-
-
Marketplace
-
Channel Resources
Articles from this Site
-
One brand, one Web site! DM Review is now the home of all the content you're used to at BIReview.com and much more. If you are registered at BIReview.com, you're already registered at DM Review. If not, take just a moment to sign up for all the free services we have for you at the new DMReview.com.
Strategic Leadership And The Domain of Business Intelligence
Organizations often seek managers who can offer 'strategic leadership,' yet it is difficult for many to define the term. In the domain of BI, 'strategic leadership' is becoming even more of an issue. The bewildering array of technical capabilities available in the marketplace and the impressive demonstrations prepared by vendors make it difficult to decide what capabilities are critical for strategic leadership in this area. An interesting visual model of the domain of BI is useful to address this issue, especially in the context of understanding the meaning of 'strategic leadership.'
Strategic Leadership
A course on 'Strategic Leadership and Decision Making,' at the National Defense University (NDU) in Washington, DC, offers an excellent definition that fits well in the context of BI. "Strategic decisions are far reaching and consequential for the organization and typically involve the commitment of vast resources. They play out over long time frames and have significant opportunity costs." With data warehousing and BI, the resources involved surely meet this criterion. Of critical importance for BI though is another part of the definition: "Strategic decisions should be made within the context of a long-term view or vision, of both the desired end-state and potentially undesired end-states brought about by the contemplated course of action." This long-term vision for the role of BI within the organization must be developed well before one is under pressure to make a short term tactical decision for buying Product A or Product B. To implement this vision, and achieve desired end-states, requires "leadership ... an interactive process [where] the collective energy of a group, organization, or nation is focused on the attainment of a common objective or goal."
A Model for the Domain of BI
To crystallize this vision we may turn to a model for this domain I developed over 15 years ago. Robust to changes in technology, it still resonates strongly with audiences in presentations around the world.
The model asserts that BI is a focus on the interaction of 'data' and 'relationships.' Either we know the data, or we do not; and either we know the relationships or we do not. We can array these options on a 2x2 matrix as shown.
The Structured Tasks
The upper left quadrant, where we know both data and expected relationships, represents the reporting mode for BI. With multi-dimensional data available for analysis, and the ability to automate structured analytical processes, here we need tools for report formatting and data visualization to deliver our standard reports or dashboard results.
The Semi-Structured Tasks
In the upper right quadrant, we know the data but are exploring to determine the relationships. For example, after advertising a product during the Super Bowl, we may have the sales data for the months before and after the big game, but do not yet know the relationships attributable to our advertising campaign. To explore the data and determine these relationships, we need tools for data analysis and decomposition, inferential statistics, heuristic modeling, and invariably data visualization.
In the lower left quadrant, we use the results of the relationships learned from exploring the previous data to apply to future time periods where we wish to forecast data we do not yet know. Our data analysis identified the factors or data dimensions that influence our forecast, and data decomposition utilized technical tools to parse out the trend, cyclicality, seasonality, and randomness in our data. To complete our forecasting, we may need tools for predictive statistics, deterministic modeling, and data visualization.
The Unstructured Tasks
In the lower right quadrant, we have no data in the structured database sense, and no known relationships. The Department of Homeland Security offers an example of this situation. With vast amounts of unstructured textual data, there is a need to find relationships between characteristics such as bits of content, sources, and timing-and quantifications of the strength of these relationships. While always in the domain of BI, this is only now becoming a viable task, and data visualization is playing an important role in presenting the results achieved.
Linking Strategic Leadership with a Vision of the Domain
With a vision of the domain enhanced by this model, we see where new technologies fit. For example, neural nets, genetic algorithms, and data mining all fit in the upper right quadrant, where they support the exploring function. On the other hand, expert systems, case-based reasoning, and fuzzy logic all fit in the lower left quadrant where they support the forecasting function.
Once set in motion, structured reporting tasks tend to run on their own and are often forgotten, but strategic leadership needs to set up a system to continually review and revise the relationships and metrics, assessing how well they support corporate strategy in a changing environment.
With the semi-structured tasks, strategic leadership should go beyond the technology focus and consider the capability of users leverage its use. It takes training - for analysts and decision makers - to make effective use of these tools. It takes leadership to assure this training is achieved.
All quadrants in the BI model have data visualization in common. A powerful tool, it is also a common source of "potentially undesired end-states" as, without training, decision-makers in the boardroom may not all gain the same understanding from the same visualization. Strategic leadership here helps assure a "desired end-state" where all members of the organization use and understand a common language of graphic communication.
A focus on the effective use of data visualization can thus be the unifying factor for corporate strategic leadership. It helps an organization to be "focused on the attainment of a common objective or goal"-as it facilitates a common language for understanding metrics and measures, the relationships and trade-offs amongst them, and the current status and decision options towards achieving a goal.
Howard A. Spielman, M.B.A., Ph.D., President of Management Semiotics International Inc., can be reached at HASpielman@ManagementSemiotics.com.
For more information on related topics, visit the following channels:
