Yield management of workforce for IT service providers

https://doi.org/10.1016/j.dss.2011.10.015Get rights and content

Abstract

Many IT service firms often suffer inadequate staffing and a possible subsequent profit loss. This may happen for a number of reasons: the high cost of timely adjusting a firm's workforce capacity, the nature of IT projects, fluctuating market demand, etc. This paper proposes a strategic use of the emerging online market mechanism, known as e-lancing, for yield management of workforce for IT service providers. By dynamically controlling admissions of online and conventional orders, a service provider can reach customers in the online channel and increase profits. The proposed model considers an IT service firm, which receives IT project orders through two channels: a conventional procurement channel and an online spot market such as Elance Online. We employ Markov decision theory to obtain optimal admission control policies. The structures of the optimal policy, which is not of threshold type, are analyzed mathematically and examples are presented numerically. The base model and its extended models, which can serve as useful tools for demand control problems of IT service providers, capture the most important characteristic of IT projects, where if a project is admitted, it seizes a random number of workers simultaneously, then it releases the workers either individually or in a group.

Highlights

► This paper proposes a new approach that helps IT service firms improve workforce utilization. ► Our Markov Decision model dynamically admits orders from traditional and online channels. ► The model captures the most essential characteristics of IT projects. ► Our results include characteristics of optimal policies and simulation results.

Introduction

Improved Internet technology that enables businesses to coordinate jobs and professionals in a global market has introduced new commercial opportunities to e-lancers by providing increased efficiency through reduced transaction costs [23], [24]. This decentralized, individual-oriented electronic market mechanism will become essential for strategic sourcing to design agile organizations by deploying resources quickly and efficiently in response to diverse market changes. Motivated by these factors favoring e-lance markets in the Internet age, this research aims to answer the following research questions: 1) how do we develop an intelligent decision tool to help information technology (IT) service providers utilize their workforce more efficiently by incorporating the e-lance market with the existing conventional channel? and 2) what are the characteristics of the optimal policy for admitting jobs from the two channels?

IT human capital is the most critical resource for an IT service provider, because there is no notion of physical products, production facilities, inventory, or supply chain in their business processes, unlike those for a physical goods manufacturer. Therefore, an efficient management of human capital is one of the main concerns in IT service firms. Specifically, adequate workforce utilization is directly related to the firm's profit and is a crucial task for an IT service provider. IT human resource has been also acknowledged as a strategic and key asset in the information systems (IS) literature [2], [10].

Utilizing workforce at an optimal level is important yet difficult for IT service providers. Many IT service firms often suffer inadequate staffing and subsequent profit loss. IT service firms cannot avoid occasionally retaining excess workforce for a number of reasons: the difficulty of timely adjusting a firm's workforce capacity, the nature of IT projects, demand uncertainty in the market, etc. Labor markets for IT firms behave very differently from standard labor markets [10], [38]. The fluctuating market demand makes it difficult for them to efficiently adjust the headcount to demand. The high demand for IS/IT during the Y2K era and dotcom boom led IT providers to aggressively recruit IT professionals. But after the dotcom boom faded, companies couldn't lay off their excess workers fast enough [13]. The nature of IT projects is another source of difficulty in optimal staffing. IT service providers want to maintain their workforce capacity at a sufficient level so that they can respond quickly to high market demand or a large project order. On the other hand, since each IT project requires a group of professionals, IT providers face the prospect of a number of idle workers when large projects are over, until they receive new projects. As such, a firm may retain idle workers when it faces low demand or the termination of a large project.

Maintaining idle workers, however, incurs significant costs to the firm without generating any revenue; the wages of the idle employees and their training costs. Myopic remedies to deal with this challenge can be ad-hoc staffing-up or laying-off. However, these ad-hoc remedies involve high initial costs or potential legal disputes, disabling the firm's agility.

