J. Eng. Technol. Manage.

35 (2015) 93–114

Contents lists available at ScienceDirect

Journal of Engineering and

Technology Management
journal homepage: www.elsevier.com/locate/jengtecman

A real options framework for R&D planning in

technology-based firms
Juite Wang *, C.-Y. Wang 1, Cheng-Yo Wu 1
Graduate Institute of Technology Management, National Chung Hsing University, 250 KuoKuang Road,
Taichung 402, Taiwan, ROC

A R T I C L E I N F O A B S T R A C T

JEL classification: The increasing uncertainty in technology and market trends makes
O32 it difficult for technology firms to capture emerging opportunities.
This paper develops an R&D planning framework based on the real
Keywords:
Technology management
options analysis to identify, develop and evaluate technological
Real options opportunities. The developed methodology can encourage R&D
R&D management organizations to proactively explore uncertainties and to identify
Technology planning managerial options to capture emerging opportunities. Further-
more, the developed real options model integrated with the Bass
diffusion model can help R&D managers evaluate and select optimal
investment decisions to maximize market payoff under different
demand structures. The case of a biochip R&D project is studied.
ß 2014 Elsevier B.V. All rights reserved.

Introduction

Since technology is an important asset for many firms in sustaining growth and retaining
competitive advantage, R&D often plays a significant role in determining a firm’s technology position.
Basically, technology underpins future products and the manufacturing processes that support these
products for a technology-reliant firm. It is very important for such a firm to clearly understand what
technologies are available to them and how to select and exploit appropriate technologies that provide
their products or services with a competitive edge. Although the importance of technology planning is

* Corresponding author. Tel.: +886 4 22840547; fax: +886 4 22859480.

E-mail addresses: rdwang@nchu.edu.tw (J. Wang), rockytallen@gmail.com (C.-Y. Wang),
s0933269@mail.ncyu.edu.tw (C.-Y. Wu).
1
Tel.: +886 4 22840547.

http://dx.doi.org/10.1016/j.jengtecman.2014.12.001
0923-4748/ß 2014 Elsevier B.V. All rights reserved.
94 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114

well understood, most technology-based firms make their R&D decisions either based on group
discussion or some strategic planning tools, such as technology roadmapping (Petrick and Echols,
2004). With increasing complexities of analyzing uncertainty in technology development and market
trends (Thorn et al., 2011), R&D decision making has become more difficult for technology-based firms
seeking to enhance their competitive position and drive their sustainable profitable growth (Pich et al.,
2002; Song et al., 2007). The presence of tremendous uncertainty has led to many failures in their R&D
investments. Therefore, determining the optimal R&D decisions for capturing emerging technological
opportunities over time has become the major challenges for hi-tech industries (Scott, 2001).
In the fast changing industrial environment, all forecasts can be expected to err in significant ways.
Technology-based firms, therefore, should incorporate more managerial flexibility into their R&D
processes in order to allow them to reassess and adjust their decisions in a timely manner according to
the impact of environmental changes (Hunt et al., 2004). Building flexibility into R&D processes,
however, always incurs costs, because it involves redundant capabilities by making smaller
investments that may miss out on economies of scale, or it may cause delays that reduce benefits
(Verganti, 1999). The management challenge is to properly choose and implement different forms of
flexibility in the R&D process and to analyze the cost–benefit tradeoff for maximizing potential market
payoffs, while minimizing downside risks.
This paper presents a real-options-based methodology that identifies potential opportunities,
develops R&D plans with managerial flexibility to capture opportunities and avoid risks, and then
evaluates the value of these opportunities. The proposed methodology is developed based on the
concept of technology roadmapping (Farrukh et al., 2003), together with the real options model by
Huchzermeier and Loch (2001) to manage technological opportunities for maximizing market payoff,
while minimizing downside losses. Since the market value of a technological opportunity is affected
by its diffusion pattern, this paper extends Huchzermeier and Loch’s real options model
(Huchzermeier and Loch, 2001) with the Bass innovation diffusion method (Bass, 1969; Mahajan
et al., 1995) to allow decision makers to evaluate and select optimal managerial options to maximize
market payoffs under different demand structures. The proposed methodology is illustrated with a
case study of a biochip R&D project. Taguchi’s robust design method (Phadke, 1989) was applied to
evaluate and select the decision that is more robust to market uncertainty.
This paper is organized as follows: the following section provides a literature review: the proposed
real options-based R&D planning framework is developed in section ‘‘A Real Options Methodology for
R&D planning’’; A case study of a biochip R&D project is illustrated in Section ‘‘Case Study of a biochip
R&D project’’ and experimental results and managerial implications are also discussed as well; Section
‘‘Conclusions’’ concludes the paper.

Literature review

This paper reviews three popular approaches in the literature for R&D planning: strategic planning
tools, decision analysis approaches, and benefit measurement methods.

Strategic planning tools

In strategic planning tools, technology roadmapping is a comprehensive approach for the

integration of technology development into product development and business strategies. It was
developed in the late 1970s and has been widely applied in industry and academia for developing
technology strategies, identifying gaps and opportunities in technology development, and resource
allocation planning (Farrukh et al., 2003; Petrick and Echols, 2004; Phaal et al., 2004). Farrukh et al.
(2003) proposed a fast-start technology roadmapping approach (T-Plan) for product planning by
practitioners, and several possible extensions, such as the exploration of new opportunities, have been
presented. Walsh (2004) modified the traditional roadmapping approach, generating a disruptive
technology roadmapping process. As products become more complex, the expertise needed to
successfully develop a new product often lies in more than one firm. Thus, Petrick and Echols (2004)
suggested that firms should extend and share their technology roadmaps with their supply chain
partners to make more sustainable new product development decisions. Although technology
 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114 95

roadmapping is fairly simple in structure and concept, the output quality is still dependent on the
strategy and planning process, which involves considerable levels of detail and depends on the
accuracy of the predicted future. There is also a lack of supporting management tools available. Several
studies have integrated technology roadmapping with management tools, such as scenario planning
(Strauss and Radnor, 2004; Drew, 2006; Saritas and Aylen, 2010), technology valuation (Thorn et al.,
2011), and patent analysis (Lee et al., 2007).

Decision analysis approaches

Since R&D planning is a complex decision making process, many studies have applied multi-
criteria decision analysis (MCDA), a discipline concerned with structuring and solving decision and
planning problems involving multiple criteria (Figueira et al., 2005) for organizing and analyzing
technological investment decisions. Several researchers have applied analytic hierarchy process (AHP)
to assess and rank technological opportunities in different industries (Sivarama Prasad and
Somasekhara, 1990; Melachrinoudis and Rice, 1991; Chang-Kyo et al., 1994; Cho and Kwon, 2004).
However, several challenges remain for applying MCDA methods in technology planning and
investment, such as dynamics and uncertainties in the long-term planning horizon, interactions
among technology alternatives, and conflicting opinions between multiple stakeholders (Salo et al.,
2003). Prabhu and Vizayakumar (2001) developed a fuzzy hierarchical decision-making framework,
which considers cross-impacts of different factors (e.g., cost, technology, and positive synergy) for
technology choice on the iron-making technology. Cho and Kwon (2004) extended AHP with a cross-
impact analysis that can assist in ranking a large number of dependent technological alternatives.
Gerdsri and Kocaoglu (2007) developed the concept of technology development envelope (TDE) to
capture the dynamic nature of technology development and applied the AHP approach to the TDE
framework for technology selection. Chen et al. (2009) applied AHP to develop a technology
assessment model and used sensitivity analysis to evaluate the stability of selection decisions.

