Knowledge Hubs in Competence Analytics:

With a Case Study in Recruitment and Selection

Emiel Caron

and Saša Batistic

School of Economics and Management, Tilburg University, Tilburg, The Netherlands

School of Social and Behavioral Sciences, Tilburg University, Tilburg, The Netherlands

Keywords: Human Resource Analytics, Human Resource DSS, Strategic Competence Analytics, Knowledge Hubs, Data

Integration, Network Analysis.

Abstract: There is a lack of consensus on the usefulness of Human Resource (HR) analytics to achieve better business

results. The authors suggest this is due to lack of empirical evidence demonstrating how the use of data in the

HR field makes a positive impact on performance, due to the detachment of the HR function from accessible

data, and due to the typically poor IT infrastructure in place. We provide an in-depth case study of Strategic

Competence analytics, as an important part of HR analytics, in a large multinational company, labelled ABC,

which potentially shows two important contributions. First, we contribute to HR analytics literature by

providing a data-driven competency model to improve the recruitment and selection process. This is used by

the organization to search more effectively for talents in their knowledge networks. Second, we further

develop a model for data-driven competence analytics, thus also contributing to the information systems

literature, in developing specialized analytics for HR, and by finding appropriate forms of computerized

network analysis for identifying and analysing knowledge hubs. Overall, our approach, shows how internal

and external data triangulation and better IT integration makes a difference for the recruitment and selection

process. We conclude by discussing our model’s implications for future research and practical implications.

1 INTRODUCTION

Human Resource (HR) analytics is a relatively young

and underexplored research field under the umbrella

of Big Data Analytics (BDA) (Batistič and Van der

Laken, 2018). HR analytics enables organizations to

use descriptive, visual, and statistical analyses of data

related to HR processes to establish business impact

and enable data-driven decision-making (Marler and

Boudreau, 2017). Empirical evidence showing the

beneficial role of HR analytics in the HR function and

in the organization, in general, is scarce (e.g., Van den

Heuvel and Bondarouk, 2016; Marler and Boudreau,

2017). Various potential pitfalls why HR analytics

fails to deliver have been identified, but there seems

to be consensus, that the HR Information Technology

(HR IT) might limit its function as an HR analytics

inhibitor: data could be inaccurate, inaccessible,

outdated, lack depth, could not have been collected at

all, or could not be integrated across function,

geography or department (Marler and Boudreau,

2017). Especially combining internal and external

data could provide challenges and pitfalls, such as

poor comparability among others (Brown and

Vaughn, 2011). Improvements should be made in HR

IT to improve its status as an enabler for HR analytics,

especially integrating different data silos (e.g.,

Rasmussen and Ulrich, 2015), which could provide

more evidence of the usefulness of HR analytics.

This scarcity of empirical evidence leads some

authors to speculate that HR analytics might be

considered a fad (Rasmussen and Ulrich, 2015;

Angrave, Charlwood, Kirkpatrick, Lawrence, and

Stuart, 2016), yet anecdotal evidence from practice

suggest that HR analytics is gaining momentum in the

everyday business. We suggest, that more evidence is

needed to provide support to the notion that HR

analytics is providing added value to the business

(e.g., Rasmussen and Ulrich, 2015).

In this paper we focus on a special class of HR

analytics and HR DSS, named strategic competence

analytics for recruitment and selection. Exploring

various issues in HR, such as recruitment and

selection, defined as the “process of seeking

applicants for potential employment” (Noe,

Hollenbeck, Wright, and Gerhart, 2014, p. 762),

Caron, E. and Batistic, S.

Knowledge Hubs in Competence Analytics: With a Case Study in Recruitment and Selection.

DOI: 10.5220/0008117805850594

In Proceedings of the 14th International Conference on Software Technologies (ICSOFT 2019), pages 585-594

ISBN: 978-989-758-379-7

585

might serve as a case for examining the additional

positive effects and usefulness of HR analytics for

business (Marler and Boudreau, 2017). When

recruiting potential employees, it is important to

know what the company is looking for in a new

employee. The description of a competence helps, as

it provides HR recruiters with a set of knowledge,

skills, and abilities (KSA’s) combined with personal

characteristics to look for in potential employees.

This competency describes important enablers for

employees to fulfill their job (Noe, Hollenbeck,

Wright, and Gerhart, 2014). Strategic competencies

are competencies that the company must obtain, in

order for them to fulfill their business strategy

(Barney, 1991).

