I would like to follow up on a comment that I left yesterday on a great post, and to drive you into one of the things I’ve been craving to find for a long time.
Life-streaming —be it the simple Twitter, the socially optimized Facebook News Feed, any more elaborate FriendFeed-like, or even video-based Seesmic-like— needs a business model; a model that is both significantly far better at finding relevancy then current models appear able, and unobtrusive. People would not accept ads between two “Such Andsuch is now ‘in a relationship’” then they would during an intense phone-gossip session.
Ads have been financing every lack of business model for more then a century, and their appear (to Twitteriffic developers e.g. at least) the best, or rather only, solution. The problems are: they don’t work, at least not for Facebook (who is providing a service far more relevant then Twitteriffic) and they are obtrusive: not too much, but imagine on the long run ads from every agent in a modern (i.e. convoluted) process: the authors’ blogging platform (WordPress), the host (FeedBurner), your RSS reader (Google Reader), the social network where your friends follow your feed (Facebook), an aggregation engine (Yahoo! Pipes), some comment engine (Freiend Feed), your cell-phone provider (T-Mobile), OS (Google Androïd), maker (Nokia), and the application developer (Twhirl). I think we had enough of ‘free’ newspapers where the paper on the right page praises the company that bought the ad on the left page.
We must not interfere with the stream of information; however, every information is an opportunity to interact, and many of those interactions can have merchantable consequences. I’m not in favor of turning every street corner into a shop, but having a shop handy when you need it is a good thing — hence the image of the butler.
Say one of your friend says he is sad: you might want to cheer him up; maybe he enjoys going out, or flowers, or simply a phone call. Having that information understood by the life-streaming platform (I’m looking at you, Facebook ‘Favorite. . .’ questions) can help develop a Right-Click feature.
Say a friend just shared how he enjoyed a music album, a book, the last episode of a TV series, a film a play, a concert. . . Wouldn’t it be nice to sample it? To have you favorite provider One-click-able to get it? Or even better: to have your friends’ calendar sorted out to find who’d like to come see it with you?
These can be rather far-fetching features, I reckon — but automatically sharing what you are listening to has been a plug-in available between most IM clients/music player for years now; having a music-streaming site play it or you is not difficult; and the sales return of that sampling are very easy to measure. If we believe Watts’ experiment on the importance of peer-effect on music appreciation, that could increase significantly music consumption.
All this demands increasingly intelligent semantic recognition; the hand-fulls of dedicated companies appearing every month makes me quite confident this is a problem being resolved. The social and economic structure of such features is however more difficult to imagine: should I (a software on the emitter’s side) encode my ‘I’m listing to Madonna’s latest album’ with a reference, or should we keep the stream-lined 140 character long messages that simple (and let my readers and their own software find out which it is)?
How many feeds should I emit, and how should they be structured? Jyri Engelström wold say by object: you might enjoy my photos, not my taste for techno; I’d argue some objects (URLs) need more refined clustering — but who am I to decide how you want to organise my insights? Shouldn’t social-filtering be allowed in there?
Finally, who should rake in the ad profits? Most news are now shared and rated by crowdsourcing or friends before they are read: if [A]lbert mentions a product on his blog hosted by [B]logger, that [C]hris submits the link to [D]igg, [E]cho makes it front-page worthy and a [F]rugal customers comments about how cheap it is if you get it from [H]omeDepot. . . Who gets what? Please note that [E] is actually a structured crowd.
My take would be that consumers would be safer if all their buying details are kept close, and to have Amazon and eTicket plug-ins associated to their own stream reader — provided they don’t have so many different readers that this become unmanageable. The operator of that stream then does the semantic filtering (no need to buy that Madonna album: I bought it on iTunes already, but I can rate it up, or simply play it). This means that a company that combines a feed-reader and a semantic recognition plug-in (like BlueOrganizer) could finance itself through the Amazon reference program, or an equivalent at iTunes, InterFlora, etc.
Pro bloggers and fashion-makers would need some of that juice to continue operate, but for that, we need a more complete integration, and this is not only far more difficult then it seems, and a privacy nightmare.
Having all the information about yourself, your likes and dislikes, what you already have read, listened to and seen essentially has to be controlled by the end-user: that way, you could bombard him with ads, as long as only those that he cleared go through his screen when he decides to see them, this prevents wasted attention. More on that soon.