Monday, September 26, 2011
VIRAL MARKETING MODEL
Ideally, you’d like your business to “go viral.” When a Web site or a product or service spreads rapidly, by word of mouth or blog of ‘net, it is said to have “gone viral,” spreading like a flu epidemic. Doctors and epidemiologist have a mathematical model to describe this, the logistic model, and what I have described here is based on that, presented as clearly as I can. It came from thinking about hits on my web site, tinandi.com, and sales of my book, Ting and I. For simplicity, I will stick with modeling the sale of the book, assuming for simplicity that one is either a buyer or a non-buyer.
We can model the rate at which the books are sold as follows:
Rate = (books/buyer)(conversion rate)(buyer fraction)(population)(1-buyer fraction)
Rate = B C F L (1-F).
Writing them next to each other means they are to be multiplied.
Let’s analyze these factors individually. B = the average number of books per buyer; it will depend on the book, the buyer and the season. Gift books at Christmastime will have a large B and encyclopedias during the summer will have a smaller B.
The conversion rate, C, will depend on how likely it is that one customer will convince another to buy the book(s), too. An attractive, inexpensive book that becomes the topic of conversation will have a much larger value of C than an unattractive, expensive, obscure tome. Note that word of mouth and blog of ‘net help to spread the news about the book, potentially raising C dramatically.
The fraction, F, of the population that have bought the book starts at F=0, where the model fails and the rate is zero. By giving out free samples, review copies, cut-rate offers, one can get F up off the floor.
The limiting population, L, is determined by many factors: communications, geography, demographics, for example, some of which you can overcome.
The medical model of epidemics assumes that once affected, the individual is thereafter immune. In our example, he has bought all of the copies of my book he is ever going to. This is the (1-F) factor. At some point, there are no new eligibles for my satisfied customers to influence. This model shows that one would expect initial sales to start slowly, as F is small, thus there is a real value to “priming the pump,” as mentioned above, with special deals, discounts, etc., to raise the fraction F who are familiar with the product from that small initial value. After this, the process “feeds on itself,” a “virtuous circle,” with old customers recruiting new ones.
If the classic medical infection model [“logistic” curve] obtains, the number of books sold versus time follows an S-shaped curve, starting near zero, rising slowly at first, then most rapidly when half the eligible, the limit, population has been affected, then slowing until it reaches that limit, F=1., when all those who might buy my book will have done so.
Creating new products or services to replace the old can keep your target population from becoming “immune.” Where the customer can be expected not only to recommend the product, site, service to others but repeat using it himself, the sky is the limit. You’ve gone “viral.”
Douglas Winslow Cooper, Ph.D., is a freelance writer and retired physicist, author of Ting and I: A Memoir of Love, Courage, and Devotion, available from Amazon, Barnes and Noble, and Outskirts Press. His email address is firstname.lastname@example.org .