Towards Organizational Self-awareness A Methodological Approach to Capture and Represent Individual and Inter-Personal Work Practices David Miguel Rolo Vicente Dissertação para Obtenção de Grau de Mestre em Engenharia Informática e de Computadores Júri Presidente: Professor Doutor Pedro Sousa Orientador: Professor Doutor José Manuel Nunes Salvador Tribolet Vogais: Professora Doutora Helena Sofia Pinto Novembro 2007
108
Embed
Towards Organizational Self-awareness - ULisboa · This thesis proposes an approach towards organizational self-awareness through, semi-automatic and real time, capture and analysis
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Towards Organizational Self-awareness
A Methodological Approach to Capture and Represent
Individual and Inter-Personal Work Practices
David Miguel Rolo Vicente
Dissertação para Obtenção de Grau de Mestre em
Engenharia Informática e de Computadores
Júri
Presidente: Professor Doutor Pedro Sousa
Orientador: Professor Doutor José Manuel Nunes Salvador Tribolet
Vogais: Professora Doutora Helena Sofia Pinto
Novembro 2007
i
ii
ACKNOWLEDGMENTS
“Behind every successful man is a woman…”, Groucho Marx
Special thanks to Professor José Tribolet for trusting me this work. Without his confidence it was
impossible to reach the maturity stage required to develop a study of this type. His abilities to
teach, discuss, and hold your interests are innumerous and impossible not to admire. More than
gratitude for all his dedication and availability that were crucial to sustain this thesis
development.
Thankfulness to Professor Marielba Zacarias for all assistance provided. Her extraordinary
availability and dedication were key issues that successfully direct this thesis. There is any
possible way to thanks her for all her indispensable advices, persistence, and comprehension.
To Dr. Luis Fialho, director of Moviflor sales department, for giving me the opportunity to
perform one of this thesis case studies in his department. Thanks to all the exceptional central
sales unit people, for their devotion and interest by my work, who were crucial to reach this
study results, and by the superb work environment provided during all time spent in Moviflor‟s
office.
To INATEL for the opportunity to implement the Portal developed during the thesis, supporting
the second case study. Thanks to change management team, and all organization people
involved, for their remarkable contribution to this project. For the friendship, support, and
advices given during INATEL case study, a special gratefulness to Eng. Carla Pereira.
My entire career has been supported by a very special person, my Mother. Considering her as a
reference has been the followed recipe to achieve many of my personal and professional
successes. For all this, tanks mom. Special thanks also to my family, particularly to my brother
and my father.
All this work would have been much more difficult without my friends. They were the ones with
who I relieve and discuss. Between them, I have to do a special gratefulness to a person who I
can‟t live without, my best friend Joana. Her honesty and amazing skills to understand and defy
me have been crucial factors for my ongoing success.
iii
To the most important women of my life, my Mother and Joana
iv
v
ABSTRACT
Nowadays organizations are subject to constant changes. Most of these are imposed by the
need to adequate their way of managing the business to market demands. Examples of these
changes are business processes optimizations or organizational restructures. To plan and
perform them it is necessary to understand organization‟s current state, only by perceiving it
become possible to correctly analyze its needs. A significant part of this knowledge is in every
single person that belongs to organization.
Each person knows what they need to work. Amongst them, documents, tools (including
computer related ones), other people and organizational departments are fundamental to
achieve goals defined for each diary task. Thus, to correctly work people must be aware of their
role in organization, as well as the role performed by others. This awareness is described as
Organizational Self-Awareness and it is studied by Organizational Engineering discipline. The
objective of this engineer field is to provide organization instruments that allow capturing and
representing people‟s knowledge of its parts, in order to synchronize the image that they have
of the whole.
This thesis proposes an approach towards organizational self-awareness through, semi-
automatic and real time, capture and analysis of individual and inter-personal work practices, as
well as their representation using enterprise models easily perceived by organizational people.
Further, these models can also be used to facilitate the investigation of possible optimizations
People, Individual and Inter-Personal Work Practices
vi
vii
RESUMO
Actualmente as organizações estão sujeitas a constantes mudanças. Estas são em grande
parte ditadas pela necessidade de adequarem a forma como gerem o seu negócio às
exigências do mercado. Constituem exemplo destas mudanças as optimizações dos seus
processos de negócio ou as alterações da sua estrutura organizacional. Para planear e realizar
aquelas mudanças é necessário conhecer e compreender o estado actual da organização,
apenas assim se torna possível analisar correctamente as suas necessidades. Parte relevante
deste conhecimento encontra-se em cada uma das suas pessoas que compõem a
organização.
Cada pessoa sabe o que necessita para realizar o seu trabalho. Documentos, ferramentas
(entre elas informáticas), outras pessoas e departamentos da organização são fundamentais
para atingir os objectivos delimitados para cada tarefa realizada diariamente. Deste modo, as
pessoas devem estar conscientes do seu papel na organização, assim como do
desempenhado pelos outros com que se relacionam. Esta consciência é designada como
Consciência Organizacional e é estudada pela Engenharia Organizacional. O objectivo deste
ramo da Engenharia é fornecer à organização instrumentos que permitam capturar e
representar o conhecimento que cada pessoa possui das suas partes, de moda a, sincronizar a
imagem que têm do todo.
Esta dissertação propõe uma abordagem que motiva a Consciência Organizacional, através da
captura e análise, semi-automática e em tempo real, de práticas de trabalho individuais e inter-
pessoais, bem como a sua representação mediante modelos facilmente perceptíveis pelas
pessoas da organização. Estes modelos pretendem igualmente facilitar o estudo de possíveis
optimizações e, posteriormente, gerir a sua implementação.
Palavras-chave: Engenharia Organizacional, Consciência Organizacional, Modelação de
Organizações, Pessoas, Práticas de Trabalho Individuais e Inter-Pessoais
viii
ix
INDEX
ACKNOWLEDGMENTS ............................................................................................................................ II
ABSTRACT ............................................................................................................................................. V
RESUMO ............................................................................................................................................. VII
Figure 39. Relation between „as is‟ and „to be‟ process presented in Portal. ............................. 75
Figure 40. Procedure list displayed for each „to-be‟ activity in Portal. ....................................... 76
Figure 41. Brief representation of INATEL approach. ............................................................... 77
Figure 42. Approach proposed by merging Moviflor and Inatel case study results. ................... 79
xiv
xv
ACRONYMS
BPM Business Process Management
BPMN Business Process Modeling Notation
EA Enterprise Architectures
IS Information Systems
OE Organizational Engineering
OSA Organizational Self-Awareness
UML Unified Modeling Language
1
INTRODUCING THE STUDY
“Processes don't work; people do", Peter Fingar
In the previous quote Peter Fingar, considered one of the Business Process Management
gurus, states that are organizational people that makes processes work. This study focus the
capture and representation of personal and interpersonal work practices as important
information that each organization person should be aware of in order to improve their
conscientious about the organization that they work for.
In this section we will introduce the thesis, describe its objectives, explain the methodology
followed, and present an overview of document structure.
1 Introduction
During the last decades many approaches came out proposing ways to improve the efficiency
and effectiveness of organizations (BPMG, 2005 pp. 8-12). Amongst them we distinguish
process reengineering, suggesting the reorganization of enterprise along its workflow, and
Business Process Management (BPM), concerning the alignment between business processes
and organizations goals. Their ambition was similar and process oriented. However, in both
approaches people were left to second plane, as a result problems such as misalignments
between business processes and Information Systems (IS) and resistance to change, just to
name a few, came out leading to unsuccessful results.
It is true that organizations should be managed considering their wider processes, instead of
each department functions, like BPM suggest. However, most of the activities that constitute
their processes are performed by people, which play specific roles in these. Thus, people‟s work
should also be considered as an important issue. In addiction, and consistent with BPM
proposes, people must have a common idea of the organization as a whole, and of their
importance in it, in order to improve their efficiency. After all, when people don‟t understand the
usage of what they are doing, their interest in it decreases and this affects negatively the quality
of their work.
Organizational Engineering (OE) emerged in order to find a solution to solve the previously
stated problems. This discipline aims to reduce the existing gap between soft and hard
sciences, being the soft related to social and management concerns, and the hard regarding the
technical aspects of the organization. As part of OE, the concept of Organization Self-
awareness (OSA) has been proposed (Zacarias, et al. a, 2007). This states that, in the same
2
way people know „who they are’ and „what, and how, are they doing‟, they should be aware of
„who they (and others) are in an organization‟ and „what, and how, they (and the others) are
doing‟.
Answering to the previous questions can influence organization people to become more aware
about their, and others, roles in organization. Enterprise modeling represents an efficient way to
achieve OSA goals (Zacarias, et al., 2007). This is justified by the fact that through models
people can easily understand the organization, promoting their awareness about it. Enterprise
Architectures are examples of these models, however usually they only illustrate organization‟s
processes widely, identifying their activities by department, instead of the roles played by each
person in them.
Every single person has distinct ways of working. This means that, in order to achieve the same
goals, each person can perform different tasks and use dissimilar resources. Thus, it becomes
important to capture and represent personal and interpersonal actions, providing through this a
real and consistent base to wonder about organization‟s actual state, commonly named „as is‟.
Another important concern in nowadays organizations is to have most of the previously
described information in real time. The speed of change imposes a real time way of managing
the business in order to easily react to changes. Thus, new approaches to provide this
information should consider real time as an important requirement while producing it. By real
time, we intend to refer to the right time, which mean the specific moment when the information
is needed by the organization e.g. to do management decisions.
In this study we propose an approach to capture, analyze, and represent personal actions using
concepts such as contexts, enterprise ontology, business processes, enterprise modeling, and
real time business. In addition, we also consider the representation of enterprise future state,
usually described as „to be‟, as an important information to avoid common change problems,
such as misalignment between processes and business (e.g. incomplete functionalities in new
ERPs), and resistance to change played by organization people directly affected by the
alterations. The approach suggested and discussed was developed during two case studies
performed in distinct organizations, in different situations.
The first case study occurred in the sales department of a furniture retail company, Moviflor,
with the aim of optimize employees work practices through their standardization. In order to
reach this, every single person registered most of their diary actions using a web form
developed to this propose. Each of the form fields, and its contents, were initially defined during
a bootstrapping phase where every person was involved and contributed. Once registers were
analyzed, personal tasks diagrams were created and presented to each person using a web
portal, where she was asked to validate if they correspond to the reality. This case study
3
reached successful results, justified by the fact that, once finished, sales department people
conducted, by their own, a reengineering of their work practices.