Benefit measurement approaches

Benefit measurement approaches usually use financial measures (e.g., cost, net present value, rate
of return, break-even period) or scoring methods to rate the benefits or costs of a technological
alternative (Tipping and Zeffren, 1995; Cooper, 2006; Mitchell et al., 2010). Since the traditional net
present value (NPV) and discounted cash flow (DCF)-based analyses often underestimate the value of a
technological investment with high uncertainty, several real options analysis (ROA) approaches, based
on financial option theory, have been developed for R&D decisions (Dixit and Pindyck, 1994). ROA
allows managers to consider the flexibility of decision making by having various options in place that
can be exercised as new information emerges (Trigeorgis, 1996). It has been applied in R&D project
valuation (Huchzermeier and Loch, 2001; Loch and Bode-Greuel, 2001; Santiago and Bifano, 2005; De
Reyck et al., 2008; Schneider et al., 2008; Cassimon et al., 2011; Thorn et al., 2011), R&D risk
management (Wang and Yang, 2012), investment timing (Kauffman and Li, 2005; Jang et al., 2013),
strategic planning (MacMillan and McGrath, 2002; Smit and Trigeorgis, 2006; Martzoukos and
Zacharias, 2013), and R&D portfolio management (Rogers et al., 2002; Wang and Hwang, 2007; van
Bekkum et al., 2009; Lo Nigro et al., 2014). Several studies have also discussed the benefits and
limitations of real options analysis in practice (Bowman and Moskowitz, 2001; Hunt et al., 2004;
McGrath and Nerkar, 2004; Krychowski and Quelin, 2010; Thorn et al., 2011).
From the above literature review, we found that few studies have focused on exploiting the
dynamics of R&D (Chen et al., 2009; Gerdsri et al., 2009) and their approaches do not consider
managerial flexibility in R&D decision making. Furthermore, the studies of decision analysis and
benefit measurement approaches are mainly on the assessment and selection of technologies or R&D
projects. Although several studies have used real options analysis to assess the value of managerial
flexibility (Huchzermeier and Loch, 2001; Loch and Bode-Greuel, 2001; Santiago and Bifano, 2005; De
Reyck et al., 2008; Schneider et al., 2008), most of them are not focused on how to explore technology
opportunities to capture upside potential while reducing downside losses. In addition, R&D decision
making should consider the impact of demand structure on market pay-off for managing demand
96 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114

uncertainty. Therefore, this paper develops a new methodology that integrates the strengths of
technology roadmapping and real options analysis, together with Bass’s innovation diffusion model, to
fill the research gap.

A Real Options Methodology for R&D planning

A typical R&D project can be divided into several phases and the milestone approach is used to
control its progress (Cooper, 2000). The real options approach has received great attention in recent
years (Krychowski and Quelin, 2010), because an initial investment of an R&D project is similar to the
purchase of an option on a future investment. The decision makers also have the option to stop or defer
the project at the end of each phase. Therefore, each phase is an option that is contingent on the earlier
exercise of other options. If the project is a technical success, then it creates the option to make a
significantly larger investment in the continuing project with relatively higher expected net benefit. If
the project fails to achieve the technical success, then there is no need to commit any further
resources, and therefore the downside risk is limited to the initial investment cost of the R&D project.
Therefore, there is a logical basis for adopting the real options approach for R&D planning.
The proposed real-option-based R&D planning involves three stages which is explain bellows.

Stage 1: Opportunity identification

The first stage is to identify potential market drivers (e.g., unmet customer needs) that may grow
market payoff in the future and to establish their linkage to products and technological solutions for
fulfilling the future market requirements. The set of product features and required performance levels
to satisfy the identified market requirements is also determined. Then, the possible technology
solutions that are embodied in a product and create benefits for its users can be recognized. The
market–product–technology linkage provides a systematic way to identify new or improved
technology solutions and interactions among them for potential commercialization. Note that it is
important to determine the desired product performance that is a fundamental factor used to link the
market drivers with technological capabilities. The potential market payoff, which might be
influenced by the demand structure, adoption speed, likelihood of blocking, and likelihood of
expropriation (McGrath, 1997), is also estimated at this stage.
In this study, the market–product matrix is used to determine important product features based on
identified market requirements. In addition, the market–product–technology linkage diagram is used
to depict the relationships among markets, products and technologies for a technological opportunity.
Both tools will be illustrated in Section ‘‘Case Study of a biochip R&D project’’.

Stage 2: Opportunity development

The key focus of this stage is on planning the technology/product development with managerial
flexibility to hedge against technical and external uncertainty for target technological opportunities.
Technical uncertainty is related to the difficulty of accomplishing technical success, while external
uncertainty is related to factors exogenous to the firm, such as market requirements uncertainty (Dixit
and Pindyck, 1994). Technical uncertainty can be resolved only by acquiring new information by
implementing an R&D project, for example, to understand the properties of new material (i.e., active
learning). External uncertainty varies unpredictably since it depends on future events that cannot be
known in advance and the prediction of external uncertainty is generally unreliable. Therefore,
incorporating managerial flexibility is very important to accommodate both technical and external
uncertainty. Although it is natural to adopt the real options approach for R&D decision making, how to
embed managerial flexibility into R&D processes to maximize upside potential, while minimizing
downside risks, still remains a challenge.
This stage first identifies the full range of contingent uncertainty that may affect new technology/
product performance in the future, and then maps the uncertainty that must be managed to
appropriate managerial actions or options. This study categorizes contingent uncertainty in R&D into
four types, based on previous studies (McGrath, 1997; Huchzermeier and Loch, 2001; Keizer et al.,
 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114 97

2005): market uncertainty, competitive uncertainty, development uncertainty, and commercializa-

tion uncertainty. Identified uncertainties and corresponding real options that have been discussed in
the literature for managing those uncertainties are listed in Table 1. This can help R&D managers to
foresee what risks may occur in the future for their technology development and to determine a
proper set of options for maximizing the value of the entire technology development based on the
industrial characteristics and the competitive advantages of the firm. The plus or minus sign behind
each uncertainty indicates the uncertainty effect on option value. For instance, the uncertainty of
adoption speed will decrease the value of options and an expansion option can be used to reduce user
adoption costs to enhance perceived customer values. It is also noteworthy that some of the options
may be compound options, in which the exercise of an associated earlier option gives rise to a later
option. For example, a strategic alliance may be used to manage the uncertainty of adoption speed,
since it provides two sequential options: the option to defer full commitment to a technological
investment and the growth option to capture potential technology opportunities (Vassolo et al., 2004).
In this study, the uncertainty–option matrix is used to determine appropriate managerial options
for overcoming each recognized uncertainty. The process diagram is used to display the various R&D
phases and various managerial options at each review gate. Both tools will be illustrated in Section
‘‘Case Study of a biochip R&D project’’.

Stage 3: Opportunity evaluation

This stage evaluates the value of R&D investment based on the R&D plan with flexibility
determined in stage 2. Since new product diffusion speed will affect the scale of potential payoffs from
the market, it is important to consider the factor of user adoption speed in the valuation of
technological investment. Slow adoption would delay revenue inflows and give more time for other
firms using established technologies or for new competitors to react. This paper extends the real
options model by Huchzermeier and Loch (2001) with Bass’s diffusion of innovation model (Mahajan
et al., 1995) based on the assumption that better product performance will increase the product
adoption willingness of users.
The real options model is described as follows. Assume that there are T + 1 review stages, t = 0, 1, 2,
. . ., T, and the new product is launched on the market at stage T. Uncertainty in technology/product

Table 1
Uncertainties and recognized options in R&D.

