IT Management Reference Guide

How well a process is automated with technology often distinguishes whether an infrastructure is world class, merely average, or even mediocre. The reason for this is that automation can frequently be a double-edged sword for IT shops. When properly applied, automation can greatly streamline processes, reduce cycle times, increase efficiency, minimize errors, and generally improve the overall service levels of an infrastructure.

When improperly applied, automation can do more harm than if never attempted in the first place. This can be due to the fact that applying automation to a poorly designed process merely results in a highly automated poor process. The output may be arrived at more quickly, but in all likelihood it will be unacceptable in quality and content. Applying the wrong type of technology to automate a process can also cause more problems than not automating at all. For example, one company attempted to automate part of their output processing with mechanical feeders that virtually eliminated the expensive and highly repetitive manual loading of paper into laser printers. Later, managers realized that only a limited type of specially bonded paper would work properly causing them to eventually go back to the manual process.

Processes should not be automated until three key conditions have been met.

1. The process should be well designed and standardized to the maximum extent achievable. Say, for example, the process you are automating is used in multiple instances or at multiple sites. If the process has been well designed and standardized at the outset, the eventual automation techniques can be migrated to the other occurrences of the process. This will help to cost justify the expense of the automation.
2. The process must be streamlined as much as possible. This is accomplished by eliminating all non-value-added steps and by ensuring all the necessary value-added steps are included. This should apply to all instances of a standardized process. This ensures that, not matter which parts of the process are selected as candidates for automation, they are in fact vital elements of the procedure.
3. Discretion should be exercised—select only those parts of a process for which automation solidly applies, and then applying the proper type of technology. In other words, this is no time to be a proving ground for new technology. Many well-intentioned attempts at automating an entire process have failed miserably due to managers rushing to be the first to use their environments as a test bed for advanced but unproven techniques. This same automation may work in small parts of a process but not the process in its entirety.

You should fully understand the type and scope of automation being planned. Automation types usually come in the form of hardware, software, or some combination of the two. For example, some large-volume, high-speed laser printers use both hardware and software to automate the functions of loading, stacking, separating, bursting, and decollating output. Similarly, tape library systems can be automated to load, mount, dismount, and store tape cartridges.

In both of these examples, both the type and scope of automation come into play. Automation of one centralized, high-speed laser printer may be appropriate for an output process that is well designed, homegrown, and highly controlled. It may not be appropriate for smaller, diverse departmental printing due to cost, complexity, and maintainability. Similarly, an automated tape library system may not be appropriate in environments of low volume, high diversity, low centralization, and multiple platforms. The message here is to thoroughly evaluate the type and scope of technology of the automation tools you might choose to use for each particular part of a process.

Understanding the need to standardize, streamline, and evaluate before you automate is only part of the puzzle. The sequence of these four activities is extremely critical to the success of any automation initiative. Many shops I have worked with want to automate first, thinking it will be bring about immediate efficiencies, cost savings, and improved quality of services. In most instances the result is just the reverse. Instead of having a poorly designed manual process, you now end up with a poorly designed automated process. Greater inefficiencies, lower levels of quality, and eventually higher costs normally follow.

The only correct sequence of activities is as follows:

1. Standardize a well-designed process.
2. Streamline that process.
3. Evaluate thoroughly the relative fit of the automation technology to the process parts.
4. Automate the appropriate pieces with proper technology

One of my more literature-oriented clients put this guideline to rhyme to help his staff remember it:

• First standardize, streamline, and evaluate,

• Before you decide to automate.

Table 1 summarizes the best and the worst possible sequence scenarios for preparing for the automation of infrastructure processes.

Table 1 Sequence Scenarios for Process Automation

References

Schiesser, Rich, IT Systems Management, Prentice Hall, 2002