The second case study was done on a social services provider, INATEL, that was performing a
drastic organizational restructure and an IS renewal. The developed work focused the
presentation of organization business processes diagrams using a web portal, similar to the one
created in Moviflor case study, and their further validation by every person involved in them. In
addiction was also included the possibility to visualize the „to be‟ state of each „as is‟ diagram.
By involving every single person in the organizational change, capturing and representing their
roles in each business process, this case study have resulted in the successful development of
organizational self awareness of INATEL employees.
1.1 Research Objectives
As the title of this thesis suggests, providing a way to enhance the accomplishment of OSA
goals is our main objective. However, in order to reach it, this was divided into the following
ones: (1) find a way to capture personal and interpersonal work practices in real time; (2)
represent each person work practices in a way that most of the people easily can understand
them; and (3) using the previous models, suggest a way to support organizational changes in
order to reduce the common difficulties associated with them, which means preparing the „to
be‟.
1.2 Thesis Methodology
The methodology defined for this thesis is based on four main steps: (1) define study goals; (2)
research related work; (3) plan and execute case studies; and (4) achieve an discuss
conclusions considering the achieved results (Figure 1). Each step is further described.
Figure 1. Thesis Methodology.
Goals were initially defined, however while researching some issues have been identified and
added to primary objectives. The reason for this change is that, before starting the research, our
initial knowledge about the state of the art was insufficient to correctly define precisely every
specific need.
Define
Goals
Research
of Related
Work
Plan and
Execute
Case
Studies
Conclude
about
Thesis
Results
4
The research of related work was done using subjects related to OE discipline, our research
context. Initially only BPM was considered, however in order to find a way to capture personal
and interpersonal work practices other important references have been studied. These were
enterprise ontology, context-based approaches, and information gathering techniques, which
have proved to be indispensable instruments to achieving our goals.
While executing case studies, most of the previously researched concepts and approaches
have been applied. Still, a constant need to go back to research stage was required. The last
step was performed along the execution of case studies. It was during them that most of the
conclusions have been achieved and described. However, when analyzing the results,
conclusions attained iteratively influenced the work that was being done.
1.3 Thesis Outline
The relevant value-added features of this thesis are the activities and results achieved during
case studies. Thus, this document structure follows each of these along their execution. It is
divided in five main parts, each separated in a set of related sections.
1. Introduction. In this part we introduce the study, describe its goals, the methodology,
and outline document structure.
2. Related Work in Organizational Engineering. This part states OE references used
during this thesis. It is composed by: (1) definition of OE, describing its goals; (2)
explanation of OSA related work, where a recent proposed framework and an approach
to apply it based on enterprise ontology and contexts are described; (3) state of BPM
art, providing a brief contextualization in this subject, including enterprise modeling and
methodologies proposed to implement BPM; and (4) analysis of some Information
Gathering Techniques, interviews and observations, these last ones based on
ethnography.
3. Problem Declaration. Here we summarize the major contributions of related work and
identify relationships and gaps between them. This is done in order to specify our
problem, and to explain the approach followed to solve it.
4. A Step Towards Organizational Self-awareness. In this part we describe each case
study. The first was performed in Moviflor, a furniture retail company, and the second in
INATEL, a social services provider.
5. Conclusions. The last part is where the new approach is proposed, and where
conclusions and issues are described.
6. Annex A – Moviflor Inquiries. In this annex we present Moviflor inquiries, used to
obtain people‟s opinion about the work developed during Moviflor case study. There are
two distinct inquiries, one answered by Moviflor sales central team, and the other filled
by sales department director.
5
RELATED WORK IN
ORGANIZATIONAL
ENGINEERING
"You can engineer the enterprise just like you can engineer anything else", John Zachman
Despite the fact that Organizational Engineering is a recent discipline, whose major growth
occurred during the last decade, there are many important references that can‟t be despaired.
This section describes the related work in Organizational Engineering used to perform this
study, focusing concepts such as organizational self-awareness, business process
management and enterprise modeling, emphasizing a context-based approach based on recent
researches. In addition to these, a brief description of information collection techniques is done.
2 Organizational Engineering Definition
Organizations are facing increasing complexity as result of the speed of change and growth of
effective and aggressive competition (Eriksson, et al., 2000). Methods and tools, to align
business processes with organization‟s strategic goals represent a new form of knowledge,
which is a needed instrument to manage these changes and improve real-time management
abilities. This knowledge is structured in the Organizational Engineering (OE) discipline, also
known as Enterprise Engineer.
Regarding organizations as complex entities that deal with contrasting concepts such as
people, value chains, business processes, information systems and technology (Sousa, et al.,
2007), OE aims to enhance the alignment between these notions in order to (1) achieve better
understanding of the enterprise, (2) grant more flexibility in organizational design, (3) provide a
solid foundation for reorganizing the business, and (4) support information systems‟
development.
People‟s knowledge about the organization and how it influences their behavior has been
concern of recent studies (Zacarias, et al. a, 2007). These focus the importance of people‟s
organizational awareness and how it can be used to optimize individual and inter-personal work
practices. As a result, the concept of organizational self-awareness (OSA) has been proposed.
6
Implementing it represents an essential pre-requisite for organizational effective action, decision
making and learning processes (Zacarias, et al. a, 2007).
3 Organization Self-Awareness
Human beings are by nature, self-awareness beings. This capacity lets us know who we are,
how we do things and what are we (and others) doing at any particular moment (Zacarias, et al.
a, 2007). Whereas this ability is innate in individuals, OSA must be built and maintained by
continuous interactions among members in order to apply it to all organization.
The reason why people‟s knowledge is the basic concern of this concept is justified by the fact
that they are the only ones capable of sense making in the organization (Zacarias, et al. b,
2007). Sense making is defined as structuring unknown contexts and/or actions and assigning
them with meaning. It is distinguished from other explanatory processes such as understanding
or interpreting by the following characteristics: the process of sense making is social, grounded
in identity construction, retrospective, focused on extracted cues, ongoing, driven by plausibility
rather than accuracy and enactive (Zacarias, et al. b, 2007). Therefore, we can describe sense
making as the consciousness that people develop of the organization as a whole and of their
place in it (Zacarias, et al. b, 2007).
The concept of OSA is characterized in two dimensions, the individual and the organizational
one. The individual dimension refers to the capacity that individual members have answering
the questions such as; who am I in this organization?, how are things done here?, what is the
organization - as a whole - doing now?. The organizational dimension refers to the combination
of human or automated agents, resources and procedures that provides organizations with the
necessary intelligence for dealing with questions such as; who are my members?, how do they
do things?, what are they doing now? (Zacarias, et al. a, 2007). An organization is self-aware
when these two dimensions are aligned.
3.1 Modeling to Organizational Self-awareness
In order to apply OSA, enterprise models are suggested (Zacarias, et al. c, 2007). Models
represent an important tool to describe the enterprise from different points of view, such as
functional, process, data, and economic (Sousa, et al., 2007). Amongst the several definitions to
describe a model, we have considered “model as a theoretical construct that represents
physical, biological or social processes, with a set of variables and a set of logical and
quantitative relationships between them” (Sousa, et al., 2007).
Enterprise modeling has been studied by Information Systems (IS) and Artificial Intelligence (AI)
fields. These models are commonly referred as Enterprise Architectures (EA) or Enterprise
7
Ontologies, respectively. They are described using more formal syntax and semantics, enabling
its processing by automated agents and reducing inconsistent interpretations (Zacarias, et al. c,
2007). EA and enterprise modeling languages are analyzed further (section 4).
As a result of recent research (Tribolet, 2006) a new meta-plane representation was defined
with the aim to clarify the relationship between OSA and enterprise models (Figure 2). This
meta-plan is represented by an agent that (1) creates and develops a model of the world, (2)
influences this world, and by other side, (3) its decisions are influenced by the world. Through
this emerge the concept of OSA as a continuous process of construction, maintenance and
change of shared images about: (i) the way the organization and its agents act between them,
(ii) the organization and agent role played in each instant, and (iii) how to control and remodel
the organization in real-time.
Figure 2. OE Meta-Plane (Tribolet, 2006)
OSA can also be related with the concept of mental models. Proposed as one of the five
disciplines described by the learning organization1 concept, mental models are assumptions,
generalizations, or even images that influence how people understand the world and react in
them.
1 The learning organization emerged during the 1980s based on a Donald Schon‟s idea that describes an organization that learns as its workers improve their knowledge, aiming to survive to the speed of change and competitive markets. This concept have been target of recent studies and become more popular with the book “The fifth discipline” (Senge, 1990).
8
Enterprise models represent an essential communication tool in supporting and enhancing
OSA. These are illustrated in the models‟ “circle”, highlighted in Figure 2, and are useful to
provide ways of thinking about the organization in order to support management principles.
Effective supports to OSA require generic models of agents‟ activities, resources, and their
relationships, but it is not enough. There is a need to know how specific agents accomplish their
work and what they are doing at particular moments, and to provide ways of evaluating how
work practices change in time. More specifically, it is necessary to answer to the following
questions (Zacarias, et al. a, 2007):
Which roles plays agent X?
How does agent X perform activity Y? Which rules follows?
Which agents interact more frequently?
How does agent X interact with agent Y? Which interaction rules govern group G?
At a particular time interval t:
o What role plays agent X? Which activity(ies) performs? Which resources use?
o With which agents interacts?
o Which event(s) trigger agent X's role?
o Which rules govern how agent X manages his different roles, activities and
resources?
In order to answer to the previous quoted questions, and to provide a way to represent,
communicate and discuss organizational conditions through appropriate enterprise models,
Zacarias et al. b (2007) propose a framework composed by an agent architecture and ontology.
3.2 A Framework Towards Organizational Self-Awareness
OSA should be considered simultaneously as an individual and collective phenomenons that
must be supported by a continuous communication among organizational agents. In order to
develop it, organizations need to improve capabilities of continuous sensing, learning and
adjusting to the dynamics of their environment.
Zacarias, et al. a (2007) suggest an architecture and ontology, departing from an ontological
position where organizations are regarded as complex and adaptive socio-technical systems.
