Team performance/productivity metrics

[b2au]

Hi all,

I am new to this forum so let me begin by saying Hello.

I have recently started my 4-month co-op placement and my objectivite is to create a scorecard/metrics to measure our team's perforamnce and productivity.

Here is a little background about our team. I am currently working in Technical Services, a department that does the IT infrastructure suppot. So the purpose of the department is to provide back end support to all our IT systems. From my understanding, we have been using ITIL for quite some time and it has matured. The service desk has created their own metrics in measuing their performance.

However, in the technical services, there is no real scorecard to show our contribution. I was hoping if there is anyone who is in a similar department who can share their experiences of the metrics that are used to measure the performance/productivity?

Although the primary objective is to measure the performance/productivity to show the value of the department, the secondary objective is to raise some KPIs to show how our department contributes value to the business. Currently, the department is not really visable to the business because the business does not deal directly with us as oppose to the service desk/PMO/business solution managers.

Your input will be much appreciated and thanks in advance for helping me grow for this work term.

Regards,
Barry

[UKVIKING]

What does the team do ?

Does it do system support ?
If so, it is involved in Incident and probably Problem management

If it does application support
The above plus release mgmt

If it does the above, then they get involved in change mgmt

So you need to differentiate the metrics/KPIs to the ITIL areas

Call being when your team gets involved
First incident

# of call per period
average time to close a call
average time to respond to call
# of major incident
% against OLA

OLA - Operations Level Agreement

An agreement between your team and the service desk/business on how quick you respond to issue

Use the historical data time wise to gauge your OLA time

remember OLA are not SLAs
_________________
John Hardesty
ITSM Manager's Certificate (Red Badge)

Change Management is POWER & CONTROL. /....evil laughter

[b2au]

thanks.

to answer some of those questions.

our team deals with system support and only critical incidents/problems will be communicated by phone/walk-in.
otherwise it is just assigned to the queue

i'm trying to prepare about 7-8 KPIs for management viewing. it will be presented in excel charts and printed on 11x17


at the moment, i only have the following two KPIs

-avg time spent on incident/problems
-mean time to repair

any more suggestions i should look into?

Thanks.

[Fabien]

If I may suggest. I'd like to push a little further than John's input because, to some extent, I don't need to know what you do to measure your performance. Instead I would focus on what you have agreed to deliver. The end is more important than the means if you will. You see, I understand that you have already defined the number of KPI's you want to track, but I wonder how you will choose the right ones, and how they will demonstrate anything if you do not start with the end in mind.

In order to function, you need objectives. For instance, you can have for objective to deliver timely support to customers.

To measure yourself against those objectives, you need to agree on a set of Critical Success Factors. You will need to agree on what it means to deliver timely support. In this case for instance, delivering timely support may mean that to be successful, you agree to deal with phone/walk-in in priority within 'n' hours, but at the same time, you shouldn't leave the queue growing beyond 'n' hours of work.

Once you have agreed on those factors, you can then agree the KPI's that will demonstrate your achievement against those critical success factors. Is it mean time to repair for phone/walk-in, but supported by number of rings to pick up the phone?... And what are acceptable levels for each of those KPIs? What are the thresholds that should trigger an alarm in your office?

... things like that....

Keep the end in mind when you measure, and agree on the meaning of the metrics in advance, otherwise you will measure a whole bunch of things that don't mean anything and you won't gain the credibility that metrics should bring you.
_________________
BR,
Fabien Papleux

Accenture
Technology Consulting | Service Excellence
Red Badge Certified

Twitter @itilgeek

[b2au]

thanks for the great input Fabien!

i know i haven't yet clearly defined the CSFs yet, and am in the processes of finding out.

being here for just one week, i have scheduled meetings with different team leads and the manager so that I can have a better idea of what they do in their area, OLAs, and CSFs.

your suggestion just made me feel like I'm somewhat on track, and hopefully when I find out more about the goal, I can get some more pointers from you and the rest.

Thanks again!
Barry

[Fabien]

[b2au]

hi again,

warning: kinda long read. sorry about that.

i just like to share some findings so that you guys can possibly give me some tips as to what I should do.

objective: to create a simple dashboard (done in excel) displaying a few (about KPI to allow the manager and team to review their performance and productivity.

department: Technical Services (IT system support)

What we do/CSFs?

