PETER KAUFMAN, SATISH JAYACHANDRAN, and RANDALL L.

ROSE*

Retailers face considerable risk in introducing new products because

of high failure rates. Given the proliferation of new products juxtaposed
against finite shelf space, retail buyers are confronted with a choice
problem. To enhance understanding of this issue, the authors examine
the role of buyer–salesperson and firm–firm relationships using data
collected in the context of actual new product selection by retail buyers at
two large grocery retailers in the United States. The findings indicate that
buyer–salesperson and firm–firm relationships have a greater influence
on new product acceptance when a new product’s attractiveness is
modest than when the new product is very unattractive or very attractive.
At modest levels of product attractiveness, the likelihood of new product
acceptance can increase by as much as 60% when the buyer has a
strong relationship with the salesperson. The study provides insights into
the complex interplay of marketing relationships and product
attractiveness in retail buyers’ selection of new products.

The Role of Relational Embeddedness in

Retail Buyers’ Selection of New Products

New products are considered critical for the long-term
success of firms. Many new products are sold to consumers
through a retail channel, and their success is contingent on
retailer acceptance and support. However, a study by the
Federal Trade Commission (2003) found that retailers face
considerable risk and costs in introducing new products
because of their high failure rate. This is a particularly vex-
ing problem for retailers because of the scarcity of shelf
space (Bloom, Gundlach, and Cannon 2000). Therefore, the
issue of new product acceptance at the retail level for the
purpose of display and sales is of great managerial and aca-
demic interest.
Researchers have investigated how retail buyers evaluate
new consumer products by focusing on product, marketing
*Peter Kaufman is an assistant professor, Department of Marketing, Illi-
nois State University (e-mail: pkaufma@ilstu.edu). Satish Jayachandran is
an associate professor (e-mail: satish@moore.sc.edu), and Randall L. Rose
is a professor (e-mail: roser@moore.sc.edu), Department of Marketing,
University of South Carolina. The authors are indebted to a coauthor, a
national sales manager for a consumer packaged goods firm, who must
remain anonymous. They also thank two major retailers for their extensive
cooperation and executives at Nabisco, Kraft, Ocean Spray, Kimberly-
Clark, and Hamilton Beach for their invaluable assistance. The authors are
grateful to the David F. Miller Center for Retailing Education and
Research, University of Florida, for providing financial support for the
study. They also thank Bill Bearden, Eli Jones, Frederick Langrehr, Bruce
Money, Subhash Sharma, Janet Wagner, Bart Weitz, Chris White, Fran-
cisco Vera, the anonymous reviewers at the Marketing Science Institute’s
Doctoral Dissertation Proposal Competition, and attendees at the Market-
ing Science Institute Conference on Academic–Practitioner Collaborative
Research hosted at Yale for their comments, suggestions, and support.
strategy, category, and environmental factors (e.g., Mont-
gomery 1975; Rao and McLaughlin 1989). However,
despite the demonstrated importance of marketing relation-
ships (Morgan and Hunt 1994), their impact on retailers’
selection of new products has not been well studied. Con-
sidering the role of relationships in new product acceptance
is critical because, as Srivastava, Shervani, and Fahey
(1998, 1999) state, these relationships are market-based
assets with the potential to create long-term value to share-
holders. Thus, the objective of this study is to examine the
role of firm–firm and buyer–salesperson relationships in
retailers’ selection of new products. We study this issue in
the context of the grocery business using data on retail buy-
ers’ selection of actual new products. Our results show that
marketing relationships have the greatest impact on new
product acceptance when these products are moderately
attractive.
Our results demonstrate the role of firm–firm and buyer–
salesperson relationships in retailers’ acceptance of new
products. Data on retail buyers’ selection of actual new
products give the findings a high degree of external validity.
This study offers insights that will help manufacturers
determine whether allocating resources to develop channel
relationships enables new product acceptance. Retailers can
benefit from this research by gaining a better understanding
of how their buyers select new products. The study also
contributes to the new product literature by highlighting the
importance of the distribution channel in new product suc-
cess. Finally, the study provides evidence of how marketing