The challenge is to model the organization for its self-awareness, using as background
organizational and IS field‟s approaches, in order to capture: (1) organizational structural and
dynamic aspects, (2) routines and decision-making processes, and (3) its formal and informal
sides. Furthermore, another important concern is to consider organization development, and
consequently motivate the continuous reverse engineering of the organization. (Zacarias, et al.
a, 2007)
9
3.2.1 Fundamental Concepts
The framework proposed is based on some fundamental concepts of the CEO framework
(Sousa, et al., 2007). This is further described in section 4.4.1.2, even though a brief
introduction is provided subsequently. CEO framework distinguishes three main concepts that
mostly contributed to the approach proposed, these are: (1) activities, (2) entities, and (3) roles.
Entities can be classified in two different kinds: resources and agents.
Activities describe what organizations do, and are identified with verbs (e.g. buy, create, and
hire). Entities (resources) are the things relevant for organization operations and are identified
with nouns. In this framework, entities are synonym of organizational resources. Resources can
be persons, machines, places, concepts or capabilities. They may be physical or abstract,
inanimate or animate, technical or social (e.g. furniture, order, employee, and programming
skill). (Zacarias, et al. a, 2007)
Joining organizational verbs (activities) and nouns (resources) form organizational predicates
(e.g. buy furniture, create tour, and hire employee). Organizational predicates provide a more
complete specification of what the organization does (Zacarias, et al. b, 2007).
Another important concept is the notion of Agents. They are regarded as physical and animate
resources with special capabilities that enable them to: (1) perform, coordinate and change
activities; (2) provide, consume, manage and change resources; and (3) monitor, coordinate
and change their own activity and the activity of their agents. The agents can be characterized
as human (persons, dyads, groups and whole organization), automated or semi-automated.
Automated and semi-automated agents are machines or human-machine combinations
(Zacarias, et al. c, 2007).
Agents are autonomous, interactive, adaptive, proactive, rational, unpredictable (capable of
non-deterministic behavior) and coordinative (agent are capable of coordinating themselves, as
well as other agents). In this ontology, agents are identified with nouns (e.g. Sales Department,
Susana, and Sales Manager) and they are the subject of organizational predicates. Agents
possess a set of capabilities, enabled by skills (know-how) or knowledge (know-what).
Linking agents (nouns), activities (verbs) and resources (nouns) we form organizational
sentences, i.e. we define what the organization does and who does it (e.g. Sales Department
buys furniture, Susana creates an order, and Sales Manager hires Susana).
Considering agents as a special kind of resources lead into the agent-resource duality. Entities,
such as the agents described before (i.e. Prof Sales Department, Susana, and Sales Manager),
may be also operate as resources. In this case, they are not the subject of the sentence.
10
Rather, they are part of the predicate. For example, in the sentence “Sales Manager hire
Susana”, the agent is the Sales Manager and Susana represents a passive resource.
Role defines the observable behavior of an entity in the scope of particular interaction contexts.
Agents play several roles and interact with other agents through these roles. (Zacarias, et al. c,
2007)
The definition of context remains dependent on its application area (Zacarias, et al. c, 2007). In
Engineering, and particularly in AI, context is viewed as a collection of things (sentences,
associated to some specific situation (environment, domain, task, agents, interactions,
conversations, etc). Analyzed in the Cognitive Sciences filed, context represents the set of all
entities that influence human (or system's) behavior on a particular situation. Although,
departing from Social Approaches, context is regarded as networks of interacting entities
(people, actors/agents and artifacts).
The considered definition in this approach defines contexts as a network of interacting agents
playing specific activities and resources. It also reflects the state of participating agents, their
interactions, and the set of rules governing these interactions.
Figure 3 illustrates the fundamental concepts of the framework and the existing
interrelationships between them. The agent plays specific roles in specific contexts. These roles
are related to some activity or resource.
Figure 3. Zacarias, et al., (2007) framework fundamental concepts.
Considering the relationship between the fundamental concepts, organizational agents can be
described by the set of roles played when related to the other concepts. The agent, by itself,
can be characterized by its ability to learn, self-coordinate and act. When related to resources,
the agent is capable to produce/consume, manage and assume a (re)design role. The capacity
11
to perform, coordinate and also (re)design activities characterizes the role played by the agent
when related to the activity concept. These relations are depicted in Figure 4.
Figure 4. Agent Basic Role Set. (Zacarias, et al. c, 2007)
3.2.2 Agent Architecture
The agent architecture proposed is based on three interdependent layers; (1) action, (2)
decision making and (3) change/learn (Figure 5). Each one of these layers describes different
faculties that the agent performs.
Figure 5. Agent Architecture. (Zacarias, et al. c, 2007)
The Action Layer aims to capture agent interaction strategies. These strategies are considered
as behavior patterns describing the guided way an agent react according to different situations.
As an example, when a person sees a cup falling, he will try to hold it because she is used to do
it.
12
The measurement of agent reflection, when facing unknown situations, is concern of the
Decision-making layer. Providing an agent dynamic representation, based on the events that
trigger state changes, this layer captures rules that define the utilization of specific interaction
strategies. These are agent activation strategies. Continuing the previous example, the reason
why the person will try to hold the cup, it‟s because once in the past she saw it falling and
braked up. In that time a bad performance measure was associated to that action.
Focusing the agent reflexive behavior, the change/learn layer aims to process the way
interaction and activation strategies change trough the time. This practice is based on strategies
performance measures defined in the previous layer. Learning only occurs when new or
redesigned strategies improve the performance of previous ones. Finishing the example stated
above, after the cup broken the person learned that letting it fall was wrong and learn to don‟t
repeat the same action again.
The relation between the three layers is illustrated in Figure 6. When acting, agents decide what
to do according to activation strategies given by its deliberation capabilities, and interaction
strategies defined by his reflexive and learning capabilities. The change of agents‟ strategies
occurs when they analyze their performance. This analysis is mainly based on subjective and
tacit performance measures, although they could be inferred through propagation, reduction
and elimination of strategies (Zacarias, et al. c, 2007).
Figure 6. Relation between Agent Architecture‟s layers. (Zacarias, et al. c, 2007)
3.2.3 Model Basic Architecture and Ontology
Organizations can be modeled as a network of interactions between autonomous resource and
activity-related agents (Zacarias, et al. c, 2007). Resource and activity-related agents, as well
13
as the interactions between them, are modeled on the basis of the agent three-layered
architecture depicted in Figure 7.
Figure 7. Model basic architecture and ontology of each layer. (Zacarias, et al. c, 2007)
The relation between agent roles and the agent architecture layers depicted in the figure above
states that in the action layer agent provide and/or consume resources and interacts according
to different contexts, while performing its activities. In terms of OSA, the action layer
representations describe how the organizations and its agents do things. (Zacarias, et al. c,
2007)
In the decision making layer the agent acts as a resource manager and activity coordinator.
These activities are performed when activation strategies, previously described also as rules,
are invoked in a particular context. This layer captures processes such as planning or
scheduling, triggered by activation rules that may cause agents to decide the activation of a
different context.
After performing the roles described in the previous layers, agent realizes change/learn roles.
Therefore he selects activities and resources according to the performance measures
previously defined for the interaction and activation strategies. Thus, the change/learn layer
aims to capture the re(design) of these strategies. The modeling of this layer has not been
address yet, currently the authors‟ concern is in the detection of changes in the previous layers.
Figure 8 depicts the ontology resulting for the two first layers, action and decision-making,
illustrating the difference between the agent strategy employed and how that influences its
14
relation with other concepts. Mediating artifacts are defined as resources that support and
constrain agent interactions. (Zacarias, et al. c, 2007)
Figure 8. Ontology details for the first two layers. (Zacarias, et al. b, 2007)
3.2.4 Applying the Framework in Different Organizational Levels
Human agents are typically studied at individual, inter-personal, group and organizational levels.
However, when considering business processes approaches they are typically related to
organizations or organizational units. As consequence, the concept of activity is related to
groups and the concept of task is associated with individual or interpersonal levels.
15
Figure 9. Basic architecture and concepts at different levels of detail. (Zacarias, et al. a, 2007)
Even considering different levels, the previously described ontology can be applied. The main
difference on the usage in each level remains on the universe of discourse i.e. a different set of
nouns, verbs, attributes, state variables and rules (Zacarias, et al. a, 2007). Verbs and nouns
used in different processes, resources and their corresponding contexts (e.g. order furniture)
are different for those defining activities (e.g. develop orders‟ application) or tasks (e.g.
elaborate order report).
Interactions at all levels are mediated by contexts that differ from layer to layer i.e. commitments
within personal action contexts obligations refer to to-do lists; in inter-personal contexts, they
mean inter-personal commitments; in group-level contexts, they signify formal or informal
agreements; finally, in organization-level contexts, commitments refer to inter-organizational
contracts. (Zacarias, et al. a, 2007)
3.3 Model Acquisition Approach
The application of the previous described conceptual framework (Architecture and Ontology of
Organizational Agents) is sustained by the Model Acquisition Approach, recently proposed by
Zacarias, et al. b (2007). This approach has already been applied in two documented case
studies (Zacarias, et al. b, 2007).
Classified as bottom-up and context-based approach, it is based on three main steps: (1) collect
action of a group of subjects, (2) identify and analyze action-layer behavior i.e. predominant
actions and resources of personal and inter-personal contexts, and (3) infer decision making
layer behavior i.e. find personal and inter-personal context activation patterns.
16
The approach is composed by five different types of activities, which are: (1) bootstrapping, (2)
action capture, (3) context discovery, (4) context analysis and (5) context integration. In addition
to these steps, the acknowledgment of the change/learn layer behavior is given by a cyclical or
continuous usage of this approach. As illustrated in Figure 10, there isn‟t any specific order
recommended to perform most of the activities involved.
Figure 10. The context-based approach proposed. (Zacarias, et al. b, 2007)
3.3.1 Bootstrapping
Started with an observation period, whose result is the definition of the different action and
resource types for each agent, this phase produces a set of actions and resources. This set is
discussed further, requiring activities such as the presentation of the set obtained to subjects,
followed by its validation. The subjects referred are agents that can be individuals or teams.
The action set is composed by verbs such as inform, read, calculate, refresh, change, and
create. Resource repository includes information items such as documents and notebooks, and
also refers to software/hardware components used to perform each action, e.g. calculator,
Microsoft Excel, SAP R/3.
At the end of this phase, the normalization of the action and resource set are done. This is a
needed step in order to support their migration into a database and facilitate future analysis.
Bootstrapping can be described bootstrapping in three main steps: (1) define the action and
resource set for each agent; (2) discuss and validate the set proposed with the subjects
involved; and (3) analyze the set of actions and resources, normalizing the items list.
