# How can I calculate performance evaluation criteria for a software team?

I am looking for a theoretical formula which can evaluate the performance of an employee. I have figured out these measurable factors on which the performance can be calculated. Can anyone tell me a feasible formula for this?

• Requests completed
• Number of bugs
• Amount of appreciation
• Number of customers assisted
• Behavioral Quotient factor (or any other measurable factors)

There are basically 4 categories, or basically the percentage for each category in which the value of the formula should lie:

• worst (25%),
• average (50%),
• good (75%),
• outstanding (100%).

Looking forward to the magic formula!

• Could you revise the question to emphasize the practical project management elements as described in How to Ask? I'm not sure the question as currently phrased is about project management.
– MCW
Oct 27 '15 at 10:44
• Yeah, this is people management as it stands. And terrible management at that. Oct 27 '15 at 11:54
• check the related questions in the right side : One of the question that might be helpful : - pm.stackexchange.com/questions/4088/… Oct 27 '15 at 13:18

Your developers need to use ability, judgement, intellegence and nuiance to do their job well.

The same is expected of you when doing your job. There is no formula and reductionist metrics are a sign of 'management smell'.

You need to both get better at your job (the how is complex) and also defer judgement of what is good work to the team, rather than to a spreadsheet.

## TL; DR

You are probably looking for a weighted average, but you need to do a lot of work to determine how important each component of your scoring system really is. Your current metrics do not appear to contain enough context or sufficiently-defined management goals to be useful without additional work.

## Analysis

The problem with any set of metrics is that you have to carefully define your context and provide meaningful goals. When you have imprecise metrics, you run the risk of measuring the wrong things or incentivizing the wrong behavior.

As an example, you want to measure "number of customers assisted." What if the value is 1?

• Is that because the person was unhelpful?
• Is it because you had very few customers that needed help?
• Are you trying to prioritize customer assistance over new features?
• Are you measuring this dimension against the rest of the team (i.e. who handled the most of the 27 tickets this period) or against some management target (e.g. Joe should handle 1/5 of all customer calls)?

There are many other contextual and analytical issues beyond these, but the preceding at least gives a good flavor for the sorts of questions you and the management team should be asking yourselves. There isn't a "one size fits all" formula because metrics require constant tuning and interpretation to be useful.