Shake up search engines - collaborative filtering

[alextj]

I hope I'm not behind the curve on this, but I've just read an article in The Economist's Technology Quarterly about collaborative filtering, and this really could turn seo on its head.

Essentially, instead of creating a popularity and relevance structure based on links and referals it uses surfers' browsing habits and history to match up users/items with similar tastes/viewers, and so recommend sites that that may interest the user based on the viewing history of that user's peers.

I believe something similar is already in use on amazon, and I believe it was the source of the problem a user had a couple of weeks ago with his site bungying around the rankings. But this definitely seems like the way to go - there are some privacy issues, but it would really help in narrowing down those 1 or 2 word searches that throw up 5 different topics using the same keyword and result in having to scroll through 5 pages to get a couple of references to what you were looking for.

It will be difficult to promote your site other than sheer popularity, at least after the browsing history is build up; but on the flip side it should be delivering a better targeted audience our way....
Look out for collaborative filtering info projects soon and forget about those off-topic and search engine links...

[cbp]

I guess it would be way too easy to manipulate - see the threads on the Alexa toolbar.

CBP

[Jayms]

Perhaps not if they can somehow apply that relevence measurement to the masses as opposed to just those with the toolbar. It can be done, but privacy issues are the biggest obstacles in my opinion.

[greeneagle]

Try optimizing for local searches.
ie.. I'd place my address/s on each page. You will most likely have better results playing that out at different levels in your presentation.
Ken