Firm strategic profile and IT outsourcing

Benoit A. Aubert & Guillaume Beaurivage &

Anne-Marie Croteau & Suzanne Rivard

Published online: 13 February 2008# Springer Science + Business Media, LLC 2008

Abstract IT outsourcing research has often been at thetransaction level, focusing on the role of the characteristicsof the IT activities as the antecedents of outsourcingdecisions. The present study extends past research effortsexamining the relationships between organizational levelvariables—in particular the business strategy—and thedecision to outsource. Results show that prospectors andanalyzers are more aggressive in their use of outsourcingfor IT operations. However, no differences between thebehaviour of the different strategic groups were found formaintenance activities, suggesting that these activities aretoo far from the core of the organization to be influenced bystrategic profile.

Keywords IToutsourcing . Strategic profile

1 Introduction

The objective of this paper is to examine the relationshipbetween business strategy and the level and nature ofinformation technology (IT) outsourcing behaviours. Pastresearch in IT outsourcing has mainly focused on the role ofthe transaction itself in explaining outsourcing decisions,without investigating the influence of strategic character-istics of the organization. The few studies looking at thestrategic profile of organizations led to inconclusive resultswhen looking at the direct effect of strategic characteristicsof organizations on outsourcing decisions. When observingorganizations, one cannot avoid noticing that organizationsof similar size and dealing in the same markets sometimesmake very different decisions with respect to IT outsourc-ing. Because major organizational decisions are likely toreflect a firm’s business strategy, it is anticipated that thedecisions to outsource IT activities will be influenced by thebusiness strategy.

The IT outsourcing phenomenon has been expandingduring the past decade, and this growth is likely to continue(Pati and Desai 2005; Young and Scardino 2006). Animportant strand of research has examined the outsourcingdecision using transaction costs theory.

Asset specificity has received a lot of attention fromresearchers (Williamson 1985). The role of asset specificityhas been supported by studies in many fields: auto parts(Monteverde and Teece 1982), aerospace (Masten 1984),and aluminum (Hennart 1988). However, results wereambiguous in the IT field (Nam et al. 1996; Aubert et al.2004). Measurement and uncertainty were also studied ininformation systems. Empirical studies showed that uncer-

Inf Syst Front (2008) 10:129–143DOI 10.1007/s10796-008-9065-8

This research was supported by the Social Science and HumanitiesResearch Council of Canada.

B. A. Aubert (*)HEC Montréal and CIRANO, 2020 University,25th floor,Montréal, Canada H3A 2A5e-mail: benoit.aubert@hec.ca

G. BeaurivageSAP Canada,380 St-Antoine West (suite 2000),Montréal, Canada H2Y 3X7e-mail: guillaume.beaurivage@sap.com

A.-M. CroteauJohn Molson School of Business, Concordia University,1455 de Maisonneuve Blvd. West,Montréal, Canada H3G 1M8e-mail: anne-marie.croteau@concordia.ca

S. RivardHEC Montréal and CIRANO,3000 chemin de la Côte Sainte-Catherine,Montréal, Canada H3T 2A7e-mail: suzanne.rivard@hec.ca

tainty played a key part in the choice of a governancemode. These studies also supported the proposition that themeasurability of the transactions strongly increases theprobability of outsourcing decisions by reducing the cost ofusing market mechanisms (Poppo and Zenger 1998; Aubertet al. 2004).

Studies relying on transaction costs, while providingseveral insights on the organization behavior, have ignoredsome intrinsic properties of the organizations. The transac-tion costs model only considers the characteristics of thetransaction itself. It does not take into account, for instance,any preferences of the managers or intent from theorganization. While ignoring these differences did notprevent transaction costs theory to explain outsourcingpatterns, taking into account these unique organizationalproperties would increase our understanding of IT out-sourcing. A similar concern arises with the resource-basedview. Its analysis remains more focused at the resourcelevel, not at the organization level, although it considers thehistorical path of the organization and the configuration itgenerated (Barney 1999).

Business strategy is one key organizational element thatdistinguishes one organization from another. It correspondsto the outcome of decisions made to guide an organizationwith respect to its environment, structure, and processesthat influence its organizational performance. In otherwords, it represents the means taken by an organization toreach its goals. Teng et al. (1995) measured the strategicorientation of firms, using the Miles and Snow’s (1978)typology, and did not find any relationship between thestrategic type and the use of outsourcing. The same authors(in Grover et al. 1994) also looked at the moderating effectof strategic type on the relationship between gaps in ISresources and outsourcing and found some significantresults, suggesting that there is a moderating effect. At thesame time, a significant body of literature in informationsystem has been devoted to the study of strategic alignment,notably alignment with the structure (see for exampleBergeron et al. 2004). Given that outsourcing is a keydecision about structure (Williamson 1985), it is perplexingthat the relationship between strategy and outsourcing hasnot been established more convincingly. These results areintriguing and call for additional investigation.

This study revisits how different strategic profiles makedifferent use of IT outsourcing. It could be described as anextension to the previous studies according to the defi-nitions of Berthon et al. (2002). It mainly differs fromGrover et al. (1994) and Teng et al. (1995) in two ways.First, the activities examined are different. The two studiesdone in the mid-nineties looked at five large groups:application development, systems operations, telecommu-nications, end-user support, and system planning. Thisstudy focuses on IS operations (including telecommunica-

tions) but details the activities with finer granularity thanprevious studies, establishing differences among the activ-ities included in each group. The second difference lies inthe measure of the strategic types. Most studies, both ininformation systems and strategy, rely on self-typing. Thestudies on strategy and outsourcing used these self-typingstrategies. The current study also offers an innovative wayto operationalize the strategic profile of the organization,using secondary data (from financial databases). Thisinformation is reported to regulatory agencies and deemedreliable. There is a possibility that biases were introduced inself-typing methods. These biases would be avoided usingpublic financial data. Therefore, this study would be amethod and context extension of past studies.

This extension leads to significant results (which werenot visible in past studies). Results show significantinfluence of strategic type on outsourcing, namely thatprospectors and analyzers are more aggressive in their useof outsourcing for IT operations. The separation ofmaintenance and operational activities also proved useful.These two activities do not appear to be managed in thesame manner. By focusing on the governance of eachactivity instead of using the percentage of budget allocatedto outsourcing, the differences between the activities wererecognized.

2 Strategic types of organizations

2.1 Miles and Snow’s typology

Approaches to identifying a business strategy can betextual, multivariate, or typological (Hambrick 1980). Thetypological approach is recognized as creating a betterunderstanding of the strategic reality of an organization,since all types of business strategies are viewed as havingparticular characteristics but a common strategic orienta-tion. While several typologies have been proposed byAnsoff and Stewart (1967), Freeman (1974), Miles andSnow (1978), or Porter (1980), the most frequently used inempirical research is Miles and Snow’s (1978) (Smith et al.1989; Zahra and Pearce 1990), which has been quoted morethan 1100 times (ISI Web of Knowledge, 1989–2007). Thistypology has successfully been used in research for morethan a quarter century (Snow and Hrebiniak 1980;Tavakolian 1989; Croteau and Bergeron 2001; Sabherwaland Chan 2001; Jusoh et al. 2006).

Miles and Snow’s typology reflects a complex view oforganizational and environmental processes, as well as theattributes of product, market, technology, organizationalstructure, and management characteristics (Smith et al.1989). The principal strength of this typology is thesimultaneous consideration of the structure and processes

130 Inf Syst Front (2008) 10:129–143

necessary for the realization of a given type of businessstrategy.

