The Word of Mouth Dynamic: How Positive (and Negative) WOM ...

DOI: 10.2501/JAR-53-1-043-060 March 2013 JOURNAL OF ADVERTISING RESEARCH 43

INTRODuCTIONIn the fields of social psychology and consumer

behavior, marketing researchers have produced

a considerable amount of theoretical work on the

effect of interpersonal communication. Such inter-

personal communication has become known as a

process of word of mouth (WOM) communication,

which now is regarded as one of the most impor-

tant and effective communications channels (Kel-

ler, 2007). Firms such as Nestlé, Procter & Gamble,

L’Oréal, Bosch, Microsoft, GlaxoSmithKline, and

Johnson & Johnson, to name just a few, increas-

ingly recognize that WOM is an extremely credible,

persuasive, and highly effective tool of informal

means of creating consumer engagement (Nielsen,

2009). Another example of the popularity of WOM

is the appearance of consultancies specializing

in this area (e.g., gaspedal.com and trnd [http://

company.trnd.com/en]). Such companies build

WOM campaigns as part of integrated marketing

communication. Likewise, the official trade asso-

ciation for the WOM marketing industry (Word

of Mouth Marketing Association, or WOMMA),

founded in 2005, promotes and advances WOM by

offering educational programs, ethical guidelines,

a standardized language, as well as pursuing a

research agenda and developing WOM metrics.

During the past decade, advances in electronic

communications technology have led to considera-

ble expansion in the number and types of informal

communications channels. Electronic newsgroups,

blogs, virtual communities, instant messaging, cell

phones, and Personal Digital Assistants (PDAs),

among other options, offer consumers instantane-

ous interactions with advertisers, fellow consum-

ers, and other market players (Allsop, Bassett, and

Hoskins, 2007; Hung and Li, 2007; Smith, Coyle,

Lightfoot, and Scott, 2007).

This uptick in activity has helped make WOM

a powerful communications channel that has an

important influence in the formation of consumer

opinions and in their purchase decisions. More-

over, this type of communication among con-

sumers is particularly interesting for advertising

practitioners, for two reasons:

• Research suggests that WOM can complement

and extend the effects of advertising (Bayus,

1985; Hogan, Lemon, and Libal, 2004), and

• Research shows that companies can stimulate

WOM through advertising (Graham and Hav-

lena, 2007).

Although previously many advertisers tradition-

ally have considered WOM as an alternative to

The Word of Mouth Dynamic: How Positive (and

Negative) WOM Drives Purchase ProbabilityAn Analysis of Interpersonal and

Non-Interpersonal Factors

RODOLFO VáZquEZ-CASIELLESUniversity of Oviedo, [email protected]

LETICIA SuáREZ-áLVAREZUniversity of Oviedo, [email protected]

ANA-BELéN DEL RíO-LANZAUniversity of Oviedo, [email protected]

This study has two main objectives: (a) to examine the relative impacts of positive and

negative word of mouth (PWOM and nWOM) on the shift in the receiver’s brand purchase

probability; and (b) to analyze the effect, direct or indirect, of a number of interpersonal

and non-interpersonal factors on the relation between PWOM or nWOM and the shift

in the receiver’s purchase probability. The data were collected from a sample of 1,035

consumers in four product/service categories. The results suggest that firms should

develop a proactive management of WOM communications that takes into account

aspects of both the sender and receiver.

44 JOURNAL OF ADVERTISING RESEARCH March 2013

THE WORD OF MOUTH DYnAMIC

advertising, WOM and advertising now

are regarded as two important communi-

cations channels that interrelate and com-

plement each other.

The main objective of the current arti-

cle is to analyze how WOM communica-

tion increases (or decreases) the receiver’s

brand-purchase probability.

WOM communication can be said to

have two dimensions (Harrison-Walker,

2001):

• “WOM activity” includes such aspects

as

– how often WOM takes place,

– the number of people with whom the

WOM sender communicates, and

– the quantity of information provided.

• “WOM valence” can be positive, nega-

tive, or neutral.

Although the content and strength of both

dimensions of the WOM condition affect

the probability of a consumer’s choosing

a particular brand, few studies have ana-

lyzed the impact of both dimensions.

The majority of studies in the WOM lit-

erature have focused on the WOM activ-

ity dimension. And, in fact, there is little

empirical evidence that helps explain how

positive and negative WOM contributes to

the shift in the probability of choosing a

brand (Assael, 2004; East, Hammond, and

Lomax, 2008).

Moreover, these studies primarily have

focused on positive WOM and its influ-

ence on which product is purchased. It also

is necessary to examine negative WOM,

which discourages purchase (Bruyn and

Lilien, 2008). To that end, this research

analyzes the WOM valence dimension

and the impact of positive and negative

WOM on the shift in the receiver’s pur-

chase probability.

Finally, this study also examines the

variables that explain—directly or indi-

rectly—the shift in the WOM receivers’

purchase probability. For this purpose, the

authors have grouped the variables into

• interpersonal factors (how actively

WOM is sought, strength of tie between

sender and receiver, and sender’s expe-

rience and strength of expression),

and

• non-interpersonal factors (receiver’s

loyalty, experience, and perceived risk).

In the current study, the authors’ research

has

• embraced the perspective of the WOM

receiver,

• compared the impacts of positive and

negative WOM, and

• determined the direct and indirect

effects of the different interpersonal and

non-interpersonal factors on the shift in

the purchase probability of the receiver

of positive (or negative) WOM.

The balance of this research

• presents the authors’ conceptual defin-

ition of WOM and the lines of research

examining how WOM works,

• describes a conceptual model, and

• presents the authors’ proposed

hypotheses.

Explanations of research design, meth-

odology, and results follow. The work

concludes with a discussion of the main

findings of the analysis and then offers

managerial implications.

LITERATuRE REVIEWThe literature defines WOM as “all infor-

mal communications directed at other

consumers about the ownership, usage,

or characteristics of particular goods and

services or their sellers” (Westbrook, 1987,

p. 261). Additionally, WOM can be any oral

and personal communication, positive or

negative, about a brand, product, service,

or organization, in which the receiver

of the message perceives the sender to

have a non-commercial intention (Arndt,

1967).

These definitions are consistent with

other WOM studies (Gruen, Osmonbekov,

Czaple, 2006; Harrison-Walker, 2001; Wan-

genheim, 2005; Wangenheim and Bayon,

2007), and share two characteristics:

• The receiver of the message must per-

ceive the sender to be unconnected with

any commercial organization. In other

words, for it to be credible, a WOM

recommendation must spring from a

natural dialogue between the two peo-

ple, and it should be the product of the

sender’s knowledge and the receiver’s

need to know.

• WOM can be either positive or negative.

Positive WOM encourages purchase,

whereas negative WOM discourages

purchase.

The WOM literature has three lines of

research.

• The first focuses on the perspective

of the WOM sender and analyzes the

reasons why people make positive (or

negative) recommendations on the basis

of their experiences with a product.

These studies conclude that a number

of forces (non-interpersonal factors)

exist and give rise to recommendations.

These forces include

– satisfaction or dissatisfaction (Ander-

son, 1998; Bowman and Narayandas,

2001; Brown, Barry, Dacin, and Gunst,

2005; Heitmann, Lehmann, and

Hermann, 2007; Maxham and Nete-

meyer, 2002), loyalty (Gounaris and

Stathakopoulus, 2004);

– commitment to the firm (Brown

et al., 2005; Dick and Basu, 1994;

Henning-Thurau, Gwinter, Walsh,

March 2013 JOURNAL OF ADVERTISING RESEARCH 45


and Gremler, 2002; Lacey, Suh, and

Morgan, 2007);

– trust (Ranaweera and Prabhu, 2003;

Sichtmann, 2007);

– service quality (Harrison-Walker, 2001);

– length of relationship with the firm

(Wangenheim and Bayon, 2004a

2004b); and

– perceived value (Matos and Vargas,

2008).

• The second line of research aims to

understand WOM receivers’ informa-

tion search behaviors or, more spe-

cifically, under what circumstances

(non-interpersonal factors) consumers

resort to WOM rather than other infor-

mation sources before making their pur-

chase decisions.

Researchers have found that con-

sumers are more likely to seek other

people’s opinions before purchasing

when they have less experience and a

stronger involvement in the purchase

of the product category (Gilly, Graham,

Wolfinbarger, and Yale, 1998) or when

they perceive greater risk in the decision

making (Bansal and Voyer, 2000).

• The third line of research also adopts

the perspective of the WOM receiver

and examines why some personal

information sources (positive and nega-

tive WOM) exert a stronger influence.