Furthermore, besides those tangible, direct expenses, underutilization of employees can substantially damage overall organizational performance. Underutilizing employee skills is the worst form of disengagement in an organization. The human resource management literature suggests correlations among employee engagement, work satisfaction, and improved organizational performance and productivity [16], [30]. The improvement in organizational performance is especially enhanced by appropriate utilization of employee's core competency, such as experience and skill [16]. Employees whose skill and experience are underutilized are likely to feel less useful and less involved in the organization. They feel uncertain about their future in the current job and organizations may face losing their competent employees. As such, employee underutilization can potentially impair employee job satisfaction and morale, and overall organizational performance. We believe that IT service providers can gain competitive advantages and enhance productivity and profitability by properly utilizing their most important assets: human capital.

The workforce utilization problem of an IT service provider has the characteristics shared in the hotel and airline industries where yield management has been successfully employed. IT service firms incur high costs for making any adjustments of initial capacities, like in the hotel and airline industries [20]. In other words, it is very expensive for them to hire and train new IT professionals or lay off the existing employees. Moreover, the inventory of IT professionals can be seen as perishable just like hotel rooms and airplane seats, in that the excess idle workforce incurs operational costs, including employees' salaries, without generating any revenue and the availability will disappear unless used now. Observing these similarities between the IT service industry and hotel/airline industries, we present a Markov decision model for yield management by effectively utilizing idle workers through e-lancing, the secondary online channel, when the market demand for the conventional channel is low.

E-lancing was first discussed in [23] in the management literature and defined as a new market mechanism comprised of freelancers (either individuals or organizations) joined in online networks to provide professional services. The most common type of market mechanism for e-lancing, considered as the secondary channel in the model, is online reverse auctions, where the clients post projects, such as software development and website design, as a form of RFP (Request for Proposal), and then IT service firms bid for them. Online auctions enable firms to efficiently outsource small projects that, mostly, involve less than six person-months of effort [41].

Examples of currently operated Web-based IT service markets include Elance Online (www.elance.com), eWork markets (www.eworkmarkets.com), Rent A Coder (www.rentacoder.com) and Guru (www.guru.com). They provide Web-based project marketplaces that connect small- and medium-sized businesses with a global pool of IT service providers. More information about several e-lancing Web sites is given in Table 1. The e-lancing marketplaces for IT services have been particularly successful among other professional services, because the majority of users are usually Web-savvy and familiar with dealing with auctions in online environments. Moreover, unlike other professional services, a delivery of IT services does not require direct physical interactions among service providers and clients.

Given growing competition in the IT service industry, the presented optimal policy for sales channel control will provide insights applicable to the management of idle manpower for IT service providers. Among IT service firms that have been already participating in and generating revenue from a popular online marketplace, Elance (www.elance.com), are NixSolutions (www.nixsolutions.com, estimated annual online revenue $1.4 million, Ukraine), SynapseIndia (www.synapseindia.com, estimated annual online revenue $1.3 million, India), XiCom (www.xicom.biz, estimated annual online revenue $1.1 million, India), and WebsiteDesignZ (www.websitedesignz.com, estimated annual online revenue $220,000, Kansas City, US). Their annual revenues and long-time existence in the online marketplace prove the sustainability of the strategy of incorporating online demand in their revenue model. We believe that more companies can benefit from incorporating the online channel in their daily operations. Those companies who haven't exploited the potential opportunities of online markets, as well as those who already utilizing the online markets, will benefit from implementing our job admission model for such two-channel operations and improve the organizational performance.

In summary, this paper proposes how IT service providers can improve productivity by integrating the e-lancing channel into their traditional business model. The specific focus is on strategic use of online service marketplaces, i.e., e-lancing markets, to manage the excess capacity of an IT service provider's labor pool. The remainder of the paper is composed as follows: We briefly review the related literature in the second section. The third section presents our analytical model for the workforce management problem with two channels, the conventional channel and the online channel, by which an IT service firm can dynamically decide whether to fulfill an incoming order. The fourth section analyzes the structure of optimal policy by defining conditions for a preferred class and proving the submodularity. Then, extended models and related analysis are presented for the case where workers are released together after completing a project. The following section demonstrates the characteristics of the optimal policies, which are obtained by numerical computations. Finally, in the last section we conclude the paper by discussing contributions and future research. Proofs of all the theorems and additional numerical examples are included in an electronic supplementary material.

