Elsevier

Decision Support Systems

Volume 42, Issue 3, December 2006, Pages 1684-1696
Decision Support Systems

A decision support model for optimal timing of investments in information technology upgrades

https://doi.org/10.1016/j.dss.2006.02.013Get rights and content

Abstract

In an environment of continuous change, organizations are faced with the challenge of deciding when to invest in information technology upgrades. While investing frequently is costly and at times risky, waiting too long can lead to lost competitiveness. Further, investing at a given time can preclude a firm from taking advantage of better technologies in the future. In the context of software upgrades, this study proposes and illustrates a decision support model to determine the optimal timing and choice of upgrades. Analysis confirms that even if continuous upgrading is feasible, it is not an optimal strategy when adoption costs are significant. Simulations show that investments in upgrades are best made when the gap between new technology and current technology reaches a critical threshold. Among other factors, this threshold is influenced by technology cost, change management cost and opportunity cost.

Introduction

In an environment of continuous technological change, organizations are frequently faced with the challenge of deciding when to invest in new and upgraded information technology (IT). Consider the releases of operating systems (OS) by Microsoft—Windows 95 in August 1995, Windows 98 in June 1998, Windows 98 Second edition in 1999, Windows ME and Windows 2000 in the year 2000 and Windows XP in 2002. With this continuous stream of upgrade releases, individuals and organizations are faced with the decision to upgrade their OS. Typically, few would upgrade to a new version every time a release is announced; instead they would leapfrog to adopting a subsequent release [40]. Making such technology investments continuously (i.e., every time a new release of a technology is announced) can be very expensive.

While investing frequently is costly and at times risky, waiting too long can put an organization at the risk of losing first mover advantages associated with introducing competitive products and services. However, by waiting, the firm can purchase a superior technology than the one available now. Clearly, in this context, it is critical for a firm to determine whether it should incur the adoption costs now to take advantage of the productivity and competitiveness gains that the currently available new or upgraded technology provides or wait for a better technology. This study examines how a firm can determine this optimal interval between adoptions and thereby its optimal choice of technology.

The primary focus is on the case of software upgrade decisions. Cases of acquiring a technology for the first time are not considered. Unlike new technology adoption, in the case of upgrade the question is more of “when” to adopt than “whether” to adopt. Furthermore, in contrast to new technologies, the uncertainties associated with benefits and costs are relatively small in the case of upgrades. First time investments in new technologies have received a lot of attention in the literature and models like net present value and options based evaluation have been developed for decision support. On the other hand, few studies have focused on upgrade situations.

The choice of upgrade situations is motivated by several reasons. First, periodic upgrade investments are increasingly becoming a significant percentage of total IT budgets and hence, an area of utmost interest to researchers and technology managers. Second, organizations seem to time the upgrades rather arbitrarily without a systematic analysis. Third, even in the case of upgrades, where reasonably accurate predictions about technological change can be made, it is not clear how different costs involved with technology adoption affect the intervals (timing) at which adoption decisions are made. Finally, despite the importance of the topic it has remained largely unexplored. To this end, the objective of this research is twofold:

  • (1)

    To develop a decision support model to determine the optimal time and choice of upgrade investment, and

  • (2)

    To study the impact of various costs on the optimal upgrade time and upgrade choice.

It is worth noting that based on the type of technology under consideration, there can be substantial differences in factors that drive adoption decisions. For example, the primary driving force behind an upgrade decision might be lack of vendor support for one firm but compatibility with competitors' or users' software might be the critical factor for another. Clearly, these are opportunity costs associated with not adopting the upgrade and can be incorporated as such in decision support models.

A decision support model must include a consideration of the dynamic context to take full account of the impact of current upgrade decisions on future ones. In a competitive landscape, technological improvements occur rapidly and investing at a given time may preclude a firm from taking advantage of better technologies at a later date. Thus, adoption decisions must consider the future impact of current decisions.

Interestingly, widely applied investment evaluation methods in information systems (IS) research like net present value (NPV) do not take into account the dynamics discussed above as these methods do not consider the impact on future decisions. In the case of continuous upgrades, it is important for firms to decide the frequency at which its technology must be replaced. Thus, unlike other types of investment decisions, firms would benefit from a long term “plan” for investment in IT upgrades. Conducting disjoint static analysis ignores the element of inter-temporal interdependence of the investments—a critical consideration for technology upgrades.

To address the interdependent nature of the decisions, this study draws upon models that have specifically addressed this issue (for e.g., see [2], [3], [11]). [3] developed a model that examined the problem of technology adoption that allows the state of nature to change at a stochastic rate. They apply an impulse-control method to determine the intervals at which decisions are made. This framework is most suitable for this study since it has been developed to solve exactly the type of dynamic problem IT decision-makers face. That is, the framework allows us to examine how a firm should optimally spread its investment over time in an environment of continuous change.

Other dynamic models have been used to study technology adoption decisions. The applicability of alternative dynamic models, such as, options-pricing and the proposed model are, however, quite different. The main motivation for using options pricing in articles such as [4] is that the benefits from investment in the technology are uncertain. By using real options, a firm delays adoption to gather valuable information regarding the investment. The question addressed in this study is quite different. This study's model is applicable when uncertainties regarding benefits or costs are not the key factors. The main concern is that firms understand that adopting something new today implies that very soon an even better technology will become available. Should it wait for that better technology or adopt now? Should it adopt every time a new or upgraded technology is released? How do the different costs involved with adoption affect the adoption decision? These questions cannot be answered by options pricing type models. In fact, the impulse control model was developed to address exactly these questions.

