Toward a Typology of Business Process Management Professionals: Identifying Patterns of Competence through Latent Semantic Analysis.

Abstract

While researchers have analysed the organisational competences that are required for successful Business Process Management (BPM) initiatives, individual BPM competences have not yet been studied in detail. In this study, latent semantic analysis is used to examine a collection of 1,507 BPM-related job advertisements in order to develop a typology of BPM professionals. This empirical analysis reveals distinct ideal types and profiles of BPM professionals on several levels of abstraction. A closer look at these ideal types and profiles confirms that BPM is a boundary-spanning field that requires interdisciplinary sets of competence that range from technical competences to business and systems competences. Based on the study’s findings, it is posited that individual and organisational alignment with the identified ideal types and profiles is likely to result in high employability and organisational BPM success.

Full text

1
Toward a Typology of Business Process Management Professionals:
Identifying Patterns of Competences through Latent Semantic
Analysis
While researchers have analysed the organisational competences that are required
for successful Business Process Management (BPM) initiatives, individual BPM
competences have not yet been studied in detail. In this study, latent semantic
analysis is used to examine a collection of 1,507 BPM-related job advertisements
in order to develop a typology of BPM professionals. This empirical analysis
reveals distinct ideal types and profiles of BPM professionals on several levels of
abstraction. A closer look at these ideal types and profiles confirms that BPM is a
boundary-spanning field that requires interdisciplinary sets of competence that
range from technical competences to business and systems competences. Based
on the study’s findings, it is posited that individual and organisational alignment
with the identified ideal types and profiles is likely to result in high employability
and organisational BPM success.
Keywords: business process management, professionals, competences,
knowledge, skills, abilities, latent semantic analysis, typology
1. Introduction
Business Process Management (BPM), an active field of research in the Information
Systems (IS) discipline, is an interdisciplinary approach to the analysis, design,
implementation, and improvement of organisational work processes and supporting
Information Technology (IT) systems (Davenport and Short 1990; Hammer and
Champy 1993; vom Brocke and Rosemann 2010). The goal of BPM is to increase
operational efficiency and effectiveness (e.g., product/service quality, compliance) by
organising a company around core business processes instead of functional departments
(Hammer and Champy 1993). Typically, BPM initiatives are associated with major
investments in enterprise information systems in order to integrate previously separated
business activities into cross-functional, end-to-end business processes (Wetzstein and

2
Leitner 2011; Altinkemer, Ozcelik, and Ozdemir 2011).
The practical relevance of BPM and related process-oriented management
approaches is undisputed (Gartner 2011; McKinsey 2011). Nevertheless, as Bandara
and colleagues (2010) state, most companies still “lack sufficient internal competencies
needed to undertake these BPM initiatives” (p. 744).
Extant research provides insights into organisational competences that are
required in BPM initiatives, such as through BPM maturity models (e.g., Van Looy, De
Backer, and Poels 2012; Rosemann and de Bruin 2005a; de Bruin and Rosemann 2007;
Hammer 2007). However, such frameworks provide only limited insights into what
process management competences are required at an individual level. Besides some rare
exceptions, such as Antonucci and Goeke (2011) and Launonen and Kess (2002), the
topic of individual BPM competences has been largely neglected in academic research.
The contemporary understanding of BPM as a holistic approach suggests that BPM
initiatives require comprehensive competence portfolios, so BPM professionals, as well
as executives and educators, face the challenge of identifying and developing
meaningful combinations of competences.
Against this background, the purpose of the present study is to clarify BPM
competence requirements at an individual level. Following related research in the IS
discipline, this goal is addressed by conducting a content analysis of job advertisements
(ads). Specifically, this study ties in with several well-recognised studies that have used
job ads to examine individual competence requirements (e.g., Gallivan, Truex, and
Kvasny 2004; Litecky and Aken 2010; Todd, McKeen, and Gallupe 1995). Job ads are
one of the most important recruitment instruments, so they can serve as a proxy for the
competences required of professionals (Todd, McKeen, and Gallupe 1995). Therefore,

3
job ads represent an appropriate data source for an examination of the required
competence sets of BPM professionals.
Latent semantic analysis (LSA), a text mining technique, is used to analyse
1,507 BPM-related job ads collected from a global online job platform and derive a
typology of BPM professionals (Doty and Glick 1994). The typology developed is
comprised of a number of ideal types and profiles of BPM professionals. It is
hypothesised that, at an individual level, matching the profile of one of the ideal types
identified will likely result in the ability to gain and retain employment (Hillage and
Pollard 1999). At an organisational level, it is posited that a complete competence
portfolio comprised of all of the individual ideal types identified is required if process-
oriented organisations are to be effective.
The remainder of this paper is structured as follows. The next section provides a
theoretical background on typologies, ideal types, and human resources in IS research.
Then the data collection and analysis approach is presented by introducing the research
method, i.e. LSA, and providing specific information on the individual steps of the
study’s research process. This section is followed by a detailed description of the
study’s findings, a discussion of the attained results against the background of existing
studies and their implications for research and practice. Finally, study limitations are
pointed out and the paper is concluded with an overview of the contributions that stem
from this research.
2. Research Background
2.1. Typologies and Ideal Types
According to Doty and Glick (1994, p. 230), “[t]ypologies are a very popular, but often
misunderstood form of theory.” Typologies have often been confused with classification

4
systems (e.g., taxonomies), which partition phenomena into mutually exclusive and
exhaustive sets (Doty and Glick 1994), while typologies are sets of “ideal” types, that is,
types that are considered to be especially effective or successful. Therefore, typologies
go beyond pure description by stating relationships between independent variables
(ideal types) and dependent variables (organisational or individual success). Many
typologies posit, for instance, that the closer an entity comes to the profile of an ideal
type, the greater its effectiveness is, although a complete match usually does not happen
in reality (Weber 1949).
One prominent example of a typology is Mintzberg’s (1979) theory on the
structuring of organisations, which describes five ideal types of effective organisational
structures: simple structure, machine bureaucracy, professional bureaucracy,
divisionalised form, and adhocracy. Each of these ideal types is described with an ideal
profile that uses a common set of descriptors, such as age, size, horizontal and vertical
specialisation of jobs, and formalisation of behaviour. These descriptors are aggregated
into the two broader categories of context and structure. For example, the profile of a
simple structure is described by young age, small size, and low horizontal and high
vertical specialisation.
Typologies can be defined theoretically or empirically (Doty and Glick 1994).
Researchers who model typologies theoretically deduce ideal types and ideal profiles
from existing theories and literature (e.g., Guillemette and Paré 2012). Ideal types and
profiles can also be derived inductively from an empirical sample (e.g., Govindarajan
1988). In this case, a researcher identifies entities from the sample that seem to be
especially effective or successful, so they resemble ideal types. Then the researcher
describes the identified ideal types by specifying ideal profiles for each ideal type.

5
This study follows the empirical approach to typology development by
analysing a large sample of BPM-related job advertisements. The idea underlying this
study is that job profiles that are in high demand on the job market represent effective
ideal types, or they would not be in high demand. As it is explained in detail below, text
mining is used to identify groups of similar jobs in high demand (i.e., ideal types) and
extract patterns of frequently used descriptive terms that describe the competence
requirements of these jobs (i.e., ideal profiles).
Independent of the process of typology development, Doty and Glick (1994)
stress the importance of using consistent and theoretically relevant descriptors to specify
ideal types. Following this recommendation, the IS literature on human resources was
reviewed in order to identify frameworks and constructs suitable for describing ideal
types of BPM professionals. The next section reports on this review.
2.2. Human Resources in Information Systems Research
The study of human resources (HR) in the IS discipline dates back to the late 1960s,
when the importance of IS personnel in organisations was first recognised (e.g., Brady
1967). Since then, numerous studies have addressed questions concerning the demand
for IS professionals and their related competences (e.g., Lee, Trauth, and Farwell 1995;
Nelson 1991; Todd, McKeen, and Gallupe 1995). A competence is defined as a work-
related knowledge, skill, or ability held by an individual (Nordhaug 1993). In broad
terms, knowledge refers to the theoretical understanding of a concept (e.g., what the
notion of a business process means), while skills relate to the practical application of
that knowledge (e.g., how to model a business process). In contrast to knowledge and
skills, which can be learned, abilities are attributes that are innate to an individual (e.g.,
a person’s ability to abstract).

6
Although there is a consensus on the conceptualisation of competence, most
empirical IS studies use the notions of work-related knowledge, skills, and abilities
interchangeably (e.g., Lee, Trauth, and Farwell 1995; Nelson 1991; Todd, McKeen, and
Gallupe 1995). The same holds true for the vast majority of job ads, which were
analysed in this empirical study. Therefore, it was decided to abstain from making strict
distinctions between the notions of knowledge, skills, and abilities and use the broader
term competence instead.
While the HR literature identifies a great number of classification systems for
organising work-related competences (Le Deist and Winterton 2005), few classification
systems are specific to BPM. Therefore, the IS literature was examined to identify
frameworks that may provide descriptors for a typology of BPM professionals. After
analysing several IS competence classifications (e.g., Tang, Lee, and Koh 2000; Litecky
et al. 2009), the scheme proposed by Todd et al. (1995) was selected for further use.
Todd et al. (1995) manually analysed job ads to determine how the competence
requirements for programmers, analysts, and IS managers changed over time. Based on
a classification scheme provided by the Association for Computing Machinery (ACM),
the authors coded the job ads by assigning the competences in them to three categories:
technical competences, business competences, and systems competences. Then they
split each of the categories into a set of two to three sub-categories, enriched with short
descriptions (Table 1).
Table 1 Classification of IS competences by Todd et al. (1995)
Category
Sub-category
Description
Technical
Hardware
Servers and personal computers. Other devices such as storage
devices, controllers, printers, and other peripherals plus networks.

