An Examination of IT Initiative Portfolio Characteristics and Investment Allocation: A Computational Modeling and Simulation Approach

Transcription

1 A Examiatio of IT Iitiative Portfolio Characteristics ad Ivestmet Allocatio: A Computatioal Modelig ad Simulatio Approach Yu-Ju Tu Uiversity of Illiois at Urbaa-Champaig yujutu@illiois.edu Ramaath Subramayam Uiversity of Illiois at Urbaa-Champaig rsubrama@illiois.edu ABSTRACT Completed Research Paper Michael Shaw Uiversity of Illiois at Urbaa-Champaig mjshaw@illiois.edu Eric Larso Uiversity of Illiois at Urbaa-Champaig ecl@illiois.edu We advace the theory pertaiig to IT goverace ad portfolio maagemet. We specifically examie how the portfolio characteristics ca ifluece IT ivestmet allocatio. We propose two portfolio characteristics flexibility ad diversity. We defie the flexibility as the umber of IT ivestmet choice, ad the diversity as the dissimilarity of IT ivestmet choice. Usig the Mote Carlo simulatio ad real-world data, we fid that ) if a firm ivests i a set of IT iitiatives with higher flexibility, it will have greater potetiality to capture a superior IT ivestmet allocatio; ) whe a firm ivests i a set of IT iitiatives with lower flexibility, it will have greater potetiality to capture a superior IT ivestmet allocatio, if the set of IT iitiatives ivolves higher diversity. Our fidigs implicate that a firm ca exercise IT iitiative flexibility ad diversify IT iitiative choice to improve portfolio ivestmet allocatio. Keywords IT portfolio, IT ivestmet, IT goverace INTRODUCTION It is widely established that the performace of IT ivestmet ca vary sigificatly across firms. I this paper, we propose oe importat explaatio for such variatios i performace the portfolio characteristics. From our observatio i practice, we otice cosiderable variatios i sets of IT iitiatives across firms. If we view a set of IT iitiatives as a portfolio ad regard it as a target for a firm s IT ivestmet allocatio, we expect that the portfolio characteristic would make a great impact o a firm s IT ivestmet allocatio, which i tur would directly ifluece a firm s IT ivestmet performace. I this paper, we specifically examie how the portfolio characteristics ca ifluece IT ivestmet allocatio. We propose two portfolio characteristics flexibility ad diversity. I our defiitio, flexibility refers to the umber of IT ivestmet choice, ad diversify refers to the dissimilarity of IT ivestmet choice. We equate the IT ivestmet allocatio with the quality of the optimal IT ivestmet choice. We the desig ad develop a decisio support model that aims to select a best set of IT iitiatives as a firm s optimal IT ivestmet choice, accordig to a firm s IT ivestmet risk limit ad budget limit. Moreover, we differetiate betwee differet flexibility degrees by simulatig sets of divisible/idivisible IT iitiatives for IT ivestmet decisio, ad differetiate betwee differet diversity degrees by simulatig high/low dispersio degrees (σ) for sets of IT iitiatives values, risks, ad costs i the distributio. We use the simulated data combied with the real-word data. At first, we collect a data set of IT iitiatives from two busiess uits of a leadig compay i isurace ad fiace idustry. Next, we use the collected data set to calibrate the variatio degree of the simulated data. Specifically, we take the mathematical momets (i.e., mea ad stadard deviatio) of the realworld data ad feed them to the Mote Carlo simulatio to propagate large sets of data o the basis of the Gaussia distributio. Our experimet results show that if a firm ivests i a set of IT iitiatives with higher flexibility, it will have greater potetiality to capture a superior IT ivestmet allocatio. Moreover, oce if a firm ivests i a set of IT iitiatives with lower flexibility, it still ca have greater potetiality to capture a superior IT ivestmet allocatio, as log as the set of IT iitiatives ivolve higher diversity. I other words, our fidigs implicate that a firm should exercise the maagerial Proceedigs of the Nieteeth Americas Coferece o Iformatio Systems, Chicago, Illiois, August 5-7, 3.

