Ask any corporate executive the secret to success and there's a good chance you'll hear that the management team must work on the same goals using the same assumptions and facts -- with transparency to the CEO and board. Yet most companies don't work that way, setting the stage for missed opportunities, hidden problems, and political gaming that uses siloed data and its analysis as a departmental weapon rather than as an enterprise asset.
One way to get closer to the goal of a company that knows how it's really doing -- which means being fully transparent and working toward the same goals with the same assumptions and data -- is through the deployment of a CPM (corporate performance management) system. Such a system analyzes data related to business performance and presents it as a collection of KPIs (key performance indicators) aggregated into a "dashboard" of simple graphical readouts, using metaphors such as stoplights and gas gauges. From the dashboard, managers can drill down to study performance data in detail.
Performance can be anything important to company strategy: employee turnover and satisfaction; defect reduction in manufacturing; net profits on new business or product lines against expectations; and so on. What's key is that the company is measuring performance -- not merely process statistics -- and is gauging it against such metrics as goals, past performance, and competitive baselines. Furthermore, to understand the complete picture, senior managers are looking at these performance areas across the entire company, not just within specific functions or departments. Typically, department managers use CPM tools within their areas, but a CPM system "rolls up" information to give senior managers a broader view of enterprise performance.
"It helps you ask the right questions and validates that the data is based on a single version of the truth," says Graham Mackintosh, vice president of strategy and business development at BI provider Cognos.
Effective CPM systems also gather and crunch data in a more targeted way than conventional BI systems do, meaning that "we don't have to say, 'Gee, I wish we'd done this two weeks ago,' " says Dan McGowan, vice president of financial reporting and analysis at Southeast Corporate Federal Credit Union, which provides financial processing services to other credit unions.
Along with helping an enterprise manage itself more effectively, CPM can also help it meet regulatory requirements -- such as Sarbanes-Oxley -- that require the CEO and other senior managers to sign off on the truth of their disclosures and to disclose "material events" within 48 hours. "How can you report to what the laws require if you don't know what is happening?" asks Jonathan Hornby, practice director of performance management at analytics provider SAS.
Known by several names -- business performance management, corporate business management, enterprise performance management, and, in some circles, real-time BI -- CPM requires a strong underlying data architecture. That requirement is one reason most CPM providers are also BI, data-analytics, or ERP vendors, such as Applix, BusinessObjects, Cognos, Computer Associates' CleverPath division, CorVu, Hyperion, Lawson Software, Outlooksoft, SAS, and SRC Software. (Many other vendors use one of the "performance management" labels but offer analysis systems only for specific tasks such as financial, IT, or supplier management.)
Only approximately 20 percent of companies have both the technology infrastructure and the commitment to transparency necessary for a successful CPM deployment, SAS' Hornby says. And as Meta Group Analyst Jonathan Poe notes, perhaps only 5 percent to 10 percent of large companies have been implementing CPM for a long time, often using homegrown tools for tapping into and analyzing their data stores.
"A lot of companies don't have the formally defined processes or a common set of metrics across the organization," adds Kevin McAuliffe, director of strategy and business performance management for the software group at IBM.
Building on a Solid Foundation
Typically, a CPM deployment builds on existing data repositories, data integration efforts, and departmental systems such as ERP, CRM, and SCM. "CPM is not a revolutionary way to reinvent your technology infrastructure," says John Colbert, vice president of service management at BPM Partners.
Instead, it usually overlays those systems and is used for gathering information across them by breaking organizational silos, says John O'Rourke, senior director of product marketing at BI provider Hyperion.
"Organizations have done a great job and siloed information. CPM takes advantage of all of these," BPM Partners' Colbert adds.
To consider a CPM deployment, organizations that haven't completed at least basic integration of key financial and operations data need to build those systems first, identifying the key data, what it means, where it resides, and how to access it. "The majority of companies have the data somewhere," Colbert says. "But the ones with financial systems and transactional systems in place are the ones that can take advantage of it."
For most others, performance data tends to exist in individuals' Excel spreadsheets, notes Mike Rost, director of product marketing at ERP provider Lawson Software.
To deploy CPM, organizations need some sort of central data repository -- one or more data warehouses, data marts, cubes, or databases with a common data architecture -- that can be used as a trusted, audited source of truth, Rost says. That's not to say that all the data must reside in one data store. Regardless, companies must know where key, validated data is and that the relevant, agreed-on metadata, business rules, and metrics are accessible to the CPM system. (Companies also have to have already integrated all or much of that data; otherwise, they won't know whether what they have is relevant, let alone accessible, Colbert notes.)
For example, product lifecycle management provider UGS uses Hyperion Essbase data cubes; utility Baltimore Gas & Electric (BG&E) pulls its data from an Oracle Financial database; credit-union services provider Southeast Corporate Federal Credit Union uses a set of Applix TM1 data cubes; and hospital group Centra Health accesses a set of SAS applications that have a common data architecture. Except for Centra, these organizations pull their data from various transaction and departmental systems, "scrubbing" the data for accuracy and consistency and placing it in validated data stores. Centra has the advantage of already having had to meet regulations that require integrated, consistent data systems, says Kim Price, decision support manager at Centra.
