TDWI | Training & Research | Business Intelligence, Analytics, Big Data, Data Warehousing

Business Intelligence and Performance Management Slowly, but Surely, Converge

While a growing number of organizations have gotten hip to the promise of PM, many are still in the PM iron age

• By Stephen Swoyer
• November 8, 2006

Before organizations take up performance dashboards and performance scorecards—and before the largely discrete practices of business intelligence (BI) and performance management (PM) formally converge—executives must first get hip to the promise of PM itself.

And while TDWI research indicates that a growing number of organizations are doing just that, implementing dashboards and scorecards, then linking them to two or more data sources—many outfits are still in the PM iron age. For these shops, PM—such as it is—typically takes the form of shot-in-the-dark ad hoc querying, multidimensional analysis (typically via Excel), and other, more sophisticated variations on straight-up operational reporting.

That’s been the experience of Patrick Hustings, president of BI consultancy Extended Results Inc., who says his clients are mostly leveraging familiar tools as they wade into the practice of performance management. “The tools range from simple Excel spreadsheets to portal solutions utilizing Microsoft SharePoint Portal and Microsoft suite of business intelligence tools like Business Scorecard Manager and ProClarity,” he comments.

And while mature PM practices are premised on finely tuned—and often highly idiosyncratic metrics or key performance indicators (KPI)—most performance management tyros start with what amount to KPI best practices, Hustings says that the measurements companies are using tend to be “typical operational KPIs: employee training globally, campaign effectiveness, customer satisfaction, and financial objectives like territory revenue vs. goals, etc.”

To a degree, many BI pros say, PM has increased in both visibility and performance as prominent mainstream vendors—such as Microsoft Corp. and Oracle Corp., among others—have pushed into the PM segment. Consider Jéjé Poudou, an R&D director with Merkurium, a Montreal, Canada-based BI ISV, who says his company has gotten more serious about PM as its own primary solution provider, Microsoft, has fleshed out its own offerings in that area. “Our main product is based on Microsoft technologies [SQL Server 2005 and SharePoint 2003]… [and] our main market is the education sector where performance management has become important,” he continues, adding that “performance management… has become more and more important.”

Enter Merkurium’s bread-and-butter product, Skopus, which exposes a dashboard interface and leverages SQL Server 2005 and SQL Server Analysis Services. “[W]e push KPIs to our end users, predefined reports and interactive dashboards. We use Sharepoint and custom templates to provide an easy-to-use interface,” Poudou explains. “Identifying KPIs and metrics is a long process due to the particular context and complexity of the education. Its’ a big challenge for us, but only close work with our clients can provide good results.”

Nor is that all, he indicates. SQL Server 2005 brings a lot more to the table in terms of connectivity—which facilitates integration and consumption of a wider variety of data sources. That’s the easy part, Poudou comments. The trick, not surprisingly, is identifying and formalizing KPIs: “Integrating multiple source is the first issue from our clients, but—most of the time—this is only a small part. The biggest part of the job is to create KPIs in front of this data and give [users] access to the source data.”

Merkurium’s customers typically interact with two types of data sources: RDBMSes, for one, and flat files, Poudou says. “The transformations required to manipulate these sources are small, and the usage of unique IDs helps us.”

In many organizations, PM has at least peripheral mindshare—i.e., it’s on the radar screen, if only as a distant bogey—but hasn’t yet gained much traction. In many cases, organizations are still fleshing out their BI infrastructures and beefing up their reporting capabilities. That’s the lot of Johan Koopmans, a BI pro with a Dutch provider of online travel services. “We're solely busy with reports. In the long term, PM might be on the radar,” Koopmans confirms.

He says there are salient differences between the two disciplines, and that a kind of ad hoc PM is already taking place in his organization. “Business intelligence is all about getting the right information to the right person at the right moment. Showing metric performance is about the same,” he indicates.

On the other hand, Koopmans says, PM seems to have more of a feedback function—as a kind of corrective, steering mechanism—into the business process itself. He says it’s this aspect of PM that’s most interesting to him. “The current reports are abstracts [aggregations] of data acquired from processes within the back office. So the current data is not solely used for BI, but also within the business processes. In this way I see a difference,” he indicates.

BI consultant Hustings, for his part, says PM effectively starts with—or departs from—BI. In most cases, he says he tries to leverage his client’s existing in-house BI assets—including, significantly, unmodified reports—as he helps them build out their PM practices.

“In all these projects, customers have literally hundreds of reports either in Crystal Reports or Reporting Services. We are not rewriting the reports. We are taking those reports, summarizing them into Objectives and KPIs, and [formatting them] into a Scorecard. That allows us to focus the organization or teams around common objectives,” he explains. “From that, they can drill down into KPIs that are Yellow or Red for causal analysis. Then, if they need to go deeper, they can drill down to the transactional reports. Basically, Scorecarding allows us to tell a story around an Objective and KPI and lead the end-user down a certain path to discover the issues. This allows us to build trust in the numbers and a common understanding of the data.”

Merkurium’s Poudou concurs: “A unified access and common vocabulary is a key point, so measuring performance without a good BI background or back-end is not possible; it’s like making decisions based on [one set of] data and really measuring another [set of] data!” he concludes. “It’s not possible—[even if it is] possible, it will provide bad results and lead to bad decisions.”

More to the point, Poudou says PM is a complement to, not as a replacement for, BI. In this respect, he echoes Koopmans’ notion of PM as a corrective feedback loop into the success or failure of an organization’s BI practice. “PM is required in front of the BI tools to ensure that the decisions and strategies are effectively pushed [to] and understood [by] the end users,” he argues. “So only a combination of [both will] ensure a successful deployment.”

More to the point, Poudou says, Merkurium hopes to encourage its customers to embrace PM as a means to transition to a proactive kind of management that simply isn’t possible using conventional BI tools. “We have to become more proactive and less reactive. We really want to improve the student success and [not have to wait] for a course failure to say ‘This student is in trouble.’ This is not an option for a good educational solution,” he confirms.