The web, and the social web in particular, has been hailed for its contribution to a new economic principle: the long tail. It was brought to our attention convincingly a couple years back by Wired editor Chris Anderson, and has since enjoyed its own long tail of popularity. The argument goes, basically, that the web’s connectedness, combined with its increasing use as communication among users, not only sustains but often accretes visibility to qualified products and services *over time*.
The marketing blitz favored by short-term and launch campaigns (the power curve) calls upon the attention supplied during a curious and enthusiastic audience. The long tail offers the opportunity for a much more slowly accumulating span of attention, which is the attention brought to the product or service by the interested user. Where the power curve delivers newsworthy results, the long tail delivers powerpoint-ready follow through.
Long tail is the discovery economy. It’s a sort of “socialization of search results” — it works by qualifying the value of results by means of social participation. The big stories in long retail have been entertainment and media consumption, e.g. Netflix, Yelp, Amazon, and so on: sites that enable discovery along the lines of “what I like” by means of meta data captured through site use and social participation along the lines of “things that are alike” and possibly “liked by members who are alike” and/or “liked by like-minded members” (the difference is associating people on the basis of what they like or whether they seem alike: interests vs. identities)…
Long tail works by talk — a kind of structured talk (because it captures the declaration of like, the level of like, the likes associated, and then of course cross-indexes those with item data and meta data such as category, genre, etc). So it works because it happens when users are interested. And a user who is interested, like a user who is searching, is a good indicator of value and relevance.
Now let’s consider the dramatic growth in conversational media: represented by tools like twitter and friendfeed. These are not tools designed to capture content in depth. Nor do they extract much by way of meta data, taxonomic relations, etc. They’re designed for simplicity and used for a faster form of talk than we get in walled-garden social networking sites. But as there is a great deal of activity in the conversation, much of it highly relevant and most of it uniquely particular to its users, it behooves us to ask: what value can be extracted from the content and relationships of conversational media?
Where in these media would an advertiser wish to be? In the tweet? Between tweets? Alongside the twitterer, his or her stream, with or without friends? Or perhaps in search results? Hashtagged? … You get the picture. Advertising always wants to be placed in the best context possible — but the context of conversational media is highly biased towards use and utility. If twitter were ham radio, the context of use would be the microphone; the context advertisers know to recognize is not the microphone, but what comes out of the speaker. What gives conversational media their utility is their functionality. To use another analogy, we don’t use phones for listening — hence our resistance to tele-marketing. Of course, I have a suspicion, unproven, that far more attention is paid by twitters to the tweeting than to the tweet reading. That’s a bias I think is shared by all posting media, and a reason for their high redundancy of communication.
There is enormous interest by third parties and by media-related businesses in the rise of conversational media. They want to know how to leverage these tools for their own purposes, be this through participation and engagement, or by monitoring and tracking. Social media marketing and advertising will mine status, news, activity, and other self-talk and conversational feeds for the kinds of valued relationships (people to people, people to things, events, etc) and associations (people in groups, audiences; things in groups, categories) they contain. The strategy here, however, may not be long tail.
Indeed the coming feed market may want to think in terms of the power curve. The personal and social news-making they are mostly used for have more in common with the power curve of news in general than they do with in-depth discovery. You might argue that discovery is surfaced through conversational tools, as in the blogs we read and then tweet to. Conversational monitors, and tools like Radian6, Buzzlogic, and smaller twitter monitoring apps, might then combine deep blog crawling for long tail value, and feed/conversational content for the power curve (breaking news, memes, viral, etc).
It will be interesting to see where this goes. I don’t think that the influencer approach most often cited as a model for conversational value is the be all and end all. It comes out of social networking models, and misses the conversational dynamics of talk tools. I suspect that the best mining applications will take a time-oriented approach over a network-based approach. (The network of followers is no network, it’s a list, and there’s no guarantee that the audience represented by a list of followers is paying attention to the stream).
Conversational media are short on content and long on activity. But if the medium is the message, the message value of conversation media may be on the envelope.