the number of cases, but also their embeddedness or not. The
objectivity of a case study rest on « multiple sources of evidence ». It seems important to remind the selection criteria
for a unique case study [15]. Three situations make the choice of a single case relevant: a critical case for testing an
established theory whose foundations and conditions of validity are well known. A revealing case to reveal a
phenomenon previously inaccessible to scientific investigation and an extreme case to understand a scarce
phenomenon.
2 Akim Berkani et al. / Procedia Computer Science 00 (2019) 000–000
1. Introduction
Almost 20 years after the agile manifesto, information systems (IS) projects are still under an increasing influence
of agile approaches [1]. Agile methods are known to be slightly formalized methodologies allowing to consider
requirements and solutions evolving throughout a project. Relying on multidisciplinary teams with self-management
principles, they are based on an iterative, incremental and adaptive development process [2].
Since the release of the agile manifesto in 2001, many methodologies appeared and gained legitimacy from
practitioner, mainly due to benefits involved. Thus, large organizations are embarking on the deployment of agile
approaches on a large scale [3]. If Information Technologies (IT) departments are usually the first entities to adopt
these methods, there are still many interrogations about this topic. One of them concerns the way an organization will
manage the adoption, adaptation and implementation of agile methods that include other departments than IT. Many
authors in information system literature characterize the adoption process as an agile transformation [4]. However,
agile transformation goes beyond the simple adoption of agile practices [5] and some organizations clearly state that
their aim is to become an "agile enterprise". Studies about agile transformation are still scarce and the term “agile
transformation” seems ambivalent.
When an agile method is adopted by an organization, the literature in management helps researchers to understand
that this adoption could be explained under the lens of management innovation adoption [ 6]. The adoption of a
management innovation has been theorized and includes several frameworks that describes different sequences such
as initiation, decision and implementation. Thus, the research question that we seek to answer is formulated as
follows: when considering the different sequences of the adoption process of a management innovation in a large
organization, what are the triggers of a decision that will enable moving from an experimentation (initiation) to a
wide generalization (implementation) of an agile method?
This paper is exploratory and seeks to investigate the events that trigger an agile transformation in a large
organization. To investigate this phenomenon, we launched a single case study of an organization rarely studied in
this context. This research deals with the IT department of the French central bank that is a public service organization.
Results contributes to understand the different decisions levels triggering a large implementation of an agile method
creating a new way of organizing a large IT department. Furthermore, the study allows to identify the phases of the
adoption process of an agile method allowing practitioner and researcher to know more about how new management
practices are adopted.
We present in next part a theoretical background about the emergence of the agile transformation topic among
researchers and we mobilize the management innovation literature to explain how we can investigate this
phenomenon. We continue by presenting the qualitative research design based on an in-depth case study, part four
presents an analysis of the case, and we finish with the practical implications and a conclusion dealing with future
works.
2. Theoretical Background
In 2016, Dikert et al., [4] proposed on their literature review to characterize the adoption of agile methods on the
large as a large-scale agile transformation. If few practitioners used the term to characterize the implementation of a
scaled framework at the project level, the agile transformation is a process where a large organization will change the
information systems development methods from a plan-driven process to an agile development on a large perspective
including changes in how different teams involved in projects are organized [7].
2.1. Agile transformation in large IT departments
Referring to Moe and Dingsoyr [8], the main questions to be addressed about large-scale agile transformation are:
Inter-team coordination, business agility, knowledge sharing and knowledge networks, and agile transformation, that
means according to them, how the deployment of agile methods can be organized? To succeed in implementing global
site adoption, organizations needs to define a process to deploy this transformation.