© 2006, American Marketing Association Journal of Marketing Research

ISSN: 0022-2437 (print), 1547-7193 (electronic) 580 Vol. XLIII (November 2006), 580–587
Relational Embeddedness and Selection of New Products
relationships work as market-based assets in enhancing
competitive advantage.
RELATIONSHIPS AND NEW PRODUCT SELECTION
Retail shelf space is a scarce resource, and its effective
deployment could determine the success of a retailer. Thus,
the selection of products that maximize the returns from
shelf space is a valuable capability for a retailer. The suc-
cess of a new product often depends on factors that go
beyond mere product features. Manufacturer marketing
support (e.g., advertising, sales promotion, display) is also a
critical determinant of new product success. In this regard,
the literature suggests that product attractiveness, a function
of new product features and other market and strategy char-
acteristics that differentiate the new product from existing
products, influences new product selection (Rao and
McLaughlin 1989). Buyers in retail organizations make an
assessment of product attractiveness when deciding to
accept or reject new products offered by manufacturers.
However, these decisions are shrouded in uncertainty and
are challenging even for experienced buyers (Kline and
Wagner 1994). Transaction cost analysis (Williamson 1985)
explains how uncertainty might even prevent transactions
from taking place or, in our context, might lead to the rejec-
tion of new products.
Institutional theory suggests that economic actions, such
as retail buyers’ evaluation and acceptance of new products,
are embedded in relationships (Granovetter 1985). The con-
cept of structural embeddedness implies that “economic
activity does not occur in a social vacuum, but rather is
nested in patterns of economic and/or social relationships”
(Dacin, Ventresca, and Beal 1999, p. 326). Embeddedness is
often viewed as a constraint that reduces economic effi-
ciency by introducing extraneous factors into market activ-
ity. However, modern viewpoints on embeddedness imply
that these relationships, at both the firm and the individual
levels, often enhance opportunities for transactions to fruc-
tify. In effect, relational embeddedness might enhance the
likelihood of retail buyers’ acceptance of new products.
Therefore, as Srivastava, Shervani, and Fahey (1998) sug-
gest, embeddedness creates the market-based asset of rela-
tionships that provides both the selling and the buying
organizations with the potential to create value by enabling
transactions. To elucidate this, we focus on the relationships
between the buyer and the seller firms and between the
salesperson and the individual buyer.
Relationship Quality
Although research into embeddedness from an institu-
tional perspective often focuses merely on the presence or
absence of relationships (Dacin, Ventresca, and Beal 1999),
the content of these ties, rather than their mere existence, is
likely to be the factor that influences economic transactions,
such as new product acceptance. The relationship marketing
literature refers to the content of the relationships that affect
economic transactions as “relationship quality.” Relation-
ship quality is defined as an overall evaluation of the
strength of a relationship and is composed of trust, satisfac-
tion, and commitment (De Wulf, Odekerken-Schröder, and
Iacobucci 2001).
Trust. Trust is present when a party has confidence in an
exchange partner’s reliability and integrity. Credibility and
benevolence are two aspects of trust that have been recog-
581
nized extensively in previous research (Doney and Cannon
1997). By developing trust, manufacturers and retailers can
take advantage of their complementary skills to reduce
transaction costs (Kumar 1996). As a key aspect of the rela-
tional embeddedness of firm–firm transactions, trust facili-
tates transactions by reducing perceived risk (Morgan and
Hunt 1994).
Satisfaction. Satisfaction is the affective state that results
from an appraisal of the relationship (Anderson and Narus
1990). A relationship in which the buyer is satisfied is likely
to reduce the risk that the buyer perceives when adopting a
new product. Thus, satisfaction with the relationship is
likely to enhance the probability of future transactions
(Heide and Miner 1992).
Commitment. Commitment is the extent to which an
exchange partner considers a relationship important and
thus is willing to work to sustain the relationship. Part-
ners in an exchange identify commitment as central to
attaining valuable outcomes for themselves (Morgan and
Hunt 1994). Thus, commitment is a critical component that
underscores the quality of the relational embeddedness of a
manufacturer–retailer association.
Overall, neither the manufacturer nor the retailer can pre-
dict new product success perfectly, in part, because of the
uncertainty related to consumer response. It is expected that
buyers will rely on the content, or quality, of their relational
embeddedness with sellers to reduce the level of uncertainty
and increase transaction efficiency. Thus, marketing rela-
tionships become market-based assets that have the poten-
tial to create value for firms by compensating for uncer-
tainty (Srivastava, Shervani, and Fahey 1998, 1999).
Therefore, the content of relationships should be an integral
part of models of corporate buyer decision making.
Hypotheses
Research has suggested that product attractiveness plays
a direct role in a retailer’s acceptance of new products (e.g.,
Rao and McLaughlin 1989). Extending this perspective, we
argue that the impact of buyer–salesperson and firm–firm
relationships on product acceptance is likely to be strongest
when the attractiveness of new products is modest. To
explain this, we rely on the heuristic–systematic model of
persuasion. The heuristic–systematic model specifies two
distinct processes by which people may be persuaded to
accept a point of view (e.g., favorably evaluate a product).
People who process heuristically “focus on the subset of
available information that enables them to use simple infer-
ential rules, schemata, or cognitive heuristics to formulate
their judgments and decisions” (Chaiken, Eagly, and Liber-
man 1989, pp. 212–13). If a heuristic is perceived as reli-
able, it may be employed more frequently because the per-
son is able to attain a sufficient level of confidence while
reducing effort (Chaiken, Eagly, and Liberman 1989).
Alternatively, systematic processing is typically viewed as
comprehensive and analytic. People access and scrutinize
all available information for its relevance and importance to
their judgment task so that the risk associated with the deci-
sion outcome is mitigated.
Buyers are trained to rely on product attractiveness while
