Journal of Personal Selling & Sales Management, vol. XXVI, no. 2 (spring 2006), pp. 95–113.
© 2006 PSE National Educational Foundation. All rights reserved.
ISSN 0885-3134 / 2006 $9.50 + 0.00.
In the sales and marketing literatures, there has been increas-
ing attention paid to the role of shared information, opera-
tional linkages, and cooperation as firms in business markets
shift to more closely coupled relationships (cf. Cannon and
Perreault 1999). Concurrent with these shifts in the relation-
ships among firms, there has been a reinvention of the tradi-
tional sales role. With an increased emphasis on relationship
marketing strategies (Anderson 1996), the sales rep has a re-
sponsibility to serve as a consultant to the customer and to
strengthen the buyer–seller relationship by helping to develop
the customer’s business and achieve customer satisfaction (Liu
and Leach 2001). To enhance sales performance and buyer–
seller relationships, firms in a wide variety of industries—
ranging from consumer package goods (CPG) and financial
services to chemicals and energy—have made substantial in-
vestments in information technology (Shoemaker 2001). Un-
fortunately, these investments continue to be vendor driven,
resulting in a high failure rate. The situation is so grim that
recent academic research characterizes sales technology imple-
mentations as a virtual “minefield” (Speier and Venkatesh
2002). Under such circumstances, it is not surprising that a
“day of reckoning” follows such investments in which astute
sales managers need to be in a position to justify their invest-
ments (Erffmeyer and Johnson 2001).
This paper proposes using a diagnostic approach for com-
posing and testing process models that link desired outcomes
with the behavioral tasks that influence them. The approach
builds on the long-standing tradition of modeling salesper-
son behaviors in the sales force and channels literatures (cf.
Behrman and Perreault 1984; Brown and Peterson 1993).
Specifically, we develop and advance hypotheses about how a
salesperson’s orientation toward information technology af-
fects two facets of performance—effectiveness in dealing with
customers and efficiency in performing internal tasks (such
as recommending how company operations can be improved).
In this regard, the model posits both indirect and direct posi-
tive impacts from a stronger technology orientation by the
In this paper, we argue that sales reps with greater technol-
ogy orientations are better able to leverage information (i.e.,
make available information more effective), which should, in
turn, facilitate sales planning and adaptive behaviors—smart
selling behaviors (tasks) that are known to be related to effec-
tive selling (cf. Spiro and Weitz 1990). Simply put, we propose
a means for assessing “how” sales technology implementations
SALES TECHNOLOGY ORIENTATION, INFORMATION EFFECTIVENESS,
AND SALES PERFORMANCE
Gary K. Hunter and William D. Perreault Jr.
Sales managers need a practical means for evaluating returns from investments in sales technology implementations
(including sales automation and sales-based customer relationship management systems). This research proposes a behav-
ioral process model approach that can be applied to evaluate sales technology implementations. We develop and test the
model with data collected from the sales force of a major consumer packaged goods company. The results indicate that a
salesperson’s technology orientation has a direct impact on internal role performance, and it affects performance with
customers through a double-mediated mechanism involving the effective use of information and smart selling behaviors
(planning and adaptive selling). Sales managers can influence sales technology orientation by providing better internal
technology support, considering technology orientation along with customer’s approval of technology in account assign-
ments, and understanding the probability of negative effects through a salesperson’s experience. In our sample, salesper-
son experience correlates with age, suggesting a “generation gap” effect on sales technology orientation.
Gary K. Hunter (Ph.D., University of North Carolina), Assistant
Professor of Marketing, College of Business Administration, Florida
International University, email@example.com.
William D. Perreault Jr. (Ph.D., University of North Carolina),
Kenan Professor of Business, Kenan-Flagler Business School, Uni-
versity of North Carolina, firstname.lastname@example.org.
The authors appreciate helpful suggestions from three anonymous
reviewers and two JPSSM editors (Greg Marshall and Ken Evans).
They acknowledge helpful comments on earlier versions of this manu-
script from Gary Armstrong, Ken Bollen, Peter Dickson, Jay
Klompmaker, Charlotte Mason, Al Segars, Jagdip Singh, Harish
Sujan, and Valerie Ziethaml. The authors gratefully appreciate sup-
port for this research from the Kenan-Flagler Business School at the
University of North Carolina, the College of Business Administra-
tion at Florida International University, the Center for Services Lead-
ership at the W.P. Carey School of Business of Arizona State University,
and an anonymous consumer packaged goods manufacturer.
96 Journal of Personal Selling & Sales Management
affect key aspects of sales performance that are important for
sales managers when evaluating salespeople in a modern rela-
SALES AUTOMATION, AND
While there is an extensive and growing literature on infor-
mation technology, sales automation, and sales-based customer
relationship management (CRM), we focus here on research
that is most relevant to this study to help frame and position
our perspectives. There is a substantial body of excellent re-
search on the impact of information technology on business
performance that is relevant to our investigation of the effects
of sales technology on sales performance. Several researchers
have addressed the relationship between information tech-
nology and organizational performance by modeling impor-
tant organizational inputs such as dollar investments in
information technology, and outputs such as financial returns
(cf. Bharadwaj, Bharadwaj, and Konsynski 1999). Most of
the work that directly models the impact of information tech-
nology implicitly treats the organization as a “black box”—so
the impact of information technology on individual tasks,
specific processes, or intermediate outcomes (such as the qual-
ity of services) is not explicitly evaluated.
A second major stream of research on the impact of infor-
mation technology relies on user evaluations, as opposed to
financial returns, to measure the success of information tech-
nology. One reason to move to this approach is that a finan-
cial investment in a particular technology does not assure that
it is used as intended by members of the organization. Thus,
there is a long-standing tradition of finding ways through
which salespeople can optimize individual sales technology
applications. JPSSM published an insightful and forward-
looking series of papers on applications in personal selling
and sales management (for examples, see Collins 1984, 1985,
1989; Comer 1981–82; Swenson and Parrella 1992) and on
the role sales managers and salespeople perform in the firm’s
marketing information systems (Evans and Schlacter 1985;
Klompmaker 1980–81). However, renewed interests in sales
technology have spawned an emerging literature that focuses
on two areas: sales–CRM (Ahearne, Srinivasan, and Weinstein
2004; Pass, Evans, and Schlacter 2004; Plouffe, Williams, and
Leigh 2004; Shoemaker 2001; Zablah, Bellenger, and Johnston
2004) and sales automation technologies (Jones, Sundaram,
and Chin 2002; Parthasarathy and Sohi 1997; Pullig,
Maxham, and Hair 2002; Schillewaert et al. 2005; Speier and
The burgeoning literature and use of CRM tools has been
so pervasive in modern practice that CRM has evolved as
both a business philosophy and a technology (Johnston and
Marshall 2005, p. 128; see Plouffe, Williams, and Leigh 2004
for varying definitions of CRM across stakeholders). CRM
often refers to “the use of technology to manage customer
interactions and transactions” (Zoltners, Sinha, and Zoltners
2001, p. 389). Meanwhile, sales (force) automation (SFA)
vendors stress that sales reps who complete routine tasks faster,
easier, or better become more effective. Recent research may
help to bridge the gap between CRM and SFA technologies
(Widmier, Jackson, and McCabe 2002) and set up agendas
for sales technology research (Tanner and Shipp 2005). The
CRM and SFA terminology does not establish a mutually
exclusive dichotomy. For example, contact management soft-
ware is a common feature across both SFA and sales–CRM
software packages and is intended to help sales reps manage
leads, track all communications with customers, schedule
follow-ups, and handle time management and planning
However, salespeople use technologies that go beyond clas-
sification as either CRM or SFA tools—including hardware
and software tools that can aid their performance of sales tasks.
For example, salespeople use cell phones for communications
and spreadsheets for analysis, and many sales managers con-
sider these tools critical to the firm’s sales technology portfo-
lio. However, few vendors or sales technology specialists
classify cell phones or spreadsheets as either CRM or SFA
Here, we refer to sales technology as information technolo-
gies that can facilitate or enable the performance of sales tasks.
As such, sales technology represents the broad range of infor-
mation technologies used by salespeople, and we consider both
sales–CRM and SFA tools as subsets of sales technology.
Jones, Sundaram, and Chin (2002) highlight the need for
salespeople to adopt sales technology in forming customer al-
liances and find that personal innovativeness, attitude toward
the new sales technology, and facilitating conditions influ-
ence sales technology infusion. While it is important to moti-
vate user acceptance of information technologies (Davis 1989;
Venkatesh and Davis 2000), information technology and sales
technology scholarship need to get beyond the adoption issue
(Ahearne, Jelinek, and Rapp 2005). One way of doing this is
to properly align the CRM program with employees, pro-
cesses, and technology (Zablah, Bellenger, and Johnston 2004).