Paasivara and co-authors [5], analysed in a recent investigation, a large-scale agile transformation of an Ericsson
product development program developing a new information technology platform. They presented the steps taken, the
challenges faced, and the mitigating actions taken. They gave four lessons learned about an international

452 Akim Berkani et al. / Procedia Computer Science 164 (2019) 449–456
4 Akim Berkani et al. / Procedia Computer Science 00 (2019) 000–000
3.1. Case study introduction
We made the choice to set up the analysis of a unique case that is mainly based on the study of an IT department
from a public service organization, with different units of analysis: people working on this entity, the Project
Management Office (PMO), and projects in interactions with other business departments (figure 1).
Founded in 1800 by Napoleon Bonaparte, the Banque de France (BDF) is an independent institution governed by
French and European law. It is a member of the Eurosystem, which is the federal system comprising the European
Central Bank and the national central banks of the euro area. Its three main missions are monetary strategy, financial
stability and the provision of economic services to the community. This is a particularly original case, a s few studies
concern a large public administration whose projects are mainly made of a normative nature and controlled by an
European institution. The analysis object of the case is the BDF IT department. The IT department alone is composed
of more than 1200 people, 600 applications and 10 European projects.
3.2. Data collection
Data collection began in January 2018 and are still ongoing. During this period,
we’ve been in immersion at the headquarters of BDF. The adoption of a
management innovation being defined as a process composed of different phases
[2] we collected process data mainly from semi-structured interviews, participant
observations, event narrations and documents. We have systematically started our
research by putting in place a retrospective analysis to understand the antecedents
as we were coming during the adoption process. As a summary, we collected 18
semi-structured interviews, 4 non-participatory observations and we collected 20
documents (with almost 100 pages in total).
3.3. Data analysis
Data has been transcribed and compiled using NVivo software. We have them coded the different verbatims and
documents by coding via a multi-thematic approach [16] (Fig.2). This methodology is well adapted for heterogeneous
material and allows to avoid the circularity effect via a certain triangulation of the different data sources. According
Dumez [16], analysis based on coding can create a circularity phenomenon. In order to favour objectivity to understand
the phenomenon of the agile transformation, we reconstituted temporal mechanisms (figure 3) and conducted a process
analysis coupled with a temporal modelling of major events, decisions and actions triggered. It is finally important to
emphasize that our communication is part of the work of a thesis, analyses are still being formalized and results
presented are preliminaries.
Fig.1: Analysis objects
Fig. 2 Excerpt of the coding process
Excerpt fro m interviews - input Actor interviewed Fact analyzed Decision taken
Level of the actor taking
the decision
"The introduction of agili ty is not that recent, the first agile projects it was more
than ten years ago and it was on based on Scrum. The team had set up a platter
to gather the team. It had been rather successful."
IT department
manager
Introduction of a new
information systems
development method in
a project
Introduction of a new
information systems
development method
Developers and project managers actors
"I want to say that when I arrived it was a little desert, we read on the site
dedicated to the method while few people consulted, but we read on thi s site that
the method practiced at the bank was SCRUM , which i s not practicable for
example with deported service centers so everyone went there on hi s own
initi ative with more or less success."
Agile coach
Harmonization of the
different
Creation of the internal
agile project
management
methodology
Operational actors
"The adoption of agility i s not quite recent, what is more recent is the desire to
deploy it more massively through traini ng and coaching. We began by
supporting the deployment of our internal method because i t was rather
opportunistic before. "
IT department
manager
Passage d'une adoption
opportuniste à
l'officialisation de
l'adoption de l'agilité
First decision to sustain
the adoption of the
internal agile project
management approach
Middle management actor
"We are working on this roadmap, [...] On september 3 during the bank's
management committee, I wi ll present them the topics we will work on or we are
working to deploy the agili ty to all projects. Senior officials need indi cators,
steps to take, everything will have to be measured. "
IT department
manager
creation of the roadmap -
Organization of the
implementation
Validation of the
implementation
Executives actors

Akim Berkani et al. / Procedia Computer Science 164 (2019) 449–456 453
Akim Berkani et al. / Procedia Computer Science 00 (2019) 000–000 5
4. Results and analysis of the case
The analysis of the data allowed us to note that the adoption trajectory of agile methods in BDF is characterized by
three sequences including a crucial decision period (figure 3). The first sequence concerns the emerging introduction
of Scrum based on local adoption with no support. The second sequence is characterized by the creation of a first agile
project management methodology. Some support is offered to projects wanting to experiment agile method. The third
sequence is the official generalization plan of agile methods to all IT projects.