This research aims to provide a practical business

case study example, which shows the potential of HR

analytics and BDA in the HR function. In doing so we

answer a call to provide more empirical evidence

showing the business value of HR analytics on a

recruitment case (van den Heuvel and Bondarouk,

2016; Marler and Boudreau, 2017, Batistič and Van

der Laken, 2018). We base our case study on

knowledge hubs, defined by Knight (2013) as a group

of individuals that focus on “… the production and

application of new knowledge which has the potential

for commercial use”. Knowledge hubs are

characterized by high connectedness and high

internal and external networking and knowledge

sharing capabilities (Evers, 2008). Hubs have three

major functions: producing knowledge, putting

knowledge to practice, and passing on this knowledge

through education and training. In HR, knowledge

hubs could be used for more effective recruitment

activities by analyzing both internal, (for example:

employee’s education, conferences visited,

publications made) and external data (for example:

graduates of a certain study, LinkedIn profiles

regarding a business function), and then linking the

internal and external data to model relationships

between the enterprise and the knowledge hubs.

Having this information, the HR department can

decide on which knowledge hubs to focus their

recruitment efforts and create an action plan. Our case

study shows how to answer typical questions in

competence analytics and derive related metrics, like:

 Where are knowledge hubs related to the

company located?

 Which knowledge hubs are relevant for the

company and what is the size of the relevant

knowledge hubs?

 How close are the knowledge hubs to the

company in terms of internal and external network

analysis?

Our study has two important theoretical

contributions. First, it provides further evidence of the

usefulness of HR and competence analytics in a

business environment and puts the emphasis on

knowledge hubs in the recruitment and selection

domain and provides a convenient practical model

that organizations can use. In doing so, we also show

how overcoming the limitations of the silo mentality,

due to lack of HR IT infrastructure (Rasmussen and

Ulrich, 2015) in place, and how a link between

internal and external databases valuable for the

recruitment process, further enhance the value of HR

analytics (Brown and Vaughn, 2011. Second, we also

contribute to the information systems (IS) literature

by adapting existing models for strategic workforce

planning from Phillips and Gully (2009) and

Makarius and Srinivasan (2017) to data-driven

competence analytics. We show the usefulness of this

model in a case study at the ABC-company.

This paper is structured as follows. In Section 2, a

model for data-driven competence analytics is

presented. This model is used in a case study in

Section 3, to identify important knowledge hubs in a

company's network of related institutions and

individuals. Finally, theoretical and practical

conclusions are drawn.

2 STRATEGIC COMPETENCE

ANALYTICS

2.1 Recruitment and Knowledge Hubs

Human Resource Management (HRM) refers to the

“policies, practices, and systems that influence

employees’ behavior, attitudes, and performance”

(Noe et al., 2014). The strategic objective of HRM is

to maximize their positive influence on company

performance. The recruitment and selection process

is an integral part of HRM. We consider a potential

employee’s KSA’s to be heavily related concepts that

HR recruiters look for in this process. In HR literature

(Marler and Boudreau, 2017), the importance of

KSA’s is stressed in the knowledge-based view of the

firm, this view sees knowledge as the most important

strategic resource of the firm for value creation.

Furthermore, social networks are important to

knowledge, as they have been shown to have an

influence on its creation, diffusion, absorption, and

use (Phelps et al., 2012). Phelps et al. (2012) define a

knowledge network or hub as “a set of nodes—

individuals or higher-level collectives that serve as

heterogeneously distributed repositories of

ICSOFT 2019 - 14th International Conference on Software Technologies

586

knowledge and agents that search for, transmit, and

create knowledge—interconnected by social

relationships that enable and constrain nodes’ efforts

to acquire, transfer, and create knowledge”.

Knowledge hubs consist of nodes that represent

actors within the network. These nodes represent

individuals or collectives such as universities,

research institutes, companies, and project teams. The

knowledge hubs can become a source of a

competitive advantage according to the knowledge-

based view. This is confirmed by Vauterin et al.

(2013), who mention that in highly competitive

markets access to, and retainment of people and ideas

is supported by strong social network relationships.

Making these knowledge hubs a source of a

competitive advantage and a structure to be exploited

in recruitment and selection (Phelps et al. 2012).