3.3.2 Capturing and Structuring Actions
Traditional modeling approaches describe tasks, activities or processes with predicates
composed by a verb followed by an object (Zacarias, et al. b, 2007), e.g. create request, print
17
invoice, solve claim. These approaches belong to BPM and are further detailed (section 4.4).
The main issue of these descriptions is the lack of subject, not including the agent.
The Model Acquisition Approach captures actions using the structure of organizational
sentences previously described (section 3.2.1). These sentences are composed by triples
subject-verb-object, where subject identifies the agent, the verb identifies the action type, and
the object identifies the resources used or produced while performing the action (Zacarias, et al.
b, 2007) e.g. Susana create order in SAP R/3 M21 menu. The structure of the actions captured
is depicted in Figure 11.
Figure 11. The structure of actions. (Zacarias, et al. b, 2007)
Each sentence is registered considering it chronological order and using the action and
resource items normalized during bootstrapping. This allows the creation of an agent action
structured and ordered repository. However, communicative actions can further be structured
using speech theory, meaning that actions can be written using a subject-verb-action form
(Zacarias, et al. b, 2007) e.g. Maria request (Susana create order in SAP R/3 M21 menu).
3.3.3 Context Identification and Display
Once analyzed the different actions and resource types registered for each agent, it is possible
to group them according to their contexts. However, constructing this sets require some
knowledge about the organization analyzed. Otherwise the contexts‟ set won‟t make any sense.
After creating the groups, they are presented to subjects that validate (and may regroup) them.
It‟s also subjects‟ responsibility to label each group according to their knowledge about the
actions contemplated. An example of this phase results is illustrated in Table 1. Some Personal
Contexts obtained during the case study. (Zacarias, et al. b, 2007)Table 1.
18
Table 1. Some Personal Contexts obtained during the case study. (Zacarias, et al. b, 2007)
3.3.4 Context-based Analysis
Identifying, characterizing and labeling contexts permit their usage as unit of analysis. The
identification of personal contexts allows a variety of context-based depictions (Zacarias, et al.
b, 2007).
Contexts are considered as networks of interacting agents playing related specific activities and
resources. Thus, analyzing the repository created with the set of agent‟s actions and resources,
labeled as contexts, allows discovering of information such as context switches (Figure 12) or
the number of actions for each context.
Figure 12. Context switches of the agent Mariana during the case study. (Zacarias, et al. b, 2007)
3.3.5 Context Integration
This phase is characterized by the production of models that facilitate the comprehension of
agents‟ tasks. The context integration is a human process whose result can be used as a tool to
actualize existing enterprise models.
19
One of the main problems of current modeling approaches (section 4.4) is the constant need for
its updating. Models that result from the context integration facilitate the construction of task
models and assist updating other existing enterprise representations, such as the ones further
described (section 4.4). Figure 13 illustrates a context-based interaction network obtained from
the analysis of a context repository.
Figure 13. A context-based interaction network obtained from case study. (Zacarias, et al. b, 2007)
3.4 Conclusion
Despite the fact that OSA is a recent area of OE, its importance demands an urgent concern
and appliance by organizations. This could be a possible way to follow in order to facilitate
organizations reaction to the constantly changing of the market. Once every single person in
organization is aware about his work, and how that influences the whole company, processes
will become more flexible and will easily adapt to changes.
However, the model acquisition approach proposed requires human intervention in most part of
the activities, making hard the achievement of real time concerns. Thus, its appliance becomes
expensive in a personal and temporal way. This represents one of the goals defined for this
study.
Enterprise ontology is an important concept described in this section, providing a way to
describe it using a common language. It will be using this ontology that we will further define a
manner to capture personal and interpersonal actions.
Justified by the fact that it is out of this study scope, there isn‟t any section dedicated to the
research advances in the automated context discovery field. Though, there are already some
studies with successful results (Zacarias, et al. b, 2007).
20
4 Business Process Management
During the last decade some approaches came out to facilitate the accomplishment of OE
goals. An important one is Business Process Management (BPM) including methods,
techniques and tools to support the design, enactment, management, and analysis of business
processes (Fingar, et al., 2003 pp. 73-97). This concepts and manners focus mainly
organization‟s business process, considering the organization as a whole and proposing the
definition of enterprise wide processes.
4.1 Brief History Contextualization
Since the eighties, process reengineering changed the way organizations manage their
processes. Connected with it, emerged the concept of Business Process Re-engineering (BPR),
defined as the re-organization of an enterprise along the flow of work, generating value sought
by the costumer (Frost, et al., 1996). Its focus are the analysis of business process
requirements, with a little transformation or rework as possible, and the description of business
process modeling techniques, which give a basis for assembling business solutions. Industrial
organizations used this method to improve their business processes‟ performance, but these
efforts were unsuccessful, failing to spot agility concerns in order to support the ongoing
change. Nowadays, reengineer is associated with drastic short term changes in processes, not
concerned with the way people react to it. This non collaborative reengineer can justify its
previous failure in many organizations.
In the last decade, Enterprise Resource Planning (ERP) systems were developed and adopted
by many organizations. Their intent was to provide everything that an organization needed to
manage its business, including the complete support of its business process. The problem was
that these systems were inflexible and most of the time obligated the change of organization‟s
business process in order to adapt to them, when what was expected, and needed, was the
reverse. The concept and usage of workflows also emerged during this decade. Their main goal
to represent all the business processes was very ambitious and important to the development of
a process-centric way of manage. One of the main issues here were the gap between what the
languages provided and what were needed to represent organizational requirements.
Recently successful efforts are being made in order to align processes with organization goals.
Process mapping tools capture and manage enterprise processes in an editable form, for more
flexible manipulation and subsequent analysis. However, these tools can't carry process models
to executions, being this one of the main ambitions proposed by the BPM.
21
4.2 Business Process Definition
In order to realize the ideas suggested by BPM, the concept of business process should be
clearly described. Business process can be defined based on several definitions, some of them
are:
Series of inter-related activities that cross functional boundaries with individual inputs
and outputs (Llewellyn, et al., 1999);
Groups of activities performed in response to a set of related business events (Frost, et
al., 1996);
A collection of activities that takes one or more kinds of inputs and creates an output
that is of value to the costumer (Hammer, et al., 2001);
The manner in which work is organized, coordinated, and focused to produce a
valuable product or service (Laudon, et al., 2006);
A collection of applications, workflows and associated IT resources that are required to
perform a set of related tasks that are major importance to a business (Associates,
2000).
The value added to customer and the help provided in order to accomplish organization‟s
business goals are common issues in these definitions. Merging both definitions, we can define
a business process as a coordinated set of activities that add value to the costumer and support
achievement of organization business goals (Sousa, et al., 2007).
Business processes importance to the organization is justified by the fact that they literally
define it, representing its critical source of competitive advantage and market differentiation
(Fingar, et al., 2003 p. 71). BPM suggests methodologies and frameworks with the aim of
helping organizations to improve the way they deal with their business processes. Examples of
these goals are the real time management capability and the alignment between organization‟s
business units.
By analyzing the concepts directly related with business process, it‟s easy to achieve a very
relevant one, the notion of business rule. To avoid misunderstandings, it‟s important to
distinguish business rules from business process: “business rules are the know and business
processes are the flow” (BPMG, 2005 p. 98). Every organized business process has its
business rules. In a practical approach, a business rule is what guides the business in running
its day-to-day operations, in order to achieve the right balance between business processes and
business goals.
Here, the concept of entity is also important. An entity is a stakeholder that plays a role in an
activity, according to several rules. This role is the observable behavior of an entity in the scope
22
of a specific collaboration context. It aims to separate the different concept that arises from the
collaborations between the entities fulfilling an activity.
Summarizing, and always considering that we are describing the business process in a wide
perspective, it‟s possible to define a business process as a set of activities, where one or more
entities plays a role (Sousa, et al., 2007), according to several rules, in order to reach business
goals.
4.3 Defining Business Process Management
BPM paradigm aims to provide the ability to use the information gathered through advanced
management techniques. This concept proposes a model to predict the impact of IT changes on
business, and by other hand, the impact of business process changes on IT environment.
Through this, BPM can be defined as a synthesis of process representation and collaboration
technologies that removes the obstacles blocking the execution of management intentions. This
concept emerged from the convergence of management theory - total quality management, Six
Sigma, business engineering and general systems thinking - with modern technologies -
application development, systems integration, computation, service-oriented architecture,
workflow, transaction management, XML and Web services (Fingar, et al., 2003 p. 71).
Briefly, it can be said that BPM strategically aims to:
Represent and align supply chain processes, project planning processes, learning
processes and product-data management processes, and manage all alike, in real-time;
Support the answer to „what if‟ questions and allows specification of the new „to-be‟
processes;
Reorients Information's Technology (IT) activities to the trajectory of the business
process, preventing the overtime projects and decrease its rate of failure.
A suggested approach to represent BPM value is the Dynamically-Stable Enterprise (DSE) from
Vasile Bucioman-Coman and Adrian Sahlean (BPMG, 2005 pp. 52-78). The DSE is a unifying
enterprise model that integrates all the enterprise elements, from the simple to the highly
sophisticated. In this model, the enterprise is seen as an abstract entity, composed by business
entities (knowledge, product2 and costumer), business processes (operations and process
management), value chain (product and information flows), decision-chain (drives change), and
2 It‟s important to refer that the product described as a business entity can be also seen as a service. The reason for this designation results from the fact that this view is based on Porter‟s value chain, which considers product development as organization‟s main interest.
23
the technology that enables processes to be effective and flexible. This value chain is driven by
the business plan model, the available resources and the technology lifecycle.
As its name suggests, the change is an important feature in DSE. In order to deal with it,
flexibility and efficiency are considered its main drivers. The flexibility challenge is to engineer
the enterprise for change. The efficiency is used as economic progress‟ main indicator. The
more stable an enterprise is, the more efficient are its operations, but the more difficult is to deal
with change. This, justify the needed alignment between these drivers, easily achieved through
BPM principles.
In order to correctly represent the importance of change and the value of BPM in the
organization, it‟s suggested to represent the enterprise as a system. This system is divided in
two dimensions, dynamic and static, revolved around three main business entities: knowledge,
product and costumer (Figure 14). The dynamic dimension is given by the behavior of the
enterprise processes, covering mainly the transformations of knowledge into product, and of
product into costumer value. The static dimension is given by the structure of processes,
including the static processes of knowledge lifecycle, product lifecycle and costumer lifecycle.
Figure 14. The DSE proposed view of the enterprise as a system (BPMG, 2005 p. 74).