Fig. 3 Temporal template of the adoption process at BDF.
The IT staff are divided into three categories: those who join the BDF after graduating, those who are recruited
after more than 15 years of experience, and external consultants. This deliberate choice helps the company to have a
blended set of competencies in a field that is always renewing itself. As an aside benefit from this HR policy, new and
proven approaches to software development methodologies and tools are constantly put in practice in the IT
department by the new recruited people. It is in this context that Agile approaches were introduced little by little in
the IT department.
4.1. Sequence 1 - Emerging introduction of different agile frameworks (from 2006 to 2014)
The genesis of the introduction of agile practices (based on Scrum and eXtreme Programming) is the first sequence
of this trajectory. It started in 2006 and runs until 2014. Projects within BDF had to apply a first methodological
approach based on the Merise2 development method (clear green of figure 2). According to an IT manager: “Before
really adopting the agility, in fact, the first experiment dates from 2006 it was simply something that had been launched
then after, although we were happy, everything had been erased”. During this period, few projects introduced Scrum
2 Merise is a general-purpose modeling methodology in the field of information systems development, software engineering and project
management.
4 Akim Berkani et al. / Procedia Computer Science 00 (2019) 000–000
3.1. Case study introduction
We made the choice to set up the analysis of a unique case that is mainly based on the study of an IT department
from a public service organization, with different units of analysis: people working on this entity, the Project
Management Office (PMO), and projects in interactions with other business departments (figure 1).
Founded in 1800 by Napoleon Bonaparte, the Banque de France (BDF) is an independent institution governed by
French and European law. It is a member of the Eurosystem, which is the federal system comprising the European
Central Bank and the national central banks of the euro area. Its three main missions are monetary strategy, financial
stability and the provision of economic services to the community. This is a particularly original case, a s few studies
concern a large public administration whose projects are mainly made of a normative nature and controlled by an
European institution. The analysis object of the case is the BDF IT department. The IT department alone is composed
of more than 1200 people, 600 applications and 10 European projects.
3.2. Data collection
Data collection began in January 2018 and are still ongoing. During this period,
we’ve been in immersion at the headquarters of BDF. The adoption of a
management innovation being defined as a process composed of different phases
[2] we collected process data mainly from semi-structured interviews, participant
observations, event narrations and documents. We have systematically started our
research by putting in place a retrospective analysis to understand the antecedents
as we were coming during the adoption process. As a summary, we collected 18
semi-structured interviews, 4 non-participatory observations and we collected 20
documents (with almost 100 pages in total).
3.3. Data analysis
Data has been transcribed and compiled using NVivo software. We have them coded the different verbatims and
documents by coding via a multi-thematic approach [16] (Fig.2). This methodology is well adapted for heterogeneous
material and allows to avoid the circularity effect via a certain triangulation of the different data sources. According
Dumez [16], analysis based on coding can create a circularity phenomenon. In order to favour objectivity to understand
the phenomenon of the agile transformation, we reconstituted temporal mechanisms (figure 3) and conducted a process
analysis coupled with a temporal modelling of major events, decisions and actions triggered. It is finally important to
emphasize that our communication is part of the work of a thesis, analyses are still being formalized and results
presented are preliminaries.
Fig.1: Analysis objects
Fig. 2 Excerpt of the coding process
Excerpt fro m interviews - input Actor interviewed Fact analyzed Decision taken
Level of the actor taking
the decision
"The introduction of agili ty is not that recent, the first agile projects it was more
than ten years ago and it was on based on Scrum. The team had set up a platter
to gather the team. It had been rather successful."