2 flexibility embedded i its IT iitiatives ad diversify its IT iitiative choice i order to icrease its IT ivestmet allocatio efficiecy ad improve performace. We orgaize the remaider of this paper as follows. I sectio, we defie the portfolio characteristics followed by our propositios. I sectio 3, we outlie the developmet of our IT portfolio decisio support model. I sectio 4, we describe our method, experimet desig, ad data collectio. I sectio 5, we summarize our results ad fidigs. I sectio 6, we discuss the maagerial implicatios. I sectio 7, we coclude this paper with cotributios. PROPOSITIONS AND RELATED LITERATURE The Characteristic of a IT Iitiative Portfolio I this paper, we defie the flexibility as the umber of choices that a firm ca select from its set of IT iitiatives to decide its IT ivestmet. I prior related literature, maagerial flexibility is regarded as oe critical role i a firm s resource allocatio strategy (Trigeorgis, 996). I ecoomics, resource meas ay asset a firm ca purchase, lease, or produce for its ow use, such as labor, lad, etc. I a firm, the completio of a IT iitiative also requires the asset icludig skill, facility, etc. which is geerated from a combiatio of the ecoomic resources. I additio, the resource allocatio of a IT iitiative is determied by a firm s IT ivestmet decisio. Therefore, we reaso that if a firm ca have a plety of ivestmet choices, it will have higher flexibility to allocate its IT ivestmet, accordig to its strategic pla. I tur, this firm will have more potetialities to better capitalize valuable opportuities or mitigate potetial losses, cotributig to a better ivestmet performace. Moreover, we view each IT iitiative as a ivestmet target i a firm. A IT iitiative is served as a umbrella to cover the associated IT projects for the same strategic objectives as well as ivestmet purposes. I this paper, a IT iitiative specifically refers to a IT depedet strategic iitiative (Piccoli ad Ives, 5). Moreover, such a iitiative is defied as havig the idetifiable objectives that leads to sustaied improvemets or limits the erosios of a firm s competitive advatage. For example, a IT iitiative ca iclude buildig the iformatio systems or applicatios, such as ERP, CRM, etc., to geerate competitive advatage; these systems are ofte complicated to develop ad malleable to chage, makig the competitive imitatio slow, difficult, ad costly. I other words, a IT iitiative is depedet o the IT projects, sice it caot be implemeted util ecessary ad associated iformatio systems have bee successfully deployed. Not every IT iitiative ca provide may ivestmet choices. I practice, we observe that there are two types of IT iitiatives. Oe type is flexible ad the other is iflexible. I particular, the two types of IT iitiatives will provide very differet umbers of IT ivestmet choices. I a firm, most fiacial ivestmets are very flexible ad able to provide may choices. Hece, for example, a firm ca cosider choices of icreasig the amout of a stock ivestmet to capture greater ivestmet retur, or decreasig it i order to alleviate the potetial ivestmet loss. To IT ivestmet, however, may IT iitiatives ca oly provide few ivestmet choices. For example, if a IT iitiative is proposed to develop a iovative CRM system, a firm perhaps caot have ay choice, but oe full ivestmet choice, i order to capture ivestmet retur. This is primarily because such a ivestmet is ot very flexible, ad ay retur ca hardly be realized util the etire iitiative is completed. O the other had, if a IT iitiative is proposed for acquirig a Cloud-based CRM, the ivestmet i that IT iitiative ca be very flexible. Oe importat feature of cloud computig is the price elasticity (e.g., pay-as-you-go). This implies that such a ivestmet ca provide may choices based o a firm s demads. Moreover, certai web iitiatives (e.g., website developmet) ca also provide may ivestmet choices, sice they are able to be divided ito cycles ad thus beefit from the completio of early cycles. Therefore, we derive a propositio as follows. Propositio : If a firm ivests i a set of IT iitiatives with higher flexibility, it will have greater potetiality to capture a superior IT ivestmet allocatio. The diversity Characteristic of a IT Iitiative Portfolio We ow tur to ivestigatig the IT ivestmet allocatio strategy whe a firm has to ivest i a set of IT iitiatives with lower flexibility. I corporate ecoomics, diversificatio refers to the degree to which a firm is able to operate i lies of busiess. It is suggested that a ret-seekig firm ca diversify ad thus beefit from allocatig its excessive ecoomic resource more efficietly. For example, a firm ca icrease the diversity of its products ad beefit from the ecoomy of scope. Most importatly, the ratioale behid such a diversificatio strategy relies o a precoditio - the resource is ot very Proceedigs of the Nieteeth Americas Coferece o Iformatio Systems, Chicago, Illiois, August 5-7, 3.