evaluating new products to reduce the risk of choosing
products that may fail. As such, systematic processing in
the context of new product choice for retail display and
582 JOURNAL OF MARKETING RESEARCH, NOVEMBER 2006
sales implies a focus on product factors. Systematic pro-
cessing is adversely affected by situational variables that
constrain the capacity for in-depth information processing.
Given the sizable number of new products to evaluate, it is
plausible that buyers may be prone to process heuristically,
with relationship quality serving as a heuristic to aid new
product acceptance. This approach is especially likely when
product information or experience is less diagnostic, a con-
dition that is more likely to hold for new products than for
established ones (Hoch and Ha 1986). Thus, relationship
quality might have a maximum impact on product accept-
ance when product attractiveness is nondiagnostic. When
the new product is clearly unattractive or attractive, product
attractiveness alone is sufficiently diagnostic (meets the
“information sufficiency threshold”; see Chaiken, Eagly,
and Liberman 1989) to enable the buyer to make a confi-
dent decision to accept or to reject, and the quality of rela-
tionships should matter less. As the notion of structural
embeddedness suggests, we focus on relationships between
the buyer and the seller firms and between the salesperson
and the individual buyer because they are qualitatively dif-
ferent in their impact on transactions. In evaluating the
impact of both relationships, our position is consistent with
that adopted in previous research (e.g., Iacobucci and
Ostrom 1996).
Buyer–salesperson relationships. Many deals in the retail
industry are initially made verbally. The acceptance of the
salesperson’s spoken word without relying on contracts is
likely to be based on the quality of the relationship between
the buyer and the salesperson. In this regard, McEvily, Per-
rone, and Zaheer (2003) observe that trust, an aspect of
relationship quality, plays the role of a heuristic that reduces
cognitive effort in decision making. It is also likely that by
diminishing the perceptions of prospective risk of a transac-
tion by evoking memories of previous successful transac-
tions, satisfaction with a salesperson may enhance the
strength of the buyer–salesperson relationship as a heuristic.
A buyer’s commitment to continuing a relationship with a
salesperson would aid product acceptance in an uncertain
environment by enhancing the importance of choosing a
product that the salesperson presents. In summary, buyers
may use the quality of the relationship they have with sales-
people as a heuristic when making new product decisions.
This is more likely to happen when a judgment of product
attractiveness is insufficiently diagnostic. In the current
context, such conditions are likely to be reflected by
“middle-of-the-road” or modest new product attractiveness
judgments because of the tendency of respondents to mark
the middle of the scale when judgments are ambiguous.
Thus:
H1: When new product attractiveness is moderate, the likeli-
hood of new product acceptance increases when the quality
of the buyer–salesperson relationship goes from low to
high.
Firm–firm relationships. Whereas the buyer–salesperson
interaction may be of shorter time duration because of
turnover or reassignments, the firm–firm relationship may
be of a longer duration and, thus, a long-term “shadow” of
the future (Heide and Miner 1992). In the firm–firm rela-
tionship situation, although the trust and satisfaction that
define relationship quality are important because expected
benefits from future transactions would be uncertain in their
absence, commitment to sustaining the relationship is likely
to play a particularly important role. This commitment,
underscored by the potential for future interactions between
the firms, could reduce the perceived risk of accepting a
product of uncertain prospects. A high-quality firm–firm
relationship will diminish the risk perceived with a moder-
ately attractive new product because of the opportunity ren-
dered by future transactions to redress losses that a poten-
tially bad product choice can cause. As such, the likelihood
of future interactions diminishes the importance of the pay-
off in a current period (Heide and Miner 1992) and lowers
risk perceptions for products of moderate attractiveness.
Thus:
H2: When new product attractiveness is moderate, the likeli-
hood of new product acceptance increases when the quality
of the firm–firm relationship goes from low to high.
METHOD
Research Context
Two grocery retailers, one with approximately 500 stores
and the other with more than 1000, collaborated in this
research to learn how product acceptance is shaped by rela-
tionships. Vice presidents at these retailers served as
research sponsors and granted access to their buying teams.
These retailers target similar consumers and have similar
new product decision-making structures, in which buyers
make decisions independently rather than by committee.
Both retailers facilitated the study by (1) mapping the new
product decision process and isolating important factors, (2)
designing the questionnaire, (3) motivating the buyers to
participate in the study (which resulted in a high response
rate of 89% and 90%), and (4) providing feedback on the
results.
Product acceptance decisions are particularly sensitive
and are time delayed. The collaborative nature of the study
allowed us to obtain the relationship and product accept-
ance data concurrently with product evaluation and, subse-
quently, to confirm the acceptance or rejection of the prod-
ucts. In addition, a senior sales manager with a supplier to
the grocery chains (1) contributed knowledge of industry
practices, (2) suggested variables that are important in prod-
uct acceptance, (3) helped develop and refine measures, and
(4) assessed the results by providing a commentary on the
final report.
Questionnaire Design
More than 35 hours were spent on personal interviews to
understand the retail buying process with managers at the
sponsor organizations and at several other retail firms. We
could not tape the interviews because of the sensitive nature
of the information. On the basis of these interviews and an
extensive review of research on buyer–seller relationships,
we developed a preliminary version of the questionnaire.
Then, eight marketing academics reviewed the question-
naire. Subsequently, the questionnaires were sent to three
buyers who had not been interviewed previously and a vice
president at each retailer that collaborated in this research.
On the basis of their feedback, we made several minor
adjustments.
Relational Embeddedness and Selection of New Products 583

