Web Exclusive

October 16, 2001

Beyond the Bucket of Reports

Too many analytic applications are merely unstructured buckets of reports. But what's of value beyond the bucket, and how do you get there?

I feel like I've seen it all. As an industry analyst and consultant, I regularly meet with all kinds of people — vendors, consultants, and IT personnel — to review their analytic applications. From the best analytic applications I've come to understand the valuable goals to which this type of software should aspire. And this understanding helps me recognize the worst analytic applications, as well as how to improve them.

For example, the most common problem I encounter is the "bucket of reports" syndrome. You recognize it immediately, when the entry screen of an application presents a long list of undifferentiated report objects. The objects typically include static reports, parameterized reports, interactive analyses, charts, advanced visualizations, multidimensional data cubes, query result sets, and so forth.

The problem is that the list has little or no structure to it, and knowledge workers are expected to click their way from object to object, praying for an epiphany. When the number of objects in the report bucket is low, most knowledge workers can cope with the chaos. But when the list reaches into the hundreds (which is quite common nowadays), it's time to aid the knowledge workers by imposing order on the bucket.

Adding Value Beyond the Bucket

Directory Structure. The first step toward order is to sort reports into a directory structure. For instance, an analytic application for supply chain intelligence may have folders for procurement, supplier ranking, and supply quality. Under the folder for supply quality might be subfolders for returns due to defect, supply failures in production, and warrantee claims per supplier component. In turn, these subfolders would contain report objects appropriate to analyzing these business entities.

Personalized Portal. The corporate portal has become a de rigueur requirement for analytic applications. First, most portal products include tools (or easily integrate with tools) for creating order-imposing directory structures. Second, most portals support personalized content, which contributes to order. Third, portals support a browser-based user interface, so you can reach knowledge workers across the enterprise and beyond.

Summary-then-Details. An arising best practice for the user interface of an analytic application is to present summary information upon entering the application. The user can then drill down into details and interactive analysis on an as-needed basis. Drilling down takes the user from one report object (or set of them) to another, so it organizes the relationships between objects.

Domain Expertise. Now, here's the hard part. The directory structure, personalized content, and summary-then-details user interface must all map directly to the business entities and processes of a knowledge worker's domain. After all, the definition of an application — analytic or otherwise — is that it encapsulates the expertise and automates the processes of a specific business domain (as with the supply chain example above). Furthermore, users don't want just any order imposed; it has to model their business domain according to their established structures and processes.

Assessment

Transforming a bucket of reports into a real analytic application isn't easy. It takes a lot of time canvassing users to learn their business domain, and it takes a lot of technical savvy to model the domain and implement it through a portal or similar platform. But imposing domain-oriented order on the chaotic bucket of reports enables knowledge workers to do their jobs faster, easier, and more effectively. And that's a value-add beyond the bucket toward which everyone should aspire.

Philip Russom, Ph.D. [www.philiprussom.com] is an independent industry analyst and consultant based in Waltham, Mass.

• Privacy
• Your California Privacy Rights
• Copyright © 2008 United Business Media LLC