Back to Blog… behavioral targeting, collaborative filtering, web 3.0 and the human touch
Amy Winehouse goes back to black and I go back to blog… after a long long break I am back full force to babble on about the beauty of technology 🙂
While I was away from the country, a mountain of papers and magazines has piled up against my post box. One of the papers were Advertising Age, which last week published a provokotiv piece around how to monetize social networks. Now, this conversation has been baking in the oven for almost a decade now with no simple straightforward solution or formula. And here comes Mr. Bob Garfield, and lays down the solution – haha of course I will not give it to you here on a silver platter! you have to go and read the entire 5000 word article… !
ok ok, I’ll give it to you here.
after much exploration and anlysis of contextual search (did anyone say web 3.0?), collaborative filtering and pattern recognition, Garfield recons the business model to sustain Facebook is….. (can you hear the drums at the background) hold on to your seats…
but wait – not just any other advertising. Trageted Advertising based on behavioral targeting and collaborative filtering: so you downloaded TV related application, you performed some quizes with your friends to compare top 2008 fav series – you will have a page offering you similar products “what you’ll like” etc. This is not a new approach. Garfield is saying hey, if it works for amazon, ebay and netflix, it will work for facebook. Garfield is making a slightly negligent deductive assumption here:
netflix and ebay are communities.
targeted behavor works there – so targeted behavior works for communities.
hence targeted behavior should work for facebook, as it is a community.
oouch, that hurt!
NOT that I am saying this CAN’T work. However, users, just like me and you, know about behavoiral targeting and collaborative filtering. They may not use these terms per se, but they know when their authentic experience has been compromised, and I will give you an example. 8 years ago I witnessed a major resistence coming from a very specific community when it didn’t like the sponsorship we offered it: We offered our Cats community – one of the leading, strongest communities – a 50% coupon to buy a specific brand of cats food. Before posting the coupon we checked the general idea, without mentioning brands, with the Community Manager to ensure the memebers won’t feel we are selling their platform to advertisers. That we’re not invading their home. ‘great idea!’ the manager said. However, the day we posted the coupon, a major shout-out came from community members – ‘either you remove the coupn, or we’re leaving to a competitor platform!’ they said. ‘But why???’ we cried ‘we checked with you and you said it’s A GREAT IDEA to use coupons!’. Community members, with great disappointment (that kind a person has from their partner after the latter forgotten their anniverary) replied ‘coupon is great, this brand however is a big NO NO’.
Turns out that specific brand also manufacturs other products, doing some experimets with animals while at it, which to that specific community was a big deal. and so targeted advertising might have worked there, if it was the exact thing the members are looking for. A Veggie Burger offered to a Carnivor. the paltform provider needs to be ultra attentive to their audience to make sure he strengthens the ground of their relationship, not breaks it.
In order to work, Behavoral targeting needs to be carefuly combined with collaborative filtering and semantic web -yes, absolutly, Garfield is correct here. But so it doesn’t kill the community, the platform provider needs to be extremely close to user and community. This is something I am afraid cannot be done by technology, yet.