|
|
| Home - Industry Article - Feb 06 Issue |
When Best Practices Aren't Good Enough
continued... page 2 |
Middle managers in charge of IT projects are often faced with similar conflicts, since often they are not technical and must depend on the judgment of the technical staff reporting to them. Moreover, since a manager's importance in a large organization is usually perceived as a function of the number of people reporting to him/her, there is little motivation for managers to advocate an innovative technical architecture, which, if successful, would significantly reduce the size of their staff. This situation is exacerbated in companies that have grown by merger and acquisition. Since a goal of these transactions is to eliminate "redundancies," IT groups often resist consolidation by what in government circles is called "malicious compliance" in order to avoid being the organization made redundant.
Moreover, these conflicts of interest reside not only within the enterprise, but across the software industry. Because technology is complex and understanding technical differentiators between product offerings requires significant background knowledge, companies have long depended on outsiders to provide guidance and advice. Industry analysts provide research and consulting services to help companies evaluate products and providers. System integrators and offshore developers provide experienced staff that can prove cost-effective while reducing risk on particular projects. While this reliance of third parties makes sense, it is important to recognize that the financial well being of these organizations in some ways depends upon technology remaining difficult to understand.
It is unreasonable for customers to expect third-party firms to point out that the next new thing promising to be the "silver bullet" is, in fact, technically almost identical to something that failed before. For example, EII products look an awful lot like the heterogeneous database management systems that people tried to build in the 80s. And knowledge management is, in many ways, artificial intelligence in a K-Mart dress. In order to get the contract, an integrator doesn't want to raise doubts about a project. At some level, then, analysts and integrators benefit from obfuscation.
What can be done?
There is a saying in the military that "what gets measured gets done," so the first thing that companies have to do is make a serious commitment to simplifying their enterprise architecture. They must also understand that this commitment will take the same kind of continual vigilance and discipline that they give to their financial reporting. There is no quick fix. Like individuals whose lifestyle has impaired their health, it's not a simple matter of finding the right salve or vitamin. Nor does the responsibility of remedying the situation lie in the province of a third-party, whether dietician, physician, or system integrator.
There are three facts of IT life that no organization should ignore:
- Like death and taxes, IT integration is, and will continue to be, a part of doing business so it is imperative that it be done efficiently. Key to efficient integration is to knowing what has already been done so that little or no time is required to rediscover information. In short, as far as integration is concerned, it is best not to depend on humans for documentation because even if they are diligent and accurate -- and diligence does not guarantee accuracy -- the sheer complexity of the information makes manual analysis a lengthy, error-prone and costly process. As a result, just as diet and exercise are key to good health, an organization must begin to use tools and processes that foster the automatic acquisition of a metadata audit trail (that is, a description of everything that is done to data in the process of moving and/or consolidating it) so that it can be queried for the purpose of impact analysis when something changes and re-used on other projects.
- Given that this legacy metadata is not electronically available today, it is not possible to anticipate how long it will take to achieve these large integration initiatives. Therefore, in order to meet deadlines on strategic integration initiatives, it is critical that companies adopt products that support a rapid integration development methodology -- that is, one that supports an efficient change cycle. An efficient change cycle not only depends upon rapid impact analysis but upon the ability to quickly implement and test changes. Thus, productivity tools must be tightly coupled with the organization's enterprise metadata strategy.
- An organization's commitment to efficient integration is as important to meeting its fiduciary responsibility as following standard accounting procedures. Not having a full metadata audit trail of everything that is done to the data when moving between applications presents the same element of risk as having "off the books" transactions.
The most important lesson to be learned is that a company should have internal staff who are capable of evaluating technology alternatives independent of third parties and that they should be given sufficient authority (and accountability) to have their recommendations "stick." That is not to say that a company should not engage analysts, consulting firms and/or system integrators, but that somebody (or a team of somebodies) in the organization should be technically knowledgeable enough about both 1) the current IT environment and 2) the technology being investigated, to make an independent judgment call as to the advisability of third-party recommendations.
Secondly, IT initiatives should be driven by business reasons. These might include providing superior customer service, addressing new markets, or availing oneself of a disruptive technology. However, the way in which the initiatives are undertaken should be driven by criteria from the top. For any purchase of new software these criteria should include the following:
- Applicability across applications and the enterprise. Since the initial purchase cost of software products is only the beginning of the costs associated with their use, the fewer products a company acquires and maintains in its IT environments the more efficient and cost-effective its IT organization will be. (It is not only the major runtime engines like dbms vendors that companies should worry about, but their productivity and desktop products as well.)
- Scalability. Often, product solutions that work well in the prototype or line-of-business environment become inefficient and/or costly when their use is expanded across the enterprise. For example, while forty minutes may be an acceptable length of time for daily updates to a data warehouse, sixteen hours (once the product has been deployed throughout the company) may not be. Similarly, if enterprise-level deployment of a solution requires a significant investment in additional hardware and/or software, it should be considered only if no other alternative is available.
- Efficient change management. If the past decade has taught us anything, it is that things change -- technology, business ownership, product offerings, and regulatory requirements, to name a few. As a result, one of the most important aspects of an IT enterprise architecture is efficient change management. This means that regardless of the source of the change, it must be possible to both 1) quickly determine the impact of that change and 2) adjust operational systems accordingly. Unless a software offering can provide this type of analysis (or enable another product to do so), it is not a good candidate solution.
- Auditability. Not unrelated to change management, it is critical that organizations be able to easily demonstrate to auditors and regulatory agencies that their IT controls are in compliance with the processes that they have documented. Any software product used in production -- and any productivity tool used in that product's implementation and deployment -- should provide some form of auditability by way of the metadata it captures. The product must also be able to automatically export this metadata in a form that makes it easy to generate compliance reports.
Conclusion
Best practices are a form of wisdom, and wisdom is acquired from mistakes. The goal of this paper was to enumerate the factors -- human and technical -- that have been the cause of shortsighted decisions in corporate IT. Given that information technology pervades every aspect of life and business, there is no going back. The only solution is to understand the importance of committing to smarter choices and using wisdom gained from past mistakes to direct the behavior of the people and institutions on which we depend for guidance. Only this way will we all truly benefit from the promise of information technology.
Kay Hammer is one of the two founders of Evolutionary Technologies International, Inc. and its CEO through December 2005. She is the author of "Workplace Warrior" (2000) and co-author of the forthcoming textbook "Fundamentals of Software Integration" (to be published by Jones & Bartlett), from which this article was excerpted. For article feedback, contact Kay at kay@eti.com
 
|
|
|