Category Uncertainty Options

Market uncertainty Demand structure (+) Expansion option, switching option, deferral
option, abandonment option, licensing-out
Adoption speed () Expansion option, strategic alliance
option (deferral & growth options), abandonment option
Competitive Blocking uncertainty () Strategic alliance option, deferral option, abandonment option
uncertainty
Matching/imitation () Expansion option, growth option, strategic alliance option,
abandonment option
Development Technology capability () In-license option (deferral & expansion or abandonment options),
uncertainty time-to-build option, growth option, outsourcing (deferral &
expansion or abandonment options), strategic alliance option
Project management Expansion option, contraction, outsource option, abandonment option
(budget, schedule)
Intellectual property () Expansion option, growth option, strategic alliance option,
abandonment option
Regulatory () Strategic alliance option, out-license option, abandonment option
Commercialization Infrastructure Strategic alliance option, abandonment option
uncertainty requirements ()
Parallel technology Strategic alliance option, abandonment option
uncertainty ()
Supply chain and Strategic alliance option, switching option, abandonment option
sourcing ()
98 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114

development can be represented by performance improvement and deterioration spread over

possible performance states. The state of the R&D plan at stage t is characterized by the expected
product performance X = (x1, x2, . . ., xn), where xk is the individual product performance, k = 1, . . ., n. Let
d be the management decision (e.g., continue, abandonment, etc.) at stage t. It is assumed that
whenever the system is in state Xi and decision d is made at stage t, the system moves to a new state Xj
at stage t + 1 with transition probability pti j ðdÞ. For development costs, an initial investment of I must
be made to start the technology development. In addition, a cost ct(d) is incurred if the management
decision d is selected at stage t.
Let XT be the performance state achieved at stage T. The market demand during stage t 2 [T+1, L] is
estimated as follows:

 
b
SðtÞ ¼ aFðX T Þ½m  Nðt  1Þ þ Nðt  1Þ½m  Nðt  1Þ þ g Nðt  1Þ (1)
m

where F(XT) = Prob(XT  D) represents the probability that performance XT exceeds the market
requirement D, m is the total market size, N(t) is the market demand through stage t, a is the
coefficient of innovation, b is the coefficient of imitation, and g is the average product consumption.
The first item of Eq. (1) denotes the innovation demand that is influenced by the firm’s product
performance relative to market requirement or competitor’s product performance. The second item
determines the imitation demand, and all of the firms in the market have similar word-of-mouth
effect. The last item calculates the repeat purchase. Parameters a, b, and m can be estimated by
analogous products or by statistical methods if initial sales data is available (Mahajan et al., 1990).
Let p be the average price per usage, q be the market share, and d be the contribution rate. The total
profit can be estimated:

X
L
P ðX T Þ ¼ pqd SðtÞ (2)
t¼Tþ1

Then the technology planning problem can be considered as a sequential decision problem that can be
formulated as a stochastic dynamic program and can be solved by the standard backward recursion.
Let r be the discount rate and nt be the total number of states at stage t + 1. The value function of
performance state Xi at stage t is written as follows:

8
>
> 1 Xnt
>
> Maxfct ðdÞ þ pti j ðdÞV j ðt þ 1Þg; t ¼ 0; . . . ; T  1
< d 1þr j¼0
V i ðtÞ ¼ Xnt Y T (3)
>
> 1
> pTij ðdÞ
: Max
>
d
fcT ðdÞ þ
1þr
ðX Þg; t¼T
j¼0

The above real options valuation model will be illustrated in the following section.

Case study of a biochip R&D project

The proposed methodology is illustrated with a case study based on an industrial biochip R&D
project of a Taiwan biochip firm (company A) whose goal is to develop low cost and high quality DNA
chip products for clinical diagnosis, plant disease diagnosis, and food quality testing. The company has
developed three technological platforms and filed more than thirty-five patents in the US, EU, Japan,
China, and Taiwan. This research took a diagnostic biochip development project for improved
detection of cervical cancer. The required data was collected through interviews with a vice president
responsible for corporate R&D and a senior R&D engineer from the case company, from secondary
sources, and from related literature. Since the R&D project data is usually confidential, we have revised
the case as necessary. Some important data is disguised and assumptions are made as necessary.
 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114 99

Opportunity identification

Background
Cervical cancer is the twelfth most common and fifth most deadly cancer in women worldwide. It
affects about 16 per 100,000 women per year and kills about 9 per 100,000 per year. In Taiwan, cervical
cancer is also the fifth leading cause of cancer death for women (9.1 per 100,000) (Liao et al., 2006). The
Pap smear (cervical cancer screening) has been widely used in developed countries and is credited
with dramatically reducing the incidence and mortality of cervical cancer. Abnormal Pap smear results
may suggest the presence of cervical intraepithelial neoplasia (potentially premalignant changes in
the cervix) before a cancer has developed, allowing examination and possible preventive treatment.
However, the Pap smear has the drawback of a high false negative rate and it can detect cancer cells
only at stage one or two. Estimates of premature deaths that could be avoided through cancer
screening vary from 3% to 35%, depending on a variety of assumptions (National Cancer Institute,
2013).
Due to increasing health awareness, one of the key strategies in cervical cancer prevention is early
diagnosis through more advanced cancer screening technologies (Cuzick, 2010). Human papilloma-
virus (HPV) has been recognized as the major cause of cervical cancer. Although there are more than
100 subtypes, fewer than 20 of them are considered ‘‘high-risk’’ for the development of cervical cancer.
Therefore, there is a great market opportunity for developing a new cervical cancer screening tool
based on the new biochip technology for multiple HPV detection and genotyping. There is also a great
potential to improve early detection (at stage 0) of cervical cancer with better accuracy, ease of use,
efficiency and cost effectiveness. In addition to the potential of reducing mortality, it may reduce the
cost of health care, since treatment for earlier-stage cancers is often less expensive than that for more
advanced-stage disease.
The company’s objective was to develop an HPV test kit capable of detecting and discriminating
between different high-risk HPV genotypes with better sensitivity and specificity, thus providing
physicians with detailed information on the nature of the infection and the appropriate follow-up
treatments at an early stage of cancer. As shown in Fig. 1, there are two prospective markets:
laboratory research and clinical diagnostics. Company A was determined to focus on the clinical
diagnostics markets in Europe, especially with France as a beachhead, because the company had
obtained CE marking approval for an existing biochip technology platform. In addition, although the
HPV test kit can be sold to research institutes without FDA approval after passing clinical sample
studies, this paper uses just the clinical diagnostics market as an example to illustrate the proposed
methodology.

Market research and key product features

From market research, the number of women taking traditional cervical cancer screening was
estimated to be about 35 million annual in France and 40% of the population was assumed to adopt the

Fig. 1. The linkage of market, product, and technology.

100 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114

new biochip technology for cervical screening. Although there were five main competitors in the same
market, none of their technology platforms were perfect in terms of detection performance and costs.
In this study, the market share of company A was assumed to be 17%, the same as its competitors. As a
result, the potential market size was estimated to be 2.4 million units, which could bring a great
revenue growth for the company in the near future. Assuming a $50 price for a single test using the
developed biochip and a gross profit margin of about 50%, based on the average gross profit margin in
the past five years, the gross profit was approximately US$ 25 million. The proposed methodology can
be integrated with the choice-based conjoint analysis for more accurate market share estimation
(Green and Srinivasan, 1990).
Based on the market drivers identified, important product features that can satisfy those drivers are
defined in Table 2. The prediction accuracy, which is highly related to the two product features,
simultaneous genotyping and high validity, is used for illustrative purposes. Simultaneous genotyping
and high validity are measured by the number of HPV genotypes and Kappa measure, respectively. The
two metrics are described below.