Miles and Snow’s typology defines four types ofbusiness strategies: prospector, analyzer, defender, andreactor. Firms choose one type rather than another accord-ing to their perception of their environment. The first threetypes are expected to enhance organizational performanceand share a continuum where the prospector strategy is atone end, the defender at the other, and the analyzer strategystands in the middle. The reactor strategy is excluded fromthe continuum since it represents an organization having nospecific strategy identified, which is likely to impedeorganizational performance. Some performance variationsamong the four strategic types have been observed depend-ing upon the industry, the sample size, or other organiza-tional constructs. Generally, it was found that prospectors,analyzers and defenders are all performing and outperformreactors (Snow and Hrebiniak 1980; Conant et al. 1987;Smith et al. 1989).

Organizations that have a prospector strategy try to reachthe largest market possible. They repeatedly make efforts toinnovate and continuously search for new product/marketopportunities. As the creators of change in their market,prospectors stress innovation and invest heavily in productR&D and environmental scanning. They usually seek forflexibility and prefer a decentralized organizational struc-ture. However, prospectors are not very efficient in theiroperations due to their strong concern with flexibility andinnovativeness.

At the opposite, organizations that adopt a defenderstrategy have a restricted market. They emphasize theefficiency of their production operations, the excellence oftheir products, the quality of their services, and their lowerprices. They have a stable and narrow niche product/marketin their industry. Focusing on the efficiency of their currentoperations, they deploy a centralized functional structurewhich facilitates the offer of high-quality and standardproducts or services at low prices. They mainly invest intheir core and most efficient technologies. However, sincedefenders rarely search outside their domain for newopportunities, they almost never make major adjustmentsin their technology, structure, or methods of operation.

Organizations that have implemented the analyzerstrategy share both prospector and defender characteristics,but in moderation. They try to be the first to introduce newproducts, yet remain in second place with certain productsfor which they offer a good quality/price ratio. With thestrengths of the above two types, they operate in a stabledomain of core products and seek new product/marketopportunities at the same time, in order to maximize growthopportunities and minimize risk. They watch their com-petitors much closely for changes and innovations, andrapidly follow the prospectors who quickly introduce

competitive products. They also copy the defenderscharacteristics when it comes to their stable domains byfostering routines and efficiency of their core operations.Overall, analyzers adopt a matrix organizational structureand deal with the challenge of simultaneously handlingsuch conflicting demands of efficiency and innovation.

Finally, organizations supporting the reactor strategy donot have the capacity to take over new opportunities ormaintain markets already acquired. They are the mostunstable organizations among the four types. They do notrespond consistently or effectively to the change anduncertainty of their environment due to the lack of strategyor misalignment between their strategy and their structure.Their incapacity to appropriately perceive the environmen-tal change results in poor performance.

2.2 Linking strategy and IT activities

An IT activity might be managed very differently depend-ing on the strategic type of an organization. Previousresearch has indeed found that each type of businessstrategy has a different portfolio of IT activities comparedto the other ones (Croteau and Bergeron 2001). Forexample, it was found that prospectors use IT applicationsto enhance their market flexibility and make quick strategicdecisions. Defenders implement IT applications to createoperational efficiencies (Tan 1997). They also improve theirbusiness performance by giving a strategic role to the ITdepartment, maintaining an open IT architecture, conduct-ing strategic scanning, and constantly evaluating theirinformation systems’ performance (Croteau and Bergeron2001). Analyzers use IT applications to conduct strategicscanning, and, therefore, improve their comprehension ofthe other organizations’ behavior (Sabherwal and Chan2001; Croteau and Bergeron 2001; Croteau et al. 2001).

Taking into account the strategic type of the organizationcan enable a finer analysis of outsourcing. While commonwisdom has often suggested that strategic activities shouldbe kept inside the firm, research has avoided the analysis ofoutsourcing taking into account specific strategic character-istics of the organization. Activities were deemed compa-rable from one organization to another. While this approachis not false, it may be improved.

2.3 Linking strategy and IT outsourcing

This study explores whether there is a link between anorganization’s strategic type and its use of IT outsourcing.More specifically, it investigates how strategic types in Milesand Snow’s typology can be related to the level of IToutsourcing activities deployed by an organization and itsorganizational performance. Asmentioned earlier, past researchon this relationship was not conclusive (Teng et al. 1995). This

Inf Syst Front (2008) 10:129–143 131

research takes a different approach in the measurement ofstrategy by using secondary data, such as financial informa-tion disclosed and annual reports, rather than mailedquestionnaires to analyze the strategy.

When examining the possible linkages between strategictype and IT outsourcing, we can formulate the followinghypotheses:

H1: Prospectors will have a higher usage of IT outsourc-ing than defenders

H2: Prospectors will have a higher usage of IT outsourc-ing than analyzers

H3: Analyzers will have a higher usage of IT outsourcingthan defenders

The justification for these hypotheses can be found in thedescription provided in Miles and Snow (1978), especiallywhen considering the adaptive cycle describing howorganizations adjust to their environment. This cycleincludes three elements: the entrepreneurial problem, whenthe choice of a given strategy is made, the engineeringproblem, which demands a specific portfolio of technolo-gies and capabilities, and finally the administrative prob-lem, which involves the choice of structure and process thatwould influence the future entrepreneurial problem. Theauthors saw these three problems as a cycle.

When looking at how defenders, prospectors, andanalyzers address these problems, many differencesemerge. When solving the entrepreneurial problem, Milesand Snow (1978) argue that defenders will grow cautiouslyand incrementally, while prospectors will grow in spurts.This would suggest that defenders can establish an internalgovernance for IT activities since the volume of activitywill be continuous and would occupy an internal workforcecontinuously. At the opposite, the prospectors would faceimportant fluctuations. One way to face fluctuating demandwould be to rely on outside help. Similarly, the environ-ment scan of the defenders would be limited. Thesecompanies would rely on their core capabilities. At theopposite, prospectors would try to introduce change andwould be more inclined to work with external partners tobroaden their operating arena.

When addressing the engineering problem, defenderswill rely on vertical integration and a single core technol-ogy while prospectors will tend to use prototyping and use

multiple technologies. This also supports our hypotheses.The vertical integration favoured by the defenders isdirectly aligned with the hypotheses. The use of multipletechnologies used by the prospectors also suggests supportto the hypotheses. Using multiple technologies simulta-neously requires a very diverse skill set. It is difficult tomaintain such skill set and outside providers are a way togain access to these skills.

Finally, the administrative problem, as described byMiles and Snow (1978) indicates that tenure in positionswould be longer and structure more stable for defenders,while prospectors would experience frequent changes andaccept people from the outside more easily than defenders.This also supports our hypotheses. Internal organizationwould be easier to maintain for defenders than forprospectors.

Another element that supports the hypotheses is the riskprofile of each type. The prospector is the risk taker inMiles and Snow’s typology. At the opposite, the defender isthe risk adverse. Entering into an outsourcing contractentails changes and contractual risks (Aubert et al. 2005).Therefore, taking into account the risk-aversion profileassociated with the strategic type of the organization wouldalso support the hypotheses.

