Team Management by Metrics

Question

Metrics are valuable tools to measure performance and to manage projects/people. Nevertheless, I have started to encounter certain issues with regards to the measurement of my team's performance.

Members of the team have started to believe that "more is better", and they are trying to get involved in additional assignments. I was pleasantly surprised with this turn of events; the team is motivated and want more challenges to keep developing their skills and grow a reputation within the Company. Therefore, even though the tasks weren't fully completed successfully, the contribution was appreciated and noted as part of their development plans.

However, following up on several 1:1 reviews with certain members of the team, I realized that the real intention behind this was the number of assignments on which they have participated rather than the quality of their performance, using these metrics to request a salary increase.

How can I implement mechanisms to measure and balance Quality over Quantity on my Team's Performance Metrics?

How can we motivate people to increase their level of involvement in the projects taking into account the quality expectations?

Answer 1

The simple rule is: you get what you measure. If you measure a number of assignments you will see a lot of them being undertaken by people. If you look for a number of bugs you'll have virtually tons of them etc.

Actually I find extremely difficult to create metrics which work universally for different people in the team. What works much better is goals. If the team shares common goal, like completing a project on time and on scope you will see more self-organized positive behaviors than you would if you tried to introduce specific metrics to incentivize specific actions.

If you try to optimize specific part of the system, e.g. using single metric, you usually get suboptimal behavior of the whole system.

In short: think more about goals, which are shared by the team and let people organize in a way which brings them closer to achieve the goal instead of measuring, and optimizing, specific fragments of the whole process.

Another trick is to make the goal achievable so people actually believe they are able to achieve it, otherwise the result will be pretty much the opposite.

Answer 2

This is not just confined to the management of people but is a rather common issue with many things that are measured. The metrics you choose to use need to be well thought out such that you are not measuring the wrong thing, creating two measures that contradict each other, incenting behavior in the wrong direction, or fixing one area but breaking another. When I observe this happening, I find it was due to two things: metrics implemented too quickly to get the program going or not having enough flexibility in the program to alter the metrics as you learn.

So in your case, it sounds like you had one metric for quantity that did its job; it motivated folks to fill their plate with work. But you did not have a corresponding metric on the quality of the performance of the work. Start from the top--the goals and objectives of the organization over the next period--and redesign your metrics for alignment. Carefully look for inconsistencies and do some if-then scenarios to ensure you are not creating any unintended consequences. Benchmark your new metrics, then press go. Re-evaluate after a period of time to see if you erred in your metric design and be flexible enough to admit it and change it.

Answer 3

I don't have a specific answer, because the specifics need to be customized to your organization and perhaps even your project. One thing I do strongly believe in is that if they are necessary because of the political environment, at least make them team-level or even division-level metrics to ensure everybody is working to the shared goal rather than their own individual goals.

Also, before defining any goals, I highly recommend you read (at least) Influencer, as it provides an incredible six-influence model to understanding how to build effective incentives to create the desired result. The main goal is not to have any competing signals, as the opposition degrades or even cancels the effectiveness of the incentives. I'd also recommend reading through Predictably Irrational and perhaps The Upside of Irrationality. These two books provide many easily-understood experiments that point out the unexpected results of our attempts to coerce people through incentives.

Answer 4

You change what you measure (I think that's a quote from Tom DeMarco or List). If you want better quality, measure quality by the # of defects/issues found in QA or overall end user satisfaction. Be careful about rewarding individuals instead of teams because that's what you will end up with. I would recommend creating a simple list of 3 items to measure for reviews, define to a degree how you will measure and keep in place a combination of qualitative and quantitative measurements. Not everything can be measured with a ruler.

Answer 5

You need to create a conflict between metrics, in order to keep them productive (and non-destructive, like in your business case just explained).

Once you have a metric that motivates people to produce more you need to create a metric that de-motivates them to produce low quality. In every industry such metric pairs could be different.

Answer 6

This is always tricky because you can't think of all the repercussions of any measurement. I've found it's more about how you describe the metric than anything else. If you say we want to increase X then you will get an increase in X - perhaps to the detriment of Y. If you say we want to increase X and then see a related increase in Y, you narrow the possibility that someone is only going to focus on grinding out more X.