Ever since Google launched the search appliance, there has been a lot of noise about the convergence of business intelligence and search. Though the business drivers are quite clear, BI search is still an evolving market. I would say search and information management are two sides of the same coin. They converge where unstructured data meets with structured data, which then flows into how information is managed.

Analyst research has shown that more than 80 percent of all business data remains unstructured. This includes both data from sources such as physical papers, scanned documents, audio recordings, videos and text from Word, PowerPoint, emails, blogs, etc. As the awareness and usefulness of BI search increases in market, the misconceptions are rapidly being built. Here, I disprove some of the misconceptions associated with BI search.

Misconception 1: BI search complicates the user interface. Though user interface designers might find it unsuitable to keep the user interface simple, Google’s instant success when it began in 1998 was driven by its very simple UI. It’s almost impossible to believe that the simple combo box UI actually tags petabytes of data today. The search UI should be the gateway through which information is accessed from the data warehouse, hence, the UI should be kept simple. However, the underlying design to access and retrieve the information should be very robust.

Misconception 2: BI search and text analytics are mutually inclusive. BI and text analytics are not mutually inclusive; they complement each other. BI search and text analytics are extensions of BI and data warehousing environments and investments, but BI search won’t replace the BI/DW. BI search is all about what to search, how to search using crawlers, and how to index it in a database so that it is in a structured format and is ready to be accessed. Whereas text analytics asks, what is there and uses a software or process that will parse text and extract facts from structured data (database management systems), unstructured data (text, email, IMs, blogs, etc.) or semistructured data (RSS Feed, XMLs, electronic data interchange docs, etc.).

Misconception 3: BI search is a new technology. Search is not new; however, searching unstructured data and preparing it for extraction to feed downstream systems is new. Similarly, unstructured data individually might not be voluminous; however, the effort to bring it in a mainstream data warehouse is a time-consuming process. Take precautions: the individual components may not be of significant volume, but as a whole the overall volume of the data warehouse and the BI systems are going to expand exponentially. The changed areas (from a data warehousing and BI standpoint) will be:

  • The way analysis is done;
  • The ways to bring in source data;
  • Data integration methodology;
  • Data storage methodology; and
  • Training will become even more important than it has been.

Misconception 4: It’s just about using text mining to access existing BI information. These days, the pure-play vendors have started building text-mining capabilities integrated in their tools. As a result, the concept of search will have to be applied differently at a process level and at a tool level. Text mining is for mature data warehouses. Hence, one should invest in it only if catering to unstructured and/or semistructured data along with the structured data. Thousands of reports are catalogued in hundreds of folders, and text mining should be able to help find  information and present it in the optimal way via performance dashboards. In the process, it has to take care of aspects such as security and permissions.

Misconception 5: BI search automatically offers immediate, enterprise-wide benefits. There’s no single correct way to approach BI search. Search technologies vary greatly – from simple keywords to relevancy ranking (like Google) to text-mining algorithms. The approach and the selection are tightly related to an individual business’s need. BI search can be used to strengthen a company’s key performance indicators by bringing in additional key information to performance dashboards that was not available before, or it can also help derive some patterns and forecasting by feeding the predictive analytic engine. Because search is both a process and technology, one has to utilize its potential according to the demands of the situation.

There is no instant benefit derived by BI search by just adopting a search technology unless you integrate it tightly with your mainstream information management. Companies investing in search should deploy the technology with a specific problem and user community in mind.

BI search, text analytics and text mining are all add-ons to the BI stack. Companies investing in these technologies should deploy the technology with a specific problem and user community in mind. If implemented carefully, these technologies are yet another innovation that will increase BI adoption. It’s all about bringing in a change in BI culture.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access