Intelligent Enterprise

Better Insight for Business Decisions

RSS
Webcasts
Whitepapers
Subscribe
Home

April 22, 2003

The BPM Drumbeat

Will business performance management live up to all the promises?

by Seth Grimes

Consultants, analysts, and writers (contributors to this magazine among them) are beating the drum for business performance management (BPM). The word is that BPM will transform how organizations — or at least those we refer to as intelligent enterprises — operate. Competitiveness supposedly requires us to pervasively measure and monitor performance and apply defect-reduction and process-optimization methodologies. We'll create systems that integrate data from disparate, distributed sources, compute key performance indicators (KPIs), and provide dashboard-style displays with "actionable" information. And let's not forget the ability to analyze historic results, forecast and simulate, and provide views tailored to individuals at a variety of levels in the management hierarchy.

BPM is being marketed as the be-all and end-all of business analytics! It's a tall order, and one that I believe is being oversold.

Methodology and Madness

We have here a massive case of "scope creep." The success of performance-management and related methodologies in important but limited contexts has led to the desire to apply them to every problem. These techniques — Baldrige, Lean Manufacturing, Kaizen, Total Quality Management, and others — were created by academic and management gurus to address specific industrial problems. Some techniques such as Kaplan & Norton's Balanced Scorecard are more generalizable than others. Balanced Scorecard calls for the creation of metrics — purposeful measurements — that capture four perspectives: customer, internal, innovation and learning, and financial. The implicit effort to extend the definition of performance beyond financial measures like profitability is essential. And Balanced Scorecard is abstract in the sense that it can encapsulate a variety of other methodologies, such as Six Sigma, that target specific business problems.

Six Sigma's analytic framework can be applied to a variety of business processes. It features two submethodologies:

DMAIC (define, measure, analyze, improve, control) for existing processes
DMADV (define, measure, analyze, design, verify) for new products and processes.

Six Sigma's goal is to reduce process-related defects to 3.4 or fewer per million cases (that is, to six adjusted standard deviations or more from the nominal value), which it does by insisting on measurable processes and by applying statistical techniques that include analysis of variance (ANOVA), Design of Experiments, and regression analysis to the defect-reduction task. (The Greek letter sigma, lowercase (σ), represents standard deviation, a measure of the degree of variation of values in a probability distribution.)

As an aside, I'm pleased to see the reintroduction of sophisticated statistics into business analytics marketed to the mainstream. Business intelligence (BI) has focused on online analytic processing (OLAP)-style slice-and-dice exploratory data analysis and data aggregation, which has meant neglect of valuable approaches to problems that involve nonlinearity, uncertainty, indeterminacy, and time complexities. Send me a note if you want a more extensive explanation.

Six Sigma certified consultants are awarded yellow, green, black, and master-black belts. This martial-arts imagery says, "Don't try this at home." But isn't avoidance of reliance on "superhero" and "cowboy" programmers an essential step for a software-development organization in advancing beyond the first level of the Software Engineering Institute's Capability Maturity Model? By analogy, adopting Six Sigma means institutionalizing an immature organizational model.