Data Collection Table 1

Typically, buyers are assigned to a product category and CONSTRUCT MEASURES AND LOADINGS
have relationships with between 5 and 50 salespeople. Buy-
ers for the nonperishables category, which includes every- Firm–Firm Relationship Quality
thing but meat, fish, bakery, produce, pharmaceuticals, and Loadings
floral departments, were identified by the two retailers that Retailer’s Trust of Manufacturer (6 items, α = .94)
participated in the study. Retailers A and B made 14 and 11 This manufacturer keeps promises it makes to our firm. .811
buyers available, respectively. Retailer A agreed to have This manufacturer is always honest with us. .897
each buyer complete 12 evaluations, and Retailer B agreed We believe the information that this manufacturer provides
to have each buyer complete 6 evaluations. us. .891
When making decisions, this manufacturer considers our
Data were collected over a six-week period for Retailer A welfare as well as its own. .873
and over a three-week period for Retailer B; each buyer We trust this manufacturer keeps our best interests in mind. .909
completed two evaluations per week. Because new product This supplier is trustworthy. .921
presentations were not always scheduled for each week, in
Retailer’s Commitment to Manufacturer (5 items, α = .94)
some cases, this period was extended. Each buyer was The relationship that my firm has with this manufacturer …
asked to evaluate a salesperson and manufacturer a maxi- is something we are very committed to. .946
mum of one time during the entire process. To minimize is very important to us. .926
carryover effects from one evaluation to the next, evalua- is something we intend to maintain indefinitely. .828
tions were ordered in two ways by varying the sequence of is something we really care about. .918
deserves our maximum effort to maintain. .878
the items. Buyers were asked to evaluate the first two new
products that were presented to them each week. To mini- Retailer’s Satisfaction with Manufacturer (5 items, α = .87)
mize response bias, buyers were promised anonymity and Our firm regrets the decision to do business with this
were told that the findings would be presented only in supplier. (R) .781
Overall, we are very satisfied with this supplier. .877
aggregate form. To encourage participation, a letter of We are very pleased with what this supplier does for us. .889
endorsement from the vice president at each participating Our firm is not completely happy with this supplier. (R) .698
retailer was sent to the buyers. The buyers were also offered If we had it to do all over again, we would still choose this
gift certificates that were redeemable at a local restaurant. supplier. .835
Because new product decisions are often time delayed, a
form was subsequently sent to buyers to confirm their deci- Buyer–Salesperson Relationship Quality
sion. A total of 210 evaluations were completed, with Buyer’s Trust in Salesperson (6 items, α = .96)
response rates of 89% and 90% for the two retailers. We This salesperson has been frank in dealing with me. .871
This salesperson does not make false claims. .862
excluded five evaluations from analysis because buyers I think this salesperson is completely open in dealing with
completed an evaluation for the same manufacturer. me. .928
This salesperson seems to be concerned with my needs. .866
Descriptive Statistics I trust this salesperson. .936
This salesperson is trustworthy. .944
Of the respondents, 75% were men. The average experi-
ence of the respondents in the grocery industry was 17.4 Buyer’s Commitment to Salesperson (5 items, α = .96)
years, and their experience in the buying function was 6.6 The relationship that I have with this salesperson …
years. The average length of buyer–salesperson relation- is something I am very committed to. .957
ships was a little more than 1.5 years. Buyers’ estimates of is very important to me. .940
is something I intend to maintain for the long-term. .934
the duration of the firm–firm relationship indicated that is something I really want to maintain. .946
most were long lasting (63% greater than 11 years). Each deserves my maximum effort to maintain. .856
buyer had an average of 13 vendor relationships in the cate-
gory under consideration. Products from companies with Buyer’s Satisfaction with Salesperson (4 items, α = .94)
I am delighted with my overall relationship with this
less than $2 million in sales made up 17% of evaluations, salesperson. .909
and products from firms with sales between $2 million and I wish more manufacturers’ salespeople were like this person. .930
$100 million and from firms with sales in excess of $100 I would like my relationship with this person to continue in
million made up 37% and 34%, respectively, of the evalua- the coming year. .912
tions (12% was unsure). Of the 205 products reviewed, 129 It is a pleasure dealing with this salesperson. .942
(63%) were accepted.
Product Attractiveness
Measures Product
Consumers who shop at our stores would judge this product
We assessed key variables using single-item and multi- to …
item formative and reflective measures. Whenever possible, possess high quality features or product composition.
we adapted measurement scales from previous research. belong to a strong brand family.
Table 1 lists the measures, and Table 2 provides descriptive be inexpensive.
statistics and correlations. have quality packaging.
have a low performance risk.
Relationship quality. We operationalized buyer percep- have a strong overall value.
tions of trust with the salesperson and the selling firm using
Doney and Cannon’s (1997) measures. We measured rela- Market Demand
tionship commitment on the firm–firm level using a scale This product will fulfill a current unmet need.
Demand is expected to be strong for this product.
adapted from Morgan and Hunt (1994). We adapted items
584
Table 1
CONTINUED
The category that this product would belong to is growing
quickly.
Economic conditions are very favorable for sales of this new
product.
Marketing Strategy
