Strategic Alignment | Course Paper

MANAGEMENT SCIENCE

Vol. 35, No. 8, August 1989

Primed in U.S.A.

STRATEGIC ORIENTATION OF BUSINESS ENTERPRISES:

THE CONSTRUCT, DIMENSIONALITY,

AND MEASUREMENT*

N.

VENKATRAMAN

Massachusetts

Institute of

Technology,

Sloan

School

of Management,

Cambridge,

Massaelmsetts

02139

This paper reports the results ofa research study aimed at conceptualizing and developing valid

measurements of key dimensions of a strategy construct—termed Strategic Orientation of Business

Enterprises. This construct is first defined by addressing four theoretical questions of scope; hi–

erarchical level; domain; and intentions versus realizations, and then conceptualized in terms of

six dimensions. Subsequently, operational indicators are developed for the six dimensions in

terms of managerial perceptions across 200 business units in a field study. An evaluation ofthe

measurement properties within an analysis of covariance structures framework indicated that the

operational measures developed here largely satisfy the criteria for unidimensionality, convergent,

discriminant, and predictive validity. Implications and lines of extensions are outlined.

(STRATEGY—BUSINESS LEVEL; MEASUREMENT SCALES; CONHRMATORY FACTOR

ANALYSIS)

1.

Introduction

Research in strategic management is at critical cross–roads today, with increased em–

phasis on developing theoretical concepts and testing empirical relationships rooted in

such concepts. Such a transformation critically depends on the conceptualization and

measurement ofthe

strategy concept

and its

derivative constructs

(Schendel and Hofer

1979;

Hambrick 1980; Venkatraman and Grant 1986). Thus, in spite of several discus–

sions on alternate approaches to operationalizing strategy (Ginsberg 1984; Hambrick

1980;

Pitts and Hopkins

1982;

Snow and Hambrick 1980), the linkage between theoretical

definitions and their corresponding measures has been generally weak. Most existing

measures for the strategy construets are either nominal (and/or single–item) scales that

have questionable measurement properties or multi–item seales whose measurement

properties (such as reliability, and unidimensionality, convergent and discriminant validity

as well as nomologieal validity) have not been systematieally assessed (Venkatraman

and Grant 1986).

This research intends to contribute to the measurement stream of strategic management

research with an explicit recognition that measure development cannot be divorced from

the broader theoretieal network (Cronbaeh

1971;

Hempel 1952).‘ The specific aim is to

develop and validate a set of operational measures for a particular eonceptualization of

strategy. The expectation is to provide an initial set of operational measures with strong

support in terms of their measurement properties that can be used by other researchers

for theory–testing as well as further extensions/refinements.

The paper is divided into four sections. The first section provides the background for

this research through a brief historical trace ofthe various approaches to the measurement

* Accepted by Vijay Mahajan; received February 18, 1988.Thispaperhasbeen with the author for 3^ months

for 3 revisions.

‘ The traditional notion ofa measurement perspective is that it is a procedure for assigning “numbers“ to

objects or events according to a particular rule. If one subscribed to this view, measurement is to be regarded

as an empirical requirement that is detached from the theoretical propositions and relationships. However, a

more current view is that measurement is an intellectual and empirical activity that provides meaning to the

theoretical variables (e.g., Hempel 1952; Carnap 1966). This paper is explicitly predicated on this view.

942

0025-1909/89/3508/0942101.25

STRATEGIC ORIENTATION OF BUSINESS ENTERPRISES 943

of Strategy constructs. In the second section, a theoretical circumscription ofthe proposed

strategy construct is presented. The research method including the data analytical approach

and results are presented in the third section. The fourth section discusses the implications

ofthe results and outlines a set of future research directions.

2.

Background

2.1.

Alternate Approaches to Strategy Measurement

Narrative Approach. This reflects the case-based tradition of business policy, predicated

on a view that the complex characterization of strategy should only be described in its

holistic, and contextual form. Since the distinctivetiess ofthe strategy concept lies in its

uniqueness to a particular setting (e.g., Andrews 1980), the implication is that strategy

can (and should) be best described verbally, and any attempt to develop a measurement

scheme will be incomplete. This reflects views such as: strategy is “an organizational

process forever in motion“ or as “the interface between the organizational process of

formulation and the organizational process of implementation“ (Andrews 1980); or that

this concept is to be viewed “more as a fluid to be worked with rather than a thing to be

actualized“ (Evered 1983, p. 61). Such philosophical abstractions ofthe strategy concept

are best captured in narrative descriptions than through finely-calibrated measuring

schemes.