IT department
manager
Introduction of a new
information systems
development method in
a project
Introduction of a new
information systems
development method
Developers and project managers actors
"I want to say that when I arrived it was a little desert, we read on the site
dedicated to the method while few people consulted, but we read on thi s site that
the method practiced at the bank was SCRUM , which i s not practicable for
example with deported service centers so everyone went there on hi s own
initi ative with more or less success."
Agile coach
Harmonization of the
different
Creation of the internal
agile project
management
methodology
Operational actors
"The adoption of agility i s not quite recent, what is more recent is the desire to
deploy it more massively through traini ng and coaching. We began by
supporting the deployment of our internal method because i t was rather
opportunistic before. "
IT department
manager
Passage d'une adoption
opportuniste à
l'officialisation de
l'adoption de l'agilité
First decision to sustain
the adoption of the
internal agile project
management approach
Middle management actor
"We are working on this roadmap, [...] On september 3 during the bank's
management committee, I wi ll present them the topics we will work on or we are
working to deploy the agili ty to all projects. Senior officials need indi cators,
steps to take, everything will have to be measured. "
IT department
manager
creation of the roadmap -
Organization of the
implementation
Validation of the
implementation
Executives actors

454 Akim Berkani et al. / Procedia Computer Science 164 (2019) 449–456
6 Akim Berkani et al. / Procedia Computer Science 00 (2019) 000–000
approach, but they had no official support.
To compete with other Central Banks bidding to the European Central Bank, the IT department decided in 2010 to
follow the CMMI3 framework to illustrate its continuous improvement commitment and to pursue a policy of
certifying its project managers following the PMBoK4 guide of the Project Management Institute (From Merise
methodology to the second project management methodology on figure 2). As started in 2006 some IT teams continued
to apply agile practices without putting a name on them. These personal initiatives were still not officially supported
by the IT department and were not part of the official methodology that teams had to follow (Ad hoc practices on
figure 2).
Two major ingredients act in this first sequence of the trajectory. The early unofficial adoption of agile practices
inspired from Scrum,eXtreme Programming by teams shows the operational introduction of agile pr actices. In 2010
the CMMI evaluation led to the creation of a new project management methodology that all new project manager had
to apply (Second predictive project management methodology on figure 2). In parallel, agile practices were present
without any internal recognition.
4.2. Sequence 2 - Experimentation and opportunistic use of the internal agile framework (from 2014 to 2018)
In 2014, a new CMMI assessment is launched and maintains the organization at a level 2. It means that projects
are "Managed", according to CMMI, this maturity level recognize that a discipline is established for each project and
materializes mainly by project plans (development insurance, quality assurance, configuration management). The
project manager has a major responsibility in level 2: he must define, document, apply and maintain his plans. From
project to project, he capitalizes and improves his project management and engineering practices.
Regarding actors, a new governor arrives end of 2015 with the desire to simplify the bank. He asked to different
middle managers to think about how their department could be simplified. Furthermore, at the same time in the IT
department, given the number of projects adopting agile practices and the heterogeneous diffusion of them, the Project
Management Office (PMO) decided to create an internal agile methodology based on different agile practices (Official
agile project management methodology on figure 2).
After being reassessed level 2 of the CMMI maturity evaluation in 2017 some measurement program was adopted.
It became clear that projects applying the official agile project management methodology were more successful. They
had fewer bugs in production and were delivering their software product with a better customer satisfaction. These
results were presented to the IT department manager who, in turn, pushed these results to the new Governor of the
Bank. It is thus interesting to note that the CMMI initiative is one of the roots enabling management understanding of
the benefits of agile practices.
It is important to notice that at this stage, the official agile project management methodology is more and more
experimented by teams. The internal framework is diffused through training and coaching. Concurrently the Waterfall
project management methodology continued to grow as it was seen by the management as a way to better control the
output of software initiatives (more documents, more metrics).