This product relative to other proposed new products in this
category has …
high planned media support (not including cooperative
advertising).
high planned couponing.
high planned product sampling/demonstrations.
strong introductory allowances.
a high slotting fee.
strong cooperative advertising funds.
a high estimated gross margin.
Notes: R = reverse coded.
Table 2
DESCRIPTIVE STATISTICS AND CORRELATION MATRIX
Variables 1 2
1. Manufacturer relationship 4.94 (.90)
2. Salesperson relationship .77 4.70 (1.04)
3. Product attractiveness .58 .55
4. New product acceptance .36 .40
Notes: Correlations greater than .13 are significant at p < .05. Means are
followed by standard deviations, which are in parentheses on the diagonal.
from this scale to measure buyer–salesperson relationship
commitment. We used Cannon and Perreault’s (1999) five-
item scale to measure relationship satisfaction with the
manufacturer, and we adapted Leuthesser and Kohli’s
(1995) five-item buyer satisfaction scale to measure the
buyer’s satisfaction with the salesperson.
Product attractiveness. The formative product attractive-
ness measure was composed of buyer perceptions of the
product (Sweeney, Soutar, and Johnson 1999), market
demand for the product, marketing strategy for the product,
and competition (Rao and McLaughlin 1989). Buyers were
asked to allocate 100 points across the four product attrac-
tiveness factors on the basis of their importance in the adop-
tion decision. We created a product attractiveness index by
averaging items within each of the four groups (product,
expected demand, marketing strategy, and competition) and
multiplying each group by the importance weight provided
by each buyer for the given product. Assigning weights for
each of these groups was important because buyers may
have different weightings for product characteristics. Miss-
ing importance data for three product evaluations led us to
assign equal weight to each group in these cases.
New product acceptance. We define a new product as a
stockkeeping unit, such as a single flavor or size, that the
retailer did not previously carry during the period of data
collection (see Rao and McLaughlin 1989). A dichotomous
measure (yes/no) of new product acceptance recorded
whether the firm actually accepted the product for display
and sales.
JOURNAL OF MARKETING RESEARCH, NOVEMBER 2006
Analysis and Measure Validation
We computed each relationship measure by taking an
average of the responses to the trust, commitment, and sat-
isfaction items.1 We subjected these measures to purifica-
tion procedures that were designed to evaluate dimensional-
ity and internal consistency (Gerbing and Anderson 1988).
Internal consistency, as determined by coefficient alpha,
exceeded .87 for each scale, and all items loaded highly on
the target construct. We compared a model with the correla-
tion of the two types of relationships constrained to one
with a freely estimated model. The constrained model
resulted in a significantly poorer fit (χ2(1) = 21, p < .05),
thus providing evidence of discriminant validity between
the two relationship constructs.
RESULTS
To determine whether either relationship had a stronger
impact on the product acceptance decision when product
attractiveness is moderate, we divided product attractive-
ness and each relationship into three levels: low, moderate,
and high. For grouping purposes, one standard deviation
below and above the product attractiveness and relationship
means delineated low and high levels, respectively. We cre-
3
ated a moderate relationship quality group using values that
fell between the poor-quality and high-quality groups (this
process is analogous to that which Aiken and West [1991]
4.42 (.87) advocate).2 The actual counts for each level of product
.68 attractiveness for each level of buyer–salesperson and firm–
firm relationship appear in Table 3, Panels A and B, respec-
tively, and the corresponding percentage acceptance rates
appear in Table 4, Panels A and B, respectively.
To evaluate H1 and H2, we used the Cochran-Armitage
procedure (a test of ordered proportions) to test for the pre-
dicted positive linear trend in the proportion of new prod-
ucts accepted across the three levels of relationship quality
(Agresti 2002, pp. 181–82). For a moderate level of product
attractiveness for both buyer–salesperson and firm–firm
relationships, acceptance rates increase when relationship
quality improves (χ2(1) = 17.33 and 5.79, respectively; p <
.05), in support of H1 and H2. Figure 1 depicts these rela-
tionships. Because of the large number of zeros in various
cells when product attractiveness is low or high, an analysis
of how relationship quality affects acceptance across all lev-
els of product attractiveness is not reliable (Agresti 2002).
However, an examination of the count data in Table 3 sug-
gests that when product attractiveness is low or high, there
is an almost 100% certainty of rejection or acceptance. This
can be considered supportive of our argument that for prod-
1We submitted all items to an exploratory factor analysis. Several
reverse-coded items from different scales formed separate factors. We
dropped these items from further analysis, as suggested by Herche and
Engelland (1996), who demonstrate that reverse-coded items regularly
load on separate factors.
2In this classification and in the ensuing analysis, we address the issue
of potential response homogeneity across buyers and products. To do so,
we undertook two steps. First, to address product-level heterogeneity, in
assessing product attractiveness, we asked the buyers to weight the differ-
ent factors, and we developed a weighted index. To address buyer-level
heterogeneity, we examined the contingency tables after controlling for the
buyer effect. We found the pattern of relationships in Table 4 to hold in this
case as well; product acceptance rates for moderately attractive products
increased for stronger buyer–salesperson and firm–firm relationships (p <
.05).
Relational Embeddedness and Selection of New Products 585

