-
Notifications
You must be signed in to change notification settings - Fork 0
Productivity metric systems
We have 3 metrics of productivity systems depending on speaker, project manager or developer roles.
Developer productivity is related with the number of history points (HP) they has realised on time. Actually, a developer can do less HP than the PM has assigned to them if he is working less than expected ─ or them could do more HP than the PM assigned to them if they helped in a task they was not assigned or has complete a task that was for the next week instead of current week.
We use then (HP accomplished on time)/(HP assigned for this time).
As a task could be miscalculated by the PM, we agree in a meeting if those tasks that was not accomplished on time were because the developer didn't work or the PM failed. In the second case, this task doesn't count.
Expected values are between 0.5 and 1.5, being 1 the ideal value, less than 0,7 bad values and more than 1 very good values. In case than all tasks of a developer were miscalculations, the productivity is 0/0, so we have no information this week about this developer. We will use then last week productivity for him.
The productivity of the project manager is voted by the developers. They should vote between 0 and 10, depending on task they had assigned had a good approximation in HP with the real time they worked (1 HP half an hour).
The productivity of the speaker is voted by his public. They should vote between 0 and 10, depending on how interesting and informative was the presentation.
Speaker and PM roles have different productivity scale than developers. As we want anonymize productivity, we should convert to the same scale.
We are using this:
- From 0 to 5 -> lineal from 0 to 0.4
- From 5 to 8 -> lineal from 0.4 to 1
- From 8 to 10 -> lineal from 1 to 1.5
For example, a 4 is 0.32, a 7.4 is 0.88, and a 8.5 is 1.125.