This second sequence shows new ingredients of the adoption trajectory. The first is the decision from the PMO to
create an official internal agile framework. Starting from the observation of many different projects who used different
agile practices, this first decision led to officialise the adoption of an agile project management methodology. Second,
the CMMI assessments makes emerge the positive impact of the internal agile project management methodology on
project that allowed to compare between the waterfall methodology. At the same time, a new Governor is appointed
at Banque de France. He quickly made “digital” and simplification of organizational processes, a strategic priority.
4.3. Implementation decision phase
After the CMMI maturity level in 2017. The waterfall methodology was still in use in most software initiatives.
3 CMMI, the acronym for capability maturity model integration, is a reference model, a structured set of best practices designed to apprehend,
evaluate and improve the activities of engineering companies.
4 Project Management body of Knowledge.

Akim Berkani et al. / Procedia Computer Science 164 (2019) 449–456 455
Akim Berkani et al. / Procedia Computer Science 00 (2019) 000–000 7
But in parallel, the internal agile methodology has been simplified, giving birth to a set of 14 practices. These 14
practices were redacted by the same people who previously wrote the agile project management methodology (from
the PMO), namely PMP and PMI-ACP certified project managers with more than 20 years of experience. This led to
an important simplification of the project process previously introduced to (falsely) reassure the upper management
used to work with waterfall approaches.
During this decision phase, a major event will impact the next stages on the adoption process. The director of IT
projects, based on the CMMI maturity assessment evidence, reported that the main quality improvements results of
projects managed under the agile approach to the Governance of BDF. The Governance body proposed that all new
projects and a third of the applications should follow an agile approach, and the governor decided to support this
strategic shift and give mandate to the director of the IT projects (who was named just before the decision sequence)
to lead this transformation and make a proposal of implementation plan.
“There is a mandate at the highest level to launch an agile transformation of the bank. The IT department is also
forced to transform to achieve this challenge” (According to the PMO manager).
The new director of IT projects department started to monitor good practices from other organizations and involved
all the different actors composing the PMO. In parallel the PMO was restructured and a new PMO manager named,
with the aim to better support agile implementation. The generalization plan also made the proposal of the creation of
an agile center of excellence and opened recruitment of agile coaches.
Different ingredients sustaining the decision process are important in this phase. The CMMI evaluation highlights
the positive impact of the agile method on projects. When these facts were presented to the Governor, he made it a
priority topic for the IT department. Another special ingredient of this phase is the reorganization of the PMO, and
the IT projects department, even if it does not depend directly from the strategic decision to adopt agility, it made
easier to introduce new approaches. Finally, this decision sequence ends with the validation of the implementation
plan by the Governor in summer 2018.
4.4. Launching the implementation
Once the generalization plan was accepted by the governor, the implementation started officially in September
2018. And, a new approach to synchronize multi-project teams in agile was adopted, the Scaled Agile Framework
(SAFe) was introduced during a seminar to experiment it for several IT project teams. To widen the agile practices
and to support the upscaling for all internal members of the IT department, PMO actors decided to experiment a scaled
framework. They were certified via the SAFe framework (SAFe Program Consultant) to try it in larger projects.
Next steps of the generalization plan are divided in 5 different pillars: organization, culture, agile governance,
business deployment and technical agility. The agile center of excellence and globally the IT project division are
preparing to address a cultural awareness campaign, business alignment with agility, and special work like
reorganizing special process such as purchasing management or human resources.
5. Practical implications:
The analysis of the case allows practitioners to identify the essential ingredients that allows to accelerate the large
implementation of an agile method. We showed how this decision is the consequence of a process that is sustained by
different ingredients such as the positive impacts of a favourable environment (viral adoption of agile practices,
positive experimentation in projects), the availability of qualitative and quantitative data (CMMI appraisal) justifying
the new approach, and newcomers at all levels supporting the agile adoption (BDF governor, IT project department,
head of IT PMO service) facilitating the promotion and validation of new approaches. Finally, the implication of the
head of the BDF governance body in the decision process created a real implementation dynamic.