2.2 Model for Data-driven Competence

Analytics

Strategic workforce planning (SWP) is used by firms

to “identify and address current and future challenges

to gaining the right talent, in place at the right time”

(Phillips and Gully, 2009). Acquiring and retaining

talent is a necessity for successfully executing a

firm’s business strategy, and a strategic and proactive

approach helps a firm to get the right talent. HR

analytics can be used in SWP to assure accurate and

real-time information in the planning process (Momin

and Mishra, 2015). This gives firms an overhand on

their competitors, which is necessary in highly

competitive talent markets. The Talent Supply Chain

Management (TSCM) model from Makarius and

Srinivasan (2017) aims to align talent supply with

talent demand by creating a relationship between the

firms and talent suppliers such as universities, a topic

which is missed in the SWP model. In order to tackle

the problem of misalignment between knowledge

supply and demand we adapt the SWP model for data-

driven strategic competence analytics. In this adapted

SWP* model, we fully integrate the strategic

knowledge sourcing process derived from Makarius

and Srinivasan (2017). Figure 2 in the Appendix

provides an overview of the SWP* model.

In the SWP* model, a step related to the

development of strategic competence data analytics

and knowledge sourcing is added after Step 3 ‘Gap

identification’. Step 4 ‘Strategic competence

analytics’ is composed out of sub steps 4a-4c:

 Development of a sourcing strategy (step 4a);

o Analyze internal HR data to determine the

skills expected of future employees.

o Analyze internal and external data to identify

knowledge hubs using social network

analytics.

 Development of hub selection criteria (step 4b);

o Determine a multi-criteria hub selection

approach. For example, by using criteria like

inter-hub competition for talent, long-term

relationship potential, cost of creating tie, etc.

 Knowledge hub evaluation (step 4c).

o Make a selection from the total number of

knowledge hubs by using the most important

criteria.

In summary, the 6-step SWP* model is an extended

version of the 5-step SWP model Phillips and Gully

(2009, with a dedicated step for strategic competence

analytics after step 3 based on the TSCM model

Makarius and Srinivasan (2017).

2.3 Social Network Analysis

To facilitate the identification of knowledge hubs in

the sourcing strategy step (4a), social network

analysis is integrated. This analysis comprises:

 Data Collection and Preparation: For the

formation of knowledge networks, three classes of

data sources are important:

o Internal data from HR information systems:

e.g. employee IS, performance management

IS, recruitment IS, etc.

o Professional Social Networks: e.g. LinkedIn

(2019), Xing (2019), etc.

o Scientometric and Bibliometric Databases:

e.g. Web of Science (2019), Scopus (2019),

Patstat (2019), Research Gate (2019), etc.

These data sources are integrated in the data

preparation phase, on the level of individuals and

organizations, and fed into a dedicated data mart

for HR and competence analytics.

 Knowledge Hub Visualization: With the HR data

mart we connect software for competence

analytics and knowledge hub visualization. Two

types of software are relevant: 1) software for

performance dashboarding (e.g. MS Power BI

(2019)) and 2) software for network visualizations

and analysis (e.g. Gephi (2019)). With 1) HR

professionals create interactive visualizations and

overall competence measures and geographic HR

data, and with 2) knowledge hubs are visualized

and analyzed.

 Knowledge Hub Analysis: In this analysis, we first

determine community structures within the

complete network, indicating structures of

competence, institutes or individuals, with the

algorithm described in (Blondel et al., 2008). This

Knowledge Hubs in Competence Analytics: With a Case Study in Recruitment and Selection

587

algorithm reports the network modularity, where

a high score represents a network with dense

connections. Network nodes have a certain

position within a knowledge hub (Phelps et al.,

2012). Therefore, another prospective way of

creating ties between hubs is the targeting of such

specific persons in the hub that have a central

position. These people can be used as an

influential link to other players within their hub,

thereby creating a connection from hub to hub.

Central individuals are recognized by using

typical centrality measures as: closeness,

betweenness, and degree centrality (Sumith et al.,

2017). Ties between actors are viewed as a means

through which information and know-ledge are

transferred. Direct ties are perceived to be more

efficient for sharing relevant and complex

information, compared to indirect ties (Singh,

2005). An individuals’ central position within a

network allows them to have more up-to-date

access to rich and diverse information, which

increases their ability to gain knowledge from

their network (Phelps et al., 2012). Therefore,

individuals or institutes with a higher value for the

centrality measures, compared to others, are

perceived to be more knowledgeable actors in a

knowledge network. Especially degree centrality,

which measures the number of direct ties, is

important in determining knowledgeable nodes in

a network (Zhang and Ma, 2016). Besides, the

centrality measures are used to cross-validate an

actors’ importance. Network eccentricity is used

to determine the distance a certain node will have

to travel to the node furthest away from it. This

measure is used to benchmark how far the

company is from other nodes.