7
Software
Application systems, operating systems, packaged products (e.g.,
databases), networking software and programming languages.
Business
Domain
Functional expertise (e.g., finance, marketing) and industry
expertise (e.g., retail, mining).
Management
General management skills including leadership, project
management, planning, controlling, training, and organization.
Social
Interpersonal skills, communication skills, personal motivation and
ability to work independently.
Systems
Problem Solving
Creative solutions, quantitative skills, analytical modeling, logical
capabilities, deductive/inductive reasoning, innovation.
Development
Knowledge of systems development methodologies, systems
approach, implementation issues, operations and maintenance
issues, general development phases, documentation, and
analysis/design tools and techniques.
Todd et al.’s (1995) framework is a comprehensive and versatile approach to
classifying job competences in the IS discipline, as it recognises both the duality and the
alignment of business and technology. Moreover, this framework has been widely
accepted and applied in numerous studies (e.g., Lee, Trauth, and Farwell 1995; Hardin,
Joshi, and Li 2002). Therefore, the categories and sub-categories (which Todd et al.
originally termed “classes” and “categories”, respectively) of this scheme were adopted
in this study as common descriptors for characterising the ideal profiles of the typology
of BPM professionals. Particular competences classified into categories and sub-
categories form an ideal profile, which describes a certain ideal type. All of the ideal
types of a particular phenomenon together form the typology — in this case, a typology
of BPM professionals.

8
3. Data Collection and Analysis
3.1. Introduction to and Example of Latent Semantic Analysis
A quantitative analysis of advertisements in the BPM job market was conducted in
order to derive ideal types and profiles of BPM jobs empirically. Specifically, the
quantitative text-mining technique LSA was used to identify patterns in job
advertisements. LSA detects patterns of word use in texts through statistical analysis
(Landauer, Foltz, and Laham 1998). The underlying idea is that the contexts (e.g.,
documents, paragraphs, sentences) in which a word appears or does not appear contain
valuable information about the meaning of both the word and the text. LSA was
originally developed in the late 1980s as an information retrieval method (Deerwester et
al. 1990), but it has since shown its usefulness in a great variety of applications and
domains.
The rationale for selecting LSA was based on its growing importance as a
methodology in the IS discipline, particularly in automated quantitative content analysis
(Evangelopoulos, Zhang, and Prybutok 2012). Content analysis is a method of
systematic analysis of texts in order to identify concepts and patterns therein (Jauch,
Osborn, and Martin 1980), traditionally done through manual coding. When LSA is
used for content analysis, the identified patterns of word use are interpreted as concepts.
For example, Sidorova and colleagues (2008) used LSA to examine 1,615 abstracts of
papers published in three leading IS journals in order to identify five core research areas
of the discipline, and Larsen et al. (2008) used LSA to cluster 14,510 abstracts from 65
IS journals into seven intellectual communities. Sidorova and Isik (2010) analysed
2,700 abstracts of papers using LSA and identified 20 distinct but interrelated sub-topics
in the field of BPM. Finally, Indulska and colleagues (2012) performed LSA on 8,544

9
abstracts from IS, management, and accounting journals to reveal core topics in these
disciplines and to show how they have evolved over the last 25 years.
LSA was also selected because it has been used successfully in the field of HR
management (Giesbers, Rusman, and Bruggen 2006; Haley et al. 2005). For example,
Laham, Bennet, and Landauer (2000) applied LSA to develop a software tool for the US
Air Force that can match job requirements with employees by using “the explicit and
implicit knowledge that already exists in extensive textual computer files of systems
documentations, training and test materials, task analyses, and service records” (p. 173).
Experiments with data comprised of more than 2,000 task descriptions and 9,000
descriptions of airmen and airwomen indicated that LSA has the potential to extract job
knowledge requirements from different types of textual documents (e.g., descriptions of
occupations, job tasks, personnel, training materials) accurately and in detail and to
determine the similarity between jobs and jobs, jobs and employees, and employees and
employees. The prototype was later extended to include additional data sources (job
listings from the US Department of Labor Occupational Network and résumés from
Yahoo and other online employment platforms) in order to serve further purposes, such
as identifying training needs (Laham, Bennett, and Derr 2002).
The proofs-of-concept for applying LSA in IS and HR management described
above indicate that LSA is a viable method for analysing job ads in the IS field. The
next part of the discussion describes the method itself.
A typical LSA is comprised of three phases. (See Tables 2 – 5 for an illustrative
example.) The goal of the first phase is to transform a collection of documents (Table 2)
into a term-document matrix (Table 3) that consists of rows that represent words and
columns that represent documents. The cells of the matrix contain the number of times a
particular word appears in a particular document. In an efficient and effective analysis,

10
documents and words typically undergo a series of pre-processing steps, such as
reducing terms to their stems (e.g., flexibility and flexible to “flexib”), filtering out
uninformative stopwords (e.g., “the”, “about”), and weighting terms according to their
relative importance. (For reasons of simplicity, term stemming and weighting are
omitted in the example in Tables 2 and 3.)
Table 2 Hypothetical job advertisements
ID
Job title
Job description (extract)
JobAd1
BPM Analyst
We look for a business analyst in the area of business process
management...
JobAd2
Workflow Developer
...you will be a developer for technical workflows...
JobAd3
Business Process
Improvement Manager
...improve business process performance...
JobAd4
SAP Process Architect
...technical process architect with SAP experience...
JobAd5
SAP Netweaver BPM
Developer
We need a developer with SAP Netweaver BPM skills...
JobAd6
Process Performance
Manager
...you will be responsible for the management of business
process performance...
Table 3 Term-document matrix for the job descriptions in Table 2 (without stopwords)
Term
Document
JobAd1
JobAd2
JobAd3
JobAd4
JobAd5
JobAd6
analyst
1
0
0
0
0
0
architect
0
0
0
1
0
0
bpm
0
0
0
0
1
0
business
2
0
1
0
0
1
developer
0
1
0
0
1
0
improve
0
0
1
0
0
0
management
1
0
0
0
0
1
netweaver
0
0
0
0
1
0
performance
0
0
1
0
0
1
process
1
0
1
1
0
1
sap
0
0
0
1
1
0
technical
0
1
0
1
0
0
workflows
0
1
0
0
0
0

11
In the second phase, the term-document matrix is analysed using a statistical
technique called Singular Value Decomposition (SVD) to reduce the dimensionality of
the term-document matrix without losing relevant information by identifying groups of
highly correlated words (i.e., words that co-occur together in documents) and highly
correlated documents (i.e., documents that contain similar words). The result of the
SVD is a set of semantic factors with associated term loadings (Table 4) and document
loadings (Table 5), which together describe specific patterns of word usage (Sidorova et
al. 2008).
Table 4 Term loadings
Factor 1
Factor 2
analyst
0.688
-0.141
architect
0.207
0.564
bpm
0.028
0.657
business
2.357
-0.381
developer
0.053
1.131
improve
0.460
-0.034
management
1.209
-0.206
netweaver
0.028
0.657
performance
0.981
-0.100
process
1.875
0.324
sap
0.234
1.221
technical
0.232
1.038
workflows
0.025
0.474
Table 5 Document loadings
Factor 1
Factor 2
JobAd1
2.416
-0.333
JobAd2
0.088
1.119
JobAd3
1.615
-0.081
JobAd4
0.726
1.332
JobAd5
0.098
1.553
JobAd6
1.828
-0.154

12
The extracted word-usage patterns are interpreted in the third phase. Depending
on which variant of LSA-based analysis is used (see Evangelopoulos, Zhang, and
Prybutok 2012 for a detailed discussion), the interpretation involves additional
statistical analyses (factor analysis, clustering, or classification) and expert judgments.
In the small illustrative example, Tables 4 and 5 (i.e., the results of the SVD) can
be interpreted right away without any further statistical analysis. Table 4 presents two
extracted factors and the associated loadings of terms taken from the job description
extracts. Assuming an exemplary factor loading threshold of 0.5, the results indicate
that Factor 1 is primarily related to the terms “business”, “process”, “management”,
“performance”, and “analyst” and that Factor 2 is mostly related to the terms “SAP”,
“developer”, “technical”, “Netweaver”, “BPM”, and “architect”. Table 5 shows the
same two factors and the corresponding document loadings. Factor 1 is primarily related
to Job Ads 1 (BPM Analyst), 6 (Process Performance Manager), and 3 (Business
Process Improvement Manager), whereas Factor 2 is primarily associated with Job Ads
5 (SAP Netweaver BPM Developer), 4 (SAP Process Architect), and 2 (Workflow
Developer). Examining the high-loading terms and the titles of the high-loading ads
together suggests that Factor 1 describes business-oriented jobs (a type of job that could
be labelled Business Process Performance Manager, for example), and Factor 2 refers to
technical jobs (which could be labelled SAP BPM Developer, for example).
The next sections outline how LSA was applied to analyse online job
advertisements on a large scale. Figure 1 and Table 9 illustrate the phases, activities,
and inputs/outputs of this study’s data collection and analysis approach . While the
basic procedure explained in the illustrative example above was followed, Figure 1 and
Table 9 contain more detailed steps required for the large-scale analysis, which are

13
explained in the next sections. In addition, it is indicated in Figure 1 and Table 9 which
steps can be executed automatically and which require human judgment.