Consistent with these research priorities, this paper builds on
previous research on information technology, sales–CRM, and
sales automation to develop and test a theoretical model that
demonstrates how sales managers can diagnose sales technol-
ogy implementations (linking technology through behavioral
selling processes to performance outcomes).
Specifically, in this study, we develop and test a normative
process model with sales force data collected with the coop-
eration of a major CPG company. Structural equation mod-
eling (SEM) is used to test the interrelationships among sales
Spring 2006 97
technology orientation, its antecedents (internal technology
support, customer approval of sales technology use, and sales-
person experience), and its consequences (information effec-
tiveness, smart selling tasks, and sales performance outcomes).
Because the process model is normative, statistically signifi-
cant links in the model provide evidence of positive effects of
sales technologies, whereas links that are not statistically sig-
nificant illuminate areas in which the seller is not realizing
expected returns. In such a case, managers should reassess
the overall fit between the sales process tasks and the sales
technology portfolio and, in turn, consider altering either
the task or the sales technology portfolio. While this approach
provides a means for diagnosing sales technology implemen-
tations, it also highlights areas where sales technology does
not have effects (as reflected by the absence of causal paths
among constructs). While the process model tested here fo-
cuses on the important new area of sales technology and is
based on constructs and theory drawn from the sales man-
agement literature, the benefits of this process modeling ap-
proach are readily adapted to measures and processes specific
to other information technology–performance relationships
Figure 1 overviews a conceptual model of key antecedents
and consequences of a sales technology orientation. In a field
sales setting, beyond the actual information technology em-
ployed by the organization, the important effects related to
successful sales technology implementations are the behav-
iors of the salesperson—and thus, we model key salesperson
behaviors here. In this section, we provide the logic,
conceptualizations, and definitions underlying this study.
While much of the extant literature on sales technology fo-
cuses on simply motivating user acceptance (Ahearne, Jelinek,
and Rapp 2005), our model centers on the salesperson’s dis-
positions toward using sales technology (sales technology ori-
entation), how sales managers can influence this important
construct, and how sales technology orientation subsequently
influences two key aspects of sales performance (including
sales force objectives that are more internally focused within
the sales organization and those that are more externally fo-
cused on buying organizations). Moreover, we propose and
test the mechanisms through which those effects occur.
Namely, we outline a process in which sales technology ori-
entation improves a salesperson’s ability to use information
effectively, which, in turn, improves smart selling behaviors
(adaptive selling and sales planning), which have been posi-
tively linked to sales performance in previous studies (Sujan,
Weitz, and Kumar 1994). Our logic draws upon and inte-
grates previous research in sales technology with perspectives
from social exchange theory (Thibaut and Kelley 1959), ex-
pectancy theory (Vroom 1964), and boundary role theory
(cf. Adams 1976; Organ 1971).
Sales Technology Orientation
The firm’s provision of sales technologies does not ensure they
will be used equally by different sales reps. It is usually at the
salesperson’s discretion to choose how much to rely on indi-
vidual technologies. Sales technology orientation refers to the
salesperson’s propensity and analytical skills for using a port-
folio of firm-provided information technologies to perform
tasks relevant to the sales role.
In addition to creating the right organizational culture to
optimize sales performance (Oliver and Anderson 1994), sales
managers influence the salesperson’s orientation toward us-
ing technology and developing the necessary analytical skills
to optimize its use. Specifically, social exchange theory
(Thibaut and Kelley 1959) suggests that salespeople’s orien-
tation toward technology can be enhanced through actions
that improve the comparison level of an existing sales tech-
nology culture to an alternative culture that more favorably
supports sales technology.
Pullig, Maxham, and Hair (2002) demonstrate the impor-
tance to sales organizations of creating the right conditions
for successful SFA implementations. Various factors (some
controlled by the firm and some not) may influence a sales
rep’s technology orientation. For parsimony, our model con-
siders three key factors suggested by existing literature as be-
ing particularly relevant to sales contexts—company internal
support for sales technology, customer approval of sales tech-
nology, and experience in the sales role.
Internal technology support is the extent to which the firm
provides the salesperson with resources needed to use sales
technology. Social exchange theory (Thibaut and Kelley 1959)
suggests that managerial actions and attitudes concerning in-
formation technology should influence salesperson behavior.
Management support for sales automation plays an impor-
tant role in ensuring that a sales force realizes performance
returns from its investments in sales technology (Jones,
Sundaram, and Chin 2002; Schillewaert et al. 2005). Such
support may take various forms, including development of
custom systems, training, and changes in the systems of evalu-
ation and compensation. For example, in a field study of bro-
ker workstations, Lucas and Spitler (1999) found that
management support and the nature of the task requirements
were important in predicting use of technology, as users’ atti-
tudes are influenced by their colleagues’ attitudes. Internal
technology support thus signals the importance an organiza-
tion places on sales technology.
Hypothesis 1: Internal technology support will positively
affect a salesperson’s technology orientation.
98 Journal of Personal Selling & Sales Management
Social exchange theory (Thibaut and Kelley 1959) sug-
gests that buyer’s attitudes and behaviors can influence the
salesperson’s orientation toward technology. Moreover, to be
effective, salespeople must, in general, be responsive to cus-
tomer needs and requests, including those related to infor-
mation. Vroom’s (1964) expectancy theory of motivation,
which proposes that individuals are motivated to conduct
behaviors that increase their expectancy, instrumentality, or
valence for rewards, is established in the marketing and sales
research (cf. Oliver 1974; Teas 1981; Walker, Churchill, and
Ford 1977). Customer approval of sales technology represents
the extent to which a customer signals expectations for tech-
nology use by the salesperson. For example, a customer who
expects a salesperson to provide an analysis of both the ben-
efits and costs of his recommendations reinforces the effec-
tive use of sales technology, whereas an “old style” buyer, who
seems content with a well-organized pitch, a joke, and a slap
on the back, does not. Thus, customer expectation about how
the salesperson uses, analyzes, and communicates data should
influence the information technologies used by sales reps (and
supported by suppliers).
Hypothesis 2: Customer approval of sales technology will
positively affect a salesperson’s technology orientation.
The model also includes salesperson experience as an ante-
cedent of sales technology orientation. Although their study’s
finding were not conclusive, Ko and Dennis (2004) argue
that highly experienced sales reps will gain the least perfor-
mance benefits from SFA system use—based on an SFA
system’s greater potential to provide useful knowledge to less
experienced reps. Others have argued that age (which corre-
lates with experience) has a negative effect on usage (Morris
and Venkatesh 2000; Speier and Venkatesh 2002). In prac-
tice, many companies struggle with the problem that sales
technologies are creating a “generation gap” among sales-
people. Younger salespeople often are more “technology liter-
ate,” based in part on more exposure to computing and
technology during the educational process (that is, younger
reps have high-technology self-efficacy). Consistent with ex-
pectancy theory (Teas 1981; Vroom 1964), we argue that sales-
people with more experience learned how to be effective
without the use of modern sales technologies; thus, they esti-
mate less instrumental fit between the tasks they need to per-
form and the use of firm-provided sales technologies. Less
instrumentality yields less motivation to adopt new technolo-
gies (Teas 1981; Vroom 1964).
Hypothesis 3: Salesperson experience will negatively affect
a salesperson’s technology orientation.
Information Effectiveness and Working Smart
Salespeople are knowledge workers. Our model reflects the
knowledge-dependent aspects of the sales role by including
Conceptual Model of Effects of Sales Technology Orientation
Spring 2006 99
information effectiveness, which is the value of available infor-
mation for working with and gaining commitment from cus-
tomers. Menon and Varadarajan (1992) highlight the vital
importance of an individual’s use of market information.
Today, salespeople have extensive access to data (including,
for example, past shipments to distributors, retail store sales,
consumer buying habits, and product performance charac-
teristics), but to be successful, they need to convert available
data into information that can be used effectively toward de-
veloping and advancing recommendations and proposals that
balance sales objectives with customer objectives. A central
purpose of information technology is to help users convert
data into effective information. When salespeople possess
analytical skills and information technology know-how, they
can transform the massive amounts of available data into use-
ful and effective information for their buyers.
Hypothesis 4: More technology-oriented salespeople will use
information more effectively.
The “smart selling” literature confirms the normative pre-
scription that salespeople should plan for specific buyer in-
teractions (Sujan, Weitz, and Kumar 1994) and tailor their
behaviors to those interactions (cf. Spiro and Weitz 1990).
Effective information is a necessary input for any meaningful
planning process. Sales technology should facilitate or enable
increased information effectiveness for a wide variety of presale
planning activities. For example, spreadsheet analysis of past
sales data can make the information more effective, which, in
turn, improves a salesperson’s planning for a sales interaction
by making sales forecasts more accurate and timely.