3 flexible for use. For example, Perose (959) idicate that the idivisibility of resources preclude a firm from fidig the best way to use resource ad thus attai at a state of rest (i.e., a equilibrium positio where a firm ca hardly use its resources more profitably). Further, Teece (98) maitais that the resource idivisibility aloe ca seldom result i the resource use difficulty postulated by Perose, uless there is a frictio for the resource exteral trasfer. For example, if the resource has high liquidity ad thus marketable, a firm ca just sell the uused resource to the market. I such a situatio, the lik betwee a firm s diversificatio strategy ad the resource idivisibility ca seldom sustai. I short, corporate ecoomic strategists argue that diversificatio is helpful to overcomig the resource allocatio challeges posted by the resource lackig divisibility ad liquidity. I this paper, we defie the diversity as the dissimilarity of choices that a firm ca select from its set of IT iitiatives to decide its IT ivestmet portfolio. From a decisio-makig viewpoit, choice diversity implicates there are more opportuities to capture a better decisio. Accordig to the classical utility theory, the similarity betwee a pair of goods is cotiget o the margial utility associated with a set of goods cosumed. I other word, the dissimilarity betwee two goods ca be viewed as the degree to which the two goods are ot substitutable (Nehrig ad Puppe, ). Thus, if we view the goods as IT iitiatives, we ca reaso that a set of more diversified IT iitiatives ca provide more o-substitutable choices, ad thus we ca have a broader search rage for a better decisio. I additio, we observe that certai IT iitiatives are idivisible ad caot be sold to the market. Thus, if we regard a IT iitiative as the resource from the view of Perose ad Teece, we reaso that we should be able to apply the aalogous diversificatio idea to improvig a firm s IT ivestmet allocatio ad decisio makig. Not every two IT iitiatives ca preset the same ivestmet attributes, such as same cost, same retur, etc. As a result, ay two sets of IT iitiatives for a firm s ivestmet cosideratio will ivolve differet diversity degrees. For example, oe commo way adopted by may firms to classify IT iitiatives is to use the class of ru-the-busiess, grow-the-busiess, ad trasform-the-busiess. I a way, they respectively correspod to the class of low retur/risk, medium retur/risk, ad high retur/risk by the fiacial ivestmet classificatio. I prior IS literature, Weill ad Aral (6) also proposes a IT asset classificatio framework icludig trasactioal, iformatioal/strategic, ad ifrastructure classes, which also implicate that a set of IT iitiatives ca be differetiated by their dissimilarities. As a result, if a firm selects a set of IT iitiatives primarily from oly sigle class, we ca expect its IT ivestmet choice diversity aturally will be lower tha the choice diversify whe a firm selects a set of IT iitiatives from the three classes. Therefore, we derive a propositio as follows. Propositio : Whe a firm ivests i a set of IT iitiatives with lower flexibility, it will have greater potetiality to capture a superior IT ivestmet allocatio, if the set of IT iitiatives ivolves higher diversity. MODEL DESIGN AND DEVELOPMENT Our model aims to maximize the value of a set of selected IT iitiatives, give the accepted IT ivestmet risk ad budget limits. I other words, our selectio ratioale is grouded o the portfolio retur-risk balace ad retur-cost balace criteria. The cost (C), risk (R), ad value (V) are the mai decisio compoets followed by the risk limit ( R ) ad the budget limit ( C ). Accordigly, if P is a vector represetig a set of selected IT iitiatives, we ca derive a coceptual model as follows. Max V (P) () Subject to: R( P) R () C( P) C (3) (This model is for maximizig the value of a set of selected IT iitiatives, give the accepted IT ivestmet risk ad budget limits) Variable xi vi Defiitio IT iitiative decisio variable The NPV of IT iitiative i Proceedigs of the Nieteeth Americas Coferece o Iformatio Systems, Chicago, Illiois, August 5-7, 3. 3