14
Data Collection and Analysis Process
Collection an d Pre-Pr oce ssing of Job AdsSingular Value DecompositionInterpretation
Collect job ads
Corpu s of job ads
Rem ov e
unwant ed job ads
(e.g., dupl icates ,
spam)
Remov e unwanted
text blocks from
job a ds (e.g.,
company
descri ption s)
Stem terms
Rem ov e
stopw ords and
uniq ue terms
Bu ild
term-do cument
mat r ix
Sea rch ter ms
(e.g., «business process»)
Dictionary with
descriptive terms
We ight
term-do cument
mat r ix
Set number of
fac t ors
Deco mp os e
term-do cument
mat r ix
Term-do cument matrix
Calculate term
and do cument
loadings
Perfor m varimax
rotation
Term Eigenvectors,
Docum ent Eigenve ctors,
Singular Values
Term loadings,
Docum ent loading s
Determine
loading
thresho lds
Inter pret and
label factors
Factors with
associated high-loading
terms and docume nts
Create ideal types
and profile s of
jobs
Ideal types,
Ideal profiles
Calculate
descri ptive
statistics
Descriptive
statistics
(e.g., number of jobs
per i deal t ype)
Rem ov e
uninformative
terms
Identify
high-loading
ter ms and
docum ents p er
fac t or
1
1716
15
141312
234 5 67 8
910 11
Numbe r
of fa cto rs
Loading
thresho lds
Sta rt
End
Figure 1 Research Process

15
3.2. Collecting and Pre-Processing Job Ads
Online job ads were analysed in order to gain a comprehensive picture of the current
demand for BPM professionals since, based on the increasing use of online job markets
and the increasing global competition for experts (Jansen, Jansen, and Spink 2005;
Douglass and Edelstein 2009), online employment platforms are a major distribution
channel for organisations’ vacancies. Against this background, job ads were collected
from a premier global online employment platform (Step 1 in Figure 1 and Table 9) and
searched for ads in English-speaking countries, downloading those that contained the
term “business process” from the platform websites in the US, Canada, the UK, and
Australia (Table 6).
Table 6 Number of hits per country for the search term “business process”
US
Canada
UK
Australia
1.000
638
480
161
Unwanted job ads (Step 2) and unwanted text blocks (Step 3) were then
removed as follows. After 45 irrelevant job search hits were removed (i.e., non-English
ads or spam), 2,279 ads remained. Manual inspection of the downloaded ads and an
LSA pre-test revealed a number of anomalies. First, a significant number of ads were
close to being duplicates. As duplicates bias the LSA without providing additional
information, a duplicate-finder tool was used to identify and remove groups of ads with
a similarity score greater than 95 per cent. Second, a considerable proportion of the ads’
content was comprised of standard company descriptions (Figure 2). Pre-tests showed
that these text blocks bias the LSA by, for example, producing semantic factors that
describe large companies, such as Deloitte or IBM, so these standard text blocks were

16
manually removed. Third, in some job ads the search term “business process” was
mentioned in the company description but not in the job description (e.g., ads in which
business process outsourcing companies were searching for a team assistant). Therefore,
after removing the standard company-description text blocks, it was checked whether
the ads still contained the term “business process” and those that did not were removed.
After pre-processing, the final dataset contained 1,507 unique BPM-related job ads.
Following well-recognised text-mining procedures (Manning, Raghavan, and
Schutze 2008), the document collection vocabulary was reduced before starting the
LSA. The complete vocabulary defined by the 1,507 job ads initially contained 11,376
terms. To make the number of terms manageable for further analysis, automatic term
stemming was applied (i.e., shortening inflected terms to their common stems, such as
shortening coordinator, coordinating, and coordinate to “coordin”). Stemming reduced
the vocabulary to 6,785 terms (Step 4). In addition, standard stop words (e.g., “the”,
“and”, “about”) and terms that occurred in only one document were removed
algorithmically (Step 5). The remaining 3,732 terms were analysed manually to identify
uninformative terms that frequently occur in online job ads (e.g., “apply”, “salary”)
(Step 6) in an effort to keep only those terms that were specific to the competence
requirements of the ads. In this process, two researchers independently examined the list
of terms and identified terms to be excluded. The two coders disagreed in about 17 per
cent of the cases. These conflicts were then reviewed and resolved by a third researcher.
Persisting conflicts were resolved in a consensus discussion that included all
researchers. The resulting controlled vocabulary consisted of 1,422 terms and was used
as a dictionary for the further analysis.

17
Figure 2 Exemplary Job Advertisement
After defining the final set of documents and terms, a 1,422-by-1,507 term-
document matrix was built (Step 7) that contained the number of times each term

18
appeared in each document. This raw matrix was then transformed by weighting terms
based on their occurrence in and across documents (Manning, Raghavan, and Schutze
2008) (Step 8). The commonly used TF-IDF (Term Frequency - Inverse Document
Frequency) weighting scheme was applied, which promotes the occurrence of rare
terms in a document and discounts the occurrence of more common terms (e.g.,
“business”, “manage”). The weighted term-document matrix built the foundation for the
subsequent SVD.
3.3. Singular Value Decomposition
After compiling the weighted term-document matrix, the SVD was performed using the
statistical computing software R. The determination of the number of factors to be
extracted is not performed by SVD itself but has to be done by the researcher. The
appropriate number of factors is heavily discussed in the literature (Evangelopoulos,
Zhang, and Prybutok 2012). Several solutions were explored in this study, including 2,
3, 5, 7, 10, and 20 factors (Step 9), decomposing for each solution the term-document
matrix into a matrix of term eigenvectors, a matrix of document eigenvectors, and a
diagonal matrix of singular values (Step 10). Multiplying the matrices of the term
eigenvectors and the document eigenvectors by the singular values matrix produces the
term and document loadings for each factor (Step 11), which signify the degree to
which a specific term or document is relevant to a given semantic factor.
3.4. Interpretation
Following Evangelopoulos et al.’s (2012) recommendation, the next step was a
procedure that is comparable to the traditional factor analysis of numerical data. High-
loading terms and documents were extracted for each factor in order to uncover the
latent semantic structure of the BPM jobs collection. Since cross-loadings of documents

19
are possible (i.e., a document can be assigned to more than one factor), a job ad may be
assigned to more than one topic (e.g., business process design and enterprise resource
planning (ERP) systems). While such cross-loadings seem to be reasonable because of
the interdisciplinary character of the BPM field, it was aimed to identify the most
distinct ideal types of BPM jobs and, therefore, it was proceeded as follows.
First, a standard statistical procedure, varimax rotation, was performed to
improve the distinctiveness of the semantic factors and to simplify their interpretation
(Sidorova et al. 2008) (Step 12). This procedure maximises the variance of the term
loadings on a factor by rotating the coordinates of the semantic space such that each
factor has a few large loadings and many small loadings, making the term-factor
association clearer and easier to interpret because a term is either descriptive (high
loading) or not descriptive (low loading) for a particular factor. To maintain the
representation of the documents in the same factor space, an identical rotation was then
performed with the document loadings matrix.
To decide whether a particular term or document should be assigned to a factor,
a specific loading threshold must be defined. Following the approach of Sidorova et al.
(2008), the threshold of a k-factors solution was determined (Step 13) by identifying the
top-1/k documents and terms. For example, in the 20-factor solution, the top 5 per cent
of high-loading terms and documents were extracted (Step 14), so that each term and
document loaded on an average on one factor. However, cross-loadings are still
possible. (See Table 10 in the appendix.)
Next, the factors were interpreted by co-examining the associated terms and
documents (Step 15). This job was done manually by three researchers, with each
researcher independently interpreting and labelling each factor by analysing the list of
extracted high-loading descriptive terms and job ads and then comparing the results. In

20
almost all cases, factor interpretation was straightforward and unambiguous; minor
disagreements in labelling were resolved during a discussion involving all researchers.
The last steps were comprised of further categorising the high-loading terms of
each factor in order to derive ideal profiles for each ideal job type (Step 16) and
calculating additional descriptive statistics (Step 17). These steps are described in detail
in the next section.
4. Results
4.1. Ideal Types
The identification of ideal types can be conducted on several levels of abstraction,
which is reflected statistically in the selection of an appropriate number of factors for
the SVD. The solutions with 2, 3, 5, 7, 10, and 20 factors were explored and it was
found that as more factors are considered, the resultant ideal types become more fine-
grained.
On the most abstract level, the 2-factor solution, job ads were distributed in
almost equal parts into more business-related jobs and more systems-related jobs. The
five highest-loading descriptive terms for the more business-related jobs, which was
labelled Business Analysts, were “manag”, “busi”, “project”, “improv”, and “process”.
Typical titles of high-loading job ads were Business Analyst, Business Process
Manager, and Project Manager. In contrast, the top five descriptive terms for the more
systems-related jobs were “busi”, “system”, “test”, “develop”, and “design”. Frequent
job titles included IT Solutions Architect, Development Manager, and Data Architect.
This broad group of jobs was labelled Systems Analysts.
Figure 3 provides an overview of the 2-, 3-, 5-, 7-, 10-, and 20-factor solutions
and illustrates how, from level to level, ideal types are added, refined, split up, and

21
dissolved. For example, the Systems Analyst ideal type is split up and dissolved into
various other positions on more detailed levels of analysis (e.g., Technical Architect,
Business Intelligence (BI) Analyst, and Enterprise Content Management (ECM)
Developer). Similarly, the more detailed the analysis, the higher the number of
differentiated Business Process (BP) Manager ideal types that emerged (BP Manager
Sales & Marketing, BP Manager Accounting & Finance, BP Manager HR, BP Manager
Supply Chain). Figure 3 also contains the relative number of job ads assigned to each
ideal type, which is an indicator of the demand for the identified ideal types in practice.
(The percentages do not add up to 100 per cent, as some job ads load on more than one
ideal type or on no ideal type at all.)