Table 1 summarizes the key elements.We can hypothesize that defenders, due to their nurturing

stability, need for control, focused activities and longertenure, will be the most cautious users of IT outsourcing,avoiding the risks associated with outsourcing. Stability intheir use of their resources will facilitate the establishmentof internal governance (Williamson 1985). At the opposite,prospectors, being the most prone to change in businessvolume, seeking flexibility, diversity, and adopting a moredecentralized (and less stable) structure, will be the highestusers of IT outsourcing. Outsourcing would be a suitableresponse to these fluctuations. Prospectors would also bemore tolerant to outsourcing risks. In all cases, analyzerswill be between the two groups.

Rockart and Leventer (1979) proposed three distinctivecategories of IT activities, which are development, oper-ations and management. Similarly, Tavakolian (1989)identified three major groups of IT activities, i.e. systemsdevelopment and maintenance, systems operations, andsystems administration. Past research in IT outsourcing has

Table 1 Main differences be-tween strategic profiles

Market Main concern

Structure Technology used

Risk profile

Prospectors High fluctuations, growth in spurts

Flexibility, change

Decentralized, shorter tenure

Varied High risk tolerance

Analyzers Defenders Stable Control Centralized,

longer tenure Focused Low risk

tolerance

132 Inf Syst Front (2008) 10:129–143

looked at IT operations and maintenance activities. Dibbernand Heinzl (2006) found that the most specific activitieswere the most strategic ones. Aubert et al. (2004) found thatmaintenance activities were less specific than IT operations.Therefore, it can be reasonably assumed that maintenanceactivities are less strategic than IT operations, all otherthings being equal. It is posited in this paper that differentcategories of IT activities will be outsourced differentlydepending on the strategic type of the organization.Operations should be more influenced by strategic typethan maintenance activities.

H4: The impact of strategic type on the outsourcingdecision will be stronger for IT operations than formaintenance activities.

3 Methodology

This section presents the methodology followed to assessthe relationship between the strategic type of an organiza-tion and its use of outsourcing. The sample of firms camefrom an existing database (Aubert et al. 2004). Theoutsourcing data are described in Section 3.1. Theprocedure used to classify the firms into one of the threecategories (prospectors, analyzers, and defenders) will bedescribed in detail, from section 3.2. It is one of the originalcontributions of this study.

3.1 Outsourcing data

The sample was taken from an existing database (Aubert etal. 2004). This database included data on 200 Canadianfirms whose CIO had participated in a survey on IToutsourcing. The database contained the firm’s identity, aswell as data on whether each of 17 IT activities wasoutsourced or performed in-house. To be included in thesample of the present study, a firm had to be publiclytraded, to ensure that its financial data and annual reportswere available. The final sample includes data about 41firms; their annual revenue varied from 15 million and 15billion dollars, and they operated in a total of 29 industries.

As described in Aubert et al. (2004), a survey wasconducted to assess the use of outsourcing from Canadiancompanies. Each company was asked to indicate, for 17activities (listed in Table 6), if it was doing the activity in-house, using its own employees, or using an outsidesupplier to perform the activities. Respondents had to selectonly one of the choices per activity, indicating the maingovernance mode adopted for each of the 17 activities. Foreach firm, two IT outsourcing scores were computed. Onewas the IT operations outsourcing score, the other the ITmaintenance outsourcing score. This score was obtained by

adding the number of activities of a given type—eitheroperations or maintenance—that were outsourced by thefirm, divided by the number of activities. The average ofthese outsourcing scores was calculated for each of thethree strategic types. The analysis will examine both thedetailed data and the outsourcing scores.

3.2 Strategic type

The literature using the typology of Miles and Snow offersseveral ways to operationalize the constructs. The simplestone was developed by Hambrick (1983) who looked at thenumber of new products brought to market by the firm.Snow and Hrebiniak (1980) developed a more encompass-ing measure where four descriptions, one by type, areintroduced to the respondent who has to select the mostappropriate one to represent the firm. This measure was usedin several studies, especially during the 1980s. Segev (1987)used an approach relying on 27 Likert-scale questions. Thismeasure was used more than 80 times in literature (ISI Webof Knowledge, 1989–2005). Finally, Conant et al. (1990)introduced a measure based on the three cycles of adaptation.This measure was also widely used.

All these measures (except the one by Hambrick 1983)were based on perceptions. Hambrick (1983), while using avery simple indicator (some will argue too simple), had theintuition that the strategy of a firm could be deducted fromobservable facts. Other authors relied on managers indicat-ing what the intentions of the management team were, orwhat their perception of the course of action was. Theseapproaches, often relying on self-typing, are vulnerable to“strategic desirability”. Managers might describe the strat-egy they would have liked to follow, or the one they wereplanning to follow in the future, rather than the one theyused to guide their action in the past.

The classification procedure proposed in this paper reliessolely on secondary data. It uses the financial informationprovided by the firm and verified by auditors. It also usesannual reports. In this sense, the procedure is based on hardfacts as much as possible. It also takes into account thecorporate intention as communicated to the shareholders inthe official documents.

Based on literature, it was possible to identify whatwould be the profiles of each strategic archetype. Themeasures of Segev (1987) helped to choose the indicators,while those of Conant et al. (1990) were used to identify therelevant elements of the annual reports to consider. Adatabase containing all variables for companies andavailable years was created. As will be described later,each firm was compared with the other firms in its industry.To do this industry comparison, all the firms in the sameindustry available in the databases described under the

Inf Syst Front (2008) 10:129–143 133

section labelled “Information sources” were used (It wasnot limited to the firms in the sample).

After the analysis of the financial data was completed,the annual report was examined to understand and validatethe elements found in the financial databases. The annualreport helped understand the environment in which theorganization operated. The indicators used and theirexpected patterns for each profile are described in thefollowing pages.

3.2.1 Information sources

Many sources were used to retrieve the information on eachfirm:

– Stock Guide: This database provides financial data andratios for the publicly listed Canadian firms.

– Mergent Online: This database provides data on activeand inactive Canadian and international firms.

– Corporate Retriever: This database provides financialdata as well as information about the type of activitiesperformed by the firm. It also indicates the businesssectors in which a company operates.

– Annual Reports: Available for every year and for everypublic company. In each one, a letter to the stock-holders is provided. It shows the picture which highermanagement tries to convey to the stockholders.Audited financial data served to validate the dataobtained from the other databases.

3.2.2 Indicators

When looking at the secondary data available, severalelements were analyzed. First, the financial informationwas surveyed: beta, debt structure, liquidity ratios, stabilityof the benefits and payment of dividends. The annual reportwas then assessed and used to understand some figures.The business volume of the firm, compared with the otherplayers of the industry, also provided information about thestrategy of the firm.

The first indicator considered was the beta of the firm.The beta coefficient is a measure of the volatility of thecompany shares compared to the market volatility. It givesan indication of the level of risk of the firm, as perceived bythe market. A firm with a low beta indicates a company lesssubject to variations than the market average. On thecontrary, a higher beta means that the value of the firmvaries more than the market. Considering this definition, adefender should have a low beta, usually under one, andunder the median of the industry. A prospector should havea high beta, considerably higher than one, and higher thanthe industry average. An analyzer should have a betaaround one, close to the median of the industry.

