Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I am wondering how all you great admins out there rate your success. I am thinking not only in a personal sense but also how you rate your employees and/or department. How do you build a business case for additional staff, ROI, new equipment, and justify your yearly expenses?

For example do you, and how do you track:

  • Trouble ticket solve rates
  • Project Success Factors
  • Project Success Rate
  • Customer Service Score (how much do your clients like working with you)
  • ROI for new employees
  • ROI for tools
share|improve this question
    
Excellent question! –  Joseph Kern Jun 15 '09 at 23:11
    
It's ROI ... not RIO ... :-) –  Joseph Kern Jun 15 '09 at 23:42
    
ROI RIO LOL Sometimes I should proof read what I post :-) –  JJ01 Jun 17 '09 at 22:08

8 Answers 8

up vote 7 down vote accepted

Heh. Mine are a little off the wall. More of how I gauge my own success :)

  • How well everything is running. Meaning, how much downtime do I incur due to unplanned outages, equipment failures, etc. Sure there's such a thing as a chip dying in a Cisco switch, or a random power supply failing in something else, etc. Force-majeure sort of things, but when I'm at the top of my game, my network should be running like a top. This leads to the second point:

  • How much preventative maintenance has been accomplished. This means basically, `If I've had time to lean then I've had time to clean.' Are all the logfiles checked? Trending has been done? Have those pesky non-critical disk errors that always pop up before a RAID array dies been investigated and checked? Am I up-to-date on making sure all applied patches for the past month are doing OK? Essentially I'm making sure that I'm being proactive with my gear. Trending and predicting points of failure is a good way to make sure everything is running 5x5.

Hmm. What else.

  • Have I `trained up' my subordinates? Meaning, if by the name of all that is holy, ALL the Exchange servers fail, can the other network admin, or the junior handle bringing email back up without a tremendous amount of drama? And so on, and so forth.

  • if I get hit by a meteorite, (hey, before you laugh you should read the news about that German kid who got tagged by one the other day! :) could my team pick up from my notes and documentation and feasibly get by, while interviewing my replacement? Meaning in a quite basic sense; is all my documentation up-to-date? Do I have my boss's trust, comfort of mine and knowledge that if need be, I could take emergency time off for a week or two if need be?

I think you get the gist of where this is going.

To summarize, if my workday consists of training co-workers and subordinates, passing on knowledge, doing a daily log check and focusing PRIMARILY on strategic IT initiatives such as security, tools to make the user's life easier and making the network run.. Oh, I dunno, faster or better or something, then I am "doing well", I suppose.

Eh, kinda rambling, but I warned you in the beginning. :)

share|improve this answer

One way to view it is how many users don't have to think about anything IT to get their jobs done.

share|improve this answer

From a user perspective, the system administrator is good if they don't have any IT issues prohibiting their productivity.

From a management perspective, the system administrator is good if they're saving the company more than they cost (ie, salary + benefits).

Both of these are hard to quantify, so you need to ask whoever you report to what kind of metrics they measure your performance against. If they're not willing to disclose that, it might be time to brush up the resumé. That's not a snarky statement; you don't want to work somewhere they don't tell you their expectations, right?

share|improve this answer

My success as an administrator? How beautiful my systems are. I cannot measure it, and it is strictly subjective.

I know, this is answer is not useful to you. I know my boss needs to care about ROI, additional staff, budget, expense cuts, etc, but I try to forget about these as much as possible. Yes, I help my boss, I understand his problems, but I don't really care about his problems internally. I know ROI cannot make me a successful administrator.

share|improve this answer

I'd track:

  • how vertical you are
  • number of employees supported
  • type of those employees

  • number of trouble tickets filed
  • time to trouble ticket resolution

  • how vertical your market is(how much you have to accept whatever software is presented)

  • By how vertical you are, I am asking if you're a jack-of-all-trades, if it blinks you're responsible, or are you exclusively a Linux sysadmin supporting this particular application. Usually you'll find this kind of admin in a large company, where there's room for several of this kind of admin.
  • By "type of employees", I mean are they Finance people, or someone technically skilled, like developers.
  • The reason that it matters how vertical your market is: If you work in a highly specialized field, you will have very little choice what software you use, and won't have the ability to move to another piece of software if there is a bug found.


None of these factors by themselves mean much; you have to take all factors into account to get any kind of meaningful result.

share|improve this answer

Since I primarily work in hosting and web applications, my key success metrics are:

  • Downtime: Are the applications available as much as they could be? How much does downtime cost vs. the cost of new equipment to reduce it.

  • Performance: How fast are pages being returned? Are there things that can be done on the server/network to improve response times?

  • Customer Satisfaction: How satisfied are the customers? If they're not happy, they won't stick around long. This is usually a direct result of the first two metrics.

For future IT purchases, we measure dozens of metrics on the servers and network gear that include CPU, memory, and bandwidth use at peak, and averages over time. We use trending tools like MRTG and Cacti to spot trends and extrapolate out to predict when we will need to upgrade or deploy new equipment.

Generally speaking though, if my inbox is empty, my phone isn't ringing, and I have some time to answer questions on Server Fault, then I'm doing something right.

share|improve this answer
    
Justin Could you take these three items a little further. What specific metrics do you use to measure Downtime, Performance, and Customer Satisfaction ? –  JJ01 Jun 17 '09 at 22:12
    
It really depends on your environment. For us, downtime is 'can we get to the Internet from our office?' and 'can customers get to the web sites that we host?'. In a simplified manner, if either of those two answers is no, then we have downtime. As for customer satisfaction, it's hard to measure specifically, but if their sites are up and running fast, they're usually satisfied. –  Justin Scott Jun 18 '09 at 19:37

Measurable: Do I comply well within the SLA.

Subjective: Do my colleagues still like me.

Personal: Do I still give a %. ;-)

% Insert phrase of your own liking

share|improve this answer
    
I like the "personal" remark. ;-) –  Cypher Sep 11 '10 at 21:49

Careful with the metrics, if you use any metric as an absolute measurement you will became a metric meeter, not a good admin.

i.e. if your (or your boss') metric is only how many tickets you close how fast, you can very easily become a drone that only shoves the tickets around without investing in long term, quality solutions and preemptive maintenance.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.