Researchers have identified forces

(interpersonal factors) that influence the

receiver’s behavior, such as

– the WOM sender’s experience and

strength of expression (Bansal and

Voyer, 2000; Gilly et al., 1998);

– the strength of the personal tie

between WOM sender and receiver

(Frenzen and Nakamoto, 1993);

– the demographic similarity (Brown

and Reingen, 1987); and

– the perceptual affinity (Gilly et al.,

1998).

CONCEPTuAL MODEL AND HYPOTHESESIn line with the aforementioned literature

on WOM, the authors have developed a

conceptual framework (See Figure 1) that

• adopts the perspective of the WOM

receiver;

• describes the relations between the type

of WOM (positive or negative) and the

shift in the brand-purchase probability;

• analyzes which type of WOM has the

most impact on the shift in the brand-

purchase probability; and

• investigates the effect—direct or indi-

rect—on the shift in the WOM receiver’s

brand-purchase probability of different

interpersonal (how actively WOM is

sought, strength of tie between sender

and receiver, and sender’s experience

and strength of expression) and non-

interpersonal factors (receiver’s loyalty,

experience and perceived risk).1

IMPACT OF POSITIVE AND NEGATIVE WOM ON SHIFT IN RECEIVER’S PuRCHASE PROBABILITYIn general, positive (or negative) WOM

is assumed to make the receiver more

1 The model hypothesised in this study is based on the theory of planned behaviour (Ajzen, 1991). In line with this theory, the authors postulate three conceptually independent deter-minants of behavioral intention (shift in the WOM receiv-er’s brand-purchase probability). The first is the attitude toward WOM. Attitude toward WOM is measured here by estimating the impact of positive and negative WOM on the shift in the receiver’s purchase probability. The second predictor is the subjective norm, which is concerned with judgments of the effects—direct or indirect—on the shift in the WOM receiver’s brand-purchase probability of different social influences (how actively WOM is sought, strength of tie between sender and receiver, and sender’s experience and strength of expression). The third antecedent is the degree of perceived behavioral control, which is concerned with perceptions of the effects, on the shift in the WOM receiver’s brand-purchase probability, of people’s confidence in their ability to deal with purchase situations (receiver’s experi-ence and perceived risk). The authors argue, however, that these three variables from the theory of planned behavior are insufficient to permit prediction of the shift in the WOM receiver’s brand-purchase probability. Anther variable—past behavior (receiver’s loyalty)—could further enhance the shift in the WOM receiver’s brand-purchase probability.

positive (or negative) about the object of

advice (East et al., 2008).

Some studies, however, have observed

the opposite response among receivers

(Fitzsimons and Lehmann, 2004): people

sometimes react against negative com-

ments and became even more committed

to the brand. Such contrary responses can

occur when

• people are directed to do things that

they do not want to do,

• the WOM receiver disagrees with the

values of the advisor, or

• when prior commitment to a brand may

prevent consumers from fully accepting

useful negative information about that

brand.

Assuming that contrary responses are not

common, then conceivably positive WOM

has a positive impact—and negative

WOM a negative impact—on the brand’s

purchase.

The authors thus offer their first

hypothesis:

H1: Positive (negative) WOM has a

positive (negative) impact on the

shift in the receiver’s purchase

probability.

It is also interesting to analyze which

type of WOM (positive or negative) has

the most impact. The literature offers little

evidence on this question.

Some studies have found that nega-

tive WOM has more impact than positive

WOM. For example, one field study found

that negative WOM has twice as much

impact on judgment or attitudes as posi-

tive WOM (Arndt, 1967). The author of

that report studied only one brand, how-

ever, and systematic research should be

based on all the brands in a category and

should include a range of categories. This

author also used a new brand about which



there could be few established beliefs.

Nevertheless, a later study confirmed the

previous results and observes that nega-

tive WOM is more influential than positive

WOM (Assael, 2004).

There is some theoretical justification for

the idea that negative information usually

has more impact on judgment than posi-

tive information. Indeed, one study found

that negative information usually was

rarer than positive information because

the latter can often be presumed (Fiske,

1980). In such instances, the relative rar-

ity of negative information surprises con-

sumers and, consequently, they pay more

attention to it.

This is the so-called negativity effect,

which has been observed in other studies

(Chevalier and Mayzilin, 2006) and can be

expressed in terms of the gap between the

position implied by the WOM message

and the receiver’s position.

This gap has diagnostic value: informa-

tion that restates what the receiver believes

may increase certainty but is unlikely to

change other aspects of a receiver’s judg-

ment (e.g., purchase probability). In many

markets, the greater amount of positive

information about the different brands

ensures that the position of most WOM

receivers is positive, so negative informa-

tion will have more impact.

In any case, the impact of positive (or

negative) WOM may differ when the

brands are familiar. For example, one

study analyzed the response of consum-

ers receiving positive and negative infor-

mation about brands and compared the

results when the consumers are familiar

or unfamiliar with the brand (Ahluwalia,

2002). According to this research, when the

brand is unfamiliar, the negative informa-

tion is perceived to have more diagnos-

ticity, but when the brand is familiar, no

significant differences exist in the impacts

of positive (or negative) information.

That study argued that brand familiarity

attenuates the perception of the greater

diagnostic value of negative information

and suggested that, under these circum-

stances, positive information may be per-

ceived to be more diagnostic than negative

information.

Such studies as those referenced here

typically measure how advice changes

judgment and/or attitude. In research on

the purchase of brands, it is more relevant

to measure the change in the purchase

probability brought about by WOM type.

From this perspective, using both role-play

experiments and surveys further meas-

ured WOM’s impact on choice as a shift

in the stated purchase probability, from

pre-WOM purchase probability to post-

WOM purchase probability (East et al.,

2008). These authors contended that, if

the pre-WOM purchase probability is

below 0.5, there is more room for change

in response to positive WOM than in

response to negative WOM.

For example, if the pre-WOM pur-

chase probability is 0.3, positive WOM

can have a maximum effect of 0.7 (up to

unity), whereas negative WOM can have

a maximum effect of only 0.3 (down to 0).

That study also sought to obtain a mean

pre-WOM purchase probability of 0.4, so

the authors argued that positive WOM

usually had more effect than negative

WOM.

In short, room for change in the brand-

purchase probability is limited by the pre-

WOM purchase probability, which could

favor the impact of either positive (or

negative) WOM, depending on the mean

pre-WOM purchase probability.

Thus, conceivably:

WOM COMMUNICATIONBETWEEN SENDER

AND RECEIVER

Positive WOMNegative WOM

SHIFT IN RECEIVER’S PURCHASE PROBABILITY

How actively WOM is sought

Tie strength

Sender’s experience

Sender’s strength ofexpression

Receiver’sexperience

Receiver’s loyalty

Receiver’sperceived risk

INTERPERSONALFACTORS

NON-INTERPERSONALFACTORS

How actively WOM is sought

Figure 1 Factors Associated with Impact of Positive and negative WOM on Shift in Receiver’s Purchase Probability



H2: Positive WOM has more impact

than negative WOM on the shift

in the receiver’s brand-purchase

probability when the pre-WOM

purchase probability is low.

INTERPERSONAL FACTORS ASSOCIATED WITH IMPACT OF POSITIVE AND NEGATIVE WOM ON SHIFT IN RECEIVER’S PuRCHASE PROBABILITYFigure 2 summarizes the characteristics

of the interpersonal factors (how actively

WOM is sought; strength of tie between

sender and receiver; sender’s experience

and strength of expression), the relation

between them, and their impact on the

shift in the receiver’s purchase probability.

HOW MARKETERS ACTIVELY SEEK WOMThe active search for information through

WOM is defined as “the process of seek-

ing and paying attention to personal

communications.”

Associated with the process of active

search for information through WOM is

the selective exposure to the messages

deriving from WOM communication,

which, in turn, implies that the consumer

has a greater predisposition toward the

WOM message (Arndt, 1967).

Thus, a message that is sought actively

will have a greater impact on the shift in

the WOM receiver’s purchase probability

than one that is received passively and is

moreover unsought and unrequested.

The authors’ third hypothesis follows:

H3: The more intense the WOM

(positive or negative) receiver’s

active search for information,

the greater the shift in the receiv-

er’s purchase probability.

TIE STRENGTHSources of WOM recommendation can

be classified according to the similarity of

the parties and the proximity or closeness

of the relationship between the WOM

receiver who must make the decision and

the WOM sender (Duhan, Johnson, Wilcox,

and Harell, 1997): in other words, in func-

tion of their tie strength. The tie strength

of a relationship is considered to be high

when the sender knows the receiver per-

sonally. Moreover, the tie strength contains

the following interpersonal dimensions

(Frenzen and Davis, 1990):

• closeness,

• intimacy,

• support, and

• association.