Considering aspects like organizational knowledge, as a result of operations, and time, through
a cyclic representation, the DSE proposed view it‟s directly related to the concept of OSA
previously described (section 3). However, in opposite to OSA framework (section 3.2), in this
view is given more emphasis to end-to-end organizational business process instead of people
roles while performing them.
24
To put BPM into practice two main steps should be taken: process-centered organizational
design and process-centric business information systems. Thus, the design of organizational
processes can only be reached through modeling techniques. The models resulting from this
exercise are used as a pre-requisite to implement the information systems.
4.4 Modeling the Organization
Models are the basic requirement to develop the important BPM process of modeling. It refers
to the process of generating a model as a conceptual representation of some phenomenon.
Enterprise Modeling applies the previous concepts to take an enterprise-wide view of an
organization, which can be used as a basis for taking decisions. In order to achieve, use, and
maintain such view, integration, communication, flexibility and support are required (Uschold, et
al., 1998). There are two categories to classify the enterprise modeling actions (Breton, et al.,
2000): descriptive modeling, as the pure act of modeling organizational behavior, and active
modeling, as the act of constructing executable systems, business process systems (BPS), from
these models. However, descriptive modeling is always a prerequisite to active modeling. BPM
focus both, with special emphasis on the last one as it is the future of it.
In the context of the DSE concept mentioned before, enterprise can be viewed as a system
where the IS are part of it. To deal with organization‟s complexity, modeling practices suggest
the simplification of the reality. This means that, as prerequisite, some information will have to
be dismissed. To deal with this loss, the correct level of abstraction should be chosen, and to do
it correctly, meta-planes, and meta-models, are becoming a common practice in OE. These new
modeling practices also emerged in order to fix the previous known disappointment of modeling
in organizations. The failures can be summarized in a Gartner Group research (1997) that
identified "Nine Reasons Why IS Organizations Do Not DO BPM3":
Business units will not make the effort
We tried CASE and it did not like it
We do not have the time
Business units tell us what to build; we do not question them
We cannot keep business and IT models in synch
Business changes too quickly to model it
Is applications development modeling not enough?
Is prototyping not enough?
Business modeling is more trouble than it is worth
3 In this context BPM means Business Process Modelling. This is a common practice in modelling approaches, easily justified by the fact that the BPM pragmatic, defined as Business Process Management, is very recent.
25
In order to avoid the reasons pointed above, the usage of meta-models has been proposed.
Enterprise Architectures (EAs) are part of them. An EA results from the continuous process of
representing and keeping aligned the elements that are required for managing the organization.
This is used to describe business structures and processes that connect them (Sousa, et al.,
2007). Process meta-models define process fundamental concepts and their relationships,
integrating static and dynamic aspects, and organizing the various models of processes. These
are also used to state who is doing what, when, how and why.
4.4.1 Enterprise Architectures
EAs aim to allow a holistic view of the organization through its decomposition in different
dimensions, and to guarantee that technologies are the ones that support the business. In order
to apply an EA many frameworks have been proposed. Each EA Framework is divided in
different levels layers, which describe the enterprise in an explicit and semiformal way from
different points of view. The different existing frameworks differ from the level of abstraction they
use, and the industry‟s branch defined as target. Only two of them are then described, focusing
the view where organization‟s business processes are located, and how they are represented.
The reason for selecting the following EAs is because they were the ones that mostly
contributed for this study. We suggest (Telematica Instituut, 2002) as a good source for others
EA descriptions.
4.4.1.1 Zachman’s Framework
The framework proposed by John Zachman, also known as “The Framework for Enterprise
Architecture” derives from equivalent structures used in other disciplines to design complex
physical products (e.g. buildings or airplanes). This framework applies to the enterprise a logical
structure for classifying and organizing its management and to develop its systems. The
generalization of the process analysis done by this framework allows its appliance to every
business area.
26
Figure 15. Zachman‟s Framework, “The Framework for Enterprise Architecture” (ZIFA, 2006).
The model proposed is based on six simple questions: “what”, “how”, “where”, “who”, “when”
and “why”. These questions are answered by five distinct perspectives: visionaries, executive
leaders, architects, engineers, implementers and workers (Figure 15). Each representation (and
each perspective) presents a different set of constraints (Goethals, 2003).
Several lines and cells should be filled in order to reply to the previous six questions and as a
guide to the tasks and activities that should be performed to employ Zachman‟s Framework. At
the end of each task, documents and artifacts should be prepared, describing the analyses
done in the best way. Models are examples of these artifacts. However, the models described in
each cells are generally more hypothetical than empirical (Telematica Instituut, 2002). The
reason for this can be justified by the fact that the main ambition of this framework is to allow its
application to a wide range of enterprise‟s areas. The absence of a common modeling
language, that can be applied to all business areas, justifies the not suggestion of one in this
framework. This would restrict its range.
Business processes can be found in the “how” column, filled with its description, model and the
IT resources that supported them. The first three cells of this column contain the list of the
enterprise processes, the model of its business processes and the application architecture that
supports the processes (Lemos, et al., 2006).
In a succinct way, this is a simple, comprehensive and neutral framework (Telematica Instituut,
2002). It‟s simple because it is a non technical and purely logic approach. Addressing the entire
enterprise makes it comprehensive, and finally, because it is totally defined apart of tools or
methodologies, it‟s neutral. This means that it does not define a meta-model to integrate the
27
information within each cell. An important issue is the fact that it doesn‟t describe how to trace
the information between cells, nor the model that should be used to describe each cell.
4.4.1.2 Organizational Enterprise Centre Framework
The Organizational Enterprise Centre (CEO) framework provides an architectural model for
process-oriented organizations. This framework uses as references some others such as
Zachman‟s framework and TOGAF. A distinct feature from the other ones is the fact that this
one uses a specific modeling language to represent the EA, i.e. UML. The EA proposed is
divided in five architecture views, represented using UML packages, which separately address
different concerns related to the organization. These architecture views are organizational,
business, information, application and technology (Sousa, et al., 2007).
The organizational architecture deals with the aspects directly related with the organization,
including concepts such as the enterprise mission, vision and strategy, and the definition of
organizational units. The business architecture defines the business processes of the
organization.
The information architecture describes what organization needs to know to run its processes
and operations as described in the business architecture. It defines a view on the business
information that is totally independent from system and technology.
The application architecture fulfils two major goals: supporting the business requirements and
allowing efficient management of the organization‟s entities. Finally, the technological
architecture represents the technologies behind application implementation, as well as the
infrastructure and environment required for the deployment of the business process support
systems.
The ambition of the previous architectural views is to describe and relate the fundamental
concepts that, as a whole, describe the enterprise architecture. Each architectural view is
composed by different parts. For example, the technological encompasses three parts: IT
infrastructure block, IT platform block and IT application block. The different architectural views
and its components are represented in Figure 16.
28
Figure 16. CEO architectural views, its components, and the relationships between them (Sousa, et al., 2007).
As mentioned before, the core concept of the business architecture is the business process. In
this architecture business process is defined as a set of activities that adds value to some
costumer, whether internal or external to the organization. The representation of the process,
using UML, is also described.
To represent the process, the framework uses the concepts of activity, role and entity. The
activity describes business roles required that are played by the organization entities, these
roles are: actor, resource, and observable state (Figure 17). Depending on the purpose, several
UML diagrams are used. Just to name a few, to represent activity done in a process, an activity
diagram it‟s used, to represent the relationships between the different roles, a package diagram
it‟s chosen, and the relation between role types it‟s done through class diagrams.
Figure 17. Relationships between activity, role and entity (Sousa, et al., 2007).
Beside all the architectures views and the way to represent its components, an important
subject is also considered, the alignment between them. This alignment is accomplished
through integrity rules defined between sets of EA concepts. The business case alignment is
motivated by the following causes: (1) business and information architecture are aligned when
29
people have information they need to run the business; (2) business and application
architectures must be aligned in order to enable the information systems to provide services to
the business.
This framework can be summarized by the fact that it emphasizes the traceability, alignment,
and assessment between enterprise concepts, facilitating their reuse and independent co
development. However, it defines no methodology to implement it, although, with the language
to model and the alignment proposed, this last issue should be chosen according to
organization‟s realities.
4.4.2 Business Process Modeling Languages
The missing aspect to describe in this modeling section is the business process modeling
languages. Modeling languages aim to represent the real process components and its behavior,
providing various concepts that, when combined, result in its representation. This representation
should be easily understood by those who are involved in the process and by those who will use
the model to improve it. Examples of these are IT people, whose main goal is to provide
efficient support to enterprise‟s business processes.
In nowadays, the goal is to create/use a language that can be applied to every type of business
reality and understood by all people involved in them. Only two languages of a huge available
list will be described. Besides technical reasons that have influenced this choice, an important
characteristic is the fact that they are implemented in accessible software, which proves their
present wide usage.
Unified Modeling Language and Business Process Management Notation
These languages are presented together because they use similar representations and, at this
time, they also belong to the same developer, OMG. The language names are Unified Modeling
Language (UML) and Business Process Modeling Notation (BPMN).
UML, as its name suggest, is a language that allows the representation of a wide range of
concepts in many different subjects. This is done through different available diagrams and
supported by the fact that it can be extended, allowing the definition of missing aspects. UML is
able to represent the different process views proposed by process meta-models, such as:
requirements, structure, content, behavior, information, stakeholder, and instance (Table 2).
Currently this language is in version 2.1.1. An example of a UML business process modeling is
depicted in Figure 18.
30
Figure 18. Example of business processes modeled using UML.
BPMN was created with the specific goal to represent business processes, allowing the
representation of activities‟ lifecycle (Figure 19). Its main limitation is associated with the restrict
process views that can be represented with it: behavior and instance. An important usage of
BPMN is the fact that it has an execution language associated with it, Business Process
Execution Language (BPEL), allowing its automatic conversion into code standards and
motivating its usage in business process systems.
Figure 19. Example of business process modeled using BPMN.
The existing relation between these two languages, its diagrams, and artifacts is represented in
Table 2 using the behavior process view proposed by Jon Holt (2005). It‟s clear that actually
these modeling languages are very similar, using the same objects to represent related objects.
Even so, in this study the chosen language was BPMN, not only justified by the fact that it‟s
specifically oriented to business process modeling, but also because it uses a cleaner
representation, easier to understand by non IS people.
