How Does Business Intelligence Work with Unstructured Data?
Suppose that you’re using an unstructured, multimedia-enabled data warehousing environment to do comparative analysis between services offered by your company (a bank) and your competitors’ corresponding offerings.
You run some basic reports and a few queries to check out market share, portfolio performance, and other measures. Or, for more advanced analysis, you use a business analysis OLAP tool to perform all kinds of drill-down analysis on the data in an attempt to fully understand the intricacies of your company’s performance with respect to your competitors.
Sometimes, though, you can’t find the answers in the numbers. Suppose that you notice a sudden increase in account closures at your bank during the past two months. What’s going on?
Types of unstructured data
You can understand the premise of business intelligence in the term itself: Get as much intelligence as possible — as fast as possible — from as many sources as possible, to help you understand what’s going on and take informed action. Under this broad definition, intelligence can easily include the following types of information that you can’t find in (or access through) a traditional data warehouse:
A competitor’s local newspaper advertisement offering no-fee checking for one year and an extra 1.5 percent earned on money-market deposits if a potential customer shows a bank statement indicating that he or she has closed an account at your bank
An advertising banner on Google that features your competitor’s same offer
A link to each of your competitors’ Web sites, where you can analyze the types of electronic banking services they offer
A transcript of an interview with a regional economic expert stating that your bank is a prime takeover target and probably won’t be in business under its current name at the same time next year
In this simple example, because the items occur locally or regionally, you might believe that you can access all this information from a multimedia-enabled data warehouse. (“A good banking analyst probably knows all this stuff anyway,” right?)
A global example of data warehousing
Think about this example on a global scale, however. Are you wondering why your company’s sales are slipping in Sweden? You might need to have these types of real-time, intelligence-gathering capabilities for a globally competitive situation.
For example, imagine a company in the chemical industry that wants the architecture for a quasi-data warehouse environment (quasi because it has only a single source of data but a huge amount of historical information that would have to be brought into the new system).
About 80 percent of the historical information was on paper, and the client was considering eventually entering that information into a document-management system. For budgetary reasons, they will deal only with the conversion of traditional historical data (character, numeric, and date information), and mapping and transforming the new incoming data. The documents would be handled later.
Imagine an environment in which you can treat all this data, which deals with the same subject matter, equally. If the data is on paper, you can scan it in as an image, index it by keyword, and make it accessible through the same environment as the traditional data. You tremendously increase the client’s business intelligence by giving them access to this information.