Convergence of Enterprise Search and BI?

Today Business Objects announced it has completed the acquisition of inxight, the developer of enterprise search tools for linguistics and text analysis.  inxight products were used in mainstream engines like K2 from Verity before a falling out over competing products and technology a couple of years back. Verity, of course, has been part of Autonomy for well over a year now - talk about consolidation!

This seems to be part of a continuing trend as enterprise search vendors discover BI as a potential market; and BI players like Business Objects discover the importance of search technology in great BI. Just a year ago, FAST Search and Transfer acquired Corporate Radar which it now markets as FAST Radar.

The real question is this: How well do BI and Enterprise Search really integrate? Andy Hayler of Kalido  asked this insightful question (no pun intended) earlier this year, and as yet I don't think anyone really has an answer. What do you think?

Comments

[Marie Risov]

Just came across this post and realized that it fully expresses the process we started implementing in our organizaiton (EDS) years ago. Search tools that provide fact extraction thus create data that is suitable to be stored in a database, together with any previously available traditional or structured data. Apply all familiar BI methods after the conversion, and you are dealing with the complete data solution. I would also suggest going beyond the traditional OLAP solutions and use advanced analytics methods, such as regression, decision trees, neural networks, etc., to assist in intelligent decision making. Clustering can also be performed on the new level, where the original structured columns and the newly extracted facts provide context for one another.

[Dave Haucke]

To date, enterprise search and BI vendors haven't done a decent job of communicating the integration roadmap. As such, we're left wondering what all the hype is about. Am I suddenly supposed to replace my database query tools with enterprise search to query my structured data? Or am I merely expected to use enterprise search to allow users to query our reports library and hope the metadata in those reports are strong enough to provide relevant results?

This story isn't about search being shoved into a BI system and replacing BI's built-in query tools. It's about how the two work alongside each other, doing what they are best at and complementing each other.

In that sense, the real story is how enterprise search does for unstructured information what BI does for structured. What intelligence can you glean from things like customer service emails, live chat logs, blog sites and other reference sites on the Internet? And then, how do you bring that back into the BI system and correlate that info with sales reports, KPIs and the like to make intelligent decisions about your business?

Enterprise search morphs from being a tool designed solely for connecting an individual user to a specific document into a solution capable of helping analyze topics of interest, measure sentiment and track trends. BI has been doing this using only numeric values, but the reason why this is a hot topic is because the really juicy stuff is found in unstructured content.

We aren't there yet, and I don't suspect we'll see a fully integrated solution that has a proven test case for a couple of years. But the dashboards of the near future certainly make this a compelling topic, and I have to say that Inxight was the perfect vendor for Business Objects.

[Ragnar Heil]

You also might want to know that FAST has aquired bwise yesterday. They are specialized in BI and co-produced FAST Data Cleansing Solution

http://www.fast.no/news.aspx?m=330&amid=9670