"Googlizing" BI with Search-Based Applications
By Eric Rogge, Sr. Director of Marketing, Exalead
Organizations are increasingly storing vast amounts of unstructured data in new Hadoop, NoSQL, and MPP analytic databases, and business intelligence tools are getting better about connecting with them.
Still, even with improving connections between BI and unstructured data stores, the challenge with today's business intelligence deployments is that they only enable quantitative analysis of a fraction of an enterprises' information assets. That's because the majority of information available to an enterprise is unstructured content held in documents, e-mail messages, collaboration forums, and on the Web. Enterprises now realize that to have a complete, 360-degree view of their operations, they need to analyze that unstructured data. That analysis involves both qualitative assessments as well as quantitative analytics. The challenge of BI isn't storing the unstructured data; it is the significant back-end development work needed to gather and quantify unstructured information sources.
Missing from an enterprise's portfolio of BI tools are search and semantic processing technology, which can efficiently process unstructured data into gists and metrics, plus handle large volumes of data from widely dispersed sources.
The effectiveness of today's BI solutions can be improved by working in conjunction with search-based applications (SBAs). SBAs are a new, emerging category of search and semantic technology that aim to improve operational productivity through processing, analysis, and delivery of key information drawn from internal and Web unstructured data. SBAs are a form of business intelligence and complement the highly quantitative analytics delivered by traditional BI products.
Search-based applications complement the ad hoc analysis and quantitative reporting typical of BI implementations. Where BI addresses the what questions, SBAs address the who, how, and why questions to give qualitative cause-and-effect explanations. They do this by collecting and co-displaying quantitative metrics and explanatory text in the same view. SBAs are also useful for extracting customer sentiment and other informational trends from the Internet -- a complex task beyond the capabilities of traditional BI.
By integrating semantic search-based applications with BI information sources (sometimes called the "Googlization of BI"), companies gain a broader understanding of their business activity that enables better business decisions to be made faster. Instead of using a single source of data as with traditional BI, SBAs can simultaneously access a wide variety of information sources while combining structured and unstructured data to provide a holistic, 360-degree view of the enterprise.
SBAs handle staggering amounts of data -- petabytes in some use cases -- while simultaneously providing Web-search-style, natural-language query interfaces that appeal to ordinary users. Today's workers, accustomed to fast and easy Google searches on the Web, can now gain the same easy-to-use tools to help them unlock information in the enterprise and gain insights for better decision making.
SBAs have a different information purpose than do BI applications. Whereas internal and external accounting standards demand focused, precise numeric precision in BI applications, many operational decisions require a broad perspective, sometimes using a collection or profile of facts such as dates, contacts, impending transactions, milestones, and opinions to provide a complete understanding. Now that audio and video are becoming common information delivery mediums, the ability to transform such multimedia files into text (and then into analytic data) is becoming important, perhaps even critical in some situations. Emerging technologies, such as voice-transcription software, are adding to the deluge of unstructured data in the enterprise, which continues to grow exponentially each year.
However, not all search-engine-derived technologies are equal. Companies looking to leverage the power of SBAs to improve BI should look for several capabilities in a solution. To most effectively boost BI, SBAs must be able to structure unstructured data (not simply index it), as well as integrate that information into the corporation's existing structured datastores. The key to this ability is semantic search technology, which analyzes the content of unstructured data to make sense of the information and rapidly identify relevant data.
In addition, companies should look for SBAs that feature service-oriented architectures (SOAs) to integrate decision tools for each user, enabling rapid deployment and simple integration within the company's information ecosystem. Effective SBA solutions for BI will also include faceted navigation, as well as the robust data security required in today's corporate environments.
SBAs combine the best of BI and enterprise search to deliver what until today has been an elusive goal for the enterprise -- that is, real-time BI with a comprehensive view of all information sources: structured and unstructured, internal and external. SBA-powered BI offers improved data scope and relevance by making better use of existing structured data and by exploiting new data channels ripe with essential information, such as e-mail messages, Office documents, and PDF files -- the vast amount unstructured data that, until now, was beyond the capabilities of BI.
As unstructured corporate data continues to grow exponentially, traditional BI will be left further behind. The efficient scalability of SBA ensures that corporations will be able to continue leveraging their growing stores of information in order to make business decisions more intelligently and more quickly.
Eric Rogge is in charge of business development and strategic marketing at Exalead. He brings 20 years of marketing, sales, and development experience in the database and business intelligence software market. Currently he is senior director of marketing for Exalead, where he is responsible for market operational and development in the U.S. Eric is a past instructor at TDWI and was a co-founder of Ventana Research. You can contact the author at firstname.lastname@example.org