You are on page 1of 5

Effect of Word of Mouth (WOM) communications of consumer behaviour - essay rubric

Word of mouth (WOM) involves the communication among consumers about products and services (Liu, 2006).
It relates to face-to-face communication as well as online conversations and user-generated content. Research
has long established the effect of WOM on consumer behaviour and purchase decisions (Brooks, 1957; Dichter,
1996); more recently however, and following the evolution of the Internet, the phenomenon has attracted
increasing attention from marketers. Through the Internet, consumers are able to engage in new forms of WOM,
and overall they are more exposed to WOM messages. Accordingly, this has affected and possibly amplified the
impact of WOM on consumer behaviour. Online WOM is also easier to track than offline communication. It
provides an unprecedented amount of data on the dynamics and effects of WOM. Marketers are looking to take
advantage of online channels and new insights to develop better WOM strategies; WOM marketing campaigns
especially, appeal as a low-cost method to generate favourable buzz and drive customer acquisition. Yet,
nonetheless, WOM remains a little understood phenomenon. Research is just starting to make use of Internet
data; and, especially, academic research is rarely translated into managerial insights.

Research on WOM is modelled on two approaches. A first approach explores the patterns of generation and
dissemination of WOM. This is done either by inferring a model of the information flow (Coleman et al. 1966), or
alternatively by investigating the psychology of WOM senders and receivers (Anderson, 1998; Dichter, 1966). A
second approach draws from the findings on WOM diffusion, but focuses on its impact on consumer behaviour.
Research of this kind usually specifies one or more features of WOM as variables, and then assesses the
influence of such variables on consumer behaviour. The present essay rubric focuses on the latter approach. It
intends to provide a rough but integrated overview on research regarding the dynamics between WOM and
consumer behaviour. In this way, it might contribute as a framework to understand the scope of WOM studies
and invite further investigations.

To begin with, research has often concluded that WOM volume has a positive correlation with consumer
behaviour; in other words, more WOM within the consumer base leads to increased customer acquisition or
sales (Anderson and Salisbury, 2003; Bowman and Narayandas, 2001; Van den Bulte and Lilien, 2001). This
research tends to rely on the method of counting WOM instances, literally, counting mentions to the product in
question. However, mere counting would not account for WOM messages exchanged repeatedly within the
same community. WOM volume can be compared to frequency in advertising: high levels of frequency might
not necessarily spread awareness, and thus increase sales, if the advert is delivered repeated to the same
individuals. Godes and Mayzlin (2004) thus introduce WOM dispersion as an additional variable that better
explains the effects of WOM on consumer behaviour. They claim that WOM disseminates quickly within a close
community, recording a high volume; but reaches a higher number of consumers when spreading across a
diverse set of communities. WOM dispersion is therefore compared to reach in advertising, and considered a
better measure of WOM impact on consumer behaviour. Their empirical study uses data from online
conversations on a sample of TV shows on Usenet groups (community forum), then compared to their TV
ratings. Data is collected throughout the TV show season and then divided between an early and a late period.
WOM volume is derived from the number of posts; WOM dispersion is derived from the entropy of

  1
conversations across groups. Findings show that there is a high correlation between early WOM dispersion and
high future ratings; correlation between volume and future ratings is not consistent. This leads to the conclusion
that WOM dispersion has a higher forecasting power than WOM volume has. Results offer valuable insights to
marketers, yet are subjected to a key limitation. The study is based on the underlying assumption that
awareness about the product is a catalyst to the purchasing process; in this context, as more consumers are
aware of the TV show, TV ratings are likely to be higher. Although findings seem to justify this assumption, this
might relate to the specific nature of the product and the network at hand. First, TV shows are a low-risk
consumption. The product is obtained at no cost; although brand differentiation is high, consumers can switch
from one TV show to another at no extra cost or effort. Therefore, the decision to watch a certain TV show is
driven mainly by behavioural factors, and often even finalised after trial of another show. Second, within the
network considered consumers are anonymous and therefore equal: any consumer has the same influence onto
any other, regardless of their demographic features or level of expertise. This type of network is likely to
generate WOM with low persuasion potential: consumers will become aware of the product but might not be
otherwise affected by WOM sources that they do not know or trust. In this scenario, awareness is more closely
associated with consumption than for other products or networks. Features of WOM that increase levels of
awareness, such as WOM dispersion, retain a high forecasting power; while WOM does not have any actual
persuasive influence on consumers. In other cases, however, WOM might be taken into account by consumers
at the evaluation phase of the purchasing process; the contents of WOM would therefore have a persuasive
influence on their decision-making.