4 ci ri C R P The cost of IT iitiative i The risk score of IT iitiative i The budget limit for the selected set of IT iitiatives The risk limit for the selected set of IT iitiatives A vector represetig a set of selected IT iitiatives Defiitios of Mathematical Notatio i the Model The Metrics ad modelig processes For IT iitiative cost estimatio, we adopt a itegrated approach. This itegrated cost will icluded the programmig costs, hardware/software costs, ad maagemet costs. Accordigly, the total cost of a set of selected IT iitiatives is as follows: C ( P ) = c x i i (4) For IT iitiative risk estimatio, we use a itegrated risk scorig approach. This itegrated risk will iclude the ecoomic risk factor (e.g., size, resource limits, etc.), the orgaizatioal risk factor (e.g., procedure chage, orgaizatioal maagemet chage, etc.), ad techological risk factor (e.g., lack of expertise, techology covexity, etc.) (McFarla, 98; Wallace ad Keil, 4). Thus, the total risk of a set of selected IT iitiatives is as follows: R( P) = r i x (5) i Regardig IT imitative value evaluatio, we apply a fiacial modelig approach - NPV. I prior IT portfolio related literature, the fiacial value is oe most importat value cosidered i a firm s ivestmet decisio (Kumar et al., 8). Hece, the total value of a set of selected IT iitiatives is as follows: V P ) = ( v x i i (6) As a result, by cosiderig fuctios (4)(5)(6), we ca derive a complete model as follows. Max V ( P ) = v i x i (7) Subject to: = i i R( P) r x R (8) = i i C( P) c x C (9) A use example is provided i appedix A Proceedigs of the Nieteeth Americas Coferece o Iformatio Systems, Chicago, Illiois, August 5-7, 3. 4

5 METHOD I this paper, we employ a itegrated research method that icorporates computatioal modelig ad data simulatio. I prior IS literature, the computatioal method is ofte adopted o the coditio that the sufficiet empirical evidece is too costly to collect (Gree at al., ; Na, ; Piramuthu ad Shaw, 998; Sikora ad Shaw, 988; Tu et al., 9). I additio, we use a certai amout of real-world IT iitiative data to iitialize the simulatio, so that we ca make our results more reliable ad more grouded i practice. The experimet desig I this paper, we seek to explore how the two portfolio characteristics flexibility ad diversity ca ifluece IT ivestmet allocatio. We desig two computatioal experimets for examiig our two propositios. I experimet, we aim to observe whether a set of IT iitiatives with high/low flexibility will ifluece the frequecy of geeratig a optimal ad superior portfolio choice from the proposed model (i.e. a choice of maximizig portfolio value, give accepted risk ad cost). We desig two scearios. I sceario, we assume that a set of IT iitiatives is composed of a set of divisible IT projects, which refers to a set of IT iitiatives with high flexibility. I sceario, we assume that a set of IT iitiatives is composed of a set of idivisible IT projects, which refers to a set of IT iitiatives with low flexibility. Accordigly, i the sceario, we set the real value for the decisio variable ( x ), ad i sceario we set the biary value for it. I experimet, we will i observe whether a set of iflexible IT iitiatives with high/low diversity will ifluece the frequecy of geeratig a optimal portfolio choice from the proposed model. We use three scearios to reflect three diversity degrees, ) the less-diversified, ) the medium-diversified, ad 3) the more-diversified sc. We operatioalize the diversity degree by varyig the dispersio degrees (σ) of IT imitative value, risk ad cost i the