While such an approach has its role for conceptual developments, it has limited use

for testing theories (see Hempel 1952) pertaining to the effectiveness of various strategies

under differing environmental, organizational, and temporal conditions. Further, if we

accept the need to adopt mid-range theorizing and/or contingency theory perspectives

(Steiner 1979; Harrigan 1983; Ginsberg and Venkatraman 1985), the narrative ap–

proaches should give way for superiot“ schemes.

Classificatory Approach. The first movement away from the idiosyncratic, narrative

descriptions of strategy is reflected in the development of strategy classifications—either

conceptual or empirical. The conceptual classifications that are inductively derived are

termed as’typologies’, and prominent ones include: Rumelt( 1974); Hofer and Schendel

(1978);

Miles and Snow (1978); Porter (1980); and Wissema, Van der Pol, and Messer

(1980).

The distinguishing feature is that such typologies are rooted in a set of parsi–

monious classificatory dimensions or conceptual criteria. While typologies are best known

for their conceptual elegance, they do suffer from an inherent weakness in that “it is

fairly easy to find a single dimension on which a typology can be based and which will

. . . support any given philosophical orientation“ (Carper and Snizek 1980, p. 70).

The empirical classifications are termed as ‘taxonomies’, and prominent ones include:

Miller and Friesen (1978) and Galbraith and Schendel (1983). These reflect empirical

existence of internally consistent configurations, but it is important to recognize that

their development is sensitive to the choice of underlying dimensions as well as the

analytical method used to extract the taxonomies (Hambrick 1984; Miller and Friesen

1984).

Thus, while it serves to capture the comprehensiveness and integrative nature of

strategy through its internal coherence, it does not reflect the ‘within-group‘ differences

along the underlying dimensions.

Comparative Approach.’^ The aim ofthe third approach is to identify and measure

the key traits (also termed as dimensions) ofthe strategy construct. Consequently, the

focus is less on categorization into one particular cell ofthe typology (or on the devel–

opment ofa parsimonious taxonomy) but on measuring the differences along a set of

^

The term, ‘comparative approach‘ has been used within management research for several different reasons.

In this paper, the use of this term is consistent with Hempel‘s (1952, p. 55) call for the use of scalars for

measuring concepts and is distinguished from the classificatory approach that seeks to measure the presence or

absence ofa phenomenon.

944 N. VENKATRAMAN

characteristics that collectively describe the strategy construct. Over the years, several

researchers have sought to develop measurement schemes to capture the differences in

strategies (Miller and Friesen 1984; Dess and Davis 1984; Hambrick 1983).

The attractiveness of this approach lies in its ability to decompose the variation that

is seen across different strategy classifications into more ‘fine–grained‘ differences along

each underlying traits (or

dimensions).

This requires that the traits have specific theoretical

content as well as adequate operational measures since this approach has its roots in the

philosophy of empirical science (Hempel 1952; Carnap 1966). Given that “all theories

in science concern statements mainly about constructs rather than about specific, ob–

servable variables“ (Nunnally 1978, p. 96), the process of construct development and

measurement is at the core of theory construction.

However, the present state of attention to construct measurement in strategic man–

agement is inadequate. Researchers continue to propose and employ measures without

corresponding tests for unidimensionality, reliability, convergent, discriminant and pre–

dictive validity (Venkatraman and Grant 1986). In the absence ofa systematic basis to

evaluate the adequacy of measurements, confidence in research results is considerably

eroded, which implies that the managerial implications derived from such results may

be questionable.

2.2.

Study

Objectives

This study adopts the comparative

approach

to strategy measurement and aims at

linking the conceptual definition and empirical indicators ofa single construct, termed

as Strategic Orientation of Business Enterprise. Specifically, the objective is to arrive at

a set of operational indicators for important dimensions of this construct that meet

minimal criteria of measurement. It is hoped that the results of such an exercise will be

of use to strategy researchers either directly in their research contexts or as a basis for

refinement and extension in the best tradition of cumulative theory building and testing.

2.3.

Premises

Premise

1:

Strategy Research Within

a‘

Variance’

Perspective Requires

Valid Measures.

