For our first column of 2007, the OpenBI Forum begins delivering on our 2006 resolution to explore the expanded reach of BI with a cross-discipline perspective. This column borrows from the academic world to provide a definition of business strategy, then attempts to show how strategy is operationalized and measured, ultimately connecting to business intelligence. Subsequent columns will drive from this foundation, addressing issues of BI validity and the development of business theories.
In a seminal Harvard Business Review article from 1996 titled, "What is Business Strategy," Michael Porter builds a simple yet remarkably sensible definition of business strategy. Based on observations from case studies, Porter first notes that strategy is often misperceived as operational effectiveness, an organizational endeavor he considers a necessary but not sufficient investment for enterprise success. Operational efficiency results when a company performs the same activities better than its rivals. Strategic positioning for differentiation and competitive advantage, however, results when a company effectively performs a different set of activities than its competitors or conducts similar activities in different ways. Operational effectiveness (OE) - Six Sigma, TQM, ABC, benchmarking, outsourcing, etc. - pushes the productivity frontier but is not sufficient for differentiation since competitors can quickly imitate, causing convergence and diminishing returns. OE raises the bar for all, ironically often initiating a race to the bottom. Competition based solely on OE is usually mutually destructive, leading to wars of attrition and subsequent marketplace consolidation.
Successful competitive strategy, on the other hand, is about being different, choosing different sets of activities to deliver unique value. A company's strategic self concept - the pieces or niches of the competitive map that the company chooses to engage - is the point of departure for strategy and requires a tailored set of activities to implement. This positioning mandates that the company make trade-offs between activities to attain competitive advantage - determining which activities are chosen and how these activities interrelate.
If operational efficiency is about attaining excellence in the performance of individual activities, strategy is about making trade-offs and combining activities to achieve an optimal set of outcomes. And the more difficult it is for competitors to imitate that collection of activities, the more imposing the moat protecting the franchise. From this perspective, a prescription for growth is to deepen rather than broaden focus, attempting to make the strategic fit of activities unique and non-reproducible. Effective management is then all about strategy: defining and communicating the unique position(s), making trade-offs, forging fit among activities, and aligning and deploying the organization.
How does the notion that strategy is about "designing and creating fit among a company's activities that involves trade-offs in positioning to gain competitive advantage" relate to business intelligence? Robert Kaplan and David Norton, originators of the Balanced Scorecard and authors of many texts on performance management, including The Strategy-Focused Organization, provide a foundation for the critical connection between strategy and business intelligence (BI). Kaplan and Norton take Porter's thinking to the next level of measurable detail, viewing the complex interplay of strategy "activities" as a series of hypotheses designed to move the organization from its current state to a desirable but uncertain future state. The segue that the organization takes, based on its positioning or self-concept and mosaic of business activities, involves a series of linked hypotheses relating activities to outcomes. Indeed, the strategic hypotheses describe a series of cause-and-effect relationships of activities and outcomes that are explicit and measurable. The activities that drive performance are called lead indicators; desirable outcomes are considered lag indicators. In this light, strategy can be seen as a family of hypotheses, linking driver activities (lead indicators) with outcomes of interest (lag indicators). The hypotheses take the form of:
- If A (lead indicator), then B (lag indicator) or
- The more/less of A (lead indicator), the more/less of B (lag indicator).
A company's composite strategy is the aggregate or map of these hypothesis linkages.
A major benefit of Kaplan and Norton's approach is that of operationalizing strategy - moving from vague and conceptual to explicit and measurable. At the point of explicit and measurable, there is a confluence of strategy and BI. The primary role of BI for both the evaluation of operational effectiveness and strategy execution is performance measurement (PM). Before addressing the usual BI functions of data integration, query/reporting and analytics, the PM BI analyst must make explicit and measurable the linkages underpinning strategy, precisely specifying relationships and outcomes. A second more subtle, but equally important, task of the PM BI analyst is one of promoting designs and methods to assure the validity of strategies so that lead and lag indicators are related as cause and effect. Rather than just observing the results of a new policy, the BI group might compare a randomized "treatment" group to a control group, a superior design to passive observation only. An example would be a retail cataloger that mails different versions of its primary catalog to randomly selected cohorts and a control group, and then measures subsequent purchasing behavior. With such a design, the organization can be confident that differences noted in the buying performance of the disparate groups are in fact due to the catalog versions they received and make strategic decisions based on the results. OpenBI Forum will have more to say on validity and design for BI in a subsequent column.
The strategic transformation of Sears, Roebuck and Company in the early and mid-90s provides a compelling illustration of both "strategy as hypotheses" and the enhanced role of BI for performance measurement. As described in the HBR article "The Employee-Customer-Profit Chain at Sears," by Rucci, Kirn and Quinn, Sears' strategy derived from the vision to make the company "a compelling place to work, to shop and to invest." To that end, the measurement team first developed a balanced scorecard for the company, called Sears Total Performance Indicators or TPI, with an additional mandate to assess the drivers of future performance with statistical rigor. The team then amassed a wealth of interview and survey data for employee attitudes and behavior as well as customer impressions, recommendations, satisfaction, etc. - and set out to relate these in a causal chain to financial performance. One such linkage that was established using survey data and causal pathway modeling, an enhancement to statistical regression analysis, was: 1) a five unit increase in employee attitude drives, 2) a 1.3 unit increase in customer impression which in turn leads to, 3) a 0.5 prcent increase in revenue growth. A larger more comprehensive model relating employee perspective with the customer view to financial performance was also built and refined. In the late '90s, TPI became the Sears mantra deployed at every level of the organization. At that time Sears assessed the strategy as highly successful, with both employee and customer satisfaction improving by 4 percent, translating to a $200 million increase in revenue as predicted by the statistical model.