As part of the Performance Review for my team member I would like to setup some indicator so quaterly I can see if we are in the right direction or not. However it's difficult for me to decide what to put there and how to measure.

I'm in charge of a department and our main goal is to support other departments. Let's say we are the techies this means that our main goal is to provide with technical solutions and help people. Most of the solutions we have to provide are long term and this makes difficult for me to come up with a way to setup. For example we need to implement a tool to automate testing. The indicator is weahter we implement it or not, But quaterly indicators ... I'm not sure how to track them.
So any idea?? How you track these indicators when the goals are long-term? What can I track?Thank you for your help

WillDuke's picture

Need more details on your business. Why are you implementing testing? If the testing is positive, is that a reflection of the service you provide? If so, can you measure your service based on the test results?

What happens to the other departments if you support them poorly? Conversely, can you measure what happens if you support them well?

Basically, it sounds like time for a review of job descriptions and department goals. You need some things to count.

bflynn's picture

One idea that comes to my mind is that you're asking to measure the "speed" of an employee. Normally speed is Units Accomplished per time. If we say that time is constant (and within the context of projects, that is sufficient), we can drop time from the equation and just measure Units Accomplished. Because most IT jobs are different, its difficult to measure different jobs directly.

For the Units Accomplished measurement, I would suggest you try to measure a ratio of Completed hours vs Actual hours. A concrete example might be the easiest way to understand what is going on. You believe part X of a project will take 160 hours. Your direct completes it in 3 weeks. The amount of work they completed is 160 - this is how much comes off the project plan. The actual time spent is 120 hours. Their ratio is 160/120 = 4/3 or 1.3. Ratios above one are good. Ratios below one are not as good.

This is just one idea and it might not be relevant to your example. Be on the lookout for employees gaming the system by setting high expectations (they're going to try to game it any time it affects their promotions or compensation.) As long as you treat KPIs as indicators and not hard science, it should be ok.

Whatever you choose - don't let this substitute for your management of your direct. These mere numbers are no substitute for people.


jhack's picture

Customer satisfaction is a good metric. It can be hard to measure, but without more details on what, exactly, your team does, I can't provide other ideas.

Check out the podcasts on "Better Internal Relations" from the end of 2006 to get some ideas on how you can find out what matters to your customers. Then figure out a way to poll them quarterly on whether you're meeting their needs.

Ideally, everything we do should ultimately generate more revenue, improve productivity, or reduce costs. If you can measure any of those, then you're in a very powerful place.

mirror42's picture

Provided that the new solutions your department delivers is managed by some form of project management (divided into tasks) you could be thinking of using KPIs for managing projects. For example to check if tasks assigned to individual team members are delivered on schedule within the set time effort, etc. Even if the tasks at hand are more research driven, I would suggest to define some level of project or task management.

Perhaps this site can provide some ideas of KPIs that you could be using:


bflynn's picture

John's post made me think of the negative that I didn't mention. What you measure is what you get. If you're measuring speed, your employees will know - they're not that dumb. And they will be faster. Quality will suffer unless you also have a KPI to measure that.

Thus, the problem with "key" performance indicators. Our jobs cannot be boiled down to just a few key things.

What you measure is what you get....


rwwh's picture

[quote="bflynn"]What you measure is what you get....[/quote]

This is so true, and it makes it so difficult to define KPI's (and incentive schemes).

I have the assignment now of defining KPI's for my R&D department.

Some examples of suggestions people have made:
* Measure the output in new patents. Risk: one could get more patents that do not pay off. We normally patent only to fence out our competition, which is a fraction of our inventions.
* Measure whether projects are in time. Risk: engineers may start to predict longer running times beforehand to make sure they are in time afterwards. Or people may start to prefer safe projects without any research risk. Not good for an innovative company. And people will complain if project priorities shift over time, making projects late without R&D having any influence on them.

Also, because of the long running time of R&D projects, statistical fluctuations from year to year are huge. This further complicates goal setting.

I have a PhD thesis from Ingrid Kerssens-van Drongelen. She made a systematic study of R&D performance measurement systems. And although one conclusion is that it is worth the efforts to measure R&D performance, the book also indicates that it is very difficult. And the examples she gives are from companies that have thousands of R&D personnel.

My fears came through when browsing to the KPI library that mirror42 refers us to above: no R&D KPI's in there either.