Tie strength, therefore, reflects the relation

and the type of tie existing between two

people.

A later work suggested that a high tie

strength will have a stronger influence

on the receiver’s behavior than a weak tie

strength (Frenzen and Nakamoto, 1993).

When the tie is strong, the WOM sender

and receiver will be more familiar with

each other, and the receiver will attribute

greater credibility to the sender. Moreover,

the receiver will be more likely to initiate

an active search for information.

In light of the preceding, the fourth

hypothesis is as follows:

H4: The stronger the tie between

the sender and the receiver, the

more actively the receiver will

seek information through the

two types of WOM (positive or

negative).

SENDER’S EXPERIENCE AND STRENGTH OF EXPRESSIONThe WOM sender’s experience conceiv-

ably will affect the way the WOM is

perceived.

Specifically, the receiver will seek infor-

mation more actively from a sender seen as

expert: in other words, someone who has

a high level of knowledge, competence,

Positive WOM

Negative WOM

WOM TYPE

H3 (+)

H4 (+)H5 (+)

H6 (+)

Tie strengthCloseness, Intimacy,Support, Association

Sender’s experienceKnowledge, Competence,

Education, Experiencein the product category

Sender’s strength of expressionConvincing arguments

How actively WOM is soughtSelective exposure to WOMPredisposition towards WOM


Figure 2 Interpersonal Factors: Impact on Shift in Receiver’s Purchase Probability



education, and experience in the product

category (Netemeyer and Bearden, 1992).

By contrast, if the receiver perceives the

sender’s knowledge, competence, educa-

tion, and experience in the product cat-

egory is low, the receiver is likely to be less

predisposed to seek or request informa-

tion from the sender to form an intention

or make a purchase decision.

The fifth hypothesis of the current study

is as follows:

H5: The greater the receiver’s per-

ception of the sender’s expe-

rience, the more actively the

receiver will seek information

through the two types of WOM

(positive or negative).

Conversely, the WOM sender’s strength

of expression can be defined as the WOM

receiver’s perception of the extent to which

the sender uses convincing arguments in

their positive (or negative) WOM.

Conceivably, the WOM sender’s

strength of expression will directly affect

the impacts of the positive and negative

WOM communication on the shift in the

receiver’s future purchase probability

(East et al., 2008).

This leads to the following hypothesis:

H6: The greater the WOM send-

er’s strength of expression, the

greater the shift in the purchase

probability of the consumer who

has received positive (or nega-

tive) WOM.

NON-INTERPERSONAL FACTORS ASSOCIATED WITH IMPACT OF POSITIVE AND NEGATIVE WOM ON SHIFT IN RECEIVER’S PuRCHASE PROBABILITYFigure 3 summarizes the characteristics of

the non-interpersonal factors (receiver’s

loyalty, experience, and perceived risk),

the relation between them, their relation

with an interpersonal factor (how actively

WOM is sought), and their impact on the

shift in the receiver’s purchase probability.

Receiver’s LoyaltyThe loyalty of the receiver of the positive

(or negative) WOM communication about

a brand can conceivably help explain the

shift in their future purchase probability.

Loyal receivers have a stronger moti-

vation for processing new positive infor-

mation (positive WOM) about the brand

they purchase habitually to reduce their

cognitive dissonance (Wangenheim,

2005) and reinforce their future purchase

behavior.

Furthermore, loyal receivers will have

a strong resistance to being persuaded

by negative information (negative WOM)

about the brand they purchase habitu-

ally and will try to convince themselves

that their previous decisions were right

and that the negative recommendation is

the result of the WOM senders’ one-off

market experience (Matos and Vargas,

2008; Sweeney, Soutar, and Mazzarol, 2008).

Thus, the conceptual framework

proposes

H7: The greater the WOM receiver’s

prior brand loyalty, the weaker

the impact of positive and nega-

tive WOM about the brand on

the shift in the receiver’s pur-

chase probability.

Receiver’s ExperienceVarious studies have suggested that a

negative relation exists between the con-

sumer’s experience and the active search

for external information before making

a purchase decision (Bansal and Voyer,

2000; Mishra, Umesh, and Stern, 1993).

This is because the expert receiver already

possesses sufficient knowledge about the

product category and has no need to con-

sult with other people before making a

decision.

Positive WOM

Negative WOM

WOM TYPEH3 (+) H7 (–)

H8 (–)

H9 (–)

H10 (+)


How actively WOM is soughtSelective exposure to WOMPredisposition towards WOM

Receiver’s loyaltyMotivation for processing positive WOM

Resistance to being persuaded by negative WOM

Receiver’s perceived riskSubjective expectation of losses

Individual characteristic

Receiver’s experienceConfidence about ability to

make an appropriate decision

Figure 3 non-Interpersonal Factors: Impact on Shift in Receiver’s Purchase Probability



By contrast, WOM receivers with little

experience or knowledge about the prod-

uct category are likely to lack confidence

about their ability to make an appropriate,

satisfying decision. Thus, they will feel the

need to consult with other consumers.

Hypothesis 8 of this work follows:

H8: The greater the WOM receiver’s

experience, the less actively they

will seek information through

the two types of WOM (positive

or negative).

Receiver’s Perceived RiskA perceived risk is defined as a “subjec-

tive expectation of losses” (Dholakia, 1997,

p. 161). The perceived risk variable is tied

to each product category, so the purchase

of different product categories is associ-

ated with different levels of perceived risk.

Perceived risk also is an individual char-

acteristic, as different people can perceive

different levels of risk when purchasing

the same product.

A relation conceivably exists between

WOM receivers’ experience in the prod-

uct category and their perceived risk in

its purchase. Consumers who are less

experienced in a particular product cat-

egory probably will perceive more risk

in that purchase and, from the informa-

tion economics perspective, they will

gain more from the information that the

WOM sender provides (Gilly et al., 1998).

The penultimate hypothesis of this work

follows:

H9: The more experienced the

receiver of the two types of

WOM (positive or negative), the

lower the perceived risk associ-

ated with the purchase of the

product/service.

In an effort to reduce the risk associated

with a purchase decision, consumers seek

information about the product. WOM is

one of the most effective sources of infor-

mation for reducing the risk associated

with the purchase of a particular product

(Guo, 2001).

People who perceive more risk in a pur-

chase situation tend to seek information

through WOM more actively than those

who perceive a lower risk (Arndt, 1967).

WOM, in fact, may be the most import-

ant information source for reducing risk

(Bansal and Voyer, 2000). It also has the

strongest impact on the receiver of the

communication, mainly because it permits

clarification of any doubts and offers the

chance of leaving feedback (Murray and

Schlacter, 1990).

The final hypothesis of the current

study:

H10: The greater the receiver’s per-

ceived risk in relation to the pur-

chase of the product, the more

actively they will seek informa-

tion through the two types of

WOM (positive or negative).

METHODOLOGYSample and Data CollectionTo test the 10 hypotheses, the authors

chose two consumer durables product

markets: mobile phones and laptop com-

puters; and two services: mobile-phone

companies and travel agencies.

These products and services were cho-

sen because

• consumers need to make evaluations

and comparisons when purchasing

them (high involvement),

• their use and consumption is very wide-

spread, and

• consumers are likely to perceive risk in

the purchase decision.

The nature of the choices helped ensure

that the authors could gather a sample

of subjects that had received positive (or

negative) WOM communications from

people unconnected to the different mar-

ket players.

The study participants were recruited

randomly in various shopping centers in

northern Spain. To ensure familiarity, the

consumers interviewed had to own and

use a mobile telephone or laptop com-

puter or use the services of a mobile-phone

company or travel agency. They also had

to have received, in the past 3 months,

positive (or negative) WOM for different

brands of the aforementioned products or

services.

After this random sampling system, the

authors obtained information from a sam-

ple of 1,100 individuals. They excluded

65 questionnaires due to incomplete data,

which resulted in a net sample size of

1,035.

Each subject was asked about only one

product or service. The sample, therefore,

consisted of 1,035 different individuals.

Furthermore, the sample consisted of

actual consumers of the products or ser-

vices analyzed, not of students.

The distribution of the subjects by prod-

ucts/services was as follows (Table 1):

• 258 subjects owned and used a mobile

phone,

• 253 owned and used a laptop,

A relation conceivably exists between WOM

receivers’ experience in the product category

and their perceived risk in its purchase.



• 268 used the services of a mobile-phone

company, and

• 256 used a travel agency.