3 CASE STUDY ON STRATEGIC

COMPETENCE ANALYTICS

3.1 ABC-company’s Supply and

Demand Profile

This case study focuses on one of the technical

competencies identified within the ABC-company, a

large technical multinational based in The

Netherlands. The name of the company is

anonymized here because of rules related to non-

disclosure. The ABC-company delivers state-of-the-

art technology in its field. ABC has a business

strategy that aims to achieve technological leadership

with an operational excellence and customer focus.

Innovation in all business segments is very important.

The company’s talent strategy starts by identifying its

core competencies. This was done through strategic

competence sourcing. A competency is identified as

a critical core competency when:

1. There is long-term demand for the competence.

2. There is high demand for the competence.

3. There is low availability of the competence in the

external labour market.

The criticality of these competencies has been

determined by looking at projected growth of

demand, and the time needed to hire a person within

the specific competency. This has led to the

identification of core competencies. This case study

focuses on one of these: competence X. Competence

X is the competency in the ABC-company for which

demand is expected to grow the most of all

competencies for the years to come, and the talent for

competence X is needed in several of its branches.

Competence X is an area of knowledge for which

the ABC-company’s demand in the long-term is high,

but the supply from the labour market is low.

Analysis shows that a several hundreds of additional

functions related to competence X are expected to

open this year. The current de-mand for competence

X in the global market is around 6,000 in 7 significant

countries identified by (Gartner, 2019): The

Netherlands, South-Korea, Japan, China, India,

Germany, and Israel. In terms of demand, the

company has a need for potential employees with a

master’s degree or a PhD. Talent supply in the same

nine countries as identified before by Gartner (2019)

offer a supply of around 35,000 competence X

talents. Currently the gap between demand and

supply is mainly experienced for master-level

competence X employees. However, it is predicted

that the overall gap for both master and PhD students

will grow in the future. Therefore, it is important to

start recruitment efforts in knowledge hubs.

3.2 Data Collection and Preparation

In order to map the ABC-company’s internal

knowledge network for competence X, a data set is

created, described in Table 1. The dataset covers 124

employees, 136 of their co-authors, and their 127

knowledge products written in the last 20 years.

Four sources of data are identified, listed in order

of importance:

1. Internal data obtained from the ABC-company’s

records. This data has been anonymized by

replacing employee numbers by identification

numbers. Age has been grouped within intervals.

ICSOFT 2019 - 14th International Conference on Software Technologies

588

Table 1: Data entities for the internal knowledge hub.

Data Entity

Employee

data

Knowledge

Products

Co-Author

data

Data Source

ABC company

Nationality

Employee

Number

Country of

Employment

Age Group

Gender

LinkedIn and Xing

Education

(Institute,

Study, Level,

Location)

(1969- 2018)

Working

Experience

(Institute,

Length,

Function,

Location)

(1988-2018)

LinkedIn, Research Gate, Xing

Skills

Product Type

(Journal Article,

Patent, Report, or

Conference)

Name

Title

Current

Employer(s)

Source

Country of

Employment

Date of

Publication

(1999 to 2018)

Skills

2. LinkedIn. LinkedIn is an online social network

used for professional networking, including job

and CV posting. Data has been obtained through

the recruitment tool offered by LinkedIn, namely

LinkedIn Recruiter. With over 530 million

members worldwide, it offers a wide base of

labour market data (LinkedIn, 2019).

3. Research Gate. Research Gate is a social

networking platform where scientists and

researchers share papers, ask and answer

questions, and find collaborators. It has over 15

million verified scientist members (ResearchGate,

2019).

4. Xing. Xing is a social networking site targeted at

professionals in German speaking countries

Germany, Austria, and Switzerland. In these

countries, Xing has over 13 million users (Xing,

2019).

Two different types of individuals were identified,

each with their own sourcing method:

1. Employees were identified through the

competence X employee list obtained from

internal company records.

2. Co-Authors were identified by gathering the

names of co-authors or co-owners belonging to

articles and patents written or owned by the ABC-

company’s personnel.

Knowledge products like journal articles, patents,

reports and conference publications serve as a link

between these two types of individuals; together

creating ABC’s internal knowledge network. The

obtained data was then accumulated in the HR data

mart and imported into the analytics software to

create visualizations.

3.3 Sourcing Strategy

The sourcing strategy consists of the following

components:

 the analysis of internal HR data to determine the

skills expected of future employees.

 the analysis of both internal and external data to

identify knowledge hubs.

Map visualizations are used to answer questions Q1

and Q2:

Q1. “Where are the competence X’s knowledge hubs

related to the ABC-company located?”, and

Q2. “Where are the external competence X’s

knowledge hubs located worldwide?”.

