Stamp of Approval for Business Intelligence
Continued
Has it been successful? Judging from user demand, yes. "We went in with 300 management licenses and have been authorized to buy 300 more for different levels of management," says Duffy. "We need this information in the hands of a lot more people."
From there, the company added call center analytics in October 2005, as well as an upgrade to Siebel Analytics 7.8.4 in early 2006. The upgrade helped Duffy streamline IT chores. For example, before the upgrade, the extract, transform, and load (ETL) process took 10 hours each night, leaving no time to produce user-requested reports overnight. Now, ETL runs in under 3 hours, and reports are ready for users by 6 a.m.
The Long View
Duffy has some challenges to conquer, however. One is the ongoing data ownership struggle between the BI data warehouse group and the conventional transactional development teams. The latter may resist BI involvement or lack confidence in the BI data, or they may simply not understand how data integration might work. For example, Duffy would like to be able to combine the 10 to 15 legacy application screens that currently populate desktops at call centers with one integrated analytics screen that connects to each separate warehouse. "It's a long-term goal," he says. "We are actually able to do this now, but there's some paranoia on the part of the CRM [customer relationship management] people not to do it. They want to keep the data in their database, but my attitude is, 'Why bother? It's sitting here ready for them now.'" He hopes in the long term to forge an alliance with the CRM team, providing users with one technical point of contact. In the end, doing so will help further the Pitney Bowes vision of business intelligence as a process that should be almost invisible to users. "It's more action orientedanalytics can drive how a business responds to a particular thing in the market," says Duffy.
Duffy also points to several important lessons learned in his long analytics odyssey. The first is to engage strong and committed project sponsors. "Continued executive sponsorship is really critical," he says, something that DecisionPath Consulting's Williams agrees with wholeheartedly. "In my experience, the odds of success go up dramatically when companies have senior-level business management sponsoring the project," she says. Next, implement with as little customization as possible. "Vanilla really works, and it makes life simpler when it comes to upgrades or integration," says Duffy. He also recommends making sure users have realistic expectations as to what BI can and cannot do, integrating with legacy systems whenever possible and understanding that analytics support is an ongoing process.
Duffy is eying further integration with CRM in the future, as well as working to extend the reach of the Siebel analytics tools to a much wider user base. Further proof of the effectiveness of the Pitney Bowes implementation? Duffy needs a single support person to cover the entire analytics deployment. "At one point, we delivered more than 400 reports to a large customer base with only one person," he states. "That's cost-effective."
For More Information
Oracle for Industrial Manufacturing
Oracle's Siebel
Carol Hildebrand is a freelance business and technology writer based in Massachusetts.