This week, both Information Builders (IBI) and FAST Search and Transfer (FAST) say they're making headway in bringing the ease of ad-hoc search to structured as well as unstructured information. IBI's advance, announced today, is called Magnify Search, a navigation tool that uses the metadata derived from Google or other search engines to index internal data stores. Users can then filter and analyze search results in what IBI calls a dynamic view (see screen shots).
Click to enlarge
Pivoting the same search result table.
Click to enlarge A tabular view of search results in Magnify Search.
"We're merging search with our Active Reports technology," says Radovan Kotorov, PhD, IBI's technical director of strategic product management. "When the results are returned, the user can display them in a tabular view within the browser, and they can analyze, chart, roll up results, essentially doing reporting within the result set." In a retail scenario, users could search for "shoes" and then create a pivot table with results that could be sorted by metadata values including brand, price, marketing campaign, discounts and so on. In a law enforcement scenario, researchers could comb through structured reports and unstructured records to find, say, relevant information on specific types of crimes committed in a given timeframe and location. Search results and associated tables can also be saved and e-mailed to others for collaborative analysis. IBI says Magnify Search has been tested with both Google Enterprise and the Lucene open-source search engine but will be adaptable to other search engines.
The whole point of search/BI exploration is to bring insight to a wider community of users who aren't familiar with structured query languages and complicated BI interfaces. FAST's foray into the BI market, the Adaptive Information Warehouse (AIW) announced yesterday, applies search to BI to enable users to explore multi-source data without normalizing in a warehouse or building OLAP cubes. AIW's two core components are FAST Radar, a BI portal that combines a dashboard and reporting front end with a search-style analysis interface, and the FAST Data Cleansing Solution, which provides real-time access to structured and unstructured data sources. The latter uses FAST's linguistic analysis technology to help match, merge and cleanse data automatically.
"Not only does FAST extract all the text in every document… text analytics can extract entities, like people, places and things, as well as concepts, so you can spot synonyms that describe the same idea," explains Sue Feldman, a search expert and analyst with IDC.
Search engines have trouble indexing structured data without the context of columns and rows, but Feldman says FAST's technology can index entire columns as well as table titles, so it supports ad hoc querying of structured data and reports. In contrast, most of the Google/BI integrations announced last year are better suited to scenarios in which you would identify your top queries and tune the search engine to deliver specific reports and documents that provide the best answers, Feldman says. While FAST may have an edge in search technology, its roots and depth in BI are largely built on Corporate Radar (the technology foundation of FAST Radar), an upstart BI vendor that had just $2 million in revenue when FAST acquired it last April.
"The ability to do complete ad hoc analysis, as opposed to being constrained by an OLAP cube, is very interesting and it may well be the future approach, but the question [about FAST as a BI vendor] is about the ability execute in the market," says Dan Vesset, a BI analyst with IDC. "Corporate Radar was a minor player in the BI market when it was acquired… Long-term, I think companies like FAST are probably good acquisition targets for larger BI vendors."