Hypothesis 5: Information effectiveness improves a sales-
Similarly, during a sales interaction, the effective use of
information improves the salesperson’s ability to anticipate
and respond to buyer concerns and objections. For example,
shelf-space management software (such as Apollo or Space-
man) allows a salesperson to recommend immediately a cus-
tom shelf arrangement for a new product. For example, a
sales rep with a higher sales technology orientation should be
able to more immediately respond to a buyer’s reactions dur-
ing consultation—whether conducting that consultation in
person or by remote means (for example, using application-
sharing software). The reason salespeople are more able to
adapt is because they possess better clarity and understanding
of customer information. In essence, more effective informa-
tion improves the salesperson’s capacity to adapt by provid-
ing patterns of meaningful insights from increasingly complex
and more readily available marketplace data.
Hypothesis 6: Information effectiveness will improve the
salesperson’s adaptive behaviors.
Key Aspects of Sales Performance
Boundary role theory (cf. Adams 1976; Organ 1971) has be-
come one of the most dominant theoretical frameworks used
to study interorganizational relationships in the sales litera-
ture over the past three decades. While boundary role theory
was developed when the prevailing view was that relation-
ships between organizations were almost inherently
adversarial—namely, organizations were forced to depend on
external constituencies to both supply inputs and consume
outputs (cf. Kahn et al. 1964; Katz and Kahn 1966)—it still
has practical relevance for conceptualizing key aspects of sales-
person performance. According to boundary role theory, to
protect its members and its unique interests, an organization
requires defined boundaries that are not readily permeable.
In sales organizations, salespeople represent the physical mani-
festation of those boundaries and are called the “linking pins”
between buyers and sellers (Adams 1976). As such, salespeople
have long played a key role in managing relationships and
information flows between selling firms and their customers.
Even though firms have new capabilities, which are often
driven by advances in information technology, the traditional
sales role in managing buyer–seller relationships remains.
Moreover, salespeople retain diverse responsibilities—some
we consider more internally focused on the selling organiza-
tion and others that are more externally focused on the sales
organization’s customers—namely, buying organizations.
Accordingly, we develop and investigate relationships that
drive two key aspects of sales performance—performance with
customers and internal role performance. Performance with
customers is the extent to which the salesperson cultivates re-
lationships with the customer organization. We define per-
formance with customers as developing an understanding of
a customer’s unique problems and concerns—marketing, tech-
nology, operations, or otherwise—and recommending solu-
tions that address those concerns. For example, in a consumer
market characterized by short life cycles, being first to market
with a new item is often strategically important to both ven-
dor and retailer. As such, quickly generating sales of new com-
pany products—before competitors roll out “me too”
initiatives—is an important aspect of successful performance
Internal role performance refers to the salesperson’s contri-
butions on issues that are predominantly internal to the
supplier’s organization. This includes things such as recom-
mending improvements in company operations and proce-
dures, acting as a special resource to cross-functional associates,
knowing the company’s products, and staying abreast of the
company’s production schedules and technological advances.
Sujan, Weitz, and Kumar (1994) show that both planning
and adaptive behaviors positively influence performance. Weitz,
Sujan, and Sujan define the practice of adaptive behaviors as
100 Journal of Personal Selling & Sales Management
“the altering of sales behaviors during a customer interaction
or across customer interactions based on perceived informa-
tion about the nature of the selling situation” (1986, p. 175).
Effective planning involves proper task prioritization, goal set-
ting, strategic thinking, and anticipation of contingencies. By
adapting the selling approach to a buyer’s unique concerns or
goals, salespeople should be more effective in overcoming ob-
jections and building commitment.
Hypothesis 7: Planning for sales interactions will improve
performance with customers.
Hypothesis 8: Practicing adaptive behaviors will improve
performance with customers.
A strong sales technology orientation can also help sales-
people to be more efficient in completing nonselling admin-
istrative tasks. In fact, such efficiency is the explicit purpose
of many sales automation software applications. For example,
time and territory management software can improve a sales-
person’s ability to coordinate administrative chores. Similarly,
effective use of order tracking systems, the ability to access
online plant production schedules, and even effective use of
e-mail can make a salesperson a better resource to associates
(e.g., logistics specialists, brand managers, financial analysts,
and the like).
Hypothesis 9: A stronger sales technology orientation will
improve internal role performance.
It is worth noting that the absence of supporting logic and
theory favors modeling other potential relationships across
constructs represented as being nonsignificant (for example,
experience has no direct effect on information effectiveness,
sales planning, adaptive behaviors, performance with custom-
ers, or internal role performance). Explicit specification of
these zero-effect hypotheses highlights the idea that the nor-
mative relationships proposed in the process model explain
the whole system of direct and indirect effects related to sales
technology and performance. In other words, relationships
not suggested by the model should not be expected.
In selecting our sample, it was desirable to identify a firm
where (1) sales technology implementation was under way,
(2) use of technology and technology skills varied among sales-
people, and (3) salespeople were involved in typical sales tasks
(i.e., both within the selling firm and selling to customer ac-
counts, but not to final consumers). Based on these criteria,
we approached the management of a well-known CPG com-
pany and asked them to allow us to collect data from the
firm’s U.S. sales force and diagnose sales technology effects.
To encourage participation and improve response rates,
the firm’s top sales executive sent each salesperson a prenoti-
fication letter as well as a cover letter with the questionnaire,
which guaranteed confidentiality to each salesperson. To fur-
ther signal anonymity, we sent questionnaires to the sales rep’s
home office addresses and asked them to return completed
questionnaires directly to the researchers’ university address.
Of 85 questionnaires distributed, 79 (93 percent) were re-
turned. We dropped one respondent from the analysis be-
cause of missing data.
The host firm is a multinational Fortune 500 manufac-
turer and distributor of CPG with sales relationships that cut
across the range of CPG business buying channels (e.g., gro-
cery chains, wholesalers, mass merchandisers, and merchan-
dising headquarters). Among other responsibilities, its
salespeople are routinely involved in category management
leadership activities for their accounts. Their use of sales tech-
nology tools includes proprietary software applications that
incorporate algorithms employed by marketing scientists in
the analysis of scanner data with decision inputs categorized
as promotion, pricing, shelving, and distribution. The firm’s
ratio of revenue per salesperson is high, an indication that its
salespeople are skillful.
Measures for Constructs
Prior to specifying a sample frame for the research to test the
model in Figure 1, we developed an initial questionnaire and
refined it based on in-depth interviews with sales executives
from four different industries. Then, we narrowed our focus
to identify a specific CPG firm to cooperate in the research
and provide access to its sales force. The questionnaire relied
on multi-item scales to develop composite measures for the
constructs in the conceptual model (Figure 1). We pretested
the questionnaire on sales managers within the host firm and
refined the directions and wording of items as appropriate.
Tables 1 and 2 present the scale items, response cues, and
relevant statistics for each of the measures.
We developed the measures for sales technology orienta-
tion, information effectiveness, and other constructs in Table
1 specifically for this research. The measures in Table 2 were
adapted from published scales that have been widely used by
For example, the scale items for internal sales role perfor-
mance and external performance with customers are from the
inventory of sales performance items that were originally de-
veloped and evaluated for reliability and validity by Behrman
and Perreault (1982; 1984). It would have been desirable to
also include other reliable quantitative measures of sales per-
formance or profit contribution provided by the company;
however, management would not release incentive compen-
sation data for individual salespeople or propriety sales
Spring 2006 101
Items and Statistics for Scales Developed in This Study: Sales Technology Orientation, Customer Approval of
Sales Technology, Internal Sales Technology Support, and Information Effectiveness
Construct Name Standard Construct Item
Items for Construct Mean Deviation GFI IFI CFI Reliability Reliability
Sales Technology Orientation
4.80 0.99 0.98 1.00 1.00 0.81
I try to link different sales technologies so that they work together well. 0.64
I have always been fascinated by advances in technology. 0.49
Compared to others in sales, I am technology oriented. 0.48
I extensively use information technologies to perform my job. 0.45
My analytical skills explain most of my success as a salesperson. 0.24
Customer Approval of Sales Technology
5.22 1.00 0.98 1.01 1.00 0.78
The buyers that I deal with do not expect me to use technology.
The buyers that I deal with are annoyed by technology.
The buyers that I deal with use information technology and expect me to. 0.42
The buyers that I deal with are much more interested in personal
relationships than data.
My customers tend to view analysis of scanner data as completely
Internal Sales Technology Support
4.31 1.13 0.95 0.92 0.91 0.73
My company adequately equips me with technology tools. 0.53
My company supplies all technologies that I need to perform my job. 0.45
My company adequately trains me on the use of sales technology. 0.43
I need more help with technology than I get.
