IBM has unveiled three packages of services and software to help organisations analyse their data for profit and improved efficiency.
The signature solutions, as IBM calls these offerings, go beyond generic analysis software to address three different specific tasks: detecting financial fraud, predicting consumer behaviour and estimating financial risk.
"Having software is important but having industry expertise and domain knowledge is also pretty essential," said Deepak Advani, IBM vice president of predictive analytics. IBM's intent behind these packages is to combine its analytic software with the lessons it has learned installing such software for clients, Advani said.
In its work of deploying systems, the IBM services arm sees a lot of the same challenges, or patterns, across different clients, Advani explained. These packages will utilise what IBM has learned deploying such systems. It also draws from the company's considerable research expertise and IBM software, such as SPSS, Cognos, Clarity and WebSphere.
Potential customers will start with a workshop to help identify which datasets they have that can be better utilised. IBM will then work with each customer to develop an analysis system, using as much of the customer's existing systems as possible. The cost for each customer will vary depending on implementation.
IBM developed the fraud detection package to help insurance agencies, health care companies and government agencies detect phony claims before they are paid out. Typically such organisations have practices in place to identify cases of fraud, but they identify the misbehaviour only after the claims are paid. An IBM system could be built that recognise the subtle hints of fraud, based on analysis of historical data, Advani said.
Another package analyses consumer actions in order to predict buying habits and other behaviours. Such a system could provide what Advani calls "next best actions" or the next behaviour a consumer might take, based on prior actions. Telecommunication companies, for instance, could use such a system to predict when a customer might drop a service, which would allow the company to make a counter-offer to keep the customer.
The third package addresses financial risk. This system is suited to chief financial officers and other financial executives. Such a system would use past performance and key metrics to predict future performance. A chief financial officer, for instance, could use the system to predict how a 10% dip in sales would affect the organisation's finances as a whole.
Beyond these three packages, IBM is also developing a number of other customised analysis offerings, though Advani did not divulge what they might be. Overall, however, the company's work is addressing a potentially huge market. IDC estimates enterprises will spend over $120 billion by 2015 on analysis systems. IBM estimates that it will reap $16 billion in business analytics revenue by 2015.