"It's about re-architecting BI and data warehouse (DW) environments that are very inefficient," said William McKnight. "Companies are looking whether they need real-time DW, real-time operational data stores, business activity monitoring, and exactly what needs to change in the current extract, transform, and load (ETL) cycle to accommodate 'right' time. Strategic decisions won't be made on intra-day information, but companies do want to monitor product movement, detect fraud, and enable call centers to make the right offer at the right time."
We discussed how this sort of effort could be complicated by the expansion of BI into more operational contexts, a trend that is diversifying the user base, as well as plausible definitions of BI's credo of "the right information, at the right time, in the right hands." Colin White offered many insightful comments about BI's increasing process orientation. "The dividing line between what is operational and what is BI is getting fuzzy," Colin said. "More and more companies want to use BI as part of their day-to-day operations. I'm finding that the group in charge of business transactions is also in charge of putting in the BI/analytics stuff from the transaction application vendor — that is, a different group from the DW group. There's a political battle going on in terms of how we integrate the two camps."
Our panel considered how some companies are choosing not to mess with these BI-transaction application packages, viewing them essentially as data marts that may be tapped for cross-business or cross-functional analytics — perhaps through new enterprise information integration (EII) software — and that this may be the ultimate role of the enterprise DW (EDW). "However, while each business or function can justify its own project, nobody wants to pay for cross systems," Colin said. "The only way to get that is to go up to the CEO level. This is where performance management comes in. It's getting much easier to say, 'okay, let's create key performance indicators across business areas and get that project funded without interfering with business-area budgets. If you can start there with the EDW, you may be able to convince the CEO to fund more cross-business analytics."
The real issue behind consolidation and even the extension of BI to a wider user base is return on investment (ROI). IT concerns often remain focused on total cost of ownership; one attendee offered an anecdote about a CIO who "didn't care" about compliance because it was the CFO's problem. His focus was on taking cost out of IT. However, as expensive BI projects and infrastructure gain a higher strategic profile, business executives are demanding a better ROI assessment. Major vendors, including Business Objects, Cognos, Hyperion, and Information Builders have been talking ROI during 2004. We'll hear more in 2005, including how to evaluate ROI against maturity scales that describe technology adoption rates and what ROI the organization should expect at each stage.
"ROI is revenue generated and money saved," said Jill Dyche. "What I'm seeing is the human cost savings BI and DW bring by eradicating the data location, gathering, cleansing, and dispersal cost. You can quantify the human labor savings; you don't have just the soft ROI of employee retention and happier customers." Jill said this was important because "as the VP of marketing, for example, moves into predictive modeling, he or she asks, "'what's it going to buy us?' There has to be a financial story before acquiring the data. This is increasingly a business conversation. Is it worth it? If not, then the ROI isn't there."
Thus, "business" will increasingly be the initiating half of the BI acronym. I'll have to leave it there for now. Thanks to Scott Humphrey for putting on a great event!
David Stodder is editorial director and editor-in-chief of Intelligent Enterprise.