The distribution by gender was 47.7 per-

cent male and 52.3 percent female. A total

of 45.4 percent of the respondents were

between ages 18 and 34 years, 40.5 percent

between 35 and 54 years, and 14.1 percent

older than 54 years. Finally, 52.6 percent

of the respondents had received positive

WOM, whereas 47.4 percent had received

negative WOM.

The authors collected the information

they required using structured question-

naires presented in the course of personal

interviews. The first part of the ques-

tionnaire asked the respondents (WOM

receivers) about their experience with the

category, their current brand, their ten-

dency to ask for advice before making the

purchase, and their perception of the risk

in the purchase.

The second part of the questionnaire

focused on the WOM received by the

individuals. The interviewees were asked

whether they had received many or only

a few positive (or negative) WOM com-

munications about a particular product or

service in recent months.

If they had received a number of

WOM communications, they were asked

to consider the single communication

they felt had had the most impact on their

decision.

The questionnaire then asked survey

participants about the following:

• the brand that was the subject of the

positive (or negative) recommendation;

• a number of points relating to the WOM

sender:

– strength of relationship with them,

– experience in the product/service cat-

egory, and

– credibility of arguments used; and

• the probability of choosing the brand

before and after the positive (or nega-

tive) recommendation.

The questionnaire concluded with ques-

tions about the age and gender of the

respondents.

Variable MeasuresUsing construct definitions and pre-

existing measures available from the liter-

ature, the authors generated a set of items

for each construct. They also consulted

five practitioners and academic experts.

Any problematic items were either deleted

or appropriately modified.

The authors presented the resulting

items to 20 consumers of the products

and services analyzed in this research in

face-to-face meetings to ensure that the

questions were worded with appropri-

ate consistency. They then revised several

items on the basis of the feedback. Finally,

to measure each of the constructs of inter-

est, they adopted single-item and multi-

item scales (Appendix).

The authors used three single-item

measures to capture information about

brand-purchase probability before receiv-

ing the WOM recommendation (receiver’s

loyalty), brand-purchase probability after

receiving the WOM recommendation, and

sender’s strength of expression.

A single-item measure is sufficient if

the construct is such that in the mind of

“raters” (e.g., respondents in a survey),

the object or attribute of the construct is

“concrete,” meaning that it consists of one

object or attribute that is easily and uni-

formly imagined (Bergkvist and Rossiter

2007, 2009).

In the current study, the object (brand)

or the attribute (how convincing/cred-

ible were the explanations of the person

who sent the WOM) is concrete, so the

authors used single-item measures to cap-

ture information about brand-purchase

probability (a 10-point scale) and sender’s

strength of expression (7-point scale).

The shift in the receiver’s brand-

purchase probability was calculated as

the difference between the probability of

choosing the brand after and before the

recommendation.

The authors also use multi-item meas-

ures to capture information about five

constructs linked to interpersonal and

non-interpersonal factors of the WOM

communication:

TABLE 1Sample Characteristics

Product – Service

Mobile phone Laptop

Mobile-phone services

Travel agency services Total

SUBJECTS 258 253 268 256 1,035

Male (%) 46.8 53.8 44.6 46.2 47.7

Female (%) 53.2 46.2 55.4 53.8 52.3

Between 18 and 34 (%) 49.4 47.8 44.5 40.1 45.4

Between 35 and 54 (%) 38.3 43.5 42.2 38.6 40.5

Over 54 12.3 8.7 13.3 21.3 14.1

Positive WOM (%) 52.4 54.2 50.8 53.1 52.6

negative WOM (%) 47.6 45.8 49.2 46.9 47.4



• how actively WOM is sought,

• tie strength,

• sender’s experience,

• receiver’s experience, and

• receiver’s perceived risk.

To evaluate each of these measures, the

authors used 7-point scales. The principal

theoretical argument for using a multi-

item measure of a construct is that a multi-

item measure captures more information

and is more likely to tap all facets of the

construct of interest (Bergkvist and Ros-

siter, 2007, 2009). Another argument for

multiple items is that it increases reliabil-

ity by allowing the calculation of the coef-

ficient alpha, which indicates the “internal

consistency” of all the items that represent

the presumed underlying construct.

This “reliability” argument, however,

needs to be qualified. The coefficient alpha

never should be used without first estab-

lishing the unidimensionality of the scales;

this can be investigated by factor analysis.

Thus, the authors ran five exploratory fac-

tor analyses for the constructs linked to

interpersonal and non-interpersonal fac-

tors of the WOM communication. These

tests confirmed the unidimensionality of

each of the five scales, which means the

coefficient alpha can be used.

The resulting coefficient alphas were

high, indicating “internal consistency”

and supporting the use of multi-item

measures to capture information about

the constructs linked to interpersonal and


communication (See Appendix).

Another problem that the authors con-

sidered was common-method bias, which

occurs when the correlation between two

or more constructs is high. To analyze

common-method bias, the authors ran

an exploratory factor analysis with all

the attributes linked to interpersonal and


communication.

The results obtained allowed the authors

to identify five factors corresponding to

interpersonal and non-interpersonal con-

structs of the WOM communication. The

factor loadings were high in each of the

five dimensions identified, and the factors

together explained 85 percent of the vari-

ance (KMO = 0.897).

As a result, the common-method bias

was not a problem for the current analysis,

so multi-item measures could be used in

subsequent analyses to obtain information

about the constructs linked to interper-

sonal and non-interpersonal factors of the

WOM communication.

Data Analysis: The Measurement ModelThe data analysis employed a two-step

procedure. The measurement model was

estimated for the entire sample (positive

and negative WOM receivers) prior to the

analysis of the structural model.2 All vari-

ables of interest were conceptualized as

reflective first-order constructs. A meas-

urement model including how actively

WOM is sought, tie strength, sender’s

experience, receiver’s experience and per-

ceived risk was subjected to confirmatory

factor analysis using structural equation

modeling (with EQS).

As a result, a 22-item, five-factor covari-

ance structure measurement model was

estimated to assess the fit, reliability, and

validity of the measurement scales of the

model constructs. In addition, the average

variance extracted (AVE), the composite

reliability coefficient (CR), and the stand-

ardized lambda parameters also were

2 The results of the structural model (causal analysis) can be seen in the Results section.

examined. Together, these tests provide

evidence of reliability and validity. See

Appendix for a summary of the scales’

psychometric properties, which were

obtained from the measurement model.3

RESuLTSDescriptive Analysis: Impact of WOM Communications on Purchase ProbabilityDescriptive analysis (using SPSS) was

employed to test the hypothesized rela-

tions H1 and H2 in the four product/ser-

vice categories. Table 2 (col. 1) shows the

categories. Columns 2 and 3 are the pur-

chase probabilities prior to positive and

negative WOM, respectively, and columns

4 and 5 are the mean shifts in the purchase

probability produced by positive and neg-

ative WOM, respectively.

According to these results, positive

WOM has a positive impact, and negative

WOM has a negative impact on the shift in

the receiver’s brand-purchase probability.

H1 is supported.

Furthermore, when the mean impacts of

positive and negative WOM are measured

as the shift in purchase probability, posi-

tive WOM produces a mean shift of 1.8542,

and negative WOM produces a mean shift

3 The measurement model fits the data well. Regarding reliability, each construct has a composite reliability and AVE greater than the recommended threshold values of 0.5 and 0.6, respectively. In addition, for all constructs, the Cronbach alpha coefficient exceeds 0.9. Convergent validity is supported as all lambda parameters are significant and greater than 0.5. Discriminant validity is supported, as the confidence intervals of the correlations between all the variables do not include 1.0 and the squared correlations do not exceed the AVE. Finally, the fit statistics indicate a good model fit (Root Mean Square Error of Approximation = 0.065; Bentler-Bonett Non-Normed Fit Index = 0.955; General Fit Index = 0.966; robust Comparative Fit Index = 0.961).

Another problem that the authors considered was

common-method bias, which occurs when the

correlation between two or more constructs is high.



of –1.6261, which means positive WOM is

14 percent more influential than negative

WOM. When absolute numbers are tested,

positive WOM has significantly more

impact than negative WOM in the means

data and across categories of products

(p < 0.001 one-tailed exact tests).

Thus, analyzing the shifts in purchase

probability, H2 is supported. Overall, posi-

tive WOM has more impact that negative

WOM.

The authors also used descriptive anal-

ysis (with SPSS) to test the hypothesized

relation H7 in the four product/service

categories. The results do show that the

greater the WOM receiver’s previous loy-

alty, the weaker the impact of both posi-

tive and negative WOM communications

on the shift in the brand-purchase proba-

bility. Table 3 and Figure 4 present the rela-

tion between previous brand loyalty and

shift in the brand purchase probability in

a more accessible form. For both positive

and negative WOM, a relatively straight

section on each plot is evident, which

then deflects toward the x-axis. These

deflections can be attributed to the effect

of brand commitment and show how this

factor constrains the impact.