22
Busin ess
Analy st
Sys tem s
Analy st
Busin ess
Analy st
Systems A nalyst
ERP Solution
Architect
Busin ess
Process A nalyst
Busin ess Proc ess
Improvement
Manage r
ERP Solution
Architect
Technical
Architect
Busin ess Proc ess
Compliance
Manage r
Busin ess
Process A nalyst
IT-Busi ness
St ra te gy
Manage r
ERP Solution
Architect
Technical
Architect
Busin ess Proc ess
Compliance
Manage r
IT-Busi ness
Strategy Ma nager
Busin ess Proc ess
Man ager Sale s &
Ma rke t in g
Business Process An alyst
Business Pro cess
Manager Sal es &
Mark etin g
ERP Solution Architect
(SAP)
Business Pro cess
Arch it ec t
Business Pro cess
Manager Accounting &
Finance
IT-Busines s Strat egy
Manager
Business Pro cess
Impr ovemen t Manag er
ECM D eveloper
Bus iness Int ellig enc e
Analy st
Business Process An alyst
(Consultant)
ERP Solution Architect
(Ora cle )
Req ui reme nts Ana ly st
Enterprise Architect
Business Pro cess
Manager Hu man
Resour ces
Bus in ess Pro ces s Cha nge
Manager
Business Pro cess
Manager Supply Chain
Technical Architect
Business Process Tra iner
User Interface
Develo per
Business Pro cess
Com pli anc e Ma nage r
50%
50%
44%
38%
17%
11%
11%
26%
21%
31%
24%
14%
17%
10%
13%
10%
11%
7%
10%
4%
13%
5%
8%
9%
15%
6%
24%
29%
1%
3%
2%
4%
6%
4%
2%
3%
5%
5%
4%
10%
4%
4%
6%
2%
3%
2%
3%
3-factor
solution
5-factor
solution
7-factor
solution
10-factor
solution
20-factor
solution
2-factor
solution
Figure 3 Overview of Ideal Types of BPM Professionals
This section reports in detail on the 7-factor solution. It was chosen to focus on
this solution, as it provides a comprehensive picture that is neither too abstract nor too
detailed. The 7-factor solution revealed seven clearly distinguishable ideal types of
BPM professionals whose profiles are analysed in detail in this section. Table 7

23
provides the number and percentage of ads assigned to a particular ideal type, which
acts as an indicator of the level at which these jobs occur in practice.
Table 7 Ideal types of BPM professionals (7-factors solution)
Ideal Type
Description
Number
of Ads
Percentage
of Ads
Business Process Analyst
… elicits, analyses, documents, and
communicates user requirements and designs
according business processes and IT systems;
acts as a liaison between business and IT
358
23.8
Business Process
Compliance Manager
… analyses regulatory requirements and
ensures compliance of business processes and
IT systems
256
17.0
Business Process Manager
Sales & Marketing
… designs sales processes and analyses
requirements for related IT systems; supports
and executes sales and marketing processes
212
14.1
Business Process
Improvement Manager
… analyses, measures, and continuously
improves business process, e.g., through
application of Lean or Six Sigma management
techniques
147
9.8
ERP Solution Architect
… implements business processes in ERP
systems
152
10.1
IT-Business Strategy
Manager
… aligns business and IT strategies; monitors
technological innovations and identifies
business opportunities
191
12.7
Technical Architect
… develops and integrates hardware and
software infrastructures
160
10.6
4.2. Ideal Profiles
The ideal profiles specify the competence requirements for each of the seven ideal types
through high-loading descriptive terms. The ideal profiles were derived using multiple
coders: three coders independently assigned the high-loading descriptive terms of each
factor to the categories and sub-categories provided by Todd et al.’s (1995) framework.
In case of coding conflicts, all researchers reviewed the disagreements and discussed

24
them until consensus was reached and all conflicts were resolved. Even though the
single terms provide only a rough indication of the exact competences required in each
profile, assigning the terms to Todd et al.’s (1995) categories and sub-categories gives
additional context that allows gaining a more precise idea of the competences actually
required by each profile.
The analysis also shows how LSA allows us to discriminate clearly among the
competences of ideal types, even though single competence-related terms load on more
than one ideal-type factor. For example, the term “analysi” loads on the ideal types
Business Process Analyst and IT-Business Strategy Manager. The term’s concrete
meaning can be determined by interpreting it in the context of the other high-loading
terms. For example, in the ideal type Business Process Analyst, the term “analysi” co-
occurs with terms like “user”, “specif”, “gather”, “elicit”, and “problem”, indicating the
need for competence in user requirements analysis, while in the ideal type IT-Business
Strategy Manager, “analysi” occurs with terms like “strateg”, “plan”, and “risk”,
indicating the need for more strategic analytical competences.
In the following the profiles of all seven ideal-type BPM professionals are
described in more detail. (The tabular ideal profiles are presented in the appendix; one
exemplary profile is presented in Table 8)
Business Process Analyst: The balanced distribution of terms across the
categories of Todd et al.’s (1995) framework indicates that a Business Process Analyst
requires a broad portfolio of competences. For example, management competences
relate to project planning and management, as the terms “project”, “lead”, “ensur”, and
“success” suggest; social competences refer to communication and collaboration (e.g.,
“stakehold”, “communic”, “collabor”, “interview”); problem-solving competences
relate to the analysis of specific user requirements (e.g., “elicit”, “user”, “specif”,

25
“issu”); and development competences mainly refer to design and documentation of
processes and systems, as suggested by terms like “document”, “map”, “model”,
“architecture”, and “design”. In addition to high-loading descriptive terms, the ideal
profile includes exemplary job titles of high-loading job ads (e.g., Business Analyst,
Business Systems Analyst, Business Process Analyst), which help us to interpret further
the nature of this ideal type of BPM professional.
Business Process Compliance Manager: The high-loading terms were mostly
assigned to domain and management competences. The terms in the domain sub-
category refer to controlling and auditing (e.g., “monitor”, “complianc”, “standard”,
“intern”, “control”, “audit”, “polici”, “qualiti”), while term stems like “coordin”, “plan”,
“perform”, “administr”, and “report” indicate that management competences are also
important in this job profile. The strong focus on compliance is also apparent in high-
loading job titles like Trade Compliance Manager, Business Process Controls Specialist,
and Audit Compliance Specialist.
Business Process Manager Sales & Marketing: Again, most of the descriptive
terms were assigned to the domain and management sub-categories, yet compared to the
Business Process Compliance Manager they refer to other competences here. Required
domain competences relate primarily to customer service and product sales (e.g., “key”,
“account”, “service”, “deliveri”, “sale”, “market”, “product”), while management
competences relate to competences like project and process management, as suggested
by terms like “lead”, “project”, “consult”, “strategi”, and “process”. A look at high-
loading job titles like CRM Business Solutions Advisor, Support Sales Account
Executive, Director International Marketing Services, and Managing Consultant
Customer Innovation supports the interpretation that this BPM ideal type of professional

26
is engaged both in the design of marketing & sales processes and systems and in the
support and execution of marketing & sales activities.
Business Process Improvement Manager: Most of the terms in this ideal profile
were assigned to the management and problem solving sub-categories (cf. Table 8).
Examples of descriptors that suggest management competences are “manag”, “process”,
“busi”, “project”, “lead”, and “organ”, indicating a need for process and project
management competences. Examples of descriptors related to problem solving
competences include “improve”, “lean”, “six”, and “sigma”, which refer to the Lean
and Six Sigma process improvement methodologies. The titles of high-loading job ads
confirm this proposition (e.g., Business Process Excellence Manager, Lean Consultant,
Director Business Process Improvement).
Table 8 Ideal profile of a Business Process Improvement Manager (7-factors solution)
Ideal Type
Business Process Improvement Manager
High-loading job titles
Business Process Excellence Manager, Lean Consultant, Operations Excellence Manager, Director
Business Process Improvement, Business Process Re-Engineering Leader
Competence requirements
Category
Sub-Category
High-loading descriptive terms (stemmed; in order of
highest loading terms)
Technical
Hardware
-
Software
tool, engin, excel
Business
Domain
oper, manufactur, qualiti
Management
manag, process, busi, project, lead, continu, organ, strategi,
leadership, plan, initi, demonstr, organiz, effect, leader,
execut, practic, strategy
Social
chang, drive, influenc, coach, facilit, cross, team
Systems
Problem Solving
improv, lean, sigma, six, belt, methodolog, techniqu,
measur, black, metric, control, key, identifi, perform, certif
Development
implement, model, develop

27
ERP Solution Architect: The high-loading terms show a balanced distribution
across software, domain, management, and development competences, and the terms tie
in well with each other. While most terms in the sub-category of software competences
refer to specific modules of SAP ERP (e.g., “sd” for sales & distribution, “fi” for
finance, “mm” for materials management, and “pp” for production planning), terms in
the development sub-category relate to the customisation of such modules (e.g.,
“configur”, “implement”, “integr”, “test”). Domain-related terms refer to areas like
supply chain management, finance, and manufacturing, the very domains that are
supported by the ERP modules mentioned above. Finally, terms related to management
and social competences refer to competences required in classic consulting projects
(e.g., “lead”, “project”, “consult”, “provid”, “plan”, “team”, “travel”).
IT-Business Strategy Manager: Most of the descriptive terms for this profile
were assigned to the management sub-category. In particular, strategic planning and
decision making are key competence areas, as indicated by terms like “strateg”, “plan”,
“risk”, “resourc”, “recommend”, “decis”, “ensur”, and “perform”. However, social
competences like collaboration and communication (e.g., “team”, “particip”, “share”,
“activ”, “communic”) also play an important role in this job profile. The aspect of IT-
business alignment becomes apparent through high-loading job titles like IT Business
Analysis Team Lead, IT Strategy Architecture Manager and IT-Business Consultant.
Technical Architect: Almost all high-loading terms were assigned to the
software sub-category. Descriptors refer mostly to specific technologies, products, and
vendors (e.g., “oracl”, “java”, “net”, “framework”, “sharepoint”, “ibm”, “microsoft",
“soa”). The high-loading terms also indicate a clear need for development competences

