How long will it be before traditional ABC1 demographic data becomes either redundant, or at least of secondary importance to the more interesting and valuable behavioural data available on customers using digital interactive channels?

Probably quite a while as whole industries will need to re-think how they buy and sell advertising, but surely it has to happen?

TV and radio audiences are fragmenting across hundreds of digital channels, with increasing program or program-genre loyalty at the expense of channel loyalty. Niche channels are offering advertisers a very clear proposition "People who are interested in gardening", "People who like sport", "People who like soaps". Some of the larger mass market broadcasters are beginning to mine their BARB data for similar sorts of viewing trends (again, "people who watch sport" etc.).

Advertisers are surely more interested in users' behaviour than necessarily who they are, how old they are, where they live, what gender they are. Traditional demographics are merely a somewhat blunt tool to enable advertisers to predict behaviour: an AB 18-15yr old male living in the South East is more / less likely to buy my product or service. Adding an interest / preference layer to this sort of information ("..who likes to watch sports", "..who like to watch gardening programs") certainly helps refine the ability to predict behaviour but falls short of what it is possible to gauge through interactive channels such as the web, wireless internet, kiosks and iTV.

Transaction data, clickstream data and user profile data, all of which can be captured through these digital channels, build a picture not just of who someone is but how they behave. If I am selling gardening products do I really care what demographic the person is as long as I know they buy lots of gardening products? If I am lastminute.com wouldn't I like to advertise to people with a propensity to surf on Friday evenings, make impulse decisions, and enjoy going away for the weekend? As an e-commerce site, wouldn't I like to focus my advertising budget on reaching people who have an interest in my products and a high propensity to buy online?

If the above is indeed desirable, the reasons that it is not yet happening to any significant degree are probably:
1. The big players, buy and sell side, don't appear to want to rock the boat too quickly
2. Lack of standards and standardised metrics for these kinds of segmentation
3. And most importantly, very few interactive properties are able to offer these kinds of user behavioural segmentations to their advertisers.

In the first phase of building their interactive property, media owners concentrated on 'getting something up there' and usually the data capture mandate was 'capture it all, everything we can'. The next phase saw more of a focus on marketing, getting people to the site. Now it is time to concentrate on really understanding those users, building relationships with them, retaining them, devising customer retention and loyalty schemes. All of these things require customer-centric data. If you have that (how and why to get it is the subject of another post), then you can also begin to cluster your customers into behavioural segments.

I want people who are interested in e-business and have a high propensity to contribute useful postings to discussion forums. Can anyone give me that?