6. Conclusion
As a reminder, the goal of this research is to answer the following question: what are the triggers of a decision that
will enable moving from experimentation to wide generalization of an agile method? The qualitative analysis based
6 Akim Berkani et al. / Procedia Computer Science 00 (2019) 000–000
approach, but they had no official support.
To compete with other Central Banks bidding to the European Central Bank, the IT department decided in 2010 to
follow the CMMI3 framework to illustrate its continuous improvement commitment and to pursue a policy of
certifying its project managers following the PMBoK4 guide of the Project Management Institute (From Merise
methodology to the second project management methodology on figure 2). As started in 2006 some IT teams continued
to apply agile practices without putting a name on them. These personal initiatives were still not officially supported
by the IT department and were not part of the official methodology that teams had to follow (Ad hoc practices on
figure 2).
Two major ingredients act in this first sequence of the trajectory. The early unofficial adoption of agile practices
inspired from Scrum,eXtreme Programming by teams shows the operational introduction of agile pr actices. In 2010
the CMMI evaluation led to the creation of a new project management methodology that all new project manager had
to apply (Second predictive project management methodology on figure 2). In parallel, agile practices were present
without any internal recognition.
4.2. Sequence 2 - Experimentation and opportunistic use of the internal agile framework (from 2014 to 2018)
In 2014, a new CMMI assessment is launched and maintains the organization at a level 2. It means that projects
are "Managed", according to CMMI, this maturity level recognize that a discipline is established for each project and
materializes mainly by project plans (development insurance, quality assurance, configuration management). The
project manager has a major responsibility in level 2: he must define, document, apply and maintain his plans. From
project to project, he capitalizes and improves his project management and engineering practices.
Regarding actors, a new governor arrives end of 2015 with the desire to simplify the bank. He asked to different
middle managers to think about how their department could be simplified. Furthermore, at the same time in the IT
department, given the number of projects adopting agile practices and the heterogeneous diffusion of them, the Project
Management Office (PMO) decided to create an internal agile methodology based on different agile practices (Official
agile project management methodology on figure 2).
After being reassessed level 2 of the CMMI maturity evaluation in 2017 some measurement program was adopted.
It became clear that projects applying the official agile project management methodology were more successful. They
had fewer bugs in production and were delivering their software product with a better customer satisfaction. These
results were presented to the IT department manager who, in turn, pushed these results to the new Governor of the
Bank. It is thus interesting to note that the CMMI initiative is one of the roots enabling management understanding of
the benefits of agile practices.
It is important to notice that at this stage, the official agile project management methodology is more and more
experimented by teams. The internal framework is diffused through training and coaching. Concurrently the Waterfall
project management methodology continued to grow as it was seen by the management as a way to better control the
output of software initiatives (more documents, more metrics).
This second sequence shows new ingredients of the adoption trajectory. The first is the decision from the PMO to
create an official internal agile framework. Starting from the observation of many different projects who used different
agile practices, this first decision led to officialise the adoption of an agile project management methodology. Second,
the CMMI assessments makes emerge the positive impact of the internal agile project management methodology on
project that allowed to compare between the waterfall methodology. At the same time, a new Governor is appointed
at Banque de France. He quickly made “digital” and simplification of organizational processes, a strategic priority.
4.3. Implementation decision phase
After the CMMI maturity level in 2017. The waterfall methodology was still in use in most software initiatives.
3 CMMI, the acronym for capability maturity model integration, is a reference model, a structured set of best practices designed to apprehend,
evaluate and improve the activities of engineering companies.
4 Project Management body of Knowledge.

456 Akim Berkani et al. / Procedia Computer Science 164 (2019) 449–456
8 Akim Berkani et al. / Procedia Computer Science 00 (2019) 000–000
on a longitudinal study shows the importance of the top management support to generalize an internal agile method.