Section snippets

Literature Review

A critical issue in any IT investment evaluation, particularly true in the case of upgrades, is the timing of adoption. The investment decision, in upgrade situations, is usually less focused on the question of “whether” to acquire a technology or not but more on “when” to acquire it. Several studies in economics have examined the timing of adoption of new technologies. Some of the seminal studies include [20], [21], [32], [33], [34]. These studies highlight the role of competition between

Model description

This study considers the technology upgrade decision of a single firm. New upgrades appear at a steady, deterministic pace. For example, every year, output can be increased by 100,000 units by implementing upgrades to the technology in use. In such an environment of continuous improvement, unless a firm adopts new upgrades immediately upon release, the gap between the technology in use and the most advanced available increases continuously. The firm has three alternatives: (i) adopt new

Overview of simulations

Simulations were conducted to arrive at numerical solutions and for analyzing the sensitivities of the results to changes in technology cost, change-management cost, and opportunity cost.3 The choice of range of values for the parameters was based on typical upgrade scenarios examined, one of which is presented here.

Operating system upgrade data from Microsoft was used to create the scenarios.4

Conclusions and areas for future research

The phenomenal rate at which new technology is available and the competitive business environment are putting increasing pressure on firms to make difficult upgrade decisions. Theoretically, upgrading technology every time a new version is released should keep firms at the cutting edge constantly. Practically, in an environment of continuous technological change, this is fraught with difficulties of managing change. Analysis suggests that even if continuous upgrading is feasible, it is not an

Acknowledgement

The authors thank the reviewers of the paper for many useful suggestions. The authors are responsible for any remaining errors.

References (40)

  • J.P. Choi

    Irreversible choice of uncertain technologies with network externalities

    RAND Journal of Economics

    (1994)
  • K. Clemons

    Evaluation of strategic investment in information technology

    Communication of the ACM

    (1991)
  • M.A. Cohen et al.

    Optimal Technology Choice in a Dynamic Stochastic Environment

    (1986)
  • G.M. Constantinides et al.

    Existence of optimal simple policies for discounted-cost inventory and cash management in continuous time

    Operations Research

    (1978)
  • B.L. Dos Santos

    Justifying investment in new information technologies

    Journal of MIS Research

    (1991)
  • J. Farrell et al.

    Standardization, compatibility, and innovation

    Rand Journal of Economics

    (1985)
  • R.K. Goel et al.

    On the adoption of new technologies

    Applied Economics

    (1997)
  • Information Week, Feb 14,...
  • P.C. Jones et al.

    Parallel machine replacement

    Naval Research Logistics

    (1991)
  • P.C. Judge et al.

    Why the fastest chip didn't win

    Business Week

    (April 28 1997)
  • Cited by (38)

    • Recommendation of startups as technology cooperation candidates from the perspectives of similarity and potential: A deep learning approach

      2020, Decision Support Systems
      Citation Excerpt :

      Nowadays, companies and organizations consistently seek emerging trends and develop corresponding technologies to survive [1].

    • Optimal software upgrade strategy: Should we sell products or premium services in the presence of piracy?

      2018, Electronic Commerce Research and Applications
      Citation Excerpt :

      These three areas constitute the versioning of software (Dey and Lahiri, 2016; Chellappa and Mehra, 2017). The existing literature on software upgrades mainly addresses two aspects: what is the upgrade pricing strategy (Kornish, 2001; Bala and Carr, 2009; Mehra et al., 2012; Zhu and Zhou, 2012) and what is the optimal timing of investments in upgrades (Mukherji et al., 2006; Mehra et al., 2014; Morgan and Ngwenyama, 2015; Ngwenyama et al., 2007). Studies have considered other aspects of upgrade strategies, such as the social and economic impacts on the diffusion dynamics of open-source software (Zaffar et al., 2011; Mehra and Mookerjee, 2012), the impacts of switching costs and IT investment strategy (Demirhan et al., 2007), and the relationship among the maintenance, upgrade policy and optimal warranty (Sahin and Zahedi, 2000).

    • A decision model and system for planning and adapting the configuration of enterprise information systems

      2017, Computers in Industry
      Citation Excerpt :

      There is a need to characterize them with specific EIS impact elements and magnitudes to estimate not only an initial cost of this type of investment, but also to estimate the cost and additional dimensions (e.g., time, human capabilities, considerations) of the upcoming implementation. Finally, Mukherji et al. [29] present a simulation model to identify the optimal moment to invest in the upgrade of an IT system. They considered technology, change management, and opportunity costs.

    • Delayed product introduction

      2012, Decision Support Systems
      Citation Excerpt :

      Therefore, to avoid getting products that could soon be obsolete, consumers may decide to wait for new products and defer making purchases. The literature refers to such a behavior as “leapfrogging” [20]. In response to the threat of leapfrogging, firms that sell multiple generations of similar products may wish to slow down the pace of new product introduction.5

    • Periodic updates relying on open source software: An impulse approach

      2012, IFAC Proceedings Volumes (IFAC-PapersOnline)
    • Diffusion dynamics of open source software: An agent-based computational economics (ACE) approach

      2011, Decision Support Systems
      Citation Excerpt :

      Similarly, the timing of upgrades from the consumer's perspective can affect the diffusion dynamics as well. Mukherji et al. state that “firms would benefit from a long term ‘plan’ for investment in IT upgrades” [33: pp. 1685]. Furthermore, it is important that long-term costs are taken into account during upgrade decisions since choice of an operating system is more like a platform decision that affects hardware, existing application portfolio, staffing/training issues etc. [14].

    View all citing articles on Scopus
    View full text