1. Production support
CSF: to keep all production system available 99.9% of the time. Schedule/planned outage does not affect the measurement.

2. Project support
CSF: to provide system support to the PMO to ensure sucessful implementation

3. Baseline activities:
- system maintenance
- checking thru logs
- R&D
CSF: continuous improvement in the service and support in the Technical Services department


What my manager is looking for?

- measurements (monthly) that will be used by the team and the manager. the purpose is to address continuous improvement. if there are any problems/concern, it will be looked after.

- he is not intereseted in measuring OLA, because the Service desk should measure that to make sure were on track. as far as i know, our team has been doing very well in the past and is able to keep availability 99.9% and meeting their OLA. however there is no measurement to prove that.

- my manager is more interested in measuring how much time we spend on production, project, and baseline activities. proactive vs reactive. the time it takes to solve incidents/problems and anything along those lines. (at least that is his vision)

so far on my drawing board i have:

1. %time spent on high value vs low value activities
high value = production support, project work, baseline activities (what were suppose to do)
low value = service request, getting clarification for incidents, incidents that are suppose to be handled by service desk, etc.

2. time spent on production, project, and baseline

3. Mean time to resolve and respond.
resolve = the time it takes to close an incident from when it gets created.
respond = the time it takes to pick up the incident from queue

and thats all I have... I'm not sure how to address the proactive vs reactive issue.

any ideas is appreciated. sorry for the long read.

Regards.

[Ed]

Hi Barry

Simple silly question

Have you actually agreed a production support OLA of 99.9% or are you planning to?

For those out there this means that on a 24/7 shop you can only have a outage time of 8 hours and 46 minutes in a year, on a 8x5 shop this translates to 2 hours and 4 minutes - ouch, ouch, ouch!

Regards

Ed

[UKVIKING]

Ed made a good point

There should not be a OLA KPI really because OLAs are 'guidelines' not hard and fast like SLAs.....

I would change the measurement abit so that the OLAs between the department get measured a little differently....

Ed also made the point about 99.99% Availabilty.

This needs to be expressed in the SLA times for the service...

If the service is a 24x7 service, then you need to calculate how much unavailabilty would go over the mark

Look to the service availability rather than device

for ex

If you have 3 web servers in a load balancing situation and 1 goes down, the web servers still have 100% availability because the other 2 are taking the load.

Only when all three are down does the availability go down

I would also have a load KPI on a 'web server' which should change when the # of servers change.

A Dashboard for reminds me of Statistics and the old saw

lies, damn lies and statistics
_________________
John Hardesty
ITSM Manager's Certificate (Red Badge)

Change Management is POWER & CONTROL. /....evil laughter

[b2au]

you guys are right.

what i meant was the OLA is to aim for service availability of 99.9%. this excludes all planned/scheduled outages.

however, i will not be measuring OLA, but possibly availability if I can't brainstorm other good metrics...

keep the ideas coming...and thanks.

[insider]

"is able to keep availability 99.9% and meeting their OLA. however there is no measurement to prove that. "

How do you know availability is at 99.9% then??? A perception of good service is different to reality.

As long as service outages are measured- and service times are clear, thenavailability calcualtions are easy to make......

[Fabien]

I would spend a bit of time narrowing down the high-level objectives to be as specific as possible. Creating a one-page simple dashboard to report on performance across all those activities is a huge task, and I don't see how that page can be easy to read.

A few thoughts.

- Think about balancing your metrics. For instance, if one of your objectives would have been to cut cost, you should measure against service level, because reaching one goal shouldn't be detrimental to the quality of service.

- In this case, for instance, how about reporting on production support directly against cost? It's not so much that 99.9% is difficult to attain, it's that from 99.5 to 99.9, you could double the cost of support/infrastructure.

- Providing system support to the PMO seems to me like an objective, not a CSF. A CSF could be: Respond to PM requests within 15 minutes for project1, project2 and project3. I also assume not all project require your undivided attention.

- Your baseline activity CSF is enormous. How do you measure specific improvements across all your operational responsibilities and report that on a third of a page?

Hope this helps you find the answer...
_________________
BR,
Fabien Papleux

Accenture
Technology Consulting | Service Excellence
Red Badge Certified

Twitter @itilgeek