Following Mohr (1982), strategy research can be divided into the ‘process‘ and the

‘variance‘ perspectives. The former

is

characterized by research studies that seek to describe

the

process

of strategy formation (Mintzberg 1978; Mintzberg and Waters 1982, 1985;

Miles and Snow 1978), resource allocation (Bower 1970), and internal corporate ven–

turing (Burgelman 1983) rather than explain the variation in a dependent variable through

a set of independent variables. In contrast, studies subscribing to the ‘variance‘ perspective

aim at theory testing through confirmation/disconfirmation ofa specific array of hy–

potheses. For instance, several PIMS–based research seek to explain variations in business

performance through a set of independent variables reflecting both strategic resource

deployments and environmental characteristics (Hambrick 1983; Prescott 1986; for re–

views,

see Ginsberg and Venkatraman

1985;

and Ramanujam and Venkatraman 1984).

The typical research practice in the variance perspective is as follows: the theory to be

tested is stated in terms ofthe language ofthe researchers as relationships among a set

of theoretical constructs; the data for testing the theory are obtained through indicators

described in the language of those providing the data (typically senior managers of target

businesses); and the assessment of correspondence between the theoretical constructs

and their empirical indicators is construct validation—which is a prerequisite for the

interpretation ofthe substantive relationships among the theoretical constructs based on

the observed relationships among the empirical indicators (Hempel 1952; Bagozzi 1980;

Schwab 1980). Thus, our first premise is that strategy research within a ‘variance‘ per–

spective is critically dependent on the quality ofthe operational measures.

Premise

2:

The

Search for

a Universal Conceptualization

of

Strategy

is Futile. Given

STRATEGIC ORIENTATION OF BUSINESS ENTERPRISES 945

the eclectic nature ofthe strategy field, it is premature to restrict the number and diversity

of approaches to conceptualize the strategy construct. However, it is important that each

conceptualization is accompanied by corresponding operational schemes. As Cameron

and Whetten noted: “Constructs such as intelligence, motivation, or leadership—whose

construct space, by definition ... is not bounded—have been better understood as

limited aspects of their total meaning have been measured“

(1983,

p. 276). The impli–

cation is that a particular conceptualization—which builds from the existing theoretical

perspectives in strategy research—should serve as the referent for the development of

operational measures.

Thus,

the output of this study would be a set of valid measures anchored in a particular

theoretical conceptualization of the strategy concept. This can be subsequently used in

different contexts to further refine/extend the conceptualization as well as the operational

measures (e.g., convergent and discriminant validity of other measures). Such an approach

reflects a cumulative theory-building perspective where progress is made by successively

testing the efficacy ofthe measures in varying theoretical networks (Cronbaeh 1971).

Premise 3: Construct Measurement Is at Least as Important as Examination of

Sub-

stantive Relationships. Any scientific research discipline can be viewed in terms of two

interrelated streams—substantive and construct validation (or measurement). The former

reflects the relationships among theoretical constructs inferred through empirically ob–

served relationships, while the latter involves the relationships between the results obtained

from empirical measures and the theoretical constructs that the measures are purported

to assess (Schwab 1980). Although these are to be treated as end-points in a continuum,

within strategic management, the former has been overemphasized with correspondingly

little attention to the latter. Since substantive relationships are fundamentally dependent

on the measurement relationships, it is necessary to recognize that construct measurement

is at least as important as the examination of substantive relationships.

While most amongst us would not quarrel with the need to assess the measurement

properties of strategy constructs, the degree of research attention to construct measurement

is far less than that for substantive issues. This could be partly attributed to the fact that

most researchers trade-off their efforts in favor of theoretical relationships among their

constructs with an implicit belief in the adequacy of their measures. This paper takes the

opposite view and devotes more attention to the development and validation of the

measures. While we recognize that measure devleopment cannot be carried out in isolation

ofthe theoretical network, it is necessary to note, at the outset, that the research questions

focus more on the correspondence between the measures and their constructs than on

the relationships among the constructs per se.

3.

Circumscribing the Theoretical Construct

A major task in conceptualizing a theoretical construct relates to the specification of

its boundaries. For strategy constructs, this is particularly complex given the wide array

of differences in terminology, disciplinary orientations as well as underlying assumptions.

The conceptual domain ofthe proposed construct is delineated by anchoring it with four

theoretical questions that are central within strategic management research. These are:

(a) Scope: should the definition distinguish between “means“ and “ends”?; (b) Hier-

archical Level: should the construct be defined at a particular level ofthe organizational

hierarchy or should it be level–free?; (c) Domain: should the domain be restricted to

some ‘parts‘ (i.e., some functional focus) or cover a broader perspective?; and (d) In-

tentions Versus Realizations: is the distinction between ‘intended‘ and ‘realized‘ strategies

relevant for conceptualizing and measuring this construct?

