When Southern California Edison broke down its IT silos, the goal was to increase the competitiveness not just of IT but of the entire organization. One of the key steps was to replace project-based analytics with an enterprise-wide analytics approach. Using this strategy, the entire company works from the same data sources, data definitions and business assumptions and, therefore, can adapt more quickly to changing market conditions.
“Before the reorganization, we had a siloed reporting approach that resulted in duplication of effort,” says, Matt Peacore, principal manager for enterprise information management at the utility. “Similar types of reports were produced multiple times by many people throughout the company, with different results.”
Now, with an enterprise-wide analytics approach, SCE’s reports are more consistent and accurate, and less effort is expended gathering information. Consequently, employees can be redeployed to more valuable tasks, such as determining what that data really means and applying those learnings to data center projects.
“By deploying an enterprise-wide approach, the organization decided that data and analytics will become a business priority at the executive level,” observes Perry Stoneman, utilities global sector and smart energy services leader at consultancy Capgemini. “This means the C-suite asks macro-level questions, which creates a different level of inquiry.”
From project-based to enterprise-wide
Changing the level of inquiry meant changing managers’ mindsets. Peacore says many SCE managers said they felt a lack of control over their data after an enterprise-wide analytics approach was implemented. “They still wanted to explain or adjust data based upon their specific business unit situations.”
Rather than allowing the disparate data sets that approach entails, SCE insists that managers use the same underlying data for their reports. This ensures, for example, that project costs are defined and measured the same way in all units. “So, while we understand there are special circumstances, we try to eliminate manipulation of the underlying data. This allows us to compare apples to apples across the business,” Peacore explains.
“Sometimes the desire for adjustments (in terms of parameters or units of measurement, for instance) were simply managerial preferences,” Peacore says, but they also often stemmed from differences in the ways business units were held accountable. For example, some units may have been charged for their power consumption, while others were not. As Peacore says, “because we wanted to standardize on one format for our budget and expenditures, we had to hold each area accountable for the same things.”
This approach helps the C-suite look beyond individual projects to see the big picture. “Project-based analytics never become truly embedded in relevant business processes,” Malene Haxholdt, manager of SAS analytics, observes. Instead, “They tend to depend on individuals and manual efforts to realize value. If those individuals leave the company, the analytical project may not be sustained.”
In contrast, an enterprise-wide analytical approach fosters a culture that provides continuous value. “A true enterprise-wide approach uses the same data foundation for multiple analytics initiatives,” Haxholdt says. For instance, “A retail company that uses point of sales (POS) data for coupon analytics also can leverage the same POS data for loss prevention activities by predicting theft, and as the foundation for inventory forecasting and automated replenishment.”
By looking broadly at data across the enterprise, Peacore elaborates, “we’ve found some interesting information about spending habits and deal structures, and so can drive different behaviors.”
Actionable data but pitfalls remain
The greatest pitfall involved with enterprise-wise analytics, Stoneman says, is that data users focus only on their own area. “They fail to look at correlations and, therefore, don’t capture trends. Reports need to embrace transient data to uncover those trends.”
That requires an IT architecture that enables high-performance real time analytics as well as enterprise-wide analytic access. Enterprise-wide analytic approaches tend to be more automated than project-oriented strategies and use more consistent definitions and business rules. Haxholdt says: “An enterprise-wide approach quickly evolves into a lot of predictive models throughout the organization.”
SCE’s Peacore says that the challenge for his power company is to “ensure business units have access to the data they need in a timely manner and to ensure we develop something that meets the needs of everybody.”
When designing an enterprise-wide analytics system, Haxholdt advises including model management capabilities that incorporate documentation and governance guidelines for the use of those analytical models. “Treat analytics as strategic, valuable assets to the company,” she says.
Tags: Data Center,Technology