I really like that paper's emphasis on overall throughput. KPIs used this way are effective. Measure what matters.
Too often you get in a game of setting up metrics silos for everyone to operate independently, and there are non-linear, emergent effects to this silo behavior.
You have to keep you head up and pay attention to how the system is actually behaving. POSIWID.
> make sure at a minimum that the measures are not encouraging local optimums
I really like that paper's emphasis on overall throughput. KPIs used this way are effective. Measure what matters.
Too often you get in a game of setting up metrics silos for everyone to operate independently, and there are non-linear, emergent effects to this silo behavior.
You have to keep you head up and pay attention to how the system is actually behaving. POSIWID.
> make sure at a minimum that the measures are not encouraging local optimums
Pretty much sums it up for me.