Ask the Right Questions
Old habits and the ingrained structure of business versus IT can make it tough to measure a project's worth. Try focusing on answering the questions.
Business intelligence is really about answering questions, not about providing better information. To get started, you should ask four questions about any business intelligence project:
- Will the improved information enhance the revenues of the enterprise?
- Will access to certain information reduce operational costs?
- Will the information help the business gain competitive advantage?
- Will improved access to a piece of business information increase the business value?
Once you understand the purpose and justification for a business intelligence project, have your team look at both the questions and the information used to answer the questions. Business intelligence capabilities are created to answer questions necessary to meet business goals. If the goals aren’t understood, you can’t measure the value of the questions, and you thus can’t justify the activity. A business intelligence project team, which ideally consists of both information technologists and business people, isn’t equipped to do its best without understanding the business goals addressed by the project, the questions that need to be addressed and what information is needed to answer the questions.
More than Better Information
Decision makers are used to defining information needs from information technologists through reports or screen layouts. The communication method addresses the information, not the questions the business side is asking. In situations where business does its own analysis, the information technologists simply provide a file or set of files to business analysts.
However, the mission of business intelligence is making better decisions to meet business goals. Business intelligence capabilities are optimized for information access, not for quickly recording and processing business events. Systems automate processes. Business intelligence systems interface with the human decision-making process. The methodologies for design, development and deployment of these two types of systems are distinctly different. One significant difference is the ability—and the need—for business and IT to work more closely together to take full advantage of business intelligence methods that trained IT professionals bring to the table.
As a business intelligence stakeholder, you must consider the roots of what is essentially a business culture issue. In most businesses, the IT function has a better track record and relationship with accounting than with sales and marketing. The rules of accounting are well-established, so when an accountant asks for information to answer questions, the questions don’t change much from one year to the next; neither does the information used to answer the questions.
However, in marketing and sales, the rules are always changing. If a marketing analyst or executive asks for information to answer a question, the question may be completely irrelevant in a matter of days or weeks.
From the IT perspective, it may appear that the marketing professional doesn’t know what he or she is doing. This is false. The difference between the accountant and the marketer is that the latter must handle a much higher rate of change when dealing with customer behavior and getting business. Once everyone recognizes this difference, all can work together to change perceptions: Sales and marketing can focus on getting business, and IT can provide the necessary access to information to answer sales and marketing questions.
In order to maximize return on investment (ROI), business intelligence capabilities must provide information to answer questions that help the business meet its goals. The ROI equation is driven by the value of the questions answered.
If business questions were needed only as part of ROI analysis, you wouldn’t need to understand the business questions once a project is justified and funded. This, however, isn’t the case. As I mentioned earlier, the primary goal of operational systems is to process and record business events; the primary goal of business intelligence capabilities is to support decision making through readily accessible information.
Operational systems are optimized to process detailed data representing distinct business events. When this primary goal of operational systems is balanced against the need to provide information for management purposes, this architecture is called a legacy system. Legacy systems are not defined by age; they are any system built based on the traditional architecture that balances these conflicting goals.
When you’re optimizing operational systems as an information technologist, you don’t need to know anything about business questions beyond the operational level. Since these systems support atomic transactions, IT focuses on supporting as many concurrent, detailed transactions as possible with fast response time for access to detail data. You need little knowledge of specific questions to optimize an operational system. In contrast, business intelligence systems are optimized for access to information, not processing and recording of operational data.
The process of optimizing business intelligence systems—specifically data marts and the information access applications used by business people—starts with understanding the questions the business needs to ask. The goal isn’t just to provide access to information, but to organize the information in such a way that questions are answered quickly and easily. The goal of the data warehouse is to provide high-quality, flexibly organized data optimized for creating data marts.
So, why is the business sometimes hesitant to provide specific questions associated with business strategies? First, the business may perceive that by specifying some questions, others are excluded. There is a grain of truth here: By optimizing for some questions, the data warehouse designer implicitly de-optimizes for others. However, unlike operational systems, the same data can be organized in different ways in multiple physical data stores to optimize for many questions. Well-designed data marts allow you to explore the information related to questions as well as answer the questions themselves.
The second reason, and most likely, is this: The business may not understand the need for specific questions.
Jeff Gentry is President of Technology to Value, LLC and Chief Strategy Officer of eScribendi LLC