Search predictions for 2007

A very Happy New Year to you and yours!

I've been tagged by Ashkan Karbasfrooshan over on the HipMojo blog, in the 2007 Predictions game  started by Mashable. Now in this case, I'm a bit ahead of the game, with my initial predictions post: Disruptive Technologies for 2007 already out for a couple of weeks. But given that it came from Ashkan and noting the other stellar players that were part of the tag, I don't want to end the game here. My only choice is to go out further on a limb and post more predictions, so here goes!

Here are a few Search predictions for 2007:

1. Vertical Search Engines: 
At least a couple of VSEs will break into the mainstream public consciousness, generating sizeable amounts of traffic and becoming juicy acquisition targets. I expect that the sites most likely to break out are those that emphasize and aggressively exploit their domain expertise, rather than focusing purely on the search aspect of things.

2. New paradigms:
At least one of the existing mainstream search engines will flirt with new paradigms for search specification or results display, or both. New types of specification can include natural language input, parametric data, search-by-example, implicit behavior and logic statements, to replace the current focus on "keyword-ese". For improvements in results display, see my earlier post on Google search results  .

3. Information Discovery:
We should start to see a larger discussion on what I call "passive searching", also known as Information Discovery. As a user, in addition to actively searching for a specific piece of information at a specific time (the classic Google search), you may also want to be able to turn on a background search that acts as an agent, filtering items of interest to you across multiple feeds/web sites and sending you a periodic report on the results. (This can be approximated today, very roughly, by setting up RSS feeds on keyword searches in the major search engines.) Such an agent may also be able to suggest high-authority feeds relevant to the areas of interest, thus providing an element of discovery rather than search.

4. Prediction Markets:
The concept of Prediction Markets is currently only used by a few large companies (Google, Yahoo, HP and IBM, to name a few) and research units. I predict that this concept will break into the next level of popularity in 2007, seeing wider adoption by the next level of smaller, but more numerous, players.

There you have it, my second round of predictions!

I'm not sure when this game ends, since 2007 has already started, but taking my turn, I will tag a few bloggers - KarmaWeb  , Emre Sokullu   and engtech; if you feel up to it, guys, write a blog post about your 2007 predictions and keep the game going!

Comments

[Jitendra]

Interesting list especially the one about prediction markets...Let me think about mine...I'll add mine shortly :-)

-Jitendra

[Andrew]

After reading your first two (2) items above I felt like you were talking about GlobalSpec; the Engineering Search Engine. They already offer a pretty cool parametric search of over 150M+ parts and components, along with an "old fashion" key word vertical search of the WWW. And with 3.3M+ registered users it would appear that have some nice gravitional pull.

[engtech]

Predictions have been made: http://engtech.wordpress.com/2007/01/04/tech-and-blogging-predictions-for-2007/

[John T. Maloney]

Hi --

You may wish to note there is an industry consortium for prediction markets.

http://www.pmcluster.com/

-j

[pandora australia]

I expect to see your new share.ma le scarpe non ti donano. meglio dei sandali con stringe più sottili e caviglia libera Perfect!

[tiffany uk]

These articles written too great,they rich contents ma le scarpe non ti donano. meglio dei sandali con stringe più sottili e caviglia libera Perfect! and data accurately.they are help to me.I expect to see your new share.

[pandora bracelets]

After reading your first two (2) items above I felt like you were talking about GlobalSpec; the Engineering Search Engine. They already offer a pretty cool parametric search of over 150M+ parts and components, along with an "old fashion" key word vertical search of the WWW. And with 3.3M+ registered users it would appear that have some nice gravitional pull.