 Number of HPV genotypes

In general, the more high-risk genotypes that can be detected, the better the product performance
is. There are about 14 high risk HPV genotypes (e.g., HPV-16, HPV-18), and individual HPV genotypes
may increase carcinogenesis more strongly (Smith et al., 2007). Also different racial groups may have
different high-risk genotypes that cause cervical cancer (e.g., HPV-16, HPV-52, HPV-58 for Asian
women). But for illustrative purposes, this research assumes that different HPV genotypes have the
same impact on increased carcinogenesis.
Furthermore, as the number of detectable genotypes increases, the development costs and lead-
time also increase. This is because each additional genotype for the new DNA fragments in the
polymerase chain reaction (PCR) might present a weak reaction, which leads to poor validity.
Therefore, it is necessary to spend more time designing new primers for improved PCR reaction in
order to obtain consistent quality of each genotype.
There were about nine global manufacturers, using different techniques (e.g., Electrophoresis, DNA
Sequencing, or biochip) for HPV genotype detection, and only five of the companies used biochips as a
diagnosis platform. The market requirement for the number of genotypes was assumed to be
characterized by a normal distribution with the mean value of 22 and standard deviation of 6 in the
following 5–7 years.

 Kappa measure (Fleiss et al., 2003)

Fleiss’ kappa statistical measure was considered as the second performance metric for assessing
the reliability agreement for the test results between traditional cervical cancer diagnosis method and
the new biochip approach. Compared to sensitivity and specificity, the kappa measure is a more robust
measure than a simple percentage agreement calculation like sensitivity and specificity, because it
considers agreements occurring by chance (Fleiss et al., 2003). Similarly, a normal distribution was
used to characterize the test reliability and the expected value and standard deviation of Kappa value
were estimated to be 0.72 and 0.12, respectively.

Table 2
The market–product matrix.

High Low product Simultaneous Portable High Reducing sample

throughput cost genotyping size validity volume

Accuracy M M H M H M
Ease of use M M M H L L
Efficiency H H H M M H
Cost effective M H H M M L

H: highly related, M: medium-related, L: low-related.

 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114 101

Although biochips can provide higher diagnostic quality, accuracy, and earlier cancer detection
than traditional approaches, there is a great uncertainty about the willingness of hospitals, clinics, and
end-users to adopt this new technology. Prices and health insurance coverage are the two main factors
in the adoption speed of the biochip platform. Therefore, reducing the cost of a biochip is one of the
main challenges faced by many biochip developers. To penetrate this emerging market, company A
also developed a new low-cost technology platform that used a biomolecule-bound substrate
technology.

Opportunity development

Identifying contingent uncertainties

The biochip R&D process can be divided into five phases: basic research, product specifications and
analytical validation, clinical sample studies, clinical evaluation and validation studies, and launch
and commercialization (see Fig. 2). Since biochip development faces great market and technology
uncertainties that may lead to project failure, it is important to identify contingent uncertainties that
may have a negative impact on project performance at this stage.
Company A identified several contingent factors at this stage: slow adoption speed, uncertain
demand structure, technical capability, and time and cost overruns. Although using the biochip for
cervical cancer screening might have great market potential in the future, the adoption speed still
remained low and the demand structure was still uncertain due to high cost, significant usage pattern
changes, and regulatory environment uncertainty. In addition, the company might not have the
technical capability to develop the test kit with performance levels that could satisfy market
requirements. Furthermore, the company did not have the permanent in-house staff and expertise to
conduct all aspects of a clinical testing and validation, which may cause time and cost overruns for the
project.

Fig. 2. The biochip R&D process.

102 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114

Identifying managerial options

After contingent factors had been identified, the company was required to determine appropriate
managerial options for each R&D phase to maximize upside potential, while minimizing downside
risks. The following managerial options were considered at certain R&D stages to overcome those
contingences (see Fig. 2 and Table 3).

 Continue/abandonment options

These two options were considered at every stage of the project. The project might be abandoned
because the biochip developed could not satisfy market requirements, failed to obtain regulatory
approval, or encountered a less promising than expected market. In this situation, selecting the
abandonment option would reduce the chance of incurring greater losses in the future. Otherwise,
selecting the continue option would advance the project into the next stage.

 License-in option

At the basic research stage, company A considered a technology license-in option to obtain a non-
exclusive license for access to required biomarkers and analytical validation methods related to
critical HPV genotypes of product H for time and cost saving. The license-in strategy provided
immediate access to more advanced technologies that might enhance the range of detectable HPV
genotypes and reliability, while saving R&D lead-time and cost. However, prior to the successful
technology transfer into company A, there might be higher uncertainty regarding the technology’s
achievable product performance in later R&D stages.

 Outsourcing options

Biotechnology R&D outsourcing is frequently used in the biotech and pharmaceutical industries,
especially at the clinical trials and manufacturing stages, and this outsourcing trend is still growing
(Piachaud, 2002). The benefits of outsourcing include reducing cost, gaining extra capacity, and
utilizing a vendor’s special expertise to enhance the competitiveness of developing products. Due to a
lack of permanent employees and the required expertise to carry out clinical trials, company A
considered outsourcing clinical studies of the developing biochip to a contract research organization
(CRO) to improve its productivity in phases 3 and 4. Alongside these benefits, however, this option also
has the drawback that the output quality of the outsourcing partner might vary due to inappropriate
process compatibility, coordination policy, and cultural fit between the two organizations (Piachaud,
2002).

 Switching options

Company A had developed a microarray-based diagnostic platform that used a traditional glass
substrate with good performance, but higher cost. The existing microarray platform could be used in
product H for the cervical cancer screening. Although the microarray-based diagnostic biochip is a
mature technology, the market is still in its infancy since its price is much higher than traditional

Table 3
The uncertainty–option matrix.

Uncertainty Option

Continue Abandonment License-in Switching Expansion Outsourcing

Demand structure x x x x
Adoption speed x x x x
Technical capability x x x x
Cost/time overrun x x x x
 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114 103

cervical cancer screening approaches. To penetrate the market, the company started to design a new
microarray prototype to lower the entire biochip cost, while maintaining similar product performance
and manufacturing performance levels. Instead of glass, the substrate of the new low-cost platform
used a macromolecular compound material that was capable of realizing the economic benefit of mass
production. However, a major disadvantage of this polymer substrate was its weak signal detection
and measurement. Therefore, the new platform still needed further improvement and required at
least one more year to finish it.
A switching option provides the right and ability but not the obligation to switch among different
sets of technologies, markets, or products (Trigeorgis, 1996). To overcome adoption speed uncertainty,
the company considered applying a switching option at stage two that delayed the platform selection
decision for one year until the new platform had been validated. If the low-cost platform was
approved, then the company could switch to the new platform and continue to develop the product
based on the new platform in the subsequent stages. However, if the new platform failed, then the
company still could use the existing platform for developing the new product.

 Expansion options

This type of option represents the possibility of adjusting the amount of investment upward,
depending on market condition and product performance. The company considered applying an
expansion option to improve product performance, if the performance could not meet market
requirements.

Opportunity evaluation

The proposed real options valuation model was applied in this case study to assess and select
appropriate managerial options at each stage to maximize market payoffs. Due to technology
uncertainty, a managerial option might result in actual performance improvement and deterioration
over possible performance levels. In the following, the uncertain performance states for managerial
options that were embedded into the project are presented.

 Continue option

The continue option means that the company conducted R&D activities by itself according to
predetermined allocated resources and schedule. The possible transition states, corresponding
probabilities, and development lead-times as well as costs are listed in Table 4. For example, at the
phase of basic research, the expected number of HPV genotypes might be improved from 22 to 24 (+2)
or 26 (+4). Alternatively, it might be reduced to 20 (2) or 18 (4). Since there is great technical
uncertainty in the early phase, the transition probabilities were equally set to 0.2. Technical
uncertainty was expected to be gradually reduced in the phases that followed. Otherwise, the
company would not enter the next phase. This was especially true at the clinical validation phase,
which involved large-scale clinical trials that incurred much higher development costs. Kappa
measures were also estimated in a similar way.

 License-in option

The company assumed that, compared to the continue option, the license-in option could save 30%
of R&D cost, while improving performance levels of detectable genotype ranges and reliability (see
Table 5). However, due to some potential difficulties in technology transfer, there was also some
chances of reduced performance.