Analyze Needs Create Order Send Order Receive Order
Process Order
Inform Client
Create Order
Client Company Supplier
available
not availableC
lien
t C
om
pan
yS
up
plie
r
available
Analyze
NeedsCreate Order Send Order
Receive Order
Process Order
Inform Client
Create Order
31
Table 2. UML 2.0 versus BPMN 2 process representation.
Process View
UML Activity Diagram BPMN Diagram
Behavior Activity. Stakeholders shown as swim lanes, activities as activity invocations, artifacts as objects.
Business Process Diagram. Stakeholders shown as lanes (pools and lanes), activities as activities, artifacts as data objects.
4.5 Implementing BPM
After answering BPM „what‟ question, this section aims to explain the methodologies proposed
in order to answer the „how‟ question. Implementing BPM is itself a definable business process,
one that has steps, order, importance, key elements, conditional routing, common objects,
known resources and business rules (BPMG, 2005 p. 111). It‟s basically a business process
that enforces empirical requirements while providing the flexibility to adapt to the specificity and
nuance of each organization's needs and situation. Depending on the author‟s base studies,
methodologies can be more IT or management oriented. However, there are some best
practices that should be considered by both (BPMG, 2005 pp. 24-26):
Start with the end in mind: a clear and a shared vision are required when starting.
Performance driven: every process, in order to be manageable, must be measurable
and driven towards the attainment of agreed performance objectives. „What you
measure is what you get'.
Stakeholder-driven: processes should focus on the stakeholders and their
requirements. The criteria for process performance must start and end both with their
needs and desires.
Criteria before decisions: early process steps in every process must be built in to get
acceptance of criteria to be applied in later ones.
First things first: after thoughtful design, all the steps should be done not skipping
conduction good planning and preparation work. When designing process the
methodology must have an adequate consideration for whole process design, not just
the visible part at the end.
4.5.1 LEARN Method (Jorge Coelho)
Jorge Coelho recommends some requirements for an effective methodology to implement BPM,
introducing concepts such as Organizational Therapy and Business Object approach (BPMG,
2005 pp. 119-130). The idea is to implement a new management business model that ensures
better performance from a learning organization perspective. This methodology concerns OSA
as an important issue to reach successfully its goals. The requirements proposed are:
Focus on the strategy and the organizational as a whole;
Use an Organizational Therapy approach;
32
Pay attention to the management of knowledge;
Model the Process-Centric Enterprise Architecture in a systematic and Business Object
Oriented way;
Deploy the strategy using a Process-Centric Enterprise Architecture;
Define a Continuous Improvement Management Model and the respective
implementation team;
Use interactive and real time techniques.
The methodology states that firstly, and before initiating any project, is necessary to understand
organization‟s strategy and its business improvement objectives. After this, these processes
should be placed in the enterprise architecture. In the case that there isn‟t any enterprise
architecture yet, two steps could be followed, a top-down approach or a middle down one.
Top-down approach would define firstly the context of the processes to improve, and later start
the modeling exercise. The middle down approach will define the first level of processes and
then drill down to find all the processes to be aligned and improved. Using this approach will
avoid the development of organizational improvements by function or department, which usually
results in the “isles of management” problem.
The point here is to avoid the identification of the „as is‟ situation to model processes and then
to analyze them. Proposing the identification primarily of the the target processes in enterprise
architecture and improve them in order to reach the project objectives defined initially.
Organizational Therapy requirement stresses the importance of the convergence between the
different views that each person have of the organization. To avoid the misalignment between
the different pictures that each person has, Jorge Coelho proposal starts with workshops where
each person do an „as is‟ analysis of the organization in order to, with all the information
obtained, fulfill the enterprise architecture.
To implement BPM, knowledge management should be considered as an important issue. It
should deal with explicit and tacit knowledge. However a big slice of this knowledge is tacit. In
order to correctly create a knowledge repository, where it‟s possible to store and manage
Human Capital, several important practices should be considered. These practices include an
important one, the creation of interactive models during workshops and interviews with the aim
to validate the knowledge that is being transmitted, and to align it with the one already existent
in this repository (Figure 20).
33
Figure 20. Central Repository proposed by Jorge Coelho‟s Methodology.
Reached the previously requirements, it‟s now time to define the processes in terms of status of
Business Objects. A Business Object is an entity that gathers all the information needed to
support to a stimulus. Basically, when a stimulus is received it‟s created a process to respond to
it, the Business Object is associated with the stimulus received and it‟s given the same name to
it. The stimuli from stakeholders give the first level of processes, which decompose the process
according to the number of business objects. When a level it‟s reached, where a process has
only one business object, this phase stops. After this the process should be decomposed into
activities, being these a set of tasks needed to manage a control point in a process.
The deployment of the strategy using a Process-Centric Enterprise Architecture has as pre-
requisite the consensus about organization actual situation, the „as is‟. This will be used to help
reaching the „to be‟ state, corresponding to the initial proposed goals. The strategy should be
obtained using the decision levels of architecture in an objective way. In this phase each
manager gets the objectives and goals associated with the processes that have been put under
his responsibility. Through this the strategy should be defined, in an objective way, where its
measures should be based on Business Object status, as they represent control points in the
processes.
Finally, it‟s important to refer that all the results provided by the previous requirements
described before must be supported by a common repository. This repository should be
available to all the intervenient to store information about the organization, therefore to support
the communication, and to reinforce a unique organization view.
The methodology proposed by Jorge Coelho is a very complete one, including a whole
description of which steps that should be taken, who should be involved in them, and how they
should be supported. Although, the enterprise architectures mentioned are not the same as the
ones previously described (section 4.4.1). Instead, the models produced by this methodology
34
only represent processes in a broad view. Thus, individual and interpersonal work practices are
not concerned, restricting its appliance in an OSA perspective.
4.6 Conclusion
BPM is a process-centric approach to the organization that aims to align important concepts,
such as supply chain processes, project planning processes, learning processes and IT. In
addiction to this, BPM propose a way to manage all the processes alike, and in real-time,
supporting the answer to „what-if‟ process and allowing the specification of the new to-be
processes.
By the concepts and definitions stated along this section, it is possible to assert that BPM came
to stay. The concerns about (i) real time management, (ii) the suggestion of organizational end-
to-end business processes definition, and (iii) processes representation using models to
facilitate its comprehension, are some of important features provided by BPM.
Gathering all notions back together, in order to fulfill some of the existing misalignment between
them, and always considering OSA as a key point, will contribute for the effective and
successful application of BPM in today‟s organizations.
5 Information Gathering Techniques
The constant need to reach the alignment between the IS and the organizational needs
contributed for the development of many techniques in order optimize requirements gathering
process. When concerning analysis of agent activities or even defining organizational business
processes, these techniques can be very useful. Nevertheless, they have in common the fact
that they all want to produce something based on the information acquired from the
organization.
Amongst all the available techniques, two were chosen. The first is a much known one and it is
used in many subjects apart from IS development, such as human resources or strategic
consultancy, usually it is named as interview. The second one, less famous and not so broad, is
entitled ethnography.
5.1 Interviews
Despite the fact that the interview technique is broadly used in many different subjects, the
reference considered, and consequently the description stated, is based on a software
requirements approach (Kotonya, et al., 1998).
35
Interviewing is a personal technique and requires the minimum presence of two people, the
interviewer and the interviewed. There are two main kinds of interviews, structured, previously
planned and strictly followed, and none structured, that is very similar to an informal
conversation. Both of these can be used in order to obtain the needed information, the choice
should be done considering aspects such as the time available, the number of persons
involved, the complexity of the information wanted, amongst others.
5.1.1 Preparing the Interview
Before starting an interview, a previous preparation phase is suggested. As an interviewer, this
should be used to (1) contextualize yourself about the your proposals through the reading of
related material; (2) establish interview goals; (3) decide who to interview; (4) prepare the
interviewed person/people; and (5) decide the type of questions and their structure.
There are two major different kinds of questions, open and closed. Open questions are
characterized by the fact that there isn‟t any predefined schedule. In closed questions the
number of answers is initially restricted, these are appropriate when precise information is
required.
The main advantages of open questions are: the interviewed gets more comfortable, uses the
vocabulary of the interviewed, generates new questions, provides more detail, and is more
interesting for the interviewed. On the other hand, this choice can result in problems, such as:
many details without relevance, losing the control of the interview, the answers taking too long,
and look like it wasn‟t prepared.
Closed questions can be good because they save time, go straight to the point, grant the control
of the interview, and allows the coverage of many relevant subjects. However, they can be
painful to the interviewed; and miss some important details.
As mentioned before, structuring the interview should also be done before starting the interview.
There are three different ways to structure an interview, pyramid, funnel, and diamond. The
pyramid starts with a specific question and ends with a generic one. The funnel is the inverse,
begins with a generic question and ends with a specific one. Finally, the diamond combines
both of the previous, consuming more time than the others.
5.1.2 Best Practices During the Interview
Once the interview started, there are some things that should be taken in consideration, such as
how to register the information gathered, how to conduct the interview, and how to behave
during the meeting.
36
There are several ways suggested to register an interview, for example an audio recorder or a
note book. When recording the conversation, the interviewed should be informed previously.
In order to collect all the needed information, there are some best practices that should be
followed by the interviewer. Before the interview he should (1) confirm the appointment; (2)
dress in an appropriate way, preferentially using the same style of the interviewed
person/people; and (3) arrive on time. Start the interview (1) acknowledging the interviewed, (2)
remembering the defined goals, and (3) informing interviewed how it will be registered. At the
end of the interview, (1) make a summary, (2) ask for doubts, and (3) acknowledge for the spent
time.
5.2 Ethnography
Genzuk (2004) presents an introduction and overview of ethnography and its applications in the
fields of social sciences. It states that “ethnography is a social science research method”, from
the fields of anthropology and sociology. Typical ethnographic research employs three kinds of
data collection: interviews, observation, and documents. These in turn produces three kinds of
data: quotations, descriptions, and excerpts of documents, having as result one product:
narrative description. (Genzuk, 2004)
Besides this definition, ethnography can be stated as having the objective of capturing customs,
beliefs, and behaviors of a group of individuals. The ethnographer spends a reasonable amount
of time (sometimes years, in case of anthropologists) with the individuals, taking detailed notes
of their actions and practices. That information is processed later on to reveal the “structure,
organization and practices” of the group studied (Sommerville, et al., 1993).
5.2.1 The ethnographic method
According to Genzuk (2004)the ethnographic method is characterized by the following features:
Individual‟s behavior and actions are studied in his natural context, without need for
virtual or simulated environments. This will be further discussed, as the “naturalism
principle” that ethnography must comply to.
