banner image for William G. Smith & Associates

DATA RESOURCE MANAGEMENT

The term "INFORMATION RESOURCES" incorporates three categories of "information stuff" essential to the modern business enterprise: its large mass of stored data (the DATA RESOURCE); the huge volume of application system program code (the APPLICATION RESOURCE); and the numerous networked hardware components, along with the operating programming that makes it work (the TECHNOLOGY RESOURCE). These three components, working together, allow the enterprise to capture, store, process, and use the DATA required to produce useful information to effectively operate and compete on a daily basis. Interestingly, the most critical, but the most neglected (and often the most costly) information resource category is the stored DATA RESOURCE. Typically, most of the IS/IT budget and effort is spent managing APPLICATION software and TECHNOLOGY. In many enterprises, there is virtually no effective DATA management. The enterprise must establish effective management control of the data resource, or it will never fully achieve the payoff: faster, cheaper, better business operations, and significantly greater business flexibility, innovation, and change. Further, data is the means by which all other resources are managed: try to imagine managing the human resources, or the financial resources of a modern enterprise with no accurate data about the employees or the accounts and money. DATA RESOURCE MANAGEMENT (DRM) is a critical, and predominant subset of the INFORMATION RESOURCE MANAGEMENT (IRM) environment.

Today, most IT/IS organizations primarily manage "applications systems", and their stored data is treated as just a hidden, under-the-hood artifact owned by each of those applications. However, for any enterprise, there is a finite set of data subjects ("entities") about which it must collect, store, and share data in order to effectively operate and compete. There is a profound misconception that each application (whether home-made, or vendor package) has "its own data", and that this data does not overlap. The truth is that the same data entities must be shared widely across organizational lines, business process lines, and application system lines. In fact, the main objective of the IRM approach is to build databases so that data can be stored once, shared by all who need it, and updated when necessary in one data store, for all users. This maximizes control of quality, timeliness, consistency, security and benefit of the data, while dramatically minimizing cost and "information float" in the enterprise. In sharp contrast to today's dis-integrated applications, all parts of the business can literally know the same thing at the same time. This was the entire purpose for the development of database technology some fifty years ago, and yet the number of enterprises that have effectively exploited this tremendous tool is astonishingly small.

Today, data about these entities are stored very redundantly in data stores for each application which happens to need that particular data; it is not uncommon to find data about a highly-shared entity such as CUSTOMER stored in literally hundreds of data stores, where it is almost impossible to ensure that the data is consistent. Data interfaces (the movement of data from the data store(s) of one application to the data store(s) of a different application which happens to need that same data) are designed, built, and maintained to try to ensure that the multiple redundant data stores are kept in some semblance of a consistent state of update, and to mimic sharing of the same data at the same time. The interfaced application system environment always involves significant "information float" - the time it takes data which is updated in one file/database to flow to all the others which depend on it. One enterprise studied a business process which normally consumed 31 days, from the time the process was triggered, to the time the end result was accomplished. Astonishingly, they found there was only 90 minutes of actual work involved in the entire 31 day duration; the rest of the time was simply information float - data moving from one disintegrated system to another. This type of situation is not the exception - it is the rule. The resulting massive data messes so prevalent in the "dis-integrated systems" environment are major constraints to faster, cheaper, better, and more flexible business operation, and any change to any data store in any application may have a significant "ripple effect" of changes which must be made to many other application data stores and program code.

Virtually every data problem (untimeliness, inconsistency, poor quality, time lags, insecurity, etc.) have their root in the practice of storing data redundantly; whenever a data fact is stored in two (or more) different data stores, the problem of keeping them all updated synchronously and properly must be confronted and resolved. If an enterprise shifts its focus to first managing the data so that it will not be stored redundantly, it is almost impossible to build a tangled mess of redundant and inconsistent application code, or of excess computer hardware. In short, managing the data first brings brings the same degree of order, organization, and optimization to all the other information resources. The converse is not so - the monster data messes that prevail in most enterprises today are affirmation of that fact. DRM is the keystone of the IRM environment.

DRM in a dis-integrated, interfaced application system environment (the antithesis of the IRM environment) is usually limited to a janitorial role, laboriously keeping track of the ever-increasing amount of redundantly-stored data, and attempting (by working with unwilling system developers/installers and development project managers) to ensure that update synchronization between systems is as timely as possible, that data inconsistency is minimized, and that the meanings of all stored data facts are reasonably well documented so they can be used intelligently. However, DRM in a true IRM environment can dramatically improve enterprise operation by ensuring the sharing of single-source data so business processes (and the people who perform those processes) can literally know the same thing at the same time, increase the speed of the process, eliminate redundant/unncecessary work, reduce errors/rework, and improve customer satisfaction. DRM can simply document the legacy data mess, or begin undoing the mess, and replacing it with a much more effective, and much less expensive shared data resource; it is largely a question of how the DRM organization is designed, the authority and resources invested in it, and the adjustment of the other SOFTWARE and TECHNOLOGY management parts of the IRM organization, so DRM truly becomes the driving objective in managing the enterprise information resources.

This seminar is based on actual experience initiating, designing, staffing, equipping, training and implementing effective and successful DRM functions and organizations (not data janitorial organizations) in large enterprises. The IS/IT organization must change to break "the way we've always done it" - it is very difficult to do "a little bit of data management". There is a threshold which must be exceeded, or the result will simply be a lot of wasted time and money, with no mearureable improvement in the state of the data resource. Exactly what must change about the existing IS/IT environment, and how to change it, is described in detail. The seminar also introduces the different levels and types of data models - the most important tools and skills used by data professionals. Finally, there is a wealth of information about the management of an enterprise Information Resource Dictionary/Repository - the intergalactic meta-database. Workshops emphasize development of effective DRM charter, policies and standards, Conceptual Data Modeling (entity-relationship modeling) and metadata management.

 
TOPICAL OUTLINE


DURATION: 4 to 5 days, depending on client's DRM situation

TARGETED AUDIENCES: (no recommended maximum number of attendees)

PREREQUISITES: Concepts of Information Resource Management


RETURN TO SEMINARS PAGE

RETURN TO TOP OF PAGE

© 2013 WILLIAM G. SMITH