5.16 1.12 0.92 0.98 0.98 0.85
Information from or about performance differences among products. 0.63
Information from or about your firm’s marketing effectiveness. 0.51
Information from or about consumer buying habits for the brand or
Information from or about product historical profitability. 0.45
Information from or about your firm’s history shipments to the
Information from or about your customer’s distribution costs. 0.41
Information from or about data collected in retail stores. 0.23
The seven-point response cues for each item were strongly disagree (1) to strongly agree (7).
Responses to this item were reverse scored.
Respondents were directed to “please indicate
how effective each of the following types of information are for earning commitment from your buyers,” with seven-point response cues from totally ineffective (1) to extremely effective (7).
102 Journal of Personal Selling & Sales Management
Items and Statistics for Scales Adapted from Previous Studies: Planning, Adaptive Behavior,
Performance with Customers, and Internal Role Performance
Construct Name Standard Construct Item
Items for Construct Mean Deviation GFI IFI CFI Reliability Reliability
5.72 0.76 0.96 0.99 0.99 0.78
I am careful to work on the highest priority tasks first. 0.51
I keep good records about my account(s). 0.47
I set personal goals for each sales call. 0.41
Each week, I make a plan for what I need to do. 0.38
I do not need to develop a strategy for a customer to get the order.
I think about strategies I will fall back on if problems in a sales
interaction arise. 0.20
5.38 0.92 0.94 0.87 0.86 0.69
I treat all of the buyers pretty much the same.
I feel that most buyers can be dealt with in pretty much the same manner. 0.52
I vary my sales style from situation to situation. 0.16
I can easily use a wide variety of selling approaches. 0.12
Performance with Customers
5.61 0.76 0.99 1.00 1.00 0.83
Convincing customers that I understand their unique problems and
Working out solutions to a customer’s questions and objections. 0.58
Quickly generating new sales of new company products. 0.45
Listening attentively to identify and understand the real concerns of
your customers. 0.42
Internal Role Performance 5.07 0.81 1.00 1.03 1.00 0.71
Acting as a special resource to other associates who need your
Recommending on your own initiative how company operations and
procedures can be improved. 0.40
Knowing the benefits and features of your company products. 0.35
Keeping abreast of all your company’s production and technological
The seven-point response cues for each item were strongly disagree (1) to strongly agree (7).
Responses to this item were reverse scored.
Respondents were directed “on each of the
following items, please rate how well you have performed relative to the average salesperson in similar selling situations,” with seven-point response cues from needs improvement (1) to
Spring 2006 103
measures on individual accounts. Beyond that, however, there
were concerns about the appropriateness of available data be-
cause of differences in account/territory potential that were
unrelated to the efforts of the currently assigned rep or were
based on team-selling efforts. In other situations, such mea-
sures could be used (and adjusted, as appropriate, by mea-
sured variables for factors beyond the control of the
salesperson). On the other hand, these are common obstacles
in research on sales performance. To help address these issues,
the Behrman and Perreault (1982) inventory of sales perfor-
mance items has been influential in the sales research (Leigh,
Pullins, and Comer 2001). Several studies have used subsets
of items to represent both relevant aspects and holistic mea-
sures of sales performance (cf. Cravens et al. 1993; Fang, Evans,
and Zou 2005; Fang, Palmatier, and Evans 2004; Oliver and
Anderson 1994; Sujan, Weitz, and Kumar 1994). In this study,
we used our construct conceptualizations and definitions to
help identify items from the inventory that we felt measured
the relevant aspects of performance proposed here. We then
subjected those scales to extensive analysis to assess their con-
vergent and divergent validity. As a result, we use four items
each to measure the two performance constructs.
The sales planning scale draws on items developed by Sujan,
Weitz, and Kumar (1994), and the items for the adaptive sell-
ing measure are from the scale developed by Spiro and Weitz
(1990). For each of these scales, we used a subset of the origi-
nal inventory of items. This was necessary to comply with
constraints imposed by the host firm’s management concern-
ing the time required to complete the questionnaire. How-
ever, the subset of items included in the questionnaire was
selected based on consistency with the original conceptual-
ization and analysis of published correlations between indi-
vidual items and the total scale.
The measure for salesperson experience is simply the
respondent’s report of the number of years in sales positions.
In addition to the items for the scales associated with the
constructs in the model, the questionnaire included a list of
different hardware and software technologies used by sales-
people. The sales rep respondents were directed to indicate
the extent of their reliance on each of these individual tech-
nologies using a rating scale anchored by 1= “not at all” and
7= “very heavily.”
Data Analysis Methods
Consistent with the recommendations of Jöreskog and
Sörbom (2001) for fitting data with our sample’s characteris-
tics, we used SEM with maximum likelihood parameter esti-
mation to assess the psychometric properties of measures,
evaluate the fit of the overall process model, and estimate
parameters for the normative relationships. In many firms,
the total number of sales reps (and thus the sample size for
technology evaluation) is limited. However, as was done here,
an effective work-around that retains the benefits of SEM
with smaller numbers of observations is to first fit confirma-
tory factor models to assess reliability as well as convergent
and discriminant validity. Estimates from the confirmatory
factor analysis (CFA) are then used to scale composite scores
and fix measurement error to estimate the structural model—
including the relationships hypothesized in Figure 1.
There is debate in the SEM literature about the strengths
and limits of different measures of fit (and no universally ac-
cepted norm). However, we relied on the comparative fit in-
dex (CFI) and incremental fit index (IFI)—statistics that are
widely accepted and also robust with small samples (Fan and
Wang 1998). Even though they are sample size dependent,
we also report the customary chi-square statistic, degrees of
freedom (df), and the goodness-of-fit index (GFI).
We supplemented our analysis using bootstrap sampling
procedures to provide more parameter and standard error es-
timates. The general bootstrap sampling approach introduced
by Efron (1979) has been widely used across various settings
and has precedence in the marketing literature (cf. Bone,
Sharma, and Shimp 1989; Crask and Perreault 1977). More-
over, because it has been adopted to and modified for use in
SEM (cf. Bollen and Stine 1993), we will not review it in
detail here. In essence, however, bootstrapping provides a
nonparametric means for establishing standard errors and
parameter estimates, which helps avoid concerns related to
distributional assumptions needed to validate the application
of asymptotically derived estimators (Efron 2000). Particular
to this study, bootstrapping has been useful for obtaining ro-
bust estimations when used with nonnormal data distribu-
tions—which are common in applied research. In this study,
we employ bootstrapping procedures to generate 1,000
samples used to empirically estimate the distributional prop-
erties of the hypothesized effects in the proposed structural
model and to calculate potential bias in standard parameter
estimates and statistical significance tests.
Evaluation of Measures
The construct reliability indices in Tables 1 and 2 are based
on the shared variance between the (observed) items and the
underlying latent construct; see Fornell and Larcker (1981)
for computational details. The reliability estimates for all the
constructs exceed 0.60, providing evidence of internal con-
sistency (Fornell and Larcker 1981) and adequate fit (Bagozzi
and Yi 1988).
To further establish the convergent validity of the eight
self-report constructs, we used the guidelines advocated by
Bollen (1989) and Anderson and Gerbing (1988). The item
reliability estimates in Tables 1 and 2 are equal to the propor-
tion of the variance in the item that is explained by its proposed
104 Journal of Personal Selling & Sales Management
latent construct; in the SEM context, this is simply the square
of the standardized factor coefficient (measurement param-
eter). Conventionally, item reliabilities greater than 0.16 (or
equivalently, parameters over 0.40) indicate that an item is
internally consistent with the other items for a scale; in Tables
1 and 2, 32 of the 39 item reliabilities are over 0.25 (λ > 0.50)
and only one is below 0.16. The lowest item reliability is for
an item from the adaptive behavior scale. However, even for
that item, the factor coefficient is 2.7 times larger than its
standard error (i.e., statistically significant). Anderson and
Gerbing (1988) offer the guideline that there is convergent
validity among items for a construct when their estimated fac-
tor coefficients are greater than two times the associated stan-
dard error. All of these constructs and items exceed those
criteria. To test the discriminant validity of the constructs, we
used chi-square difference tests to compare one-factor models
(e.g., covariance constrained to one) to two-factor models in
a pairwise fashion across the combinations of constructs pro-
posed here. For all comparisons, the two-factor models had
better fit than their one-factor alternatives, suggesting diver-
gence between all pairs of constructs.
Tables 1 and 2 also provide goodness-of-fit statistics for
the confirmatory factor model for each construct and its as-
sociated items. The fit indices in general provide strong sup-
port for constructing scales based on these items. The lowest
GFI is 0.92, which suggests a good fit of the measurement
model across the analyses. Similarly, the IFI and CRI are
strong, except for the adaptive behaviors construct, which
indicates a marginal fit. Based on this analysis, we conclude
that we have nine distinct constructs representing those pro-
posed in the normative model.