The debt structure of the firm was analyzed with fourfinancial elements: debt/equity ratio, debt ratio, interest paidper $100 of sales, and intensity of use of long-term assets.This helps understand whether the firm is financed by debtor by equity. A defender will favour stability by having alow debt which will strive toward zero, pay little interest,have low debt/equity ratio. Prospector will show oppositeprofiles. These firms take risks to generate high yields anduse debt to favour expansion (debt ratio over 1). Aprospector will show high interest payments per sale.Finally, the analyzer will have a moderate level of debt(around 30% of sales amounts), pay some interests andshow a median debt and debt/asset ratios to favour bothexpansion and stability. Of course, any ratio has to becompared to the industry average for a firm. On its own, theratio has limited usefulness.

The amounts invested in research and development areanother source of information on the strategic type of theorganization. The amounts allocated to research anddevelopment were taken into account and compared to theindustry averages. It is expected that a defender will havelower R&D expenses than a prospector. The analyzer willexhibit R&D expenses close to the industry median.

Sales expenses ratio illustrates the effectiveness of thefirm in leading its business activities, and its marketingaggressiveness. Lower expenses, under 60%, will point outa competitive firm, with a lean approach, which is ratherassociated with a defender. High expenses, above 85%,point out a firm which either experiences difficulty incontrolling its expenses or invests aggressively in market-ing activities, what would be associated with a prospector.Moderate expenses, between 60% and 85%, will point outto an analyzer. This ratio varies considerably from oneindustry to another. Again, firms were compared to theirindustry average. The percentages given are an illustrationfor the “typical firm”.

Concerning liquidity ratio, three financial ratios wereconsidered: cash ratio, current ratio (also called workingcapital ratio) and EBITDA-to-interest coverage ratio. The firstone is the amounts in cash or cash equivalents divided by theliabilities; the second one is computed by dividing short-termassets by short-term debt. The last one measures the capacityof the firm to cover its interest payments using its benefits.

Defenders will have high liquidity ratios, providingstability to the firm. A cash ratio of 1.5 and a current ratioabove 1 will be considered to be associated with thedefender. The EBITDA-to-interest coverage ratio will behigh, or there will not be interest payments. This analysiswill also have to take into account the industry averages.Prospectors will show low liquidity ratios since short-termassets are reinvested continuously to favour expansion ofthe firm. The analyzer will have liquidity ratios around 1(or around its industry average).

134 Inf Syst Front (2008) 10:129–143

The stability of the benefits was also considered. Milesand Snow (1978) maintained that each strategic type(prospector, defender and analyzer) will be linked toperformance. However, the pattern of profits will differfrom one type to another. Three financial indicators wereconsidered: the variation of raw and net benefits andtreasury movements generated by the company.

A defender will have a stable pattern of benefits whencompared to the average of its industry, while theprospector will show highly variable levels of benefitsfrom 1 year to another and when compared to the industry.Analyzers will follow the industry average.

The dividends paid on regular shares as well as theremittance of past dividends were analyzed. For defenders,the payment of dividends will be positive and constantsince, generally speaking, they will be able to balance theirportfolio of activities and pay predictable dividends. Forprospectors, the payment of dividends is not part (ingeneral) of their preferences and, as a result, there will befew or punctual payments. Expansion will be favoured,which reduces the possibility of dividends. Availableresources will be reinvested in the company, not returnedto the shareholders. For analyzers, the firm will begenerally distributing dividends corresponding to themedian of the industry.

Finally, the annual report of the firm has to beconsidered at three levels, the first two will be forvalidation of the previous analysis and the last, for analysisas an indicator. First, a validation of key figures wasperformed to ensure that the data obtained from the variousdatabases was accurate, errors like a difference in the fiscalyear or territory was to be detected. Second, someexplanation of the numbers could be obtained. For instance,one defender might have surprisingly low liquidity, but wewill understand in the annual report that the purchase of acompetitor was realized to save the niche which will bealigned with the defender’s strategy. Third, the section «Letter to the stockholders » allowed an understanding of thestrategy which the firm officially claims it is following. Adefender will speak about reducing expenses, about

focusing on the effectiveness of the firm, about quality,about delivery, or about vertical integration. A prospectorwill discuss expansion, innovation and strategies to getcloser to the customer. An analyzer will seek both types ofobjectives in moderate manner. This served as an additionalindicator.

In summary, Table 2 shows the categories of indicatorsused and the expected trends.

Firms with no coherent pattern, which would not fit inany of the profiles, would be classified as reactors. Thesefirms would react to outside events without a guiding lineof conduct. They lack cohesiveness in their responses to theenvironment and, therefore, a pattern would be impossibleto identify for reactors.

3.2.3 Final database and validation

A database comprising the data about the firms in thesample and the firms in their industry was constructed. Thefirms were coming from 29 industries. Data over 15 yearswere extracted to adequately understand each industry. Thiswas done to avoid potential exceptional events in a givenindustry. Once the portrayal of each industry was estab-lished, the data from the companies in the sample werecompared to their corresponding industry data. A descrip-tive table was established for each firm in the sample.

In order to validate the classification of the organizationsin their appropriate strategic category, two activities wereperformed. First, the indicators and the expected trendswere validated by an academic expert and an accountingexpert. Then, once the firms were classified into theirrespective categories, ten firms were selected for a secondevaluation. The accounting expert reclassified the firms andconfirmed all but one decision.

3.3 Examples

In order to illustrate how the secondary data were used toidentify the strategic types of the organizations, threeexamples are provided. They are representative of a

Table 2 Indicators

Indicator or group of indicators Defender Analyzer Prospector

Beta Low Industry average HighDebt structure Little or no debt Industry average High levels of debtResearch and development Lower than industry average Industry average Higher than industry averageSales expenses Low Median HighLiquidity ratios High Industry average LowStability of the benefits Very stable Fluctuates with industry trends Higher variability than industryPayments of dividends High Industry average Little, irregular or no dividendsStrategy in the annual report Cost, effectiveness, quality, delivery,

vertical integrationSeek a median strategy Expansion, innovation, customer

Inf Syst Front (2008) 10:129–143 135

defender, an analyzer, and a prospector. All the firms in thesample were classified using the same procedure.

3.3.1 Defender

The first example is Cameco, the world’s largest supplier ofuranium at the time of the study. Cameco has a beta lowerthan the one of its peers in the industry. Its debt/equity ratiowas at 0.17, its debt ratio was at 0.14, and $1.24 of interestwas paid per $100 of sales. The intensity of use of long-term assets was at 5.84. Cameco is in a better position thanthe firms in its industry. Its interest expenses are among thelowest and the debt ratio of the industry is at 0.27. Theseindicators all suggest a defender.

The research and development in the mining industry islimited (and often centered on finding new mining sites).Half of the firms have no R&D outside the exploration ofnew sites. Cameco has limited expenses of R&D outsideexploration (0.029% of revenues) which would be betweenanalyzer and defender. The sales expenses (0.49) are verylow compared to the other firms in the industry (0.76),which is the mark of a defender.

The liquidity ratios also bear the stamp of a defender,with the exception of the cash ratio. The cash ratio is at0.77. However, the working capital is at 1.96 and theinterest coverage ratio is at 19.58 (half of the firms in theindustry were showing losses, thus showing negativeratios). The firm is profitable, although profits are variable(but much less than the other firms in its industry).Interestingly, over several years, the profits before taxes ofCameco are more stable than the net profit, which mightsuggest adjustments in accounting rules. Cameco is veryprofitable. At the current 12% rate of return, it is actually alow point for several years.