Question 1 is answered by comparing the

geographical maps produced by the analytics

software. These show the location of employees’

current and past employers, location of current

employers of the co-authors, and location of the

employees’ past educational institutes, respectively.

Looking at the map, we can see the areas around the

ABC-company locations are most heavily covered by

related knowledge players in their hub. Especially

Eindhoven, in the overall Europe area and the areas

around San Diego and San Jose in America.

The co-author employer locations, depicted in

Figure 1, show an interesting overview, suggesting

additional locations such as: Israel, United-Kingdom,

India, Turkey, China, Russia, Germany and Greece as

possibly interesting knowledge locations. The

Netherlands and East- and West-Coast of the United

States remain large suppliers of knowledge players.

Question 2 is answered in the same manner. Here

maps are used to show the location of external

competence X talents’ current and past employers,

and the location of the external’s past educational

institutes, respectively.

Knowledge Hubs in Competence Analytics: With a Case Study in Recruitment and Selection

589

Figure 1: This map shows the locations of the current and

past employers of the ABC employees’ co-authors. The

color of the labels indicates the country it is part of, the size

of the labels indicates the number of co-authors that have

worked for that employer.

3.4 Hub Selection Criteria

In order to select the hubs most relevant to ABC, a

multi-criteria hub selection approach has been

formulated based on the requirements of ABC. The

requirements are the result of structured interviews

with HR professionals. The knowledge hubs of

interest are selected based on the following selection

criteria, listed in order of importance to ABC:

1. The occurrence of skills requested. These skills

ensure that the individuals in the hub, when

selected or recruited, fulfil the different aspects of

their job.

2. The number of central actors, based on the

centrality measures: degree, betweenness, and

closeness. The centrality of actors within a hub

gives an indication of their knowledge base.

Furthermore, the more connections an actor has,

the bigger its impact will be when it is added to

the internal network of ABC.

3. The size of the hub. The size of the hub is based

on the number of actors that are part of the hub.

The larger the number of actors, the more

interesting the hub will be. Any recruitment

activities will be able to yield a larger number of

possible employees.

4. The competition for individuals in the hub. This

selection criterion is applied on the hubs that

remain after the first three criteria. Gartner’s

TalentNeuron (2019), a platform used by ABC

providing talent analytics insights, is used to gain

information on the competition for competence X

talent in the geographical areas of the selected

hubs. When competition is high, it will be more

difficult for ABC to recruit from these hubs.

Other possible selection criteria are based on the

proximity to current locations of ABC, and the

existence of connections to ABC:

5. The proximity to actual locations of ABC

branches is currently not an important selection

criterion to ABC. ABC has specified that it is

willing to open branches in new locations, if the

size of the talent supply there gives them reason

to.

6. The existence of direct connections to ABC is

currently not an important selection criterion to

ABC. Already having a connection to important

hubs will make it easier to approach them for

recruitment. However, it is more important to get

insight in all hubs that could be of interest to ABC,

not just the ones where ABC already has some

kind of relationship with.

3.5 Internal Network Analysis

Figure 3 in the Appendix depicts the internal network

for competence X. The centrality measures allow us

to identify the most ‘knowledgeable actors’ in its

network. Table 2 gives an overview of the ten most

important institutions and their centrality measures,

ordered by degree, and Table 3 gives an overview of

the ten most important individuals and their centrality

measures, again ordered by degree.

Table 2: Top-10 centrality measures for internal network

institutions.

Institute

Centrality measure

Degree

Betweennes

Closeness

Eindhoven Univ. of

Technology

20 0.011904 0.325911

University of Arizona 16 0.007337 0.310212

Tech. company XYZ 11 0.002154 0.312621

University of Twente 10 0.001427 0.304348

Delft University of

Technology

9 0.001763 0.312621

Utrecht University 7 0.003895 0.315068

TMC 7 0.001033 0.306667

Imperial College

London

6 0.005639 0.346237

University of

Groningen

6 0.000558 0.297048

Hellas (FORTH) 6 0.000168 0.26094

ICSOFT 2019 - 14th International Conference on Software Technologies

590

Table 3: Top-10 centrality measures for individual actors.