Table 3
COUNT DATA

A: Buyer–Salesperson Relationship
Decision
Relationship Product Attractiveness Reject Accept Percentage
Low Low 17.00 .00 .00
Medium 6.00 6.00 50.00
High .00 2.00 100.00
Medium Low 14.00 1.00 6.67
Medium 35.00 80.00 69.57
High .00 11.00 100.00
High Low .00 .00 —
Medium 4.00 17.00 80.95
High .00 12.00 100.00

B: Manufacturer–Retailer Relationship
Decision
Relationship Product Attractiveness Reject Accept Percentage
Low Low 21.00 .00 .00
Medium 7.00 11.00 61.11
High .00 1.00 100.00
Medium Low 7.00 1.00 12.50
Medium 34.00 74.00 68.52
High .00 11.00 100.00
High Low 3.00 .00 .00
Medium 4.00 18.00 81.82
High .00 13.00 100.00

Table 4 modest. We observed a similar pattern for the quality of the

PERCENTAGE OF ACCEPTANCE firm–firm relationship, in support of H2. When the product
is judged to be unambiguously weak or strong, the buyer
A: Across Levels of the Buyer–Salesperson Relationship When Product possibly perceives that little risk reduction can be obtained
Attractiveness Is Moderate through relying on relationship quality.
Decision Theoretical and Managerial Implications
Salesperson Impact of relationships on new product acceptance. We
Relationship Reject Accept Percentage show how retailers’ selection of new products is influenced
Low 6.00 6.00 50.00 by the embeddedness of the decision in relationships, both
Medium 35.00 80.00 69.57 at the firm and at the individual decision-maker level. The
High 4.00 17.00 80.95 study contributes to institutional theory by demonstrating
how the content of relationships, not their mere existence,
B: Across Levels of the Manufacturer–Retailer Relationship When influences organizational actions. The study contributes to
Product Attractivness Is Moderate the relationship marketing and new product literature
Decision streams by providing evidence of how relational embedded-
Retailer
ness can influence new product success. Our results indi-
Relationship Reject Accept Percentage cate that when the manufacturer does not offer a clear
competitive advantage (i.e., product attractiveness is mod-
Low 7.00 11.00 61.11
Medium 34.00 74.00 68.52 est), the likelihood that the buyer will accept the new prod-
High 4.00 18.00 81.82 uct increases by more than 60% when the manufacturer’s
sales representative has established a strong relationship
with the buyer. A similar, but somewhat weaker, benefit is
uct acceptance, relationship quality matters most when gained under the same product attractiveness conditions
product attractiveness is less diagnostic. when the firm–firm relationship is strong (i.e., approxi-
mately 33%). Thus, the benefits of relationship building
DISCUSSION seem clear in any situation in which a clearly superior prod-
We examine how retailers’ choice of new products is uct is not being introduced. Overall, our results support Sri-
embedded in firm–firm and buyer–salesperson relationships vasatava, Shervani, and Fahey’s (1998, 1999) argument that
by focusing on the role of relationship quality as a heuristic relationships serve as market-based assets that enhance
that simplifies decision making in an uncertain environ- transaction likelihood and create value.
ment. In support of H1, product acceptance increased sig- In light of these findings, manufacturers that are inter-
nificantly as a function of the quality of the buyer– ested in influencing new product acceptance rates may wish
salesperson relationship when product attractiveness was to assess their training practices. Although our data are
586 JOURNAL OF MARKETING RESEARCH, NOVEMBER 2006
Figure 1
ACCEPTANCE RATES FOR MODERATE PRODUCT
ATTRACTIVENESS BY LEVEL OF RELATIONSHIP
cross-sectional, the quality of a buyer’s relationship with a
salesperson was positively associated with the duration of
the relationship (p < .05). The consequences of shorter
buyer–salesperson relationships caused by turnover are not
easily quantifiable, but our findings indicate a clear benefit
for new product acceptance when relationships are strong.
This benefit is worth considering because an average pack-
aged goods company spends less than 4% of sales on its
sales force (De Vincentis and Kotcher 1995).
Practitioner feedback. The retailers and the national sales
manager who collaborated on our study provided feedback
on the results. The key points from the feedback are as
follows:
•Although practitioners expected relationships to play a role in
new product selection, it was not clear to them that this influ-
ence was restricted to moderately attractive products.
•Most buyers are spread too thinly across multiple categories
and cannot perform a lot of category analysis and administra-
tion. Thus, they rely on manufacturers to act as category advis-