To distinguish between WOM influence at various stages, De Bruyn and Lilien (2008) adopt a multi-stage model
of WOM analysis. This model allows to discriminate between WOM that spreads awareness and WOM that
affects consumers' decisions. They argue that the WOM enacts an informative or a persuasive potential
depending on the social ties between WOM sender and receiver. They consider four variables to define the
nature of social ties, that are, tie strength, perceptual affinity, demographic dis/similarity and source expertise.
They then observe how each variable affects consumer awareness, interest and ultimately decision to consume.
The empirical study uses data collected from an Internet-based viral marketing experiment, which involves an
unsolicited email suggesting to complete a survey and to then recommend it to someone else. The phases of
the experiment (e-mail is opened, link is clicked, then survey is completed) reflect the phases of the purchasing
process. Findings show that tie strength between sender and receiver generates awareness, but is not related
to ultimate decision-making; also, perceptual affinity influences interest in the product, but again not decision-
making. Demographic dissimilarity is the only variable that causes WOM to have an effect on all awareness,
interest and purchase. The study, therefore, shows that the relationship between the WOM source and the
receiver determines whether WOM merely drives awareness or performs a persuasive power; this points to
limitations in other research, such as Godes and Mayzlin’s (2004), that is based on a network that only enables
the awareness factor. If De Bruyn and Lilien’s study (2008) highlights an interesting distinction between types of
WOM impact, it by no means provide the only explanation of WOM differential effects. For instance, the study
does not find relevance for the variable of source expertise, arguably due to the mere fact that the product
considered did not required any expert information. For other categories, WOM source expertise might enhance
WOM persuasion. Research on the role of opinion leaders (Richins and Root-Shaffer, 1988; Katz and

  2
Lazarsfeld, 1955) has long debated this effect.

In light of WOM two-fold potential, other studies have turned to the persuasion features of WOM, and in
particular focused on the effects of WOM content on consumer behaviour. Findings are here more
contradictory. Liu (2006) argues that WOM valence, so the proportion of positive or negative information about
the product, is not related to changes in consumer behaviour over a period of time. Basing his argument on
WOM on films found on online message boards and the box office sales, he claims that WOM volume remains
the most accurate variable to forecast future sales. Yet one can argue that Liu chooses a scenario where WOM
does not have a high persuasive potential; similarly to Godes and Mayzlin (2004), the opinions expressed on
online message boards might not define consumers’ ultimate purchasing decision. The empirical study would
then reflect patterns of awareness, whereas it would not shed light on patterns of persuasion and preference.

Chevalier and Mayzlin (2006) dispute Liu’s results. They support the notion that WOM valence has a differential
impact on consumer behaviour, and additionally they claim that the negative effect of negative WOM is greater
than the positive effect of positive WOM. Their empirical study focuses specifically on WOM that has a high
persuasion potential, that is, online reviews: online reviews are not related to awareness as they are specifically
sought by consumers to make a purchasing decision. Data is gathered from the websites of two booksellers at
three points in time; user star ratings are used to determine the positive or negative character of WOM. Findings
show that average star rating has a direct impact on book sales; additionally when ratings are not flattened by
they mean value, the relative effect of one-star reviews is significantly higher than the relative effect of five-star
reviews. This proves that, in this scenario, WOM content does enact a persuasion power that affects consumer
behaviour, and that negative WOM has a greater weight in decision-making than positive WOM. Arguably, the
research is not fully safe from bias. For instance, it does not take into account a number of factors that could
influence the audience perception of the reviews, such as the contents of the text associated with the star rating
or the order of presentation. This might lead to underestimating WOM messages that, for features other than
valence, had a strong impact on the consumers. Moreover, the study is centred on a one-off purchase, and
thus is not able to account for changes in WOM persuasive impact over time. Godes and Mayzlin (2004) as well
as Liu (2006) show that WOM has a greater impact on consumer behaviour early in the product lifecycle; the
correlation decreases when the product becomes more common or has been once adopted by the consumer.
As earlier shown, however, these studies only consider WOM as a driver of awareness, which obviously
becomes less valid as more consumers know about the product. On the opposite, WOM persuasion potential
should not decline in a similar way. Arguably, WOM valence should still have an impact on consumers later in
the product lifecycle or after initial adoption.

East et al. (2008) attempt to tackle this issue. They do not specifically mention phases of the product lifecycle,
but introduce pre-purchase probability as a factor that can moderate the effects of WOM valence. In other
words they take into account the fact that, unless the category is completely unprecedented on the market,
consumers will possess a positive or negative inclination towards the product prior to receiving the WOM
message. This inclination will necessarily change the extent to which WOM can affect their behaviour. The
empirical study uses data gathered through a survey, so contrary to the research presented so far it did not use
Internet-based data. On the questionnaires, subjects reported instances of WOM across a wide range of