28
(e.g., “design”, “architectur”, “implement”, “configur”, “integr”, “test”). High-loading
job titles like Technical Analyst and Java Architect underscore the technical
competences required in this job profile.
Comparing and contrasting various profiles shows that their focus differs
substantially. The relative importance of each competence sub-category was calculated
by determining the proportion of high-loading terms in a given sub-category to the total
amount of high-loading terms assigned to the ideal type (Figure 4). This analysis shows
major differences in the competence profiles of the seven ideal types. While it can
generally be said that BPM professionals require a mix of technical, business, and
systems competences, the relative importance of these three categories differs
significantly among the identified seven ideal types. Typically, two to three of the seven
detailed competence sub-categories defined by Todd et al. (1995) play a major role in a
particular BPM job.
For example, for the Technical Architect ideal type, around 60 per cent of the
high-loading terms were classified as technical (hardware and software), while
approximately 15 per cent of the terms were placed into the business category and 25
per cent into the systems category. In contrast, the position Business Process Manager
Sales & Marketing focused on business competences (domain, management, and
social), as around 80 per cent of the high-loading terms were assigned to this category,
while technical and systems competences were approximately equally distributed
among the remaining 20 per cent of the terms. Looking at the Business Process Analyst
ideal type, around 45 per cent of the high-loading terms were allocated to the systems
category (problem solving and development), while another 45 per cent of the terms
referred to business competences and around 10 per cent to technical competences.

29
In sum, the comparison of the ideal profiles shows major differences regarding
their scope and focus. In fact, many of the BPM positions analysed revealed one core
competence area. For instance, the IT-Business Strategy Manager shows a clear
concentration on business competences. However, there are also ideal types which
require an equal distribution of competences among the three main categories. The
Business Process Improvement Manager shows an almost equal distribution of high-
loading terms across the technical, business, and systems competence categories.
Figure 4 Overview of Ideal Profiles of BPM Professionals
4.3. Propositions
The results of the empirical analysis in this study suggest a number of propositions
regarding the ideal types and profiles of BPM positions and their consequences for
individual professionals and organisations.
Ideal types and profiles can be identified on several levels of abstraction.
Typologies ranging from 2 to 20 ideal types were empirically derived. While the 2-
factor solution resulted in equal partitioning into business- and systems-related jobs, the
20-factor solution revealed more specific job types. Hence, it is posited that ideal types
of BPM professionals can be defined on several levels of abstraction, ranging from a

30
coarse-grained distinction between Business Analysts and Systems Analysts to a fine-
grained differentiation among positions responsible for the various parts of the BPM
lifecycle (e.g., analysis, improvement, compliance), types of business processes (e.g.,
sales & marketing, supply chain, accounting & finance), and classes of supporting IT
systems (e.g., ERP, BI, ECM) (Proposition #1a).
Because of the lack of classification systems for BPM-specific competences, the
IS knowledge and skill framework developed by Todd, McKeen, and Gallupe (1995)
was applied to define ideal profiles of BPM professionals. The framework proved to be
a valuable lens through which to analyse and classify BPM-related competences. In
particular, the analysis revealed that the ideal types of BPM professionals identified are
T-shaped or π-shaped, that is, general competences in all three categories of
competences and a specialisation in one or two specific sub-categories are required.
Hence, it can be stated that each ideal type of BPM professional is associated with an
ideal profile that consists of a broad range of technical, business, and systems
competences and a specialisation in one or two specific sub-categories of competences
(e.g., domain, management, software) (Proposition #1b).
It can also be argued that, in order to be highly employable, individual BPM
professionals should possess competences that are aligned with one of the ideal profiles
identified. The rationale behind this argument is that employability can be broadly
defined as an individual’s ability to get and keep fulfilling work, and “for the individual,
employability depends on the knowledge, skills and attitudes they possess” (Hillage and
Pollard 1999, p. 2). Hence, as the ideal profiles have been derived from an empirical
analysis of the demands of the BPM labour market, it can be posited that individuals
who possess competences that are aligned with any ideal profile of a BPM professional

31
are likely to be more employable in the BPM field than are individuals who possess
competences that are not aligned with any of the ideal profiles (Proposition #2a).
At an organisational level, it can be argued that, in order to achieve BPM
success, process-oriented organisations should have a complete portfolio of all the ideal
types of BPM professionals from one level of abstraction. Following the logic of
contemporary BPM maturity models, the underlying assumption of this argument is that
BPM success (i.e., efficient and effective business processes) is a function of the depth
and width of organisational BPM capabilities (e.g., Rosemann & de Bruin 2005b; de
Bruin & Rosemann 2007). According to the resource-based view of the firm,
organisational capabilities stem largely from having valuable, rare, inimitable, and non-
substitutable resources in the firm (Wright 2001). It can be argued that the ideal types of
BPM professionals that were identified in this study represent such resources because
they have been derived from analysing and aggregating competences that are in high
demand on the labour market. Hence, it is posited that organisations that have a
complete portfolio of BPM professionals that is aligned with the overall set of ideal
types for a given level of abstraction have better business process performance than do
organisations that have a portfolio that is not aligned with the overall set of ideal types
(Proposition #2b).
It can also be argued that each organisation faces specific contextual factors that
influence the type of required BPM personnel. Large companies with a highly mature
BPM tend to employ more BPM-specific personnel than do smaller companies with less
BPM maturity (Rosemann and de Bruin 2005b; Hammer 2007). Likewise, more BPM
maturity is typically associated with a broader portfolio of BPM capabilities (Jesus et al.
2009). Hence, it is posited that organisational context factors (e.g., company size, BPM

32
maturity) determine the appropriate level of abstraction of the ideal types of BPM
professionals (Propositions #2c).
5. Discussion
The objective of this study was to clarify individual BPM competence requirements.
Comparable empirical research in the field of BPM is scarce, but two studies related to
the research objective of this study were identified and the results were compared and
contrasted.
Launonen and Kess (2002) investigate the roles and skills relevant for business
process re-engineering (BPR) teams based on five case studies of manufacturing units
of electronics companies. They apply the team roles classification proposed by Platt et
al. (1988) and argue that a BPR team should have eight functional skills: innovation,
resource investigation, organising, teamwork, meeting, finishing, evaluation, and
project work. While the identified roles and skill requirements reflect the one-time
project nature of classic BPR, the positions and competence profiles generated in this
study are based on understanding BPM as a holistic and continuous management
approach, so they also include continuous positions and associated competence
requirements (e.g., Business Process Manager Sales & Marketing).
Another recent study of BPM positions and responsibilities, by Antonucci and
Goeke (2011), is based on the Gartner report, “Role Definition and Organisational
Structure: Business Process Improvement” (Melenovsky and Hill 2006). The authors
identify four BPM positions—Business Process Director, Business Process Consultant,
Business Process Architect, and Business Process Analyst—and related responsibilities
using a survey-based approach. These four positions are well in line with the results of
this study. Most of the responsibilities of a Business Process Director coincide with the

33
competences required from an IT-Business Strategy Manager, and the Business Process
Consultant role involves competences, which in this study are assigned to a Business
Process Improvement Manager and a Business Process Analyst. The responsibilities of
a Business Process Architect correlate to those identified, for example, for Technical
Architects and ERP Solutions Architects, and the Business Process Analyst role largely
corresponds to the Business Process Analyst and Business Process Compliance
Manager profiles. Therefore, the seven ideal types of BPM professionals determined in
this study do not contradict the results attained by Gartner and verified in Antonucci and
Goeke (2011) but provide a more detailed and comprehensive view of this topic.
5.1. Implications for Research
This study opens several opportunities for future research. First, a typology of BPM
professionals from an empirical analysis of job advertisements in the area of BPM was
inductively constructed, but empirical testing of this typology is needed to complete the
theorising process. Such testing requires operationalising the typology constructs,
measuring those constructs in practice, and testing the stated propositions. For example,
the fit between the identified ideal profiles and the competence profiles of actual BPM
professionals must be measured (Doty and Glick 1994) by developing a measurement
scale that operationalises the categories and sub-categories of the ideal profiles. This
scale could then be used to measure the degree to which a candidate possesses the
competences in those categories. Testing the stated propositions at an individual level
also requires measuring and comparing the employability of BPM professionals
(Rothwell and Arnold 2007), while at an organisational level, testing the propositions
requires analysing the alignment between a company’s BPM competence portfolio and
the identified ideal types and measuring the relationship between a portfolio of ideal

34
types of BPM professionals and a company’s BPM success (Karimi, Somers, and
Bhattacherjee 2007).
A second opportunity for future research arises from having used LSA to
analyse the content of job ads and derive ideal types and profiles of BPM professionals,
a process that included both automated statistical analysis and manual coding and
interpretation. The automated part of LSA allowed us to uncover the latent semantics of
texts and address the synonymy problem in natural language processing. These are two
major advantages compared to simpler information retrieval models, such as Simple
Boolean Matching (SBM) and the Vector Space Model (VSM) (Manning, Raghavan,
and Schutze 2008). Compared to text-analysis techniques that rely on formal semantics
(e.g., enhanced Topic-based Vector Space Model (eTVSM) (Polyvyanyy and Kuropka
2007)), LSA requires considerably less manual effort. However, formally defining
concepts and relationships of a domain (e.g., in an ontology) would consider more
linguistic phenomena (e.g., collocations and word ordering) and allow for logical
reasoning (e.g., making inferences). Future research may try to combine the efficiency
of the approach proposed here with the accuracy of more formal semantic techniques.
Manual coding of job ads, another alternative to the approach applied in this
study, has been the standard approach to analysing job ads in the past, but it is subject to
risks, such as bias from personal interpretations, so it requires triangulation of multiple
researchers in order to yield reliable results. Moreover, because of its cost, manual
analysis is usually restricted to small samples of ads and is less likely to be repeated
regularly. In an attempt to combine the strengths of automated and manual analysis, the
LSA was conducted in iterative cycles, where multiple researchers reviewed the
plausibility of the automated statistical analysis and interpreted its results. A major part
of these manual interventions focused on pre-processing job ads and filtering out