The logic in using these questions is to provide a systematic framework for concep–

tualizing the construct, and use it as a referent for operational schemes. We believe that

946 N. VENKATRAMAN

these four questions collectively address important aspects of c o n c e p t u a l i z i n g the strategy

construct. Further, we recognize the controversial nature of these questions, where the

opposing viewpoints are valid for alternate strategy conceptualizations. Thus, our purpose

in selecting these questions is

not

to reconcile these opposing views but to take a particular

stance to guide the measure development

exercise.

In other

words,

our aim is to explicitly

link the conceptual domain (in terms of t h e s e four questions) with the operational domain.

This ensures that the isomorphic nature ofthe linkage between construct definition and

construct measurement is maintained.

3.1.

Scope: ‘Means’

Versus

‘Means and Ends’

Some authors prefer to separate strategies, viewed as means from goals, viewed as ends

(e.g., Hofer and Schendel 1978; Schendel and Hofer 1979), while others prefer to treat

strategy as a comprehensive concept encompassing both goals (ends) and means (e.g.,

Andrews 1980). Within a general management perspective (Andrews 1980)—where

formulation and implementation are intertwined and goals and means interlinked, there

are no particular reasons to separate the two. However, the newer strategic management

research paradigm (Schendel and Hofer 1979) explicitly separates goals (goal structure

and goal formulation) from strategies (formulation, evaluation, and implementation).

In this vein, MacCrimmon (forthcoming) views strategic management in terms of three

components: ends (goals); means (actions); and conditions (contexts). Thus, a relevant

question for strategy research study could be to examine the efficacy of a particular

strategy (means) to attain certain ends (goals) within a particular setting (conditions).

This study separates strategies from goals, consistent with the views of Hofer and

Schendel (1978), Schendel and Hofer( 1979), MacCrimmon (forthcoming), and Mintz–

berg and Waters (1985). Thus our focus is on the means adopted (i.e., resource deploy–

ment patterns) to achieve the desired goals (i.e., ends, purposes, and objectives). This

choice is advantageous in that it provides a

restricted scope

of domain for the construct

such that the measures of strategy can be used to examine relationships between strategies

and goals in different contexts. Had the domain ofthe strategy construct included goals,

the link between strategy and goals would have been isomorphic, resulting in an inability

to examine the nature of linkages between strategies and goals as well as explore notions

of equifinality (i.e., the possibility of alternate combinations of

strategies

and conditions

to achieve the same ends).

3.2.

Hierarchical

Level

There is reasonable consensus on the three-level categorization ofthe strategy concept—

corporate, business, and functional (Grant and King 1982; Hax and Majluf 1984; Hofer

and Schendel 1978). Strategy at the corporate level is primarily concerned with answering

the question: “In what set of businesses is this organization engaged in?“ and is viewed

in terms ofthe pattern of linkages among the different businesses constituting the corporate

profile (Rumelt 1974). Strategy at the business level (also termed as business–unit, or

strategic business unit, SBU) is concerned with the following question: “how do we

compete effectively in each of our chosen product–market segment?“ Accordingly, the

theoretical issues at this level relate to the requirement of matching environmental op–

portunities and competitive threats with the efficient deployment of organizational re–

sources (Bourgeois 1980; Grant and King 1982; Hofer and Schendel 1978). In addition,

the general theme of competitive strategy and the attainment of competitive advantage

is an important issue at the business strategy level (Porter 1980, 1985). Strategy at the

functional level focuses on the maximization of resource productivity within each ofthe

specified functions (such as operations or marketing) and is generally derived from the

business level strategy (Hofer and Schendel 1978).

STRATEGIC ORIENTATION OF BUSINESS ENTERPRISES 947

The construct is defined at the level ofthe business for the following reasons. Since

strategic management aims to integrate key functions toward adopting a general man–

agement perspective (Schendel and Hofer 1979), the functional level is not particularly

important. Similarly, as organizations increasingly diversify (Chandler 1962; Rumelt

1974) and operate in multiple product-market segments, the corporate level is too ag–

gregated for understanding the strategic responses to environmental influences such as

competitive moves, technological changes, as well as entry and exit of competitors. Thus,

business unit is an appropriate and useful level of analysis.

3.3.