Individual

Centrality measure

Degree

Betweennes

Closeness

Employee 4051 41 0.413872 0.509494

Co-Author 34 19 0.003991 0.350763

Employee 4111 16 0.015800 0.431635

Employee 4117 15 0.012473 0.428191

Employee 4112 12 0.011973 0.427056

Co-Author 44 12 0.001183 0.346237

Co-Author 38 12 0.000896 0.346237

Employee 4115 10 0.005898 0.423684

Co-Author 40 10 0.002686 0.345494

Co-Author 42 10 0.000632 0.345494

The ten most important individuals have been

selected based firstly on their degree centrality, as

direct relationships have been assumed to contribute

most to an actor’s knowledge (Phelps et al., 2012).

The betweenness and closeness centrality measures

have been normalized, which makes it easier to see

the difference between actors.

Inspecting Table 2, we see institutions that are

important according to all centrality measures, the

top-3 being: Eindhoven University of Technology,

University of Arizona, and Tech. company XYZ.

Looking at the most important individuals in the

network depicted in Table 3, we see Employee 4051

and 4111 in the top-3 based on all centrality

measures, and Co-Author 34 based on degree

centrality. These institutions and individual actors

play an important role in ABC’s knowledge network,

and since these are all part of ABC’s internal network,

the relationships of this network are used in

approaching the institutions and individuals

connected to them in the external network.

The identified institutions can be associated with

via partnerships and the universities might be

partnered with to secure future supply of talent. The

companies are important to ABC for the purpose of

knowledge sharing. All institutions are candidates for

recognizing suitable employees in the selection

process, when they occur in an applicant’s curriculum

vitae. In the future, the data on the identified

individuals of a network is used to create an average

persona to which future candidates can be compared

in the selection process. Furthermore, the recognition

of central individuals in the network of a strategic

competence makes it easier to recognize whom to

turn to when questions regarding this competence

arise. Since degree centrality is a measure of direct

connections, and with that an indicator of knowledge,

it suggests that, although Co-Author 34 might not

hold much knowledge of competence X on its own,

this person is able to refer to the right people and can

create a team of knowledgeable actors.

3.6 Knowledge Hub Evaluation &

Recommendations

Using the concept of network modularity, 38 hubs or

sub-networks are detected in the network for

competence X. The modularity score of the

competence X network is high with 0.846, indicating

strong community structure within the network. In

order to answer the question: “Which of these

knowledge hubs are relevant to ABC?” we apply the

selection criteria identified in coordination with ABC

to the 38 identified hubs. Selection criterion one, the

occurrence of demanded skills, has been met through

the data collection method. The externals’ skills are,

like the employees and co-author’s skills, directly

related to the demanded skills. The internal ABC

network has been confirmed to contain skills that are

related to the skills demanded by ABC. The external

network consists of individuals that have at least ten

of these skills. Therefore, all 38 hubs satisfy the first

selection criterion.

Selection criterion two, the number of central

actors, evaluates the 38 remaining hubs according to

their number of central actors. The actors are deemed

‘central’ when at least two centrality measures

identify them to be among the ten most central

individuals or institutes. In the competence X

network analysis, for example, the ‘Eindhoven

University of Technology’-hub is evaluated as a

central institute and Employee 4051 is evaluated as a

central individual. The remaining hubs are then

evaluated according to their size expressed in number

of actors. This analysis indicates, for example, that

the ‘Eindhoven University of Technology’-hub and

the ‘Indian Institute of Technology’-hub are

relatively large in terms of size. Lastly, we determine

competition for talent in the locations of the central

actors of these five hubs. We use TalentNeuron

(Gartner, 2019) to determine the demand for

competence X personnel in the locations of these

hubs. After this analysis several hubs are excluded

because the number of direct employers competing is

very high. This last selection criterion leaves, for

example, Eindhoven and Kanpur as most desirable

hubs after applying the SWP* model.

In order to answer the question: “How close are

these knowledge hubs to ABC?”, we define the metric

Knowledge Hubs in Competence Analytics: With a Case Study in Recruitment and Selection

591

weighted number of ties (WNT). This metric counts

the number of ties from ABC with the resulting hubs

from the hub selection process. The WNT is defined

as the weighted total of: 1) the # of ABC employees

belonging to that hub (w

=1), 2) the # of institutes

ABC has a partnership with (w

=1), and 3) the co-

authors of ABC employees (w

=0.5). The co-authors

have a smaller weight, since these actors are often not

directly connected to ABC. The WNT determines the

approach of action that has to be taken towards the

hub. A low WNT indicates that the first step should

be creating partnerships and creating awareness of

ABC within the hub. A high WNT indicates a

relationship exists already, therefore these can be

used to source knowledge talent from these hubs. In

this case the ‘Eindhoven University of Technology’-

hub has a high WNT close to 50. Next, we connect

the hubs with a high WNT to the identified central

actors. For these hubs and actors, the recruitment

actions are, for instance: to build a ‘persona’ based on

data on the individuals within the hub that recruitment

will be targeted towards, and to build

partnerships/relationships with central institutions

and individuals. And as a final step, it is important to

continuously monitor the strategic competence

network, and evaluate the selection and recruitment

activities’ positive and negative effects per hub. By

doing this, up-to-date network information is ensured,

and the recruitment and selection activities are fine-

tuned to the specific hubs.