When we compare the fit of the structural equations for
our hypothesized model with the fit of an extension of that
model that explicitly specifies a same-source factor (see Bagozzi
1984 for computational details), the fit is not improved (χ
10.0 with 7 df, p = 0.19). Thus, while same-source bias is
always a potential problem with self-reports, this test provides
evidence that it was not a factor in these results.
Model Fit and Parameter Estimates
Table 3 provides product moment correlations among all the
constructs in the model. The fitting criteria for estimation of
the model are based on maximizing the fit of the observed
sample covariances and the implied covariance structure pre-
dicted based on the structural equations. The overall fit sta-
tistics for the structural model indicate an excellent fit (χ
25.4 with 23 df, p = 0.33; GFI = 0.93, IFI = 0.98, CFI =
0.98). Figure 2 summarizes the results of the structural model
estimates. Consistent with the model specification, a single-
headed arrow depicts a hypothesized relationship between
constructs, and the numbers next to an arrow are the stan-
dardized parameter estimate (path coefficient) and the prob-
ability level for the test of the null hypothesis that the
parameter is zero. Double-headed arrows represent the esti-
mated (free) covariances among constructs. Omitted paths
represent zero constraints.
Path coefficients for eight of the relationships specified in
the model have the hypothesized sign (direction) and are sta-
tistically significant (p < 0.05). The path coefficient for the
other relationship, between adaptive behaviors and perfor-
mance with customers, is positive as hypothesized, but the
probability level (p < 0.10) is not significant at the 0.05 level.
The path coefficients in Figure 2 consider the direct effect(s)
of each independent variable. For completeness, Table 4 pro-
vides the standardized total effect for each independent vari-
able on each endogenous variable. The total effect is the sum
of direct and indirect effects (Bollen 1989). For example, the
total effect of customer approval to use sales technology on
information effectiveness is 0.15. This reflects its indirect ef-
fect through sales technology orientation and is equal to the
product of the two related path coefficients (0.27 and 0.59).
Thus, on average, in the host company, a customer approval
rating that is one standard deviation higher stimulates a higher
sales technology orientation and a 0.15 increase in informa-
tion effectiveness. In general, Table 4 provides a means for
relating constructs on the left side of the model in Figure 2 to
those on the right side.
Sales Technology Orientation and Information
The effects of the antecedent constructs—internal technol-
ogy support, customer approval of sales technology, and sales-
person experience—collectively explain 24 percent of the
variation in sales technology orientation. The path coefficient
for customer approval of sales technology (0.28) is similar to
the coefficient for internal technology support (0.25). This
suggests that a salesperson’s technology orientation may be
influenced as much by the rep’s effort to respond to buyers’
approval of sales technology as it is to respond to the firm’s
own investments or encouragement to use technology. How-
ever, the correlation between customer approval and internal
support is negative but marginally insignificant (–0.18, p <
0.10) if one uses a two-tailed test. This suggests that the
company’s focus on technology support is not consistent with
the technology needs expected from its customer base. More-
over, the –0.35 coefficient for experience in the sales role is
larger in comparison and provides evidence of a statistically
significant experience gap with respect to sales technology.
Clearly, evaluations of technology investments that do not
take into consideration individual differences among users
Spring 2006 105
Correlations Among Scales for Constructs in Model
Technology Customer Internal Salesperson Information Adaptive with
Scales Orientation Approval Support Experience Effectiveness Planning Behavior Customers
1. Sales Technology
2. Customer Approval of
Sales Technology 0.09 1.00
3. Internal Support for
Sales Technology 0.20 –0.13 1.00
4. Salesperson Experience –0.33 0.19 –0.19 1.00
Effectiveness 0.40 0.26 0.20 –0.25 1.00
6. Planning 0.38 0.18 0.24 –0.10 0.49 1.00
7. Adaptive Behavior 0.21 0.16 0.09 –0.04 0.41 0.28 1.00
8. Performance with
Customers 0.23 0.03 –0.03 0.02 0.16 0.33 0.28 1.00
9. Internal Role
Performance 0.38 0.12 0.05 0.06 0.23 0.20 0.20 0.54
Note: Correlation coefficients greater than 0.21 result in a statistically significant t-test (probability less than or equal to 0.05) and are shown in boldface.
106 Journal of Personal Selling & Sales Management
Summary of Structural Model Estimates
Notes: Fit statistics suggest a strong overall fit for the hypothesized model: (minimum fit function χ
= 25.4 (p = 0.33, df = 23), normal theory weighted
least squares χ
= 26.01 (p = 0.30), Satorra–Bentler χ
= 23.58 (p = 0.43), χ
corrected for nonnormality = 30.23 (p = 0.14), GFI = 0.94, IFI = 0.98,
CFI = 0.98, RMSEA = 0.04). Overall model fit statistics for this alternative specification (indicated by significant modification indices noted in Table 5)
provide a modest improvement: (minimum fit function χ
= 16.9 (p = 0.72, df = 21), Satorra–Bentler χ
= 16.27 (p = 0.75), GFI = 0.95, IFI = 1.03,
CFI = 1.03, RMSEA = 0.06). Solid lines indicate statistically significant effects supported both parametric (maximum likelihood estimation) and
nonparametric (bootstrapping) estimation, based on probability values for one-tailed significance effects on path coefficients. The squared multiple
correlation appears at the upper right edge of the circle for each endogenous variable.
Standardized Total Effect (Based on Sum of Direct and Indirect Effects) for
Each Independent Variable on Each Dependent Variable
Dependent Variable for Each Path Estimated in Overall Model
Sales Performance Internal
Independent Technology Information Adaptive with Role
Variable Orientation Effectiveness Planning Behaviors Customers Performance
Technology Support 0.25 0.14 0.09 0.07 0.04 0.12
Customer Approval of
Sales Technology 0.28 0.15 0.09 0.08 0.04 0.13
Salesperson Experience –0.35 –0.20 –0.12 –0.10 –0.05 –0.16
Orientation 0.55 0.34 0.29 0.15 0.46
Information Effectiveness 0.61 0.53 0.28
Adaptive Behaviors 0.18
Spring 2006 107
and buying accounts have the potential to misdiagnose why
and when the investments are effective.
Sales technology orientation has a strong and statistically
significant effect (0.55) on how effectively the salesperson uses
information with customers. In turn, information effective-
ness is positively related to the two working smart constructs.
Information effectiveness explained 37 percent of the vari-
ance in planning and 29 percent of the variance in adaptive
behaviors. Thus, information effectiveness did increase plan-
ning and the practice of adaptive behaviors in this sales force.
In combination, adaptive behaviors and planning explained
16 percent of the variance in performance with customers.
The effect of planning is greater (a path coefficient of 0.31
versus a coefficient of 0.17 for adaptive behavior) and statis-
tically significant. Sales technology orientation accounts for
21 percent of the variance in internal role performance.
From Table 4, note that the total effect of sales technology
orientation on performance with customers is 0.15. Thus,
the total effect of sales technology orientation on internal role
performance (0.46) is greater than on performance with cus-
tomers (0.15). So, while there are sales technology orienta-
tion returns on both performance outcomes, this suggests that
the selling organization is realizing greater returns on its in-
vestments in sales technology from internal role performance
outcomes, or efficiency gains in contrast to external effective-
Taken as a whole, the results support the hypothesized in-
direct effects of sales technology orientation on performance
with customers through information effectiveness and its sub-
sequent effects through planning (paths H5 and H7) and
adaptive behaviors (paths H6 and H8). Also, the results are
consistent with the focus of past research on smart selling
(H8 and H9), grounding this study within the theoretical
context of previous sales management literature.
Test of Zero Constraints Across Model Constructs
and Bootstrap Sampling Results
The model tests single-, double-, and triple-mediated con-
straints among constructs in the model. Analysis of those con-
straints, given an overall excellent fit for the model, provides
added insights into the mechanisms through which sales tech-
nology orientation affects key aspects of performance. For
example, a company can invest heavily in internal technology
support but will not realize significant direct gains in internal
role performance except through sales technology orientation
(a single-mediated process). Furthermore, such investments
would yield no return on performance with customers except
through sales technology orientation that increases informa-
tion effectiveness and planning or adaptive behaviors (a triple-
mediated process). Thus, using this process, it becomes much
clearer that managers must consider the tasks and processes
through which sales technology tools can affect performance
outcomes. If this is not done, the opportunity cost may be
In addition to providing alternative model specification
fit statistics, Table 5 summarizes the comparisons between
our bootstrapping sampling results and the parametric pro-
cedures used to obtain estimates.