In some cases, organizations can tap into real-time transaction systems rather than use intermediate data stores, as long as they know the data is valid and under what circumstances it's OK for managers to see KPIs based on incomplete or unverified data, IBM's McAuliffe says. He also concedes that managers need to be wary of spending too much time tracking minor variations in unaudited, real-time data.
Sorting the Essential From the Tangential
After an enterprise has a sufficiently robust data infrastructure in place, the big challenge becomes figuring out what data is meaningful, trustworthy, and relevant. Ditto on the metrics used to contextualize that data.
For example, in its CPM deployment, UGS found that, although it could get lots of data from its SAP order-entry and service-billing systems and its homegrown CRM system, some vital data just didn't exist in the SAP data structures, such as revenue from new customers, recalls Eric Kline, director of IT applications at UGS. So, the company had to develop logic rules to create this information from other data. "Additional fields may need to be created or captured" in existing systems, he notes.
"We really had to work through what the targets and benchmarks are," Centra's Price notes. That process led to questions such as "If you are at 99 percent [of a target], is that a negative? Is it 100 percent all the time?" she recalls of the hospital group's SAS Strategic Performance Manager-based CPM effort. "It's also hard to choose the things where we don't look so good," despite the fact that targeting those areas usually leads to performance improvements, Price says.
A CPM effort can also expose -- or be torpedoed by -- internal politics. Because a truly effective CPM system allows no one to hide their performance or to maintain silos of proprietary information, accessing the right data and metrics can require navigating difficult political issues. Board members, senior managers, and other authorized users suddenly have access to the same information at the same time. Because of this, report massaging goes away -- as it must for true transparency. Meta Group's Poe suggests that IT staff are well-suited to navigate these waters because they "know how to dissect the unknown -- they're used to change and [to] being transparent from having to share with others on a team." But strong executive support is ultimately required to ensure real sharing, Lawson's Rost says. Otherwise, "you'll find some people trying to protect their turf," UGS' Kline adds.
More Analysis, Less IT
After management has bought into a CPM effort, enterprises often discover that proper data discovery and analysis requires a strong partnership between IT and business staff -- and that effort can easily take 60 percent to 80 percent of the deployment time, Meta Group's Poe notes. "Don't rush to code," he advises.
Otherwise, "the garbage [data] just comes out a lot faster."
Sometimes companies use existing metrics such as the Six Sigma defects analysis often used in manufacturing or the Balanced Scorecard approach that weighs financial performance, customer trends, internal processes, and internal education, notes Tony White, an analyst at Gartner. But because such metrics are usually part of applications with hard-coded assumptions, "IT needs to investigate these, which can be a real headache," he says.
Another challenge is finding the right predictive metrics, Southeast Corporate's McGowan says. "Regression analysis works quite well in some areas, and in others, it doesn't," he adds.
Such work requires different skills than many IT organizations have, Gartner's White says. That's what BG&E found in its deployment of an Outlooksoft Everest CPM system, which initially targeted operations, management, and capital planning across 110 project groups, pulling data mainly from an Oracle Financial system. The company added a second functional analyst and has increased the workload of both the IT and functional staff, says Tom Lowe, director of business strategy at BG&E.
And analysts and providers alike agree that the role of business staff must be greater in a CPM deployment than in most IT efforts. "You really need the subject-matter experts empowered and well-trained on the tools," Southeast Corporate's McGowan says.
Jim O'Connor, OLAP systems supervisor at cable provider Adelphia Communications, says he's noted that the transition to CPM is harder for traditional IT people -- especially database administrators -- than it is for business staff. For example, as Adelphia continues its deployment of a Clarity-based CPM system, O'Connor sees his staff's role changing from generating reports and managing databases to developing the underlying data architecture and training business staff how to use CPM-based tools.
This shift can also reduce the need for some IT staff: As the CPM effort broadens across the company, O'Connor expects to see significant overall savings, given that the need to create reports will diminish. After all, as Bill O'Connor, CIO of semiconductor company Zarlink, says, "You only do the report once" in a CPM system.
Phasing in CPM
According to analysts and executives, deployment of a CPM system should be done in stages. "Don't bite off the whole thing," Hyperion's O'Rourke advises. "You need a clear decision on what metrics you're going to start with -- and not try to do everything for everybody," says Sherman Mink, manager of financial and reporting applications at UGS. "It's an iterative process," Zarlink's O'Connor concurs.
Most organizations start with financial performance management because most already have the data and metrics in place and are trying to speed up the reporting of their financial reports, BPM Partners' Colbert says. Next are operational performance issues such as reducing service calls or improving distribution, he says. That's because those areas are more easily quantified and often use data-based metrics for process improvement and auditing, he adds.
Often, new data streams will be needed to assess performance in areas such as marketing and human resources, where most companies haven't gathered meaningful data consistently, Colbert says.
It's also important to target your indicators and metrics. "You can't manage a thousand indicators," Price observes.
"Most organizations have too much information and tend to get into firefighting because of that," SAS' Hornby adds. "Each individual should be looking at 10 to 20 metrics, all in alignment with the company's goals," he says. Otherwise, they're just looking at a dashboard all day, rather than thinking about and acting on the business performance issues that CPM is supposed to illuminate in the first place.