E-Business
Issue No. 5 - April/June 2002
Mining the Back Paddock - IT Boosts Business Intelligence
by David Twiss
Business involves making decisions, and Business Intelligence (BI) is about using systems to support business decisions.
Every business gathers lots of information in its online systems—information on financials, personnel, sales process, inventory, manufacturing, production and logistics. Most of the information is in separate systems which exist for specific purposes in the business. Best practice in BI recognises that the aggregate of this information has value beyond the sum of the value of each piece.
But what are we going to do with this sum of knowledge?
Imagine running a bank. Wouldn’t it be handy to know which loan officers tend to write the most risky car loans? How about being able to recognise account activity patterns that show when customers are about to close their accounts? You could target them for special offers and maybe retain their business. How about knowing which customers fit a certain demographic profile that would qualify them for a new loan product that you're about to launch?
The BI techniques one would use to answer these questions are known as online analytical processing (OLAP) and ‘data mining’. OLAP is all about aggregating data from various sources to form a multi-dimensional data model, and then using tools to form aggregates, summarise and ‘slice and dice’ the data to form new views that provide an intuitive feel for trends across the business. Data mining is about discovering previously unknown, and potentially useful information from data.
An interesting data-mining example is from the US Internal Revenue Service (IRS). The IRS aggregated their data with other data from health, motor vehicle registration and various other agencies and went looking for relationships. Amongst other things they concluded that people with personalised number plates are more likely to cheat on their tax return. This has enabled the IRS to better target their audit program.
<...