 Outsourcing option

The company considered outsourcing the clinical sample and validation studies because they
expected to leverage the clinical experience of a contract research organization (CRO) to improve the
104 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114

Table 4
Transition probabilities, R&D lead times and costs for the continue option.

Phase Performance state Duration Cost

Transition probability (yr) (US$)

Basic research Genotype 4 2 0 +2 +4

Prob. 0.20 0.20 0.20 0.20 0.20 1.0 0.440 M
Kappa 0.05 0 +0.05
Prob. 0.33 0.34 0.33

Product specifications & Genotype 4 2 0 +2 +4

analytical validation Prob. 0.10 0.20 0.40 0.20 0.10 2.5 0.846 M
Kappa 0.05 0 +0.05
Prob. 0.30 0.40 0.30

Clinical sample studies Genotype 4 2 0 +2 +4

Prob. 0.10 0.15 0.50 0.15 0.10 0.5 0.250 M
Kappa 0.05 0 +0.05
Prob. 0.27 0.46 0.27

Clinical evaluation & Genotype 4 2 0 +2 +4

validation studies Prob. 0.00 0.08 0.90 0.02 0.00 3.0 1.650 M
Kappa 0.05 0 +0.05
Prob. 0.10 0.80 0.10

Launch & – – – – – 10 1.900 M

commercialization

Table 5
Transition probabilities, R&D lead times and costs for the license-in option.

Phase Performance state Duration (yr) Cost (US$)

Transition probability

Basic research Genotype 4 2 0 +2 +4

Prob. 0.10 0.15 0.30 0.25 0.20 1.0 0.352 M
Kappa 0.05 0 +0.05
Prob. 0.250 0.400 0.350

reliability of cancer screening. The company assumed that outsourcing might slightly improve kappa
measures, while reducing development costs by 10–15%. Table 6 shows the details for outsourcing
options at different stages.

 Switching option

The company assumed that using a new low-cost platform could reduce the cost to $25, thus
increasing consumer adoption by 20%. However, the new platform was still under development and
the company estimated that it would increase R&D costs by 15% to achieve the same performance with
the existing platform. Table 7 shows details of the switching option.

 Expansion option

The company considered improving the level of product performance to satisfy market
requirements through 20–30% increases in development budgeting in phases 2 and 3. The improved
effort in phase 2 was focused on the number of detectable genotypes (see Table 8).
The proposed methodology was implemented by MS Excel with VBA on a PC platform. In addition,
the ordinal least square regression method (Hair et al., 2009) was used to estimate the two main
parameters of the Bass diffusion model using the global disease diagnosis market data for microarray
biochip (BCC Research, 2008). The coefficients of innovation and imitation were estimated to be
0.0131 and 0.2235, respectively.
 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114 105

Table 6
Transition probabilities, R&D lead times and costs for the outsourcing option.

Phase Performance state Time (yr) Cost (US$M)

Transition probability

Clinical sample studies Genotype 4 2 0 +2 +4

Prob. 0.10 0.15 0.50 0.15 0.10 0.5 0.225 M
Kappa 0.05 0 +0.05
Prob. 0.20 0.50 0.30

Clinical evaluation & Genotype 4 2 0 +2 +4

validation studies Prob. 0.00 0.08 0.90 0.02 0.00 3 1.403 M
Kappa 0.05 0 +0.05
Prob. 0.05 0.80 0.15

Table 7
Transition probabilities, R&D lead times and costs for the switching option.

Phase Performance state Time (yr) Cost (US$)

Transition probability

Product specifications & Genotype 4 2 0 +2 +4

analytical validation Prob. 0.10 0.20 0.40 0.20 0.10 3.5 0.973 M
Kappa 0.05 0 +0.05
Prob. 0.30 0.40 0.30

Table 8
Transition probabilities, R&D lead times and costs for expansion option.

Phase Performance state Time (yr) Cost (US$M)

Transition probability

Product specifications & Genotype 2 0 2 +4 +6

analytical validation Prob. 0.150 0.300 0.350 0.150 0.050 2.5 1.100 M
Kappa 0.05 0 +0.05
Prob. 0.300 0.400 0.300

Clinical sample studies Genotype 4 2 0 +2 +4

Prob. 0.050 0.150 0.350 0.350 0.100 0.5 0.300 M
Kappa 0.05 0 +0.05
Prob. 0.250 0.500 0.250

Experimental results, discussion, and managerial implications

Project evaluation and optimal decisions

Firstly, we show the benefits of adopting managerial flexibility in terms of expected market payoff.
If there was no managerial flexibility considered in the R&D process, then the expected market payoff
was 0.39 M and the investment would be abandoned. However, as the concept of managerial
flexibility was applied, the expected market payoff was improved up to 0.72 M and the option value
was about 1.10 M. Fig. 3 shows that market payoffs regarding different levels of product performance
were all non-negative, because the project with a poor performance state was abandoned in a certain
R&D stage. This result validates the benefits of embedding managerial flexibility into R&D investment.
Secondly, optimal managerial options for all phases were suggested to decision makers, as shown
in Figs. A1–A7. For example, in phase 2, which involved several critical decisions in the early phase of
biochip development, four types of managerial options were considered (see Fig. A1). When the
expected product performance was in the ‘‘acceptable’’ area, wherein the expected number of
106 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114

Fig. 3. Market payoffs for different product performance levels.

genotypes and the kappa value were within the adequate performance levels, then the continue
option that followed the predetermined allocated resources and schedule was chosen because the
performance levels still could be further improved in latter R&D stages.
When the expected performance was in the ‘‘marginal’’ area, wherein the received product
performance might not be sufficient to fulfill future market requirements, either the expansion option
or the switching option was recommended, depending on whether or not the existing biochip
platform could be further improved to satisfy future market requirements by making extra R&D
investment. If the kappa value was worse than the mean market requirement, then the expansion
option was selected to further improve the performance levels, especially on the range of detectable
genotypes that would cause cervical cancers. Otherwise, the switching option was recommended for
applying a new low-cost platform that could offer a lower product price to attract more biochip
adopters and increase sales. Furthermore, when the product performance was in the ‘‘rejection’’ zone,
wherein the obtained performance levels were poor and it was impossible to make sufficient
improvement in latter R&D stages to satisfy market requirements, the abandonment option was
suggested to prevent an ineffective investment. Similarly, the optimal decisions for the rest of review
stages are shown in Figs. A2–A7. Note that Figs. A5–A7 show optimal decisions for phases 3–5, after
the switching option had been chosen at phase 2.

The value of flexibility and market uncertainty

Since market requirements may be difficult to estimate accurately, it is worth considering the
impact of its variability on the value of managerial flexibility. Fig. 4 shows market payoffs with
managerial flexibility regarding different levels of variability in market requirements. No strict
relationship was found between variability of market requirements and market payoffs. In addition,
Fig. 5 shows that the option value strictly decreased with increased variability of market
requirements. This is because the option value diminishes and the decision with managerial
flexibility is similar to the decision with the NPV criterion (Huchzermeier and Loch, 2001; Wang and
Yang, 2012). This finding indicates that, when high market uncertainty exists, incorporating
managerial flexibility still cannot avoid the risk of failing to match market requirements. The company
still needs to put more effort into reducing market uncertainty by collecting more reliable market
information or it needs to wait until the market trend becomes clear in order to fully explore the
benefits of managerial flexibility. In addition, the company also needed to enhance its technical
capability to increase its ability to fulfill market requirements.

Robustness of investment decisions

We are also interested in determining the robustness of obtained solutions. Taguchi’s robust
design method was applied to evaluate and select the decision with less sensitivity to sources of
 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114 107

Fig. 4. Market payoffs regarding different levels of variability in market requirements.

