Corporate Information Factory (CIF) Resources by Bill Inmon, Inmon Data Systems

> home > view content

Profiling The DSS Analyst

By W. H. Inmon

The DSS analyst is the individual on whose behalf an entire sub industry of information processing has emerged. In the mold of tools to help the DSS analyst discover and probe the mysteries that are buried in the data of the corporation, there are display tools, analysis tools, access tools, storage tools, and the like. But who really is this person - the DSS analyst?

First and foremost, the DSS analyst is someone who is a businessperson who may or may not be technically literate. The DSS analyst usually works in the financial, marketing, or sales department. Typically the DSS analyst is someone who holds a staff position, reporting to a high level manager. It is the job of the DSS analyst to cruise the data the corporation has collected over the years and to discern meaningful patterns and trends that will help management make better decisions.

But there is another very different characterization that can be made of the DSS analyst community. DSS analysts can be divided into two distinct classes - explorers and farmers. A DSS analyst who is an explorer is someone who occasionally scans the data warehouse in search of who knows what - a data surfer in today's parlance. The explorer has a mindset of - "I am not sure what I want or how to find it, but when I see it I will surely know it." The explorer operates on a random basis. Much of the time spent by the explorer DSS analyst browsing data leads nowhere particularly productive. But occasionally the explorer comes up with a brilliant insight.

The other kind of DSS analyst can be called a farmer. The DSS analyst who is a farmer is someone who regularly reaps information from the data warehouse in a predictable manner. The farmer rarely goes looking for things that he/she does not know is in the data warehouse. The farmer repetitively calculates and analyzes the same data over a lengthy period of time. The DSS analyst farmer is a very predictable person, as much as the explorer is an unpredictable person.

Both explorers and farmers are very important members of the DSS analysis community. Both make an important contribution. And there is a symbiotic relationship between the two groups.

The iterative life cycle of development that is typical of the data warehouse DSS environment fits well with the notion of two kinds of DSS analysts. During the early stages of the iterative development process - when the formulation of the DSS process is being established, it is the explorer who is leading the way. But once the iterative analysis pattern starts to turn into a predictable, regular pattern of DSS processing, then the farmers of the DSS community take command.

When a shop starts to build the data warehouse DSS environment for the first time and where there has been no previous DSS processing, the predominant pattern of processing is of the free-form, explorer variety. However, over time as the DSS development life cycle manifests itself, the explorer work turns to farmer work.

As time passes there is more repetitive processing that occurs in the DSS environment. The amount of explorer work that occurs stays constant over time. The repetitive processing grows at a fairly constant rate.

Of course, in the event that a shop has previously built DSS processes and reports, then at the outset of data warehouse DSS processing the shop already has a pattern of DSS work established.

Implications

There are some interesting implications to the growth of DSS systems. Where a shop has done no previous DSS processing, then the design of the DSS systems and data bases cannot be optimized around any predominant pattern of access. Said another way, when a shop brings up its first DSS processing, there is no way for the data base designer to know how to optimize the design, since there is no knowledge of how the DSS system and the data warehouse will be used. You cannot ask the explorer how he/she will be using the data warehouse because the DSS explorer simply doesn't know. But when a pattern of usage starts to emerge - i.e., as systems start to mature from explorer- type DSS processing into farmer- type of DSS processing, then the data base designer can start to optimize the design of the system in favor of the predominant pattern of usage.

Another interesting implication of the difference between explorer and farmer DSS analysts is that of the role of metadata. In the case of explorer DSS processing, metadata plays a very important role. Metadata becomes the "card catalog" that is used to guide the explorer through the many possibilities of analysis. In a word, metadata is the best friend of the DSS analyst and is the place where explorer analysis starts.

On the other hand, metadata is important to the farmer DSS analysts, but not nearly so important as metadata is to the explorer DSS analysts. The farmer DSS analysts operate by rote and after the first two or three times of execution of the same procedure pay little or no attention to the metadata surrounding their DSS processing.

Summary

The world of DSS processing is led by an individual called the DSS analyst. DSS analysts can be broken into two categories - explorers and farmers. Explorer DSS analysts have a random pattern of access and analysis. As such explorers are very unpredictable. On the other hand, farmer DSS analysts exhibit a very predictable pattern of access and analysis. The difference between the approaches taken by farmers and explorers explains why the data base designer has such a hard time optimizing (or even stabilizing) the design of the DSS environment at the outset of design.