BI Requirements Gathering: Leveraging What Exists

 

In my recent columns, I described two of the most common techniques for gathering BI requirements – interviewing and facilitation. Another frequently used technique is building on what already exists. This incorporates three distinct approaches: analysis of existing queries and reports, inventory of spreadsheets and inventory of current extracts.

Analysis of Existing Queries and Reports

Businesspeople who are looking for business intelligence capabilities typically are not starting from a clean slate. Over time, they have established a series queries and reports that are executed on an ad hoc or regular basis. These reports contain data that they receive and purportedly use. Understanding these provides both advantages and disadvantages when gathering requirements. The major advantage is that using the existing deliverables helps to provide a basis for discussion. Commenting on something concrete is easier than generating new ideas. With the existing reports in hand, key questions to ask include:

* Why do you need this report?

* Do you regularly use this report?

* How do you use this report?

* Is there data in the report you don’t use?

* Is there other data you use in conjunction with the report?

The first of these questions is possibly the most important. A goal of business intelligence is to enable action based on information. Often people work with something familiar rather than looking at new ways to achieve their objectives. Understanding why the report is (perceived to be) needed may lead to a discussion that helps the business user move from a report environment to an analytics environment. For example, if the person says that he or she uses the report to anticipate product demand, then the discussion can easily move to information and analytic capabilities that are needed, thereby helping the person graduate from a report user to a true business intelligence user.

There are disadvantages to using existing reports as well. First and foremost is the tendency to adopt a myopic view. Concentrating on the existing report may keep the user from considering additional needs. Further, the content of the report may have been dictated by ancient needs, system limitations and development time constraints. Awareness of the disadvantages enables the analyst to use the existing queries and reports as one of the means for gathering requirements.

Inventory of Spreadsheets

Spreadsheets are one of the most common tools used by business analysts. Regardless of the information source, information often finds its way to a spreadsheet. Understanding the spreadsheets that people use provides insight into both the data that is being analyzed and the way it is analyzed. (Still to be understood is the “why” as indicated above.) The problem in getting an inventory of spreadsheets is that these are typically not catalogued. Techniques for building the inventory include:

* Contact key business analysts and ask them to list the spreadsheets they are using.

* During a session with business analysts, ask them to open Excel and then look at the recent history of spreadsheets that were used.

The first approach is less threatening, but may not yield a full list. The second approach, particularly if it is done at different points in the business cycle (e.g., end-of-month, budget time, etc.) is more invasive, but is likely to yield a much more comprehensive set.

Inventory of Extracts

Business analysts need information. Sometimes, their preference is to use traditional channels and get the IT department to provide them with data marts and reports. When these approaches don’t meet their needs, they often resort to getting data extracts which they then use to populate (independent) data marts or spreadsheets. While extracts are very effective, often these are created as one-off solutions. As a result, the number of extracts expands over time. The extracts often contain similar data, leading to significant data redundancy. Compounding this issue is that each extract is done at a particular point in time and is based on a set of assumptions that meet the individual requestor’s needs.

An inventory of extracts provides information on the data that people are using for business analysis. Providing an architecturally sound method for delivering the information not only meets the business needs, it also enables a reduction in environment complexity and data redundancy and an improvement in the quality and consistency of the data used for decision-making.

This third column on business intelligence requirements gathering explored how we can build on the information people are already getting to help us gather BI requirements. In the next column, we will explore a related technique – observation. Please keep in mind that these techniques are to be used in conjunction with each other. Not all techniques are needed for a single initiative, but all should be considered to determine the best approach.

I welcome your input – please send me your thoughts at jgeiger@intelsols.com.

Jonathan G. Geiger is executive vice president at Intelligent Solutions, Inc.

This story originally appeared on Information Management, a SourceMedia publication.