Bollen (1989, pp. 267–268) notes that no hard fast rule
for sample size exists, but it is desirable to have at least several
cases per free parameter estimated. Similarly, Bentler and Chou
(1988, p. 172) refer to their own widely used 5:1 ratio for
sample size to number of free parameters estimated as an “over-
simplified guideline” and not an absolute. Nonetheless, many
SEM researchers use the 5:1 ratio of sample size to parameter
estimates as an “absolute” guideline, but there really is no ab-
solute minimum sample size or even an absolute minimum
We fit the proposed model to 1,000 bootstrap samples and
used the estimates from those fittings to develop robust stan-
dard error estimates. Bootstrap estimates are provided with
the estimated percentage of bias in the maximum likelihood
parameter estimates and standard errors. For all hypotheses
but one, the results of the bootstrap provide incremental evi-
dence that supports the results of the maximum likelihood
estimates (e.g., seven of the nine hypotheses are supported,
one is rejected, and one is only “marginally” supported). By
dividing the lowest magnitude parameter estimate (either
maximum likelihood or bootstrap) by the highest magnitude
standard error (which is always the bootstrap standard error),
we form a statistic that can be compared to t-distribution
tables (α = 0.05) for statistical significance testing. The boot-
strap suggests the relationship between the firm’s support for
information technology and sales technology orientation is
weak (only marginally supported by the data).
In SEM, each endogenous variable has a structural error
associated with its prediction by other variables in the model.
The alternative model modifications represent inclusion of
statistically significant (nonzero) path estimates that are hy-
pothesized in the proposed model to be nonsignificant.
Sörbom (1989) proposed the use and interpretation of modi-
fication indices obtained when fitting the proposed structural
model as a means for developing and testing alternative model
specifications. Only two of the 45 zero-constrained relation-
ships (18 in the β matrix of η to η effects, 15 in the γ matrix
of ξ to η effects, and 12 in the structural error term covari-
ance matrix, ψ) had significant modification indices, suggest-
ing a better model fit could be obtained by freeing these
elements. Specifically, modification indices suggested two se-
quenced changes to the proposed structural model: (1) freeing
108 Journal of Personal Selling & Sales Management
Summary of Structural Model and Bootstrap Results
Full Information Bootstrap
Maximum Bootstrap Estimated
Likelihood Full Information Mean Parameter
Unstandardized Maximum Unstandardized and Direct
Parameter Likelihood Parameter (standard Effect
Estimate Standardized Estimate error) Hypothesis
Dependent Variable Hypothesized (standard Parameter (standard Bias Supported
Independent Variable Effect
error) t-Value Estimate p-Value error) (in percent) (
α = 0.05)?
Performance with Customers
Planning + 0.30 0.29 3.3 Yes
(0.13) 2.36 0.31 < 0.05 (0.15) (–7.1)
Adaptive Behaviors + 0.14 0.17 –21.4 No
(0.14) 1.01 0.19 < 0.10 (0.15) (–1.6)
Internal Role Performance
Sales Technology Orientation + 0.37 0.38 2.7 Yes
(0.09) 3.93 0.59 < 0.001 (0.11) (22.2)
Information Effectiveness + 0.42 0.42 0 Yes
(0.10) 4.22 0.61 < 0.001 (0.11) (–10.0)
Information Effectiveness + 0.44 0.45 –2.3 Yes
(0.11) 4.04 0.48 < 0.001 (0.12) (–9.1)
Sales Technology Orientation + 0.62 0.63 1.6 Yes
(0.14) 4.51 0.59 < 0.001 (0.15) (–7.1)
Sales Technology Orientation
Internal Technology Support + 0.22 0.22 0 Marginal
(0.12) 1.85 0.27 < 0.05 (0.14) (–16.7)
Customer Approval + 0.27 0.27 0 Yes
(0.13) 2.07 0.27 < 0.01 (0.15) (–15.4)
Salesperson Experience – –0.25 –0.24 4 Yes
(0.08) –2.98 –0.35 < 0.001 (0.09) (–12.5)
Spring 2006 109
Alternate Model Specifications
Index Hypothesis Supported?
Structural Error Term Covariances
Sales Technology Orientation and Information Effectiveness 26.9 No
Information Effectiveness and Sales Planning 15.1 No
Other 43 Zero-Constrained Hypotheses in β, γ, and ψ Matrices < 5.0 Yes
Notes: Fit statistics suggest a strong overall fit for the hypothesized model: (minimum fit function χ
= 25.4 (p = 0.33, df = 23), normal theory weighted least squares χ
= 26.01 (p = 0.30),
= 23.58 (p = 0.43), χ
corrected for nonnormality = 30.23 (p = 0.14), GFI = 0.94, IFI = 0.98, CFI = 0.98, RMSEA = 0.04).
The hypothesized effects summarized here
include only the direct effects conceptualized herein and shown in Figure 1. The alpha level chosen for significance testing was 0.05 and, in the most conservative test, we would choose the
parameter estimate with the lowest magnitude and divide that by the bootstrap standard error estimate to construct t-values for testing. The hypothesized structural model constrains to zero all
other potential effects among variables whose path coefficient is not freely estimated. See Table 4 for a summary of total effects (direct and indirect).
The overall model fit and modifications
indices (Sörbom 1989) were used to test all nonzero relationships, including, for example, the double- and triple-mediated processes through which the three exogenous variables (internal
technology support, customer approval, and in-role expertise) influence the two aspects of performance (performance with customers and internal role performance). All hypothesized zero-
constrained paths (constrained feedback loops) in the original model were supported (i.e., none have effects that statistically differ significantly from zero). Exogenous variables were allowed to
covary, and the three covariance estimates obtained from the analysis were –0.18 between internal support and customer approval, –0.22 between internal support and salesperson experience,
and 0.21 between customer approval and salesperson experience.
The alternative model modifications represent inclusion of statistically significant (nonzero) path estimates that are hypoth-
esized in the proposed model to be nonsignificant. Sörbom (1989) proposed the use and interpretation of modification indices obtained when fitting the proposed structural model as a means
for developing and testing alternative model specifications. Two of the 45 zero-constrained relationships (18 in the β matrix of η to η effects, 15 in the γ matrix of ξ to η effects, and 12 in the ψ
matrix of structural error term covariances) had significant modification indices, suggesting a better model fit could be obtained by freeing these elements. The overall model fit statistics for this
alternative specification provide a modest improvement: minimum fit function χ
= 16.9 (p = 0.72, df = 21), Satorra–Bentler χ
= 16.27 (p = 0.75), GFI = 0.95, IFI = 1.03, CFI = 1.03,
RMSEA = 0.06).
110 Journal of Personal Selling & Sales Management
the structural error covariances for the error terms associated
with modeling sales technology orientation and information
effectiveness (modification index = 25.9) and (2) freeing the
structural error terms associated with the information effec-
tiveness and sales planning constructs (modification index =
15.1). The overall model fit statistics for this alternative speci-
fication provide a modest improvement: minimum fit func-
= 16.9 (p = 0.72, df = 21), Satorra–Bentler χ
(p = 0.75), GFI = 0.95, IFI = 1.03, CFI = 1.03, root mean
square error of approximation [RMSEA] = 0.06.
DISCUSSION AND CONCLUSIONS
The results of the study presented here are generally consis-
tent with the parsimonious model proposed and, at the same
time, provide diagnostic insights about the role of sales tech-
nology. The structural model reveals that the data are consis-
tent with this conceptualization. The model explains 16
percent of the variance in performance with customers and
21 percent of the variation in internal role performance. There
is, of course, unexplained variance from other factors, such as
effort, but these results compare favorably with past research
in sales performance. For example, in their meta-analysis of
75 years of sales performance research, Churchill et al. (1985)
note that, on average, no single variable accounts for more
than 10 percent of the variation in salesperson performance.
Yet the two working smart predictors explain 16 percent of
the variance in performance, while customer and sales tech-
nology orientation alone explains 21 percent of the variation
in internal role performance.
From a statistical theory perspective, this study poses two
primary limitations—within-firm design and sample size. The
use of a within-firm design warrants caution concerning the
applicability of our results across a broad spectrum of firms—
although generalizations to other firms within the CPG in-
dustry are less remote. On the other hand, these limitations
provide an opportunity to demonstrate an appropriate appli-
cation of the SEM process recommended here within a small
sample context. Because many sales organizations have less
than 100 employees, this is a reality that many sales managers
would face when carrying out this approach. However, SEM
methods are used in academic research for similarly sized
samples. For example, a recent content analysis of over 500
publications of SEM studies in psychology journals points
out that about one-fifth had sample sizes smaller than 100
(MacCallum and Austin 2000). Thus, our application in this
sample size is not unique (although our work goes well be-
yond the norm in providing extensive bootstrapping results
across various estimators to demonstrate the stability of the
parameter estimates). This added step adds more credibility
and confidence toward the internal validity of our study—
and would do the same for others implementing this approach.