Cameco pays regular dividends (approximately 4% ofthe gross sales). This is in an industry where less than halfof the firms pay any dividends at all. In the annual report,we notice that Cameco relies on very long-term contracts.This emphasizes the stability of the firm. The annual reportindicated that contracts were very long term. The firmemphasized its ability to offer stability and low prices for itsproducts, outlining the quality of its assets (mining sites).The majority of the indicators point toward a defenderprofile (Table 3).

3.3.2 Analyzer

The second example is Alcan, evaluated as an analyzer.Alcan is a leader in the aluminum industry. At the time, ithad revenues of $11 billion. Alcan had a beta in the highestquintile of its industry, which indicates that the investors

saw the firm as risky. However, its financial structure wasmore conservative. Its debt ratio was 0.19, the debt/equityratio was at 0.3, and $1.30 of interest was paid per $100 ofsales. The intensity of use of long- term assets was 2.23. Inits industry, half of the firms had a debt ratio under 0.2. Theaverage was at 0.27. The interest paid was under theaverage (3.31). The financial structure leans toward the oneof an analyzer in this industry.

The research and development expenses also pointtoward an analyzer. The average of the industry is slightlyover 4% of revenues. Alcan is at 1.85%. However, half ofthe firms showed near-zero R&D expenses in this industry.Alcan is, therefore, in a middle ground. The sales expensesprovide a similar picture. The industry average is at 0.76and Alcan is at 0.77.

The liquidity ratios also suggest that Alcan is ananalyzer. The cash ratio is at 1.39, the working capital ratiois at 2.28, and the interest coverage ratio is at 8.46. Lookingat the industry, the cash ratio is above 1, the working capitalratio is around the median (2.12). The interest coverageratio is high (especially considering that half of the firms inthis industry show a negative interest coverage ratio).

The industry shows variable profitability. In fact, theaverage of the sector shows a loss and the median shows a5% rate of profit. The variance is very high. Compared tothe industry, Alcan shows a stable profit pattern and a 6%margin. This oscillates between a defender and an analyzer.Alcan always paid regular dividends. Only half of the firmsin this industry pay dividends. This would lean toward adefender. Finally, the annual report indicates shows that theorganization was establishing a risk management programwhile pursuing expansion through vertical integration. Thissuggests a carefully managed expansion in areas in whichthe firm has knowledge and expertise. Table 4 summarizesthe situation of Alcan:

The overall strategy clearly is one of analyzer. While afew elements would point toward prospectors’ or defend-ers’ behaviour, Alcan showed a very balanced profile.

Table 3 Defender

Indicator or group of indicatos Defender Analyzer Prospector

Beta CamecoDebt structure CamecoResearch and development CamecoSales expenses CamecoLiquidity ratios CamecoStability of the benefits CamecoPayments of dividends CamecoStrategy in the annual report Cameco

136 Inf Syst Front (2008) 10:129–143

3.3.3 Prospector

The company GEAC computer is an example of aprospector. The year before the study, it showed revenuesof $190M. These revenues grew to $366M the year of thesurvey and to $629M the following year. This is an averageyearly rate of growth of 180%.

Not surprisingly, the beta is at 1.24, which is over 1 butnot extremely high. However, the debt ratios are high. Thedebt ratio is at 0.63, the debt/capital ratio is at 1.73 (theindustry is at 0.38), and $1.72 is paid per $100 of sales. Theintensity of use of long-term assets is at 2.45. When lookingat the high-growth pattern of the company, it shows thatthis growth is aggressive and financed through debt. This isthe mark of a prospector.

The sales expenses ($69) are above the industry average($60). The R&D expenses in this industry are usually high(around 20% of sales). GEAC’s expenses are lower thanthose of the other firms (10%) but have doubled (in percentof sales) from the previous years. An additional 2% wasadded the following year. This suggests higher reliance onresearch and development (moving toward prospector).

The liquidity ratios are low. The cash ratio is at 0.54. Theworking capital ratio is at 0.57. The interest coverage ratiois negative (−8.82). The firm is in the lower end of itsindustry distribution on liquidity measures, which is themark of a prospector.

The firm was not profitable the year of the survey. Itposted a loss. The following year, a profit (twice asimportant as the previous loss) was posted, followed byanother year of loss, followed by another profitable year….This pattern shows high variability and is the mark of aprospector. The firm has never paid dividends. The strategyoutlined in the annual report reflected a prospector’sbehaviour, insisting on offering a complete solution to itscustomers and on growth. The firm had just completed anacquisition that tripled its revenues.

The indicators, summarized in Table 5, suggest that thefirm has adopted a prospector’s strategy. The firm isacquiring other companies in order to complete its offeringto its clients.

4 Results

4.1 Strategic profiles

The 41 firms of the sample were found to include 11prospectors, 14 analyzers, 13 defenders, and 3 reactors.Firms showed coherent sets of characteristics that corre-sponded well to one of the categories. The few deviationsare explained in the paragraphs below.

4.1.1 Prospectors

The data on the 11 prospector firms suggested a manage-ment that was not very conservative. These firms paiddividends only when they were operating in an industrywhere all players paid dividends. Even then, the prospectorspaid lower dividends than their competitors. There was ahigh level of coherence among the different indicators forthis category of firms. No prospector had a ratio of debtunder the average of industry, and none had sales expenseslower than the average of its industry. The liquidity ratios ofall firms were considerably lower than the average of theirindustry. Annual reports introduced a clear picture of aprospector.

4.1.2 Analyzers

Fourteen firms were classified as analyzers. While ana-lyzers can show some degree of behaviours similar toprospectors’ or defenders’, a majority of their indicatorswere in the center of the scale. These firms’ managementwas in general conservative. The ratios of the analyzers

Table 5 Prospector

Indicator or groupof indicators

Defender Analyzer Prospector

Beta GEACDebt structure GEACResearch and development GEACSales expenses GEACLiquidity ratios GEACStability of the benefits GEACPayments of dividends GEACStrategy in the annual report GEAC

Table 4 Analyzer

Indicator or groupof indicators

Defender Analyzer Prospector

Beta AlcanDebt structure AlcanResearch and development AlcanSales expenses AlcanLiquidity ratios AlcanStability of the benefits AlcanPayments of dividends AlcanStrategy in the annual report Alcan

Inf Syst Front (2008) 10:129–143 137

tended to follow the industry average. Moreover, in theannual reports, these firms talked as much about quality andabout competitive prices as they did about innovation andabout closeness to the customers. They addressed issues ofexpansion and management of risks since the analyzers arerisk averse. Horizontal or vertical integration was also partof their strategy.

4.1.3 Defenders

Thirteen firms were classified as defenders. This group wasnot as homogeneous as the two previous ones. These firmshad a conservative, rigid management and showed riskaversion. Of course, not all the elements considered wereperfectly aligned toward the “defender” category. However,a majority of indicators converged toward the sameconclusion. Some facts could explain why some indicatorsmoved away from the typical defender profile.

One firm was more into debt than the average of theindustry, while still being a defender. The debt was due tothe acquisition of a company in financial difficulties. Fivefirms had higher sales expenses than the industry average,while having all other indicators toward the defenderprofile. Two reasons were believed to explain this situation.Firstly, among all analyzed ratios, sales expenses isprobably the most controversial since the annual reportsdo not always include the same types of expenses in thiscategory from one company to the other, even for firmsworking in the same industry. It seems that accounting rulesallow some flexibility to the firms when reporting theseexpenses. Secondly, some firms incur these expenses torespond to quality norms or certifications, thus increasingthe expenses even when it is only to protect access to amarket.