Fig. 5. Option values regarding different levels of variability in market requirements.

uncontrollable noise factors, such as market uncertainty, with a minimum number of experimental
runs (Phadke, 1989). It has been applied to various industrial fields, such as electronics, automobiles,
biotechnology, and telecommunications, to improve the desired performance of the product while
minimizing the effect of noise. We apply the concept of robust design to evaluate and determine the
robustness of R&D planning decisions under uncertainty. The basic concept of robust design is
outlined below.
In the product design, a number of variables can affect the performance of a product and they can
be classified into control factors, noise factors, and performance metrics. Control factors are decision
variables to be varied in a controlled manner during the experiment, while noise factors are variables
that cannot be explicitly controlled during the manufacturing and operation of the product.
Performance metrics are the product specifications or target performance levels that are expected to
be achieved during the experiments. The concept of robust design is simple. For a given target
performance, many combinations of parameter values may be possible to yield the desired result.
Some combinations, however, may be more sensitive to uncontrollable variation than others. This
variation is typically modeled by noise factors. Signal-to-noise ratio (S/N ratio), which is the ratio of
the signal (mean) over the noise (variability), is used to measure robustness. The larger the S/N ratio,
the more robust the performance. The purpose of robust design is to use an experimental approach in
order to choose the combination of parameter values that maximizes the S/N ratio.
For the illustrative purpose, the decision at review stage t = 1 with the mean performance levels,
genotype = 25 and kappa = 0.7, was used. The current optimal decision was to select the expansion
108 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114

option. We also evaluated the decision robustness with respect to four uncertain parameters of the
Bass model, i.e., a, b, m, and g. Three levels, including maximum, minimum, and current value, were
chosen for each parameter (see Table 9). Since there are have four parameters and three levels, the L4
orthogonal array was selected and nine different experiments were conducted to examine decision
robustness. The S/N ratio was calculated for each experiment to determine the effect each parameter
has on market payoff. Since the objective of the proposed approach was to maximize the market
payoff, the larger-the-better S/N ratio was used to evaluate the robustness of decision i (Phadke, 1989):
2 3
1 Xn
1
SN i ¼ 10 log10 4 5 (4)
n j¼1 y2i j

where n is number of trials and yij is the performance characteristic of decision i in trial j. Table 9 shows
the result of the robustness analysis, including the potential market payoff regarding each of the nine
possible market scenario and the S/N ratio for each of the four decision options. Among the four
decision options, the expansion option which had the largest S/N ratio is recommended to hedge
against the market uncertainty.
In addition, we found that though the switching option that applied the new low-cost platform
might attract more adopters, the unit profit margin was low and the increased market size was not
able to compensate for the losses. We also conducted sensitivity analysis to examine in what situation
the expansion option is replaced by the switching option as the optimal decision regarding other key
factors, such as R&D costs, product price, and user adoption level. The result is shown in Table 10. For
example, in the current setting, the switching option that intends to develop a new low-cost
technological platform will increase R&D cost by 15% (0.973 M). From the sensitivity analysis, we
found that as the increased R&D cost is varied within 10–20%, the expansion option still remains as the
optimal decision. However, for the other two influential factors, new product price and user adoption
level, the expansion option will not remain as the optimal decision, as the new product price is varied

Table 9
Experimental results of robustness analysis for different options at t = 1.

Experiment No. 1 2 3 4 5 6 7 8 9 S/N ratio

Noise factors
a 0.0118 0.0118 0.0118 0.0131 0.0131 0.0131 0.0144 0.0144 0.0144
b 0.2012 0.2235 0.2459 0.2012 0.2235 0.2459 0.2012 0.2235 0.2459
m 1.6 2.4 3.2 1.6 2.4 3.2 1.6 2.4 3.2
g 0.3 0.4 0.5 0.3 0.4 0.5 0.3 0.4 0.5

Decision options
Continue $0.00 $1.71 $6.36 $0.02 $2.06 $7.06 $0.05 $2.36 $7.73 94.13
Abandon $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 0.00
Switching $0.00 $1.91 $7.17 $0.04 $2.22 $7.82 $0.02 $2.54 $8.38 96.57
Expansion $0.00 $2.00 $6.97 $0.03 $2.38 $7.71 $0.08 $2.75 $8.39 98.99

Table 10
Sensitivity analysis for the expansion and switching options at stage t = 1.

Key factors for switching option Current value Possible variation Optimal decision

R&D cost 0.973 M [0.930 M, 1.2 M] Expansion

(+15%) (+10% to +20%)

Product price for the new platform $25 [$20, $30] $20 to $25: Expansion
(50%) (+60% to 40%) $26 to $30: Switching

User adoption level 2.88 M [2.76 M, 3.0 M] 2.76 M to 2.91 M: Expansion

(+20%) (+15% to +25%) 2.92 M to 3.0 M: Switching
 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114 109

from $26 to $30 or the user adoption level is varied from 2.93 M to 3.0 M. Therefore, decision makers
should be careful to examine the possible ranges of both factors before making the final decision.

Managerial implications
This study also offers several managerial implications. First, although the proposed proactive R&D
planning methodology is able to improve the value of an R&D project under uncertainty, higher
market uncertainty will still reduce the option value, decreasing the attractiveness of the proposed
methodology. Therefore, it is imperative for technology-based firms to improve their market research
capability (e.g., estimating market requirements from the target market segment) in order to take full
advantages of the real options-based R&D planning methodology.
Second, since emerging technologies usually involve great technological and market uncertainties,
it is not easy to accurately estimate parameter values for the developed project valuation model in the
opportunity evaluation stage. In order to make appropriate investment decisions based on the
valuation model, R&D managers need to be convinced that the parameter values are sufficiently
reliable (Thorn et al., 2011) or that the selected decisions have enough robustness over the noise
factors. Therefore, the developed opportunity evaluation model needs to be used carefully in order to
avoid making erroneous decisions. In addition, it is worth noting that the value of real options analysis
is in the implications for project design rather than in the actual planning evaluation (Bowman and
Moskowitz, 2001; Adner and Levinthal, 2004). Technology-based firms are encouraged to proactively
explore market and technological uncertainties and managerial options to overcome them in the
opportunity identification and development stages. It is also important to regularly reassess the
project for unforeseen uncertainties and deviations from the original plan that may occur as a result of
unexpected market conditions or emerging technological challenges (Sommer et al., 2009).
Third, to avoid over-committing to a technological opportunity, the firm can also break up the
investment into a series of smaller sequential R&D projects that can gradually explore and solve
technological challenges, while waiting for market uncertainty to be resolved. Therefore, major
commitments are made only if circumstances are favorable (McGrath, 1997). Some small to medium
technology-based firms hesitate to invest in high-risk projects, even though such projects could lead
to high profits. Real-options-based R&D planning can give firms confidence they need to pursue future
growth opportunities because this methodology maximizes market payoffs while reducing downside
risks.
Finally, it is difficult to for managers to identify the latent uncertainties and managerial options
within their firms. Some creative problem solving methods, such as the cross-functional team
approach and brainstorming techniques, can be used to identify opportunities, recognize
uncertainties, and determine managerial options. The quality of teamwork in the early stages of a
project has a profound impact on later project performance (Hoegl et al., 2004).

