Search technology encompasses everything from Google and Yahoo online tools to enterprise software from vendors such as Autonomy and Fast Search & Transfer. And it's quickly becoming a critical element of the information management toolbox within many companies. Search generally isn't a substitute for business intelligence tools, but it can be better suited for some tasks, such as when the specific source of needed information isn't known.
Business intelligence tools are designed to help users looking for a specific answer or wanting to get a view of a data set, says Gartner analyst Betsy Burton. They generally support structured queries into known information sources such as a company database. Search tools, in contrast, are more free-form. Users may not know exactly what they're looking for--the process of discovery, in fact, can be just as important as getting the final answer--and the data source may not be known.
My introduction to BI tools came years ago when I went to work for EDS as a tech writer/analyst where I helped put all system documentation online, and then participated in the development of tools to search the data. At my next company, I managed the implementation of front-end apps to Pacific Bell's massive consumer datawarehouse, where superusers accessed the system with DSS Agent, BusinessObjects, and several SAS modules. The problem with these large systems and tools was that they could only be accessed by the superuser through well-formed, structured queries. The data was then sorted, parsed, and served up in pre-defined reports for the masses. The cost of these tools was/is high, which limited what small to mid-sized companies could do with their data.
What's exciting about the advances in search technology is that what was once a light-weight, general search medium is becoming just as powerful as the more expensive BI suite. While search will never replace the BI capability, the gap between superuser and the common end user will decrease dramatically. By pairing BI with search, search results become more targeted and accurate, and, ultimately, businesses can get more value out of their application investments.
Over the next five years, search-engine vendors such as Google will establish relationships with online information providers, such as the Securities and Exchange Commission's Edgar database, the U.S. Census Bureau, and Major League Baseball, predicts Gartner's Burton. Those relationships will involve building links between their search tools and the metadata that underlies those databases, which will make search more like a BI tool in that it will be searching structured data for specific information and answers.
Business intelligence software developers already are taking steps to tie their products to search technology. This week, Information Builders will unveil WebFocus Intelligent Search, which, using the vendor's iWay integration software, lets users search structured and unstructured data using the Google Search Appliance and combine the data for analysis using the WebFocus BI tool. Cognos, SAS, and SPSS are among vendors that link their BI software to IBM's OmniFind, a search engine for perusing unstructured information in documents, E-mail archives, Web pages, and other sources, using IBM's WebSphere Information Integrator.
The pervasiveness of search is bolstering the BI space across the board. However, the skill gap between superuser and end user is still the major impediment to widespread adoption.
"The biggest problem is lack of skills," Graham says. "Some of the tools are pretty hard to use." Some products still require a great deal of IT intervention, such as developing reports for users. "If you're going to roll out BI throughout the company, you've got to get IT out of the middle," she says.
It's no wonder that ease of implementation (cited by 68%) and ease of use for a broad range of workers (67%) topped survey respondents' lists when asked about the most important features in a business intelligence product.
Advances in search capability should shorten that gap in the coming years.