Thus, this work provides support for

H7: the receiver’s loyalty reduces the

impact of positive and negative WOM on

the brand-purchase probability.

Causal Analysis: Structural Model EvaluationThe authors used structural-equation

modeling (with EQS) to test the remaining

hypotheses in the two subsamples: con-

sumers who had received positive WOM

or negative WOM (Table 4).

Most of the proposed relations are

supported in both subsamples: consum-

ers who had received positive (or nega-

tive) WOM. Thus, the more intense the

receiver’s active search for information

and sender’s strength of expression, the

greater the shift in the brand-purchase

probability (See Table 4).

Thus, H3 and H6 are supported.

The results also confirm the influence

of sender’s experience, receiver’s experi-

ence, and receiver’s perceived risk on how

actively WOM is sought, and H5, H8, and

H10 are supported.

The results also show that H9 is sup-

ported in both subsamples: the greater the

experience of the receiver of the positive

(or negative) WOM communication, the

lower the risk perceived in the purchase of

the product/service.

TABLE 3Mean Shift in Brand Purchase Probability as Function of Loyalty

Purchase probability prior to WOM (receiver’s loyalty)

Mean shift in purchase probability after positive WOM

Mean shift in purchase probability after negative WOM

X-axis Y-axis Y-axis

0 1.2206 0.0000

1 1.5778 −0.3846

2 1.9029 −0.9526

3 2.2857 −1.2806

4 2.4240 −1.6907

5 2.5563 −2.1852

6 1.9337 −2.5957

7 1.3508 −2.5222

8 0.7143 −1.6870

9 0.6471 −0.7143

10 0.0000 −0.4667

TOTALS 16.6132 −14.4796

The relation between previous brand loyalty and shift in the brand purchase

TABLE 2Impact of Positive and negative WOM on Shift in Receivers’ Purchase Probability

Product/Service Category

Purchase probability Shift in purchase probability

Prior to positive WOM

Prior to negative WOM

Shift after positive WOM

Shift after negative WOM

Mobile phone 4.4336 4.5245 1.7310 −1.4951

Laptop 4.4389 4.6631 1.9005 −1.6631

Mobile-phone services 3.6892 3.7330 1.7149 −1.5519

Travel agency 4.6347 4.9896 1.8767 −1.7187

TOTAL 4.2984 4.4721 1.8542 −1.6261

Descriptive analysis: impact of receivers’ loyalty on purchase probability



The results, however, only partially

support the influence of the strength of

the tie between sender and receiver on

how actively WOM is sought. The param-

eters obtained are significant for the

subsample of consumers who received

positive WOM but non-significant for the

subsample of consumers who received

negative WOM.

H4 is supported for positive WOM but

not for negative WOM.

DISCuSSION AND CONCLuSIONSWOM long has been recognized as a pow-

erful force affecting consumers’ attitude

and choice. The results of the current study

enhance current understanding of how

WOM influences the receiver’s choice.

This article, unlike previous studies,

represents an attempt to explicitly test the

differential impact of positive and nega-

tive WOM on the mean shift in the brand-

purchase probability.

Furthermore, this work differs from

most earlier studies on WOM in that the

authors investigated various brands in a

range of categories of products (mobile

phones and laptops) and services (mobile-

phone companies and travel agen-

cies). From this perspective, this study

makes various contributions to the lit-

erature on WOM communications in

marketing.

Specifically:

0,0000

0,1250

0,2500

0,3750

0,5000

0,6250

0,7500

0,8750

1,0000

1,1250

1,2500

1,3750

1,5000

1,6250

1,7500

1,8750

2,0000

2,1250

2,2500

2,3750

2,5000

2,6250

2,7500

2,8750

3,0000

0 1 2 3 4 5 6 7 8 9 10

Positive WOM Negative WOM

Purchase probability prior to WOM (receiver’s loyalty)

Mea

n sh

ift in

pur

chas

e pr

obab

ility

aft

er p

ositi

ve a

nd n

egat

ive

WO

M

Figure 4 Mean Shift in Purchase Probability after Positive and negative WOM



• The empirical analysis shows that posi-

tive (negative) WOM has a positive

(negative) impact on the shift in the

receiver’s brand-purchase probabil-

ity. The results also show that positive

WOM has a stronger impact on brand-

purchase probability than negative

WOM.

An explanation for positive WOM’s

stronger effect in this study is that the

prior purchase probability tends to be

below 5 on a 10-point scale. In particu-

lar, the prior purchase probability is

4.2984 for the positive WOM subsample

and 4.4721 for the negative WOM sub-

sample. This situation leaves more room

for change in response to positive WOM

than in response to negative WOM.

Thus, the results suggest that nega-

tive WOM is less diagnostic than posi-

tive WOM. There is a “positivity effect”

(using Fiske’s gap explanation), with

positive WOM having more impact than

negative WOM.

• The results also show that the same

interpersonal factors govern the impact

of both positive and negative WOM

on the shift in the receiver’s brand-

purchase probability. The authors can

conclude that the sender’s strength of

expression has the greatest influence for

both positive and negative WOM, fol-

lowed by how actively WOM is sought.

The findings of this study suggest that

when the sender’s strength of expres-

sion is high and when WOM (positive or

negative) is actively sought, WOM will

have a significant influence (positive or

negative) on the shift in the receiver’s

brand-purchase probability.

Thus, marketing strategies designed

to promote interpersonal communica-

tion will reach more senders/receivers

and be more efficient if they are directed

at senders with strength of expression

and receivers who are motivated to seek

information through WOM.

Firms also should pay particular

attention to the potential influence of

negative WOM, as these communica-

tions reduce the purchase probability.

Consequently, companies should also

adopt decisions about marketing strat-

egies directed at senders and receivers

with the objective of minimizing the

sending of negative WOM and/or the

effects of exposure to negative commu-

nications of people motivated to seek

advice actively.

Nevertheless, the effect of both

interpersonal factors on the receiver’s

decision is stronger when the WOM

information is positive (positive WOM)

than when it is negative (negative

WOM).

TABLE 4Structural Models: Parameter Estimates

Hypothesised pathsPositive WOM subsample Supported

Negative WOM subsample Supported

H3: How actively WOM is sought → Shift in brand purchase probability 0.226** YES −0.065* YES

H4: Tie strength → How actively WOM is sought 0.096** YES 0.015 nO

H5: Sender’s experience → How actively WOM is sought 0.155** YES 0.135** YES

H6: Sender’s strength of expression → Shift in brand purchase probability 0.489** YES −0.440** YES

H7: Receiver’s loyalty → Shift in brand purchase probability −0.325** YES −0.426** YES

H8: Receiver’s experience → How actively WOM is sought −0.159** YES −0.126 YES

H9: Receiver’s experience → Receiver’s perceived risk −0.218** YES −0.219** YES

H10: Receiver’s perceived risk → How actively WOM is sought 0.646** YES 0.659** YES

Bentler-Bonett non-normed Fit Index (BBnnFI) 0.910 0.915

Comparative Fit Index (CFI) 0.919 0.923

General Fit Index (GFI) 0.924 0.928

Root Mean Square Error of Approximation (RMSEA) 0.075 0.080

Satorra-Bentler Scaled Chi Square S-Bc2

(probability value)1231.1561 (0.000)

1279.0923 (0.000)

Standardised parameters are shown (**p < 0.01)



• The current findings support the

hypothesis that level of receiver’s loy-

alty (a non-interpersonal factor) reduces

the impact of both positive and nega-

tive WOM on the shift in the receiver’s

purchase probability. Thus, the effect of

WOM (positive or negative) is condi-

tioned by the receiver’s previous loyalty.

As the receiver’s level of loyalty

toward a brand increases, positive and

negative WOM about that brand will

have less impact on the shift in the

future purchase probability. Looking at

the plots in Figure 4, positive messages

(positive WOM) clearly have more

impact when the receiver’s pre-WOM

loyalty is from 0 to 6, whereas negative

messages (negative WOM) have more

influence in the range 4 to 7 (See Figure 4).

Thus, the potential impact of WOM

(positive or negative) can be estimated

for any segment of consumers if the

mean pro-WOM loyalty can be assessed

using purchase records or management

judgment, for example.

• Various factors directly influence how

actively WOM is sought and indirectly

affect the shift in the receiver’s purchase

probability.

The current findings indicate that,

when senders are perceived as knowl-

edgeable, the receivers are motivated

to actively seek information (positive or

negative WOM) from them. Thus, a sig-

nificant positive relation exists between

the two constructs.

