See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/308608900

Human Resource Management in the Age of Big Data

Technical Report · September 2016

DOI: 10.13140/RG.2.2.13683.94246

CITATION READS

1 2,087

1 author:

Benjamin Erb
Ulm University
32 PUBLICATIONS   74 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

chronograph – A Distributed Platform for Event-sourced Graph Computing View project

Chatbots in Clinical Psychology and Psychotherapy View project

All content following this page was uploaded by Benjamin Erb on 25 September 2016.

The user has requested enhancement of the downloaded file.

Human Resource Management in the Age of Big Data

Benjamin Erb

Seminar Applied Management Diagnostics

Summer Term 2016

Dept. Individual Differences and Psychological Assessment

Institute of Psychology and Education
Faculty of Engineering, Computer Science and Psychology
Ulm University
Human Resource Management in the Age of Big Data

Author Benjamin Erb

Course Psychology (Bachelor)
Contact benjamin.erb@uni-ulm.de

Lecture Applied Management Diagnostics

Term Summer term 2016
Course ID PSY71104.002
Department Individual Differences and Psychological Assessment
Institute of Psychology and Education
Faculty of Engineering, Computer Science and Psychology
Ulm University

Version September 25, 2016

Typesetting LATEX, apa6 Package

This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this license,

visit http://creativecommons.org/licenses/by/4.0/ or send a letter to Creative Commons, PO Box 1866, Mountain

View, CA 94042, USA.

Title Image

Skygarden by Negative Space, released under Public Domain (CC0 license).

http://negativespace.co/photos/skygarden/
HUMAN RESOURCES & RECRUITING IN THE AGE OF BIG DATA 3

Abstract

Big data and data analytics enable novel kinds of data-driven decision making
and management strategies in human resources. Approaches such as people analytics
have already been implemented by corporations and replace traditional procedures for
recruiting, hiring, and staff development. Due to the disruptive force of automation and
the momentum of artificial intelligence in other fields, we argue that human resource
management will increasingly be transformed and reshaped by these developments.
There are arguments that heavily data-driven approaches are beneficial for both the
workforce and companies and that they are even necessary for the work environment of
the future. On the other hand, there are also cautions against such rule of algorithms
due to privacy implications and ethical issues.
HUMAN RESOURCES & RECRUITING IN THE AGE OF BIG DATA 4

Contents

Introduction 5

Background 5
Big Data, Data Mining & Machine Learning . . . . . . . . . . . . . . . . . 5
People Analytics: Data Analytics in Human Resources . . . . . . . . . . . 6

Opportunities and Challenges for Human Resource Management 7

The Benefits of Data-Driven Human Resources . . . . . . . . . . . . . . . 7
Machine Bias and the Threat of Algocracy . . . . . . . . . . . . . . . . . . 8
Issues in Big Data for Human Resource Management . . . . . . . . . . . . 9

Conclusion 10
HUMAN RESOURCES & RECRUITING IN THE AGE OF BIG DATA 5

Introduction

Big data is considered to be an inherent part of the fourth industrial revolution.

Cloud computing, cyber-physical systems, advances in artificial intelligence, and
automation are not only reshaping everyday life, they are also going to heavily disrupt
human work life. The availability of massive amounts of data and the ability to derive
and gain reasonable insights thereof also affects human resource (HR) management.
In this work, we briefly introduce the notion of big data and accompanying data
processing approaches. Next, we show how these approaches are already used in
human resource management. Furthermore, we will discuss the opportunities and
challenges of data-driven human resource management, including a brief outlook on
open challenges and a critical perspective towards ethical aspects and privacy issues.
We conclude our work with a glimpse at the future of human resource management,
taking into account automation and artificial intelligence.

