As the IT departments of public corporations burn the midnight oil to meet the deadlines of Sarbanes-Oxley (SOX), many are asking the same question: "Why was this so hard?"
Indeed, the primary difficulties faced by IT have been not the individual technical requirements of SOX, but the scope of the legislative mandate. In SOX, every change to the company's data must be accounted for, and security potholes that once allowed ad hoc modifications must be filled. Moreover, changes to the software itself must be rigorously documented, along with the identities of authors of changes and the notations of the authorizations that permitted those changes. Simple things, like modifying the length of a database field to fix a problem--once a relative snap on the iSeries--have been transformed by the SOX auditing requirements into non-trivial IT management headaches.
Change Management and Documentation Management Tools
Some of the most significant tools that are helping IT managers through this nightmare are the Change Management and Documentation Management suites that IT vendors have been selling for years. These tools have evolved significantly since the days of simple source/object control. Many suites today provide a good framework for accounting for and auditing of software changes. Moreover, the implementation of these tools--on homogeneous computing platforms like the iSeries i5 or across multiple server platforms in a typical IT shop--can really help IT to better communicate the IT implications of business decisions that management makes. Today, some tools can actually assist in the estimation of change projects themselves.
I know! On first blush, this sounds like one of those calls for project management tools that chew up budget and slow down development. But consider what the real requirements for IT have become in recent years. Consider, for instance, how many organizations moved in the 1990s to the eXtreme Programming (XP) model for application development.
The XP Model--A Model of Chaos?
Under the XP model, new solutions are developed with small teams consisting of users and programmers who work together to swiftly build new applications. Each team is seen as a microcosm of the organization and an agent of change. This development model encourages the rapid implementation of new code and the ability to progressively change that code to meet the evolving demands of the users.
XP is an excellent way to get a lot of new code built quickly, but it's also a really lousy way to build a comprehensive, integrated information system. The modern application development tools place a lot of power into the hands of individuals who may or may not have the business acumen to vouchsafe the company's best interests. The best intentions of these teams may, in fact, create a system that is running counter to the fiscal requirements against which the organization may be measured by auditors.
The Parable of the Quick and Dirty Fix
An example of this was brought home to me when I worked at an apparel manufacturer. The fashion department needed a "quick and dirty" way to calculate the labor costs for building a certain line of garments. When the fashion director approached IT with his proposal, IT informed him that the project would require at least a month of study to identify the implications to IT's ERP system. That's right! A month, just to see if it could be done!
This was, of course, not the answer that the fashion director wanted to hear. So, instead, unbeknownst to IT, the fashion director hired his own IT consultant to quickly build a solution to his department's problem. By all accounts, the consultant did an excellent job coordinating the requirements: In short order, he had developed a solution that completely filled the needs of the fashion department.
Today's Hero May Be Tomorrow's Rogue
This fashion director was proclaimed to be a hero by the upper management team, and it quickly adopted the reports of the new system as an adjunct to the IT department's ERP system.
Within no time at all, many more such adjunct systems were being implemented throughout the organization. Soon, IT's entire ERP system was rapidly unraveling in a classic case of conflicting data sets. Management was asked to choose which data sets or information systems it would be using and which systems it would be integrating into IT's ERP. When the choices were made, it then took more than a year to identify how to bring all of these adjuncts into IT's control. That's right! More than a year!
The Auditor's Response
At the end of the first and second fiscal years, the company's auditors were perplexed. Where did these sales projections come from? Where did the cost projections come from? How did those numbers figure into the financial calculations that were derived from the central ERP database? No one could say! The auditors noted the problems and added their own set of requirements for fixing the problems. IT's ERP requirements grew and grew, while more adjunct systems continued to promulgate through the company.
By the end of the fourth year, the IT manager was forced out. Upper management had finally lost patience with IT's slow response to the burgeoning demands of auditors and line managers for accurate information.
But that wasn't the end of the problem. New adjunct systems continued to be built, and the ERP system continued to unravel. New IT managers came and went. New requirements by auditors continued to build in IT's backlog. Upper management had completely lost control over its own information system and was operating by the seat of its pants.
The End of the Company
In the end, upper management determined that the mess of conflicting data no longer made it economically feasible to sustain an internal IT organization. It decided to outsource all of its basic IT requirements to a large IT consulting organization. Five years later, the company was sold to a larger clothing manufacturer.
What was the cause of this IT meltdown? Was it the fault of the guerilla actions of the fashion director? Was it the fault of the XP development model? Was it the fault of upper management? Maybe!
What Went Wrong?
But maybe it was also IT's inability to provide a reasonable estimate of time to create the fashion director's initial request. Without the change management and documentation management tools available to estimate the extent of the requested changes, the IT manager had ultimately been forced to fall upon his own sword.
Today, the complexity of modern information systems requires us, as managers, to treat the software applications and networks of hardware as more than mere libraries of instructions and databases housed in racks of hardware. In order to respond rapidly to the requirements of our fellow managers, we must organize our understanding of these resources into a structure of information that we too can rapidly access and query.
We need to be able to quickly judge the impact and implications of simple changes to the information system as a whole, and we need to develop better ways to communicate those implications to our upper management. Long gone are the days when a simple IT task list would meet the needs of communication. Each change we make to a single database must now be accurately reflected throughout the entire system. Just identifying those changes can take a significant amount of analysis. Without automated tools, IT's guess at the extent of a simple change can lead to tremendous delays in the ability of the organization to respond to changes in the business model.
The Need for Better Tools
Change management tools and documentation management tools are key resources that IT must have to meet these requirements. These are analysis tools that can inform us of not only what has changed in the past, but what must be changed in response to a new initiative in the future.
The task of implementing SOX is only now making this need known to us, but that need has existed for years. As soon as we acknowledge this need--not only for IT, but for the company as a whole--the sooner we'll be able to resolve future legislative mandates that follow SOX and make our IT infrastructure meet the needs of our organizations.
Thomas M. Stockwell is Editor in Chief of MC Press Online, LP.
LATEST COMMENTS
MC Press Online