Among the many factors swaying consumer behaviour, word-of-mouth (WOM) ranks at the top. Many studies show that the number one source of information shaping purchasing behaviour is what others think about brands. As a result, positive and negative word-of-mouth are critical to the success or downfall of products and services. Industry research shows that a seven percent increase in positive word-of-mouth messages can increase a company’s revenues by as much as one percent. Conversely, another study shows that an increase of 1,000 word-of-mouth complaints can cost the airline industry an accumulated loss of as much as $8.1 billion over 20 months.
Particularly important for marketing executives is to find who, in their customers’ social networks, is most effective at spreading positive word-of-mouth about their brands and products. Given friends’ well-known tremendous influence on what consumers think, feel and buy, aiming to encourage word-of-mouth with friends seems a good idea. For instance, friends impact a wide range of individual preferences ranging from smoking habits to eating behaviour and even political preferences. As such, many marketing textbooks or advisors recommend to target customers’ close connections and encourage consumers to spread the word about their products and services among friends. This perspective assumes that one’s friends are disposed to speak favourably about products and services. But is it always the case?
My recent paper, Sharing with Friends versus Strangers: How Interpersonal Closeness Influences Word-of-Mouth answers this critical question by looking at the extent to which the content customers share with a message recipient is swayed by how close they feel to him or her. In particular we look at whether varying closeness between a WOM sender and a recipient changes the amount of positive and negative information communicated. For instance, do people share the same amount of positive and negative information when sharing with a friend or an acquaintance?
Not quite, it turns out. In fact, customers demonstrate a systematic tendency to share negative information about a product or a service to someone with whom they feel a strong bond, and more positive information to someone with whom they feel a weak bond. Therefore, surprisingly, while positive information tends to spread more easily across weakly tied connections, negative word-of-mouth tends to spread among close ties, making them potentially dangerous for marketing executives. This distortion emerges in both online and face-to-face interactions across contexts and product categories.
Why does this happen? The rationale stems from the activation of different motivations when addressing close versus distant others. Talking to a close friend activates a motive to protect him or her from any harm or hardship, making message senders more likely to share negative information. However, talking to someone distant activates a motive to self-enhance, making message senders more likely to share positive information to build their image and avoid being perceived as someone who complains.
Digital connections and the sharing of word-of-mouth
In social networks, feelings of closeness vary both within a platform (e.g., a customer might be connected with both friends and acquaintances) and between platforms (e.g., people tend to use platforms like Facebook for friends, and platforms like LinkedIn for more distant acquaintances).
In another study, we asked managers to share a message featuring the pros and cons of a service on LinkedIn with either a close friend or a distant acquaintance and found that they shared more cons than pros with their close contacts but more pros than cons with their distant ones. In fact, we found that the number of overlapping connections positively predicted message valence: the more friends in common between a sender and a recipient, the more negative she tended to be.
Building on this study, we asked young adults to share a message featuring pros and cons of a product to someone they knew on LinkedIn vs. Facebook. Within this consumer group, Facebook is typically used to foster and maintain personal connections whereas LinkedIn is used for professional connections. We found that participants, overall, included a greater number of pros than cons when sharing on LinkedIn but the reverse pattern when sharing on Facebook – confirming our hypothesis.
When does this effect occur?
We also investigated when the effect seemed stronger versus weaker. To do so, we varied in one study whether participants talked about either a novel or an established product. Why?
New product offers give consumers particularly rich grounds to build their image because they make them look more interesting, smart and in-the-know. At the same time, talking about the same new products also gives consumers particularly good reasons to protect others, because new products are typically associated with more potential drawbacks. As a result, one might expect the effect to amplify for new products with people sharing even more positive information with acquaintances but more negative information with friends. However, the effect should weaken when the product or service talked about is well-established and thus provides fewer reasons to self-enhance or to protect others.
Confirming our intuition, talking about a novel product produced both more positive word-of-mouth with an acquaintance and more negative word-of-mouth with a friend. However, the effect disappeared when the product was well-established. Overall, highlighting a product’s novelty amplified the effects of relationship strength, prompting participants to share more positive information with distant others, but more negative information with those close to them.
Shifting feelings of closeness in word-of-mouth contexts
Our findings hold a few key implications for brand managers. First, brands can encourage the sharing of positive (or negative) word-of-mouth messages by adjusting the context of their communication to trigger different feelings of either closeness or distance depending on the objective. This could mean displaying a photo of a potential recipient to make a customer relate or highlighting dissimilarities between the sender and the recipient.
In languages that have different pronouns to address someone close versus someone distant such as French, Spanish or Italian, varying pronouns to enforce closeness or distance with a customer can be a powerful vector to influence word-of-mouth valence. For example, in French, tu is used when addressing close others, but vous, when addressing distant others. In fact, in another experiment part of this research program, we manipulated whether people shared a message using the T-form or the V-form and found they were more negative in the former case than the latter. This is a small change that could help brands tune the sharing of information in the desired direction.
Greater attention should also be paid to the type of platform on which the organisation shares the information. LinkedIn may be a better bet for starting a campaign related to professionals whose ambition to self-enhance will help their message. For new products or innovations, extra care should be taken, as customers amplify negative information to their close groups, such as on Facebook. If marketers want to leverage key influencers such as bloggers, it would also serve them well to take a look at the size of their audience and closeness of their relationships.
Finally, our research speaks to the potential dangers of advising a company to leverage customers’ friends to spread ideas or products. This might not be the most effective way to build positive momentum; in fact, in some contexts, it might just do the opposite.
David Dubois is Assistant Professor of Marketing at INSEAD and a Programme Director of Leading Digital Marketing Strategy, one of INSEAD’s executive development programmes.
-
View Comments
(3) -
Anonymous User
I wonder: Are marketing teams responsible for discovering key influencers taking into account not only absolute reach but also the average size of the clusters / communities in which influencers are active? For example, I may be active in 20 LinkedIn groups of 30 members each, but more likely to generate negative WOM relative to someone active in 1 LinkedIn group with 200 members, despite having a larger absolute reach...
Krishna Kumar Silvery
Wow, i wasn't really expecting it.
-
Leave a Comment
Anonymous User
05/04/2016, 04.10 pm
Hello David,
This comes as a total surprise and how this insight can lead to measure to amplifying the positive and curtailing the impact of negative WOM.