Following the classification given by [17], it appears that the agile transformation studied in this paper can be
described as a gradual introduction approach in a subset of the enterprise followed by a big bang approach imposed
to the full enterprise. Finally, this work questions the linearity of the adoption process of a management innovation.
Initially considered as purely sequential and linear, we could apply in this case the analysis of the adoption process to
each identified project management methodology that would be a more cyclic process.
Acknowledgments
We gratefully acknowledge the helpful availability of all the actors interviewed and we thank all the team of the
Banque de France Project Management Office giving us time to share their experiences.
References
[1] Diegmann, P., Dreesen, T., Binzer, B., & Rosenkranz, C. (2017) “Journey Towards Agility: Three Decades of Research on Agile Information
Systems Development.” International Conference on Information Systems. Seoul, Korea.
[2] Boehm, B. (2002) “Get ready for agile methods, with care.” Computer 35(1): 64–69.
[3] Hobbs, B., & Petit, Y. (2017) “Agile Methods on Large Projects in Large Organizations.” Project Management Journal 48(3): 3–19.
[4] Dikert, K., Paasivaara, M. and Lassenius, C., (2016) “Challenges and success factors for large-scale agile transformations: A systematic
literature review.” Journal of Systems and Software (119): 87–108.
[5] Paasivaara, M., Behm, B., Lassenius, C., & Hallikainen, M. (2018) “Large-scale agile transformation at Ericsson: a case study.” Empirical
Software Engineering 23(5): 2550-2596.
[6] Damanpour, F. and Wischnevsky, D. J. (2006) “Research on innovation in organizations: Distinguishing innovation-generating from innovation-
adopting organizations.” Journal of Engineering and Technology Management 23(4): 269-291.
[7] Paasivaara, M. et al. (2013) ‘Integrating global sites into the lean and agile transformation at Ericsson” Proceedings - IEEE 8th International
Conference on Global Software Engineering, Bari, Italy.
[8] Moe, N.B. and Dingsøyr, T., (2017) “Emerging research themes and updated research agenda for large-scale agile development.” In Proceedings
of the XP2017 Scientific Workshops on - XP ’17 1–4.
[9] Gerster, D. and Dremel, C. (2019) “How Enterprises Adopt Agile Structures: A Multiple-Case Study’, in HICSS 2019 Proceedings 4957–4966.
[10] Alqudah, M. and Razali, R. (2016) “A Review of Scaling Agile Methods in Large Software Development”, International Journal on Advanced
Science, Engineering and Information Technology 6(6), 828-837.
[11] Damanpour, F. and Schneider, M. (2006) “Phases of the Adoption of Innovation in Organizations: Effects of Environment, O rganization and
Top Managers.” British Journal of Management 17(3): 215-236.
[12] Hecker, A. & Ganter, A., (2013) “The influence of product market competition on technological and management innovation: Firm-level
evidence from a large-scale survey.” European Management Review, 10(1): 17–33.
[13] Damanpour, F. (1991) “Organizational Innovation: A meta-analysis of effects of determinants and moderators.” Academy of Management
Journal 34(3): 555-590.
[14] Eisenhardt, K. M. (1989) “Building Theories from Case Study Research.” The Academy of Management Review 14(4): 532-550
[15] Yin, R. (2009), Case study research: design and methods (4th Revised edition ed.): SAGE Publications Inc.
[16] Dumez, H. (2016). Méthodologie de la recherche qualitative: Les questions clés de la démarche compréhensive. (ed) Vuibert.
[17] Julian B., Noble J., Anslow C. (2019) Agile Practices in Practice: Towards a Theory of Agile Adoption and Process Evolution. In: Kruchten
P., Fraser S., Coallier F. (eds) Agile Processes in Software Engineering and Extreme Programming. XP 2019. Lecture Notes in Business
Information Processing, (355). Springer, Cham