The Long Tail & Recommenders

Chris Anderson explains his “long tail” perspective as “the mass market turned into a million niches.” A variety of new media platforms (what I call “interest aggregators” – iTunes, Amazon, etc.) are providing distribution channels that are so big and niche-centric that products and services traditionally represented by low demand or low sales volume are now collectively making up market share that rivals or exceeds the relatively few bestsellers and blockbusters.

What does this mean in terms of marketing? If you can generate word-of-mouth endorsement that spreads through online communities and if these endorsement get reflected in interest aggregators (e.g., iTunes “playlists,” Netflix “friends’ queues”), the underdog can win (or at least compete). You should be able to use niche communities and targeted messaging / experiences to locate markets and sales channels for products and services that would otherwise never be able to compete in a mainstream media, mass market environment.

This looks good on paper, but how does this marketing vision (utopian?) play out in an ever splintering, and exponentially growing, media landscape? Take the blogosphere. According to Technorati, there are over 7.8 million weblogs, and 937 million links. And it’s doubling in size about once every 9 months. Yikes! A million niches for a million niches?

Basically, the burden will fall squarely on “recommender engines” to make sense of this crazy world of ubiquitous editorial content and distribution channels. I like to think of these engines as "pushed search" based on personalization algorithms. At the most basic level, a recommender engine is like Amazon’s “customers who bought this book also bought…” When I search for Malcolm Gladwell’s new book “Blink” in Amazon, search results include recommendations to purchase “The Tipping Point,” “Collapse,” and the “Wisdom of Crowds,” each based on the aggregation of prior purchasing behavior.

The trouble with this model is that it biases recommendations towards the big hits and most popular titles. So, a recommender engine in this regard poses a significant challenge to the underdog who is searching for markets and sales across the long tail of consumer demand. Moreover, recommendations based on an algorithm of similar purchasing behavior will never be as accurate as an algorithm that reflects true psychographic variables and preferences. The BrightCove and Many2Many blogs both referenced this dynamic recently and provide some good examples of other recommendation engines based on other algorithms, such as the Upto11.net “Popularity Slider.”

There will probably be a natural selection and extinction process among recommender engines – both stand alone and built into existing distribution channels. I’m thinking of companies such as ChoiceStream that will survive and evolve. A key trait associated with survival will be related to portability and seamless integration with portals and interest aggregators. Imagine real-time preference updating across a variety of platforms, almost like Plaxo, but for recommender engines and portals instead of email address books. The ones that survive and evolve will be based on simple but powerful principles.

A couple years ago, Philips Research USA produced a white paper that identifies three: accuracy, ease-of-use, and trust. I’ve talked about this elsewhere, but trust – or source credibility – is becoming one of the strongest predictors of online consumer behavior. I imagine this will continue to be the case, especially as the credibility of mainstream sources continues to tank and opportunities for self-selection of media content continue to grow.

The more accurate, credible and easy-to-use that recommender engines become, the more access marketing and PR strategists will have to leverage the long tail for our clients. In the meantime, it behooves us to develop targeted messages and experiences that help saturate relevant niche communities with genuine word-of-mouth content and reciprocal linking. Working within the existing search context, the long tail strategy has to be focused on pluncking off key niche distributors, owning these channels (in terms of message and brand experiences), and nurturing word-of-mouth communications (napsterizing niche-centric brand experiences). CMO Magazine has a nice case study about how BTI Communications Group was able to use blogs to compete with AT&T in the VoIP space despite the lopsided media presence garnered by the big bell.