distributio. Moreover, we use the value/risk ratio to decide the superiority of IT ivestmet allocatio. This ratio is aalogous to the sharp ratio (Sharpe 994), which is a commo portfolio performace idex i fiace. First, we defie a threshold ratio, the ratio of the origial set of IT iitiatives before selectio. Next, we compute the ratio of the optimal IT ivestmet choice selected by the proposed optimizatio model. Fially, oly if the ratio of the optimal choice is greater tha that of the threshold ratio, we cout it as a superior IT ivestmet allocatio. Besides, we use differet budget cosideratios ad potetial risk tolerace levels. First, we radomly set the budget limits. A portfolio choice will be cosidered oly if its total cost is uder the budget. Secod, we assume that a firm will ted to gover its IT ivestmet risk ad thus we simulate a firm s potetial accepted risk levels as high, medium, ad low. If a firm s accepted risk for its IT ivestmet is low, ay portfolio choice whose risk is greater tha the 5% of the maximum possible portfolio risk will ot be take as superior; if a firm s accepted risk for its IT ivestmet is medium, ay portfolio choice whose risk is greater tha the 5% of the maximum possible portfolio risk will ot be take as superior; eve if a firm s accepted risk for its IT ivestmet is high, ay portfolio choice whose risk is greater tha the 75% of the maximum possible portfolio risk will still ot be take as superior. Data collectio For experimet iput, we use the simulated data combied with the real-word data. At first, we collect a data set of IT iitiatives from the busiess uits of a leadig compay i isurace ad fiace idustry. Each IT iitiative icludes value, risk, ad cost iformatio; the value iformatio is estimated by the compay s fiacial experts, ad risk ad cost iformatio are estimated by the maager of the busiess uit that proposes the IT iitiative for a strategic purpose. This compay also has a MIS uit for supportig the compay s geeral IS/IT fuctios. Next, we use the collected data set to calibrate the variatio degree of the simulated IT iitiatives. Specifically, we calculate the mathematical momets3 (i.e., mea ad stadard deviatio) of the IT iitiatives ad feed them to the Mote Carlo simulatio to propagate data, accordig to the ormal distributio (Gaussia distributio). We use a persoal computer (Itel Duo.8G CPU with 8GB RAM) to implemet the simulatio. RESULT The frequecy of geeratig a superior IT ivestmet choice for allocatio i the experimet ad experimet are show i table ad table. We ru 3 differet data sets for each of the iteratios, ad each set is composed of 3 IT iitiatives. The details of diversity scearios are listed i appedix B 3 The details of momets of real-world data are listed i appedix Proceedigs of the Nieteeth Americas Coferece o Iformatio Systems, Chicago, Illiois, August 5-7, 3. 5

6 I the bottom lies, we compute the overall average frequecies, their stadard deviatios i the paretheses, ad the overall average frequecy. Low accepted risk Medium accepted risk accepted risk Lower Lower Lower Iteratio Iteratio Iteratio Iteratio Iteratio Iteratio Iteratio Iteratio Iteratio Iteratio Sub Avg.(Std.). (.).47 (.3). (.).6578 (.63). (.) Overall =., Low Flexibly =.583 Table. Superior IT Ivestmet Choice Selectio Frequecy for Experimet Result.67 (.9) (Note: This result table is for examiig our propositio : if a firm ivests i a set of IT iitiatives with higher flexibility, it will have greater potetiality to capture a superior IT ivestmet allocatio. The Low, Medium, ad accepted risks refer to a firm s differet possible accepted risks toward IT ivestmet). Low accepted risk Medium accepted risk accepted risk Diversity Low Diversity Diversity Low Diversity Diversity Low Diversity Iteratio Iteratio Iteratio Iteratio Iteratio Iteratio Iteratio Iteratio Iteratio Iteratio Sub Avg. (Std.).633 (.3).933 (.4).7967 (.49).4333 (.).867 (.63) Overall Diversity =.7456, ad Low Diversity =.3678 Table. Superior IT Ivestmet Choice Selectio Frequecy for Experimet Result.3767 (.74) (Note: This result table is for examiig our propositio : whe a firm ivests i a set of IT iitiatives with lower flexibility, it will have greater potetiality to capture a superior IT ivestmet allocatio, if the set of IT iitiatives ivolves higher diversity. The Low, Medium, ad accepted risks refer to a firm s differet possible accepted risks toward IT ivestmet) Proceedigs of the Nieteeth Americas Coferece o Iformatio Systems, Chicago, Illiois, August 5-7, 3. 6

7 Summary of Fidigs The experimet results (table ) support our propositio. We propose that if a firm ivests i a set of IT iitiatives with higher flexibility, it will have greater potetiality to capture a superior IT ivestmet allocatio. As see i our summary of fidig (figure ), we ca fid that if a set of IT iitiatives ivolve high flexibility, the frequecy of capturig a superior IT ivestmet choice will be clearly higher tha the frequecy whe a set of IT iitiatives ivolve low flexibility. Figure. Summary of Experimet Result The experimet results (table ) also support our propositio. We propose that whe a firm ivests i a set of IT iitiatives with lower flexibility, it will have greater potetiality to capture a superior IT ivestmet allocatio, if the set of IT iitiatives ivolves higher diversity. As see i our summary of fidig (figure ), we ca fid that if a set of IT iitiatives ivolve low flexibility but high diversity, the frequecy of capturig a superior IT ivestmet choice will be clearly higher tha the frequecy whe a set of IT iitiatives ivolve low flexibility ad diversity. Figure. Summary of Experimet Result DISCUSSION Exercisig IT Iitiative flexibility to improve portfolio ivestmet allocatio From the corporate ecoomic perspective, a firm s value creatio is critically cotiget o both the strategy to allocate resources ad the strategy to evaluate ivestmet alteratives (Trigeorgis, 996). I prior IS literature, there is a stream of studies ivestigatig how to use the fiacial real-optios approaches to evaluate the alteratives (optios) of a IT ivestmet (Bearoch ad Kauffma, ; Fichma, 4). Oe well-kow example is that Bearoch ad Kauffma () apply the Black-Scholes model to reflect the uderestimated ivestmet value of a firm s electroic bakig etwork. I other words, the maagerial flexibility cocepts i these studies are directly liked to the optios of geeratig additioal Proceedigs of the Nieteeth Americas Coferece o Iformatio Systems, Chicago, Illiois, August 5-7, 3. 7

8 value (i.e., the future value) for a IT ivestmet. I fact, as Trigeorigis (996) argues, the maagerial flexibility provided by a IT ivestmet would brig to a firm aother beefit, i.e., the improvemet of ivestmet resource allocatio. I this paper, our fidigs show that, i terms of IT ivestmet allocatio, a set of flexible IT iitiatives is better tha that of a set of iflexible IT iitiatives. This implicates that a firm ca cosider exercisig the maagerial flexibility of deivestmet/re-ivestmet for certai IT iitiatives to improve the overall IT ivestmet allocatio for a portfolio. Namely, our fidigs complemet to the prior IS literature. Our fidigs reflect that the maagerial flexibility of a IT ivestmet is ot oly related to the future value for a IT ivestmet but also relevat to the overall ivestmet allocatio for a portfolio of IT iitiatives. Diversifyig IT iitiative choice to improve portfolio ivestmet allocatio For maagig IT iitiatives, the idea of diversificatio is ot oly helpful for reducig the risk, but also for allocatig a firm s ivestmet more efficietly i a portfolio. I prior IS research, the fiacial portfolio theories, such as the moder portfolio theory (Markowitz, 959) are ofte applied to addressig