35
irrelevant words. Several attempts to automate this process were made, including trying
to identify text blocks with standard company descriptions algorithmically (e.g., by
identifying HTML tags that surround such blocks or using a plagiarism finder to detect
recurring standard phrases). Systematically exploring strategies for reducing the manual
effort of the applied approach is a worthwhile topic for future research.
Finally, future studies may also use the methodology developed in this study to
analyse other areas of the IS job market (e.g., Business Intelligence, Enterprise
Resource Planning).
5.2. Implications for Practice
This study provides several implications for practice. First, its results can be used to
assess and develop BPM capabilities at both the individual and the organisational level.
At an individual level, the results show pathways for career choices and decisions
regarding continuing education, and the competence sets of the ideal types provide
guidance for individuals’ professional development. At an organisational level, the
identified ideal types and profiles can be used as blueprints for strategic HR
management (e.g., establishment of a BPM Centre of Excellence) and staffing decisions
(e.g., for BPM projects).
Our findings can also be used for educational purposes, as the ideal types of
BPM professionals provide direction for the development of specific educational
programmes. For example, curricula may be developed on the basis of the identified
BPM competences that are required in practice.
5.3. Limitations
The conducted research also contains some limitations. First, there may be limitations
concerning the data source. As this study considers only current ads on BPM-related

36
jobs, no conclusions on the historical development of jobs in the BPM field can be
drawn, only insights into the current state of employment requirements can be provided.
Future research may determine how BPM-related job profiles change over time in an
attempt to predict future requirements. In addition, the study considered only ads from
one employment platform and only jobs in the US, Canada, the UK, and Australia.
Therefore, the used data set may be culturally biased, as it covers only Anglo-Saxon
countries, so the generalisability of the attained results is limited to this culture. Future
research should consider job ads from other cultural backgrounds.
Another limitation refers to the use of job ads as input data for the analysis of
BPM competences, as job ads may include biases. For example, they may be written to
suit a particular individual, they may ask for more competences than can be reasonably
expected from one applicant, and they may be based on competence categories that HR
thinks the business needs (or applicants will understand), whether it does or not. While
it is acknowledged that such biases may exist in the analysed data, it is believed that
they are not significant for the research results, as the number of job ads that were
examined should be sufficiently large to minimise the effect of biases in a few ads. The
processing of such a broad data source as that used in this research gives a particular
advantage to the here applied approach over other research methods, such as interviews,
because it diminishes the risk of biases caused by specific contextual backgrounds.
Finally, limitations arise from the limitations of LSA. LSA is based on a bag-of-
words model that ignores word orders and collocations, which can lead to words being
taken out of context. For example, if the high-loading terms “Oracle” and “database”
were extracted, it could be inferred that the related jobs require some kind of
competence in the area of Oracle databases, but it is still not known whether the
candidate should possess basic knowledge in Oracle databases, ten years of practical

37
experience in the area, or an official Oracle certification. Making such conclusions
requires the co-occurrence of other descriptive terms, such as “experience” or
“certification”.
In addition, LSA only partly addresses the problem of polysemy (i.e., the
problem that some words have more than one meaning). From a computational
perspective, polysemy is represented in LSA by descriptive terms that load high on
multiple factors (Deerwester et al. 1990), so inspecting the terms that co-occur with a
specific term helps to determine the context-specific meaning of that term. For example,
when “bank” co-occurs with “money”, it likely relates to financial institutions, while it
probably refers to geography when it co-occurs with “river”. Recognising and
interpreting such patterns is not straightforward and requires substantial domain
knowledge.
As LSA is based upon a bag-of-words representation of documents, it is less
suitable for analysing texts that contain metaphorical languages (e.g., poems), causal
reasoning (e.g., legal texts), or temporally or logically ordered lists (e.g., technical
manuals) (Wolfe and Goldman 2003; Kintsch 2001). However, such linguistic
phenomena are rare in job advertisements, so they played a minor role in the empirical
analysis. Instead, the ads were characterised by the occurrence of very specific
vocabulary with numerous technical terms and proper names denoting technologies or
products, a situation in which LSA usually operates well (Giesbers, Rusman, and
Bruggen 2006). In fact, experiments have shown that in domains with very specific
vocabularies (e.g., science and technology), the performance of LSA can rival the
performance of human coders in terms of accuracy (Landauer et al. 1997). Nonetheless,
although the required effort is substantially lower, the heuristic approach of identifying
patterns of BPM competences taken in this study does not reach the level of accuracy

38
that can theoretically be reached by manually enriching natural language job
descriptions with formal semantic annotations (e.g., RDF, OWL).
In order to address the above mentioned limitations, future research should focus
on triangulating the here presented results with findings generated by other empirical
methods. This process can take three forms: First, one can compare the LSA results
with results obtained by human coders analysing the same data set. At this, standard
content analysis procedures and metrics (e.g., inter-coder reliability) can be used to
judge the reliability and quality of the LSA results.
Second, one can ask human experts to assess the relevance of the LSA results
(without analysing the underlying data set). Recently, a number of researchers have
outlined processes for such evaluations. Chang et al. (2009), for example, have
conducted experiments in which they have intentionally inserted random factors into the
overall set of latent semantic factors and asked participants to identify these “intruders”.
Likewise, they also inserted random terms into the list of high-loading terms associated
with a latent semantic factor and again asked participants to identify the “intruders”.
Low ability of human judges to identify intruders indicated a low coherence and
relevance of the derived factors and terms.
Third, the here presented results can be compared and contrasted with findings
obtained by applying other empirical methods. For example, one could conduct a
Delphi study with BPM experts from various industries and regions to define typical
BPM jobs and related profiles. This method relies on the use of expert opinions to
obtain reliable consensus via a series of questionnaires with controlled feedback
(Dalkey and Helmer 1963). It is typically applied to structure group consensus finding
processes in complex issues which require diverse backgrounds, as for example, in the

39
case of identifying BPM competences that are required in different companies,
industries or regions (Linstone and Turoff 1975).
6. Conclusion
Our study provides a comprehensive understanding of BPM competence requirements
as seen from a practitioner’s perspective. More than 1,500 current BPM job
advertisements were examined through application of a state-of-the-art text-mining
technique (LSA) and identified distinct ideal types and ideal profiles of BPM positions
on several levels of abstraction. The analysis revealed that the competence requirements
differ markedly among the various job types, yet all of them consist of a mix of
technical, business, and systems competences. While most BPM jobs have a particular
focus on one or two areas, some jobs require an equal distribution of competences in all
areas. This study contributes to the existing IS body of knowledge on BPM by (1)
proposing a typological theory of ideal types of BPM professionals, (2) giving practical
guidance for building individual and organisational BPM competences, (3) informing
the development of BPM curricula, and (4) demonstrating the usefulness of LSA for
exploratory theory-building research.

40
7. References
Altinkemer, K, Y. Ozcelik, and Z.D. Ozdemir. 2011. “Productivity and Performance
Effects of Business Process Reengineering: A Firm-Level Analysis.” Journal of
Management Information Systems 27 (4): 129–161.
Antonucci, Y.L., and R.J. Goeke. 2011. “Identification of Appropriate Responsibilities
and Positions for Business Process Management Success: Seeking a Valid and
Reliable Framework.” Business Process Management Journal 17 (1): 127–146.
Bandara, W., D. R. Chand, A. M. Chircu, S. Hintringer, D. Karagiannis, J. Recker, A.
Van Rensburg, C. Usoff, and R. J. Welke. 2010. “Business Process Management
Education in Academia: Status, Challenges, and Recommendations.”
Communications of the AIS 27 (1): 743–776.
Brady, R.H. 1967. “Computers in Top-Level Decision Making.” Harvard Business
Review 45 (4): 67–76.
Chang, J, JL Boyd-Graber, S Gerrish, C Wang, and DM Blei. 2009. “Reading Tea
Leaves: How Humans Interpret Topic Models.” In Neural Information Processing
Systems (NIPS) Conference. Vancouver.
Dalkey, N, and O Helmer. 1963. “An Experimental Application of the Delphi Method
to the Use of Experts.” Management Science 9 (3): 458–467.
Davenport, T H, and J E Short. 1990. “The New Industrial Engineering: Information
Technology and Business Process Redesign.” Sloan Management Review 31 (4).
De Bruin, T., and M. Rosemann. 2007. “Using the Delphi Technique to Identify BPM
Capability Areas.” In Proceedings of the Australasian Conference on Information
Systems. Toowoomba.
Deerwester, S., S. T. Dumais, G. W. Furnas, T. K. Landauer, and R. Harshman. 1990.
“Indexing by Latent Semantic Analysis.” Journal of the American Society for
Information Science and Technology 41 (6): 391–407.
Doty, D.H., and W.H. Glick. 1994. “Typologies as a Unique Form of Theory Building:
Toward Improved Understanding and Modeling.” Academy of Management
Review 19 (2): 230–251.
Douglass, John Aubrey, and Richard Edelstein. 2009. “The Global Competition for
Talent. The Rapidly Changing Market for International Students and the Need for a
Strategic Approach in the US.” Research and Occasional Papers Series, Center for
Studies in Higher Education, UC Berkeley.
Evangelopoulos, N., X. Zhang, and V. R. Prybutok. 2012. “Latent Semantic Analysis:
Five Methodological Recommendations.” European Journal of Information
Systems 21 (December 21): 70–86.