ers. In such cases, retailers use manufacturers’ expertise to
learn to manage inventory, promote products, and arrange their
planogram (a diagram that shows how retail products should
be placed on retail shelves) more effectively. This level of sup-
port possibly entices the buyer to rely on the relationship as a
heuristic when evaluating moderately attractive products from
a trusted manufacturer.
•Relationships matter in the context of moderately attractive
products because they have a higher probability of failure than
clearly attractive products. Retailers know that if a product is
accepted despite its modest appeal and performs modestly as
well, suppliers and salespeople with whom they have strong
relationships will help mitigate their losses. This belief also
enhances the likelihood of the use of relationships as a heuris-
tic in new product decisions.
•On the negative side, retailers mentioned the need to disassoci-
ate the personal side of buyer–salesperson relationships from
the business side and ensure that the personal side does not
dominate the new product acceptance decision. However,
given the relatively stronger impact of the buyer–salesperson
relationship than that of the firm–firm relationship on new
product acceptance, there is a likelihood that this disassocia-
tion might not occur. If so, retailers might end up with inferior
product assortments that would be to the detriment of con-
sumers and, in the long run, to the retailers. Thus, although
relationships, as market-based assets, have the potential to gen-
erate value, improper reliance on this asset might lead to value
destruction.
Limitations and Directions for Further Research
First, a study based on self-report data has the potential
to be influenced by common method bias. The only certain
way of overcoming this problem would be to collect data
for independent and dependent variables from different
sources. However, this is a difficult task in organizational
research in which corporate buyers are the decision makers.
However, method bias is less likely to be a problem in
investigations of complex relationships, as we test in this
study.
Second, this study involves buyers at two large grocery
retailers; thus, the results may not generalize to the entire
population of corporate buyers at other retailers. Corporate
buyers in smaller retailers may behave differently. For
example, large manufacturers may have more distant rela-
tionships with smaller retailers because of the higher costs
involved with courting smaller retailers. This may manifest
in larger firms choosing to use wholesalers that possess
long-standing relationships with customers (retailers) to
blanket lower-volume accounts, thus sacrificing any chance
of building their own relationships. In addition, corporate
buyers in different industries may also behave differently.
Third, the data collection process in this study limited our
ability to examine the underlying process measures that
would help us ascertain whether relationship quality was
indeed used as a heuristic when new products were of
uncertain attractiveness. Experimental and qualitative
approaches are necessary to examine and confirm the
underlying processes that explain the influence of relation-
ship quality on new product selection.
Several additional areas for further investigation emerge
from this research. First, as Rao and McLaughlin (1989)
advocate, it would be meaningful if corporate buyer deci-
sions were linked to product performance. In this regard,
diagnostic information can be provided to buyers to
enhance decision making and help address the important
question of whether strong relationships impede or enhance
the accuracy of new product acceptance decisions. Another
avenue for research would be to examine the influence of
these relationships in group buying situations.
Second, additional research should determine how rela-
tionships and other factors affect the quality of acceptance
as determined by shelf placement and the frequency and
magnitude of in-store displays. A firm may be happy to
have its product accepted for distribution at a retailer but
dismayed to learn that it will be shelved in a suboptimal
location, substantively reducing its chance of success. Pre-
liminary qualitative research indicates that buyers may be
more likely to give preferred shelf space to manufacturers
that provide outstanding service (Cooper, Dröge, and
Daugherty 1991). Additional research examining how rela-
tionships and other factors influence a buyer’s decision to
delist or remove a product from the category might also
prove valuable.
Relational Embeddedness and Selection of New Products 587