Background

Big Data, Data Mining & Machine Learning

The term big data describes a shift of paradigms for data processing that emerged
when storage capacities had been exceeded to an extent where traditional mechanisms
became inappropriate (McAfee, Brynjolfsson, Davenport, Patil, & Barton, 2012). Big
data is often defined using multiple dimensions of data (Laney, 2001): volume (i.e.,
the quantity of data), variety (i.e., the structure of data), and velocity (i.e., growth
of data). Recently, the three dimensions have been complemented with additional
properties which represent challenges for (big) data processing: variability (i.e., varying
consistency of various data) and veracity (i.e., varying quality of various data).
Big data has been enabled mainly by a number of advances in computer science
(e.g., distributed computing and data mining), information technology (e.g., cloud
computing and virtualization), and data storage technology (e.g., inexpensive mass
storage devices). For instance, Google’s MapReduce paradigm (Dean & Ghemawat,
2008) provides a scalable approach for computations on Petabytes of data.
HUMAN RESOURCES & RECRUITING IN THE AGE OF BIG DATA 6

Apart from the pure scale of data, big data has been accompanied by a number of
advances in data analytics, also known as knowledge discovery, data mining, and to
some extent, artificial intelligence. The application of statistical methods to large-scale
data sets aims for the extraction of information and the identification of insights (Chen,
Chiang, & Storey, 2012). This includes a lot of different techniques and methods (Han,
Kamber, & Pei, 2011; Leskovec, Rajaraman, & Ullman, 2014), such as text mining
(analyzing textual data; e.g., sentiment analysis of user comments), clustering (grouping
of similar items; e.g., detecting customer groups), classification (categorizing novel
data into known structures; e.g., spam detection), real-time analytics (low-latency
processing; e.g., high-frequency trading), or anomaly detection (the identification of
outliers in data patterns; e.g., credit card fraud detection).
Data mining techniques are also used to create, feed and tune predictive models.
The case of a Target customer who has been identified as being pregnant due to
slight changes in her buying behavior (Hill, 2012) is a textbook example for predictive
analyses and major implications thereof. Machine learning represents another approach
for devising models in data analysis by feeding data as input and generating predictive
models like artificial neural networks.

People Analytics: Data Analytics in Human Resources

The thorough success of big data and data analytics in many business sectors also
reached HR since the early 2010s, mainly fueled by the availability of widely collected
HR data as well as strong expectations regarding valuable predictions for HR. Data
to be considered for such analyses range from detailed candidate profiles and various
individual performance indicators to complex behavioral data like mail communication
and interaction patterns (Peck, 2013). Especially the fusion of different data sets,
e.g., the identification of many-sided, multifaceted prediction variable sets, provides
extremely relevant insights and differs from simpler, traditional regression analyses.
So-called people analytics covers data analytics and predictive analysis of the entire
workforce and in all stages of employment (Waber, 2013). This includes predictions in
terms of skill analysis, selection and hiring of employees, talent identification, attrition
HUMAN RESOURCES & RECRUITING IN THE AGE OF BIG DATA 7

probabilities, or workplace health management.

For instance, Davenport, Harris, and Shapiro (2010) list six major use cases
for talent analytics: (i) human-capital facts, (ii) analytical HR, (iii) human-capital
investment analysis, (iv) workforce forecast, (v) talent value model, and (vi) talent
supply chain.
Petersen (2015) lists more general use cases for people analytics: (i) the early
detection of quitting intentions of employees, (ii) the projection of future leadership
ranks, (iii) stress level analysis for workplace health management, (iv) detailed KPI
metrics of individual employees, (v) network analyses of internal communication,
and (vi) the creation of personas for modeling successful employees based on actual
employee data.
People analytics has originally found its way into HR for recruiting, and was initially
adopted by highly innovative tech companies and dedicated start-ups that identified
people analytics as a substantive business model (Peck, 2013). Google represents the
poster child of a tech company adopting people analytics. Google does not only rely
on data mining mechanisms to fine-tune their hiring procedures (Dahl, 2016), they
also crunch data to identify common traits and habits of successful employees and
managers (internal project Oxygen, see Bryant, 2011; Garvin, 2013; Sullivan, 2013).

