When Ford was trying to boost the image of the Fiesta, its European-style compact car, in the US, it signed up 100 consumers to use the vehicles on the condition that they shared their experiences through the YouTube, Flickr, Facebook and Twitter accounts Ford created for the campaign.
Within six months they had posted 60,000 items, garnered 4.3m YouTube views and, according to Ford, generated 50,000 sales-leads and 35,000 test-drives. All for a fraction of the cost of a traditional advertising campaign.
Stories like this have multiplied in recent years as marketers fall in love with the dream of word-of-mouth marketing. Why pay media owners vast sums of money to send a message if you can get customers to do it for free, or at least at a low cost?
The great marketing educator Philip Kotler wrote in his textbook, Marketing Management: 'Word-of-mouth communication from an immediate and trusted source is typically more influential than media coverage from a remote and trusted source, despite the prestige of the latter.
'This should normally incline the marketer to favour personal influence channels over mass media ... the mass communicator may accomplish his message dissemination more efficiently by using a lower advertising budget and directing it specifically at opinion leaders.'
Given the attractiveness of this proposition, and the recent explosion of fresh means of magnifying word of mouth via various forms of social media, you would have thought that it would be top dog in the marketing mix by now. Yet, despite many compelling case studies like Ford's, it has not triggered far-reaching strategy shifts.
Membership of organisations such as WOMMA, the UK's word-of-mouth marketing association, has levelled out. So what's going on? Could it be, as a popular T-shirt in Silicon Valley suggested: 'Psssst. Word-of-mouth marketing doesn't work. Pass it on.' Here are some reasons for caution. First, that Kotler quote is almost 40 years old, from his 1972 textbook. Marketers have been saying the same things about word of mouth for decades.
The trouble is, as Kotler pointed out, that 'unfortunately the marketer has little direct control over social influence channels'.
That is supposed to have changed with the era of social media, where everything is so much more measurable and trackable. But has it? Two years ago, Duncan Watts, principal research scientist at Yahoo! Research, decided to track 39m Twitter chains started by 1.6m users. He wanted to discover why some tweets 'go viral' and others don't, but his main finding was that 98% didn't cascade at all. We notice the ones that do. They become 'case studies', but they represent a vanishingly small proportion of the population as a whole.
Watts also tested the claim that you can set a viral hare running by seeding it with carefully targeted 'influentials' - credible, confident, natural networkers. He recruited 61,000 people via a website and asked them to send a message to 18 targets worldwide - a bigger repeat of Stanley Milgram's original work that identified 'six degrees of separation'.
The study confirmed Milgram's original finding. It also found something else, however: even though this group had its own cohort of 'superconnectors' with big networks, still, 95% of the messages did not pass through them. Watts said: 'Influentials don't govern person-to-person communication. We all do.'
Last year, his findings were backed by research from the University of California that tried to identify the nature of influential users in internet social networks. None of a long list of attributes - whether they were friends, descriptors from their profiles, gender, and so on - could be used to predict who was going to be influential at any point in time.
Watts went on to design a computer model to test how messages pass through big networks populated by nodes with varying numbers of connections. Many trends flowed through the network, but the vast majority were created by average Joes.
Even when Watts gave the 'influentials' 40 times more networking power than the average, the same result applied. The critical factor was not the networking power of superconnectors, but how willing people were to be influenced. 'If society is ready to embrace a trend, almost anyone can start one - and if it isn't, then almost no one can,' said Watts.
Another constraint on word-of-mouth marketing is the brands and categories people talk about. Research by Keller Fay shows that people talk far more about some categories, such as food, dining and media, than others, such as household products (see graph, below left).
Furthermore, research by Robert East, professor of strategy, marketing and entrepreneurship at Kingston Business School, suggests that rates of word-of-mouth recommendation are correlated with brand share: people talk more about bigger brands, because more people own them.
Another way to understand word of mouth is via sociology and anthropology. Paul Adams, formerly of Google and now at Facebook, is a believer in the power of social networks. 'For many years, we have considered people as isolated independent actors,' he wrote in his blog. 'We now know that this isn't true. Peoples' networks influence almost every aspect of their lives. What they do. Where they go. What brands they prefer. What products they buy.'
He added, however, that even in a world where some have hundreds of Facebook 'friends', 'people tend to have a trusted 'inner circle' of fewer than five people. 'People turn to their closest friends for advice.' This is apparently supported by Fay, whose latest 'TalkTrack' word-of-mouth research reports that 81% of it takes place face-to-face, with another 10% by phone. Only 9% takes place online, including email, texting and social networking.
Also, the 9% that does take place online can be highly unrepresentative, warns Richard Anson, chief executive of the peerreview site Reevoo. It takes the form of a 'u' shape, dominated by extreme detractors or advocates. The real curve tends to be 'n'shaped, with most people somewhere in the middle, but it is harder to get them to express their opinions. Perhaps that's why research by Lightspeed for Marketing two years ago found forums and social networking to be the least-trusted sources of advice (see bar chart, left).
Does this mean word-of-mouth marketing is a dead duck? Not necessarily. First, word of mouth is how brands actually grow strong: we call it reputation. This is not created by a separate activity called 'word-of-mouth marketing', however. It's a by-product of doing everything else right: delivering excellent quality and value, providing good customer service, treating customers with respect, and so on.
Second, as word-of-mouth marketing practitioners such as 1000Heads point out, there are still many things that brands can do to listen to what people are saying and to get them talking. A word in your ear, though. Don't expect it to solve all your marketing problems for you.
Alan Mitchell is a respected author and a founder of Ctrl-Shift and Mydex.
Not everything you hear about word of mouth is true
- Most messages pass through 'average Joes' not 'superconnectors' or 'influentials'.
- The vast majority of recommendations are made offline, not online - and they are usually made by close friends.
- Degree of influence has more to do with 'willingness to be influenced' than the influencer.
- Word of mouth is closely related to brand share: people talk more about big brands than little ones.
- Online recommendations are 'u'-shaped - mostly extreme detractors or advocates. The real recommendation curve is 'n'-shaped, with most people somewhere in the middle.