Likewise, when the tie between send-

ers and receivers is strong, the receivers

are motivated to actively seek positive

WOM information (empirical evidence

for negative WOM was not found).

Conversely, the receiver’s experience

was also found to be a significant indi-

cator of how actively WOM is sought.

The more knowledgeable people are

or the more experience they possess,

the less intense will be the active search

for information (positive or negative

WOM).

Furthermore, the greater the receiv-

er’s experience, the less risk they will

perceive in the purchase; and the greater

the perceived risk, the more active the

search for WOM information (positive

or negative WOM).

• For both positive and negative WOM,

the receiver’s perceived risk has the

strongest positive effect, followed by the

receiver’s experience (negative effect),

sender’s experience (positive) and, to

a lesser extent, strength of tie between

sender and receiver (positive influence

only for positive WOM).

The practical implication of these empiri-

cal results is that companies should pay

particular attention to the consumers who

are most motivated to seek advice actively

(less experienced consumers who perceive

more risk in the purchase) to maximize

their exposure to positive communications

from senders perceived as knowledgeable

and with whom they have strong ties and

minimize their exposure to negative com-

munications from senders perceived as

experts.

Management ImplicationsWOM is one of today’s most powerful

marketing tools. It is reported to be one

of the fastest growing sectors in market-

ing and media services. Smart marketers

have an opportunity to become a part of

the consumer-driven WOM conversation

through well-planned, well-researched,

and well-executed WOM marketing pro-

grams—at which time, they will be well

positioned to influence consumers’ pur-

chase intentions.

For marketers, the findings of this study

suggest that companies should develop

marketing strategies to encourage positive

WOM messages and increase their effec-

tiveness. At the same time, companies

should clearly also make efforts to avoid

negative WOM.

Thus, academics and marketing direc-

tors should pay more attention to the

management of WOM, as WOM can

complement the firm’s policy of advertis-

ing communication and, hence, increase

its efficacy. Likewise, not only are both

types of communication complementary,

which improves the firm’s performance,

but traditional media and marketing com-

munications have a significant role to play

in influencing conversations. Firms can

use advertising to spread messages in the

media that stimulate consumers to speak

favorably about their brands and say good

things about their products.

This ideal situation, however, will hap-

pen only if the brand is credible, the firm’s

products/services are reliable, and its

marketing activities are believable. In such

cases, customers are very likely to initi-

ate more positive conversations about the

brand than negative ones.

Given the need to obtain a positive WOM

flow from senders to receivers, companies

should aim to keep the senders satisfied

by providing a good product or service,

effectively practicing sender-centered rela-

tionship marketing. In this context, WOM

programs should create experiences for

consumers and convey information that

encourages influential individuals or

groups to talk freely, authoritatively, and

credibly with others.

Enabling consumers to co-create brand

meaning and tell stories is essential to

WOM. Bain & Co. has reported there

is no better force to drive sales growth

than strong customer advocacy. Indeed,

its research shows that the most recom-

mended company in its category grows

2.5 times more than the category average.

Likewise, Booz & Co., a leading consul-

tancy has advised, “Make your consumer



an advocate: shift marketing objectives

from sending a message to facilitating con-

versations with and between consumers.”

Thus, companies should use marketing

activities to incentivize satisfied consum-

ers with a high degree of loyalty to act as

senders of positive WOM messages, par-

ticularly to those people (receivers) with

whom they maintain a strong tie. It is not

just a question of encouraging an “endog-

enous WOM,” which is characterized by a

conversation that occurs naturally among

consumers as a function of the senders’

positive experiences with the product or

service. Firms also should develop a proac-

tive management of customer-to-customer

communications. In other words, they cre-

ate an “exogenous WOM” as a result of

its actions (effecting meaningful positive

WOM).

For this purpose, companies should use

loyalty programs to incentivize satisfied

customers to become effective dissemi-

nators of information. In fact, companies

should put the satisfied customer at the

center of a communications campaign and

encourage them to offer potential con-

sumers detailed positive information—

convincingly argued—about the brand’s

good points.

The satisfied senders’ experience and

strength of expression can help them to be

perceived as “opinion leaders.” It is in this

context that a company needs to identify

the most dynamic consumers in the rec-

ommendation activity, as they can help it

attract new customers with a minimum

economic investment. Moreover, the eco-

nomic value that an active customer can

generate via WOM is a good means of

segmenting customers in function of their

capacity to influence the rest—the goal of

all marketers.

One other aspect of interest to practition-

ers in the strategic management of WOM:

in what environment is WOM communi-

cation more commonly generated? Since

2006, advertising-agency/WOM consult-

ants Keller Fay has compiled data since

2006 that consistently show that WOM

conversations happen most often in off-

line environments. Clearly, practitioners

cannot ignore the off-line conversations

that take place between people, but they

should also try to exploit to the maximum

the opportunities opened up by the Inter-

net (Fong and Burton, 2006) and encourage

people to support WOM so the company

can benefit from it as soon as possible.

Much of the importance of the Internet

for WOM communication is grounded in

the links between people in a social net-

work. A recommendation is more valu-

able in groups that have stronger links.

People with closer relationships—married

couples or friends, for example—tend

to interact more frequently than mere

acquaintances.

Knowing this, companies might target

“evangelical” customers—the most inter-

connected groups in the Internet—by

analyzing social networks, cliques, or the

groups that such contacts create and iden-

tify the most influential people in these

groups.

Companies also should remember that

the most socially connected individuals

are more liable to offer recommendations

and that physical proximity is not a sig-

nificant variable. It is critical, therefore,

that companies understand the social net-

works in which consumers operate (their

links and how they operate in them) to

extend positive WOM and minimize the

effects of any negative WOM.

In short, the diffusion of opinion lead-

ers’ messages in both traditional and mod-

ern communications media (e.g., social

networks such as Facebook, Twitter, and

LinkedIn) will help persuade positive

WOM receivers to increase their brand-

purchase probability.

In parallel, companies also should

stimulate the desire in the receiver to

obtain advice about the product or service.

Firms can use market research to identify

consumers (receivers) with a moderate

purchase probability or loyalty, and distin-

guish segments with high (or low) experi-

ence in the product or service category.

The aim is for the receivers to seek infor-

mation from satisfied customers so their

future purchase probability increases. An

advertising campaign could encourage

both segments of consumers to seek infor-

mation from other people: for example,

to seek information from satisfied, loyal

expert senders with whom they have a

strong tie or who are considered an object-

ive, third-party source. Receivers with

experience in the product/service cat-

egory and moderate brand loyalty could

begin to doubt and eventually conclude

that they do not possess as much experi-

ence as they initially had thought.

Another possibility to increase the

effectiveness of a positive WOM mes-

sage is to do market research to classify

potential consumers (receivers with a

moderate purchase probability and low

experience) into segments according to the

risk that they primarily associate with the

purchase decision. The resulting groups

could be addressed with differential com-

munications and interpersonal influence

strategies.

For example, potential customers high

in functional risk could be encouraged

to connect with satisfied, loyal expert

senders about the superiority of a firm’s

product/service. Customers with high

social/psychological risk perception

could be recommended to senders in their

peer group with whom they maintain a

strong tie and perceive as being similar to

themselves.

Limitations and Future Lines of ResearchThis study has a number of important

limitations, which can be seen as starting

points for future research.



• The work uses retrospective data in that

the receivers must remember positive

(or negative) WOM messages that have

had an impact on their brand choice and

indicate the probability of choosing the

brand before and after the positive (or

negative) recommendation.

It would be advisable in future work

to subject the receivers to actual posi-

tive and negative WOM situations and

record their actual purchase decision.

• This study focuses on receivers who

actively seek information: in other

words, receivers who already are inter-

ested in the product or service category.

This analysis is important, but it does

not shed any light on why some WOM

communications have no influence at

all, whether positive or negative.

• The subsamples for the four categories

analyzed have low or moderate mean

pre-WOM purchase probabilities, which

explains why negative WOM is less

diagnostic than positive WOM. Future

research should investigate situations

where negative WOM has a stronger

impact.

• The consumers in the current study

were familiar with the brands consid-

ered for the categories analyzed, and

the innovations that the companies had

launched in the market represented

incremental levels of novelty linked to

product re-launches.

It would be interesting to investigate

the impact of positive and negative

WOM in the case of radical innovations

(adoption of new products in their intro-

duction stage).

• The current study analyzed only the

final outcome of the WOM (shift in the

purchase probability). Another obvious

possibility for further research would

be to explore the impact of positive and

negative WOM on the intermediate

stages in the receiver’s decision-making

process (motivation and desires, search

for information, and attributes consid-

ered in the purchase).