Conclusions

This paper developed a real options approach for R&D planning to maximize upside potentials,
while minimizing downside risks. The proposed methodology identified potential opportunities,
developed an R&D plan with managerial flexibility, and then evaluated the value of the opportunity
with the developed real options analysis model. The proposed methodology was illustrated with a
biochip R&D project for cervical cancer screening. The obtained results show that the developed R&D
planning framework can improve the expected market payoff of an R&D project through fully
exploring contingent uncertainties and managerial options to overcome those uncertainties. The
proposed methodology also generates optimal managerial options regarding various scenarios for
product performance that can help R&D managers to improve their R&D investment decisions to
capture emerging technological opportunities under uncertainty. In addition, to fully utilize the power
of managerial flexibility, it is important for an R&D organization to develop its market research
capability to reduce market uncertainty and to enhance its technical capability for increasing the
likelihood of satisfying uncertain market requirements. Finally, Taguchi’s robust design method was
applied to evaluate and select the decision that was more robust under market uncertainty.
110 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114

The contributions of this paper are three-fold. Firstly, the methodology developed here is able to
assist decision makers in linking R&D planning with market needs, while avoiding the major weakness
of technology push without proper consideration of whether a user need is satisfied. Next, it also
encourages R&D organizations to proactively recognize contingent uncertainties along with exploring
technology opportunities and to identify possible managerial options to maintain future
opportunities, while avoiding downside risks. Finally, the developed real options model, integrated
with an innovation diffusion model, can help R&D managers to evaluate and select optimal managerial
options to maximize market payoffs under different demand structures.
Since market and technological uncertainties can have great impacts on the quality of R&D
planning decisions, it is worthwhile to apply alternative methodologies to improve R&D decision
making under uncertainty. In addition to Taguchi’s robust design method and sensitivity analysis used
in this paper, further research will try to integrate the developed methodology with fuzzy sets theory
(Zadeh, 1978; Collan et al., 2009; You et al., 2012) or Monte-Carlo simulation methods (Willigers and
Hansen, 2008; Mitchell et al., 2010) to fully explore the impact of uncertainty on R&D investment
decisions.

Acknowledgements

This research is supported by grant nos. NSC 99-2221-E-005-050 and NSC 100-2221-E-005-072
from the National Science Council of the Republic of China. We also want to thank anonymous referees
for their constructive comments on this paper.

Appendix A. Optimal decisions for different performance levels at different review stages

Figs. A1–A4 show optimal decisions for different performance states from stage 2 to stage 5.
Figs. A5–A7 show optimal decisions for different performance states from stage 3 to stage 5, if the
switching option had been chosen.

Fig. A1. Optimal decisions for different performance levels at t = 1.

Fig. A2. Optimal decisions for different performance levels at t = 2.

 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114 111

Fig. A3. Optimal decisions for different performance levels at t = 3.

Fig. A4. Optimal decisions for different performance levels at t = 4.

Fig. A5. Optimal decisions with the switching option for different performance levels at t = 2.
112 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114

Fig. A6. Optimal decisions with the switching option for different performance levels at t = 3.

Fig. A7. Optimal decisions with the switching option for different performance levels at t = 4.

References

Adner, R., Levinthal, D.A., 2004. What is not a real option: considering boundaries for the application of real options to business
strategy. Acad. Manage. Rev. 29 (1) 74–85.
Bass, F.M., 1969. A new product growth model for consumer durables. Manage. Sci. 15, 215–227.
BCC Research, 2008. DNA diagnostics: technologies and global markets. BCC Research. Available from http://www.bccresearch.
com/report/BIO063A.html.
Bowman, E.H., Moskowitz, G.T., 2001. Real options analysis and strategic decision making. Organ. Sci. 12 (6) 772–777.
Cassimon, D., De Backer, M., Engelen, P.J., Van Wouwe, M., Yordanov, V., 2011. Incorporating technical risk in compound real
option models to value a pharmaceutical R&D licensing opportunity. Res. Policy 40 (9) 1200–1216.
Chang-Kyo, S., Eui-Ho, S., Kwang-Churn, B., 1994. Prioritizing telecommunications technologies for long-range R&D planning to
the year 2006. IEEE Trans. Eng. Manage. 41 (3) 264–275.
Chen, H., Ho, J.C., Kocaoglu, D.F., 2009. A strategic technology planning framework: a case of Taiwan’s semiconductor foundry
industry. IEEE Trans. Eng. Manage. 56 (1) 4–15.
Cho, K.-T., Kwon, C.-S., 2004. Hierarchies with dependence of technological alternatives: a cross-impact hierarchy process. Eur. J.
Oper. Res. 156 (2) 420–432.
Collan, M., Fullér, R., Mezei, J., 2009. A fuzzy pay-off method for real option valuation. J. Appl. Math. Decis. Sci.. (Article ID:
238196). Available from: http://dx.doi.org/10.1155/2009/238196.
Cooper, R.G., 2000. Product Leadership: Creating and Launching Superior New Products. Perseus Books Group, Cambridge, MA.
Cooper, R.G., 2006. Managing technology development projects. Res. Technol. Manage. 49 (6) 23–31.
Cuzick, J., 2010. Long-term cervical cancer prevention strategies across the globe. Gynecol. Oncol. 117 (2 Suppl.) S11–S14.
De Reyck, B., Degraeve, Z., Vandenborre, R., 2008. Project options valuation with net present value and decision tree analysis.
Eur. J. Oper. Res. 184 (1) 341–355.
 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114 113

