> you’d do your homework first and then come to the table with "35% of our [insert some key metric here] are caused by the lack of X functionality" Tangential,

> you’d do your homework first and then come to the table with "35% of our [insert some key metric here] are caused by the lack of X functionality" Tangential, how do you decide whether to collect a new metric? Use data-driven thinking? But how, you don't have data until you invest in building that collector.