The main source of information comes from observations and informal conversations
with the individuals under study.
There is no pre-formed plan and assumptions about the data collected. Data is
collected in raw. Only afterwards it will be mined and information can be processed out
of it. This is strongly related to another fundamental principle, “discover”, which will also
be discussed further on.
The group being studied is typically small and restrict to a particular setting.
37
The analysis of the data produced, implies the interpretation of the meanings and
intentions of the actions observed. In conformity with the third principle,
“Understanding”, that will be presented later on.
Apart from the previous described features, there are two different types of roles that can be
performed by the observer: participant or onlooker observer. This is a major decision that
influences the whole process of ethnography. Besides the two extremes, being a participant or
a mere spectator, all the other values of the spectrum are possible. Being a participant observer
implies sharing the life and activities of the people under study. The researcher gains a much
more profound insight of what‟s happening.
Nevertheless, to accomplish a good ethnography practice, the researcher must be able to keep
up with the task of observing and taking note of those observations. So, it is quite
understandable that becoming a participant deeply compromises the whole process and it will
be very difficult for an ethnographer to cope with both, the observations and his activities as a
participant. Some believe that it is neither necessary nor convenient to become a participant. In
fact, they say that the ethnographer must become as distant of the practices as possible, even if
they seem familiar. One must always rationalize and remain aware during the whole process.
5.2.2 Fundamental Principles
Ethnography thinking is rooted in three fundamental principles: naturalism, understanding and
discovery. (Genzuk, 2004)
Naturalism “is the view that the aim of social research is to capture the character of naturally
occurring human behavior”. This implies that no virtual or simulated environment is suitable to
achieve the purpose that ethnography aims to reach. Contrary to interviews and experiments,
only ethnography can show us the actual actions and behaviors of the group under study. To
accomplish, the researcher must be aware that his presence may affect the environment in
study. The ideal observation is the one where the subjects will not know that they are being
observed, though this may raise ethnic issues, such as the right to privacy. As not all
occurrences can be observed, an ethnographer must try to document enough situations to
extrapolate, later on, the fundamental patterns, key actions and procedures.
Another principle, which we will state as the ability to produce valid explanations for the
behavior of its members, is the principle of “Understanding”. In order to produce “valid
explanations”, one must gain knowledge of the cultural aspects that can influence the actions
observed. In this situation being participant at some degree clearly helps to develop cultural
insight, but as discussed before, participation may also blur the capacity of true objectivity. This
is also true when the observer is already naturally emerged in the culture of the observed.
38
Indeed, the behaviors may seem familiar and the danger of misunderstanding is particularly
great.
The “Discover” principle, that should be respected, states that the researcher should approach
the study with a clean mind instead of “being limited to the testing of explicit hypotheses”. In
fact, this erroneous attitude, will most certainly lead to the formulation of pre-assumptions that
could blind and misguide the researcher through the whole ethnography process.
5.2.3 Ethnography and Information Systems Design
Narrative descriptions, the classic output of the ethnographic method, lack the objectivity and
formalisms necessary for the engineering process (Genzuk, 2004). This technique has proven
to be useful before in the fields of Information System Design as we‟ll describe following.
Through years, as more research and experiments were made, the problems of integrating
ethnography with software methodologies became well defined. Viller, et al. (1999) identifies
and summarizes those issues:
Time: Ethnography, when applied to social research, usually takes a long time, months
or even years. The Requirements Engineering process needs to be much faster.
Results: The output of the ethnographical method, long textual descriptions, is difficult
to integrate in the existent methodology of system design.
Culture: As a result of sociologist‟s and requirements engineer‟s different backgrounds,
communication between the two groups can become a complex issue.
Abstraction: Abstraction does not suit ethnography very well. Ethnographic outputs are
too much directed to the details of a particular situation.
Skill: Because ethnography lacks a fixed step-based methodology, the results are
greatly dependent on the ethnographer‟s skill.
To address these issues, three different approaches on managing the integration of both
techniques were proposed by Button, et al. (1996): Concurrent ethnography, Informed by the
ethnographic account, and ethnographically informed method.
In concurrent ethnography the ethnographer goes to field, the work place of the group under
study, and undertakes the ethnographic process. Although an account may result, it is not the
account that will serve as input for the rest of the design process, but the ethnographer himself.
The idea is for the ethnographer to work as proxy between the group under study and the
requirements engineers.
39
The informed by the ethnographic account technique is characterized by the ethnographer
production of an account as a result of its study. This time, however, it is the account that will be
passed to the design process, instead of the interaction between the design engineers and the
ethnographer.
Finally, the ethnographically informed method, initially proposed by Button, et al. (1996), is
described under the category of the Coherence methodology (Viller, et al., 1999). In this case,
the design process is informed by the ethnographic methodology itself. The idea is to create a
new process which is influenced by ethnography and integrates ethnographic principles. This
way, instead of having two distinct methodologies working together, ethnography and the
design process, the purpose is to have only one process, which merges ethnography into the
design process.
5.2.3.1 Presentation Framework
Amongst others, the presentation framework structures ethnographic data in terms of three
dimensions of work: distributed coordination”, plans and procedures” and awareness of work”.
Distributed Coordination is concerned with how the tasks are performed within the broader
context of the organization (Viller, et al., 1999). It basically defines ‟who does what‟. It‟s
extremely useful for defining the roles played by each individual as they interact with each other.
Plans and procedures focuses on how the organization describes its processes and structure.
This information is usually described in documents and basically reflects how organizations see
themselves.
The third dimension, awareness of work refers to how individuals perform their tasks so that
what they are doing is made ‟visible‟ or ‟available‟ to others (Viller, et al., 1999).
As a result, it‟s possible to state that the presentation framework demonstrated a great potential
in structuring fieldwork notes and making them available and more accessible to the design
process.
The reason why this framework was chosen is justified by its special concern about work
practices, being through this aligned with the concept of OSA, and easily related with the agent
architecture and ontology previously described (section 3.2).
5.3 Conclusion
The success of interview techniques appliance in many different subjects contributed for their
broad usage of this famous method. Concerning a conversation between, at least, two people,
where one of them is asking questions, this procedure requires a previous preparation and the
40
follow of important best practices. However, being a technique that demands people presence
its success does also dependent on the interviewers‟ personality and ability to get what he
wants.
As main disadvantages, interviews usually take too long, requiring many resources and
representing an expensive method. In addition to it, they can lead into wrong interpretations or
even information omissions.
Ethnography is a social sciences technique. Though, lately software development fields have
also applied this method to requirements gathering. There are several ways proposed to
observe and register. They state how the ethnographer should behave and what should he
register, amongst other things.
This technique has the main advantage of preventing information omission, stated as one of the
interview disadvantages. However, the presence of the observer can influence the way people
behave, resulting in the gathering of wrong information. The time is a common disadvantage is
both methods.
41
PROBLEM DECLARATION
“There is nothing so easy to learn as experience and nothing so hard to apply”, Josh Billings
The fundamental theoretical bases used in this study are OE discipline notions, frameworks,
methods and approaches stated in the previous sections. In addition, a guideline has been
defined, OSA through enterprise modeling, concerning the capture of each person way of
working and real-time issues.
This section relates the theories previously described, states thesis‟ problem, and describes the
research approach followed to achieve our goals, and as a result solve the problem defined.
6 Problem Statement
The importance of people‟s awareness about organization and how that knowledge influences
their way of working, and consequently assist the optimization of organization processes, is one
of our major concerns.
Enterprise models have demonstrated their effectiveness in acquiring OSA (section 3.1 and
4.4). However, the framework and approach stated (section 3) only suggest a way to capture,
structure actions, and identify contexts. It doesn‟t mention how this information can be
represented using action-based models, which are important instruments to communicate. On
the other hand, BPM modeling proposals and Jorge Coelho‟s LEARN method (section 4) don‟t
consider people‟s work practices as relevant information, by considering mostly organization‟s
macro processes.
In this study we aim to find a way to use enterprise models that represent each organizational
person tasks, illustrating their personal and interpersonal work practices. The motivation to do
this it is justified by the importance that these models have in synchronizing the image that
people have of organization. Capturing and representing this information in real time is also one
of our ambitions, in order to speed up the achievement of OSA goals and provide people within
organizations instruments that allow them to wonder in real time about possible optimizations
and support its further implementations.
To summarize, this study concerns finding a way to answer the following OSA questions:
In a „as is‟ based approach:
o Which tasks perform agent X?
42
o How does agent X performs each task?
o What resources are used by agent X to perform these tasks?
Regarding the „to be‟, after the reengineering of work practices:
o Which tasks will be performed by agent X?
o How will agent X tasks be performed in the future?
o What resources will be used by agent X to perform these tasks in the future?
In addiction, and as already stated before, the inexistence of a method that allows real-time
capturing of the information needed to answer these questions, and its further representation in
a way that organizational people can understand them, is the problem addressed.
The results attained by solving this problem will allow the synchronization of representations
(models) that people have about organizations. Thus, this will become an efficient way to
perform management actions, support change management, and plan IS, amongst others.
7 Research Approach
In order to solve the problem stated above, an approach that encloses all OE described
concepts has been proposed. This is based on the relationship and gaps existing between
these.
All the previously described subjects are connected. In the framework proposed towards OSA
(section 3.2) enterprise models are referred as a key instrument. These models are also
mentioned by BPM (section 4), which proposes their design using EA and business process
modeling languages.
To put the OSA framework into practice, using Model Acquisition Approach (section 3.3), or to
apply BPM to organizations, with LEARN method (section 4.5.1), it becomes necessary to
gather information about the organization concerned. This can be done through specific
techniques, described in section 5, where it is possible to distinguish interviews and
ethnography.
Joining all the existing relationships between the three stated areas, OSA, BPM, and
information gathering techniques was the approach followed to reach the objectives of this
study. It can be summarized as an approach that merges the different concerns of these areas,
validating its benefits and efficiency in reaching OSA goals (Figure 21), and fulfill the existing
gaps between them in order to promote organizational people awareness about it.
43
Figure 21. Relation between OE concepts and the Approach proposed.
The approach proposed has been defined along time. Representing case studies a major part
of the developed work, each one will be described further. Here all the important decisions will
be explained, as well as the approach formulation, as a result of a reflection over the developed
work.