It is important to note that one of our objectives was not
only to propose the means for managers to diagnose sales
technology implementations but also to do so in a manner
that provides a tractable within-firm diagnostic tool. To be a
tractable within-firm diagnostic tool, the data needed for mod-
eling the relationships needed to be accessible to the sales man-
agers—and a survey of their own salespeople meets that
criterion. In sum, the methods used in this study are tractable
to within-firm investigations, but they are also analytically
sophisticated, and care should be taken when interpreting these
results—particularly with regard to inferences concerning
cause and effects.
At the same time, the results in Table 4 highlight the fact
that the total effect of sales technology orientation on inter-
nal role performance here is 0.46, as contrasted with a total
effect of 0.15 for performance with customers. From a diag-
nostic standpoint, this implies that in this company, sales reps
with an inclination and the analytical skills for applying in-
formation technologies to their sales tasks are having more
effect on internal processes and operations than they are in
resolving customer problems and concerns. In the modern
era of relationship marketing, that is a finding worthy of
management’s consideration. Specifically, from a normative
standpoint, there is an opportunity for our host firm to put
more emphasis on technologies that improve performance
There is not evidence from this study on which to argue
that this pattern would apply across other CPG supply situa-
tions (i.e., with other suppliers), but we speculate that this is
the case. Case studies in the popular press tend to emphasize
sales automation applications, the focus of which tends to be
on cutting sales force costs or making more efficient the flow
of information needed by the supplier company. Further, these
applications focus on existing tasks rather than on enabling
tasks that previously were not performed (or performed well).
On the other hand, the ability and effort required of a sales
rep in applying information technology to come up with in-
tegrative, win-win solutions for both the company and the
retailer are less structured and tend to require more adaptive,
custom efforts. Yet it is this type of application where sales
technology may have a greater impact on the revenue-gener-
ating side of category management efforts.
The process modeling approach presented here relies on
estimated effects rather than on direct user evaluation or judg-
ments of which sales technologies are most useful in com-
pleting different sales tasks. It is not our intent to suggest that
user evaluations of information technologies are not poten-
tially useful but, rather, that estimation of the impact of tech-
nologies via a process model provides complementary
information. To elaborate on this difference, we asked respon-
dents to rate their reliance on various sales technologies, cal-
culated the mean and standard deviation of the respondents’
Spring 2006 111
ratings, and correlated the reliance ratings with the constructs
in our proposed model. In lieu of providing all descriptive
data here, to illustrate our point, we isolate one of those tech-
nologies—shelf-space management software. From a diagnos-
tic standpoint, this was one of the least frequently used
applications, yet it also had the highest bivariate correlation
with both internal role performance and performance with
customers. Although reliance on this sales technology shows
a relationship to process and performance outcomes, it is not
correlated with internal support. This finding suggests that
the process modeling approach may add incremental infor-
mation beyond user evaluations that could improve manage-
ment decision making, although further research is needed.
The behavioral process modeling approach used in this
study offers a flexible method for diagnostic modeling of the
technology-to-performance relationship and how it works
through behavioral mechanisms. To apply the approach to
diagnose sales technology implementations, managers can
develop normative models of their sales technology processes,
beginning with desirable outcomes, identifying behavioral
mechanisms that affect those outcomes, and then mapping
out antecedent influences of those outcomes. This approach
can be used across a wide variety of settings through which
technology is thought to affect behavioral processes. While
our model specification is parsimonious and focuses on in-
formation effectiveness and smart selling tasks, in other con-
texts, the model specification could be expanded to include
other relevant antecedents, technologies, process tasks, and
Adams, J. Stacey (1976), “The Structure and Dynamics of Be-
havior in Organizational Boundary Roles,” in Handbook of
Industrial and Organizational Psychology, Marvin D.
Dunnette, ed., Chicago: Rand McNally, 1175–1199.
Ahearne, Michael J., Ronald Jelinek, and Adam Rapp (2005),
“Moving Beyond the Direct Effect of SFA Adoption on
Salesperson Performance: Training and Support as Key
Moderating Factors,” Industrial Marketing Management, 34
———, Narasimhan Srinivasan, and Luke Weinstein (2004),
“Effect of Technology on Sales Performance: Progressing
from Technology to Technology Usage and Consequences,”
Journal of Personal Selling & Sales Management, 24, 4 (Fall),
Anderson, James C., and David W. Gerbing (1988), “Structural
Equation Modeling in Practice: A Review and Recom-
mended Two-Step Approach,” Psychological Bulletin, 103
Anderson, Rolph E. (1996), “Personal Selling and Sales Man-
agement in the New Millennium,” Journal of Personal Sell-
ing & Sales Management, 76 (November–December), 5–15.
Bagozzi, Richard P. (1984), “A Prospectus for Theory Construc-
tion in Marketing,” Journal of Marketing, 48, 1 (Winter),
———, and Youjae Yi (1988), “On the Evaluation of Structural
Equation Models,” Journal of the Academy of Marketing Sci-
ence, 16 (Spring), 74–94.
Behrman, Douglas N., and William D. Perreault, Jr. (1982),
“Measuring the Performance of Industrial Salespersons,”
Journal of Business Research, 10 (September), 355–369.
———, and ——— (1984), “A Role Stress Model of the Per-
formance and Satisfaction of Industrial Salespersons,” Jour-
nal of Marketing, 48 (Fall), 9–21.
Bentler, Peter M., and Chih-Ping Chou (1988), “Practical Issues
in Structural Modeling,” in Common Problems/Proper Solu-
tions: Avoiding Error in Quantitative Research, J. Scott Long,
ed., Newbury Park, CA: Sage, 161–192.
Bharadwaj, Anandhi S., Sundar G. Bharadwaj, and Benn R.
Konsynski (1999), “Information Technology Effects on Firm
Performance as Measured by Tobin’s q,” Management Sci-
ence, 45 (July), 1008–1024.
Bollen, Kenneth A. (1989), Structural Equation Modeling with
Latent Variables, New York: John Wiley & Sons.
———, and Robert A. Stine (1993), “Bootstrapping Goodness-
of-Fit Measures in Structural Equation Models,” in Testing
Structural Equation Models, Kenneth A. Bollen and J. Scott
Long, eds., Newbury Park, CA: Sage, 111–135.
Bone, Paula F., Subhash Sharma, and Terrence A. Shimp (1989),
“A Bootstrap Procedure for Evaluating Goodness-of-Fit In-
dices,” Journal of Marketing Research, 26, 1 (February),
Brown, Steven P., and Robert A. Peterson (1993), “Antecedents
and Consequences of Salesperson Job Satisfaction: Meta-
Analysis and Assessment of Causal Effects,” Journal of Mar-
keting Research, 30 (February), 63–77.
Cannon, Joseph P., and William D. Perreault, Jr. (1999), “Buyer–
Seller Relationships in Business Markets,” Journal of Mar-
keting Research, 36 (November), 439–460.
Churchill, Gilbert A., Neil M. Ford, Steven W. Hartley, and
Orville C. Walker, Jr. (1985), “The Determinants of Sales-
person Performance,” Journal of Marketing Research, 22
Collins, Robert H. (1984), “Artificial Intelligence in Personal
Selling,” Journal of Personal Selling & Sales Management, 4,
1 (May), 58–66.
——— (1985), “Enhancing Spreadsheets for Increased Produc-
tivity,” Journal of Personal Selling & Sales Management, 5, 2
——— (1989), “Unleash the Power of Desktop Presentations,”
Journal of Personal Selling & Sales Management, 9, 1 (Spring),
Comer, James M. (1981–82), “Invited Essay: Sales Management
and the Computer: Prospects for the 1980’s,” Journal of
Personal Selling & Sales Management, 2, 1 (Fall–Winter),
Crask, Melvin R., and William D. Perreault, Jr. (1977), “Valida-
tion of Discriminant Analysis in Marketing Research,” Jour-
nal of Marketing Research, 14 (February), 60–68.
112 Journal of Personal Selling & Sales Management
Cravens, David W., Thomas N. Ingram, Raymond W. LaForge,
and Clifford E. Young (1993), “Behavior-Based and Out-
come-Based Salesforce Control Systems,” Journal of Mar-
keting, 57 (October), 47–59.
Davis, Fred D. (1989), “Perceived Usefulness, Perceived Ease of
Use, and User Acceptance of Information Technology,” MIS
Quarterly, 13 (September), 319–340.
Efron, Bradley (1979), “Bootstrap Methods: Another Look at
the Jackknife,” Annals of Statistics, 7 (1), 1–26.
——— (2000), “The Bootstrap and Modern Statistics,” Journal
of the American Statistical Association, 95, 452 (December),
Erffmeyer, Robert C., and Dale A. Johnson (2001), “An Explor-
atory Study of Sales Force Automation Practices: Expecta-
tions and Realities,” Journal of Personal Selling & Sales
Management, 21, 2 (Spring), 167–175.