  3
product categories, and also specified WOM valence and intensity, their attitude towards the brand, as well as a
number of other factors. Findings show that pre-purchase probability does in fact affect how WOM valence is
received. When subjects were not committed to an alternative brand, and thus more likely to purchase the
product, positive WOM had a greater impact on their behaviour. Disputing Chevalier and Mayzlin’s study, in this
context positive WOM would be more likely to lead to purchase than negative WOM to discourage it. East et al.
are able to overcome many limitations highlighted in previous studies as the survey covers a wide range of
products and variables. Findings are more easily applicable to several contexts; yet they remain liable of the
sources of error intrinsic to self-reporting. In other words, the subjects’ perception and report of their behaviour
does not necessarily reflect their actions. Also, although the study does observe a correlation between WOM
intensity and consumer behaviour, its result are still expressed through a binary distinction between positive and
negative WOM. Additional studies advance the hypothesis that discourse dynamics within WOM messages
have a more variegated impact on consumer behaviour. Kozinets et al. (2010), for instance, show that the WOM
messages are appropriated within the narrative of the WOM source, and even though this narrative does not
necessarily changing WOM valence, it might affect the receiver’s perception. In this regard, research is still
developing; further research that integrates discourse analysis with data on consumer behaviour is required to
validate this hypothesis.

Overall, the present essay rubric has presented a distinction between two roles of WOM in the purchasing
process: WOM as a driver of awareness, measured by volume and dispersion, and WOM as a persuasive
power, dependent on the content of the messages. Studies here included link the two to consumer behaviour
and therefore provide valuable insights to marketers looking to manage WOM or design a WOM marketing
campaign. Yet, the comparison between studies also shows that, depending on the nature of network, product
and consumer base, the effectiveness of these two aspects of WOM varies. Marketers are therefore required to
evaluate their specific circumstances and objectives before relying on the findings.

It must be here noted that the essay rubric falls short of accounting for the potential differences between organic
WOM and WOM marketing initiatives. The dynamics and effects of WOM when instigated by the company
might not necessarily reflect this of naturally occurring WOM, as consumers might perceive WOM messages as
less trustworthy or valuable. This might halt WOM diffusion through the network; or alternatively, it might erase
WOM persuasion potential and even reverse the intended effects. Kozinets et al. (2010) point out that with
WOM campaigns marketers are often intervening on pre-existing social dynamics, and therefore the resulting
WOM messages can only form narratives that are coproduced by both firm and consumers. Marketers must be
able to cultivate the company’s presence in the consumers’ narrative, so that it complies with the consumers’
social practises. In light of the coproduction framework, further research might compare different approaches
that firms can take to interact with their consumer base. Findings of this kind would prove extremely informative
both for WOM seeding programmes and social media management.

References:

Anderson, E.W. (1998). Customer satisfaction and word of mouth. J. Service Res., 1, 5-17.

  4
Anderson, E. and L.C. Salisbury (2003). The formation of market-level expectations and its covariates. Journal of Consumer
Research, 30, 115-124.

Bowman, D. and D. Narayandas (2001). Managing customer-initiated contacts with manufacturers: the impact of share on
category requirements and word-of-mouth behaviour. Journal of Marketing Research, 38, 281-297.

Brooks, R.C. (1957). 'Word of Mouth' Advertising in Selling New Products. Journal of Marketing, 22, 154-61.

de Bruyn, A. and G. L. Lilien (2008). A multi-stage model of word-of-mouth influence through viral marketing. International
Journal of Research in Marketing, 25, 151-163.

Chevalier, J. and D. Mayzlin (2006). The Effect of Word of Mouth on Sales: Online Book Reviews. Journal of Marketing
Research, 43, 345-354.

Coleman, J.S., E. Katz and H. Menzel (1966). Medical Innovation: a Diffusion Study. Indianapolis: Bobbs-Merrill.

Dichter, E. (1966). How word-of-mouth marketing works. Harvard Business Review, 44, 148.

East, R., K. Hammond and W. Lomax (2008). Measuring the impact of positive and negative word of mouth on brand
purchase probability. International Journal of Research in Marketing, 25, 215-224.

Godes, D. and D. Mayzlin (2004). Using Online Conversations to Study Word-of-Mouth Communication. Marketing Science,
23, 545-560.

Katz, E. and P. Lazarsfeld. (1955). Personal Influence. Glencoe (IL): Free Press.

Kozinets, R., K. de Valck, A.C. Wojnicki and S.J.S. Wilner (2010). Networked Narratives: Understanding Word-of-Mouth
Marketing in Online Communities. Journal of Marketing, 74, 71-89.

Liu, Y. (2006). Word of Mouth for Movies: Its Dynamics and Impact on Box Office Revenue. Journal of Marketing, 70, 74-89.

Richins, M. and T. Root-Shaffer (1988) ,The Role of Evolvement and Opinion Leadership in Consumer Word-Of-Mouth: an
Implicit Model Made Explicit, NA - Advances in Consumer Research,15, 32-36.

Van den Bulte, C. and G. Lilien (2003). Two-stage partial observability models of innovation adoption. Working paper,
Wharton School of Business, Philadelphia.

  5

You might also like