Gartner PCC Summit 2007 (part 10)

"So not done: Search Technology’s Strategic Future will surprise you” (by Whit Andrews from Gartner) at PCC 2007 Summit.

- what is driving the pervasiveness of information access technology?
- how will enterprises deploy the power of information access technology in unexpected ways?
- how are vendors responding to increased popularity of information access technology?

The search box bridges users and data of all kinds is built up by users, relevancy and data indices/interfaces.

In YE12, search logic in more than 75% of new installations will contain some internal user identity.

Search comprises everything on the continuum, structured to unstructured data/information.

Users have no idea that something is not searchable.

Support for extending the searched: the spider and the ant.
Spidered data is collected and the index remains iterated but static invoked addressed return current pages. (Pro: stable/known - Con: Can go stale, link rot, enormous storage need)
A query-driven ant knows the proper path to travel to collect a fresh version of what may likely be relevant data. (Pro: fresh - Con: extra application load)
Whit says the ‘spider model’ is not enough in this real-time environment.

In 2008 20% of the result data in new info access projects will be refreshed at the time of the search.

By 2008, query federation will play a key role in more than 50% of strategic new info access projects in the Global 2000.

More and more a single best way of defining relevancy is outdated. People ask for ways to tweak the relevancy based on their needs.

If you’re looking for search “pick a platform” with great breadth and establish tactical alternatives.

By 2012 no Global 2000 enterprise will standardize absolutely on a single search vendor. This gives enterprises flexibility.

The necessity remains to access structured and unstructured information from the same search platform. However, users don’t understand that this is difficult.

Query-corpus mismatches will be important too. This gives companies a way to define what impact not-finding information at a certain point for a sales deals, for instance.

Lessons from Web 2.0 sites/vendors:
- non-textual results (LivePlasma, Grokker)
- live query refinement (Clusty, Pandora, Acoona)
- user tagging (del.icio.us, NY Times!, flickr)
- Users links with each other and publish to each other (Jeteye, LinkedIn, Rollyo)
And of course there’s overlap between these.
Advice: investigate what social search could mean for your organization.

So, you have two main approaches: straight-forward search and social search. They’ll meet in the middle. But also pay attention to the niche players in search, such as business intelligence search.

Magic Quadrant for Information Access Technology, 2006
Autonomy, IBM, Fast Search & Transfer, Endeca and ZyLab are in the upper-righthand corner of the Quadrant.

5 most powerful vendors:
1. Google (ease of use)
2. Microsoft (good-enough, due to Sharepoint and Vista)
3. IBM (broad product line)
4. Adobe (will acquire Autonomy or FAST)
5. Oracle (due to their database technologies)

There is new vigor in unconventional delivery models: quality of service, agility and economics.

Recommendations:
Today:
- list what you think you’re searching
- inventory your search vendors
- learn what you can about relevancy
Next year:
- begin to make search more pervasive in your enterprise. Integrate, do not unify, search and business intelligence; search and CRM; search and multiple RDBMS and so forth
- add 50% of what you thought you were searching - but weren’t
- add 50% of what your workers wish they could search
- select your search platform
- inventory and value your relevancy strategies
The years after that:
- add the other 50% you thought you thought you were search
- add what the users wish they could search