But now a long-taken-for-granted central principle of viral marketing -- that large-scale changes in behavior can begin like disease epidemics, with just a few highly connected people -- is facing its toughest challenge yet. At the center of a growing fray is an unlikely figure: an Australian-born sociology professor at Columbia University named Duncan Watts, who comes armed with mathematical models that, he believes, unsettle much of what you think you know about viral marketing.
For a few years now, Mr. Watts looked at how big cascades of behavior begin, focusing most recently on the theory of influencers, those people disproportionately capable of triggering big change and featured in books such as 2003's "The Influentials." What he's found, through computer modeling as opposed to real-world research, is that they're not necessarily all that effective.
The crux of Mr. Watts' argument is that even if influentials are several times as influential as a normal person, they have little impact beyond their own immediate neighborhood -- not good when you're trying to create a cascade through a large network of people, as most big brands do. In those cases, he argues, it's best to skip the idea of targeting that treasured select group of plugged-in folks and instead think about that group's polar opposite: a large number of easily influenced people. He calls this big-seed marketing.
Sounds a lot like mass marketing, doesn't it?
"Most of what's happening is happening in the network," Mr. Watts said. "If the network is conducive to contagion, then it doesn't matter where you put a match to it. Ordinary people can do much better than you think."
With the emergence of easily-observed online networks like MySpace and its precious scaled and engaged audience, this kind of thinking takes on a new urgency. But Mr. Watts disagrees with any suggestion that new social media suggests more complexity than what goes on in the actual world. "The difference is that you can see it. People are starting to realize how complex the social world is. Read something like Malcolm Gladwell's book. He makes it sound so easy."
In reality, tipping -- experiencing that exponential growth -- is very difficult. Just ask Jonah Staw, the founder of Little Miss Matched, a three-year-old company that sells mismatched socks to tweens. Little Miss Matched has tripled in size in each of those years, but it hasn't seen the exponential growth its founders expected. He freely admits that, despite his company's growth and some clever promotional ploys, it has not tipped.
"I know that Malcolm Gladwell would tell me I haven't figured out the right thing to do yet, but each thing you do takes a lot of resources -- of staff, time, and funds," he said.
Though Mr. Staw isn't familiar with Mr. Watts' body of work, he inadvertently described one of its conclusions that emerged from it. An irony of our age is that, though everyone acknowledges consumers are in control, marketers still believe they're running the show, right down to trying to plan for virality as any creative told to "just go make a viral video" will lament. Virality is an outcome, not a channel to be planned. In Mr. Watts' chaotic conception of the world, you might as well try to plan for a terrorist attack or some other random event.
"We cannot predict what is going to happen," he said. "Things happen randomly. You want strategies that don't depend on being right, but do depend on being able to measure things very well. You throw things out there, with as low cost as you can manage and with as great a diversity as you can stand and then you see what gets taken up."
This thinking, if you buy it, has dramatic implications for a marketing business most believe is entering a post-mass world. Media is fragmenting and consumers are more skeptical and harder to reach, leaving viral and word-of-mouth as the most attractive cost-efficient alternative to paid advertising. A way to slip into culture without paying millions for media, it transfers the work of distribution to consumers themselves, which is not only efficient in terms of cost but it also grants viral content or ideas more credibility since it came from a friend. In short, advertising is expensive and hard-to-believe; viral is cheap and credible. It is the un-mass.
But Mr. Watts' research poses all sorts of questions about how things "go viral." What if you should be planting content all over creation, turning your content-distribution plans into one big hedge against the vast improbability of any of it actually catching on, i.e. "going viral"? What if it requires something not totally unlike mass marketing?
Not everyone agrees, of course. Ed Keller, co-author of "The Influentials," said Mr. Watts is creating a straw man.
"The way we think about influencers -- and the way, quite frankly, the marketplace thinks about them -- isn't that they should be the only group you should target," Mr. Keller said in an interview. "But targeting them does help optimize word-of-mouth about their products and services. Duncan also never tells us how you find those easily-influenced people."
In reality, Mr. Watts doesn't claim influentials are entirely bogus, but that any assumption that disproportionately influential people are more important needs to be checked out empirically. The debate so far has been prone to devolving into a pissing match between people who essentially speak different languages, between an academic researcher and a marketing researcher. But one thing it has accomplished is refocusing discussion around just how social networks work at a crucial time.
In asking these questions, Mr. Watts isn't wholly alone, despite the private claim of one big-name word-of-mouth marketer who dismissed Mr. Watts as "an isolated thinker." Readymade word-of-mouth networks such as Boston's BzzAgent show that just about anyone with a yen for the insidery good feeling you get from shilling new products can be part of a viral movement.
And in April, a trio of researchers published some findings on an online retailer's recommendation network in paper in the journal "Transactions on the Web." They found that there are limits to how long recommendation chains last and just how influential those most influential nodes are -- "over a few of their friends and not everyone they know" -- and concluded that "viral marketing was found to be in general not as epidemic as one might have hoped."