Trickle-down BI

BI data and analysis are escaping their silo and delivering value to all corners of enterprise operation

BI is shedding its staid reputation as a report-generating tool for an elite squadron of executives and financial managers. Although no one is forsaking the value of those reports for evaluating business goals and forecasting growth, data culled using big BI apps from vendors such as Business Objects, Cognos, and Hyperion -- along with mini-BI applications embedded in other enterprise systems, including ERP and CRM -- are bringing BI to the masses, from line-of-business managers to call-center or support-desk workers.

Results from the InfoWorld Business Intelligence Report 2005 bear this out: During the next year, 70 percent of respondents plan to increase the number of employees who have access to BI solutions.

Rather than BI being used as a separate, disconnected mechanism for analysis -- the ultimate example being CPM (corporate performance management) systems -- BI applications integrated with other apps are delivering the capability to display and interact with BI data in its native form, in real time. A call-center upsell application with an embedded BI app, for example, can predict which of a handful of products would be best-suited for a particular customer based on that customer’s recent transactions and credit history, as well as on the company’s inventory.

Bill Gassman, principal analyst at Gartner, observes, “BI is showing up at deeper levels of the organization. It’s a shift toward directed BI, where you’re guiding people through decisions.” The InfoWorld Business Intelligence Report reveals that 45 percent of companies surveyed now use BI solutions to guide employees through decision-making processes.

Our survey, administered via the Internet between Feb. 24 and March 9, also revealed that 32 percent of respondents think that the most important feature of their companies’ BI solution is prepackaged integration with existing enterprise applications. Analysts expect those numbers to grow.

Click for larger view.

Alaska Airlines is forging ahead toward its goal of getting “business intelligence down to the customer-facing level” by using Siebel Business Analytics, says James Archuleta, director of CRM at Alaska Airlines. The challenge is controlling the flow. “We’re still working on the metadata layer to define what all the business rules are because we don’t want ad hoc BI gone crazy,” he says.

The trickling down of BI is reflected in several terms. Noted expert Keith Gile, principal analyst at Forrester Research, calls it operational BI: analytic functionality built into the procedural interface of an enterprise application -- the screens and applications used by sales personnel, for example -- that makes or recommends decisions for end-users, thereby shrinking operational response time to minutes or seconds.

“We are witnessing a shift away from merely seeing BI as tactical or strategic” at high levels in the organization, Gile says. In some instances, BI apps are embedded to add value to the existing enterprise apps; in other instances, data from BI apps is “surfaced as a meaningful component of the enterprise app,” he says.

By any name, BI’s expanding role is changing work cultures across the enterprise by delivering timely information to frontline workers, typically through customized dashboards, industry observers say (see “Customized Dashboards Deliver,” page 50).

Click for larger view.

Keeping up with the times

The proliferation of operational BI is getting a boost from a rising demand for sophisticated composite applications. Vendors such as Actuate, Cognos, and IBM are focusing efforts on delivering BI services and components to allow them to embed their core technologies within external apps from the likes of Oracle, SAP, and Siebel Systems.

Siebel Business Analytics is one platform that leverages operational BI. “We have Web services interfaces so you can take any analytical result from Siebel Analytics and inject it into a Java application or .Net application, and so on. That lets us display analytical results in another application,” says Paul Rodwick, vice president of marketing for Siebel Business Analytics at Siebel. But aside from that, the secret to success is “rich, deep integration of analytics into [CRM apps] to direct workflow and get more real-time information flowing,” he says.

Application developers are relying on EII (enterprise information integration) software, which provides an abstraction layer over information assets that allows for a single, composite view of data derived from disparate sources. EII technology allows for virtual joining of data from disparate, unrelated sources for the purpose of surfacing information directly from enterprise apps or non-data-warehousing sources.

For example, a BI tool such as Cognos ReportNet uses Composite Software’s Composite Information Server, which allows ReportNet to extract information from data repositories spread across an organization and compile it into a single report.

But it’s important not to play fast and loose with the word “report” when discussing operational BI. Until recently, “BI” and “report” have been joined at the hip. But not everyone needs pages and pages of analysis. “You don’t want to give reports to a call-center person,” Rodwick says. “You want to give them a nugget of information that changes what their screen tells them to do.”

And that information “has to be so simple, my mother can use it,” Forrester’s Gile quips. “Because who do we have in the call centers? Generally, they aren’t trained in BI, nor should they have to be.”

Rodwick suggests that operational BI turns the venerable software category on its head. “Most traditional BI solutions are about driving masses of reports to people who have to go through them, figure out what interests them, and then decide what action to take,” he says. “The major trend for BI [applications] is to drive select information to individuals when they need it, be sure that it’s already relevant to what they have to accomplish, and then guide them toward what action to take.”

IT rolls up its sleeves

For IT, operational BI requires an exhaustive understanding of the requirements of workers in addition to myriad business processes. “It all goes wrong if IT makes the wrong decisions about what to provide users,” says Michael Corcoran, vice president and chief communications officer at Information Builders. His company recently unveiled WebFocus 7, a product designed to push BI to the operational level and to tailor it accordingly.

“We’re no longer talking about power users, so when IT pulls together its test group, it has to be absolutely sure [the members] represent the needs and skill level” of users in the real world, Corcoran says. Software to serve this diverse group also needs to be capable of pulling in data from across the enterprise. WebFocus 7 provides native access to more than 200 data sources and data formats, including relational and legacy data. WebFocus 7, which is scheduled to ship on April 25, strengthens ties with connectivity software from Information Builders subsidiary iWay Software, a vendor of application adapters.

Click for larger view.

Gartner’s Gassman urges IT to pay careful attention to how data from disparate systems is captured and integrated. “This is one of the inhibitors” of embedded BI, he says. “If you say, ‘I’m going to buy an analytic application from XYZ Company and use it to pull data together from Siebel and SAP to make wonderful decisions,’ immediately you end up with a data-quality issue. If you just throw data from operational systems together and the data’s not accurate, you’ll start to run into trouble when people make decisions based on Canadian dollars instead of U.S. dollars.”

Click for larger view.

In a recent report on best practices for developing BI applications for broad use, Forrester’s Gile says IT should approach rollouts by separating technical functionality -- such as OLAP, visualization, or predictive analytics -- from a BI application’s roster of functional requirements -- such as customer segmentation, supply-chain optimization, or forecasting.

This simplifies the process of targeting the appropriate organizational layer. Where employees are characterized by minimal training and maximal turnover, the report continues, analytic functionality built into the procedural UI goes a long way in assisting operations such as customer service, shipping, and manufacturing. It also reflects a gradual shift toward putting relevant data into the hands of more workers.

That evolution may be gradual, but the cumulative effect on BI and its expanded range of users is dramatic. “It’s changing whole cultures, but I think it’s a positive change,” says Rich Clayton, vice president of product marketing at Hyperion, a BI platform vendor and BPM software provider. “With all this accessible data comes a tremendous amount of clarity. When a salesperson knows that the manager can look at day sales outstanding, down to the minute, a whole different mind-set develops. That kind of visibility is a powerful development.”

Copyright © 2005 IDG Communications, Inc.

How to choose a low-code development platform