44
A STEP
TOWARDS ORGANIZATIONAL
SELF-AWARENESS:
CASE STUDIES
"If you hold a cat by the tail you learn things you cannot learn any other way." Mark Twain
Most of the developed work during this study was performed in real organizations. This allowed
us to put into practice many concepts, theories, and approaches stated before, validating their
appliance in real case studies. This practical effort has also contributed for the development of
new tools and to state a new approach that motivates OSA, allow the reverse engineering of
individual and interpersonal work practices, and provide a knowledge base to support „to be‟
analysis.
In this section two case studies are described. Despite the fact that both of these cases are
based on the research objective, spread the “virus” of OSA along organizations and motivate
with this the optimization of organization‟s processes, approaches followed differ from one case
study to the other. This difference is justified by each organization‟ goals defined for the work
developed and by the universe of organizational agents involved during the case studies. Each
case is further described, chronologically ordered.
8 Optimizing Moviflor Personal and Interpersonal Tasks
The first case study was developed in a furniture retail company, Moviflor. This is a Portuguese
and familiar organization, in the market along the last 30 years selling home products, from
furniture to decoration and illumination, carpets, textiles, amongst others. Nowadays Moviflor
employs more than 1000 workers in 20 stores and 5 warehouses distributed all along
Portuguese continental district and both islands, Madeira and Azores.
In 2001, Moviflor carried out a huge reform that consisted in intern reorganizations and on the
adoption of a new IS. This restructuring aimed to enforce and refresh organization‟s market
identity and redefine its missions and strategic goals. The main IS change was the
implementation of the ERP SAP R/3, whose objective was to provide a way to support widely
45
organization‟s business processes. After a first trembled period, SAP is correctly working in
most of Moviflor‟s organizational units. Their business is also running well, justified by the
constant grow of their sales during the last years.
This case study was centered in Moviflor sales department. This is composed by three main
groups (Figure 22): (1) category managers, responsible for product management and suppliers‟
agreements; (2) sales central, in charge of buying products, acquiring services from suppliers,
and manage the relation with suppliers after the product/service acquisition; and (3)
importations, responsible for buying and managing the orders provided by suppliers from other
continents whose merchandise is transported by containers. Firstly, the project started in sales
central, and once finished, a second study was performed in importations business unit.
Figure 22. Moviflor sales department organizational structure.
8.1 Defining Objectives
In order to start the work, it becomes necessary to identify main goals that would drive it.
Basically, in this phase there was a need to recognize project guidelines and understand
organization expectations. Similar as in LEARN method proposed by Jorge Coelho (section
4.5.1), the objectives were defined during a meeting where the sales department director point
out which were his priority issues.
The initially defined goals were based on sales central business unit. The objective was
optimize communication procedures with suppliers in order to (1) reduce the deficient
accomplishment of contracted points, such as late deliveries; (2) decrease the number of
incomplete or damaged products stored in warehouses; and (3) optimize employees work
practices through their standardization.
To accelerate the project, as suggested by LEARN method (section 4.5.1), an initial
contextualization in organization business and history was also obtained. This was done
Sales Manager
Category Managers
Sales Central
Importations
46
through informal visits to Moviflor stores and by reading organization documentation, e.g.
previous studies and company training information.
8.2 Delineating a Strategy
Once goals were clarified, it was time to define a strategy to reach them. This step is also part of
LEARN method (section 4.5.1), aiming to answer questions such as: (i) which people and when
will be involved in the project?; (ii) how long will it takes to provide results?; and (iii) how long
will the project takes?. As answer to these questions, initially was proposed a short observation
and interviews phase, followed by a group meeting where the subsequent steps would be
defined. This means that the answer to the last questions was rescheduled to the next reunion.
Being the sales central composed by 6 people, who work hardly diary, the participation of each
one in the project was gradual. The project started with only one person, a key user that was
already working in Moviflor for about five years. This person organizational awareness facilitates
a rapid contextualization in Moviflor‟s business and processes. After this first stage, other
people had gradually joined the project.
8.3 Instantiating Model Acquisition Approach
The following described performed steps are related to the ones proposed by the model
acquisition approach described in section 3.3. Thus, the subsequent phases can be expressed
as an instantiation of this approach.
Instead of using the term „agent‟, we will use the word „people‟. Being the following steps based
on each people actions, its comprehension becomes clear switching these terms. However,
when referring to other organizational levels, such as organizational units, the term agent is
applied.
8.3.1 Bootstrapping
Bootstrapping begins with an observation period that aims to define different action and
resource types for each person. This took about one week and consisted on interviews and
observations.
8.3.1.1 Interviews
Using the principles and methods previously referred about interviewing (section 5.1), some
questions were done to the chosen key user. These questions intended to realize, in a generic
way, which different tasks were performed diary in sales central and how they were
accomplished.
47
Generally, the used questions can be characterized as open ones, such as “What are your diary
main subjects?” or “Tell me about your work”. However, along the interview process there was a
need to funneling, changing to more specific questions, e.g. “How do you create an order?”.
This interview structure is usually classified as pyramid. As suggested in interviews best
practices, each took about 1 hour and was done in an informal way, like an ordinary
conversation, habitually while the interviewed person was working.
During the interview course, some methods were used to confirm that the interviewer correctly
understood what has been told him. As examples of these, two main practices were performed.
A typical way to validate something was to repeat. Another way was to design what people had
said using simple representations, e.g. connected boxes, to represent flows, hierarchical
graphics to represent relations, etc.
8.3.1.2 Observations
Observations of people‟s work were performed using some principles of information gathering
ethnography technique (section 5.2). Ethnographic principles realize aspects such as the
observer behavior, the register support and the results presentation.
Observer behavior was more transparent as possible, in order to avoid influencing people‟s
work. While observing the ethnographer was quiet, in silence, and only writing what he was
seeing.
Given this case ambition, the main concern was to register people‟s actions, and resources
used to perform them. This was done by writing this information in a notebook. It‟s important to
refer that while observing and registering, the ethnographer didn‟t filter which important tasks to
note. This process should only be done after, when analyzing the information gathered. The
way how actions were written and how the results were presented is detailed further.
Before observing, an initially clarification of the work proposes were described to each person.
This helped them to be more comfortable, and performing their working naturally.
8.3.1.3 How People React to Ethnography
Even not being a recent technique, it is not frequently applied in organizations. Thus, it
becomes interesting to state the way observed person react, and also how observer feels while
performing it.
Initially, after a brief explanation to observer about we were about to do, usually he reacted
saying that there was no problem to him, and that he will be comfortable while is being
observed. However, it was easy to notice that during the first instants the observed person got a
little bit constrained by the fact that someone is sited behind him noting everything that he was
48
doing. Normally, this constraint only happened during the first hour. After it, observer “forgot”
that was being examined. A good way to optimize the relation with the observer was to show
him the notes written and clarify the aim of the work proposed.
The constraint is also part of the observer feeling while performing this practice. However,
similar to observed person, after a short period of time observer becomes more comfortable.
Stated feelings and behaviors can depend from person to person. However the described
impressions were based on the observations of six different people, whose reactions didn‟t
differ radically.
A curious fact was that for Moviflor people the observing technique was one of the preferred
ones (Annex A). When informally questioning people about this unexpected result, they
answered that it was the only method that required less effort for them, “we only have to work
normally…while you register what you want without disturbing”.
8.3.2 Registering Personal Actions
The observed actions were registered in an organizational sentence format (section 3.2.1).
However, in this phase they weren‟t normalized yet. During observation people‟s actions were
described in a draft form, e.g. Pedro confirm order by email, Susana create order in SAP ME21
menu, João updates warehouse delivers in excel file.
After each observation day, actions registered were analyzed and normalized in order to
complete the following fields: time, actor, action, description, subject or keywords, and
resources. Apart for the subject, all the other fields are part of the organizational sentences
format.
Subject field, also described as key words, was used to group actions according to the context
definition, a network of interacting agents playing related specific activities and resources
(Zacarias, et al. b, 2007). Each action was contextualized, grouping it according to its keywords.
Table 3 exemplifies each field contents, except for the time.
49
Table 3. Draft organizational sentences registered during observation.
Sentence Actor Action Description Subject or keywords
Resources
Pedro confirm order by email
Pedro Confirm Confirmation of
previously created order with supplier
Order Confirmation Email
Susana create order in SAP ME21 menu
Susana Create New order creation
Order Creation SAP (ME21)
João updates warehouse
delivers in excel file
João Update
Update of warehouse list
where are saved all the delivers
Order Deliver Excel (delivers.xls)
After the observation period a set of actions, resources, and subjects normalized were defined.
As suggested in bootstrapping phase, the set was presented to the agents involved in order to
validate them. Moviflor sales director was also present during validation, using his wide
knowledge about the department work to support this validation. The validated set for each field
is further provided.
It was during the validation meeting that were defined the next work steps. The proposal was to
capture personal actions and represent them in order to, after this, recommend optimizations.
The followed approach is described next.
8.3.3 Capturing and Structuring Actions
This phase, as it names suggest, aims to capture and structure actions. The proposal is to
obtain actions through organizational sentences composed by triples subject-verb-object
(section 3.2.1).
Being the team constituted by six people, each one with different work practices, but only with
slightly different responsibilities, it becomes necessary to develop a tool in order to support, and
speed up, the project. This tool was a web application, named “Activity Register Portal”4, and
initially came out only to with the ambition of providing a semi-automatic way to get structured
and normalized organizational sentences.
Basically, the application developed was a web form where people register each action after
perform it (Figure 23). The form was composed by the following fields: subject, action,
technological resources, personal resources, other resources, delegation and other information,
as a free text field. Except for the free text, all the others fields were inserted by choosing one of
the presented options. However, there was also a free input associated to each field.
4 The chosen name for the Portal doesn‟t illustrate correctly its proposal in a technical way. However, the name is justified because it facilitates the explanation of applications objectives to ordinary people.
50
Figure 23. Action form available in the Activity Register Portal developed.
Subject was used to select the context of the action performed. When introducing Portal to
people, the instructions for fulfilling this field were explained using the email message analogy
i.e. imagine that you have to send an email message to your director explaining an action that
you have done; which subject will you choose for it?.
The options available in subject filed were: calculate products needs (provisions), products
codes, catalog product, suppliers‟ contacts, containers, new order creation, order devolution,
order state, damaged products, incomplete products, pendent orders, publicity, and client claim.
Each of these was presented and validated during bootstrapping phase before making part of
the set.
The notion of action is similar to the one considered by model acquisition approach
organizational sentences (section 3.3.2). They are the verbs that identify the action type.
Available options in the set action were: analyze, ask, calculate, change/update, confirm,