4 CONCLUSIONS

4.1 Theoretical Contribution

By providing a case study of the usage of analytics in

knowledge hubs related to the recruitment and

selection process, we provide two main theoretical

contributions. First, we provide much needed

empirical evidence on how the right usage of HR

analytics can have beneficial effects on the

organization business (Marler and Boudreau, 2017)

and especially in the recruitment and selection

domain of the HR function. We provide an example

on how HR decision making benefits by triangulating

internal and external data (Brown and Vaughn, 2011),

and show how beneficial it is for the HR function to

overcome its internal focus and look for useful data

even outside the function itself (Rasmussen and

Ulrich, 2015). The proposed model SWP* contributes

to the literature by providing an example on how the

use of strategic competence analytics and overall HR

analytics, helps the organization recognize and utilize

competencies in and around the firm. For example,

the results and proposed model show, how recruiters

identify important talents or important institutions in

their knowledge network, which in the end provide

the organization with unique human capital – as they

can focus to recruit specific individuals with unique

SKA’s (Lepak and Snell, 1999). Second, we also

contribute to the general IS literature by further

developing a model for data-driven competence

analytics, inspired by the work of Phillips and Gully

(2009) and Makarius and Srinivasan (2017), in

producing and validating dedicated analytics for HR.

In the model, knowledge hubs are identified

algorithmically as community structures and

subsequently analysed with centrality measures.

Hence, concepts from network analysis are applied in

the context of strategic competence analytics. In

addition, the presented method builds also on the

visualization strength of network analysis, which

provides a simple and effective overview of the most

important actors in the knowledge hub. This approach

enables recruiters to identify hubs that fulfil their

criteria of interest, recognize the important players in

these hubs, identify their connections to these hubs of

interest, and adapt their recruitment strategy

accordingly.

4.2 Practical Implication

Instead of relying purely on social capital and

intuition to find the right talents, recruiters, can use

the model provided in this paper. The model provides

an easy to understand network visualization, which

potentially points out individuals and institutions that

have a strong connection with the company. Such

networks provide data-based insights into central

actors within the company network, which can be

targeted for recruitment.

4.3 Limitations and Future Work

A limitation of this paper is the fact that only a

confined number of data sources has been used as

sources for the network data. For example, LinkedIn

has high coverage of the labour market of the USA,

but low coverage of the East-Asian and African

labour market (LinkedIn, 2019). Therefore, the

representation of population by the data is not entirely

accurate. Future research should include multiple

sources of data to obtain more reliable and global

results, such as Facebook (Brown and Vaughn, 2011).

In addition, for future research we want to:

 Develop appropriate models for data cleaning,

ICSOFT 2019 - 14th International Conference on Software Technologies

592

to better facilitate the integration of HR data

sources and align with legal issues.

 Finetune the set of relevant metrics for

competence analytics, and

 Validate and develop the presented model at more

companies and organizations.

ACKNOWLEDGEMENTS

The authors would like to thank Demy van Hassel for

her work on this topic and the ABC-company for their

contribution.

REFERENCES

Angrave, D., Charlwood, A., Kirkpatrick, I., Lawrence, M.,

and Stuart, M. (2016). HR and analytics: Why HR is set

to fail the big data challenge. Human Resource

Management Journal, 26(1), 1-11.

Barney, J. (1991). Firm resources and sustained competitive

advantage. Journal of management, 17(1), 99-120.

Batistič, S., and Van der Laken, P. (2018). The history,

evolution, and future of big data & analytics: a

bibliometric analysis of its relationship to performance

in organizations. British Journal of Management. In

press.

Blondel, V., Guillaume, J., Lambiotte, R., and Lefebvre, E.

(2008). Fast Unfolding of Communities in Large

Networks. Journal of Statistical Mechanics Theory and

Experiment. 10.1088/1742-5468/2008/10/P10008.

Brown, V. R., and Vaughn, E. D. (2011). The Writing on

the (Facebook) Wall: The Use of Social Networking

Sites in Hiring Decisions. Journal of Business and

Psychology, 26(2), 219-225.