• The current work focused mainly on the

influence of WOM. The study could be

expanded to compare the effects of posi-

tive and negative WOM with those of

other types of information obtained by

the receivers through different commu-

nications media (e.g., advertising, direct

marketing, and the firm’s sales force).

• The current work did not allow the

authors to establish whether the receiv-

ers obtain the information from face-

to-face relationships or through social

networks. Given the importance of elec-

tronic WOM, it would be useful to study

the dynamics of online WOM and prod-

uct sales.

The aim would be to study e-WOM

via consumer opinion platforms and

investigate what motivates senders and

receivers to articulate themselves on the

Internet.

RodoLFo vázquez-CasieLLes is a Professor of Marketing at

the University of Oviedo (Spain). His current research

interests include consumer behavior and perceptions

of justice in the services markets. Other research

interests include interfirm relationships in distribution

channels, retail management, brand equity, and market

orientation. His recent articles have been published

in the European Journal of Marketing, Industrial

Marketing Management, International Journal of

Research in Marketing, International Journal of

Service Industry Management, Journal of Applied

Social Psychology, Journal of Business Research,

Journal of Business and Industrial Marketing,

Marketing Letters, Supply Chain Management, Service

Industries Journal, and Psychology and Marketing,

among other journals.

LetiCia suáRez-áLvaRez is an Assistant Professor of

Marketing at the University of Oviedo (Spain). Her current

research interests include relationship marketing,

tourism marketing, consumer behavior, and service

recovery strategies in the service sector. She has

presented papers at various academic meetings and

conferences. Her work has appeared in the Journal

of Applied Social Psychology, Journal of Marketing

Management, Journal of Travel Research, Management

Research, Service Industries Journal, and Psychology

and Marketing, among other journals.

ana BeLén deL Río-Lanza is an Associate Professor in

Marketing at the University of Oviedo (Spain). Her current

research interest focuses on brand equity, consumer

behavior, service failure, service recovery strategies, and

perceptions of justice in the services markets. She has

presented papers at various academic meetings and

conferences. Her recent articles have been published

in the British Food Journal, Journal of Business

Research, Journal of Product and Brand Management,

Journal of Applied Social Psychology, and Marketing

Letters, among other journals.

ACknOWLEDGMEnTS

The authors thank the Spanish Ministry of Sci-

ence and Innovation for the financial support

provided for this research under the 2008–2011

Call for R&D Projects (MICINNECO2008-03698/

ECON).

It would be advisable in future work to subject the

receivers to actual positive and negative WOM

situations and record their actual purchase decision.



REFEREnCES

AhluwAliA, r. “How Prevalent Is the Negativ-

ity Effect in Consumer Environments?” Journal

of Consumer Research 29 (2002): 270–279.

AJZen, i. “The Theory of Planned Behavior.”

Organizational Behavior and Human Decision Pro-

cess 50, 2 (1991): 179–211.

Allsop, D. t., B. r. BAssett, and J. A. hosKins.

“Word-of-Mouth Research: Principles and

Applications.” Journal of Advertising Research 47,

4 (2007): 398–411.

AnDerson, e. w. “Customer Satisfaction and

Word-of-Mouth.” Journal of Service Research 1, 1

(1998): 5–17.

ArnDt, J. “Role of Product-Related Conversa-

tions in the Diffusion of a New Product.” Journal

of Marketing Research 1, 4 (1967): 291–295.

AssAel, h. Consumer Behaviour: A Strategic

Approach. Boston, MA: Houghton Mifflin Com-

pany, 2004.

BAnsAl, h., and p. A. Voyer. “Word-of-Mouth

Processes within a Services Purchase Decision

Context.” Journal of Service Research 3, 2 (2000):

166–177.

BAyus, B. “Word of Mouth: The Indirect Effects

of Marketing Efforts.” Journal of Advertising

Research, 25, (June/July) (1985): 31–39.

BergKVist, l., and J. r. rossiter. “The

Predictive Validity of Multiple-Item versus

Single-Item Measures of the Same Constructs.”

Journal of Marketing Research 44, May (2007):

175–184.

BergKVist, l., and J. r. rossiter. “Tailor-Made

Single-Item Measures of Doubly Concrete Con-

structs.” International Journal of Advertising 28, 4

(2009): 607–621.

BowmAn, D., and D. nArAyAnDAs. “Man-

aging Customer-Initiated Contacts with Man-

ufacturers: The Impact on Share of Category

Requirements and Word-of-Mouth Behaviour.”

Journal of Marketing Research 38 (2001): 281–297.

Brown, J. J., and p. h. reingen. “Social Ties

and Word-of-Mouth Referral Behaviour.” Jour-

nal of Consumer Research 14, 3 (1987): 350–362.

Brown, t. J., t. e. BArry, p. A. DACin, and r.

F. gunst. “Spreading the Word: Investigating

Antecedents of Consumers’ Positive Word-of-

Mouth Intentions and Behaviours in a Retailing

Context.” Journal of the Academy of Marketing Sci-

ence 33, 2 (2005): 123–138.

Bruyn, A., and g. lilien. “A Multi-Stage

Model of Word-of-Mouth Influence Through

Viral Marketing.” International Journal of

Research in Marketing 25, 3 (2008): 151–163.

CheVAlier, J. A., and D. mAyZlin. “The

Effect of Word-of-Mouth on Sales: Online Book

Reviews.” Journal of Marketing Research 43, 3

(2006): 345–354.

DholAKiA, u. m. “An Investigation of the

Relationship Between Perceived Risk and

Product Involvement.” In Advances in Con-

sumer Research, M. Brucks and D. MacInnis, eds.

Provo, UT: Association for Consumer Research,

1997.

DiCK, A., and K. BAsu. “Customer Loyalty:

Toward an Integrated Conceptual Framework.”

Journal of the Academy of Marketing Science 22

(1994): 99–113.

DuhAn, D. F., s. D. Johnson, J. B. wilCox,

and g. D. hArrell. “Influences on Consumer

Use of Word-of-Mouth Recommendation

Sources.” Journal of the Academy of Marketing Sci-

ence 25, 4 (1997): 283–295.

eAst, r., K. hAmmonD, and w. lomAx. “Meas-

uring the Impact of Positive and Negative

Word-of-Mouth on Brand Purchase Probabil-

ity.” International Journal of Research in Marketing

25 (2008): 215–224.

eAst, r., K. hAmmonD, and m. wright. “The

Relative Incidence of Positive and Negative

Word-of-Mouth: A Multi-Category Study.”

International Journal of Research in Marketing 24

(2007): 175–184.

FisKe, s. t. “Attention and Weight in Per-

son Perception: The Impact of Negative and

Extreme Behaviour.” Journal of Personality and

Social Psychology 38, 6 (1980): 889–906.

FitZsimons, g. J., and D. r. lehmAnn. “When

Unsolicited Advice Yields Contrary Responses.”

Marketing Science 23, 1 (2004): 82–95.

Fong, J., and s. Burton. “Electronic Word-of-

Mouth: A Comparison of Stated and Revealed

Behavior on Electronic Discussion Boards.”

Journal of Interactive Advertising 6, 2 (2006),

61–70.

FrenZen, J. K., and h. l. DAVis. “Purchasing

Behaviour in Embedded Markets.” Journal of

Consumer Research 17 (1990): 1–12.

FrenZen, J. K., and K. nAKAmoto. “Structure,

Cooperation and the Flow of Market Informa-

tion.” Journal of Consumer Research 20 (1993):

360–375.

gilly, m. C., J. l grAhAm, m. F. wolFin-

BArger, and l. J. yAle. “A Dyadic Study of

Interpersonal Information Search.” Journal of the

Academy of Marketing Science 26, 2 (1998): 83–100.

golDsmith, r., and D. horowitZ. “Measuring

Motivations for Online Opinion Seeking.” Jour-

nal of Interactive Advertising 6, 2 (2006): 1–16.

gounAris, s., and V. stAthAKopoulus. “Ante-

cedents and Consequences of Brand Loyalty:

An Empirical Study.” Journal of Brand Manage-

ment 11, 4 (2004): 283–306.



grAhAm, J., and w. hAVlenA. “Finding the

‘Missing Link’: Advertising’s Impact on Word

of Mouth, Web Searches and Site Visits.” Journal

of Advertising Research 47, 4 (2007): 427–435.

gruen, t. w., t. osmonBeKoV, and A. J.

