|Sign-Up for Free Exclusive Services:||Portals|||||eNewsletters|||||Web Seminars|||||dataWarehouse.com|||||DM Review Magazine|
|Covering Business Intelligence, Integration & Analytics||Advanced Search|
The 5 Principles of High-Impact Analytics
IDC recently completed a major study -- The Financial Impact of Business Analytics, based on interviews at 43 organizations in North America and Europe. (See www.idc.com/analyticsroi.) This was the first comprehensive investigation of the return on investment (ROI) derived from analytic applications. The study found that organizations which have successfully implemented and utilized analytic applications have realized returns ranging from 17 percent to more than 2,000 percent with a median ROI of 112 percent.
In addition to the quantitative results, the interviews identified a series of best practices for successfully implementing and integrating analytics in a business process context. IDC identifies these lessons learned as the five principles of high-impact analytics.
1. Recognize the Application Imperative
A common thread in each business analytics case covered in the study is a laser focus on a critical business issue. Analytic applications are specific to a business function (such as marketing, finance or product development) within an industry (such as banking, retail or chemical manufacturing). Analytic applications focus on specific decision processes within the function such as scheduling of personnel in airport operations or making cross-selling recommendations to customers.
The consistent advice from sites was not to tackle everything at once. Address an issue that is business critical and where there is a potential for significant payback. In the words of one application manager, "Create a small-scale project and move it along. Such a project doesn't require much to get started." Many sites talked about starting with one application and one function and then adding functions in a staged implementation.
However, the "start small" concept has a caveat attached to it. Organizations with foresight execute the initial analytics project with an eye to the future. They build in processes for data quality, providing a robust data infrastructure upon which the current application runs and to which future applications can be added. In fact, a number of sites were extending existing data warehouses to accommodate a new application for users not previously served.
2. Democratize Information Assets
Business analytics is not equivalent to an executive information system (EIS). The premise behind an EIS was that decision-making was a semi-regular senior management activity, with the resulting decisions passed down the chain of command. Hence, only senior managers required information for decision support.
Today, however, decisions are made frequently and at all levels of the organization. This fundamental shift requires the timely delivery of relevant, targeted information to each person who makes decisions in an area of responsibility. The broadening of the base and increased frequency of decision making raises the potential for ROI in improving these processes.
This organizational shift is coincident with a technology shift toward the Web. The pervasive use of the Web for delivering and accessing information is the basis for the democratization of corporate information assets. Many more employees are information consumers who understand how to use Web browsers and search engines to locate information. Analytic applications should exploit simple Web-based interfaces that capitalize on these skills, broadly opening access to information. The economics are straightforward: the more users who gain benefit from an analytic investment, the greater the potential ROI.
3. Build Discipline in Decision-Making Processes
To achieve ROI, it is not sufficient that an application is widely used. The issue is how an analytic application is used and whether it drives improvement in decision making. As one manager stated, "Start with the process and the desired best practices before even looking at technology." What are the tasks that need to be done in order to arrive at a, decision? What information must be analyzed and made available at each step?
The goal is to improve by driving consistency in decision making. Capture best practices of,experienced performers who use information to identify root causes of problems or select wisely between alternative courses of action. The preservation of expertise is a basic premise of knowledge management and was in evidence at a number of sites. This is particularly important for those types of operational decisions that are made repeatedly by different individuals in an organization.
For example, at an academic publishing house, the arrangement of information and the sequence of screens in the application remind users of all of the financial considerations needed when making a project investment decision. This brought greater consistency in the decision-making process with a heightened sense of accountability for the quality of decisions and the accuracy of forecasts.
4. Recognize New Skills Required for Knowledge Workers
Business analytics changes the way that people have to think. They must think analytically, and that's not always something you can teach people. For example, marketing professionals must be able to apply segmentation and cross-selling models to determine the next wave of a marketing campaign.
Train users well. One interviewee observed, "This is not a tool for the average user. They made [the analytic software] very user friendly, but it's like giving a weapon to someone who doesn't know how to use it. They would hurt themselves."
Business analytics impacts people -- from the business process to the job requirements and, ultimately, to the hiring process. In short, business analytics is redefining the knowledge worker. As the knowledge bar is raised, we expand the knowledge worker's reach, but we also expand the level of skills required for a knowledge worker to get a job done.
5. Deal with Complexity: Closed-Loop, Adaptive Systems
What goes around, comes around. From a technology perspective, the success of a closed-loop system hangs on the slender thread of data quality. From an organizational perspective, success depends on the willingness of employees to accept the adjustments to operational rules that stem from the analysis.
The purpose of business analytics is to analyze the feedback from operations -- what we are purchasing, shipping or managing. Based on the analysis, adjustments are made to operations, and the cycle continues. We learn, and then we correct. Hence, the combination of analytics and operations comprises an adaptive system that helps an organization deal with complexity.
This cycle of correction and improvement was pervasive in the business analytics cases. For example, a forecaster of product demand analyzed the accuracy of past forecasts in a successful effort to improve his next forecast. A more accurate forecast meant that there needed to be fewer last-minute changes or escalations to the manufacturing schedule.
Analytic applications are not standalone systems. Each business process (such as procurement or customer service) can be supported by a system that has grown organically: specialized operational systems interoperating with purpose-built analytic applications. Getting through the complexity crisis requires that these systems work with one another to deliver maximum results. A majority of sites recognized the value of a data warehouse and its accompanying processes to enable the analytics to work on a base of quality data. A closed-loop system intensifies this requirement.
Across the board, IDC found that the implementation of business analytics, guided by these five principles, produced outstanding return on investment for organizations across a wide range of industries.
© 2003 Henry Morris and IDC
Check out DMReview.com's resource portals for additional related content, white papers, books and other resources.
Henry Morris is the group vice president and general manager for for IDC's Integration, Development and Applications Strategies (IDeAS) solution research group. Dr. Morris started the Analytics and Data Warehousing research service at ICD and coined the term "analytic applications" in 1997. Currently, he is exploring the relationship between business intelligence and business process automation. Morris may be contacted at firstname.lastname@example.org.