Dixit, A.K., Pindyck, R.S., 1994. Investment Under Uncertainty. Princeton University Press, Princeton, NJ.
Drew, S.A.W., 2006. Building technology foresight: using scenarios to embrace innovation. Eur. J. Innovat. Manage. 9 (3)
241–257.
Farrukh, C., Phaal, R., Probert, D., 2003. Technology roadmapping: linking technology resources into business planning. Int. J.
Technol. Manage. 26 (1) 2–19.
Figueira, J., Greco, S., Ehrgott, M., 2005. Multiple Criteria Decision Analysis: State of the Art Surveys. Springer-Verlag, Boston.
Fleiss, J.L., Levin, B., Paik, M.C., 2003. Statistical Methods for Rates and Proportions, third ed. Wiley, New York.
Gerdsri, N., Kocaoglu, D.F., 2007. Applying the analytic hierarchy process (AHP) to build a strategic framework for technology
roadmapping. Math. Comput. Model. 46 (7/8) 1071–1080.
Gerdsri, N., Vatananan, R.S., Dansamasatid, S., 2009. Dealing with the dynamics of technology roadmapping implementation: a
case study. Technol. Forecast. Soc. 76 (1) 50–60.
Green, P.E., Srinivasan, V., 1990. Conjoint analysis in marketing: new developments with implications for research and practice.
J. Market. 54 (4) 3–19.
Hair, J.F., Black, W.C., Babin, B.J., Anderson, R.E., 2009. Multivariate Data Analysis, seventh ed. Prentice Hall, New York.
Hoegl, M., Weinkauf, K., Gemuenden, H.G., 2004. Interteam coordination, project commitment, and teamwork in multiteam
R&D projects: a longitudinal study. Organ. Sci. 15 (1) 38–55.
Huchzermeier, A., Loch, C.H., 2001. Project management under risk: using the real options approach to evaluate flexibility in
R&D. Manage. Sci. 47 (1) 85–101.
Hunt, F., Mitchell, R., Phaal, R., Probert, D.R., 2004. Early valuation of technology: real options, hybrid models and beyond. J. Soc.
Instrum. Control Eng. Jpn. 43 (10) 730–735.
Jang, Y.S., Lee, D.J., Oh, H.S., 2013. Evaluation of new and renewable energy technologies in Korea using real options. Int. J. Energy
Res. 37 (13) 1645–1656.
Kauffman, R.J., Li, X., 2005. Technology competition and optimal investment timing: a real options perspective. IEEE Trans. Eng.
Manage. 52 (1) 15.
Keizer, J.A., Vos, J.-P., Halman, J.I.M., 2005. Risks in new product development: devising a reference tool. R&D Manage. 35 (3)
297–309.
Krychowski, C., Quelin, B.V., 2010. Real options and strategic investment decisions: can they be of use to scholars? Acad.
Manage. Perspect. 24 (2) 65–78.
Lee, S., Kang, S., Park, Y., Park, Y., 2007. Technology roadmapping for R&D planning: the case of the Korean parts and materials
industry. Technovation 27 (8) 433–445.
Liao, C.C., Wang, H.Y., Lin, R.S., Hsieh, C.Y., Sung, F.C., 2006. Addressing Taiwan’s high incidence of cervical cancer: factors
associated with the nation’s low compliance with Papanicolaou screening in Taiwan. Public Health 120 (12) 1170–1176.
Lo Nigro, G., Morreale, A., Enea, G., 2014. Open innovation: a real option to restore value to the biopharmaceutical R&D. Int. J.
Prod. Econ. 149, 183–193.
Loch, C.H., Bode-Greuel, K., 2001. Evaluating growth options as sources of value for pharmaceutical research projects. R&D
Manage. 31 (2) 231–248.
MacMillan, I.C., McGrath, R.G., 2002. Crafting R&D project portfolios. Res. Technol. Manage. 45 (5) 48–59.
Mahajan, V., Muller, E., Bass, F.M., 1990. New product diffusion models in marketing: a review and directions for research. J.
Marketing Res. 54, 1–26.
Mahajan, V., Muller, E., Bass, F.M., 1995. Diffusion of new products: empirical generalizations and managerial uses. Market. Sci.
14 (3 Suppl.) G79–G88.
Martzoukos, S.H., Zacharias, E., 2013. Real option games with R&D and learning spillovers. OMEGA Int. J. Manage. Sci. 41 (2)
236–249.
McGrath, R.G., 1997. A real options logic for initiating technology positioning investments. Acad. Manage. Rev. 22 (4) 974–996.
McGrath, R.G., Nerkar, A., 2004. Real options reasoning and a new look at the R&D investment strategies of pharmaceutical
firms. Strategic Manage. J. 25 (1) 1–21.
Melachrinoudis, E., Rice, K., 1991. The prioritization of technologies in a research laboratory. IEEE Trans. Eng. Manage. 38 (3)
269–278.
Mitchell, R., Hunt, F., Probert, D., 2010. Valuing and comparing small portfolios. Res. Technol. Manage. 53 (2) 43–54.
National Cancer Institute, 2013. Fact sheets. Available from: http://www.cancer.gov/cancertopics/factsheet.
Petrick, I.J., Echols, A.E., 2004. Technology roadmapping in review: a tool for making sustainable new product development
decisions. Technol. Forecast. Soc. 71 (1–2) 81–100.
Phaal, R., Farrukh, C.J.P., Probert, D.R., 2004. Technology roadmapping – a planning framework for evolution and revolution.
Technol. Anal. Strategic Manage. 71 (1) 5–26.
Phadke, M.S., 1989. Quality Engineering Using Robust Design. Prentice Hall, Englewood Cliff, NJ.
Piachaud, B.S., 2002. Outsourcing in the pharmaceutical manufacturing process: an examination of the CRO experience.
Technovation 22 (2) 81–90.
Pich, M.T., Loch, C.H., De Meyer, A., 2002. On uncertainty, ambiguity, and complexity in project management. Manage. Sci. 48 (8)
1008–1023.
Prabhu, T.R., Vizayakumar, K., 2001. Technology choice using FHDM: a case of iron-making technology. IEEE Trans. Eng. Manage.
48 (2) 209–222.
Rogers, M.J., Gupta, A., Maranas, C.D., 2002. Real options based analysis of optimal pharmaceutical research and development
portfolios. Ind. Eng. Chem. Res. 41 (25) 6607–6620.
Salo, A., Gustafsson, T., Ramanathan, R., 2003. Multicriteria methods for technology foresight. J. Forecast. 22, 235–255.
Santiago, L.P., Bifano, T.G., 2005. Management of R&D projects under uncertainty: a multidimensional approach to managerial
flexibility. IEEE Trans. Eng. Manage. 52 (2) 269–280.
Saritas, O., Aylen, J., 2010. Using scenarios for roadmapping: the case of clean production. Technol. Forecast. Soc. 77 (7)
1061–1075.
Schneider, M., Tejeda, M., Dondi, G., Herzog, F., Keel, S., Geering, H., 2008. Making real options work for practitioners: a generic
model for valuing R&D projects. R&D Manage. 38 (1) 85–106.
114 J. Wang et al. / Journal of Engineering and Technology Management 35 (2015) 93–114

Scott, G.M., 2001. Strategic planning for technology products. R&D Manage. 31 (1) 15–26.
Sivarama Prasad, A.V., Somasekhara, N., 1990. The analytic hierarchy process for choice of technologies: an application. Technol.
Forecast. Soc. 38 (2) 151–158.
Smit, H., Trigeorgis, L., 2006. Strategic planning: valuing and managing portfolios of real options. R&D Manage. 36 (4) 403–419.
Smith, J.S., Lindsay, L., Hoots, B., Keys, J., Franceschi, S., Winer, R., Clifford, G.M., 2007. Human papillomavirus type distribution in
invasive cervical cancer and high-grade cervical lesions: a meta-analysis update. Int. J. Cancer 121 (3) 621–632.
Sommer, S.C., Loch, C.H., Dong, J., 2009. Managing complexity and unforeseeable uncertainty in startup companies: an empirical
study. Organ. Sci. 20 (1) 118–133.
Song, Y.I., Lee, D.H., Lee, Y.G., Chung, Y.C., 2007. Managing uncertainty and ambiguity in frontier R&D projects: a Korean case
study. J. Eng. Technol. Manage. 24 (3) 231–250.
Strauss, J.D., Radnor, M., 2004. Roadmapping for dynamic and uncertain environments. Res. Technol. Manage. 47 (2) 51–57.
Thorn, V., Hunt, F., Mitchell, R., Probert, D., Phaal, R., 2011. Internal technology valuation: real world issues. Int. J. Technol.
Manage. 53 (2–4) 149–160.
Tipping, J.W., Zeffren, E., 1995. Assessing the value of your technology. Res. Technol. Manage. 38 (5) 22–39.
Trigeorgis, L., 1996. Real Options: Managerial Flexibility and Strategy in Resource Allocation. The MIT Press, Cambridge, MA.
van Bekkum, S., Pennings, E., Smit, H., 2009. A real options perspective on R&D portfolio diversification. Res. Policy 38 (7)
1150–1158.
Vassolo, R.S., Anand, J., Folta, T.B., 2004. Non-additivity in portfolios of exploration activities: a real options-based analysis of
equity alliances in biotechnology. Strategic Manage. J. 25 (11) 1045–1061.
Verganti, R., 1999. Planned flexibility: linking anticipation and reaction in product development projects. J. Prod. Innovat.
Manage. 16 (4) 363–376.
Walsh, S.T., 2004. Roadmapping a disruptive technology: a case study: the emerging microsystems and top-down nanosystems
industry. Technol. Forecast. Soc. 71 (1–2) 161–185.
Wang, J., Hwang, W.L., 2007. A fuzzy set approach for R&D portfolio selection using a real options valuation model. OMEGA Int. J.
Manage. Sci. 35 (3) 247–257.
Wang, J., Yang, C.-Y., 2012. Flexibility planning for managing R&D projects under risk. Int. J. Prod. Econ. 135 (2) 823–831.
Willigers, B.J.A., Hansen, T.L., 2008. Project valuation in the pharmaceutical industry: a comparison of least-squares Monte Carlo
real option valuation and conventional approaches. R&D Manage. 38 (5) 520–537.
You, C.J., Lee, C.K.M., Chen, S.L., Jiao, R.J., 2012. A real option theoretic fuzzy evaluation model for enterprise resource planning
investment. J. Eng. Technol. Manage. 29 (1) 47–61.
Zadeh, L.A., 1978. Fuzzy sets as a basis for a theory of possibility. Fuzzy Sets Syst. 1, 3–28.