Yesterday I spent an illuminating hour with the folks behind Wowd, a still-private-beta search upstart that is taking a new approach to, well, just about everything in search as we traditionally understand it.
In a odd coincidence, this morning Venturebeat published a thoughtful piece on how search might shift when more data, in particular social data, is added to the mix.
I point these two links out as a marker of sorts, I’ve got a much longer piece brewing in me about Wowd’s approach to search, and how the Big Guys might respond should an upstart like Wowd get traction. More on that soon. Meanwhile, the Wowd guys posted on Tim and my Websquared paper here.
5 thoughts on “Thinking About New Models for Search”
This is definitely interesting and challenging. Social networking sites (except twitter) are not opening up. So search engines doesn’t have access to social graph or behavioral data. Twitter data can be right away be used by search engines. For example, i have written a small tool for myself to find out latest news on iphone apps. ( http://www.appsd.com/twitter/ ) For this i crawl twitter every 5 minutes and collects all links posted by users and rank these links based on number of tweets. This itself i find very useful. All the top articles are really interesting. Number of tweets are a kind of rank and it could be one signal for ranking.
The WOWD demo looks interesting but they’re just repackaging old concepts under a new interface. Popularity, freshness, and history with in-results drill-down have all been tried in various ways. The old Open Text search engine took on some of those concepts, as did Northern Light.
I think the next step in Web content development will borrow something from trends analysis. People will be looking for (probably already are looking for) derivative value and methods for distilling useful and interesting information from a huge glut of user activity.
I don’t mean for advertising analysis, but rather for showing visitors content creation. Think of every aspect of a Web page as a dynamic poll. You create the news when you visit the page and click on a widget (15% of all visitors like this widget!).
We could say we’ll conceptualize our conversations by expressing ourselves more abstractly.
Imagine trying to build a search engine for that.
But as more sites try to engage users through easy, simple content contribution words will become less important and aggregation will become more important.
Then the science fiction writers will get out their pens and their voice recorders and start debating which is more important: the social gestalt or the preservation of individual expression.
It would be cheap to say there is room for both on the Web. Of course there is. But the Internet is not the Web. The Internet is all of us working together to share information of all kinds. I suspect a lot of people will take the easy path more often than not.
You might be also interested in FAROO’s attempt to utilize the collective power of the crowd by Peer-to-peer Search and Real-time Social Discovery: http://blog.faroo.com/?p=181
In thinking about new models for search (local) consider this one(beta):
Provide extremely local data WITHOUT entering GPS,zip code,city-state, bookmarks,clicking a menu or map,or nearest cell tower approach.
Using a proprietary data base of new address codes for a search entry, that are geographically identifiable, easy to apply, and return a single answer- that micro community you are searching for on a single mini-portal that links to anything else- social, weather, news, business, etc.
Most important are the formatting of the unique terms that after you read 14 words of instruction,you can use them anywhere you want( without looking up a single term)keeping things simple and on site and not at the mercy of the search engines to find the answer.It may be simple, but it works.
You are right, Michael, a lot of people choose the take the easy path, that is why we chose to categorize the results on surchur and not just display them in random order.
I am looking forward to see wowd in action but until then I recommend you to use our real time search engine.