Tuesday, January 25, 2011

Performance Metrics at Work


I'm in the middle of reading Management 3.0, a book about Agile leadership and management. The book has been thought provoking, and I would like to think and write about dozens of different topics. The subject that I've been thinking about today is "Performance Metrics," measuring individual or group performance at work.

In my experience, Agile usually just focuses on team performance, e.g., the Balanced Scorecard method described in the book Succeeding with Agile. The author in Management 3.0 argues for measuring metrics at multiple levels of an organization including individuals, teams, and higher levels.

I've worked in the software industry for over 20 years, and I've only seen the use of reviews for individuals with annual corporate "performance reviews." It's an ritual I always despise, both as a manager having to do reviews and as a person being reviewed.  I despise the multiple choice "satisfactory / unsatisfactory" questions. Right now I'll get angry even trying to remember them.  I liked having of list annual objectives in theory, but they never worked in practice: 1) a year is a long time, and there was usually no review or follow-up during the year; 2) what I'd written at the beginning of the year was often irrelevant because the direction of the company had changed so radically in the intervening time. I would have liked to see reviews or metrics done at higher levels of the organization. The top management usually rolled out nonsense goals every year like my favorite from Spectrum Holobyte, "Ship hits on time."

It makes me wonder what a good individual performance review process would look like. I've heard of "360" reviews, which are mentioned in , Management 3.0 and that sounds like a good idea particularly if done in a group meeting. Among Management 3.0's 8 tips for measuring performance, I was most taken aback by

  • "Never create ratings yourself: The value of your opinion as a manager about the performance of a person or a team is very, very, very small. ..."

Upon reflection, I find I agree with this. Direct interaction with my manager is often been a lot less significant and frequent than my other work relationships: customers of my code and tools, departmental directors I'm supporting, etc.

Individual evaluation criteria and performance metrics I've seen have almost always been qualitative. I'm curious what useful quantitative ratings for individual software developers could look like, e.g., for programmers, producers, artists, designers. I've never worked in sales so "dollar sales" has never been relevant. I've been a fan of various agile team measurements: velocity, burn-down and burn-up charts, etc. Without measurement, I find both self-discipline and continuous improvement to be more challenging.

No comments:

Post a Comment