Showing results for 
Search instead for 
Did you mean: 
Create Post

Big brother is watching you

Level 11

Management wants to be able to track employees productivity and their performance to use in periodic employee evaluations. Performance improvements can lead to keeping your job, bonuses, and pay increases, where declines in performance lead to pay decreases, black marks in your HR file, and possibly losing your job. When accurate metrics are used in evaluations, they can be beneficial to the organization and individual.

One type of metric is the Service Level Agreement, or SLA. This defines how IT, as a group, is going to respond to the issue that a user reports. Some organizations define the SLA to be the time it takes to resolve the issue. While some issues can be resolved quickly, like a password change, other issues can take a long time to diagnose and correct. The SLA can also be measured by how long it takes for the initial response to the request. While I believe this is a better metric than time to resolution, this can be abused. The person assigned the ticket can easily make initial contact but not make any progress in diagnosing or resolving the issue.

Another metric is customer satisfaction. One way of getting this metric is using surveys sent to the requestor after the issue is resolved. Not all surveys make it to their destination and if they do, many get ignored or deleted. The questions on surveys are written in multiple choice for easy analysis, but don’t really provide much for collecting real feedback from the requestor. If an issue was handled by multiple people, then who does the survey reflect upon and does the requestor realize this.

Managers want to know how well their employees are doing and a way to accurately measure the employee. Ticketing systems have some metrics that can be used to track how well employees are doing. How do you accurately measure employees, in particular Help Desk employees?  What are the metrics that really matter? Can all of these metrics be tracked in one system? How would you like to be measured?

Level 13

In theory, customer satisfaction would be the truest measure of performance. That said, there are areas that are going to be unfairly hurt by that (the NOC, for example, who might only be engaged during a major outage), and areas where such a metric might create a conflict of interest (security, where a denial of a request is almost certain to result in a poor survey). Those types of areas likely need to be measured differently, or against different thresholds.

Level 17

Specifically those chained to the desk have easy metrics to make sure they are at least working. Then others to determine possibly how well you are doing throughout and as an entirety of the call.

With that said, there is nothing better than someone in a training capacity who can listed to calls and coach your HD employee's. It's one thing to watch an old 80's video on phone and service etiquette but having someone to walk you through your own real life examples can really shed light on someone's demeanor and tone when on the phone.

Someone who can help your folks understand process related to empathy and a need to direct or 'run' the call while letting the caller feel like they are in control + a technical background is a hard combination some times... and more often you want THAT Guy on the phone.

Metrics start with # of calls - Broken down to calls/hour

From that you get a % logged => # of tickets generated/# of total calls

   make sure they at least log the calls. If you are not logging them all you will miss an important call.

Next you take the % Closed (or first call resolution issues taken) =>  # of tickets closed / # of total tickets

Track the time of the calls - so you can calculate an Average Time per Call (there should be software doing this)

Depending on your information gathering, and possible "logged In" or variations of this status you may have.

total Time logged shown / day => to calculate Avg Time Logged In (Active) / Day

If After Call Status (to close tickets), and Aux Mode (helping someone else with a call, Technical Issues, Etc) you can also pull the Average time they were in "After Call" status. Also AVG Time in Aux Mode / Day


So you get;

How many calls

AVG Time Per Call

Avg Time After Call (Finishing up ticket/routing/details) - I like this as an average so total AFTER CALL TIME / # of Calls for the average per call.

AVG Aux Time (per hour/day/week, depending on how your metrics show)

# of tickets

% logged

% Closed (first call resolve)

And of course the Survey and Customer feedback. The ticket staying closed is sometimes a good one. Constant Re Opened tickets may be an indication of training - either the skill set or the Ticketing system.

I know there are more I am missing but that should supply basics to start with for anyone trying to build this in.

Level 16

This is a touchy subject in many of the places I have worked.  So many of these metrics are dependent on other items.  For example, the tech that takes the more difficult issues will have less tickets closed and take longer to complete them, but that tech is very valuable due to the type of ticket they close.  Hours worked or logged in is rough because time is not necessarily productive time.

I think it comes down to trusting and relying on your team leads, co-workers and managers.  Statistics are great, but humans should be judged based on many items and not just comparative graphs.

I've seen anonymous survey's concerning the employee be very good.  You find out all sorts of good things when people are anonymous.


SLA'a are 4 letter words in some shops.  If set up properly and for the right things they actually help you. 

I have also seen some overly tight ones that are designed to make you fail.

Regarding the customer statisfaction...I find that a survey as part of the ticket process really helps that part along.

There you can get some honest feedback on the work performed...sometimes a better gauge than raw numbers...

Level 13

I don't see much in the comments about metrics above that apply to some of what I do.  System monitoring, reviewing logs, improving systems, design.  Those are things not measured by help desk tickets and the reality is, these things make up the majority of my job.  Don't get me wrong, I do get ticket, if they apply to systems I work with, but those don't come in a volume to fill my time, and I am expected to do other things, that may not have a good measure.  Or would have a bad measure if something fails.

Today I was asked to help review an RFP for a new product/service and look to see if it is complete.  Even my supervisor and CIO agree this is uncharted waters in technology for us, so we are not sure if it is complete.  How do you measure that for my employee evaluation?  Is it just success or failure?