Then, four firms had weak liquidity ratios for defenders.In fact, some firms, having very stable income, financialstructure funded by equity and stable sale expenses, stillhad very little liquid assets. It might be due to the fact that itis not necessary to maintain high levels of liquidities toensure their financial stability since all of the other elementsin their environment are stable. Another firm showed a loss.Its industry faced difficult times. Finally, the analysis of theannual reports was mostly in line with the financial data.Four defenders spoke about expansion, what, a priori,would contradict the defender profile. However, expansionis necessary for some defenders, to secure their niche. Theyhave to grow and secure their market.

4.1.4 Reactors

Three firms were classified as reactors. One of the firmsrecorded no income. This firm seemed to be liquidating itsassets or transferring them to another company. Another

firm was in severe deficit and trying to straighten itsfinances before being acquired by foreign investors. Finally,in an industry in full growth, a third firm could not succeedin generating profits. It would be in transition andrestructuring before being acquired by foreign investors.These firms were discarded from the sample.

4.2 IT outsourcing patterns

For each firm, data were available on whether the firm hadoutsourced or not each of the 17 IT activities. Table 6reports the percentage of firms from each category thatoutsourced a given activity. The results show a patterngenerally aligned with the hypotheses. Indeed, ten of the 17activities exhibit the expected pattern. Only one of the 17 isopposite to expectation (micro-computer maintenance). Thecheck mark (√) in the right column indicates that theexpected pattern was found for the activity.

As expected, prospectors and analyzers have muchhigher outsourcing scores than the defenders for theoperations (activities 1 to 11). Seven of the 11 activitiesshow the expected pattern for IS operations. Anotheractivity–Data entry–exhibits the expected pattern, althoughto a lesser extent: while a significantly larger proportion ofprospectors than defenders outsourced this activity, thepercentage of analysers and of defenders that outsourced itwas the same (hence the ± indication mark in Table 6).

When looking more closely at the results, it appears thatIT operations and maintenance activities are not managed inthe same way. First, there is more outsourcing ofmaintenance activities than of IT operations activities.Second and surprisingly, the outsourcing scores in themaintenance activities are at times higher for defenders thanfor prospectors and analyzers (although these differencesare not significant).

It is important to remain cautious when looking forstatistical significance for such a small sample. The analysisof these results is mainly aimed at identifying the mostimportant differences. It is not done to “prove” a result withthe idea that it can be generalized.

When comparing analyzers and prospectors using t tests,these two groups do not show any significant difference.However, when comparing the activity means of defenderswith the corresponding means of analyzers and prospectorscombined, several activities show significant differences(p<0.05). Several activities are significantly less outsourcedby defenders than by both analyzers and prospectors. Theseactivities are shown in Table 7.

For each firm, two IT outsourcing scores were computed.One was the IT operations outsourcing score, the other theIT maintenance outsourcing score. These scores wereobtained by adding the number of activities of a giventype—either operations or maintenance—that were out-

138 Inf Syst Front (2008) 10:129–143

sourced by the firm, divided by the number of activities.The average of these outsourcing scores was computed foreach of the three strategic types.

A t test was conducted and the results indicate that theaverage IT operations outsourcing score was significantlylower for the defenders than for the prospectors and for theanalysers. No significant difference between the strategictypes was found for the IT maintenance outsourcing scores(Fig. 1).

5 Discussion

Two different types of results were obtained, according tothe types of IT activities. For IT operations, both H1 and H3

were supported. Both prospectors and analyzers relied moreon outsourcing for their IT operations than defenders did.H2 was not supported; there was no significant difference

between prospectors and analyzers. For maintenanceactivities, none of the hypotheses about the influence ofstrategy on outsourcing profile was supported. Therefore,H4, stating that the impact of strategic type on theoutsourcing decision will be stronger for IT operations thanfor maintenance activities, was supported.

5.1 Distinction between strategic profiles

When comparing the strategic profiles, the immediateobservation is the similarities between the outsourcingpattern observed for both prospectors and analyzers. Thenumbers are extremely close. Defenders showed a differentpattern. Their use of outsourcing for IT operations isextremely limited. In a way, this challenges the idea thatthe continuum applies for all the elements of the strategicprofiles. Traditionally, analyzers are portrayed as a middleground between the two extremes. However, if analyzersare thought as a combination of prospectors’ and defenders’characteristics, maybe outsourcing is one of the prospec-tors’ behaviours that analyzers adopt more readily, at leastfor IT operations. The results suggest that it might beworthwhile to further investigate this aspect.

Table 7 Significant differences

Activities that are significantly less outsourced by the defenders thanby the analyzers and prospectors combined (p<0.05)

1. Scheduling of operations2. Control of operations3. Technical support services7. Operation of client–server systems9. Printer operation10. Disk space management12. Operating system maintenance

IT operations

Prospectors ≈ Analyzers > Defenders

Maintenance activities

Prospectors ≈ Analyzers ≈ Defenders

Fig. 1 Outsourcing scores

Table 6 Patterns of IT outsourcing

Percentage of firms outsourcing the activity

Activities Prospector (%) Analyzer (%) Defender (%) Expected pattern(n=11) (n=14) (n=13)

IT Operations 1. Scheduling of operations 18.18 14.29 0.00 √2. Control of operations 18.18 21.43 0.003. Technical support services 45.45 25.00 7.69 √4. Software operation 18.18 14.29 7.69 √5. Operating system operation 27.27 21.43 7.69 √6. CPU operation 27.27 28.57 7.697. Operation of client-server systems 9.09 25.00 0.008. Operation of telecom software 40.00 38.46 25.00 √9. Printer operation 40.00 28.57 7.69 √10. Disk space management 27.27 21.43 0.00 √11. Data entry 22.22 7.69 7.69 ±

Maintenance 12. Operating system maintenance 45.45 30.77 7.69 √13. Hardware maintenance 81.82 71.43 92.3114. Micro-computer maintenance 54.55 78.57 84.6215. Network maintenance 54.55 42.86 38.46 √16. Printer maintenance 72.73 71.43 84.6217. Telecom lines maintenance 81.82 64.29 75.00

Inf Syst Front (2008) 10:129–143 139

“Classical” outsourcing, namely the main data centeroperations (software, OS, and CPU operations), was usedby approximately a quarter of prospectors and analyzers.Almost no defenders used outsourcing for these operations.This pattern is also observable when looking at themanagement activities. These activities provide anotherdistinction between the outsourcing behaviour of thedifferent profiles. These activities involve the managementof the operations. Control and scheduling imply decisionsabout how and when IT operations are conducted. Supportservices also imply control activities since they include theactivities associated with the introduction of softwaremodifications or new functionalities into production.

None of the defenders had outsourced scheduling orcontrol of operations. Only one had outsourced technicalsupport services. While prospectors and analyzers were notmassively outsourcing these activities outsourcing was stillpresent. Almost half of the prospectors had outsourced theirtechnical support services.

The only activity that defenders seemed to outsourcewithin the IT operations was the operation of telecommu-nication software. Twenty-five percent of the defendersoutsourced this activity. This percentage was still lowerthan the one observed for the other firms, which wasaround 40%.