Evers, H. (2008). Knowledge hubs and knowledge clusters:

Designing a knowledge architecture for development.

ZEF Working paper Series, 1-21.

Gartner. (2019). Talent Neuron. https://www.gartner.

com/en/human-resources/talentneuron

Gephi. (2019). https://gephi.org.

Knight, J. (2013). Education hubs: international, regional

and local dimensions of scale and scope, Comparative

Education, 49:3, 374-387.

Lepak, D. P., and Snell, S. A. (1999). The human resource

architecture: Toward a theory of hu-man capital

allocation and development. Academy of management

review, 24(1), 31-48.

LinkedIn. (2019). https://www.linkedin.com.

Makarius, E., and Srinivasan, M. (2017). Addressing skills

mismatch: Utilizing talent supply chain managment to

enhance collaboration between companies and talent

suppliers. Business Horizons, 60(4), 495-505.

Marler, J., and Boudreau, J. (2017). An evidence-based

review of HR analytics. The International Journal of

Human Resource Management, 28(1), 3-26.

Microsoft Power BI. (2019). https://powerbi.

microsoft.com/en-us.

Momin, W., and Mishra, K. (2015). HR Analytics as a

Strategic Workforce Planning. International Journal of

Applied Research, 1(4), 258-260.

Noe, R., Hollenbeck, J., Wright, P., and Gerhart, B. (2014).

Human Resource Management: Gaining a Competitive

Advantage (9th Global Edition ed.). Irwin: Mcgraw-

Hill Education.

Patstat (2019). https://www.epo.org/.

Phelps, C., Heidl, R., and Wadhwa, A. (2012). Knowledge,

Networks, and Knowledge Networks: A Review and

Research Agenda. Journal of Management, 1115-1166.

Phillips, J., and Gully, S. (2009). Staffing forecasting and

planning. Alexandria: Society for Human Resource

Management.

Rasmussen, T., and Ulrich, D. (2015). Learning from

practice: how HR analytics avoids being a management

fad. Organizational Dynamics, 44(3), 236-242.

ResearchGate. (2018). Recruit Scientists and Researchers.

Retrieved from ResearchGate.com: https://

solutions.researchgate.net/recruiting/.

Scopus. (2019). Elsevier, https://www.scopus.com.

Singh, J., (2005). Social Networks as Determinants of

Knowledge Diffusion Patterns. Management Science.

51. 756-770. 10.1287/mnsc.1040.0349.

Sumith, N., Annappa, B., and Bhattacharya, S. (2017).

Social Network pruning for building optimal social

network: A user perspective. Knowledge-Based

Systems, 101-110.

Vauterin, J., Michelsen, K., and Linnanen, L. (2013).

Attracting and retaining student talent from around the

world: The lived experience in university-industry

collaboration. Industry and Higher Education, 27(3),

163-178.

Van den Heuvel, S., and Bondarouk, T. (2017). The rise

(and fall?) of HR analytics: a study into the future

application, value, structure, and system support.

Journal of Organizational Effectiveness: People and

Performance, 4(2), 157-178.

Web of science. (2019). Thomson Reuters, http://

apps.webofknowledge.com.

Xing. (2019). Facts and Figures. Retrieved from Xing.com:

https://corporate.xing.com/en/about-xing/facts-and-

figures/.

Zhang, Y., and Ma, D. (2016). Modeling the evolution of

collaboration network and knowledge network and

their effects on knowledge flow through social network

analysis. Journal of Digital Information Management,

14(4), 246-254.

Knowledge Hubs in Competence Analytics: With a Case Study in Recruitment and Selection

593

APPENDIX

Business Strategy

Talent Strategy

Develop Action

Plan(s) to Address

Forecasted Talent

Gaps

Identify Gaps

(Projected Labor

Surpluses or

Shortages)

Monitor, Evaluate,

and Revise Forecasts

and Action Plans

Forecasted Demand for

Labor

Forecasted Supply of Labor

Strategic

competence

analytics

Step 1 Step 2 Step 3 Step 4 Step 5 Step 6

(a)

Sourcing

strategy

(b)

Hub selection

criteria

(c)

Hub

evaluation

Figure 2: Model for data-driven strategic competence analytics (SWP*). Adapted from Phillips and Gully (2009).

Figure 3: This knowledge network shows the internal network of ABC. The size of the nodes is determined by degree

centrality, and its color by the modularity class. Nodes with a low degree centrality are filtered out to improve visibility of

important actors.

ICSOFT 2019 - 14th International Conference on Software Technologies

594