Modern Middle Manager
Primarily my musings on the practical application of technology and management principles at a financial services company.
Executive's Guide to IT -- Chapter 7, IT Operations

Saturday, July 12, 2003  

IT operations are near and dear to my heart. No, really, they are. Operational issues, performed poorly, will destroy an IT department's reputation. What are operational issues? Areas such as the help desk, server/network/telecom management, e-mail, databases, backups and standard daily tasks...failures in any of these areas are a black eye. The benchmark is perfection, with good reason -- few, if any, companies in America today can function if the computer systems, and the applications they support, go down. This chapter in the Executive's Guide to Information Technology is about setting up IT operations to perform as efficiently and optimally as possible. To do that, the authors focus on standard procedures, IT staffing ratios and evaluating infrastructure investments.

The authors spend a lot of time discussing standard tasks -- how to identify them, how to proceduralize them, how to document and test them. I've found a lot of value in doing this for routine tasks such as setting up a new server, creating a new desktop PC, performing backups, installing programs, etc. My favorite technique is to have the person assigned the task write up the procedure and then test it by having someone else perform it. If he/she botches it, rewrite the procedure based on the gaps found in the test. Then test with another staff member. We've cleaned up several of our tasks this way while adding redundancy to the (usually mind-numbing) task. Daily tasks I rotate between all of my department members so that they all know how to perform backups, split files to imaging, etc. These are tasks that also become candidates for automation through scripting. I'll save that for the review of Chapter 8, Application Management.

IT staffing ratios based on a recent Gartner group study are discussed as well. The study tried to come up with a rough idea of what staffing ratio is needed to support end-users. I used that study about a year ago to finally hire a help desk/desktop support person. Because of that, I was able to work with him specifically on desktop support issues and root cause analysis (another big point made in the staffing section) to reduce help desk calls by about 50%. What I found when we were understaffed by that one position is that symptoms were treated but the underlying problems weren't. Hence we'd have about 400 calls in a month, whereas now we're barely over 200 and dropping. A harried, understaffed IT operations function maybe be cheaper on paper than my staffing levels, but the level of service will decline, perhaps catastrophically, in the long run.

Speaking of staffing levels, my department consists of two programmers, one desktop support person, one network engineer and one systems engineer. Despite the specializations, there's a certain overlap to everyone's job -- for example, I expect everyone to spend time at the help desk covering for someone who is on vacation or out in the field. This model supports a 100 person, $22 million a year financial services company with about $2 billion in fiduciary assets under management. I get the expertise in areas the company needs (servers, networking, programming) while still covering basic tasks well.

Evaluating investments in the IT infrastructure is not done very formally. We do not have an IT steering committee. Because of this, I have one hurdle to clear to get the money I need every year for infrastructure investments. Those decisions are made in the fourth quarter during our budget process. It is at that point that the projects are prioritized, assigned a budget and, if necessary, approved by both the COO and CEO. I usually just have to clear the COO. The authors provide a list of potential benefits that should be identified and quantified before moving forward with a project. Without any coordinated strategy, my IT operational goals and budget priorities are simple -- cut costs and reduce downtime. Hence most of our budget is spent on projects to increase availability and reliability, increase capacity to meet projected needs and reduce costs/increase service levels (such as outsourcing our imaging system). Projects considered outside the normal IT project/budget process usually don't have a prayer unless they have a payback within 12 months.

As the authors say, a well-run IT operations function will result in positive perceptions and good end-user satisfaction. The implied opposite is much pain and suffering for the IT director and a hostile relationship between end-users and the IT department.

posted by Henry Jenkins | 7/12/2003 07:28:00 PM

Comments: Post a Comment
the author
open source