All these elements outline the tight control (internal) onoperations that defenders keep over their key IS operations.This appears to be an extreme stance. The other twoprofiles show moderate usage of outsourcing. As will bedescribed in the next section, these differences were notobserved for the maintenance activities.

One hypothesis that can be formulated to explain thesedifferences would be linked to the fact that the risktolerance of prospectors and analyzers is higher than theone of the defenders. Prospectors and analyzers might bemore subjected to internal resource gaps (as described byGrover et al. 1994) compensated by the use of outsourcing.Alternatively, they might simply be more tolerant tooutsourcing risk, looking for increased flexibility with orwithout resource gaps. Current data do not provideinformation to further investigate this aspect.

5.2 Distinction between IT operations and maintenanceactivities

As mentioned before, maintenance activities are signifi-cantly more outsourced than IT operations. Two possibleexplanations can be offered to explain these differences.The first one pertains to the more systemic nature of IToperations when compared to maintenance activities. IToperations are more tightly integrated with the organiza-tion’s activities. Information system operations are process-ing the information associated with the day-to-day activities

of the organization. Sales, purchases, production activitiesare all included in the list of activities supported by the ISoperations. Therefore, these IT operations vary in synchwith the company operations. Information system opera-tions probably present similar patterns in variety andvolume as the company operations.

On the other hand, maintenance activities are moreautonomous than IT operations. Maintenance activities areperformed according to a calendar that is not necessarilylinked to the operations of the organization. Some mainte-nance activities are performed at regular time intervals (forinstance on equipment) or according to a calendar deter-mined by the vendor (for instance on the operating system).It is unclear, looking at the description of maintenanceactivities (Ng et al. 2002) if the maintenance activities aremore or less stable in volume and skills required than ISoperations. However, these activities are very similar fromone organization to another since many tasks are vendordependent and vendor initiated (Ng et al. 2002). Therefore,the economies of scale available for these activities shouldbe more significant than for IS operations (vendorsperforming standard tasks for every clients). This couldexplain why defenders, analyzers, and prospectors all adoptoutsourcing patterns that are similar.

As illustrated in Fig. 2, the second explanation lies in theproximity to the core of the organization. IS operationssupport the core activities of the organization. Maintenanceactivities are very indirectly linked to the core activities.The maintenance activities are not as specific as IToperations (Aubert et al. 2004). This would lead us toexpect that they are not as strategic as IT operations(Dibbern and Heinzl 2006). Maintenance activities arefurther away from the core of the organization than ISoperations. This would mean that these activities are too faraway from the organizational core to be affected by thestrategic orientation of the organization. The choice ofgovernance mode for these activities would be determinedprimarily by the characteristics of the activities, not byorganizational level variables.

CORE

Maintenance activities

IS operations

CORE

Maintenance activities

IS operations

Fig. 2 Distance from the core

140 Inf Syst Front (2008) 10:129–143

5.3 Limitations

Prior to discussing the contributions of this study andsuggesting avenues for future research, some limitations ofthe study must be acknowledged. These limitations mainlyconcern sample size, measure of the dependent variable,and the variance of characteristics among the activities.

The first limitation is that of the small sample size.Indeed, the fact that each of the four sub-samples includedless than 15 firms puts some limitations to the extent ofgeneralization that can be made from this study. Because ofdata availability, the sample size was limited to 41organizations. However, detailed information was obtainedon each of these firms. The limited sample size prevents thestatistical detection of small differences between twostrategic types that share some characteristics. It might bewhy, when considering H2, no difference was observedbetween prospectors and analyzers. Although the patternsobserved in Table 6 provided some support for H2, thesedifferences were not significant.

A second limitation is related to the measure of thedependent variable. In this study, respondents were asked toindicate the main governance mode for each of 17information systems operations activities. It might beargued that such a measure is weak because a firm thatoutsources a small portion of a given activity—say 10% ofits hardware maintenance—would fall in the same categoryas another firm that does not outsource any of its hardwaremaintenance activities. Other researchers have alleviatedthis issue by weighing the importance of outsourcing agiven activity relative to the budget of the IS function(Grover et al. 1994 for instance).

The other limitation of the study is the fact that all the ITactivities studied were IT operations. It would have beeninteresting to include software development in order toincrease the variety of activities and the diversity of theactivity characteristics. Some software development activ-ities might be closer to the core of the organizations. Itwould have provided additional information about the“closeness to the core” proposition.

6 Contributions and avenues for future research

Not with standing the limitations outlined in the previoussection, the results indeed support the idea that there is alink between the strategic profile of an organization and theuse of outsourcing by the organization. However, this linkis not systematically present for all activities. Past research(Teng et al. 1995) had not been able to identify this directlink. By extending the work done in the past and using adifferent measure of strategic profile, a direct link could beestablished. The results highlighted the importance of

retaining a sufficient level of granularity for the activitiesconsidered when studying outsourcing. Aggregated meas-ures of outsourcing might not have revealed the differencesobserved in the current results.

The other contribution is the measure of strategic profile.While past studies had relied on perceptions of themanagers concerning the overall organizational strategy,this study used secondary data, publicly available, to drawthe strategic profile of the organization. Using financialinformation, organizations could be compared to acceptedaccounting standards as well as to their peers in theirindustry. This enabled a classification into strategic catego-ries that was devoid from some limitations associated withself-typing (notable social desirability and local perceptionsor imperfect knowledge of the organization as a whole,especially in the case of large organizations).

The results suggest several avenues for future research.First, the extension of the study to include softwaredevelopment would be very interesting. One could presumethat software development would be closer to the core ofthe organization than both IS operations and maintenanceactivities. If it is the case, and if the “distance” from thecore is affecting the relationship between the strategicprofile of the organization and the outsourcing level, astronger relationship between software development out-sourcing and strategic profile should be expected.

An interesting extension would be to include firmperformance in the model. It would be interesting to see iffirms that do not follow the outsourcing pattern suggestedby the theory perform less than firms adjusting theirbehaviour to the theoretical propositions. A larger samplesize would enable the evaluation of a model of fit betweenstrategic profile and outsourcing profile.

The research could also be extended to non-IS activities.By increasing the variety of activities considered, theunderstanding of the relationship between outsourcing andstrategy would be increased. The other avenue suggested bythe results is the extension of the study to offshoring. Onthe one hand, similar results could be expected. On theother hand, since a key advantage argued by practitionersfor offshoring is the cost reduction, the offshoring optionshould be particularly attractive for defenders, who alwaystry to find ways to reduce their costs.

References

Ansoff, H. I., & Stewart, J. M. (1967). Strategies for a technology-based business. Harvard Business Review, 45(6), 71–83.

Aubert, B. A., Patry, M., & Rivard, S. (2004). A transaction costmodel of IT outsourcing. Information & Management, 41, 921–932.

Aubert, B. A., Patry, M., & Rivard, S. (2005). Taking stock on IToutsourcing risk. Database, 36:4, 9–28.

Inf Syst Front (2008) 10:129–143 141

Barney, J. (1999). How a firm’s capabilities affect boundary decisions.Sloan Management Review, 40(3), 137–145.

Bergeron, F., Raymond, L., & Rivard, S. (2004). Ideal patterns ofstrategic alignment and business performance. Information &Management, 41(8), 1003–1020.

Berthon, P. L., Pitt, M., Weing, , & Carr, C. (2002). Potential researchspace in MIS: A framework for envisioning and evaluatingresearch replication, extension, and generation. InformationSystems Research, 13(4), 416–427.