the issues related to IT portfolio maagemet. From the perspectives of such theories, a firm should diversify ivestmets i order to strike a average for mitigatig the risk of ivestmet, i.e., ot puttig all your eggs i oe basket. Strictly speakig, this beefit of ivestmet diversificatio is o the basis of a set of related ivestmets that are characterized by the mutual iterdepedecy. However, there is aother situatio where a set of ivestmets could be urelated ad characterized by the pooled iterdepedecy. For example, i corporate ecoomics, it is ofte argued that a firm should seek for diversifyig the urelated lies of busiess i order to capture the beefits of resource allocatio efficiecy. I this paper, our fidigs complemet the prior IT portfolio studies ad highlight the importace of diversifyig urelated IT ivestmet. Our result implicates that a diversified set of urelated IT iitiatives ca geerate a better portfolio ivestmet allocatio, sice each IT iitiative is mutually idepedet but share the similar pools of risk costrait ad budget costrait i our experimets. More importatly, we provide a importat reaso why a firm eeds to adopt a diversificatio strategy for a better portfolio performace, eve if its IT iitiatives have very low mutual iterdepedecies. CONCLUSION AND CONTRIBUTION This paper makes several cotributios to research ad practice. First, our portfolio-level fidigs, i terms of characterizig ad maagig of IT iitiatives as a collective, ca strogly complemet the prior IS research focused o idividual IT project ad thus advace theories pertaiig to IT goverace ad portfolio maagemet. Secod, our experimetal results highlight that a firm ca achieve better IT ivestmet allocatio by employig the maagerial flexibility of IT iitiative or diversifyig the selectio choice of IT iitiative. Third, we create the decisio model to support IT iitiative portfolio decisio-makig. I the past, the primary decisio model that firms could rely o was the fiacial portfolio models. Such models are ofte difficult to accommodate a firm s IT ivestmet cotext. A most obvious weakess, for example, is that the iput metrics of the fiacial portfolio models rely o the fiacial asset s historical price, which fails to traslate to the commo IT iitiative maagemet routies i a firm. Fially, we employ a itegrated research methodology that icorporates computatioal modelig, real-world data, ad simulatios, which would strogly complemet existig methodologies established i IS. I the future, we pla to ivestigate the characteristic-directed strategies for maagig IT iitiative portfolio ad it is a research topic of great eeds from both academic ad busiess commuities. REFERENCE. Bearoch, M., ad Kauffma, R.J. () Justifyig electroic bakig etwork expasio usig real optios aalysis, MIS Quarterly, 4,, Fichma, R. G. (4) Real optios ad IT platform adoptio: Implicatios for theory ad practice, Iformatio Systems Research, 5,, Geoffrio, A.M. Lagragia Relaxatio ad its Uses i Iteger Programmig., Mathematical Programmig Study,, pp:8-4, Gree, A., Kaa, K., ad Krisha, R. () O Evaluatig Iformatio Revelatio Policies i Procuremet Auctios: A Markov Decisio Process Approach, Iformatio System Research,,, Proceedigs of the Nieteeth Americas Coferece o Iformatio Systems, Chicago, Illiois, August 5-7, 3. 8