Domain: Distinguishing Between ”Parts” Versus “Holistic” Perspectives

Following Hambrick (1980), a distinction is made between “parts“ of strategy and

strategic typologies reflecting a more holistic or interconnected perspective. For example,

the categorization proposed by Buzzell, Gale, and Sultan (1975) and Porter (1980) reflect

the competitive or the product-market sector of the overall strategy concept, while def–

initions and conceptualizations of strategy adopted by Miller and Friesen (1978), Miles

and Snow (1978) and Mintzberg (1978) reflect a much broader and holistic perspective.

This study focuses on a broader notion of strategy based on two reasons. One, a focus

on one or two areas such as marketing (e.g., market share positions, new product intro–

duction frequency or geographical coverage) or manufacturing (e.g., research and de–

velopment and product quality) taps only a functional orientation which does not truly

reflect the general strategic orientation of a business. The second reason is that although

many holistic definitions of strategy have been offered (Andrews 1980; Chandler 1962;

Grant and King 1982; Miles and Snow 1978; Mintzberg 1978), measurement schemes

have not adequately operationalized such definitions. Thus, by operationalizing the con–

struct in holistic terms, the results of this study can be used to more directly employ such

richer definitions in strategy research.

3.4. Intentions Versus Realizations

The fourth issue to be addressed in developing the theoretical framework ofthe construct

relates to the distinction between ‘intended‘ and ‘realized‘ strategies (Mintzberg 1978).

As argued by Mintzberg and Waters “conceiving strategy in terms of intentions means

restricting research to the study ofthe perceptions of what those who, it is believed, make

strategy intend to. And that kind of research—of intentions devoid of behavior—is simply

not very interesting or productive“ (1982, p. 465). By viewing ‘realized‘ strategy as a

“pattern in a stream of decisions“ (Mintzberg 1978), strategies become consistencies in

the behavior of organizations (Mintzberg and Waters 1982). In line with our earlier

decision to view strategy as a pattern of critical decisions, along a holistic perspective,

and in conjunction with the arguments by Mintzberg and Waters, the focus is on the

realized strategies.

Thus,

these four issues serve to delineate the domain ofthe construct. The next task

is to identify critical dimensions within this domain, and this is addressed below.

3.5.

Specifying the Construct’s Dimensionality

A priori versus a posteriori. Given the complexity ofthe strategy concept, it is logical

to assume that Strategic Orientation of Business Enterprises (STROBE) is a multidi–

mensional construct. However, its dimensionality can be arrived at one of two different

ways.

One is a priori, namely develop the different dimensions ofthe construct based

on the theoretical perspectives that guided the construct definition. Thus, the dimen–

sionality is pre–specified and subsequently validated by testing against data, where the

data analytic scheme is viewed as a means towards confirmation or rejection of the

948 N. VENKATRAMAN

theoretical dimensions derived. Illustrative examples

are:

Dess and Beard‘s (1984) study

on Alddch’s (1979) dimensions of organizational environments; and the Aston Program

on organization structure (Pugh and Hiekson 1976).

The other approach is not to pre-specify the dimensions but to empirically derive them

a posteriori

through data analytic techniques such as factor analysis or multi–dimensional

scaling (e.g., Blackburn and Cummings 1982). This approach is generally considered to

be ‘theory-free‘ and is adopted only in those cases where little theoretical basis exists for

a

priori

deriving the dimensions. In such cases, the danger is that the dimensions may

not be interpretable for use in substantive research and that they may not be stable over

different study settings. More importantly, the data analytic scheme occupies a central

role in the conceptualization and operationalization ofthe construct.

Given our intention to develop operational measures for important theoretical di–

mensions of strategic orientation rather than uncover dominant dimensions through

data-analytic methods, we decided to specify the dimensions a priori. Accordingly, six

important dimensions of strategic orientation are identified for this study. A brief de–

scription ofthe dimension is provided in the following paragraphs.

Aggressiveness. This dimension refers to the posture adopted by a business in its

allocation of resources for improving market positions at a relatively faster rate than the

competitors in its chosen market. These may be based on product innovations and/or

market development (Miles and Cameron 1982) or high investments to improve relative

market share and competitive position (e.g., Buzzell, Gale and Sultan 1975; Fruhan

1972;

Hofer and Schendel 1978). It also reflects the notions of’explosion‘ (i.e., improve

competitive position in the short-run) as conceptualized by Wissema et al. (1980), and

the strategy of’multiplication’, i.e., expansion of market share by multiplying as noted

by Vesper (1979) and the pursuit of market share

as

an important path towards achieving