Keeping Pace

Informatica strives for real-time analysis

Informatica is one of those software houses that has quietly built its reputation by extracting data from other systems, transforming it into some scrubbed and predefined form, and then analyzing it in an application built on top of an analysis engine.

But in the Internet era, the company is moving closer to the pace of events.

"Over the last year, weve made many changes that reflect near-real-time [data analysis]," says Girish Pancha, senior vice president and general manager of Informaticas Platform Business unit.

The emphasis is less on capturing data and more on making the analysis of operational information available to end users as quickly as possible, as befits an era dominated by Internet timeframes.

For example, clothing retailer J. Crew uses Informaticas PowerMart suite for building data warehouses and the applications that analyze the warehouse data. J. Crew captures information from its store sales, catalog sales and online traffic. It can then measure which goods are selling best in each channel. J. Crew can use that information to determine whether Web sales are cannibalizing its usual catalog sales, and in what instances the Web site and catalog complement each other, Pancha says.

Informaticas customers also get a faster way to analyze customer data through its linking software, which connects its data loading and analysis engine and the back-end systems that capture the data.

The company has added an $80,000 to $100,000 option, PowerConnect for IBM MQ Series, to its platform. Messages passed by MQ Series middleware can be received by Informaticas data integration software, PowerCenter 5, which in turn supplies the necessary parts of the message to an analysis application, giving it the data recently captured from transactions and other operational systems, such as purchasing or billing.

Connectors already exist between PowerCenter 5 and PeopleSoft, SAP and Siebel Systems Enterprise Resource Planning applications, which contain customer relationship data. The $200,000 PowerCenter software may also receive eXtensible Markup Language-tagged data or clickstream data and send it to an analytical application.

"When something happens in a transaction system, our platform allows you to understand what that means immediately," Pancha says. Data-mining and data-analysis systems have typically waited for overnight batch updates or several days worth of data before receiving the update to the underlying data warehouse or localized data warehouse, known as a data mart.

Last August, Informatica acquired Zimba, a supplier of data access applications for mobile devices. Later this year, it will integrate that technology with its data analysis platform and offer mobile workers to obtain alerts or summaries from their data analysis systems, as fresh data streams in.

"You set up the rules and thresholds in the analytic application. If such-and-such a transaction occurs, send me an e-mail. Or if this situation develops, send me a Web page, " Pancha says.

The future is personalizing analytical systems to respond to events, notify someone, or even trigger some planned response, Pancha says. "Not only does it alert me through a phone message from a synthesized voice, but the analytical application can take action through instructions to another application," he says.