If you had the chance to redesign your IT environment from scratch, with no history of custom application or skill development costs to justify and with no installed-base hardware or middleware constraints, how would it differ most from what you have today? The answer, I suggest, defines your most important goals for the year thats just begun.
Before you read any further, stop a moment and think about your actual answer to my opening question. Imagine that it was asked, not by me, but by a far-sighted senior manager who understands the potential of IT and who has the power to write big checks.
If you can get away with it, close your eyes. (But if anyones watching, it might be safer to take a walk.) Somehow, find a way to give the hindbrain a chance to climb out of the nested boxes of what you subconsciously treat as unsolvable problems. Consider both the challenge and the opportunity. Then come back, and well compare notes on what rises to the top of our respective lists.
Welcome back. I wonder if your list includes, somewhere near the beginning, a dramatic reduction of manual steps in starting and managing your routine business processes.
Theres data to suggest that improved process automation should be among your high-priority goals. A survey released last month by AppWorx indicates that human operators are in the process loop at two enterprise sites out of every five, with more than one-third of sites reporting that the burden of manual effort is one of their significant barriers to achieving IT performance goals.
The process automation technologies that have been grafted on top of our legacy IT platforms are notoriously brittle—and costly to deploy and maintain. Batch files are fragile in the face of routine system maintenance or reconfiguration, and clock-driven systems may not even try to address real-world business calendars with their holidays, accounting periods, and other special dates and deadlines.
Dependencies among various steps of a process are often important but difficult to address, especially when a process involves multiple applications on weakly coupled platforms. If step N+1 of a process takes place on schedule, using data from step * thats invalid due to a delay in running that predecessor task, the corruption may spread and be difficult to identify, let alone to reverse. Audit trails may be difficult to implement at all, and most attempts to do so will overwhelm operators with confirmations of normal results that obscure the important exceptions. Sarbanes-Oxley implications are left as an exercise for the reader.
I have my own reality check for any proposed solution to such problems, dating back to the appearance of the Microsoft mantra “information at your fingertips.” I ask people to imagine an enterprise with eastern and western operating regions and routine reports that require rolling up results from both regions into consolidated numbers. I ask people to next imagine that business in the West grows far more quickly than in the East, to the point that the western region is split into Northwest and Southwest units.
How is this reflected in the various scripts, database queries and other IT artifacts that perform the routine processing of enterprise operating numbers? Anyone who claims to have an object-oriented solution to this had better be able to give me a brief and technology-independent answer to that question; the longer and more technology-specific the answer must be, the greater the costs and likelihood of error in making it work over the long haul.
AppWorx, of course, has what it believes to be an effective answer, and I hope to be reviewing the next release of the companys flagship product in the next few months. I suspect that a case study, rather than (or in addition to) a lab review, may be required to get a good handle on the products capability to wrap an object abstraction around a real-world IT installation.
In the meantime, I hope youll give me the benefit of your answers to the questions that Ive asked this week.
Peter Coffee can be reached at [email protected]