Section snippets

Literature review

With the growing popularity of using Business-to-Business (B2B) online reverse auctions for procurement, a number of studies have been recently published from various perspectives. Schoenherr and Mabert [39] offer procurement managers managerial insights and best practices obtained from case study companies which used online reverse auctions for sourcing. Greenwald, Kannan, and Krishnan [15] examine the procurer's information revelation policy in multiple auction sessions and the influence of

Model development

The workforce management problem of an IT service firm is modeled as a dynamic admission control problem in a two-class Markovian loss system with multi-servers receiving random batches. The assumptions made in the model are as follows:

  • 1)

    There is neither staff augmentation nor loss during the period considered in the analysis. That is, the total number of IT workers in the firm remains constant.

  • 2)

    The pool of IT workers is composed of homogeneous developers/programmers both in terms of their skills

Results on the preferred class and submodularity

In this section, we will derive sufficient conditions for class 1 and class 2 jobs to be preferred. Let ei be the i-th unit vector which has 1 at the i-th position and 0 elsewhere. We denote Bn(0i)(x) = Vn(x)  Vn(x + ei) as the expected loss (burden) to the system that an additional class i job creates when the system is in state x and there are n remaining transitions. Similarly, we denote Bn(ki)(x) = Vn(x + ek)  Vn(x + ei) as the expected burden to the system if a class k job is replaced by a class i

Extended models: when workers in a team are released together

The model we described so far allows individual workers to complete modularized tasks at their service rate and be released independently. This way, the availability of workers is maximized because workers do not have to wait for others in the team to complete their tasks. Nevertheless, we note that it may not be always feasible to implement this, due to the difficulty of modularizing a project, workers may have to be released together, and consequently a different formulation, which we will

Numerical computations

This section examines the structures of the optimal policies for the base problem discussed in Section 3 and presents several numerical solutions obtained by the Value Iteration algorithm [34], [42]. We employ this fundamental Markov decision programming algorithm for its straightforward implementation; proposing or proving faster algorithms is beyond the scope of this work. However, note that recent advances in algorithms for Markov Decision Processes have made many efficient alternative

Discussion and conclusion

This paper addresses the need for improving workforce utilization in IT service firms in response to the fluctuating market demand. We presented a novel approach in which IT service providers can maximize revenue by taking advantage of the emerging e-lancing marketplaces and dynamically controlling project orders in both traditional and online channels. Our study verifies the effectiveness of the new revenue model and provides optimal policies to successfully implement it.

Using Markov decision

Joung Yeon (J.Y.) Kim is an Assistant Professor of Accounting at School of Business, Indiana University Kokomo. She received her Ph.D. degree in Management from Purdue University's Krannert School of Management. Her research focuses on understanding user behavior in online marketplaces, revenue management of IT service firms, and accounting information systems. She holds both master's and bachelor's degrees in chemical engineering from Korea Advance Institute of Science Technology.

References (43)

  • R. Caldentey et al.

    Revenue management of a make-to-stock queue

    Operations Research

    (2006)
  • S.M. Carr

    Note on online auctions with costly bid evaluation

    Management Science

    (2003)
  • S. Carr et al.

    Optimal admission control and sequencing in a make-to-stock/make-to-order production system

    Operations Research

    (2000)
  • P. Dai et al.

    Topological value iteration algorithm for Markov decision processes

  • T.W. Ferratt et al.

    IT human resource management configurations and IT turnover: theoretical synthesis and empirical analysis

    Information Systems Research

    (2005)
  • G. Gallego et al.

    Optimal dynamic pricing of inventories with stochastic demand over finite horizons

    Management Science

    (1994)
  • Gartner-Research

    Flexible Staffing Models: Breaking the Staff up/Layoff Cycle?

    (April 2001)
  • Y. Gerchak et al.

    Reservation planning for elective surgery under uncertain demand for emergency surgery

    Management Science

    (1996)
  • A. Greenwald et al.