Level 15

I am in a similar situation whereby I do not receive a lot of NOC tickets but only the occasional ticket.  The bulk of my time is spent working on new projects and technology within my designated assignments.  Somewhere in the neighborhood of 50-60 systems, applications, and hardware projects.  We value time spent on projects both in the hours applied to said project as well as the success of the project in terms of budget and on-time.  Even if a project is termed before completion or I am assigned to fact finding missions, those are recorded in the project log.  Ultimately, this is reviewed as part of the evaluation.

I have not found a good formula or company that performs employee evaulations in an absolutely objective manner.  It is always subject to the whims of the manager.

Level 15

Tracking people's social networks (within industry and enterprise social network) is turning out to be the best way to find the star employees. These "connectors" or "hubs" are the ones highly connected between teams and often the go-to person for many people whether it's just to reconnect the requestor with another colleague or network.

Level 11

Any time you deny the requestor what they are asking for will result in negative feedback. How do you compensate for that?

Level 11

So how do you determine from these statistics that one employee is doing an excellent job while another is doing poor.

Level 11

Do you get more surveys completed if they are anonymous?

Level 11

I understand that SLAs are to give the business a timeframe for a response or resolution, but I agree that many times they are setup for failure.

In regard to the SLAs you have seen, which are typically better? Ones that are based on response or resolution? What time frames?

Level 11

When I was a network administrator, most of my work was project related. I would brought in on tickets that were fairly complex or the problem was elusive. My annual evaluations were completely up to the perception of my manager.

You could get anonymous surveys from your peers, but that depends on how you interact with them. One of the perceptions I have noticed over the years is that the most recent events, especially bad ones, tend to influence feedback. You could have done an excellent job over the entire year, but then that problem you didn't resolve right away last week because of a number of other legitimate issues puts a black mark on you.

Level 11

I like the idea of tracking all projects and tasks, then review them at the end of the period. But like you said it is at the whim of the manager. Perhaps if along side the list surveys can be included. What about surveys sent to those involved in the project after it has been completed? That would get more accurate feedback since it is closer to the project. Those surveys then be reviewed along with the list of projects and tasks to lessen the whims?

Level 11

What social networks are you referring to? Ones like Twitter & Facebook, or Thwack & support forums?

Level 15

Both traditional social networks and also enterprise collaboration tools

Level 13

That was my point; that those types of areas must be measured by metrics other than customer satisfaction survey feedback. For those types of areas, I would look at how performance is measured in areas like internal audit and legal for guidance.

Level 13

Speaking for myself, I prefer SLAs based on response. While there is the potential for abuse (responding quickly to the ticket and then ignoring it from there), that type of abuse usually gets uncovered pretty quickly in the user/customer satisfaction surveys.

Level 15

I like that thought but unfortunately I have worked for far too many managers that hold equal weight to his subjective review versus the objective review of the stakeholders.  Even though, everything progressed forward and the stakeholders are happy with the result.  If the manager did not like you, you never got a fair evaluation.  I have chalked it up to the frailties of man. 

Level 17

Use them together. You always have the 'Type of call' Issue - it may be someone hanging up, or needing to call back later. Others will take days if you let them. The Help/Service Desk is also a malleable institution depending on your industry. Some are going to be scripted, and other a little more free form. Free form allowing more ability to help and get a 1st level resolution. A more script based, or 'Strict' setup will provide you with more details when using these metrics. But in a health care environment - dealing  with Doctors and Patients - you have to be able to adjust, and any place where going off script is an issue could create service issues. For that you may rely more on your metrics as a whole or the average - rather than just trying to meet specific marks.

Take your metrics over time to come up with an average of what the team does.

    From there you can see your high end performers, and low end. Of course use of a trainer or someone to review some calls and work with your analyst will improve your metrics,

    Trying to determine who lingers the most after calls isn't just the #'s, it may be that it takes them longer to type, or maybe the put in more details into the ticket that helps the next level of support.

Level 14

I tend to lean towards customer satisfaction as long as their trouble tickets closed rate is not too low. 

Level 11

I agree that SLAs need to be based on response. What timeframe do you think there should be on that response? And how often?

I have opened cases with some manufacturers and I get a response quickly. Typically they ask for information I have already provided or I reply back immediately with the information they requested and then I don't hear back for several days. In these situations I typically give negative feedback on the communication portion of the surveys, but it is very frustrating when I am trying to resolve a problem and I don't get very good service from the support engineers.

Level 11

I left my last $dayjob because of my manager. He was always very negative and critical of the people that worked for him and gave praises to people on other teams.

About the Author
I am a network and UC engineer for a mainly Cisco reseller. I have worked in the networking industry for about 14 years and started as a network administrator for a small CLEC (carrier) where I did it all in IT and worked on the carrier network. After the CLEC, I went to work for a large healthcare organization in the Houston area and stayed with them for about three and a half years. Now I work for a reseller in the professional services part of the organization. I am currently studying for the CCIE in Routing and Switching lab. You can find me on the Twitter @twidfeki or I blog on Packet Pushers and