SAN FRANCISCO -- One key to a successful CRM analytics deployment lies in providing information to your sales department -- more information than they know what to do with, according to a speaker at Oracle OpenWorld.
"By flooding them with too much information, they want to see less and less," said Bill Duffy, data warehouse project manager with Stamford, Conn.-based Pitney Bowes, during a session at the 42,000-strong user conference being held here this week. "They really need to be focused."
Duffy is a veteran of multiple Siebel implementations at the postage meter manufacturer, including a move to Siebel Analytics 7.8.4 early last year. He shared his experiences, along with some tips and pitfalls at the conference.
Since 2002, Pitney Bowes has deployed Siebel Marketing, two Sales modules, Call Center analytics and four Field Service deployments for its 1,500 field service agents. A $5.5 billion company, Pitney Bowes first turned to Siebel and its marketing technology to segment its customers. Siebel analytics now essentially runs the sales organization, Duffy said.
"Now executives can't get enough of it and can't be without it," he said.
Pitney Bowes' significant investment in Siebel has not been without its problems, however. Before the company upgraded to Analytics 7.8.4, the extract transform and load process took 10 hours each night, Duffy said. After the upgrade, that process takes just a few hours.
"This is really because Siebel realized the first process was really rotten," Duffy said. "It's 100% better. Precached reports are really the only way to fly."
Before putting analytics in place, Pitney Bowes had generated lengthy, cumbersome paper reports that provided little actionable information, according to Duffy. When IT began deploying Siebel Analytics and issuing reports, it opened the floodgates of information. One resource was developing 400 reports, Duffy said.
"We have to foist ourselves on these sales folks," he said. "They key was limited ongoing support."
Overwhelmed with information and without extensive support from IT, executives began to determine what information they really needed on their own. Now, they rely on, at most, 12 reports.
During the deployment, it was vital for the analytics team to work with the CRM side of the house.
"Analytics is not the same as CRM," Duffy said. "We don't care about the transactions, we care about the information. CRM will resist BI involvement. They don't think about information."
CRM stakeholders have to be involved with the data warehouse people because all changes to CRM affect the data warehouse, he warned.
As with so many IT projects, executive sponsorship is critical. Additionally, companies should avoid customization as much as possible.
"Vanilla really works," Duffy said. "Pitney Bowes is really Neapolitan, but some of our applications are close to vanilla, and they make analytics a breeze. If you can live with it out of a box, live with it."
Managing expectations is also critical. Analytics can only do so much, and people will ask for the world if they think they can get it, Duffy warned.
Organizations should connect directly to legacy applications wherever they can, rather than trying to replicate them.
Duffy is a big proponent of extending analytics to the enterprise and would in fact one day like to see everyone in the company with analytics on their desktop. Yet, when deploying analytics it's important to put controls on users. For example, you don't want impatient users running the same queries on their own that take 12 hours and bring the system down. Pitney Bowes has managed to stem that by developing analytic dashboards that alleviate that need.
Duffy also suggested keeping Siebel, or now Oracle, in the process as much as possible, using its metadata model, attending every class it offers and engaging Oracle's expert services.
"Make them look over your shoulder," he said. "Keep their skin in the game. There's a lot of integrators out there."
Companies deploying analytics should pick integrators carefully, he warned. Some are better than others, and who is good and who is not often changes because expertise is so in demand.
Finally, wait on the upgrades, bringing analytics up after the other systems.
"I am always behind," Duffy said. "To do it at the same time is crazy. That's why we were successful with our sales implementation."