Evans, Kenneth R., and John L. Schlacter (1985), “The Role of
Sales Managers and Salespeople in a Marketing Informa-
tion System,” Journal of Personal Selling & Sales Manage-
ment, 5, 2 (November), 49–58.
Fan, Xitao, and Lin Wang (1998), “Effects of Potential Con-
founding Factors on Fit Indices and Parameter Estimates
for True and Misspecified SEM Models,” Educational and
Psychological Measurement, 58 (October), 701–735.
Fang, Eric, Kenneth R. Evans, and Shaoming Zou (2005), “The
Moderating Effect of Goal-Setting Characteristics on the
Sales Control Systems–Job Performance Relationship,” Jour-
nal of Business Research, 58 (September), 1214–1222.
———, Robert W. Palmatier, and Kenneth R. Evans (2004),
“Goal-Setting Paradoxes? Trade-Offs Between Working
Hard and Working Smart: The United States Versus China,”
Journal of the Academy of Marketing Science, 32 (2), 188–202.
Fornell, Claes, and David F. Larcker (1981), “Evaluating Struc-
tural Equation Models with Unobservable Variables and
Measurement Error,” Journal of Marketing Research, 18 (Feb-
Johnston, Mark W., and Greg W. Marshall (2005), Relationship
Selling and Sales Management, Boston: McGraw-Hill Irwin.
Jones, Eli, Suresh Sundaram, and Wynne Chin (2002), “Factors
Leading to Sales Force Automation Use: A Longitudinal
Analysis,” Journal of Personal Selling & Sales Management,
22, 3 (Summer), 145–156.
Jöreskog, Karl, and Dag Sörbom (2001), LISREL 8: User’s Refer-
ence Guide, Chicago: Scientific Software International.
Kahn, Barbara E., and Leigh McAlister (1997), Grocery Revolu-
tion: The New Focus on the Consumer, Reading, MA:
Kahn, Robert L., Donald M. Wolfe, Robert P. Quinn, J. Diedrick
Snoek, and Robert A. Rosenthal (1964), Organizational
Stress: Studies in Role Conflict and Role Ambiguity, New York:
John Wiley & Sons.
Katz, Daniel, and Robert L. Kahn (1966), The Social Psychology
of Organizations, New York: John Wiley & Sons.
Klompmaker, Jay E. (1980–81), “Incorporating Information from
Salespeople into the Marketing Planning Process,” Journal
of Personal Selling & Sales Management, 1, 1 (Fall–Winter),
Ko, Dong-Gil, and Alan R. Dennis (2004), “Sales Force Auto-
mation and Sales Performance: Does Experience and Ex-
pertise Matter?” Journal of Personal Selling & Sales
Management, 24, 4 (Fall), 311–322.
Leigh, Thomas W., Ellen B. Pullins, and Lucette B. Comer
(2001), “The Top Ten Sales Articles of the 20th Century,”
Journal of Personal Selling & Sales Management, 21, 3 (Sum-
Liu, Annie H., and Mark P. Leach (2001), “Developing Loyal
Customers with a Value-Adding Sales Force: Examining
Customer Satisfaction and the Perceived Credibility of
Consultative Salespeople,” Journal of Personal Selling & Sales
Management, 21, 2 (Spring), 147–156.
Lucas, Henry C., Jr., and V.K. Spitler (1999), “Technology Use
and Performance: A Field Study of Broker Workstations,”
Decision Sciences, 30, 2 (Spring), 291–311.
MacCallum, Robert C., and James T. Austin (2000), “Applica-
tion of Structural Equation Modeling in Psychological Re-
search,” Annual Review of Psychology, 51, 201–236.
Menon, Anil, and P. Rajan Varadarajan (1992), “A Model of
Marketing Knowledge Use Within Firms,” Journal of Mar-
keting, 56 (October), 53–71.
Morris, Michael G., and Viswanath Venkatesh (2000), “Age Dif-
ferences in Technology Adoption Decisions: Implications
for a Changing Work Force,” Personnel Psychology, 53 (Sum-
Oliver, Richard L. (1974), “Expectancy Theory Predictions of
Salesmen’s Performance,” Journal of Marketing Research, 11
———, and Erin Anderson (1994), “An Empirical Test of the
Consequences of Behavior- and Outcome-Based Sales Con-
trol Systems,” Journal of Marketing, 58 (October), 53–68.
Organ, Dennis W. (1971), “Linking Pins Between Organiza-
tions and Environments,” Business Horizons, 14 (6), 73–80.
Parthasarathy, Madhavan, and Ravipreet S. Sohi (1997), “Sales-
force Automation and the Adoption of Technological In-
novations by Salespeople: Theory and Implications,” Journal
of Business and Industrial Marketing, 12 (3–4), 196–208.
Pass, Michael W., Kenneth R. Evans, and John L. Schlacter
(2004), “Sales Force Involvement in CRM Information
Systems: Participation, Support, and Focus,” Journal of Per-
sonal Selling & Sales Management, 24, 3 (Summer),
Plouffe, Christopher R., Brian C. Williams, and Thomas W. Leigh
(2004), “Who’s on First? Stakeholder Differences in Cus-
tomer Relationship Management and the Elusive Notion
of ‘Shared Understanding,’” Journal of Personal Selling &
Sales Management, 24, 4 (Fall), 323–338.
Pullig, Chris, James G. Maxham, III, and Joseph F. Hair, Jr.
(2002), “Salesforce Automation Systems: An Exploratory
Examination of Organizational Factors Associated with Ef-
fective Implementation and Salesforce Productivity,” Jour-
nal of Business Research, 55 (May), 401–415.
Schillewaert, Niels, Michael J. Ahearne, Ruud T. Frambach, and
Rudy K. Moenaert (2005), “The Adoption of Information
Technology in the Sales Force,” Industrial Marketing Man-
agement, 34 (May), 323–326.
Spring 2006 113
Shoemaker, Mary E. (2001), “A Framework for Examining IT-
Enabled Market Relationships,” Journal of Personal Selling
& Sales Management, 21, 2 (Spring), 177–185.
Sörbom, Dag (1989), “Model Modification,” Psychometrika, 54
Speier, Cheri, and Viswanath Venkatesh (2002), “The Hidden
Minefields in the Adoption of Sales Force Automation Tech-
nologies,” Journal of Marketing, 66 (July), 98–111.
Spiro, Rosann L., and Barton A. Weitz (1990), “Adaptive Selling:
Conceptualization, Measurement, and Nomological Valid-
ity,” Journal of Marketing Research, 27, 1 (February), 61–69.
Sujan, Harish, Barton A. Weitz, and Nirmalya Kumar (1994),
“Learning Orientation, Working Smart, and Effective Sell-
ing,” Journal of Marketing, 58 (July), 39–52.
Swenson, Michael J., and Adilson Parrella (1992), “Cellular Tele-
phones and the National Sales Force,” Journal of Personal
Selling & Sales Management, 12, 4 (Fall), 67–74.
Tanner, John F., Jr., and Shannon Shipp (2005), “Sales Technology
Within the Salesperson’s Relationships: A Research Agenda,”
Industrial Marketing Management, 34 (May), 305–312.
Teas, R. Kenneth (1981), “An Empirical Test of Models of Sales-
persons’ Job Expectancy and Instrumentality Perceptions,”
Journal of Marketing Research, 18 (May), 209–227.
Thibaut, John W., and Harold H. Kelley (1959), The Social Psy-
chology of Groups, New York: John Wiley & Sons.
Venkatesh, Viswanath, and Fred D. Davis (2000), “A Theoreti-
cal Extension of the Technology Acceptance Model: Four
Longitudinal Field Studies,” Management Science, 46 (Feb-
Vroom, Victor H. (1964), Work and Motivation, New York: John
Wiley & Sons.
Walker, Orville C., Jr., Gilbert A. Churchill, Jr., and Neil M.
Ford (1977), “Motivation and Performance in Industrial
Selling: Present Knowledge and Needed Research,” Journal
of Marketing Research, 14 (May), 156–168.
Weitz, Barton A., Harish Sujan, and Mita Sujan (1986), “Knowl-
edge, Motivation, and Adaptive Behavior: A Framework
for Improving Selling Effectiveness,” Journal of Marketing,
50 (October), 174–191.
Widmier, Scott M., Donald W. Jackson, Jr., and Deborah Brown
McCabe (2002), “Infusing Technology into Personal Sell-
ing,” Journal of Personal Selling & Sales Management, 22, 3
Zablah, Alex R., Danny N. Bellenger, and Wesley J. Johnston
(2004), “Customer Relationship Management Implemen-
tation Gaps,” Journal of Personal Selling & Sales Manage-
ment, 24, 4 (Fall), 279–295.
Zoltners, Andris A., Prabhakant Sinha, and Greggor A. Zoltners
(2001), The Complete Guide to Accelerating Sales Force Per-
formance, New York: AMACOM.