9 5. Kumar, R., ad Ajja, H., ad Niu, Y. (8) Iformatio techology Portfolio Maagemet, Iformatio Resource Maagemet Joural,,3, McFarla, F. W. (98) Portfolio approach to iformatio systems, Harvard Busiess Review, 59, 5, Markowitz, H.M. (959) Portfolio Selectio: Efficiet Diversificatio of Ivestmets. New York: Joh Wiley & Sos. 8. Na, N. () Capturig Bottom-Up Iformatio Techology Use Processes: A Complex Adaptive Systems Model, MIS Quarterly, 35,, Nehrig, K. ad Puppe, C. () A Theory of Diversity, Ecoometrica, 7, 3, Perose, E.T. (959) The Theory of the Growth of the Firm, New York: Joh Wiley&Sos,. Piccoli, G., ad Ives, B. (5) IT-Depedet Strategic Iitiatives ad Sustaied Competitive Advatage: A Review Ad Sythesis Of The Literature, MIS Quarterly, 9, 4, Piramuthu, S., H., M. J. Shaw (998) Usig Feature Costructio to Improve the Performace of Neural Networks, Maagemet Sciece, 44, 3, Sharpe, W. F. (994) The Sharpe Ratio, Joural of Portfolio Maagemet,,, Sikora, R. ad Shaw, M. (988) A Multi-Aget Framework for the Coordiatio ad Itegratio of Iformatio Systems, Maagemet Sciece, 44,, Teece, D.J. (98) Ecoomics of Scope ad the Scope of the Eterprise, Joural of Ecoomic Behavior ad Orgaizatio,, Trigeorgis, L. (996) Real-optios: Maagerial ad Strategy i Resource Allocatio. Cambridge, MA: MIT Press 7. Tu, Y.., Zhou, W, ad S. Piramuthu. (9) Idetifyig RFID-Embedded Objects i Pervasive Healthcare Applicatios, Decisio Support Systems, 46,, Wallace, L., ad Keil, M. (4) Software Project Risks ad their Effect o Outcomes, Commuicatios of the ACM 47, 4, Weill, P. ad Aral, S. (6) Geeratig Premium Returs o Your IT Ivestmets. MIT Sloa Maagemet Review, 47,, Proceedigs of the Nieteeth Americas Coferece o Iformatio Systems, Chicago, Illiois, August 5-7, 3. 9

10 APPENDIX A This example is adapted from a real-world case. We demostrate a small set of IT iitiatives, ad it is collected from a fiacial service uit of a isurace compay i the Midwest durig 9-. Id. Iitiative ame NPV Risk 4 Cost x JEE platform migratio $8 8 $49 x Mobile paymet pla $ 5 $38 x3 Cotract maagemet system upgrade $4 5 $46 x4 Payroll system upgrade $ 5 $39 x5 Uderwritig system upgrade $ 5 $58 x6 Life ad auto policy web iterface $5 4 $8 x7 Auity policy modificatio $ $9 x8 Cliet e-otice system $8 $9 x9 Partership e-credit pla $6 5 $3 x Electroic moey trasferrig gateway $6 7 $8 x Debt/ledig data aalysis pla (BI) $7 6 $3 Sum $9 46 $7 Table 3. The IT Iitiative Ivestmet Data Table I this example, we wat to select two best sets of IT iitiatives (a aggressive portfolio choice ad a coservative portfolio choice for the compay s ivestmet cosideratio; thus, the risk costraits ( R ) are set as 5% ad 75% of the highest risk score (46). I additio, we wat to cut ito half the total cost of the ivestmet, so the cost costrait ( C ) is set as 5% of the highest cost ($7). These settigs just follow the heuristics, ad the settigs i differet compaies would be very cotiget. Moreover, by usig our proposed model (7)(8)(9), we ca derive the followig optimizatio problem. Max (x*8+x*+ x*7)/9 () s.t. (x*8+x*5+ x*6)/46 R () (x*49+x*38+ x*3)/7 C () We the use a optimizatio software (LINDO) ad a persoal computer to solve ()()() ad derive the two best sets of IT iitiatives for a compay s aggressive portfolio choice ad coservative portfolio choices as follows. Portfolio choice Selected IT iitiatives NPV Risk Cost Coservative {x5,x6,x7,x8} $67 85 $4 Aggressive {x,x,x4,x5,x6,x7,x8,x9,x} $8 35 $34 Table 4. The IT Iitiative Portfolio Choice Results 4 The risk is estimated by usig a scorig approach (-). The lowest value o this spectrum idicates the iitiative is very low risk. The highest value o this spectrum idicates the iitiative is very risky; such scores should be reflective of the iitiative risk from a holistic ivestmet perspective (i.e., iclusive of the busiess, IT, etc. perspectives) Proceedigs of the Nieteeth Americas Coferece o Iformatio Systems, Chicago, Illiois, August 5-7, 3.

11 APPENDIX B Value ($) Cost ($) Risk (score) (µ, σ) (, 6) (, ) (5, 3) mi to max Table 5. Statistical descriptios of real-world data -Diversity: {Value ~N(µ, σ )= (, 6*6*3), *3), Risk score ~N(µ, σ )=(5, 3*3*3)} Cost ~N(µ, σ )=(, Low-Diversity: {Value ~N(µ, σ )= (, 6*6*.33), **.33), Risk ~N(µ, σ )=(5, 3*3*.33)} Cost ~N(µ, σ )=(, Table 6. The diversity scearios Proceedigs of the Nieteeth Americas Coferece o Iformatio Systems, Chicago, Illiois, August 5-7, 3.