Defining and using KPIs in a successful business intelligence system

In this book excerpt, author David Loshin discusses how to select key performance indicators and turn data into actionable knowledge for BI success.

This is an excerpt from Chapter 2, "The Value of Business Intelligence," from the book Business Intelligence, Second Edition: The Savvy Manager's Guide by David Loshin. Loshin is president of Knowledge Integrity Inc., a BI and data management consulting company. In this chapter, he explains how key performance indicators (KPIs) can aid in measuring BI and analytics performance. He also discusses the challenges of defining and measuring performance in terms of meeting business objectives, and goes on to explain how to use "actionable knowledge" to generate value from a successful business intelligence system.

KPIs can be collected together to provide a conceptual scorecard for a business and can be associated with a number of different business activities, especially within our four value driver areas such as financial value, productivity, risk and trust. In fact, a large number of KPIs can be defined in terms of measuring performance associated with many different BI analytical activities.

Another conceptual value of BI is the ability to capture the business definitions of the key performance indicators, manage those definitions as part of the corporate knowledge base and then provide a visualization dashboard that reflects those KPI measurements, presented in a form for management review. This BI dashboard displays the results of the analytics required to configure the KPIs in a succinct visual representation that can be understood instantaneously or selected for drill-down. A BI Dashboard will not only provide real-time presentation of the selected KPIs, but will also hook directly into the BI components that allow for that drill-down.

Copyright info

This excerpt is from the book Business Intelligence, Second Edition: The Savvy Manager's Guide, by David Loshin. Published by Morgan Kaufmann Publishers, Burlington, Mass. ISBN 9780123858894. Copyright 2012, Elsevier BV. To download the full book for 30% off the list price, visit the Elsevier store and use the discount code SAVE3013 through Jan. 31, 2014.

By looking at some sample performance metrics, we can become comfortable with engaging the business users to assess their query and reporting needs as well as determine the degree to which existing data sets can address those needs. And the categorization of business value drivers that has been presented earlier in this chapter supports the BI process by helping to clarify general business objectives and corresponding performance metrics and indicators.

Improving the way the business is run as a result of integrating a BI framework goes beyond the technology -- key stakeholders must specify what their perception of "performance" is, provide the performance measures and then define achievable targets and use the tools to inform the decision-making processes. These measures are put in place to assess, measure and control the degree to which the business objectives are being met. Specific programs can be designed and developed around improvements within any of these key categories. Consider these examples:

  • Revenue generation via customer profiling and targeted marketing. Business intelligence reports and analyses reflecting customer transactions and other interactions enable the development of individual customer profiles incorporating demographic, psychographic and behavioral data about each individual to support customer community segmentation into a variety of clusters based on different attributes and corresponding values. These categories form the basis of sales and profitability measures by customer category, helping to increase sales efforts and customer satisfaction.
  • Risk management via identification of fraud, abuse and leakage. Fraud, which includes intentional acts of deception with knowledge that the action or representation could result in an inappropriate gain, is often perpetrated through the exploitation of systemic scenarios. Fraud detection is a type of analysis that looks for prevalent types of patterns that appear with some degree of frequency within certain identified scenarios. Reporting of the ways that delivered products and services match what had been sold to customers (within the contexts of their contracts/agreements) may highlight areas of revenue leakage. Both of these risks can be analyzed and brought to the attention of the proper internal authorities for remediation.
  • Improved customer satisfaction via profiling, personalization and customer lifetime value analysis. Customer lifetime value analysis calculates the measure of a customer's profitability over the lifetime of the relationship, incorporating the costs associated with managing that relationship as well as the revenues expected from that customer. Employing the results of customer profiling can do more than just enhance that customer's experience by customizing the presentation of material or content. Customer profiles can be directly integrated into all customer interactions, especially at inbound call centers, where customer profiles can improve a customer service representative's ability to deal with the customer, expedite problem resolution and perhaps even increase product and service sales.
  • Improved procurement and acquisition productivity through spend analysis. Spend analysis incorporates the collection, standardization and categorization of product purchase and supplier data to select the most dependable vendors, streamline the RFP and procurement process, reduce costs, improve the predictability of high-value supply chains and improve supply-chain predictability and efficiency.

Each of these examples can be viewed in both the operational and the strategic business perspectives. The operational view provides insight into existing conditions and performance comparing existing activities to expectations. From the strategic perspective, we can evaluate the degree to which any potential measurements impact future corporate value.

Using actionable knowledge

It is important to recall that you can only derive the value from information if you are able to make positive changes based on that information. This means that some investment will be required to build the environment where data can be turned into knowledge, but the real benefit occurs when that knowledge is actionable. That means that an organization cannot just provide the mechanics for creating knowledge; it must also have some methods for generating value using that knowledge.

This is not a technical issue -- it is an organizational one. Delivering actionable knowledge is one thing, but to take the proper action requires a nimble organization with individuals empowered to take that action. And despite the costs, the senior managers must be convinced that the investment will yield results. Therefore it is in the best interests of the organization to consider the types of costs inherent in developing a BI platform for comparison with the anticipated benefits.

This includes analyzing costs in relation to increased performance for any value driver related to the activity, such as:

  • The fixed costs already incorporated into the BI infrastructure (e.g., database or query and reporting tool purchases);
  • The variable costs associated with the activity (e.g., are there special software components required?);
  • The ongoing costs for maintaining this activity;
  • The value of the benefits derived by taking actions when expected knowledge is derived from the activity;
  • The costs and benefits of other BI components that need to contribute to this business activity;
  • The value model expected from this activity;
  • The probabilities of successful applications of these actions to be applied to the expected value;
  • The determination of the time to break even as well as a profitability model.

For further discussion of this chapter, see this Q&A with author David Loshin.

Next Steps

Learn the truth about common BI misconceptions

Dig Deeper on Business intelligence management

Data Management
Content Management