Opportunities and Challenges for Human Resource Management

The Benefits of Data-Driven Human Resources

The followers of people analytics argue that data-driven management of HR

provides advantages for all parties involved (Davenport et al., 2010; Peck, 2013;
Waber, 2013). Companies and their HR departments benefit from more reasoned HR
decisions. Promising candidates can be identified, recruited, and hired based on the
knowledge derived from large amounts of data. Insights from internal HR data can
be used to improve management (de Romree, Fecheyr-Lippens, & Schaninger, 2016).
Smallest trends in the workforce can be quickly recognized and proactively reacted
to. This is especially interesting for the prevention of attrition and the prediction of
HUMAN RESOURCES & RECRUITING IN THE AGE OF BIG DATA 8

retention problems (Sullivan, 2013). Similar to digital recruiting and hiring processes,
solutions for people analytics require higher initial expenditures, but provide scalable
mechanisms and cost-effective solutions in the long run (de Romree et al., 2016).
In a data-driven HR economy, employees receive a more accurate placing as well
as fair opportunities according to their potential (Peck, 2013). Another common
argument that favors data-driven HR over traditional HR is the neutrality of data and
algorithms, compared to humans. Even the most experienced HR managers cannot
entirely set aside their exposure to a wide range of human biases. When selection and
hiring algorithms are configured to focus solely on the key requirements of applicants,
the algorithms yield teams more diverse and functioning than human recruiters would
choose (de Romree et al., 2016; Peck, 2013). People analytics is also a valuable tool
for supporting change management (Waber, 2013). Furthermore, predictive analytics
known from the medical sector also can be used for the early detection and prevention
of work-related conditions, such as occupational burnouts (Waber, 2013).

Machine Bias and the Threat of Algocracy

A smaller countermovement opposes the idea of bias-free algorithms and brings the
role of algorithms into question. Their arguments are not specific to people analytics,
but apply to other aspects of big data related to human behavior, predictive analytics,
and computational decision making.
While algorithms are per se free of any biases, their programmers are not. Hidden
assumptions can very well influence the actual behavior of the algorithm compared
to the intended and designated behavior. A more imminent threat is the latent
information hidden in the data used as input. For predictive analytics, algorithms
rely on complex models to project future states. These models are typically built and
generated using data from the past. If the input data is subject to a bias, this bias
inevitably becomes part of the model. Even worse, the predictions will then again be
subject to the underlying bias. A substantial machine bias has been investigated by
Angwin, Larson, Mattu, and Kirchner (2016) for crime prediction, where algorithms
that are already in use today showed a consistent bias against minorities.
HUMAN RESOURCES & RECRUITING IN THE AGE OF BIG DATA 9

A deep neural network is a machine learning approach where a program learns from
vast amounts of training data by building complex artificial neural networks. This
approach has recently achieved unprecedented success in certain problem domains
previously not solvable for computers, e.g., recognizing complex patterns or playing the
board game Go (Mnih et al., 2015). A downside of this approach is the general absence
of a human comprehensible model representing result of the learning process. When
such an algorithm is fed with 100,000 candidate profiles and their respective hiring
outcomes, once trained, the algorithm can then be used to predict the outcome of
new candidates. However, the rationale of the decision — while being 100% reasonable
(based on the training data) — is not necessarily obvious to any human. It is the sole
outcome of a highly connected network of weighted, artificial neurons reacting to a
stimulus (i.e., the profile data). As opposed to other algorithmic approaches, the
results of a machine learning mechanism may be even more prone to latent biases in
the data sets, as the repercussions of biases become less obvious to humans.
Morozov (2013) provides a comprehensive perspective on the implications of
algorithmic decision making and argues that a rule by algorithm, also called algocracy
represents a serious threat to democracy. Especially the usage of algorithms for
policing increasingly affects our everyday lives (Hickman, 2013). Data sovereignty and
the prerogative of interpretation are among the primary issues (Danaher, 2014), but it
is primarily the untransparent, and closed character of decision making processes based
on data mining, that represents the real menace (Morozov, 2013). Recent examples
of potential consequences became apparent in stock market fluctuations caused by
algorithmic trading or the manipulation of public opinion by corporate algorithms
that modify and filter social networking contents.

Issues in Big Data for Human Resource Management

Apart from a dystopian fear of the rule of algorithms, people analytics also
raises several issues regarding privacy, ethics, and lawfulness. Data is the fuel of data
analytics, and in the case of people analytics, such data is always highly personal (Peck,
2013). So there is a thin line between detailed, individual performance metrics collection
HUMAN RESOURCES & RECRUITING IN THE AGE OF BIG DATA 10

and total surveillance of employees (Morozov, 2013). Personal data of employees is

not only subject to a trade-off between individual privacy and decent monitoring by
the employer, but also to legal directives regarding worker rights. Implications of
workplace surveillance on perceived distributive and procedural justice are discussed
by Introna (2000) and Alge (2001).
The idea of algorithmic placing of candidates based on big data also raises ethical
issues. The scoring of candidates based on computational behavioral analytics and
data mining is considered as fairer by their supporters. Others argue that an assess-
ment without a strong human in the loop constrains the chances of an individual
because of their data profile and estimation models of similar individuals (Peck, 2013).
Theoretically, algorithms could be implemented to assign improbable fits with a certain
likelihood. This is still different from a human who develops faith in a candidate during
an interview and gives a person a chance based on intuition. A negative consequence
on talent prediction is the potential devaluation of less promising candidates. Being
permanently characterized as “low potential” based on the personal profile and an
algorithmic comparison with similar profiles may negatively affect motivation and
commitment of human beings as well as their emotional well-being.

Conclusion

People analytics does not only complement traditional psychological assessments

of individuals, it also starts to replace prevalent methods for recruitment, selection,
hiring, performance evaluation, and staff development.
In order to assess the impact of people analytics for the future of HR, it is also
important to embed it into the workplaces of tomorrow. The raise of automation,
robotics, and artificial intelligence is about to heavily disrupt human work in almost all
aspects (Ignatius, 2016; Smith & Anderson, 2014), although some researchers expect a
longer time span for this transition (e.g., Autor, 2015) to happen. According to Chui,
Manyika, and Miremadi (2016), about 45% of the work activities in the United States
are endangered by automation. At the moment, automation is mainly focusing on
physical work, and data collection and processing tasks. Due to advances in artificial
HUMAN RESOURCES & RECRUITING IN THE AGE OF BIG DATA 11

intelligence, other tasks such as unpredictable physical work or the application of

expertise will soon also become subject to automation, at least to a certain extent (Chui
et al., 2016; Smith & Anderson, 2014). In the long run, automation will even affect
mid- to high-skilled professions, such as complex activities of medical doctors.
This transition yields vast economic challenges regarding education, occupational
retraining, reallocation of jobs (Ignatius, 2016; Smith & Anderson, 2014), and eventu-
ally, social challenges regarding the value of human capital and the distribution of
income (Autor, 2015). At least for the former issues, people analytics may become
beneficial. The identification of detailed skills required in jobs that are less prone to
automation and the development of appropriate job training systems are a first step,
followed by specific programs to reallocate retrained workforces.
Ironically, people analytics itself could represent a precursor of HR automation.
If people analytics is implemented and rolled out on large scale in companies, HR
departments could be scaled down to a bare minimum. Remaining HR jobs could
include a chief human resources officer for long-term strategic management, data
scientists that operate people analytics systems, and a very small HR skeleton staff
for remaining tasks that are not fully automatable (e.g., specific interviews).
HUMAN RESOURCES & RECRUITING IN THE AGE OF BIG DATA 12

References

Alge, B. J. (2001). Effects of computer surveillance on perceptions of privacy and