41
Gallivan, M. J., D. P. Truex, and L. Kvasny. 2004. “Changing Patterns in IT Skill Sets.
A Content Analysis of Classified Advertising.” DATA BASE for Advanced in
Information Systems 35 (3): 64–87.
Gartner. 2011. “Gartner CIO Survey 2011.”
http://www.gartner.com/it/page.jsp?id=1283413.
Giesbers, B, E Rusman, and J Van Bruggen. 2006. “State of the Art LSA”. Heerlen.
http://lnx-hrl-075v.web.pwo.ou.nl/handle/1820/685.
Govindarajan, V. 1988. “A Contingency Approach to Strategy Implementation at the
Business-Unit Level: Integrating Administrative Mechanisms with Strategy.”
Academy of Management Journal 31 (4): 828–853.
Guillemette, M.G., and Guy Paré. 2012. “Toward a New Theory of the Contribution of
the IT Function in Organizations.” MIS Quarterly 36 (2): 529–551.
Haley, DT, P Thomas, A De Roeck, and M Petre. 2005. “A Research Taxonomy for
Latent Semantic Analysis-Based Educational Applications.” International
Conference on ….
Hammer, M. 2007. “The Process Audit.” Harvard Business Review 85 (4): 111–123.
Hammer, M., and J. Champy. 1993. Reengineering the Corporation: A Manifesto for
Business Revolution. Harper Business. New York: HarperBusiness.
Hardin, A.M., K.D. Joshi, and X. Li. 2002. “Business as Usual? IS Job Skill
Requirements During Internet Era.” In Proceedings of the Americas Conference on
Information Systems. Dallas, Texas.
Hillage, J., and E. Pollard. 1999. “Employability: Developing a Framework for Policy
Analysis.”
https://www.education.gov.uk/publications/eOrderingDownload/RB85.pdf.
Indulska, M., D. S. Hovorka, and J. Recker. 2012. “Quantitative Approaches to Content
Analysis: Identifying Conceptual Drift across Publication Outlets.” European
Journal of Information Systems 21: 49–69.
Jansen, Bernard J., Karen J. Jansen, and Amanda Spink. 2005. “Using the Web to Look
for Work: Implications for Online Job Seeking and Recruiting.” Internet Research
15 (1): 49 – 66.
Jauch, LR, RN Osborn, and TN Martin. 1980. “Structured Content Analysis of Cases: A
Complementary Method for Organizational Research.” Academy of Management
Review 5 (4): 517–525.
Jesus, L., A. Macieira, D. Karrer, and M. Rosemann. 2009. “A Framework for a BPM
Center of Excellence.” BPTrends: 1–9.

42
Karimi, J., T.M. Somers, and A Bhattacherjee. 2007. “The Impact of ERP
Implementation on Business Process Outcomes: A Factor-Based Study.” Journal
of Management Information Systems 24 (1): 101–134.
Kintsch, W. 2001. “Predication.” Cognitive Science 25 (2): 173–202.
Laham, D, W Bennett, and M Derr. 2002. “Latent Semantic Analysis for Career Field
Analysis and Information Operations.” In The Interservice/Industry Training,
Simulation & Education Conference (I/ITSEC). Orlando.
Laham, D., W. Bennet, and T. K. Landauer. 2000. “An LSA-Based Software Tool for
Matching Jobs, People, and Instruction.” Interactive Learning Environments 8 (3):
171–185.
Landauer, T. K., D Laham, B Rehder, and ME Schreiner. 1997. “How Well Can
Passage Meaning Be Derived without Using Word Order? A Comparison of Latent
Semantic Analysis and Humans.” In Proceedings of the Nineteenth Annual
Conference of the Cognitive Science Society. Stanford.
Landauer, T.K., P.W. Foltz, and D. Laham. 1998. “An Introduction to Latent Semantic
Analysis.” Discourse Processes 25 (2-3): 259–284.
Larsen, K. R., D. E. Monarchi, D. S. Hovorka, and C. N. Bailey. 2008. “Analyzing
Unstructured Text Data: Using Latent Categorization to Identify Intellectual
Communities in Information Systems.” Decision Support Systems 45 (4): 884–896.
Launonen, Martti, and Pekka Kess. 2002. “Team Roles in Business Process Re-
Engineering.” International Journal of Production Economics 77 (3) (June): 205–
218.
Le Deist, Françoise Delamare, and Jonathan Winterton. 2005. “What Is Competence?”
Human Resource Development International 8 (1) (March): 27–46.
Lee, D.M.S., E.M. Trauth, and D. Farwell. 1995. “Critical Skills and Knowledge
Requirements of IS Professionals: A Joint Academic/Industry Investigation.” MIS
Quarterly 19 (3): 313–340.
Linstone, HA, and M Turoff. 1975. The Delphi Method: Techniques and Applications.
London: Addison-Wesley.
Litecky, C., and A. Aken. 2010. “Mining for Computing Jobs.” IEEE Software 27 (1):
78–85.
Litecky, C., A. Aken, B. Prabhakar, and K. Arnett. 2009. “Skills in the MIS Job
Market.” In Proceedings of the Americas Conference on Information Systems. San
Francisco.
Manning, C. D., P. Raghavan, and H. Schutze. 2008. Introduction to Information
Retrieval. New York: Cambridge University Press.

43
McKinsey. 2011. “McKinsey Global Survey Results: A Rising Role for IT.”
https://www.mckinseyquarterly.com/PDFDownload.aspx?ar=2900.
Melenovsky, M.J., and J.B. Hill. 2006. “Role Definition and Organizational Structure:
Business Process Improvement”. Stamford. http://www.gartner.com/id=733320.
Mintzberg, Henry. 1979. The Structuring of Organizations. Englewood Cliffs, NJ:
Prentice Hall.
Nelson, R.R. 1991. “Educational Needs as Perceived by IS and End-User Personnel: A
Survey of Knowledge and Skill Requirements.” MIS Quarterly 15 (4): 503–525.
Nordhaug, O. 1993. Human Capital in Organizations: Competence, Training, and
Learning. Oslo: Scandinavian University Press.
Platt, S., R. Piepe, and J. Smyth. 1988. Teams - A Game To Develop Group Skills.
London: Gower Press.
Polyvyanyy, A., and D. Kuropka. 2007. “A Quantitative Evaluation of the Enhanced
Topic Based Vector Space Model.”
Rosemann, M., and T. de Bruin. 2005a. “Towards a Business Process Management
Maturity Model.” In Proceedings of the European Conference on Information
Systems. Regensburg.
———. 2005b. “Application of a Holistic Model for Determining BPM Maturity.”
BPTrends: 1–21.
Rothwell, Andrew, and John Arnold. 2007. “Self-Perceived Employability:
Development and Validation of a Scale.” Personnel Review 36 (1): 23–41.
Sidorova, A., N. Evangelopoulos, J. S. Valacich, and T. Ramakrishnan. 2008.
“Uncovering the Intellectual Core of the Information Systems Discipline.” MIS
Quarterly 32 (3): 467–482.
Sidorova, A., and O. Isik. 2010. “Business Process Research: A Cross-Disciplinary
Review.” Business Process Management Journal 16 (4): 566–597.
Tang, H. L., S. Lee, and S. Koh. 2000. “Educational Gaps. Knowledge, as Perceived by
IS Educators: A Survey of Knowledge and Skill Requirements.” Journal of
Computer Information Systems 41 (2): 76–84.
Todd, P.A., J.D. McKeen, and R.B. Gallupe. 1995. “The Evolution of IS Job Skills: A
Content Analysis of IS Job Advertisements from 1970 to 1990.” MIS Quarterly 19
(1): 1–27.
Van Looy, Amy, M De Backer, and G Poels. 2012. “A Conceptual Framework and
Classification of Capability Areas for Business Process Maturity.” Enterprise
Information Systems iFirst: 1–37.

44
Vom Brocke, Jan, and M. Rosemann. 2010. Handbook on Business Process
Management 1 & 2. Heidelberg: Springer.
Weber, M. 1949. The Methodology of the Social Sciences. New York: Free Press.
Wetzstein, B, and P Leitner. 2011. “Identifying Influential Factors of Business Process
Performance Using Dependency Analysis.” Enterprise Information Systems 5 (1):
79–98.
Wolfe, MBW, and SR Goldman. 2003. “Use of Latent Semantic Analysis for Predicting
Psychological Phenomena: Two Issues and Proposed Solutions.” Behavior
Research Methods, Instruments, & Computers 35 (1): 22–31.
Wright, PM. 2001. “Human Resources and the Resource Based View of the Firm.”
Journal of Management 27 (1): 701–721.

