An entrepreneur approached us recently, puzzled and a little frustrated. She had what was, on the face of it, a great idea for a personal gift business. Even better, one tailor-made for viral marketing: it was the kind of thing that you'd expect people would pass along as an affordable and useful service to family, friends, neighbors, co-workers, fellow church-goers... But despite doing the de rigueur marketing du jour -- paid search and telling folks she new about it -- she wasn't making any appreciable gains toward the 100,000 user milestone she had set for herself and for investors as a gate for further effort and funding.
Viral marketing is conceptually easy to get, but tactically hard to do. One way to understand it better is to actually approach it the way epidemiologists do, which is to understand disease vectors and transmission mechanisms. The former refers to the paths through which an illness spreads, while the latter refers to how sickness gets passed along, and by whom. Listening to the entrepreneur it became clear that while she had a highly nuanced view of her ultimate target customer segments, she had a very uniform view of the folks who would get the message to them.
To unpack the challenge, we suggested the following exercise. Assume for a second that the virus will spread in a simple 1:10 pattern -- that is, each person tells ten, who then each tell ten in turn. In six "hops" you get to 100,000 people (ignore the decay rate for the moment). Question is, how do you sustain signal strength and consistency across each hop? The answer lies in figuring out who in advance your likely carrier for each hop might be, based on formal and informal roles; next, understand his tribe, his mechanisms for engaging it, and his motivations for doing so; then package your "virus" for maximum effect in that context.
Here's how this might work in practice. Let's take a popular example: "most-emailed articles". Today, the tools used by popular services like NYT Online include a form element where you can add your comments on an article you are forwarding to an acquaintance. One essential lesson from "Structured Collaboration" best practices is that blank text fields are some of the least effective elements for enabling people to interact. Most people, confronted with the "comments" field, might type something like "Thought you'd find this interesting" if they type anything at all. I presume they wouldn't have sent the article if they didn't think the recipient wouldn't find it interesting -- making the comment gratuitous, and missing the opportunity to convey something else.
That "something else" could be information that could provide *context* to the recipient. Let's assume the form element was, instead of a comment, a drop-down list or set of check boxes that described the recipient, or recipient group in some way. For example, recipients could be "work", or "school", or "church", or "government", etc. What would this allow us to do? It would allow us to convey the article in the context of other related articles that together might point to a theme.
This synthesis could help to convey a point much more powerfully than sending along the isolated example might. (To sustain usage of a mechanism like this, the email might provide only a short article summary and a link that would bring each successive recipient back to a site from which the enhanced forwarding mechanism could be used again.)
Taking this one step further, add tagging. Other readers subscribing to those tags (via email or RSS, doesn't really matter) will be more "susceptible" to reading those forwarded articles knowing they have been categorized into a bucket they care about (at least until tag spam causes this increased receptivity to decay).
Now, let's add a motivation layer. For example, let's assume that if you forward an article, or for that matter any other viral marketing element, you and the recipient are automatically entered in a sweepstakes drawing, or get a coupon. Providing context to the recipient also provides information to the sponsor/ marketer that allows these offers to be more closely tailored.
On a related front, my friend Bill Ives just signed on to help another friend, Peter Gloor, at Peter's social network analysis software firm iQuest Analytics.
I'd be interested in hearing about any applications people know of that demonstrate these ideas today.