Conant, J. S., Mokwa, M. P., & Wood, S. D. (1987). Managementstyles and marketing strategies: An analysis of HMOs. HealthCare Management Review, 12(4), 65–75.

Conant, J. S., Mokwa, M. P., & Varadarajan, P. R. (1990). Strategictypes, distinctive marketing competencies and organizationalperformance: A multiple measures-based study. Strategic Man-agement Journal, 11(5), 365–383.

Croteau, A.-M., & Bergeron, F. (2001). An information technologytrilogy: Business strategy, technological deployment and organi-zational performance. Journal of Strategic Information Systems,10(2), 77–99.

Croteau, A.-M., Raymond, L., & Bergeron, F. (2001). Comportementsstratégiques, choix et gestion des systèmes d’information:Contribution à la performance. Systèmes d’Information etManagement, 6(4), 5–26.

Dibbern, J., & Heinzl, A. (2006). Selective outsourcing of informationsystems in small and medium sized enterprises», informationsystems outsourcing-enduring themes. In R. Hirschheim, A.Heinzl, & J. Dibbern (Eds.) New perspectives and globalchallenges ((pp. 70–82)2nd ed.). Berlin: Springer.

Freeman, C. (1974). The economics of industrial innovation. Har-mondsworth, England: Penguin.

Grover, V., Cheon, M., & Teng, J. T. C. (1994). An evaluation of theimpact of corporate strategy and the role of informationtechnology on IS functional outsourcing. European Journal ofInformation System, 3(3), 179–190.

Hambrick, D. C. (1980). Operationalizing the concept of business-level strategy in research. The Academy of Management Review,5(4), 567–575.

Hambrick, D. C. (1983). Some tests of the effectiveness andfunctional attributes of Miles and Snow’s strategic types.Academy of Management Journal, 26(1), 5–26.

Hennart, J.-F. (1988). Upstream vertical integration in the aluminiumand tin industries: A comparative study of the choice betweenmarket and intrafirm coordination. Journal of Economic Behav-ior & Organization, 9(3), 281–299.

Jusoh, R., Ibrahim, D. N., & Zainuddin, Y. (2006). Assessing thealignment between business strategy and use of multipleperformance measures using interaction approach. The BusinessReview, 5(1), 51–60.

Masten, S. (1984). The organization of production: Evidence from theaerospace industry. Journal of Law and Economics, 27(2), 403–418.

Miles, R. E., & Snow, C. C. (1978). Organizational strategy,structure, and process. New York: McGraw Hill.

Monteverde, K., & Teece, D. J. (1982). Supplier switching costs andvertical integration in the automobile industry. The Bell Journalof Economics, 13, 206–213.

Nam, K., Rajagopalan, S., Rao, H. R., & Chaudhury, A. (1996). Atwo-level investigation of information systems outsourcing.Communications of the ACM, 39(7), 37–44.

Ng, C., See, P., Gable, G., & Chan, T. (2002). An ERP-client benefit-oriented maintenance taxonomy. The Journal of Systems andSoftware, 64, 87–109.

Pati, N., & Desai, M. S. (2005). Conceptualizing strategic issues ininformation technology outsourcing. Information Management &Computer Security, 13(4), 281–296.

Poppo, L., & Zenger, T. (1998). Measurement explanations for make-or-buy decisions in information services. Strategic ManagementJournal, 19(9), 853–877.

Porter, M. E. (1980). Competitive strategy: Techniques for analyzingindustries and competitors. New York: Free Press.

Rockart, J. F., Leventer, J. S. (1979). Centralization vs. decentraliza-tion of information systems: A preliminary model for decisionmaking, CISR working paper, Sloan School of Management, MIT.

Sabherwal, R., & Chan, Y. E. (2001). Alignment between IT andbusiness strategies: A study of prospectors, analyzers, anddefenders. Information Systems Research, 12(1), 11–33.

Segev, E. (1987). Strategy, strategy-making, and performance in abusiness game. Strategic Management Journal, 8(6), 565–577.

Smith, K. G., Guthrie, J. P., & Chen, M. J. (1989). Strategy size andperformance. Organizational Studies, 10(1), 63–81.

Snow, C. C., & Hrebiniak, L. G. (1980). Strategy distinctivecompetence and organizational performance. AdministrativeScience Quarterly, 25, 317–336.

Tan, F.B. (1997). Strategy type, information technology and perfor-mance: A study of executive perceptions, proceedings of the AISAmericas conference, Indianapolis, Indiana. August, pp. 848–850.

Tavakolian, H. (1989). Linking the information technology structurewith organizational competitive strategy: A survey. MIS Quar-terly, 13(3), 309–317.

Teng, J. T. C., Cheon, M. J., & Grover, V. (1995). Decisions tooutsourcing information systems functions: Testing a strategy-theoretic discrepancy model. Decision Sciences, 26(1), 75–103.

Williamson, O. (1985). The economic institutions of capitalism. NewYork: Free Press.

Young, A., & Scardino, L. (2006). Outsourcing more but enjoying itless: What’s the real problem. Gartner Report, 7 pp.

Zahra, S. A., & Pearce II, J. A. (1990). Research evidence on theMiles–Snow typology. Journal of Management, 16(4),751–768.

Benoit A. Aubert is currently President and Chief Executive Officerof the CIRANO (Centre inter-universitaire de recherche en analyse desorganisations) and Professor at HEC Montreal. He was Director ofResearch at HEC Montréal from 2004 to 2007. Benoit Aubert wasawarded the professorship in Governance and Information Technologyat HEC Montréal in 2003. His main research areas are outsourcing,risk management, and new organization forms (virtual, network,alliances, etc.). He published several articles, book chapters, confer-ence proceedings, and reports on these topics. He also publishedpapers on trust, ontology, and health care information systems. Hefrequently acts as expert consultant on outsourcing decisions, ITstrategy and enterprise reorganizations.

Guillaume Beaurivage is currently Consultant for SAP Canada. Hiswork focuses primarily on Business Intelligence. He completed aMaster of Science at HEC Montreal during which he worked on therelationships between outsourcing and business strategy. His work

142 Inf Syst Front (2008) 10:129–143

was supported by the SSHRC (Social Science and HumanitiesResearch Council) of Canada.

Anne-Marie Croteau is associate professor of MIS at the JohnMolson School of Business at Concordia University. She was thefounder and director of the Graduate Certificate in e-Businessprogram. She is the current director of the Executive MBA and theInternational Aviation MBA programs. Her research focuses oninterorganizational governance of IT, strategic management of IT,and organizational transformation. Her research has been published invarious scientific journals such as Journal of Strategic InformationSystems, Journal of Information Technology, Industrial Management& Data System, IEEE – Transactions on Engineering Management,

Canadian Journal of Administrative Sciences, and InternationalJournal of Knowledge Management as well as in various nationaland international proceedings.

Suzanne Rivard is the HEC Montreal Endowed Chair in StrategicManagement of Information Technology and is a fellow of the RoyalSociety of Canada. Her research interests are in the areas ofoutsourcing of information systems services, IT implementation,software project risk management, and strategic alignment of IT.Suzanne Rivard is a fellow of the Royal Society of Canada. Her workhas been published in such journals as Communications of the ACM,the Journal of Information Technology, the Journal of ManagementInformation Systems, MIS Quarterly, Organization Science and others.

Inf Syst Front (2008) 10:129–143 143