45
8. Appendix
8.1. Data Analysis
Table 9 Data collection and analysis process
Phase
Step
Activity
Input
Output
Tool
Automated/manual
Collection and Pre-processing
of Job Ads
1
Collect job
ads
Search terms
(e.g.,
“business
process”)
Corpus of
2,324 job
ads
Python
script
Automated
2
Remove
unwanted job
ads (e.g.,
duplicates,
spam)
Corpus of
2,324 job
ads
Corpus of
1,507 job
ads
Duplicate
finder
Manual/Automated
3
Remove
unwanted text
blocks (e.g.,
company
descriptions)
Corpus of
1,507 job
ads
Corpus of
1,507 job
ads
Text editor
Manual
4
Stem terms
Corpus of
1,507 job
ads,
Dictionary
with 11,376
descriptive
terms
Corpus of
1,507 job
ads,
Dictionary
with 6,785
descriptive
terms
R script
Automated
5
Remove stop
words and
unique terms
Corpus of
1,507 job
ads,
Dictionary
with 6,785
descriptive
terms
Corpus of
1,507 job
ads,
Dictionary
with 3,732
descriptive
terms
R script
Automated
6
Remove
uninformative
terms
Corpus of
1,507 job
ads,
Dictionary
with 3,732
descriptive
terms
Corpus of
1,507 job
ads,
Dictionary
with 1,422
descriptive
terms
Spreadsheet
tool
Manual
7
Build term-
document
matrix
Corpus of
1,507 job
ads,
Dictionary
with 1,422
descriptive
terms
Term-
document
matrix
R script
Automated
8
Weight term-
document
matrix
Term-
document
matrix
Term-
document
matrix
R script
Automated

46
Singular Value Decomposition
9
Set number of
factors
-
Number of
factors
-
Manual
10
Decompose
term-
document
matrix
Term-
document
matrix,
Number of
factors
Term
eigenvectors,
Document
eigenvectors,
Singular
Values
R script
Automated
11
Calculate
term and
document
loadings
Term
eigenvectors,
Document
eigenvectors,
Singular
Values
Term
loadings,
Document
loadings
R script
Automated
Interpretation
12
Perform
varimax
rotation
Term
loadings,
Document
loadings
Term
loadings,
Document
loadings
R script
Automated
13
Determine
loading
thresholds
Term
loadings,
Document
loadings
Loading
thresholds
Spreadsheet
tool
Manual
14
Identify high-
loading terms
and
documents
per factor
Term
loadings,
Document
loadings,
Loading
thresholds
Factors with
associated
high-loading
terms and
documents
Spreadsheet
tool
Automated
15
Interpret and
label factors
Factors with
associated
high-loading
terms and
documents
Labelled
factors with
associated
high-loading
terms and
documents
Spreadsheet
tool
Manual
16
Create ideal
types and
profiles of
jobs
Labelled
factors with
associated
high-loading
terms and
documents
Ideal types,
ideal profiles
Text editor
Manual
17
Calculate
descriptive
statistics
Labelled
factors with
associated
high-loading
terms and
documents
Descriptive
statistics
(e.g. number
of jobs per
ideal type)
Spreadsheet
tool
Manual

47
8.2. Ideal Profiles of the 7-factors Solution
Table 10 Factor cross-loadings of job ads
Loading
Number of job ads
Percentage of job ads
on 1 factor
511
33.9
on 2 factors
200
13.3
on 3 factors
103
6.8
on 4 factors
38
2.5
on 5 factors
18
1.2
on 6 factors
1
0.1
on 7 factors
1
0.1
on no factor
635
42.1
Table 11 Ideal profile of a Business Process Analyst
Ideal type
Business Process Analyst
High-loading job titles
Business Analyst, Business Systems Analyst, Business Process Analyst
Competence requirements
Category
Sub-category
High-loading descriptive terms (stemmed; in order of
highest loading terms)
Technical
Hardware
technic, technolog
Software
system, softwar, data, tool, flow, workflow, product, excel,
interfac, web, environ, object, visio
Business
Domain
qualiti, standard, industri
Management
busi, project, process, manag, plan, provid, effect, perform,
conduct, present, scenario, organ, inform, defin, expert,
lead, initi, ensur, success, senior, recommend, definit,
priorit, execut, desir
Social
team, stakehold, communic, facilit, collabor, interview,
group, workshop, assist, particip, interact, liaison, translat,
chang, accept, verbal, activ, session
Systems
Problem Solving
analyst, analysi, user, specif, gather, analyz, elicit, problem,
identifi, analyt, complex, valid, evalu, request, scope, issu,
research, enhanc, solv, interpret, gap
Development
document, test, case, develop, model, design, methodolog,
implement, creat, write, techniqu, diagram, support, life,
cycl, map, architectur, descript, unit, sdlc, agil

48
Table 12 Ideal profile of a Business Process Compliance Manager
Ideal type
Business Process Compliance Manager
High-loading job titles
Financial Control Analyst, Trade Compliance Manager, Business Process Controls Specialist, Audit
Compliance Specialist
Competence requirements
Category
Sub-category
High-loading descriptive terms (stemmed; in order of
highest loading terms)
Technical
Hardware
technic
Software
system, excel, sql, databas, microsoft, offic, access, oracl,
softwar, ms
Business
Domain
control, oper, procedur, data, account, audit, intern, financ,
servic, monitor, polici, financi, complianc, depart, qualiti,
environ, product, qualif, standard, extern
Management
report, manag, ensur, busi, perform, process, provid,
administr, project, inform, coordin, plan, effect,
recommend, assign, complet, prepar, effici, profici
Social
assist, communic, team, activ, staff
Systems
Problem Solving
user, issu, review, improv, analyst, problem, analysi,
analyz, analyt, identifi
Development
support, document, develop, test, maintain, program,
implement, creat, write

49
Table 13 Ideal profile of a Business Process Manager Sales & Marketing
Ideal type
Business Process Manager Sales & Marketing
High-loading job titles
Internal Sales Manager, CRM Business Solutions Advisor, Support Sales Account Executive, Director
International Marketing Services, Managing Consultant Customer Innovation
Competence requirements
Category
Sub-category
High-loading descriptive terms (stemmed; in order of
highest loading terms)
Technical
Hardware
technic, technolog
Software
excel, system, softwar
Business
Domain
sale, market, servic, deliveri, account, key, deliv, financi,
intern, product, industri, oper, environ, plan, record, sell,
commerci, outsourc, valu, global, partner, direct, line,
contract, programm, report
Management
manag, busi, project, organis, process, consult, ensur, lead,
strategi, success, demonstr, provid, proven, director, track,
stakehold, drive, focus, present, effect, execut
Social
team, relationship, chang, communic, activ
Systems
Problem Solving
analysi, improv
Development
develop, support, build, implement

50
Table 14 Ideal profile of a Business Process Improvement Manager
Ideal type
Business Process Improvement Manager
High-loading job titles
Business Process Excellence Manager, Lean Consultant, Operations Excellence Manager, Director
Business Process Improvement, Business Process Re-Engineering Leader
Competence requirements
Category
Sub-category
High-loading descriptive terms (stemmed; in order of
highest loading terms)
Technical
Hardware
-
Software
tool, engin, excel
Business
Domain
oper, manufactur, qualiti
Management
manag, process, busi, project, lead, continu, organ, strategi,
leadership, plan, initi, demonstr, organiz, effect, leader,
execut, practic, strategy
Social
chang, drive, influenc, coach, facilit, cross, team
Systems
Problem Solving
improv, lean, sigma, six, belt, methodolog, techniqu,
measur, black, metric, control, key, identifi, perform, certif
Development
implement, model, develop

51
Table 15 Ideal profile of an ERP Solutions Architect
Ideal type
ERP Solutions Architect
High-loading job titles
Business Solutions Specialist, ERP Business Analyst, SAP Solution Architect, Oracle Functional HR
Consultant, SAP Business Process Expert Finance
Competence requirements
Category
Sub-category
High-loading descriptive terms (stemmed; in order of
highest loading terms)
Technical
Hardware
-
Software
sap, modul, system, erp, sd, fi, mm, oracl, pp
Business
Domain
suppli, chain, financ, global, order, product, expertis,
manufactur
Management
consult, busi, project, manag, process, provid, plan, lead
Social
team, chang, travel
Systems
Problem Solving
user
Development
configur, implement, support, integr, test, design, develop,
lifecycl

52
Table 16 Ideal profile of an IT-Business Strategy Manager
Ideal type
IT-Business Strategy Manager
High-loading job titles
IT Business Analysis Team Lead, IT Strategy Architecture Manager, IT-Business Consultant
Competence requirements
Category
Sub-category
High-loading descriptive terms (stemmed; in order of
highest loading terms)
Technical
Hardware
technolog, technic
Software
enterpris, object, sourc, system, data
Business
Domain
oper, valu, expert, unit, servic, expertis, extern, program,
standard, agreement, intern, monitor
Management
strateg, strategi, plan, risk, inform, resourc, recommend,
decis, manag, organ, initi, direct, busi, ensur, perform,
provid, potenti, trend, goal, leadership, organiz, effect,
establish, impact, leader, project, situat, lead, prioriti, align,
practic, execut, demonstr, approv, probe, achiev, action,
set, wide, perspect, conduct, process, cost, signific, defin,
present, outcom, propos
Social
partner, team, relationship, group, collabor, particip,
anticip, facilit, share, chang, cross, influenc, activ,
communic, member, conflict, peer, interact, engag, assist,
critic
Systems
Problem Solving
identifi, issu, complex, evalu, assess, review, problem,
determin, analyz, solicit, improv, research, analysi
Development
develop, implement, integr, support, maintain

53
Table 17 Ideal profile of a Technical Architect
Ideal type
Technical Architect
High-loading job titles
Technical Analyst, Windows Server Developer, Java Architect, Desktop Virtualization Engineer,
SharePoint Developer
Competence requirements
Category
Sub-category
High-loading descriptive terms (stemmed; in order of
highest loading terms)
Technical
Hardware
server, technolog, technic, infrastructure, network
Software
oracl, enterpris, sql, servic, web, data, databas, java, ibm,
system, softwar, compon, platform, window, tool,
sharepoint, net, xml, secur, framework, webspher, standard,
microsoft, soa, suit, environ, bi
Business
Domain
-
Management
manag, busi, bpm, consult, provid, perform, deliveri
Social
team
Systems
Problem Solving
-
Development
architectur, design, integr, architect, develop, model,
support, implement, configur, test, program