Scorecards, dashboards are two different things

American Suzuki Motor Corp., Brea, Calif., serves its customers through a network of 1,700 dealerships in 49 states. Whether it's a motorbike, an automobile, an outboard engine, or any other Suzuki product, Suzuki's promise is the same: if there's a problem with the product during the warranty period, take it back to the dealer, and the dealer will fix it.

It's good customer service, but it's also a hostage to fortune. If the dealer does a lousy job or takes too long to get Suzuki's approval to undertake a warranty repair, then it's Suzuki's name that suffers. But giving too much autonomy to dealers—to self-authorize repairs, for example—is like signing a blank check: fix what you like, and just send Suzuki the bill. This potential combination of expensive repairs, tight sales margins, and requirements for rapid response form a particularly unholy trio.

It's a dilemma—and to resolve it, American Suzuki turned to a business intelligence (BI) solution from Hyperion, Sunnyvale, Calif. Called the Self-Authorization Project by American Suzuki, the solution feeds data relating to warranty claims into an IBM iSeries server running a DB2 database. SPSS, analytical software, then replicates that data into a cube, which Hyperion Intelligence—a solution Hyperion acquired when it bought out former partner Brio in July 2003—then analyzes and presents the results in a dashboard format.

Key metrics include warranty repair costs by vehicle and dealer, defects per vehicle, total claim analysis by region and district, duplicate repairs by vehicle (the number of times the same fault is fixed), and externally sourced customer satisfaction data. To gain self-authorization of warranty repairs, says Suzuki Data Warehouse Specialist Claire Ashby, a dealer must meet a certain number of criteria—including speed with which they undertake repairs, accuracy and timeliness of their warranty billing process, and extent to which a fault stays fixed.

"It's not exactly the kind of dashboard that people were talking about a few years back, but it offers the same functionality," says Ashby. "It collects information, presents it on a single screen, and allows drill down for detail on specific issues."

A traditional dashboard has all the needed information ready to go, concedes Srikant Gokulnatha, Hyperion director of product marketing and strategy. "The dashboard in use at Suzuki has a richer ability for drill downs, to see data in different contexts. It lets end users customize the data view based on their own requirements."

And the benefits are clear. Customer service improved, potential for over-charges were minimized, productivity of head office employees was boosted, and—best of all—there's the prospect of increased market share, estimated at 0.25 percent. What's more, as Suzuki implements other BI projects, cross-analyses become possible. For example, explains Ashby, a subsequent project to monitor dealer performance permits Suzuki to look at under-performing dealers and examine how well their warranty claims have been processed.

Welcome to the brave new world of BI—a class of application that crashed MSI's Top 100 big time this year. But why? Some say manufacturers are waking up to a new generation of user-friendly features such as dashboards, scorecards, and 'off-the-shelf' analytics that deliver greater insight into business performance than ERP systems can. And do it more quickly and across a wider range of activities than ever before.

At Zarlink Semiconductor, a manufacturer of specialist semiconductors headquartered in Ottawa, Ontario, CIO Bill O'Connor says he's generated significant savings by replacing just about every report coming out of its SAP enterprise system with equivalent—or better—information from a BI solution.

The impetus, he explains, was the severe business downturn beginning in 2001. Three years ago, says O'Connor, his budget was $20 million annually, and he had a staff of 112—35 of whom were focused on SAP. Now he has just 34 people total, only one of whom is assigned to SAP. The IT budget runs $5.5 million annually. SAP is the transaction engine—but the business is managed using technology from Cognos.

Beginning with billings and backlog, and then moving to expense reporting, headcount, compensation, forecasting, accounts receivable, and cash, the old reporting system was replaced with reports from Cognos' Visualizer tool. "Users determine what information they want to see, the reports are graphical and interactive, and management can drill down as they like," says O'Connor. There's also a scorecard, containing what O'Connor describes as "20 or so comprehensive reports that management can cover in about 10 minutes."

Over time, the nature of the engagement with Cognos has changed. "Once we'd built up credibility with the users, it moved from a push to a 'pull' process, with users asking for more and more reports, says O'Connor."

Once purely the province of power users, BI clearly has gone mainstream. "BI applications were bought in dribs and drabs, one or two at a time, across departments and functions," says Patrick Morrissey, senior director of enterprise performance management at San Jose, Calif.-based Business Objects. "We're now seeing deal sizes getting bigger, and the purchase decision is becoming board level."

While user-friendliness and pre-built analytics play a role, Morrissey also points to BI applications' role in joining disparate data sources and systems—especially the latter. The greater the number of systems from which meaningful inferences must be drawn, "the more likely a BI application is the right way forward."

"BI provides an overlying harmonization layer," says Morrissey. "Take determining compliance with service level agreements for example: was the delivery time met, or was it not? Without that cross-supply-chain harmonization, it can be difficult to tell."

But harmonizing what, exactly? While monitoring metrics might be informative, it's not necessarily going to maximize insight. No matter how user-friendly or pretty the dashboard, if it's not showing the right metrics, inferences will be lost. Effective BI, in short, requires—well, intelligence.

According to Paul Hoy, director of manufacturing solutions at Cognos, effective BI is a three-stage process. "First, plan the performance in terms of deciding objectives and how these objectives support the corporation. Second, decide which metrics determine progress towards objectives. Third, use BI to understand performance achieved."

Although the terms 'scorecards' and 'dashboards' often are used interchangeably, says Hoy, for BI professionals their meanings are very different: "Scorecarding is the process of comparing actuals to targets and plotting trends over time. Dashboarding is a digital cockpit, if you will. If you want to see a snapshot for a moment in time, and then drill down, you use a dashboard. If you want to see a trend, and then drill down, you use a scorecard."

It's important that a BI application sticks to delivering intelligence, believes Anant Jhingram, BI director at IBM's Silicon Valley Laboratory in San Jose, Calif. Jhingram says it isn't the role of a BI application to issue alerts or escalate problems. "BI isn't a front- line application," he says. "It's role is to supply what's been missing in the past—monitoring and analyzing those other applications."

Trouble is, 'those applications' increasingly regard BI as part of what they do. The result, concedes John Hagerty, a VP at AMR Research, Boston, Mass, is confusion. "There's a proliferation of dashboards and scorecards," says Hagerty. "As a result, there's tension. Which do I use? Do I use several, but at different layers and for different purposes? There's no one right answer; a lot depends on individual circumstances."

Take Sunnyvale, Calif.-based Network Appliance Corp., the $1.2 billion manufacturer of data storage devices, which uses BI capabilities in what typically would be considered a planning and scheduling tool: Rapid Response from Webplan, Ottawa, Ontario. According to John Sicard, a Webplan VP, Rapid Response actually has two BI tools.

The first, he explains, is a 'resolution engine,' so companies see the consequences of scheduling decisions for inventory, customer service levels, capacity utilization, and margin. The second is a 'live scorecard'—capable of incorporating top-level business objectives, extracted by Metric Manager from Cognos, a Webplan business partner—which highlights decision results in scorecard format. The crucial difference: BI provides an historic view of the consequences of decisions; Webplan's view is forward looking.

Network Appliance, says IT Director Michael Peters, uses Webplan primarily to assess the impact of sourcing decisions. When the company only had one manufacturing facility in California, sourcing wasn't an issue, he says. But these days, with contract manufacturers in the U.S. and overseas, and a second facility in Scotland, life is more complicated—especially when products from different facilities need to be shipped so as to arrive together.

Thanks to information provided by Webplan, the consequences of any sourcing decision and the impact on the company's top-level metrics are clearer, and—given Rapid Response's simulation capabilities—can be discerned before the decision actually is made.