An Intelligent Approach to Business Intelligence: Part One

There’s a lot of information about analytics and business intelligence. In fact, you could argue that there is too much information. Business executives and IT professionals scour the internet to research “how to do it” and find everything from relevant blogs to international management consulting firms specializing in the topic. There are a lot of authors and solutions, but most subject-matter experts don’t blog and write articles; they are deep in the process of creating technology enablement and freeing up the data that can be used by knowledge workers.

Part one of “An Intelligent Approach to Business Intelligence” will focus on providing insight into the challenges of capturing, understanding and leveraging data. Part two will offer a suggested approach using the latest reliable infrastructure offered within computer, network, and communications technology.

Background

In the early nineties, Bill Inman introduced applicable ways of aligning data and business reporting towards end-user needs. His concept was called Data Warehousing and the information technology community quickly adopted his theories. New terms such as “star schema,” “metadata,” and “data cleansing and transformation” were being used as components of this fresh approach. Many Fortune 1,000 companies and large institutions funded large initiatives to build new repositories of data using technologies from Oracle, Microsoft, IBM Cognos, Business Objects, and others. Companies quickly leapt into the field with few skills and little training or expertise and, as a result, most of these initiatives never reached their original goals and objectives.

Business Need

The last six years of economic expansion have created a higher number of transactions in most businesses, thus, doubling the data captured by corporations every three years. The need to analyze this data and make sound business decisions is more important than ever in today’s global economy. Consistent key metrics across the enterprise can become a competitive advantage overnight. On the flip side, companies lacking this level of insight will find the competition can quickly get ahead without warning.

In addition, the sheer lack of “knowledge workers,” defined as the people who are empowered to make decisions in companies, creates a struggle to make sound decisions without the proper data. Furthermore, the fact that it takes a significant investment in time and resources to ramp up new knowledge workers prohibits additional resources from making sound, timely decisions at the lowest levels of the organization. Too often, the decision’s timeframe has come and gone while the organization searches for the correct data. If this is the norm, the company could fail to make proper decisions resulting in less than favorable economic results over time.

More and more decision makers are using data outside of their internal systems to make decisions as well. Using outside data for comparative analysis also requires an internal repository for “just in time” analysis. Again, this need can be fulfilled by the implementation of a business intelligence system.

The Business Problem

Why do business intelligence systems fail to live up to their expectations? To use an old Clintonesque phrase, “It’s the operations and enhancements, stupid.” Building a data warehouse, or what is now called a business intelligence or analytics platform, is a complex venture. A comprehensive business intelligence or analytics system cannot be purchased off the shelf as a unified whole. Due to the iterative nature of building this environment, most initial attempts weren’t clearly defined, clearly funded, clearly staffed, or clearly managed, even though the project received a lot of attention from executive end-users and their IT executive counterparts.

What is the issue? After the platform is built, the original team that built the system becomes reassigned to another IT project and the operation of the data warehouse is parsed out to the database support group: the Windows or UNIX systems support people, the PC support desk, and other support groups. These groups are well intended, but they are not trained on new techniques in data warehousing and how to manage the data warehouse from a user or technical perspective. In addition, the various support groups lack continuity in providing the user community with support in terms of data warehousing issues. Are they database experts? Yes. UNIX? Sure. Tools? Certainly. But business intelligence and analytics? No way!

As a result, the new data warehouse or analytics system ends up being deployed and used to a lesser degree than what the executive leadership intended. It is not necessarily a failed project, but it is not a success either.

Are we saying that this is characteristic of all projects in IT? Absolutely not. However, the data warehouse does have operational characteristics that differ significantly from an ERP system or other transactional systems. An intelligence system is a living system, so the data starts to change soon after the initial production system is released. The way in which reports and analysis are done changes in a heuristic fashion. The initial questions and measurements that the users originally documented are different within 90 days after implementation due to today’s short business cycles.

As an example, consider how sales and marketing need to be aligned with manufacturing and distribution in order to get the highest efficiency of production and order management. In all organizations, sales, marketing, manufacturing, and distribution (individually or collectively) reorganize at least once every 12 months. This requires new data analysis, new formats, and new reports.

Now that we understand a bit about the challenges, we’ll look at a new approach to business intelligence and analytics in part two.

To learn more view our webinar, “Do you know what it takes to be a more data driven BI organization?”



« Back to blog