REFERENCES Contact on Buyer-Seller Cooperation,” Academy of Manage-

ment Review, 35 (2), 265–91.
Agresti, Alan (2002), Categorical Data Analysis, 2d ed. New
Herche, Joel and Brian Engelland (1996), “Reversed-Polarity
York: John Wiley & Sons.
Items and Scale Unidimensionality,” Journal of the Academy of
Aiken, Leona S. and Stephen G. West (1991), Multiple Regres-
Marketing Science, 24 (Fall), 366–75.
sion: Testing and Interpreting Interactions. Newbury Park, CA:
Hoch, Stephen J. and Young-Won Ha (1986), “Consumer Learn-
Sage Publications.
ing: Advertising and the Ambiguity of Product Experience,”
Anderson, James C. and James A. Narus (1990), “A Model of Dis-
Journal of Consumer Research, 13 (September), 221–33.
tributor Firm and Manufacturer Firm Working Partnerships,”
Iacobucci, Dawn and Amy Ostrom (1996), “Commercial and
Journal of Marketing, 54 (January), 42–58.
Interpersonal Relationships: Using the Structure of Interper-
Bloom, Paul N., Gregory T. Gundlach, and Joseph P. Cannon
sonal Relationships to Understand Individual-to-Individual,
(2000), “Slotting Allowances and Fees: Schools of Thought and
Individual to-Firm, and Firm-to-Firm Relationships in Com-
the Views of Practicing Managers,” Journal of Marketing, 64
merce,” International Journal of Research in Marketing, 13 (1),
(April), 92–108.
53–72.
Cannon, Joseph P. and William D. Perreault Jr. (1999),
Kline, Barbara and Janet Wagner (1994), “Information Sources
“Buyer–Seller Relationships in Business Markets,” Journal of
and Retail Buyer Decision Making: The Effect of Product-
Marketing Research, 36 (November), 439–60.
Specific Buying Experience,” Journal of Retailing, 70 (1),
Chaiken, Shelly, Alice H. Eagly, and Akiva Liberman (1989),
75–88.
“Heuristic and Systematic Information Processing Within and
Kumar, Nirmalya (1996), “The Power of Trust in Manufacturer-
Beyond the Persuasion Context,” in Unintended Thought, J.S.
Retailer Relationships,” Harvard Business Review, 74 (6),
Uleman and J.A. Bargh, eds. New York: Harcourt Brace
92–107.
Jovanovich, 212–52.
Leuthesser, Lance and Ajay K. Kohli (1995), “Relational Behavior
Cooper, Bixby M., Cornelia Dröge, and Particia J. Daugherty
in Business Markets: Implications for Relationship Manage-
(1991), “How Buyers and Operational Personnel Evaluate Ser-
ment,” Journal of Business Research, 34 (3), 221–33.
vice,” Industrial Marketing Management, 20 (1), 81–85.
McEvily, Bill, Vincenzo Perrone, and Akbar Zaheer (2003), “Trust
Dacin, Tina M., Marc J. Ventresca, and Brent D. Beal (1999), “The
as an Organizing Principle,” Organization Science, 14 (1),
Embeddedness of Organizations: Dialogue and Directions,”
91–103.
Journal of Management, 25 (3), 317–56.
Montgomery, David B. (1975), “New Product Distribution: An
De Vincentis, John R. and Lauri Kien Kotcher (1995), “Packaged
Analysis of Supermarket Buyer Decisions,” Journal of Market-
Goods Salesforces—Beyond Efficiency,” The McKinsey Quar-
ing Research, 23 (August), 255–64.
terly, 1, 72–85.
Morgan, Robert M. and Shelby D. Hunt (1994), “The
De Wulf, Kristof, Gaby Odekerken-Schröder, and Dawn Iacobucci
Commitment–Trust Theory of Relationship Marketing,” Journal
(2001), “Investments in Consumer Relationships: A Cross-
of Marketing, 58 (July), 20–38.
Country and Cross-Industry Exploration,” Journal of Market-
Rao, Vithala R. and Edward W. McLaughlin (1989), “Modeling
ing, 65 (October), 33–50.
the Decision to Add New Products by Channel Intermediaries,”
Doney, Patricia M. and Joseph P. Cannon (1997), “An Examina-
Journal of Marketing, 53 (January), 80–88.
tion of the Nature of Trust in Buyer–Seller Relationships,” Jour-
Srivastava, Rajendra K., Tasadduq A. Shervani, and Liam Fahey
nal of Marketing, 61 (April), 35–51.
(1998), “Market-Based Assets and Shareholder Value: A Frame-
Federal Trade Commission (2003), “Slotting Allowances in the
work for Analysis,” Journal of Marketing, 62 (January), 2–16.
Retail Grocery Industry: Selected Case Studies in Five Product
———, ———, and ——— (1999), “Marketing, Business Pro-
Categories,” FTC Matter No. P001201, (November).
cesses, and Shareholder Value: An Organizationally Embedded
Gerbing, David W. and James C. Anderson (1988), “An Updated
View of Marketing Activities and the Discipline of Marketing,”
Paradigm for Scale Development Incorporating Unidimension-
Journal of Marketing, 63 (Special Issue), 168–79.
ality and Its Assessment,” Journal of Marketing Research, 25
Sweeney, Jillian C., Geoffrey N. Soutar, and Lester W. Johnson
(May), 186–92.
(1999), “The Role of Perceived Risk in the Quality-Value Rela-
Granovetter, Mark (1985), “Economic Action and Social Struc-
tionship: A Study in a Retail Environment,” Journal of Retail-
tures,” American Journal of Sociology, 91 (3), 481–510.
ing, 75 (1), 77–105.
Heide, Jan B. and Anne S. Miner (1992), “The Shadow of the
Williamson, Oliver (1985), The Economic Institutions of Capital-
Future: Effects of Anticipated Interaction and Frequency of
ism. New York: The Free Press.