procedural justice. Journal of Applied Psychology, 86 (4), 797.
Angwin, J., Larson, J., Mattu, S., & Kirchner, L. (2016). Machine bias. Retrieved
from https://www.propublica.org/article/machine-bias-risk-assessments-in-
criminal-sentencing
Autor, D. H. (2015, August). Why are there still so many jobs? the history and
future of workplace automation. Journal of Economic Perspectives, 29 (3), 3–30.
doi:10.1257/jep.29.3.3
Bryant, A. (2011, March 12). Google’s quest to build a better boss. The New York
Times. Retrieved from http://www.nytimes.com/2011/03/13/business/13hire.
html
Chen, H., Chiang, R. H., & Storey, V. C. (2012). Business intelligence and analytics:
from big data to big impact. MIS quarterly, 36 (4), 1165–1188.
Chui, M., Manyika, J., & Miremadi, M. (2016). Where machines could replace humans—
and where they can’t (yet). McKinsey Quarterly, 7.
Dahl, M. (2016, June 9). Google expects different answers from men and women
on a question used to predict job success. nymag.com. Retrieved from http:
//nymag.com/scienceofus/2016/06/one-question-google-uses-to-predict-the-
success-of-its-male-and-female-job-candidates.html
Danaher, J. (2014). Rule by Algorithm? Big Data and the Threat of Algocracy.
Retrieved from http://ieet.org/index.php/IEET/more/danaher20140107
Davenport, T. H., Harris, J., & Shapiro, J. (2010). Competing on talent analytics.
Harvard Business Review, 88 (10), 52–58.
de Romree, H., Fecheyr-Lippens, B., & Schaninger, B. (2016). People analytics reveals
three things hr may be getting wrong. McKinsey.
Dean, J. & Ghemawat, S. (2008). Mapreduce: simplified data processing on large
clusters. Communications of the ACM, 51 (1), 107–113.
HUMAN RESOURCES & RECRUITING IN THE AGE OF BIG DATA 13

Garvin, D. A. (2013). How google sold its engineers on management. Harvard Business
Review, 91 (12).
Han, J., Kamber, M., & Pei, J. (2011). Data mining: concepts and techniques. Elsevier.
Hickman, L. (2013, July 1). How algorithms rule the world. The Guardian. Retrieved
from https://www.theguardian.com/science/2013/jul/01/how-algorithms-rule-
world-nsa
Hill, K. (2012). How target figured out a teen girl was pregnant before her father did.
Retrieved from http://www.forbes.com/sites/kashmirhill/2012/02/16/how-
target-figured-out-a-teen-girl-was-pregnant-before-her-father-did/
Ignatius, D. (2016, August 11). The brave new world of robots and lost jobs. The
Washington Post. Retrieved from https://www.washingtonpost.com/opinions/
the-brave-new-world-of-robots-and-lost-jobs
Introna, L. D. (2000, December). Workplace surveillance, privacy and distributive
justice. SIGCAS Comput. Soc. 30 (4), 33–39. doi:10.1145/572260.572267
Laney, D. (2001). 3D data management: controlling data volume, velocity, and variety.
META Group.
Leskovec, J., Rajaraman, A., & Ullman, J. D. (2014). Mining of massive datasets.
Cambridge University Press.
McAfee, A., Brynjolfsson, E., Davenport, T. H., Patil, D., & Barton, D. (2012). Big
data. Harvard Business Review, 90 (10), 61–67.
Mnih, V., Kavukcuoglu, K., Silver, D., Rusu, A. A., Veness, J., Bellemare, M. G., . . .
Ostrovski, G., et al. (2015). Human-level control through deep reinforcement
learning. Nature, 518 (7540), 529–533.
Morozov, E. (2013). The real privacy problem. MIT Technology Review, 116 (6), 32–43.
Retrieved from https://www.technologyreview.com/s/520426/the-real-privacy-
problem/
Peck, D. (2013). They’re watching you at work. The Atlantic, 312 (5), 72–84. Retrieved
from http://www.theatlantic.com/magazine/print/2013/12/theyre-watching-
you-at-work/354681
 HUMAN RESOURCES & RECRUITING IN THE AGE OF BIG DATA 14

Petersen, M. (2015). People Analytics: 6 spannende Anwendungsfaelle fuer daten-

basierte Personalentscheidungen. Retrieved from http://t3n.de/news/people-
analytics-projekte-609331/
Smith, A. & Anderson, J. (2014). AI, robotics, and the future of jobs. Pew Research
Center. Retrieved from http://www.pewinternet.org/2014/08/06/future-of-jobs/
Sullivan, J. (2013). How google is using people analytics to completely reinvent hr.
TLNT: The Business of HR, 26.
Waber, B. (2013). People analytics: how social sensing technology will transform
business and what it tells us about the future of work. FT Press.

View publication stats