Enterprise Search Summit

Enterprise Search Summit

This morning’s Enterprise Search Summit in New York began with Daniel Tunkelang who has worked with Endeca, Google and is currently Head, Query Understanding, at LinkedIn.  Interesting talk.  I loved the term “crowd sourced curation” he used to talk about building shortcuts (using go/links in LinkedIn).  I liked the idea of mining search logs of queries to build labels and recycle the logs.  Daniel also referred to Peter Morvilles Search Patterns — something we will likely hear more about when Peter speaks in the fall at Internet Librarian 2013, Oct 28-30 in Monterey CA.

Following Daniel’s talk, Helge Legernes, CTO, Findwise reported on his organization’s second survey on enterprise search adoption.  What spoke to me was that people still do not find it easy (55%) to find the right information.  AND, that two of the top obstacles for not finding the right information, are related to tagging — the lack of it, and the inconsistency of it when applied.  We will definitely here more about this in the fall at Taxonomy Boot Camp, Nov 5-6 in DC.

Long time search industry watcher, Sue Feldman, CEO of Synthexis led a panel on search trends.  Sue has a new book out called The Answer Machine  (& here’s an earlier published piece on the topic).  Her list of search trends for the future:

  • Cognitive computing
  • Context, conversation, infoApps
  • Big data & analytics (she referred to Tom Davenport‘s 2007 book, Competing on Analytics)
  • Integrated information platforms, integrated information, integrated technologies
  • Learning systems

Comments from Sue’s panel: Jeremy Bentley, CEO, Smartlogic: “With 50% of people still unhappy with enterprise search, it is primitive and needs to evolve more.”Others added, perhaps with graph search.  Bottom line with the panel: search technologies are not as good as using multiple approaches to search.  So no silver bullet for enterprise search yet.  What disruptive technology do you think will cause something to shake up enterprise search?