    On evaluating information revelation policy in procurement auctions: a Markov decision process approach

    Information Systems Research

    (2010)
  • F. Heller

    Decision Making and Leadership

    (1992)
  • M. Jorgensen et al.

    A systematic review of software development cost estimation studies

    IEEE Transactions on Software Engineering

    (2007)
  • Cited by (9)

    • Operational risk: Emerging markets, sectors and measurement

      2015, European Journal of Operational Research
      Citation Excerpt :

      In Garcıa-Dastugue and Lambert (2003) discuss the use of internet specific decision support systems in reducing operational risk facing companies in supply chain related risks. In Kim, Altinkemer, and Bisi (2012) yield management decision support systems reduce operational losses but also improve workforce utilisation. Despite the acknowledged importance of decision support systems in improving operational risk management in firms, it has been recognised in various studies that emerging market firms typically do not have the same level of decision support systems as in developed markets.

    • Decision-making in IT service management: A simulation based approach

      2014, Decision Support Systems
      Citation Excerpt :

      Simulation modeling is widely used to support decision-making in different business areas and management tasks. An extensive variety of practical questions can be addressed with simulation, such as strategic management, planning, control and operational management, process improvement and technology adoption [16,24,26,36,39,50,82,95,108]. The growing global diffusion of ITSM frameworks and standards due to the fact that organizations that implement them obtain important benefits [15,61,62].

    • Organizing knowledge workforce for specified iterative software development tasks

      2014, Decision Support Systems
      Citation Excerpt :

      For human resources like workforce, Faraj and Sproull [22] suggested that expertise in software development teams must be managed and coordinated in order to leverage its potential, and knowledge workforce needs to be efficiently utilized in order to achieve desired team performance. As a result, excess or under-utilized workforce is usually regarded as a loss to an organization [35]. In this study, we investigate efficiency in terms of workforce utilization during the software project development as another process performance measurement.

    • A model to support IT infrastructure planning and the allocation of IT governance authority

      2014, Decision Support Systems
      Citation Excerpt :

      This study continues research into how decision support techniques can facilitate IT decisions, building on prior work in a knowledge warehouse setting [27], electronic market setting [29], and Internet server-based setting [5]. Our use of a Markov decision problem model is consistent with recent research that uses MDP for other IT governance issues such as workforce and data management [12,21]. While these studies support improved performance of specific IT infrastructure, this study broadens and extends prior research by offering a decision support model to ensure that overall IT investments are aligned with firm strategy.

    View all citing articles on Scopus

    Joung Yeon (J.Y.) Kim is an Assistant Professor of Accounting at School of Business, Indiana University Kokomo. She received her Ph.D. degree in Management from Purdue University's Krannert School of Management. Her research focuses on understanding user behavior in online marketplaces, revenue management of IT service firms, and accounting information systems. She holds both master's and bachelor's degrees in chemical engineering from Korea Advance Institute of Science Technology.

    Kemal Altinkemer is an Associate Professor at the Krannert Graduate School of Management at Purdue University. He received his PhD from University of Rochester, Rochester NY. He was a Guest Editor in Telecommunication Systems, Information Technology and Management. He is an Associate Editor in seven journals. His research interests are in Design and analysis of computer networks, infrastructure development, distribution of priorities by using pricing as a tool, infrastructure for E-commerce and pricing of information goods, bidding with intelligent software agents, strategy for Brickandmortar business model. He has published more than 40 articles in journals such as Operations Research, Operations Research Letters, Management Science, INFORMS Journal on Computing, Transportation Science, EJOR, Computers and OR, Annals of Operations Research, and more than 40 others in various conference proceedings. He belongs to ACM, AIS and INFORMS.

    Arnab Bisi is an Associate Professor at the MYRA School of Business, India. He has been an Assistant Professor at the Krannert School of Management at Purdue University, USA. His research and teaching interests include stochastic models, statistics, inventory and supply chain management. He received a Ph.D. in Mathematics and Statistics from Hong Kong University of Science and Technology, an M.Stat degree from Indian Statistical Institute and a B.Sc. in Statistics from the University of Calcutta.

    View full text