CZAplewsKi. “eWOM: The Impact of Customer-

to-Customer Online Know-How Exchange on

Customer Value and Loyalty.” Journal of Busi-

ness Research 59, 4 (2006): 449–456.

guo, C. “A Review on Consumer External

Search: Amount and Determinants.” Journal of

Business and Psychology 15, 3 (2001): 505–519.

hArrison-wAlKer, l. J. “The Measurement of

Word-of-Mouth Communication and Investi-

gation of Service Quality and Customer Com-

mitment as Potential Antecedents.” Journal of

Service Research 4, 1 (2001): 60–75.

heitmAnn, m., D. r. lehmAnn, and A. her-

mAnn. “Choice Goal Attainment and Decisions

and Consumption Satisfaction.” Journal of Mar-

keting Research 44 (2007): 234–250.

henning-thurAu, t., K. p. gwinner, g.

wAlsh, and D. D. gremler. “Understanding

Relationship Marketing Outcome: An Integra-

tion of Relational Benefits and Relationship

Marketing.” Journal of Service Research 4, 3

(2002): 230–247.

hogAn, J., K. lemon, and B. liBAi. “Quantify-

ing the Ripple: Word-of-Mouth and Advertising

Effectiveness.” Journal of Advertising Research 44,

3 (2004): 271–280.

hung, K. h., and s. y. li. “The Influence of

eWOM on Virtual Consumer Communities:

Social Capital, Consumer Learning, and Behav-

ioral Outcomes.” Journal of Advertising Research

47, 4 (2007): 485–495.

Keller, e. “Unleashing the Power of Word of

Mouth: Creating Brand Advocacy to Drive

Growth.“ Journal of Advertising Research 47, 4

(2007): 448–452.

lACey, r., J. suh, and r. m. morgAn. “Differen-

tial Effects of Preferential Treatment Levels on

Relational Outcomes.” Journal of Service Research

9, 3 (2007): 241–256.

mAtos, C. A., and A. VArgAs. “Word-of-

Mouth Communications in Marketing. A

Meta-Analytic Review of the Antecedents and

Moderators.” Journal of the Academy of Marketing

Science 36 (2008): 578–596.

mAxhAm, J. g., and r. g. netemeyer. “A

Longitudinal Study of Complaining Custom-

ers’ Evaluations of Multiple Service Failures

and Recovery Efforts.” Journal of Marketing 66,

4 (2002): 57–71.

mishrA, s., u. n. umesh, and D. e. stern.

“Antecedents of the Attraction Effect: An

Information-Processing Approach.” Journal of

Marketing Research 30 (1993): 331–349.

murrAy, K. B., and J. l. sChlACter. “The

Impact of Services Versus Goods on Consum-

ers’ Assessment of Perceived Risk and Variabil-

ity.” Journal of the Academy of Marketing Science

18, 1 (1990): 51–65.

netemeyer, g. r., and w. o. BeArDen. “A

Comparative Analysis of Two Models of Behav-

ioural Intention.” Journal of the Academy of Mar-

keting Science 20, 1 (1992): 49–59.

nielsen. “Trust, Value and Engagement in

Advertising.” Nielsen Global Online Consumer

Survey, July, 2009. Retrieved March 14, 2011

from http://blog.nielsen.com/nielsenwire/

wp-content/uploads/2009/07/trustinadvertis-

ing0709.pdf.

rAnAweerA, C., and J. prABhu. “On the Relative

Importance of Customer Satisfaction and Trust

as Determinants of Customer Retention and

Positive Word-of-Mouth.” Journal of Targeting,

Measurement and Analysis of Marketing 12, 1

(2003): 82–90.

siChtmAnn, C. “An Analysis of Antecedents

and Consequences of Trust in a Corporate

Brand.” European Journal of Marketing 41, 9/10

(2007): 999–1015.

smith, t. J. Coyle, e. lightFoot, and A. sCott.

“Reconsidering Models of Influence: The Rela-

tionship between Consumer Social Networks

and Word-of-Mouth Effectiveness.” Journal of

Advertising Research 47, 4 (2007): 387–397.

sweeney, J. C., g. n. soutAr, and t. mAZ-

ZArol. “Factors Influencing Word-of-Mouth

Effectiveness: Receiver Perspectives.” European

Journal of Marketing 42, 3/4 (2008): 344–364.

wAngenheim, F. “Postswitching Negative

Word-of-Mouth.” Journal of Service Research 8, 1

(2005): 67–78.

wAngenheim, F., and t. BAyon. “Satisfac-

tion, Loyalty and Word-of-Mouth within the

Customer Base of a Utility Provider: Differ-

ences Between Stayers, Switchers and Referral

Switchers.” Journal of Consumer Behaviour 3, 3

(2004a): 211–220.

wAngenheim, F., and t. BAyon. “The Effect

of Word-of-Mouth on Service Switching: Meas-

urement and Moderating Variables.” European

Journal of Marketing 38, 9/10 (2004b): 1173–1185.

wAngenheim, F., and t. BAyon. “The Chain

from Customer Satisfaction Via Word-of-Mouth

Referrals to New Customer Acquisition.” Jour-

nal of the Academy of Marketing Science 35, 2

(2007): 233–249.

westBrooK, r. A. “The Dimensionality of

Consumption Emotion Patterns and Consumer

Satisfaction.” Journal of Consumer Research 18, 1

(1987): 84–91.



APPENDIXResults of Measurement Model: Psychometric Properties of the Scales

ConstructsStandardised loading (l)*

Indicate your level of agreement with the following statements (Likert: 1 = total disagreement, 7 = total agreement)

1. Receiver’s experience (AvE = 0.972; CR = 0.950; a = 0.955). Similar items can be found in: Bansal and voyer (2000)I know this product/service category very well

I am competent and capable in things concerning this product/service category

I am very familiar with the current features of this product/service category

I am very experienced in the purchase of this product/service category

I think I have enough information about this product/service category

0.880

0.910

0.923

0.847

0.889

2. How actively WOM is sought (AvE = 0.704; CR = 0.877; a = 0.906). Similar items can be found in: Bansal and voyer (2000)Before buying I seek advice from people unconnected to the firm

Probability of accepting advice from other people (1 = very low; 7 = very high)

Explicit requirement of opinion of other people to take decisions (1 = very low; 7 = very high)

0.850

0.821

0.846

3. Receiver’s perceived risk (AvE = 0.767; CR = 0.908; a = 0.901). Similar items can be found in: Bansal and voyer (2000) and Wangenheim

and Bayon (2004b, 2007)Thinking about buying this product worries me because of the possibility of taking a risk

I think it would be a mistake if I didn’t seek the opinions of other people unconnected to the firm to avoid risks

I feel that buying this product is risky and I can avoid these risks if I seek advice from other people unconnected to the firm

0.817

0.870

0.936

4. Tie strength (AvE = 0.846; CR = 0.975; a = 0.975). Similar items can be found in: Bansal and voyer (2000) and Wangenheim and Bayon

(2004b, 2007)I speak to this person frequently

I have a trusting relationship with this person

This person understands me, shares my concerns, supports me

I have a strong personal relation with this person

I have interests and pastimes in common with this person

This person has social values and lifestyle like mine

I think that this person generally behaves very similarly to my way of seeing life

0.874

0.926

0.943

0.952

0.902

0.914

0.924

5. Sender’s experience (AvE = 0.846; CR = 0.956; a = 0.950). Similar items can be found in: Bansal and voyer (2000) and Wangenheim and

Bayon (2004b, 2007)This person is very knowledgeable about this product/service category

This person is very competent in things concerning this product/service category

This person has previously bought this product/service category

This person is very experienced in this product/service category

0.924

0.930

0.914

0.910

Other scales developed for this study

6. Sender’s strength of expression. Similar items can be found in: East et al. (2007 and 2008)

How convincing/credible do you think were the explanations of the person who gave you this recommendation? (1 = not at all convincing/credible;

7 = very convincing/credible)

7. Probability of choosing brand.** Similar items can be found in: East et al. (2007 and 2008)

What was the probability that you would choose this brand before receiving the recommendation? (receiver’s loyalty) (0 = zero; 10 = very high)

What was the probability that you would choose this brand after receiving the recommendation? (0 = zero; 10 = very high)

* All standardised loadings are significant (p < 0.01); ** The shift in the WOM receiver’s brand-purchase probability is calculated as the difference between the probability of choosing the brand after and before the recommendation

Copyright of Journal of Advertising Research is the property of Warc LTD and its content may not be copied or

emailed to multiple sites or posted to a listserv without the copyright holder's express written permission.

However, users may print, download, or email articles for individual use.

Date post:	03-Jan-2017
Category:	Documents
Upload:	hoangdieu
View:	216 times
Download:	2 times

The Word of Mouth Dynamic: How Positive (and Negative) WOM ...

Documents